Cases in Primary Care

CASES IN
PRIMARY CARE
Best practice in clinical diagnosis, treatment, and management
July 2017
CASES IN PRIMARY CARE
The presentation of interesting cases has a long tradition as an educational tool and contributes to
lifelong learning. Discussing the clinical course and management of individual patients enhances our
understanding of disease and provides a framework for learning about medical advances.
The New England Journal of Medicine publishes several case-based series including Case Records
of the Massachusetts General Hospital, Clinical Problem-Solving, and Clinical Practice, plus
our online Interactive Medical Cases. We have chosen the cases in this collection based on their clinical
relevance to primary care.We hope you find this collection engaging and instructive.
Edward W. Campion, MD
Executive Editor and Online Editor
The New England Journal of Medicine
800.843.6356 | f: 781.891.1995 | nejmgroup@mms.org

860 winter street, waltham, ma 02451-1413
nejmgroup.org
Cases in Primary Care NEJM.org
Table of Contents
CASE RECORDS OF THE MASSACHUSETTS GENERAL HOSPITAL

5 Case 10-2017 A 6-Month-Old Boy with Gastrointestinal Bleeding and Abdominal Pain
L.M. Allister and Others
Mar 30, 2017
14 Case 2-2017 An 18-Year-Old Woman with Acute Liver Failure
K.R. Olson and Others
Jan 19, 2017
CLINICAL PROBLEM-SOLVING
26 Making the Connection
A.R. Schulman and Others
Feb 2, 2017
33 Back to the History
M.W. Montgomery and Others
May 4, 2017
CLINICAL PRACTICE
40 Screening for Colorectal Neoplasia
J.M. Inadomi
Jan 12, 2017
48 Heart Failure with Preserved Ejection Fraction
M.M. Redfield
Nov 10, 2016
INTERACTIVE MEDICAL CASES
59 Under Pressure
E.M. DeFilippis and Others
Mar 30, 2017
60 Dissecting a Case of Abdominal Pain
J. Casey and Others
Oct 27, 2016
2017 Copyright Massachusetts Medical Society. NEJM Group is a division of the Massachusetts Medical Society.
All rights reserved.
3
CASE RECORDS OF THE MASSACHUSETTS GENERAL HOSPITAL

The Journal has been publishing Case Records of the Massachusetts General Hospital since 1923. These reports
of clinicopathological conferences are one of the most popular medical teaching tools in the world. They describe
actual cases that expose the process of medical decision making, and range in focus from common conditions to
medical mysteries, exploring advances in challenging differential diagnosis and treatment.
Back to Table of Contents

The n e w e ng l a n d j o u r na l of m e dic i n e
Case Records of the Massachusetts General Hospital
Founded by Richard C. Cabot

Eric S. Rosenberg, M.D., Nancy Lee Harris, M.D., Editors
Virginia M. Pierce, M.D., David M. Dudzinski, M.D., Meridale V. Baggett, M.D.,
Dennis C. Sgroi, M.D., JoAnne O. Shepard, M.D., Associate Editors
Emily K. McDonald, Sally H. Ebeling, Production Editors
Case 10-2017: A 6-Month-Old Boy

with Gastrointestinal Bleeding
and Abdominal Pain
Lauren M. Allister, M.D., Ruth Lim, M.D., Allan M. Goldstein, M.D.,
and Jochen K. Lennerz, M.D.
Pr e sen tat ion of C a se
Dr. Akash Gupta (Medicine and Pediatrics): A 6-month-old boy was seen in the emer- From the Departments of Emergency
gency department of this hospital because of gastrointestinal bleeding and ab- Medicine (L.M.A.), Radiology (R.L.), Sur
gery (A.M.G.), and Pathology (J.K.L.),
dominal pain. Massachusetts General Hospital, and the
The patient had been in his usual state of health until 2 days before presenta- Departments of Emergency Medicine
tion, when his parents noted that he began to have intermittent episodes of ab- (L.M.A.), Radiology (R.L.), Surgery (A.M.G.),
and Pathology (J.K.L.), Harvard Medical
dominal pain. During these episodes, some of which woke the patient from sleep, School both in Boston.
he cried and pulled his legs up toward his chest while lying on his back. His
N Engl J Med 2017;376:1269-77.
parents reported that they palpated his abdomen during some of the episodes and DOI: 10.1056/NEJMcpc1616020
it felt rigid; they suspected that he might be having discomfort related to excessive Copyright 2017 Massachusetts Medical Society.
intestinal gas. He continued to eat and drink normally without vomiting. The next
day, the patient had two bowel movements, and the stools had reddish discolor-
ation. With the first bowel movement, the redness seemed to be present in a small
amount and only on the outside of the stool; with the second bowel movement,
the amount of redness increased. The patients mother attributed the stool discolor-
ation to beet consumption, since bowel movements with reddish stools had also
occurred in the past after the patient had eaten beets. Intermittent episodes of
apparent abdominal pain continued, and between the episodes, the patient be-
haved normally. On the morning of presentation, he had a third bowel movement
with reddish stools. His parents took him to day care, where he continued to have
occasional periods of crying and pain, followed by a bowel movement that ap-
peared to consist almost entirely of blood, including a large clot. After this bowel
movement, he was reportedly pale and diaphoretic. The day care provider called
the patients mother, who picked him up and took him to the emergency depart-
ment of another hospital.
On examination at the other hospital, the temperature was 36.5C, the pulse
178 beats per minute, the blood pressure 95/52 mm Hg, the respiratory rate 24 breaths
per minute, and the oxygen saturation 100% while the patient was breathing am-
n engl j med 376;13 nejm.org5 March 30, 2017 1269

Table 1. Laboratory Data.*

stools became more firm, and daily bowel move-
ments were associated with straining. He received
Reference Range, On Presentation, cholecalciferol, and he had begun using an
Variable Age-Adjusted This Hospital
unspecified over-the-counter teething gel and
Hematocrit (%) 33.039.0 17.5 unspecified homeopathic teething tablets 1 week
Hemoglobin (g/dl) 10.513.5 5.7 earlier. Immunizations were current through
Reticulocyte count (%) 0.52.5 7.6 4 months of age; vaccines (including the second
Whitecell count (per mm3) 600017,500 22,200
dose of live, oral humanbovine reassortant
pentavalent rotavirus vaccine) had been adminis-
Differential count (%)
tered 6 weeks earlier. There were no known aller-
Neutrophils 1749 38 gies. The patient lived with his parents, attended
Lymphocytes 6777 59 day care, and had no known sick contacts. His
Monocytes 411 3 parents were from Brazil; he was born in the
Redcell count (per mm 3) 3,700,0005,300,000 2,070,000 United States and had not traveled outside the
Prothrombin time (sec) 11.014.0 12.4 country. There was no family history of bleeding
disorders.
Prothrombintime international 0.91.1 1.0
normalized ratio On examination, the temperature was 36.3C,
Activated partial thromboplastin 22.137.0 19.4
the pulse 160 beats per minute, the blood pres-
time (sec) sure 98/47 mm Hg, the respiratory rate 32
Total protein (g/dl) 6.08.3 5.4 breaths per minute, and the oxygen saturation
Albumin (g/dl) 3.35.0 4.1
99% while the patient was breathing ambient
air. He appeared well. Bowel sounds were pres-
Globulin (g/dl) 1.94.1 1.3
ent; the abdominal examination was otherwise
Iron (g/dl) 45160 19 limited because the patient was crying. The dia-
Ironbinding capacity (g/dl) 230404 351 per contained melena and a small amount of
stool. The remainder of the examination was
* To convert the values for iron and ironbinding capacity to micromoles per liter,
multiply by 0.1791. normal.
Reference values are affected by many variables, including the patient popula Dr. Ruth Lim: Thirty-five minutes after the pa-
tion and the laboratory methods used. The ranges used at Massachusetts General tients arrival in the emergency department, an
Hospital are ageadjusted, for patients who are not pregnant and do not have
medical conditions that could affect the results. They may therefore not be ap ultrasound examination of the abdomen was
propriate for all patients. performed. There was no evidence of intussus-
ception, appendicitis, a focal lesion, or abnor-
mally dilated bowel loops. Bowel peristalsis was
bient air. The weight was 9.1 kg. On palpation present.
of the abdomen, there was diffuse tenderness, Dr. Gupta: On examination after ultrasonogra-
which was greater on the right side than on the phy, the pulse was 168 beats per minute, and the
left, and no masses. There were no external blood pressure 94/36 mm Hg. The patient ap-
anal fissures, and the remainder of the physi- peared pale. The abdomen was soft, without
cal examination was normal. Two hours after distention, tenderness, or masses, and bowel
arrival at the other hospital, the patient passed sounds were present. Results of the physical
a dark-red stool that was described as resem- examination were otherwise unchanged. Blood
bling currant jelly. Intravenous normal saline levels of electrolytes, glucose, aspartate amino-
(5 ml per kilogram) was administered, and he was transferase, alanine aminotransferase, alkaline
brought by ambulance to the emergency depart- phosphatase, total bilirubin, direct bilirubin,
ment of this hospital for further evaluation and and C-reactive protein were normal, as were the
treatment. anion gap, platelet count, red-cell indexes, and
The patient had a history of infantile colic and results of renal-function tests. The results of
gastroesophageal reflux, which had previously other laboratory tests are shown in Table 1.
been treated with ranitidine. He received a low- Packed red cells were transfused, and pantopra-
lactose cow milkbased formula. Pureed fruits zole and famotidine were administered intrave-
and vegetables had recently been introduced into nously.
his diet, after which constipation developed, his A diagnosis was made.
1270 n engl j med6376;13 nejm.org March 30, 2017

Case Records of the Massachuset ts Gener al Hospital
Differ en t i a l Di agnosis overdose that leads to liver failure and gastroin-

testinal bleeding, or nonsteroidal antiinflamma-
Dr. Lauren M. Allister: This 6-month-old boy pre- tory drugs, which can cause irritation of the
sented with gastrointestinal bleeding manifested gastric mucosa and subsequent gastrointestinal
by hematochezia, along with intermittent ab- bleeding? Although these exposures are unlikely
dominal pain and one episode of melena. He underlying causes of this patients illness, they
appeared ill and had tachycardia. Pertinent fea- warrant further consideration and, possibly, toxi-
tures of the history include gastroesophageal cologic testing.
reflux, a possible milk-protein allergy (since he
was receiving a low-lactose formula), and expo- Diagnostic Evaluation
sure to unspecified teething tablets and a homeo- My diagnostic considerations fall into three broad
pathic teething medication. It is important to categories: common, less common, and poten-
note the absence of fever, forceful vomiting, and tially life-threatening. Among the common diag-
hematemesis. Because care in the emergency noses, ileocolic intussusception seems to be the
department is more process-driven than outcome- most likely possibility; the patient presented at a
driven, the evaluation in this case can be con- typical age (since intussusception most com-
densed into the following steps: rapid assess- monly occurs during infancy or early childhood)
ment, stabilization, and diagnostic evaluation. and had colicky abdominal pain and worsening
gastrointestinal bleeding, with stool described as
Rapid Assessment resembling currant jelly. Meckels diverticulum
This ill patient had tachycardia, pallor, profound is the most common congenital malformation
anemia with ongoing blood loss, and intermit- of the gastrointestinal tract, and if the diverticu-
tently abnormal findings on abdominal examina- lum contains ectopic or heterotopic mucosa, it can
tion. He presented with gastrointestinal bleed- cause gastrointestinal bleeding.1,2 Of the clinical
ing manifested by hematochezia and melena. My findings associated with Meckels diverticulum,
initial diagnostic considerations include causes bleeding is one of the most common in chil-
of lower gastrointestinal bleeding, although the dren.1,3,4 Since Meckels diverticulum is classically
description of melena gives me reason to think associated with painless bleeding, this patients
that this patient could have bleeding from both apparent abdominal pain is difficult to reconcile
upper and lower gastrointestinal sources. Less with this diagnosis.5,6 However, if Meckels diver-
likely is an isolated, massive upper gastrointesti- ticulum is associated with obstruction caused by
nal bleed with rapid transit time through the intussusception, volvulus, or perforation, then
infants gastrointestinal tract. pain can be a complicating feature.3 I would also
consider an inflammatory or allergic gastritis
Stabilization or colitis, because these are common causes of
The patients airway was intact, and his breath- lower gastrointestinal bleeding among children
ing was unlabored. However, his circulation was who present to the emergency department.7 The
compromised; he was pale and had tachycardia, presence of mild gastritis plus colitis related to
and the hematocrit was 17.5% with ongoing a milk-protein allergy could explain both the
blood loss. He required volume resuscitation hematochezia and melena (mixed upper and
with the administration of isotonic fluids and lower gastrointestinal bleeding), as well as the
the transfusion of packed red cells, which was associated pain. Infectious colitis seems unlike-
performed in the emergency department. ly, given the absence of fever, sick contacts, and
The patient was neurologically intact. A bed- travel. Other common causes of lower gastro-
side glucose measurement may have been useful intestinal bleeding, such as a fissure or polyp,
in determining whether poor feeding with resul- are not typically associated with such a severe
tant hypoglycemia contributed to his unwell presentation, so these diagnoses are easily ruled
appearance. The reported use of homeopathic out in this case.
and unspecified teething medications raises con- In a 6-month-old infant, the less common
cerns about an unintentional toxic exposure. diagnoses that cause lower gastrointestinal bleed-
Could the teething tablets or medications have ing include vascular malformations of the gastro-
contained acetaminophen, which can cause an intestinal tract, atypical lymphonodular hyperpla-
n engl j med 376;13 7

nejm.org March 30, 2017 1271

sia, the hemolyticuremic syndrome, inflammatory two common diagnoses are ruled out. In addi-
bowel disease, toxin-mediated processes, and tion, I am worried about the possibility of poten-
underlying bleeding diatheses. I would give these tially life-threatening hemorrhagic shock, given
causes careful consideration only after the com- the patients continued blood loss and profound
mon diagnoses have been ruled out. anemia.
In this case, several diagnoses must be con- In the emergency department, emphasis is
sidered because they are potentially life-threat- placed on providing the best possible systematic
ening if missed. These diagnoses can be con- care during the period leading up to the diagno-
sequences of either the common or the less sis rather than conclusively determining the diag-
common conditions and include a perforated nosis; nevertheless, I think the diagnosis in this
viscus, an acute abdomen, obstruction, hemor- case is Meckels diverticulum. In an infant who
rhagic shock, septic shock, and the presence of has massive lower gastrointestinal bleeding with
associated upper gastrointestinal bleeding while resultant hemodynamic compromise and for
the patient is being evaluated for lower gastroin- whom intussusception has been ruled out on the
testinal bleeding. Serial physical examinations basis of ultrasonographic findings, the most
and diagnostic testing are critical in identifying likely diagnosis is Meckels diverticulum, and this
any of these potentially life-threatening pro- possibility needs to be investigated before other
cesses. diagnoses can be considered.7 The abdominal
pain is one aspect of this patients clinical pre-
Diagnostic Testing sentation that does not totally fit with the diag-
The findings ascertained through diagnostic test- nosis of Meckels diverticulum, although an ob-
ing that are the most important in developing a struction or perforation could introduce pain
differential diagnosis in this case are the hema- into the clinical picture. The description of me-
tocrit of 17.5%, the elevated white-cell count of lena is not consistent with Meckels diverticulum
22,200 per cubic millimeter (which is nonspe- but could be explained if the bleeding mucosa
cific but worrisome), and the absence of intus- from the diverticulum was proximal enough for
susception and other notable findings on ultra- resultant blood to undergo partial digestion.6 It
sonography. The normal electrolyte levels, liver is also possible that the single stool described
profile, and coagulation indexes are reassuring, as melena was not truly melena but stool with
and they argue against some systemic disease darker or maroon blood that originated from a
processes that would typically be associated lower, rather than an upper, gastrointestinal
with abnormalities in one or more of these mea- source. Mixed gastritis and colitis is less likely
sures. However, a few additional studies would than Meckels diverticulum overall, and bleeding
help to narrow the differential diagnosis. Be- related to allergic gastrointestinal disease is un-
cause of the possibility of a toxic exposure, I likely to be as acute and severe as the bleeding
would perform a serum toxicology screen. In seen in this case.8 In the emergency department,
addition, I would perform blood and stool cul- it is more straightforward to obtain a scan to
tures to evaluate for infection, as well as abdomi- assess for Meckels diverticulum than to perform
nal radiography to assess for bowel perforation, upper and lower endoscopy; the scan mandates
given the multiple days of gastrointestinal symp- coordination of fewer hospital resources, does not
toms and the worsening clinical appearance. To require the administration of anesthesia, and is
rule out upper gastrointestinal bleeding, I would noninvasive. If a scan were nondiagnostic, I
consider performing gastric aspiration. would consider other studies, such as endoscopy
In view of the available test results, the ab- or abdominal computed tomography.
sence of intussusception on abdominal ultraso- Dr. Virginia M. Pierce (Pathology): Dr. Baldwin,
nography, and the patients ongoing blood loss, what was your impression when you evaluated
two diagnoses from my list of common diagno- this patient?
ses remain most likely: Meckels diverticulum and Dr. Katherine R. Baldwin (Pediatric Gastroenter-
gastritis plus allergic colitis. Many other diagno- ology): Our first step was to localize the source
ses have been effectively ruled out through diag- of blood loss. Melena is classically thought to
nostic testing, and several less common causes reflect upper gastrointestinal bleeding (proximal
would not be seriously considered until these to the ligament of Treitz), but it can also be
1272 n engl j med 376;13 nejm.org March 30, 2017

8

caused by more distal lesions, such as lesions ining was performed in a lateral view. An abnor-
the small bowel and right colon.9 We considered mal focus of radiotracer accumulation was seen
both upper gastrointestinal sources (including in the right paramedian region of the abdomen
variceal bleeding, vascular malformations, and that gradually increased in intensity over time
ulcer) and lower gastrointestinal sources (includ-
(Fig. 1); this finding is consistent with ectopic
ing colitis, Meckels diverticulum, and vascular gastric mucosa in a Meckels diverticulum.
malformations). We thought that the subacute Technetium-99m pertechnetate normally accu-
tempo of this patients clinical presentation, the
mulates in any gastric mucosa, including ectopic
large volume of blood loss, and his age were gastric mucosa; therefore, this radiotracer is use-
most consistent with Meckels diverticulum. ful in the evaluation of a suspected Meckels di-
Although Meckels diverticulum is commonly verticulum. False positive scans can occur. Tech-
thought to be a painless lesion, pain can result netium-99m pertechnetate is excreted by the
from intussusception (with the diverticulum serv-urinary system, and activity is normally seen in
ing as the lead point), intermittent volvulus the bladder and kidneys. Radiotracer activity in
around associated fibrous bands, or torsion. the stomach can pass distally into the duodenum
We recommended that the patient undergo and small bowel. Premedication with a histamine
prompt evaluation by a pediatric surgeon and H2-receptor antagonist can reduce the release of
that a scan to assess for Meckels diverticulum radiotracer from the stomach. Bowel or urinary
be obtained after the administration of a hista- activity is suggested by movement of focal radio-
mine H2-receptor antagonist to help retain radio-tracer activity over time, whereas focal accumu-
tracer in the gastric mucosa. We did not think lation in a Meckels diverticulum should remain
that endoscopy would be immediately useful; fixed in position. A lateral view of the abdomen
although the performance of upper gastrointes- can be helpful in distinguishing urinary activity
tinal endoscopy is standard for a large volume ofin the ureters, which are located in a posterior
blood loss because of the potential for diagnos- position. A false positive scan can also result
tic and therapeutic intervention, most causes of from inflammation, intussusception, bowel ob-
lower gastrointestinal bleeding do not require struction, or vascular lesions.
colonoscopy. Furthermore, colonoscopy in a pa- A false negative scan can result from the
tient with acute severe bleeding may be techni- presence of too little or no gastric mucosa in a
cally challenging because of difficulty with visu-
Meckels diverticulum; approximately 20% of
alization.10 Meckels diverticula do not contain gastric mu-
cosa. Other causes of false negative scans include
recent ingestion of barium or perchlorate, move-
Cl inic a l Di agnosis
ment of the Meckels diverticulum, and brisk
Gastrointestinal bleeding due to Meckels diver- gastrointestinal bleeding.11
ticulum.
Discussion of M a nagemen t
Dr . L aur en M. A l l is ters
Di agnosis Dr. Allan M. Goldstein: As a result of the clinical
presentation and the findings on the scan, the
Gastrointestinal bleeding due to Meckels diver- infant was brought to the operating room. A
ticulum. short transverse incision was made in the right
lower quadrant, and the diverticulum was identi-
fied (Fig. 2). Inflammation and scarring were
Im aging S t udie s
present at its base; these findings are consistent
Dr. Lim: After the patient received premedication with ulceration in the small intestine, at its junc-
with intravenous famotidine, a technetium-99m tion with the diverticulum. A segmental small-
pertechnetate scan was obtained to assess for the bowel resection, which included the diverticu-
presence of a Meckels diverticulum. Immedi- lum and the presumed area of ulceration, was
ately after the intravenous injection of 0.97 mCi performed, followed by a hand-sewn end-to-end
of radiotracer, anterior planar imaging was per- anastomosis.
formed continuously for 1 hour. Additional imag- A variety of operations can be performed to
n engl j med 376;13 nejm.org March 30, 2017 1273

9

A Anterior
Right Left
5 min 10 min 15 min
20 min 25 min 30 min
35 min 40 min 45 min
B Lateral
%
56
0
Right Left

10

Figure 1 (facing page). Technetium-99m Pertechnetate

Scan of the Abdomen.
A technetium99m pertechnetate scan of the abdomen
was performed to assess for Meckels diverticulum.
Anterior planar images (Panel A), which were obtained
continuously for 1 hour, show an abnormal focus of ra
diotracer accumulation in the right paramedian region
of the abdomen that gradually increases in intensity
over time (black arrow). Physiological radiotracer accu
mulation is present in the stomach (white arrow), bow
el (arrowhead), and bladder (asterisk). Lateral planar
images (Panel B) confirm that the abnormal focus of
radiotracer accumulation (arrow) is in a location that is
compatible with bowel activity and is not consistent
with urinary activity, which would be more posterior.
treat a Meckels diverticulum that causes gastro-

intestinal bleeding. These include simple diver- Figure 2. Intraoperative Photograph.
ticulectomy, wedge resection of the diverticulum After a short transverse incision was made in the right
lower quadrant, the diverticulum was identified. Inflam
and the small cuff of adjacent ileum at its base,
mation and scarring are present at the base of the diver
and segmental small-bowel resection, which was ticulum; these findings are consistent with ulceration
done in this case. The primary cause of bleeding in the small intestine, at its junction with the diverticu
is the presence of acid-producing ectopic gastric lum. Photograph courtesy of Dr. David Lawlor.
mucosa in the diverticulum, which leads to the
development of an ulcer in adjacent normal mu-
cosa. The ulcer can be present in the diverticu- and other causes of bleeding have been ruled
lum itself but is usually located at the junction out, laparoscopy should be considered to assess
of the diverticulum and the ileum,12 as appeared for Meckels diverticulum.
to be the case in this patient. Although remov-
ing both the ectopic mucosa and the ulcer would Pathol o gic a l Discussion
seem to be the best approach, removing the ec-
topic mucosa alone may be sufficient, since the Dr. Jochen K. Lennerz: We received a segment of
ulcer would probably then heal. However, it is small bowel (1.7 cm by 1.5 cm by 1.5 cm) with
essential to remove all ectopic gastric mucosa, an attached intact, blind-ending diverticulum
which cannot be reliably detected from the out- (2.5 cm by 0.8 cm by 0.8 cm) for pathological
side. Therefore, a reasonable approach is to per- examination. The serosa near the small intestine
form a simple diverticulectomy for a diverticu- showed patchy fibrinous inflammation (Fig. 2)
lum with a narrow base but to perform a wedge and was otherwise mildly hyperemic; the tip of
or segmental resection for a diverticulum with a the diverticulum had no attached bands. The
broad base, since ectopic tissue may be left be- sections showed an average wall thickness of
hind if the diverticular base is not fully excised. 0.2 cm and normally folded mucosa with red
If the area of ulceration is apparent, as in this discoloration toward the small bowel. In con-
case, then resecting it with the diverticulum is trast to the mucosal herniation through the
also reasonable. bowel wall that is present in diverticular disease,
An important scenario to consider is whether this diverticulum contained all three layers of
this patient would have received different treat- bowel wall. Given the anatomical location of
ment if the scan had been negative, which could the diverticulum on the antimesenteric surface
have easily occurred, given the imperfect sensi- of the mid-ileum, these findings represent per-
tivity of the test.13,14 Meckels diverticulum needs sistence of a proximal part of the vitelline duct
to be included in the differential diagnosis for (omphalomesenteric duct), or Meckels diver-
any child being evaluated for hematochezia. If a ticulum.
technetium-99m pertechnetate scan is negative A histotopogram allowed us to perform a
n engl j med 376;13 nejm.org

11 March 30, 2017 1275

mucous neck cells, parietal cells, and basal chief

A
cells (Fig. 3C). This composition is diagnostic of
a Meckels diverticulum containing a large amount
of terminally differentiated, heterotopic gastric
mucosa of the body (fundic type). There was no
evidence of heterotopic pancreatic tissue, dyspla-
sia, or cancer.
It is unusual for a Meckels diverticulum in a
6-month-old patient to be completely lined by
HCI gastric mucosa that is most likely secreting a
large amount of acid into the small intestine.
The mucin-secreting surface foveolar epithelial
cells in this Meckels diverticulum protected the
B C underlying diverticular mucosa from the acid
secreted by the underlying parietal cells. Thus,
the pathophysiological cascade in this patient
can be described as follows: the secreted, non-
neutralized acid produced by the heterotopic gas-
tric mucosa and the secreted chief-cellderived
enzymes in the diverticulum led to peptic ulcer-
ation in the adjacent intestinal mucosa, which
caused gastrointestinal bleeding. The conversion
of the hemoglobin in the gastrointestinal bleed
into melena was presumably related to the large
Figure 3. Histotopogram. load of digestive chemicals secreted by the
A histologic section of the diverticulum shows all three layers of bowel wall Meckels diverticulum. The serosal fibrinous ad-
(Panel A); serosal fibrinous exudate is present (white arrow). The intestinal hesions may have caused the pain.
mucosa at the opening of the diverticulum shows mucosal erosion, acute Dr. Gupta: After surgery, the patients course
inflammation, and underlying granulation tissue, findings consistent with an was briefly complicated by ileus, but by the third
ulceration (Panel B). The gastric mucosa lines the entire diverticulum and
is composed of surface foveolar cells, mucous neck cells, parietal cells, and
postoperative day, his diet was regular and he
basal chief cells (Panel C). The heterotopic gastric mucosa secretes hydro was discharged home. Nine days later, he was
chloric acid, which causes peptic ulceration in the adjacent intestinal mucosa seen for follow-up by the pediatric surgeon and
(Panel A, black arrow). was doing well; he had no pain, hematochezia,
or melena, and the abdominal examination was
normal.
histologic examination in the spatial context of
the entire diverticulum (Fig. 3A). There were two A nat omic a l Di agnosis
key mucosal findings. First, the intestinal mu-
cosa at the opening of the diverticulum showed Meckels diverticulum with heterotopic gastric
epithelial erosion and underlying granulation mucosa and associated peptic ulceration in the
tissue, findings consistent with mucosal ulcera- intestinal mucosa.
tion (Fig. 3B). The acute inflammation extended This case was presented at Pediatric Grand Rounds.
Dr. Allister reports receiving consulting fees from Medscape
through the muscularis propria and was associ- Consult. No other potential conflict of interest relevant to this
ated with fibrinous serositis. Second, this diver- article was reported.
ticulum was remarkable because the mucosa of Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
the entire diverticulum was made up of at least We thank Dr. Rebecca Cook, Chief Resident in Pediatrics, for
four distinct cell types: surface foveolar cells, her assistance with organizing the conference.
References
1. Sagar J, Kumar V, Shah DK. Meckels 2. Pepper VK, Stanfill AB, Pearl RH. Di- diverticulum. Surg Clin North Am 2012;
diverticulum: a systematic review. J R Soc agnosis and management of pediatric ap- 92:505-26.
Med 2006;99:501-5. pendicitis, intussusception, and Meckel 3. Park JJ, Wolff BG, Tollefson MK, Walsh

12

EE, Larson DR. Meckel diverticulum: the 7. Teach SJ, Fleisher GR. Rectal bleeding tinal Endoscopy guidelines. J Pediatr Gas-
Mayo Clinic experience with 1476 patients in the pediatric emergency department. troenterol Nutr 2017;64:133-53.
(1950-2002). Ann Surg 2005;241:529-33. Ann Emerg Med 1994;23:1252-8. 11. Spottswood SE, Pfluger T, Bartold SP,
4. Yahchouchy EK, Marano AF, Etienne 8. Arvola T, Ruuska T, Kernen J, Hyty et al. SNMMI and EANM practice guide-
JC, Fingerhut AL. Meckels diverticulum. H, Salminen S, Isolauri E. Rectal bleeding line for Meckel diverticulum scintigraphy
J Am Coll Surg 2001;192:658-62. in infancy: clinical, allergological, and mi- 2.0. J Nucl Med Technol 2014;42:163-9.
5. Itriyeva K, Harris M, Rocker J, Goch- crobiological examination. Pediatrics 2006; 12. Cobb DB. Meckels diverticulum with
man R. Not just painless bleeding: Meck- 117(4):e760-e768. peptic ulcer. Ann Surg 1936;103:747-68.
els diverticulum as a cause of small bowel 9. Zuckerman GR, Trellis DR, Sherman 13. Al Janabi M, Samuel M, Kahlenberg A,
obstruction in children two cases and TM, Clouse RE. An objective measure of Kumar S, Al-Janabi M. Symptomatic pae-
a review of the literature. Case Rep Emerg stool color for differentiating upper from diatric Meckels diverticulum: stratified
Med 2015;2015:938346. lower gastrointestinal bleeding. Dig Dis diagnostic indicators and accuracy of
6. Poley JR, Thielen TE, Pence JC. Bleed- Sci 1995;40:1614-21. Meckels scan. Nucl Med Commun 2014;
ing Meckels diverticulum in a 4-month- 10. Thomson M, Tringali A, Dumonceau 35:1162-6.
old infant: treatment with laparoscopic JM, et al. Paediatric gastrointestinal endos- 14. Tseng YY, Yang YJ. Clinical and diag-
diverticulectomy a case report and re- copy: European Society for Paediatric nostic relevance of Meckels diverticulum
view of the literature. Clin Exp Gastroen- Gastroenterology, Hepatology, and Nutri- in children. Eur J Pediatr 2009;168:1519-23.
terol 2009;2:37-40. tion and European Society of Gastrointes- Copyright 2017 Massachusetts Medical Society.
Lantern SLideS Updated: CompLete powerpoint SLide SetS from the CLiniCopathoLogiCaL ConferenCeS
Any reader of the Journal who uses the Case Records of the Massachusetts General Hospital as a teaching exercise or reference
material is now eligible to receive a complete set of PowerPoint slides, including digital images, with identifying legends,
shown at the live Clinicopathological Conference (CPC) that is the basis of the Case Record. This slide set contains all of the
images from the CPC, not only those published in the Journal. Radiographic, neurologic, and cardiac studies, gross specimens,
and photomicrographs, as well as unpublished text slides, tables, and diagrams, are included. Every year 40 sets are produced,
averaging 50-60 slides per set. Each set is supplied on a compact disc and is mailed to coincide with the publication of the
Case Record.
The cost of an annual subscription is $600, or individual sets may be purchased for $50 each. Application forms for the current
subscription year, which began in January, may be obtained from the Lantern Slides Service, Department of Pathology,
Massachusetts General Hospital, Boston, MA 02114 (telephone 617-726-2974) or e-mail Pathphotoslides@partners.org.
n engl j med 376;13 nejm.org March 30, 2017 1277

13

Case Records of the Massachusetts General Hospital
Founded by Richard C. Cabot

Eric S. Rosenberg, M.D., Nancy Lee Harris, M.D., Editors
Virginia M. Pierce, M.D., David M. Dudzinski, M.D., Meridale V. Baggett, M.D.,
Dennis C. Sgroi, M.D., Jo-Anne O. Shepard, M.D., Associate Editors
Emily K. McDonald, Sally H. Ebeling, Production Editors
Case 2-2017: An 18-Year-Old Woman

with Acute Liver Failure
Kristian R. Olson, M.D., Amir H. Davarpanah, M.D.,
Esperance A. Schaefer, M.D., M.P.H., Nahel Elias, M.D.,
and Joseph Misdraji, M.D.
Pr e sen tat ion of C a se

From the Departments of Medicine Dr. Carolyn A. Boscia (Medicine and Pediatrics): An 18-year-old woman was seen in
(K.R.O., E.A.S.), Pediatrics (K.R.O.), Radi- the emergency department of this hospital 11 weeks after the birth of her first
ology (A.H.D.), Surgery (N.E.), and Pathol-
ogy (J.M.), Massachusetts General Hos- child because of acute liver failure.
pital, and the Departments of Medicine The patient had been well until 1 week before this presentation, when rhinor-
(K.R.O., E.A.S.), Pediatrics (K.R.O.), Radi- rhea, sore throat, and cough developed. On the fourth day of illness, she was seen
ology (A.H.D.), Surgery (N.E.), and Pathol-
ogy (J.M.), Harvard Medical School in an urgent care clinic because of worsening cough, wheezing, and dyspnea.
both in Boston. Bronchitis was diagnosed, and promethazinedextromethorphan syrup and a
N Engl J Med 2017;376:268-78. 5-day course of oral azithromycin were prescribed. The patient returned home.
DOI: 10.1056/NEJMcpc1613467 Over the next 3 days, abdominal discomfort, nausea, vomiting, diarrhea, and
Copyright 2017 Massachusetts Medical Society.
vaginal bleeding developed. The patient also noted progressive yellowing of her
skin and eyes. When she woke up on the morning of the current presentation,
she felt light-headed. When she arose from bed, syncope occurred; the patient fell
and had a laceration of the chin. Her boyfriend called emergency medical services
(EMS), and a team was dispatched to the patients home.
On assessment by EMS personnel, the patient had jaundice and diaphoresis.
She appeared fatigued. The pulse was 120 beats per minute, the blood pressure
82/56 mm Hg, the respiratory rate 22 breaths per minute, and the oxygen satura-
tion 100% while she was breathing ambient air. Nystagmus occurred on right lateral
gaze. The abdomen was distended, and tenderness was present in the right lower
quadrant. The capillary blood glucose level was 121 mg per deciliter (6.7 mmol per
liter), and an electrocardiogram showed sinus tachycardia. Intravenous fluids and
supplemental oxygen (through a nasal cannula at a rate of 2 liters per minute) were
administered, and the patient was transported to the emergency department of
another hospital.
On arrival at the other hospital, the patient reported burning abdominal pain,
which she rated at 10 on a scale of 0 to 10, with 10 indicating the most severe
pain. The temperature was 37.0C, the pulse 88 beats per minute, the blood pres-
268 14 nejm.org
n engl j med 376;3 January 19, 2017

sure 107/42 mm Hg, the respiratory rate 24 breaths ter delivery because of unspecified abnormal
per minute, and the oxygen saturation 100% laboratory test results. Her medications were
while she was breathing ambient air. The abdo- albuterol as needed, azithromycin, and prometha-
men was soft, with tenderness on the right zinedextromethorphan syrup; she did not take
side, and there was trace edema of the legs. The herbal remedies or supplements and had no
results of the examination were otherwise unknown allergies. Immunizations were reportedly
changed. The blood carbon dioxide level was current.
21 mmol per liter (reference range, 24 to 34), and Six weeks before this presentation, the patient
the blood glucose level was 104 mg per deciliter had moved to an urban area of New England
(5.8 mmol per liter; reference range, 70 to 100 mg where she currently lived with her daughter, boy-
per deciliter [3.9 to 5.6 mmol per liter]). The friend, and boyfriends parents. She was of South-
anion gap and blood levels of sodium, potassium, east Asian descent and had been born in the
and chloride were normal, as were the results of United States. She did not smoke tobacco, use
renal-function tests and a serum toxicology screen, illicit drugs, or drink alcohol. A grandmother
which included a test for acetaminophen. Other had unspecified liver disease.
laboratory test results are shown in Table 1. On examination, the patient appeared tired
Dr. Amir H. Davarpanah: A chest radiograph and and had marked jaundice. The temperature was
a computed tomographic (CT) scan of the head 37.6C, the pulse 90 beats per minute, the blood
(obtained without the administration of intrave- pressure 100/58 mm Hg, the respiratory rate 18
nous contrast material) were normal. Ultrasonog- breaths per minute, and the oxygen saturation
raphy of the abdomen revealed mildly increased 100% while she was breathing ambient air. Con-
hepatic parenchymal echogenicity (Fig. 1A). This junctival icterus was present. The abdomen was
finding, although nonspecific, could reflect he- soft, with mild epigastric tenderness; abdominal
patic steatosis or diffuse parenchymal disease. guarding, rebound tenderness, and Murphys sign
No focal liver lesions were identified. The gall- were absent. There was a 1.5-cm laceration on
bladder was distended, with apparent wall ede- the chin. The remainder of the physical exami-
ma, a small amount of pericholecystic fluid, and nation was normal. Examination of a peripheral-
layering sludge (Fig. 1B). The common bile duct blood smear revealed smudge cells, burr cells,
was normal in diameter, with no intrahepatic basophilic stippling, rouleaux formation, dysplas-
biliary ductal dilatation (Fig. 1C). The spleen was tic neutrophils, and 1+ polychromasia. The anion
mildly enlarged, to a greatest longitudinal diam- gap, venous blood-gas measurements, and results
eter of 13 cm (normal, 12). of renal-function tests were normal, as were
Dr. Boscia: Intravenous fluids, piperacillintazo- blood levels of sodium, potassium, chloride, car-
bactam, morphine, ondansetron, and N-acetyl- bon dioxide, magnesium, glucose, amylase, lipase,
cysteine were administered, and packed red cells and fibrinogen; additional laboratory test results
were transfused. Five hours after arrival at the are shown in Table 1. Testing for urinary human
other hospital, the patient was transferred to the chorionic gonadotropin was negative. A serum
emergency department of this hospital. toxicology screen was negative, and a urine
In the emergency department, the patient re- toxicology screen was positive for opiates and
ported that the abdominal pain and light-headed- negative for all other analytes. Urinalysis re-
ness had resolved and the nausea had decreased. vealed slightly cloudy, amber-colored urine with
She recalled that during the past several days, a specific gravity of 1.019, a pH of 6.0, 2+ bili-
her gums had bled easily when she brushed her rubin, 2+ urobilinogen, and 1+ occult blood by
teeth and her urine had been tea-colored. She dipstick; there were 0 to 2 white cells and 0 to
had a history of mild asthma. Eleven weeks ear- 2 red cells per high-power field. An electrocar-
lier, she had given birth to her first child; after diogram showed sinus tachycardia.
an otherwise uncomplicated pregnancy, preterm Dr. Davarpanah: Doppler ultrasonography of
labor developed and was complicated by placen- the abdomen revealed persistent evidence of wall
tal abruption, and vaginal delivery occurred at edema and intraluminal sludge in the gallblad-
32 weeks of gestation. The patient reported that der. Murphys sign was absent, although this
she had remained in the hospital for 1 week af- finding was not reliable because of the prior
n engl j med 376;3 nejm.org January 19, 2017 269

15

Table 1. Laboratory Data.*
Reference Range, On Presentation, Reference Range, On Presentation,

Variable Other Hospital Other Hospital This Hospital This Hospital
Hematocrit (%) 3648 19.6 36.046.0 23.4
Hemoglobin (g/dl) 12.015.8 6.4 12.016.0 7.9
White-cell count (per mm3) 4.811.2 10.1 4.513.0 11.1
Differential count (%)
Neutrophils 4585 72 4062 77
Band forms 08 5
Lymphocytes 1545 10 2740 9
Monocytes 012 6 411 10
Eosinophils 07 1 08 1
Myelocytes 0 4 0 2
Metamyelocytes 0 2 0 1
3
Platelet count (per mm ) 150,000400,000 140,000 150,000400,000 175,000
Red-cell count (per mm3) 3,600,0005,400,000 1,720,000 4,000,0005,200,000 2,210,000
Nucleated red-cell count (per 100 white cells) 0 0 0 1
Mean corpuscular volume (fl) 8298 113.8 80.0100.0 105.9
Mean corpuscular hemoglobin (pg/red cell) 27.035.0 37.4 26.034.0 35.7
Mean corpuscular hemoglobin concentration (g/dl) 32.037.0 32.9 31.037.0 33.8
Red-cell distribution width (%) 9.017.9 22.1 11.514.5 24.4
Direct antiglobulin test Negative Negative
Prothrombin time (sec) 11.014.0 24.5
Prothrombin-time international normalized ratio 0.91.1 2.6 0.91.1 2.1
Activated partial-thromboplastin time (sec) 22.035.0 43.4
d-Dimer (ng/ml) <500 591
Calcium (mg/dl) 8.710.5 7.2 8.510.5 7.2
Phosphorus (mg/dl) 2.64.5 1.8
Total protein (g/dl) 6.48.6 5.5 6.08.3 5.3
Albumin (g/dl) 3.44.8 2.1 3.35.0 2.2
Globulin (g/dl) 1.94.1 3.1
Alanine aminotransferase (U/liter) 045 20 733 24
Aspartate aminotransferase (U/liter) 040 152 932 152
Alkaline phosphatase (U/liter) 40150 22 15350 14
Total bilirubin (mg/dl) 0.21.2 19.7 01.0 26.3
Direct bilirubin (mg/dl) 00.4 21.9
-Glutamyltransferase (U/liter) 536 116
Lactate dehydrogenase (U/liter) 110210 344
Lactic acid (mmol/liter) 0.52.2 3.9 0.52.2 1.7
Ammonia (mol/liter) 1248 49
* To convert the values for calcium to millimoles per liter, multiply by 0.250. To convert the values for phosphorus to millimoles per liter,
multiply by 0.3229. To convert the values for bilirubin to micromoles per liter, multiply by 17.1. To convert the values for lactic acid to milli-
grams per deciliter, divide by 0.1110. To convert the values for ammonia to micrograms per deciliter, divide by 0.5872.
Reference values are affected by many variables, including the patient population and the laboratory methods used. The ranges used at
Massachusetts General Hospital are for adults who are not pregnant and do not have medical conditions that could affect the results. They
may therefore not be appropriate for all patients.
270 n engl j med 376;3 nejm.org January 19, 2017

16

A B
C D
E F
Figure 1. Abdominal Ultrasound Images.

Grayscale ultrasound images show mildly echogenic parenchyma of the liver (Panel A) and distention of the gall-
bladder, with layering sludge, wall edema, and a small amount of pericholecystic fluid (Panel B). A color Doppler
ultrasound image shows that the common bile duct is normal in diameter (Panel C). Pulsed-wave Doppler ultra-
sound images show normal flow of the hepatic artery (Panel D), portal veins (Panel E), and hepatic veins (Panel F).
administration of analgesic agents. Pulsed-wave portal veins (Fig. 1E) and hepatofugal flow in
Doppler ultrasonography revealed normal arte- the hepatic veins (Fig. 1F).
rial flow in the main hepatic artery (Fig. 1D). Dr. Boscia: The administration of N-acetylcys-
There was also normal hepatopetal flow in the teine was continued, and the chin laceration was

17

sutured. Additional laboratory tests were per- 21 years of age), up to 50% of patients who pre-
formed, and a diagnosis was made. sent with acute liver failure do not present with
encephalopathy.2 Modified criteria for the diag-
nosis of acute liver failure in children include
Differ en t i a l Di agnosis
evidence of acute liver injury and severe coagu-
Dr. Kristian R. Olson: This previously healthy 18- lopathy (INR, >2.0) in the absence of encepha-
year-old woman presented 11 weeks after the lopathy. Given this patients age, I would argue
birth of her first child with evidence of worsen- that she meets the criteria for acute liver failure.
ing liver failure after a nonspecific 7-day illness. A specific diagnosis is important in deter-
To develop a differential diagnosis, it is impor- mining prognosis, guiding treatment, and coun-
tant to determine whether the patients liver seling the patients relatives. Using data derived
dysfunction is consistent with a diagnosis of from several large, multicenter series involving
acute liver failure. the Pediatric Acute Liver Failure Study Group
and the Acute Liver Failure Study Group, we can
Acute Liver Failure construct lists of recognized causes of acute
Acute liver failure in adults is characterized by a liver failure in children older than 10 years of
sudden loss of hepatic function without evidence age and in adults.1,3 Because this patient is at the
of preexisting liver disease. Criteria for the diag- threshold of adulthood, we need to consider the
nosis include the presence of coagulopathy (inter- causes in each population (Fig. 2).
national normalized ratio [INR], >1.5), hepatic This patient has nonspecific symptoms and
encephalopathy, and an illness of less than 24 findings on physical examination, and so it might
weeks duration.1 This patient has evidence of seem futile to arrive at a specific diagnosis.
liver injury and an INR well above 1.5. She does However, the presence or absence of relatively
not have features of encephalopathy, such as al- elevated values on routine laboratory tests can
tered consciousness, compromised intellectual help immensely. She has an aspartate amino-
functioning, tremors, or asterixis, and thus she transferase level that is nearly 5 times the upper
may meet only the criteria for acute liver injury. limit of the normal range, whereas the alanine
However, in the pediatric population (which can aminotransferase level is not elevated at all. The
be considered to include patients who are up to direct bilirubin level is more than 26 times the
Other cause, 2
Veno-occlusive disease, 2
Drug-induced Other cause, 7 Indeterminate
Multiple causes, 1
liver injury, 7 cause, 14
Hepatitis A or B
Shock or ischemia, 3 virus infection, 10
Viral hepatitis, 5
Indeterminate
cause, 32 Drug-induced
Autoimmune liver injury, 11
disease, 10
Ischemia, 4 Acetaminophen
Acetaminophen Autoimmune exposure, 47
Metabolic disease, 9
exposure, 29 disease, 5
Wilson's
disease, 2
3%Yr of Age
Children >10 Adults
Figure 2. Causes of Acute Liver Failure in Children and Adults.

Data are adapted from Lee et al.1 and Squires and Alonso.3 The numbers shown are percentages. Among children
older than 10 years of age, Wilsons disease accounted for 90% of metabolic disease.
272 n engl j med 376;3 nejm.org

18 January 19, 2017

upper limit of the normal range, with the major- of acute liver failure are caused by drug-induced
ity being conjugated bilirubin (22 mg per deciliter liver injury. This patient did not report use of
[374.5 mol per liter]). Cholestasis, which repre- over-the-counter medication; however, it is im-
sents a decrease in bile flow caused by either portant to confirm that she includes dietary and
impaired secretion of hepatocytes or obstruc- nutritional supplements as over-the-counter med-
tion, is heralded by prominent elevations in the ications. She began taking azithromycin and
bilirubin level and in the alkaline phosphatase promethazinedextromethorphan 3 days before
or -glutamyltransferase level. Despite the direct the current presentation. Antimicrobial therapy
hyperbilirubinemia in this patient, the alkaline is the most frequent cause of drug-induced liver
phosphatase level is below the normal range. injury, and in particular, the incidence of azith-
However, with a -glutamyltransferase level of romycin-associated hepatic injury is increasing.
more than 3 times the upper limit of the normal Drug-induced liver injury affects women in 72%
range, evidence of a cholestatic pattern remains. of cases and results in hepatocellular injury in
61% of cases.6 A cholestatic pattern can arise but
Acetaminophen Exposure typically does so 1 to 3 weeks after the patient
Although a serum acetaminophen level was un- has started a new medication. In addition, eosino-
detectable in this patient on presentation, it is philia and fever are typical features of drug-
important to maintain suspicion for either inad- induced liver injury that are not present in this
vertent chronic ingestion or an acute one-time patient. Furthermore, the time between the ini-
ingestion several days before presentation. The tiation of azithromycin therapy and the develop-
RumackMatthew nomogram, which is used to ment of acute liver failure in this patient was
assess for potential hepatotoxicity after an acet- only a few days, which makes the diagnosis of
aminophen exposure, is developed to assess only drug-induced liver injury unlikely.
for acute, single ingestions. The prevalence of
postpartum depression is approximately 10%,4 Pregnancy
and young age and unplanned pregnancy have During pregnancy, dilutional hypoalbuminemia
been identified as risk factors. Although this and elevation of the placental-derived alkaline
patient could have ingested acetaminophen 48 to phosphatase level can lead providers to falsely
72 hours before presentation and had an unde- assume that the patient has liver disease. How-
tectable level on presentation, there is no history ever, the elevations of the aspartate aminotrans-
of a psychiatric illness, which might suggest the ferase level, INR, and -glutamyltransferase level
possibility of an intentional overdose. In addi- in this patient are uniformly abnormal. Eclamp-
tion, the mechanism of acetaminophen hepato- sia affects 2 to 8% of pregnant women and can
toxicity is centrilobular necrosis, which is caused occur up to 6 weeks post partum, but this pa-
by the accumulation of N-acetyl-p-benzoquinone tients symptoms developed later. Furthermore,
imine. The hallmark of liver injury is markedly in pregnant women, the aminotransferase levels
elevated aminotransferase levels, which are usu- are typically 10 to 20 times the upper limit of the
ally in the thousands and frequently 400 times normal range and the bilirubin level is typically
the upper limit of the normal range. This bio- less than 5 mg per deciliter (85.5 mol per liter);
chemical feature is not consistent with this pa- also, the alkaline phosphatase level in this pa-
tients laboratory test results. However, she re- tient is higher than would be expected during
ceived treatment with N-acetylcysteine, which pregnancy.
results in increased survival even among pa- The HELLP syndrome (hemolysis, elevated liv-
tients with acute liver failure that is unrelated to er enzyme levels, and a low platelet count) occurs
acetaminophen exposure.5 in less than 1% of pregnant women, and only
one third of cases occur after delivery.7 In this
Drug-Induced Liver Injury case, examination of a peripheral-blood smear
Idiosyncratic hepatic reactions to medications did not reveal typical features of hemolysis, al-
other than acetaminophen and to complemen- though the presence of indirect hyperbilirubine-
tary or alternative therapies are referred to as mia may suggest a hemolytic process. However,
drug-induced liver injury. In adults, 11% of cases the patients platelet count was normal, and thus

19

the diagnosis of the HELLP syndrome is un- Patients with autoimmune hepatitis typically
likely. present with nonspecific symptoms, including
fatigue, lethargy, malaise, anorexia, nausea, ab-
Ischemic Hepatopathy dominal pain, and itching.10 Symptoms may first
Although the patient had hypotension when the become evident during pregnancy, and postpar-
EMS team arrived, the hemodynamic insult in tum exacerbations do occur. However, some fea-
ischemic hepatopathy is normally present well tures of autoimmune hepatitis are absent in this
before evidence of liver injury. In addition, the patient, including coexisting autoimmune condi-
typical biochemical profile of ischemic hepatopa- tions, associated small-joint arthralgias, and the
thy includes a dramatic rise in aminotransferase typical pattern of markedly elevated aminotrans-
and lactate dehydrogenase levels and normal or ferase levels. In addition, she does not have ele-
only mildly abnormal hepatic synthetic function. vated globulin levels (which typically correlate
The BuddChiari syndrome, or hepatic venous with IgG levels, even in the setting of acute liver
outflow obstruction, is another consideration failure).11 Given that the patient is female and
because the pooled prevalence during pregnancy her ratio of alkaline phosphatase (IU per liter) to
and the puerperium is approximately 6.8%.8 aspartate aminotransferase (IU per liter) is lower
However, acute liver failure develops in less than than 1.5, her globulin level is not elevated, and
5% of patients with the BuddChiari syndrome. there is no evidence of illicit-drug or excessive
In addition, although the aminotransferase levels alcohol use, we can calculate that she has a
may be only moderately elevated (as in this patient), score on the scoring system of the International
the bilirubin level is seldom higher than 7 mg Autoimmune Hepatitis Group of 7 (on a scale
per deciliter (119.7 mol per liter), whereas the ranging from 20 to 31, with a score of 10 to 15
bilirubin level in this patient is higher than 20 mg indicating probable autoimmune hepatitis and
per deciliter (342.0 mol per liter). This patient a score higher than 15 indicating definite auto-
also had a normal vascular Doppler ultrasound immune hepatitis).9 Although the information
evaluation, which rules out the diagnosis of the we are given at this point is inadequate to allow
BuddChiari syndrome. us to completely calculate the score and defini-
tively rule out the possibility of autoimmune
Viral Infection hepatitis, it makes this diagnosis unlikely.
Viral hepatitis is the cause of acute liver failure
in 10% of cases in developed countries. It is inter- Wilsons Disease
esting to note that this patients grandmother Wilsons disease, also known as hepatolenticular
had unspecified liver disease and that the pa- degeneration, is an autosomal recessive disease
tient is of Southeast Asian descent. In the United characterized by impaired copper metabolism
States, the rate of chronic hepatitis B virus infec- due to a defective ATPase. The mean age at onset
tion is 6% among pregnant women of Asian ranges from 12 to 23 years, and this patients
descent versus only 0.6% among pregnant white age falls within that range. Patients with Wilsons
women. An exacerbation of hepatopathy can oc- disease may present with chronic liver disease,
cur as the result of the relative immunosuppres- acute liver failure, hemolysis, and psychiatric or
sion associated with pregnancy. However, if this neurologic manifestations. The Leipzig criteria
patient received prenatal care, she would have for Wilsons disease might be helpful in deter-
been screened for hepatitis B virus. In addition, mining the diagnosis in this patient, but we are
she had no fever and few risk factors for acute not given the results of biochemical tests for
hepatitis B virus infection, and viral hepatitis copper or genetic testing.12
typically results in aminotransferase levels that Fortunately, rapid diagnostic criteria for Wil-
are more than 25 times the upper limit of the sons disease can be used in patients who pre-
normal range. sent with acute liver failure. A screen that shows
a ratio of alkaline phosphatase (IU per liter) to
Autoimmune Hepatitis total bilirubin (mg per deciliter) of lower than
Autoimmune hepatitis is a chronic, progressive 4.0 and then subsequently shows a ratio of aspar-
disorder, but it can also cause acute liver failure.9 tate aminotransferase (IU per liter) to alanine
274 n engl j med 376;3 20

nejm.org January 19, 2017

aminotransferase (IU per liter) of higher than nant Wilsons disease.15 In this patient, there-
2.2 has been described as 100% sensitive and fore, either the antecedent illness or treatment
specific for the diagnosis of Wilsons disease.13 with azithromycin may have played a role.
According to these criteria, this patient has a The revised Wilsons disease prognostic index
presumptive diagnosis of Wilsons disease. Fur- is highly accurate in predicting death due to
thermore, patients with acute liver failure who fulminant Wilsons disease in both children and
have Wilsons disease have a median alkaline adults.16,17 A score higher than 11 portends death
phosphatase level of 20.5 U per liter, as com- if the patient does not undergo transplantation,
pared with a median level of 146.5 U per liter and in this patient, the score was 14. Given the
among patients with acute liver failure who do patients vanishingly low likelihood of survival,
not have Wilsons disease. This patients alkaline we recommended admission to the intensive
phosphatase level was 22.0 U per liter. care unit (ICU) and immediate evaluation for
In Wilsons disease, acute liver failure devel- orthotopic liver transplantation.
ops in the setting of subclinical chronic liver After the patient was admitted to the ICU, the
disease. If liver transplantation is not performed, 24-hour urinary copper level was obtained. She
acute liver failure due to Wilsons disease is fatal. underwent a slit-lamp examination, and there was
I believe that serum copper and 24-hour urinary no evidence of KayserFleischer rings. Because
copper levels were most likely obtained to con- she had intact renal function, chelation therapy
firm the diagnosis of Wilsons disease in this with penicillamine was initiated to promote uri-
patient and that, if she survived, she underwent nary copper excretion as a bridging measure
liver transplantation. while she awaited transplantation.
Dr. Virginia M. Pierce (Pathology): Dr. Schaefer, During hospital days 2 through 4, additional
what was your impression when you evaluated testing was performed. The patients serum
this patient? copper level was normal (0.96 g per milliliter
Dr. Esperance A. Schaefer: The hepatology service [15.1 mol per liter]; reference range, 0.75 to
was consulted after the patients arrival at the 1.45 [11.8 to 22.8 mol per liter]), her cerulo-
emergency department. The pattern of liver injury plasmin level was low (8 mg per milliliter; refer-
including minimal elevation of aminotrans- ence range, 20 to 60), and her 24-hour urinary
ferase levels, marked hyperbilirubinemia, and a copper level was markedly elevated (1419 g per
low-to-normal alkaline phosphatase level did specimen; reference range, 15 to 60). Her coagu-
not fit neatly into a hepatocellular or cholestatic lopathy worsened, and confusion and hyperam-
pattern. These biochemical findings, combined monemia developed. Shortly after the patient was
with the parenchymal changes observed on ultra- placed on the liver transplantation list, a donor
sonography, suggested preexisting chronic liver was identified, and the patient underwent ortho-
disease with superimposed acute liver injury. topic liver transplantation that day.
Given that this patient was taking azithromy-
cin, we considered the diagnosis of drug-induced Cl inic a l Di agnosis
liver injury.13,14 However, several clinical features
in this case strongly suggested an alternative Fulminant hepatic failure due to Wilsons disease.
diagnosis. The patients age, sex, possible hemo-
lytic anemia, and low alkaline phosphatase level Dr . K r is t i a n R . Ol sons
raised strong clinical suspicion for Wilsons Di agnosis
disease. We used the rapid diagnostic criteria for
Wilsons disease,13 and the ratio of alkaline Wilsons disease (hepatolenticular degeneration).
phosphatase to total bilirubin was 0.5 and the
ratio of aspartate aminotransferase to alanine Pathol o gic a l Discussion
aminotransferase was 6.3; these findings sug-
gest that a diagnosis of Wilsons disease could Dr. Joseph Misdraji: On examination of the ex-
be made with 100% sensitivity and specificity. It planted liver, the cut surface was mottled and
has been previously noted that viral infection or had a subtle nodular appearance, with scattered
drug toxicity may serve as a trigger for fulmi- regenerative nodules that varied in color from

21

A B
C D
E F
Figure 3. Explanted Liver.

A photograph of the cut surface of the liver shows a mottled appearance, with several regenerative nodules that
vary in color from tannish-brown to yellow (Panel A). Hematoxylin and eosin staining shows that the parenchyma
has a nodular appearance and widespread inflammation (Panel B). At higher magnification, the hepatocytes have
relatively abundant eosinophilic cytoplasm and cytoplasmic vacuolization at the periphery of the cell (Panel C).
Several central veins have endophlebitis, with mononuclear infiltrates in the endothelium of the veins (Panel D).
A trichrome stain for collagen highlights a portal tract with pale, bluish-gray septa that are indicative of collapse
and fibrosis (Panel E). A rhodanine stain for copper shows copper deposition (reddish granules) in hepatocytes
in a nodule (Panel F).

22

tannish-brown to yellow (Fig. 3A). On micro- undergo transplantation.22 Wilsons disease is

scopic examination, the parenchyma was nodu- the only cause of acute liver failure that allows a
lar and had moderate inflammation, and some patient with preexisting liver disease to be listed
nodules were steatotic (Fig. 3B). At higher mag- as status 1A. Therefore, making this diagnosis
nification, the hepatocytes had abundant eosin- is essential, as was shown in this case.
ophilic cytoplasm, and many of them showed This patient received a good-quality liver do-
vacuolization of the cytoplasm (Fig. 3C). Portal nated from a deceased 23-year-old man who had
tracts showed mononuclear inflammation and been declared brain dead because of penetrating
occasional plasma cells. Several hepatic veins head trauma and hemorrhagic shock. We per-
showed endophlebitis, with mononuclear cells formed liver transplantation with the use of the
infiltrating the endothelium (Fig. 3D); this par- piggyback technique, in which the surgeon
ticular feature has been described only rarely in performs hepatectomy with preservation of the
Wilsons disease.18 A trichrome stain highlighted inferior vena cava and then performs anastomo-
septa, which did not have the dense, blue stain- sis to attach the donors suprahepatic inferior
ing typical of established cirrhosis but rather had vena cava to the recipients inferior vena cava at
pale, bluish-gray staining, which is suggestive of the level of the hepatic veins. The graft cold
fibrosis and collapse (Fig. 3E). A rhodanine stain ischemic time was 5 hours, and the graft warm
for copper was performed, and copper was iden- ischemic time was 27 minutes. Total blood loss
tified in hepatocytes in a few nodules (Fig. 3F). was 800 ml. The patient was extubated on day 1
The pathological diagnosis of Wilsons disease after transplantation and transferred to the in-
is generally based on the presence of compatible patient transplant unit the following day. She
histomorphologic features and results of stain- received maintenance therapy with a triple im-
ing for copper, including a rhodanine stain. How- munosuppression regimen (tacrolimus, mycophe-
ever, staining for copper in tissue is unreliable, nolate mofetil, and prednisone) and was dis-
since the presence of copper in the cytoplasm of charged home on day 9 after transplantation.
hepatocytes might not be detected on a rhodanine Since the transplantation, the patient has
stain.19 Therefore, in patients with suspected Wil- done well, with the exception of three episodes
sons disease, copper quantification performed of allograft dysfunction; the first was relatively
on either a dedicated core-biopsy specimen or a minor and occurred during an upper respiratory
paraffin-embedded tissue sample is considered tract infection, and the second and third were
to be the best available diagnostic test.20 A value due to acute rejection during the administration
of 30 g per gram of dry weight is normal, and of subtherapeutic tacrolimus levels 11 and 15
a diagnosis of Wilsons disease is highly likely months after transplantation. The two episodes
when the value exceeds 250 g per gram. In this of rejection resolved after treatment with intra-
case, copper quantification was performed on a venous methylprednisolone and an increase in
formalin-fixed, paraffin-embedded tissue sample immunosuppression maintenance therapy. The
from the explanted liver, and a value of 978 g outcomes associated with liver transplantation
per gram was obtained, confirming the diagno- for acute liver failure induced by Wilsons dis-
sis of Wilsons disease. ease are excellent, if transplantation is per-
formed prior to neurologic deterioration.23-25
M a nagemen t a nd Fol l ow-up
A nat omic a l Di agnosis
Dr. Nahel Elias: Identifying the cause of acute liver
failure in this patient was critical in determining Fulminant hepatic failure due to Wilsons disease.
her place on the liver transplantation list.21 In This case was presented at the Internal Medicine Case Con-
this country, the highest priority (United Net- ference.
work for Organ Sharing status 1A) is reserved No potential conflict of interest relevant to this article was
reported.
for patients with liver failure who have a life Disclosure forms provided by the authors are available with
expectancy of less than 7 days if they do not the full text of this article at NEJM.org.

23

References
1. Lee WM, Squires RH Jr, Nyberg SL, 9. Czaja AJ. Diagnosis and management plantation in adult patients with Wilsons
Doo E, Hoofnagle JH. Acute liver failure: of autoimmune hepatitis. Clin Liver Dis disease. Liver Transpl 2007;13:55-61.
summary of a workshop. Hepatology 2008; 2015;19:57-79. 18. Stromeyer FW, Ishak KG. Histology of
47:1401-15. 10. Krawitt EL. Autoimmune hepatitis. the liver in Wilsons disease: a study of 34
2. Squires RH Jr. Acute liver failure in N Engl J Med 2006;354:54-66. cases. Am J Clin Pathol 1980;73:12-24.
children. Semin Liver Dis 2008;28:153-66. 11. Tanaka S, Okamoto Y, Yamazaki M, 19. Johncilla M, Mitchell KA. Pathology
3. Squires RH, Alonso EM. Acute liver Mitani N, Nakqjima Y, Fukui H. Signifi- of the liver in copper overload. Semin
failure in children. In: Suchy FJ, Sokol RJ, cance of hyperglobulinemia in severe Liver Dis 2011;31:239-44.
Balistreri WF, eds. Liver disease in chil- chronic liver diseases with special ref- 20. Ludwig J, Moyer TP, Rakela J. The liver
dren. 4th ed. New York: Cambridge Uni- erence to the correlation between serum biopsy diagnosis of Wilsons disease:
versity Press, 2012:32-50. globulin/IgG level and ICG clearance. methods in pathology. Am J Clin Pathol
4. Banti S, Mauri M, Oppo A, et al. From Hepatogastroenterology 2007;54:2301-5. 1994;102:443-6.
the third month of pregnancy to 1 year 12. Ferenci P, Caca K, Loudianos G, et al. 21. Reddy KR, Ellerbe C, Schilsky M, et al.
postpartum: prevalence, incidence, recur- Diagnosis and phenotypic classification Determinants of outcome among patients
rence, and new onset of depression of Wilson disease. Liver Int 2003;23:139- with acute liver failure listed for liver
results from the Perinatal Depression 42. transplantation in the United States. Liver
Research & Screening Unit Study. Compr 13. Korman JD, Volenberg I, Balko J, et al. Transpl 2016;22:505-15.
Psychiatry 2011;52:343-51. Screening for Wilson disease in acute liver 22. Organ Procurement and Transplanta-
5. Hu J, Zhang Q, Ren X, Sun Z, Quan Q. failure: a comparison of currently avail- tion Network. Policy 9: allocation of livers
Efficacy and safety of acetylcysteine in able diagnostic tests. Hepatology 2008;48: and liverintestines. Washington, DC: De-
non-acetaminophen acute liver failure: 1167-74. partment of Health and Human Services,
a meta-analysis of prospective clinical tri- 14. Hadem J, Tacke F, Bruns T, et al. Eti- 2016 (https://optn.transplant .hrsa.gov/
als. Clin Res Hepatol Gastroenterol 2015; ologies and outcomes of acute liver failure governance/policies/).
39:594-9. in Germany. Clin Gastroenterol Hepatol 23. Eghtesad B, Nezakatgoo N, Geraci
6. Chalasani N, Bonkovsky HL, Fontana 2012;10(6):664-9.e2. LC, et al. Liver transplantation for Wilsons
R, et al. Features and outcomes of 899 15. Roberts EA, Schilsky ML. Diagnosis disease: a single-center experience. Liver
patients with drug-induced liver injury: and treatment of Wilson disease: an up- Transpl Surg 1999;5:467-74.
the DILIN Prospective Study. Gastroen- date. Hepatology 2008;47:2089-111. 24. Guillaud O, Dumortier J, Sobesky R,
terology 2015;148(7):1340-52.e7. 16. Dhawan A, Taylor RM, Cheeseman P, et al. Long term results of liver transplan-
7. Bacak SJ, Thornburg LL. Liver failure De Silva P, Katsiyiannakis L, Mieli-Vergani tation for Wilsons disease: experience in
in pregnancy. Crit Care Clin 2016;32:61-72. G. Wilsons disease in children: 37-year ex- France. J Hepatol 2014;60:579-89.
8. Ren W, Li X, Jia J, Xia Y, Hu F, Xu Z. perience and revised Kings score for liver 25. Medici V, Mirante VG, Fassati LR, et al.
Prevalence of Budd-Chiari syndrome dur- transplantation. Liver Transpl 2005;11: Liver transplantation for Wilsons disease:
ing pregnancy or puerperium: a systematic 441-8. the burden of neurological and psychiatric
review and meta-analysis. Gastroenterol 17. Petrasek J, Jirsa M, Sperl J, et al. Re- disorders. Liver Transpl 2005;11:1056-63.
Res Pract 2015;2015:839875. vised Kings College score for liver trans- Copyright 2017 Massachusetts Medical Society.
Lantern SLideS Updated: CompLete powerpoint SLide SetS from the CLiniCopathoLogiCaL ConferenCeS
Any reader of the Journal who uses the Case Records of the Massachusetts General Hospital as a teaching exercise or reference
material is now eligible to receive a complete set of PowerPoint slides, including digital images, with identifying legends,
shown at the live Clinicopathological Conference (CPC) that is the basis of the Case Record. This slide set contains all of the
images from the CPC, not only those published in the Journal. Radiographic, neurologic, and cardiac studies, gross specimens,
and photomicrographs, as well as unpublished text slides, tables, and diagrams, are included. Every year 40 sets are produced,
averaging 50-60 slides per set. Each set is supplied on a compact disc and is mailed to coincide with the publication of the
Case Record.
The cost of an annual subscription is $600, or individual sets may be purchased for $50 each. Application forms for the current
subscription year, which began in January, may be obtained from the Lantern Slides Service, Department of Pathology,
Massachusetts General Hospital, Boston, MA 02114 (telephone 617-726-2974) or e-mail Pathphotoslides@partners.org.

24

CLINICAL PROBLEM-SOLVING
Appearing in the first issue of each month, this Journal feature presents particulars about real patients in stages
to experts, who respond to the information, sharing their clinical reasoning with the reader.
25

Clinical Problem-Solving
Caren G. Solomon, M.D., M.P.H., Editor
Making the Connection

Allison R. Schulman, M.D., Ali Tavakkoli, M.D., Christopher C. Thompson, M.D.,
Amy L. Miller, M.D., Ph.D., and Joseph Loscalzo, M.D., Ph.D.
In this Journal feature, information about a real patient is presented in stages (boldface type) to an expert
clinician, who responds to the information, sharing his or her reasoning with the reader (regular type).
The authors commentary follows.
From Departments of Medicine (A.R.S., A 41-year-old man with a weight of 159 kg and a body-mass index (BMI; the weight
C.C.T., A.L.M., J.L.) and Surgery (A.T.), in kilograms divided by the square of the height in meters) of 49.1 presented for con-
Brigham and Womens Hospital, Harvard
Medical School, Boston. Address reprint sideration of bariatric surgery. He had been morbidly obese since childhood; he had
requests to Dr. Miller at the Department tried several commercial weight-loss programs in addition to dieting on his own but
of Medicine, Brigham and Womens Hos- had had little long-term success.
pital, 75 Francis St., Boston, MA 02115, or
at almiller@partners.org.
Bariatric surgical procedures are a well-established approach to the treatment of
N Engl J Med 2017;376:476-82.
DOI: 10.1056/NEJMcps1602495 morbid obesity, offering sustainable weight loss and a reduction in the risk of
Copyright 2017 Massachusetts Medical Society. conditions related to obesity. Candidacy is stratified according to BMI. Adults with
a BMI of 40 or higher are potential candidates for the procedure. Patients with a
BMI of 35.0 to 39.9 are generally considered to be eligible if they have at least one
serious coexisting condition, such as obstructive sleep apnea, type 2 diabetes, or
hypertension.
The patient had a history of hyperlipidemia and had undergone cholecystectomy. His
medications included rosuvastatin and ezetimibe. He had no known drug allergies.
He lived with his wife and three children and worked in information technology. He
was a current smoker, with a 30-year history of one to three packs per day. He had a
remote history of excessive alcohol use and had been abstinent for the past 15 years.
He reported no illicit drug use. He had no known family history of gastrointestinal
disorders. After consultation, he decided to undergo Roux-en-Y gastric bypass.
Roux-en-Y gastric bypass is one of the most common bariatric surgical treatments
for obesity. The procedure involves the creation of a small upper gastric pouch (with
a capacity of approximately 30 ml), stapling of the stomach, and division of the small
intestine at the proximal jejunum, with anastomosis of the distal portion (the ali-
mentary, or Roux, limb) to the gastric pouch. The proximal end of the divided jeju-
num (the biliopancreatic limb) is anastomosed farther down the jejunum (Fig. 1).
Pancreatic and biliary secretions come into contact with food below this anasto-
mosis, in the common channel of the small intestine. Although the initial the-
ory was that Roux-en-Y gastric bypass led to weight loss through mechanical re-
striction of the upper gastric pouch and a reduction in the surface area for nutrient
absorption, it is now known that the surgery also causes neurohormonal changes
that may influence metabolism and weight.
476 n engl j med26

376;5 nejm.org February 2, 2017

The patient quit smoking several months before

surgery. He underwent an uncomplicated lapa-
Gastric
roscopic Roux-en-Y gastric bypass that included pouch
the creation of an 80-cm Roux limb. In the first
year after surgery, he lost more than 80% of his
excess weight, with his total weight declining to
83.9 kg. After that first postoperative year, the
patient had some weight gain, which prompted
him to resume smoking. His weight eventually sta- Common
Gallbladder bile duct
bilized at 95.3 kg (BMI, 29.4).
Pancreas
Some weight regain after Roux-en-Y gastric bypass
is not uncommon and may result from progressive
nonadherence to the prescribed diet or from
anatomical changes, including dilation of the
gastrojejunal anastomosis or enlargement of the
gastric pouch. Gastrogastric fistula, a complica-
tion seen more commonly after open Roux-en-Y
gastric bypass than after laparoscopic procedures, Roux limb
is another potential cause of weight regain. In

this case, weight gain may occur because the chan- Biliopancreatic
limb
nel between the gastric pouch and the excluded Food Digestive
remnant stomach facilitates the passage of food juices
through the previously bypassed foregut. Exces-
sive weight regain in patients with altered anat-
omy may require consideration of another inter-
vention, after appropriate steps are taken to
re-educate the patient with regard to appropriate
dietary choices.
Common
Eight years after undergoing Roux-en-Y gastric channel
bypass, the patient presented to his primary care
physician with pain in the epigastric region and
the left upper quadrant of the abdomen. The pa-
tient rated the pain as 5 on a scale of 1 to 10 (with
Figure 1. Schematic Representation of Procedure for Roux-en-Y Gastric
higher scores indicating more severe pain), and he Bypass.
said that the pain was exacerbated by eating. He During Roux-en-Y gastric bypass, the gastric pouch is restricted and is
had no melena or hematochezia. The physical ex- anastamosed to the distal jejunum (Roux limb). The distal stomach and
amination was notable for abdominal tenderness, biliopancreatic limb of the jejunum are also anastamosed, thereby creating
without rebound or guarding. The results of a basic a common channel.
metabolic panel, blood count, and liver-function
tests were unremarkable.
sis, which can be triggered by rapid weight loss
There are several potential causes of abdominal after Roux-en-Y gastric bypass; and disorders of
pain in patients who have undergone Roux-en-Y the remnant stomach.
gastric bypass. These include internal hernias,
which often occur at defects in the mesentery Computed tomography of the abdomen revealed
created at the time of bypass surgery; ulceration no evidence of an internal hernia but did show
of the gastrojejunal anastomosis, which can re- inflammation of the gastrojejunal anastomosis.
sult when acid injures the jejunum; cholelithia- An ulceration on the jejunal side of the surgical
n engl j med 376;5 nejm.org 27February 2, 2017 477

cific dietary triggers. Episodes were associated

with mild urgency but no tenesmus or fecal in-
continence. The stool was pale, without melena or
hematochezia, often floated, and left a residue
after flushing. The patient also noted an increase
in flatus and daily feculent eructation.
The symptoms are suggestive of a malabsorptive

process, which is typically manifested in pale,
greasy, voluminous, foul-smelling stools that float
and are difficult to flush, in addition to weight
loss despite adequate food intake. In this spe-
cific case a secondary osmotic diarrhea is likely,
with the poorly absorbed solutes causing water
retention in the bowel. Osmotic diarrhea resolves
Figure 2. Endoscopic View of the Jejunal Aspect with fasting and thus does not typically occur
of the Gastrojejunal Anastomosis, Showing Ulceration. overnight. Disorders that may cause malabsorp-
A marginal ulceration can be seen (arrows). tion after Roux-en-Y gastric bypass include bac-
terial overgrowth in the small intestine and en-
anastomosis was identified on esophagogastro- teroenteric fistula; the latter limits the surface
duodenoscopy (Fig. 2). Suture material was re- area for intestinal absorption and is also known
moved from the site to prevent further irritation of to cause feculent eructation.
the mucosa. Although enteroenteric fistulas are rare and
are most commonly associated with inflamma-
High-dose proton-pump inhibitors are the main- tory bowel disease, ulceration causes a predispo-
stay of therapy and should be initiated, and the sition to their formation. Other causes of malab-
patients nutrition status should also be re- sorption in the general population (which can also
viewed to make sure that it will allow for ade- be present after bariatric surgery) include new
quate tissue healing. Discontinuation of smoking, lactose intolerance, chronic pancreatitis, celiac
avoidance of nonsteroidal antiinflammatory drugs disease, bile-salt deficiency, and chronic infections.
(NSAIDs), and the removal of endoscopic sutures
or staples at or near the site of marginal ulcer- The patient did not initially pursue further evalu-
ation are associated with an increased likelihood ation, but after 4 months of persistent symptoms,
of successful healing. With appropriate treatment, including progressive generalized weakness, my-
these ulcers heal in the majority of patients. algias, and muscle cramps, he presented to the
emergency department for evaluation. His weight
The patient was treated with 40 mg of omeprazole had decreased from 95.3 kg to 72.6 kg.
twice daily, but despite excellent adherence to the
regimen, his abdominal pain did not abate. Su- This patient has chronic diarrhea, as defined by
cralfate, 1 g per 10 ml administered four times per persistently loose stools for more than 4 weeks,
day, was added as adjunctive therapy, but serial occurring at an average of more than three times
repeat endoscopies conducted at 2-month inter- per day. His associated symptoms suggest clini-
vals revealed that the ulcer was persistent. Within cally significant electrolyte abnormalities, which
6 months after the diagnosis of marginal ulcer- should be addressed expeditiously.
ation, the patient suddenly began to have foul- Biopsy of the colonic mucosa should be pur-
smelling, watery diarrhea six to seven times a day, sued. Normal-appearing mucosa will be observed
without nocturnal diarrhea. These episodes were in many processes that cause chronic diarrhea,
most voluminous after meals and often had visi- which makes histologic examination an essential
ble particulate matter that was consistent with component of the evaluation. Although this ex-
what he had recently eaten. There were no spe- amination could be accomplished by means of
478 n engl j med 376;5 nejm.org February 2, 2017

28

flexible sigmoidoscopy, the fecal eructation is sug- Tests for antibodies to tissue transglutaminase
gestive of an enteroenteric fistula, which makes IgA and the human immunodeficiency virus were
colonoscopy a better choice, since it allows visu- negative; the total IgA level was normal. Stool
alization of the entire colon. In general, upper studies were negative for bacterial infection and
endoscopic evaluation should be considered in for ova and parasites. The fecal fat level was elevat-
patients with chronic diarrhea if the lower endo- ed, and fecal leukocytes were rare.
scopic evaluation is unrevealing. This recommen-
dation is particularly appropriate in this case given Whereas excessive loose, sagging skin can be seen
the possibility of an enteroenteric fistula; if the after substantial weight loss, the diminished skin
fistula involves the upper gastrointestinal tract, turgor and dry mucous membranes in the patient
it may be more easily visualized through upper indicate interstitial volume depletion. The severe
endoscopy. Breath testing for small intestinal bac- hypokalemia and hypomagnesemia are conse-
terial overgrowth should also be considered. quences of his profound diarrhea. The low levels
of albumin and prealbumin suggest that he is se-
The patients temperature was 37.0C, pulse 97 beats verely malnourished. His anemia may reflect poor
per minute, blood pressure 113/69 mm Hg, respi- iron absorption or could be due to gastrointesti-
ratory rate 16 breaths per minute, and oxygen nal blood loss that was not detected on a single
saturation 99% while he was breathing ambient guaiac test. The prolongation of prothrombin
air. He appeared to be chronically ill and cachectic time is probably due to the depletion of vitamin
but was not in acute distress. The oropharynx was Kdependent coagulation factors, and the elevat-
clear, with dry mucous membranes and feculent ed level of C-reactive protein, although nonspe-
breath. His thyroid was not palpable. The cardiac cific, suggests inflammation. Although tissue
and pulmonary examinations were normal. His levels of antibodies to transglutaminase IgA have
abdomen was nondistended, with minimal ten- high sensitivity and specificity for celiac disease
derness to deep palpation in the epigastrum and in the presence of normal total IgA levels, nor-
left upper quadrant. Excessive loose, sagging skin mal findings do not definitively rule out celiac
was noted around his midabdomen. The skin disease. However, celiac disease would not ac-
turgor in his legs was diminished. The results of count for the feculent eructation in this patient.
neurologic and rectal examinations were unremark- Fecal fat excretion can be increased in diarrheal
able, and a stool test for guaiac was negative. illnesses even in the absence of fat malabsorp-
The potassium level was 2.2 mmol per liter, tion. The standard for the diagnosis of increased
magnesium 1.6 mg per deciliter (0.7 mmol per li- fecal fat excretion in patients with malabsorption
ter; reference range, 1.7 to 2.6 mg per deciliter is quantitative measurement with a biochemical
[0.7 to 1.1 mmol per liter]), and phosphorus 1.5 mg assay. Normal excretion is less than 6 g per day,
per deciliter (0.5 mmol per liter; reference range, assuming that the patient is consuming 70 to
2.4 to 4.3 mg per deciliter [0.8 to 1.4 mmol per li- 120 g per day of dietary fat. Collection periods
ter]). Sodium and chloride levels were normal. of 3 days or longer increase the sensitivity of
The blood urea nitrogen level was 5 mg per decili- detection.
ter (1.8 mmol per liter) and creatinine 0.85 mg per
deciliter (75 mol per liter). The hematocrit was The patient was admitted to the hospital for hydra-
35.4%. The white-cell count and platelet count tion and electrolyte supplementation, diagnostic
were normal. The thyrotropin level was 2.96 mIU testing, and monitoring. After hydration and elec-
per liter. The level of albumin was 1.5 g per decili- trolyte repletion, a colonoscopy revealed a small
ter, and prealbumin 9.3 mg per deciliter (reference area of erythema near the hepatic flexure; no other
range, 20 to 40). The erythrocyte sedimentation abnormalities were noted. Biopsy specimens of the
rate was 16 mm per hour (reference range, 0 to 13), colon and terminal ileum were normal. Upper
and the C-reactive protein level 13.4 mg per liter endoscopy revealed normal esophageal mucosa,
(reference range, 0 to 5). The prothrombin time with marginal ulceration and severe inflamma-
was elevated at 15.9 seconds (reference range, tion at the gastrojejunal anastomosis; when pres-
12.0 to 14.4). Liver-function tests were normal. sure was exerted on one of the areas of inflamma-
n engl j med 376;5 nejm.org February 2, 2017 479

29

Esophagus
Gastric pouch
Fistula
Transverse
colon
Fistula
Descending
colon
Figure 3. Upper Gastrointestinal Series.

Contrast material can be seen traveling down the esophagus and through the gastric pouch and fistulous tract to
the transverse colon, from which it traverses the descending colon. No contrast material is seen within the small
bowel or the proximal large bowel, findings that are diagnostic of a gastrocolic fistula.
tion, the endoscope passed through a fistulous anatomy of the gastrointestinal tract. Although
tract, terminating in the transverse colon. surgical repair is indicated, the patients poor
nutritional status makes him a poor candidate
The endoscopy revealed an enteroenteric fistula. for such a procedure.
Symptoms vary depending on the location of the
fistula and the extent of the bowel that is by- Total parenteral nutrition was provided for 8 weeks
passed. Enteroenteric fistulas in which only a before the endoscopic removal of the gastrocolic
short segment of the bowel is circumvented may fistula and the revision of the gastrojejunal anas-
be asymptomatic. Conversely, fistulas that bypass tomosis and gastric pouch. After surgery, the pa-
a larger segment of the bowel typically cause di- tient began to have normal formed bowel move-
arrhea but may also be associated with weight ments and transitioned back to oral nutrition.
loss or abdominal pain. Gastrocolic fistula is a Complete resolution of the fistulous connection
well-recognized but rare complication of peptic was confirmed on a repeat upper gastrointestinal
ulcer disease, stomach and colon cancer, Crohns series. At follow-up 5 months after surgery, the
disease, and surgical procedures. The combina- patient had returned to a weight of 95.3 kg.
tion of feculent eructation and vomiting, diar-
rhea, and malnutrition is the classic symptom C om men ta r y
complex.
This patient initially presented with abdominal
An upper gastrointestinal series showed contrast pain, for which the differential diagnosis was
material traversing the gastric pouch into a fistu- broad given his history of gastric bypass surgery.
lous tract that terminated in the transverse colon. The marginal ulceration identified at that time
The contrast material did not pass into the small failed to heal and was eventually complicated by
bowel or the proximal large bowel (Fig. 3). a profound malabsorptive diarrhea, with resul-
tant dehydration. Although dumping syndrome
These findings confirm the diagnosis of a gas- and bacterial overgrowth in the small intestine
trocolic fistula and also help to delineate the are common causes of diarrhea after gastric by-
480 n engl j med 376;5 30

nejm.org February 2, 2017

pass, other causes, including enteroenteric fis- ing to oversew the ulceration has been proposed
tula, must be considered in this patient popula- in the treatment of recalcitrant marginal ulcer-
tion, especially when fecal eructation is present. ation, but more information regarding efficacy
These patients should therefore undergo anatomi- and durability is needed.12
cal evaluation before their symptoms are attrib- Fistulous complications have also been re-
uted to the more common causes. In this case, a ported after Roux-en-Y gastric bypass and are
full diagnostic workup ultimately led to the diag- often associated with a chronic leak from the
nosis of gastrocolic fistula. surgical anastomosis or with a marginal ulcer
The increasing prevalence of severe obesity that does not heal. Gastrogastric fistulas
and associated conditions has resulted in a sub- which occur when a channel develops between
stantial increase in the number of bariatric pro- the gastric pouch and the excluded stomach
cedures performed.1-5 Roux-en-Y gastric bypass is remnant and allows ingested food to enter the
one of the most commonly performed weight-loss bypassed foregut are most commonly associ-
surgeries. Outcomes of bariatric procedures have ated with Roux-en-Y gastric bypass procedures
improved substantially in the past decade, but that are performed with an open approach, al-
complications can occur; vigilance and thorough though fistulas may also occur after laparoscopic
assessment of patients presenting with symptoms procedures. Gastrocolic fistulas are a rare com-
is required. plication of the formation of a gastrojejunosto-
The cause of marginal ulceration is usually my and are typically initiated by ulceration at the
multifactorial.6-8 Poor perfusion of local tissue is site of the anastomosis. This type of fistula for-
believed to contribute to the occurrence of ul- mation has been reported as late as 20 years after
cers, which can also be caused by excessive acid the surgical Roux-en-Y gastric bypass has been
exposure in the gastric pouch; analysis of a performed.13 Depending on the size and location
consecutive series of patients suggests that for- of the fistula, consideration of endoscopic clo-
eign materials (e.g., sutures or staples), a history sure (including endoscopic suturing or over-the-
of smoking, and the use of NSAIDs are associ- scope clips) or surgical closure should be consid-
ated with an increased incidence of marginal ered. In rare instances, other types of fistulas
ulceration.9 The potential role of Helicobacter py- have been reported after gastric bypass surgery,
lori infection as a predisposing factor remains including gastrobronchial, gastropericardial, and
controversial. jejunocolic fistulas.14
Retrospective data support the view that the The current case underscores the importance
majority of patients with marginal ulcerations have of considering late complications of bariatric sur-
a response to medical therapy with high doses of gery as causes of new symptoms in patients who
proton-pump inhibitors and sucralfate8; the ad- have had this surgery. Although gastrocolic fis-
dition of sucralfate has been recommended in tula is an uncommon complication of Roux-en-Y
patients who are already taking a proton-pump gastric bypass, the combination of malabsorptive
inhibitor at the time of diagnosis.10 In addition, diarrhea, fecal eructation, and previous marginal
smoking cessation and indefinite discontinua- ulceration provided important clues that pointed
tion of NSAIDs are recommended. The removal to this complication.
of foreign material and testing for H. pylori (as well
as treatment, if test results are positive) should Disclosure forms provided by the authors are available with
also be considered.11 The use of endoscopic sutur- the full text of this article at NEJM.org.
References
1. Buchwald H, Avidor Y, Braunwald E, 3. Colquitt J, Clegg A, Loveman E, Royle 5. Buchwald H, Oien DM. Metabolic/
et al. Bariatric surgery: a systematic re- P, Sidhu MK. Surgery for morbid obesity. bariatric surgery worldwide 2011. Obes
view and meta-analysis. JAMA 2004;292: Cochrane Database Syst Rev 2005;4: Surg 2013;23:427-36.
1724-37. CD003641. 6. Azagury DE, Abu Dayyeh BK, Green-
2. Arterburn DE, Olsen MK, Smith VA, et 4. Perry CD, Hutter MM, Smith DB, Ne- walt IT, Thompson CC. Marginal ulceration
al. Association between bariatric surgery whouse JP, McNeil BJ. Survival and chang- after Roux-en-Y gastric bypass surgery:
and long-term survival. JAMA 2015;313: es in comorbidities after bariatric sur- characteristics, risk factors, treatment, and
62-70. gery. Ann Surg 2008;247:21-7. outcomes. Endoscopy 2011;43:950-4.
n engl j med 376;5 nejm.org February 2, 2017 481

31

7. Dallal RM, Bailey LA. Ulcer disease 10. Carr WR, Mahawar KK, Balupuri S, device to treat recalcitrant marginal ul-
after gastric bypass surgery. Surg Obes Small PK. An evidence-based algorithm ceration (with video). Gastrointest Endosc
Relat Dis 2006;2:455-9. for the management of marginal ulcers 2012;76:435-9.
8. Rasmussen JJ, Fuller W, Ali MR. Mar- following Roux-en-Y gastric bypass. Obes 13. Fronza JS, Martin JA, Nagle AP. Jeju-
ginal ulceration after laparoscopic gastric Surg 2014;24:1520-7. nocolic fistula after Roux-en-Y gastric
bypass: an analysis of predisposing fac- 11. Steinemann DC, Bueter M, Schiesser bypass. Surg Obes Relat Dis 2012;8(5):
tors in 260 patients. Surg Endosc 2007;21: M, Amygdalos I, Clavien PA, Nocito A. e60-2.
1090-4. Management of anastomotic ulcers after 14. Sakran N, Assalia A, Keidar A, Goit-
9. Coblijn UK, Lagarde SM, de Castro SM, Roux-en-Y gastric bypass: results of an ein D. Gastrobronchial fistula as a com-
Kuiken SD, van Wagensveld BA. Symptom- international survey. Obes Surg 2014;24: plication of bariatric surgery: a series of
atic marginal ulcer disease after Roux-en-Y 741-6. 6 cases. Obes Facts 2012;5:538-45.
gastric bypass: incidence, risk factors and 12. Jirapinyo P, Watson RR, Thompson Copyright 2017 Massachusetts Medical Society.
management. Obes Surg 2015;25:805-11. CC. Use of a novel endoscopic suturing
clinical problem-solving series

The Journal welcomes submissions of manuscripts for the Clinical Problem-Solving
series. This regular feature considers the step-by-step process of clinical decision
making. For more information, please see authors.NEJM.org.
482 n engl j med 376;5 nejm.org February 2, 2017

32

Back to the History

Mary W. Montgomery, M.D., Sigal Yawetz, M.D., Bruce D. Levy, M.D.,
and Joseph Loscalzo, M.D., Ph.D.
In this Journal feature, information about a real patient is presented in stages (boldface type) to an expert
clinician, who responds to the information, sharing his or her reasoning with the reader (regular type).
The authors commentary follows.
An 82-year-old man presented to the emergency department with a 6-month history From the Department of Medicine,
of worsening back and left hip pain. He had been well until 6 months earlier, when Brigham and Womens Hospital and Har-
vard Medical School, Boston. Address
acute fever and cough developed. Pneumonia was diagnosed, and a 10-day course of reprint requests to Dr. Levy at Brigham
levofloxacin was prescribed. His respiratory symptoms resolved. At about the same and Womens Hospital, Harvard Insti-
time that the respiratory symptoms resolved, low back pain began to develop. The tutes of Medicine Building, 77 Ave. Louis
Pasteur (HIM855), Boston, MA 02115, or
patient did not recall any antecedent trauma, fall, or heavy lifting. Over the course of at blevy@partners.org.
the ensuing months, his back pain continued to worsen, and left hip pain developed;
N Engl J Med 2017;376:1783-8.
he presented to the emergency department after walking became painful. The pain DOI: 10.1056/NEJMcps1607608
did not radiate down his legs, but it occasionally woke him from sleep and was worse Copyright 2017 Massachusetts Medical Society.
after activity. He reported no relief from bending forward. He reported no morning

stiffness, fever, chills, or night sweats. He noted that he had lost 9 kg (20 lb) of weight
over the course of the preceding few months. He had no bowel or bladder inconti-
nence.
This patient has chronic (duration of >12 weeks) back pain. Demographic and
clinical features, including age, the presence of constitutional symptoms, pain awak-
ening him from sleep, and persistence and worsening of the pain, indicate the need
for imaging. This history arouses concern for an underlying systemic illness, such
as metastatic cancer, a systemic infection, or an autoimmune process. Polymyalgia
rheumatica can cause hip pain and systemic symptoms such as weight loss, but it
is associated with morning stiffness, particularly in the shoulders and hips.
The patients medical history included the sick sinus syndrome for which a pace-
maker had been placed years earlier, diastolic heart failure, pulmonary fibrosis, hyper-
lipidemia, hypertension, gout, herpes zoster with postherpetic neuralgia, and basal-
cell carcinoma for which the patient underwent Mohs micrographic surgery. His
medications included allopurinol, aspirin, metoprolol, simvastatin, gabapentin, and
recently, oxycodone and acetaminophen for the low back pain. He had no known
drug allergies. He was retired but had worked previously as a university professor.
He lived with his wife and had grown children. He had smoked cigars for many years
but had quit 28 years before the current presentation. He reported drinking one glass
of alcohol per night until 6 months before this presentation, when his current illness
began. He reported no history of illicit-drug use. He had traveled internationally in
the distant past, including trips to Asia and South America, and had traveled
throughout the United States. His family history was notable for congestive heart
failure in his mother, coronary heart disease in his father, and non-Hodgkins lym-
phoma in his brother.
n engl j med 376;18 nejm.org33May 4, 2017 1783

The most important risk factor for metastatic does not rule out vertebral osteomyelitis since
cancer to the spine is a personal history of cancer. only one quarter of patients with vertebral osteo-
The patient has a history of basal-cell carcinoma, myelitis present with this finding. Focal tender-
but this type of cancer rarely causes distant metas- ness on examination can also be absent in early
tases and is unlikely to be related to his current metastatic spinal disease. Spondyloarthropathy is
presentation. He has a remote smoking history, unlikely, given his age, the normal hip examina-
which puts him at risk for lung cancer. His age tion, and his history of worsening pain with
is a risk factor for prostate cancer, which can activity (rather than with rest).
metastasize to the vertebrae. He has had no re-
cent device implantation, surgery, or trauma that The white-cell count was 8000 per cubic milli-
could have introduced infection. His previous meter, with 87% neutrophils, 8% lymphocytes,
travel to Asia and South America could put him 2% monocytes, and 0% eosinophils. The hemato-
at risk for tuberculous spondylitis. crit level was 39.1%, and the platelet count was
231,000 per cubic millimeter. A comprehensive
On physical examination, he appeared to be in no metabolic panel and coagulation studies were
acute distress. His temperature was 37.1C (98.7F), normal. The erythrocyte sedimentation rate was
the blood pressure 110/60 mm Hg, the pulse 64 mm per hour, and the C-reactive protein level
60 beats per minute, the respiratory rate 16 breaths was 47.9 mg per liter (normal range, 1 to 3).
per minute, and the oxygen saturation 98% while
he was breathing ambient air. The oropharynx The elevations in erythrocyte sedimentation rate
was clear, without thrush or oral ulcerations. The and C-reactive protein are of concern and sug-
lungs were clear to auscultation. A cardiovascular gest an inflammatory, infectious, or neoplastic
examination revealed a regular rate and rhythm cause of the back pain. These tests have high
without murmurs, rubs, or gallops. A pacemaker sensitivity in vertebral osteomyelitis and spinal
generator was palpated over the left pectoralis metastatic disease but are nonspecific. The nor-
muscle, and no erythema, fluctuance, or tender- mal white-cell count does not rule out vertebral
ness was noted. The abdomen was soft and non- osteomyelitis, since more than one third of pa-
tender, with normal bowel sounds and without tients with this condition present without leuko-
hepatosplenomegaly. There was no palpable cytosis. When systemic disease is suspected, as in
cervical, axillary, or inguinal lymphadenopathy. this case, the preferred imaging method is mag-
A neurologic examination revealed that the patient netic resonance imaging (MRI), unless a contra-
was oriented to person, place, and time. Cranial indication exists. Pacemakers are a relative con-
nerves II to XII were normal. He had no muscle traindication to MRI, although some devices are
wasting in his legs, and the muscle strength in his MRI-compatible.
legs was 5/5 both proximally and distally. The
patellar reflexes were slightly diminished in both Because the patient had a pacemaker, MRI was
knees. A sensory examination showed that sensa- not performed. Computed tomography (CT) of the
tions to light touch and vibration were intact. No lumbar spine with contrast enhancement revealed
saddle anesthesia was present, and the rectal tone end-plate irregularity and destruction of the L4
was normal. He had no point tenderness over any and L5 vertebrae, with a surrounding left paraver-
portion of his spine or his sacroiliac joints. The tebral and ventral epidural soft-tissue collection
straight-leg raising maneuver revealed no radicu- extending into the right foramen. There was no
lar symptoms in either leg. A bilateral hip exami- evidence of nerve-root compression (Fig. 1).
nation, including flexion, internal rotation, and
external rotation, was normal. No Janeway lesions The imaging findings arouse suspicion for verte-
or Osler nodes were present. bral osteomyelitis. Involvement of two contigu-
ous vertebrae is more consistent with osteomyeli-
The patient has no neurologic deficits to suggest tis than with cancer. Vertebral osteomyelitis is
spinal cord compression or the cauda equina syn- primarily the result of hematogenous seeding,
drome. The negative results of the straight-leg direct infection after spinal surgery, or extension
raising test make lumbosacral radiculopathy un- from an infected adjacent tissue. The most com-
likely. The absence of focal spinal tenderness monly isolated organism is Staphylococcus aureus,
1784 n engl j med 376;18 nejm.org May 4, 2017

34

followed by gram-negative bacilli and streptococ- enzyme was normal. Three sets of blood cultures
cus. Coagulase-negative staphylococci and Propi- were negative. Chest radiography showed multi-
onibacterium acnes are often causes of osteomyeli- ple calcified mediastinal and hilar lymph nodes.
tis after spinal surgery with fixation devices. CT of the chest confirmed this finding and also
This patient has had no recent spinal surgery; showed parenchymal scarring in the right upper
however, the history of pacemaker placement lobe and a trace pleural effusion on the right side
could be relevant. Persistent bacteremia from (Fig. 2).
pacemaker lead infections with nonvirulent organ-
isms, such as coagulase-negative staphylococci, The patient has extensive calcified mediastinal
can lead to vertebral osteomyelitis that manifests lymph nodes, a finding that is suggestive of previ-
subacutely. ous granulomatous disease. Disease processes
that can lead to calcification include tuberculosis,
Given the CT findings, more thorough travel and histoplasmosis, coccidioidomycosis, and sarcoid-
dietary histories were obtained to assess the pa-
tients epidemiologic risk for atypical causes of
A
vertebral osteomyelitis. The patients travel history
included trips to Thailand, Malaysia, Venezuela,
Belize, Nicaragua, Australia, and New Zealand,
as well as locations throughout the United States.
He resided in Boston during the summer and
spent the winters in Arizona, which is where he
was residing when the back pain first developed
6 months earlier. He had never worked or lived on
a farm and did not consume unpasteurized dairy
products.
His extensive travel history broadens the differ-

ential diagnosis to include infection with Myco-
bacterium tuberculosis and infection with brucella
species, both of which may manifest subacutely.
This patient reports no exposure to livestock or
to unpasteurized dairy products, which makes the
diagnosis of brucellosis unlikely. Other possi-
bilities include infection with endemic fungi, B
such as Cryptococcus neoformans, Histoplasma capsu-
latum, Blastomyces dermatitidis, or Coccidioides im-
mitis. Rare cases of vertebral osteomyelitis have
also been reported with Burkholderia pseudomallei,
the causative agent of melioidosis, which is en-
demic to parts of Australia and Southeast Asia.
Overall, the patient appears well, without any
neurologic deficits, so there is no immediate need
for surgical intervention or empirical antibiotics.
A definitive diagnosis determined by laboratory
evaluation, including blood cultures, followed by
aspiration biopsy if the blood cultures are nega-
Figure 1. CT Scans of the Lumbar Spine.
tive, should be pursued before therapy is initiated.
CT scans of the lumbar spine with contrast enhance-
ment show end-plate irregularity and destruction of
Testing for human immunodeficiency virus (HIV) the L4 and L5 vertebrae (Panel A, arrow), with a sur-
antibody was negative. The prostate-specific anti- rounding left paravertebral and ventral epidural soft-
gen level was 0.7 ng per milliliter (reference range, tissue collection (Panel B, arrow.)
0 to 4). The blood level of angiotensin-converting
n engl j med 376;18 nejm.org May 4, 2017 1785

35

ly outside the lungs and seed the spine soon

after the initial exposure.
Pathological analysis of the vertebral disk-biopsy

specimens revealed chronic granulomatous in-
flammation with giant cells, plasma cells, and
necrotic bone, as well as round, thick-walled,
nonbudding fungal forms that were 15 to 25 m
in diameter, as best seen on fungal stain (Fig. 3).
A fungal culture grew mold, which was identified
as C. immitis. Serologic testing for coccidioidal
antibodies was positive. The complement-fixing
Figure 2. CT Scan of the Chest. antibody titer was 1:64. The patient was offered
A CT scan of the chest shows multiple calcified medi- initial treatment with amphotericin, but he de-
astinal and hilar lymph nodes.
clined this intravenous therapy and was instead
started on oral itraconazole. Because he had side
effects with itraconazole, he was switched to
osis. Sarcoidosis and histoplasmosis are rare high-dose fluconazole. Six months later, he had
causes of vertebral osteomyelitis; therefore, tuber- progression of his coexisting conditions and
culosis and coccidioidomycosis are the more chose to discontinue the antifungal therapy; he
likely diagnoses. died after transitioning to hospice care.
A purified protein derivative skin test was nega- C om men ta r y

tive. A test for urinary Histoplasma capsulatum anti-
gen was negative. Antibiotics were not initiated, Coccidioidomycosis, which is caused by the fungi
and the patient underwent CT-guided L4L5 disc C. immitis and C. posadasii, is endemic to the south-
aspiration and core-needle biopsy. Fluid could not western United States. Infections occur after the
be aspirated on the initial attempt; 2 to 3 cc of inhalation of spores. The majority of infections
sterile saline were injected and reaspirated. This are acquired in southern Arizona (as was prob-
sample and two core bone-biopsy specimens were ably the case with this patient), central Califor-
sent for culture and histopathological examina- nia, New Mexico, and Texas.1
tion. The results of initial staining for bacteria, Up to two thirds of persons with coccidioido-
fungi, and acid-fast bacilli were negative. The core mycosis have an asymptomatic or mild illness
bone-biopsy specimens were sent to the state lab- and do not present for medical care. Of persons
oratory to test for M. tuberculosis by means of a who seek care, the majority present with a self-
polymerase-chain-reaction assay; the results were limited subacute pulmonary process or flulike
negative. illness, often referred to as valley fever. Many
patients have evidence of infiltrates on chest
One of the leading possible diagnoses before the imaging, which makes it difficult to differentiate
biopsy was performed was tuberculous spondy- valley fever from community-acquired pneumo-
litis. The negative results of the purified protein nia. In endemic regions, coccidioidomycosis ac-
derivative skin test, acid-fast bacilli stains, and counts for up to one quarter of the cases of
polymerase-chain-reaction assay for tuberculosis community-acquired pneumonia.2 This patient
make tuberculosis an unlikely diagnosis. had a history of pneumonia that antedated the
Another diagnostic possibility is coccidioido- development of back pain; in retrospect, his respi-
mycosis, given that the back pain developed while ratory symptoms were probably caused by coccidi-
the patient was in Arizona, soon after pneumo- oidomycosis.
nia was diagnosed. Primary coccidioidomycosis Most respiratory infections resolve without
is often misclassified as community-acquired treatment, and extrapulmonary dissemination
pneumonia since it also manifests as fever, from the hematogenous spread of coccidioido-
cough, and an infiltrate on chest imaging. Coc- mycosis develops in less than 0.5% of immuno-
cidioides species can disseminate hematogenous- competent patients. The most common sites of

36 May 4, 2017

dissemination are the skin, meninges, joints, and on test kit and laboratory performance).9 The IgM
vertebrae. Factors that increase the risk of dis- and IgG antibodies in most patients with coc-
semination include HIV, lymphoma, diabetes, cidioidomycosis revert to negative as the illness
solid-organ transplantation, pregnancy, and treat- resolves; therefore, a positive antibody test usu-
ment with high-dose glucocorticoids or tumor ally signifies a recent infection.9 The concentra-
necrosis factor inhibitors.3,4 The risk of dissemi- tion of complement-fixing antibodies is associat-
nation also appears to be higher among men ed with the extent of disease and can be checked
than among women and higher among persons during treatment to monitor the response to
of African or Filipino descent than among per- treatment. A titer higher than 1:16 often corre-
sons of other ethnic groups.5 Other than being lates with extrapulmonary dissemination; this
male, this patient had no apparent risk factors.
Older age, when adjusted for coexisting condi-
tions, has not been associated with an increased A
risk of dissemination.3
Coccidioidal vertebral osteomyelitis can involve
multiple vertebrae, as was the case with this
patient, in whom the L4 and L5 vertebrae were
affected.5 The intervertebral disk is often spared
early in the course of coccidioidal vertebral osteo-
myelitis, as it is in the course of tuberculous
osteomyelitis, and the absence of intervertebral
disk involvement can help differentiate this dis-
ease from bacterial causes.6 Unlike tuberculosis,
coccidioides does not cause a gibbus deformity
(a form of structural kyphosis that may result
from vertebral compression).5 As the disease pro-
gresses, paraspinal soft-tissue abscesses or epi-
dural abscesses can develop and may lead to
neurologic complications. Disseminated spinal B
coccidioidomycosis is important to recognize
since it is associated with a high risk of compli-
cations and death if it remains untreated.7
Most patients present with subacute or chron-
ic back pain and often do not have systemic signs
of infection, such as fevers or night sweats.5 As a
result of the indolent nature of this infection, the
diagnosis is often delayed by weeks or months, as
was the case with this patient. The diagnosis of
disseminated coccidioidomycosis is made by the
isolation of coccidioides species through culture
or direct microscopy or by serologic detection
of coccidioidal antibodies. Laboratory personnel
should be alerted that the differential diagnosis
includes coccidioidomycosis, because the inhala-
tion of spores can lead to infection.8
Coccidioidal antibodies can be detected with Figure 3. Core-Needle Biopsy of the Vertebral Disk.
the use of several techniques. In most patients, A specimen that was obtained from a core-needle biopsy
of the vertebral disk and was stained with Gomori methe-
antibodies develop within 3 weeks after the on-
namine silver (Panel A) shows a spherule (arrow) with-
set of symptoms. Immunodiffusion kits that de- out visible budding. The same specimen, stained with
tect antibodies have the highest specificity (100%), periodic acid-Schiff (Panel B), also shows a spherule
and enzyme-linked immunoassays have the high- (arrow).
est sensitivity (range, 68.5 to 100%, depending
n engl j med 376;18 nejm.org May 4, 2017 1787

37

patient had a titer of 1:64.10 A urine coccidioides fluconazole.12 Adjunctive surgical dbridement
antigen enzyme immunoassay was found to have or stabilization is often indicated in these cases
a sensitivity of 71% in patients with profound and is crucial if there is cord impingement or
immunosuppression, although more data are spine instability.13 The combination of medical
needed to assess its performance in immuno- and surgical treatment has been associated with
competent patients.11 better outcomes, including symptom relief and
On the basis of limited data from case series, control of the disease, than medical therapy alone.5
treatment of severe coccidioidal osteomyelitis, Most cases of coccidioidal vertebral osteomyeli-
defined as extensive vertebral involvement, spine tis require more than 1 year of antifungal treat-
instability, or risk of spinal cord impingement, ment, and some require indefinite treatment.9
often consists of initial administration of am- This case highlights an uncommon cause of
photericin B, followed by administration of itra- vertebral osteomyelitis that manifested as chron-
conazole, fluconazole, posaconazole, or voricon- ic low back pain in an elderly man. It is important
azole. Liposomal amphotericin B can be used as to consider the possible presence of C. immitis in
an alternative to amphotericin B in patients who the differential diagnosis of vertebral osteomy-
are at risk for drug-induced renal toxic effects. elitis, even in areas in which coccidioidomycosis
Moderate-to-severe disease is often treated with is not endemic, since patients may have previously
an oral azole alone.9 A randomized, controlled lived in, or even briefly visited, endemic areas.
trial that compared fluconazole with itraconazole Presented at the Clinical Pathological Conference Series, De-
for nonmeningeal coccidioidomycosis showed no partment of Medicine, Brigham and Womens Hospital, and
Harvard Medical School both in Boston.
significant difference in response rates between No potential conflict of interest relevant to this article was
the two treatments; however, a subgroup analy- reported.
sis suggested that patients who have skeletal Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
infections are more likely to have a response to We thank Dr. Danny Milner for assistance with the patho-
therapy with itraconazole than to therapy with logical analysis and images.
References
1. Galgiani JN, Ampel NM, Blair JE, et al. 7. Crum NF, Lederman ER, Stafford CM, coccidioidomycosis. Clin Microbiol Rev
Coccidioidomycosis. Clin Infect Dis 2005; Parrish JS, Wallace MR. Coccidioidomy- 2013;26:505-25.
41:1217-23. cosis: a descriptive survey of a reemerging 11. Durkin M, Connolly P, Kuberski T, et al.
2. Valdivia L, Nix D, Wright M, et al. disease: clinical characteristics and cur- Diagnosis of coccidioidomycosis with use
Coccidioidomycosis as a common cause of rent controversies. Medicine (Baltimore) of the Coccidioides antigen enzyme im-
community-acquired pneumonia. Emerg 2004;83:149-75. munoassay. Clin Infect Dis 2008;47(8):
Infect Dis 2006;12:958-62. 8. Stevens DA, Clemons KV, Levine HB, e69-e73.
3. Brown J, Benedict K, Park BJ, Thomp- et al. Expert opinion: what to do when 12. Galgiani JN, Catanzaro A, Cloud GA,
son GR III. Coccidioidomycosis: epidemi- there is Coccidioides exposure in a labo- et al. Comparison of oral fluconazole and
ology. Clin Epidemiol 2013;5:185-97. ratory. Clin Infect Dis 2009;49:919-23. itraconazole for progressive, nonmenin-
4. Stevens DA. Coccidioidomycosis. 9. Galgiani JN, Ampel NM, Blair JE, et al. geal coccidioidomycosis: a randomized,
N Engl J Med 1995;332:1077-82. 2016 Infectious Diseases Society of double-blind trial. Ann Intern Med 2000;
5. Martirosyan NL, Skoch JM, Zaninovich America (IDSA) clinical practice guide- 133:676-86.
O, Zoccali C, Galgiani JN, Baaj AA. A para- line for the treatment of coccidioidomy- 13. Kakarla UK, Kalani MY, Sharma GK,
digm for the evaluation and management cosis. Clin Infect Dis 2016;63(6):e112- Sonntag VK, Theodore N. Surgical man-
of spinal coccidioidomycosis. Surg Neurol e146. agement of coccidioidomycosis of the
Int 2015;6:107. 10. Nguyen C, Barker BM, Hoover S, et al. spine: clinical article. J Neurosurg Spine
6. Mirochnik BD, Lev S, Weingarten EP. Recent advances in our understanding of 2011;15:441-6.
Case 209: Disseminated coccidioidal spon- the environmental, epidemiological, im- Copyright 2017 Massachusetts Medical Society.
dylodiskitis. Radiology 2014;272:914-8. munological, and clinical dimensions of
clinical problem-solving series

The Journal welcomes submissions of manuscripts for the Clinical Problem-Solving
series. This regular feature considers the step-by-step process of clinical decision
making. For more information, please see authors.NEJM.org.
1788 n engl j med 376;18 nejm.org May 4, 2017

38

CLINICAL PRACTICE
This Journal feature begins with a case vignette highlighting a common clinical problem. Evidence supporting
various strategies is then presented, followed by a review of formal guidelines, when they exist. The article
ends with the authors clinical recommendations.
39

Clinical Practice
Screening for Colorectal Neoplasia

John M. Inadomi, M.D.
This Journal feature begins with a case vignette highlighting a common clinical problem. Evidence
supporting various strategies is then presented, followed by a review of formal guidelines, when they exist.
The article ends with the authors clinical recommendations.
A 79-year-old woman visits your office for routine health maintenance. She has nor- From the Division of Gastroenterology,
mal daily bowel movements without rectal bleeding. Her medical history is notable Department of Medicine, University of
Washington School of Medicine, and the
for osteoarthritis but no other medical conditions. She takes nonsteroidal anti- Department of Health Services, Univer-
inflammatory medication and multivitamins. Her maternal uncle received a diagno- sity of Washington School of Public
sis of colorectal cancer at 65 years of age, but she has never undergone screening for Health both in Seattle. Address reprint
requests to Dr. Inadomi at the Division of
colorectal cancer. Would you advise this patient to undergo screening for colorectal Gastroenterology, Department of Medi-
cancer, and if so, which screening strategy would you recommend? cine, University of Washington School of
Medicine, 1959 N.E. Pacific St., Seattle,
WA 98195, or at jinadomi@uw.edu.
The Cl inic a l Probl em N Engl J Med 2017;376:149-56.
C
DOI: 10.1056/NEJMcp1512286
olorectal cancer is the third most commonly diagnosed can- Copyright 2017 Massachusetts Medical Society.
cer and cause of death from cancer in the United States1; however, it can
be detected in asymptomatic patients at a curable stage, and several ran-
domized, controlled trials have shown lower mortality among patients who undergo
screening than among those who do not. Screening can also detect precancerous
polyps that can be removed during colonoscopy, thereby reducing the incidence of
cancer. This review focuses on screening patients at average risk for the develop-
ment of colorectal cancer.
An audio version
of this article
S t r ategie s a nd E v idence is available at
NEJM.org
Screening Options
Multiple strategies are available to screen patients who are at average risk for the
development of colorectal cancer, including fecal occult blood testing (with the use
of guaiac-based or immunochemical tests) alone or in combination with stool
DNA examination, endoscopy (flexible sigmoidoscopy or colonoscopy), radiologic
examination (computed tomographic [CT] colonography), and testing for blood-
based molecular markers, such as circulating methylated septin 9 gene (SEPT9)
DNA. Each strategy has differing characteristics with respect to accuracy, invasive-
ness, interval, costs, and quality of evidence supporting its use. The advantages
and disadvantages associated with each screening strategy are summarized in
Table 1. Colorectal cancer screening involves not only the one-time use of a screen-
ing test, but also repeated testing over a persons lifetime (programmatic screening).
In addition, if colonoscopy is not performed as the primary screening test, all
other screening strategies require colonoscopy as follow-up to a positive test.
n engl j med 376;2 nejm.org 40

January 12, 2017 149

Key Clinical Points
Colorectal Cancer Screening

Patients at average risk for the development of colorectal cancer should begin screening at 50 years of age.
Screening between 76 and 85 years of age should be tailored, and screening should stop after 85 years
of age.
There is no best strategy for colorectal cancer screening; therefore, the most effective strategy is the
one that a patient can adhere to consistently.
Screening strategies for colorectal cancer among patients at average risk include annual fecal occult blood
testing (with the use of highly sensitive guaiac-based or immunochemical tests) or colonoscopy every
10 years. If these strategies are declined, other screening options may include flexible sigmoidoscopy,
computed tomographic colonography, or blood or stool DNA testing.
Although the harms of specific tests vary greatly, the overall risks of complications from screening are
related mainly to colonoscopies (performed up front or as follow-up to other positive screening tests)
and polypectomies; these risks are low and similar among all strategies.
Fecal Screening Tests studies used guaiac-based testing, almost all cur-
Fecal screening can be divided into two broad rent population-based screening programs use
categories: those that detect blood from ulcerated immunochemical tests because their accuracy is
colonic mucosa resulting from cancer or large greater and dietary restrictions are not required.
polyps and those that detect molecular markers Molecular markers including abnormal DNA
that are shed from cancerous epithelial cells. from cancerous cells can be detected in stool.
Fecal occult blood tests include guaiac-based FIT combined with a stool DNA test (FIT-DNA)
tests and immunochemical tests that use anti- has been approved by the Food and Drug Ad-
bodies to detect human blood. The fecal immu- ministration (FDA) for colorectal cancer screen-
nochemical test (FIT) has the advantage of not ing. One study showed that one-time FIT-DNA
requiring dietary restrictions; such restrictions had a higher sensitivity for detection of colorec-
are necessary with guaiac-based tests because tal cancer than one-time FIT alone (92.3% vs.
they may be falsely positive in the presence of 73.8%), but specificity was lower (86.6% vs.
blood from red meat or food that reacts with the 94.9%).11 The screening interval differs between
guaiac (e.g., raw horseradish, turnips, or broccoli). FIT (annual) and FIT-DNA (interval unknown,
FIT should be performed at home on successive although the U.S. Preventive Services Task Force
bowel movements; digital rectal examination in recommends 1 or 3 years), which makes a com-
the clinic does not provide an adequate stool parison of the effectiveness of programmatic
sample for testing. One-time FIT has been re- screening difficult. Data from studies evaluating
ported to have a sensitivity of 79% (95% confi- the colorectal cancer mortality benefit of screen-
dence interval [CI], 69 to 86) and a specificity of ing FIT-DNA are lacking.
94% (95% CI, 92 to 95) for detection of cancer,
and two or three samples do not significantly Endoscopic Screening
increase the accuracy over a single sample.2 Flexible Sigmoidoscopy
However, many types of immunochemical tests Randomized trials have shown that screening
are available, and their test characteristics vary with flexible sigmoidoscopy, followed by colonos-
widely. copy if precancerous polyps are detected, reduces
High-quality evidence supports a strategy of colorectal cancer mortality. Although not all tri-
fecal occult blood testing every year or every als have shown a significant benefit with respect
2 years to screen for colorectal cancer, with colo- to reducing mortality (mortality benefit),15,16 the
noscopy used as follow-up to a positive test. intention-to-treat analyses in several large, ran-
Several randomized, controlled trials have shown domized, controlled trials have confirmed the
up to 32% lower mortality from colorectal can- effectiveness of one-time and periodic (every 3 to
cer with this strategy than with no screening, 5 years) sigmoidoscopy, with a 26 to 31% lower
with up to 30 years of follow-up (Table S1 in the mortality from colorectal cancer among patients
Supplementary Appendix, available with the full who underwent flexible sigmoidoscopy screening
text of this article at NEJM.org).3-7 Although these than among those who underwent no screening

41

Table 1. Comparison of Key Features of Screening Strategies.*
Strategy and Effect Detection of Precancerous

Cases in Primary Care
on Cancer Mortality Quality of Evidence Interval Cost-Effectiveness Convenience and Requirements Neoplasia
Guaiac FOBT and FIT: Multiple RCTs have shown a mortali- Annual May be more effective and Performed at home Does not reliably detect
32% lower mortality ty benefit (reduction in mortality) less expensive than no precancerous neoplasia
for guaiac FOBT2-7; although FIT screening; total costs
is more accurate than guaiac lower than no screen-
FOBT, RCTs evaluating FIT are ing, because of the high
lacking expense of late-stage
cancer treatment with
biologic agents
Flexible sigmoidoscopy: RCTs have shown a mortality benefit8,9 Every 5 yr Cost-effective as compared Limited bowel preparation as Can detect precancerous
27% lower mortality with no screening and compared with colonoscopy neoplasia
other strategies
Flexible sigmoidoscopy plus A single RCT showed that flexible Annual (FIT) and Cost-effective as compared Strategy that combines endo- Can detect precancerous
FIT: 38% lower mortality sigmoidoscopy plus FIT reduces every 10 yr (sig- with no screening and scopic and stool testing neoplasia
n engl j med 376;2

cancer mortality more than sigmoidoscopy) other strategies
moidoscopy alone10
FIT-DNA: unknown effect Data from studies showing a mortal- Every 1 or 3 yr Less effective and more Performed at home Does not reliably detect
on mortality ity benefit are lacking; studies costly than FOBT, FIT, precancerous neoplasia
nejm.org
were limited to the detection of or colonoscopy
cancer and precancerous polyps
Clinical Pr actice
by FIT-DNA as compared with

colonoscopy11

Colonoscopy: 68% lower A prospective cohort study showed Every 10 yr Cost-effective as compared Requires full bowel preparation; Can detect precancerous
mortality a mortality benefit12 with no screening and usually requires sedation neoplasia
42 January 12, 2017

other strategies and an escort
CT colonography: unknown Data from studies showing a mortal- Every 5 yr Less effective and more No sedation required but re- Can detect precancerous
effect on mortality ity benefit are lacking; studies costly than FOBT, FIT, quires bowel preparation neoplasia
were limited to the detection of or colonoscopy
cancer by CT colonography as
compared with colonoscopy13
Circulating methylated Data from studies showing a mortal- Unknown Unknown A blood test may be associated Does not reliably detect
SEPT9 DNA: unknown ity benefit are lacking; studies with greater adherence than precancerous neoplasia
effect on mortality were limited to the detection of that with other screening
cancer by circulating methylated tests
SEPT9 DNA as compared with
colonoscopy14
* CT denotes computed tomography, FIT fecal immunochemical test, FIT-DNA fecal immunochemical test combined with stool DNA test, FOBT fecal occult blood test, and RCT randomized,
controlled trial.
The effect on mortality represents a comparison of the strategy with either no screening or other strategies.
Cost-effectiveness was determined as the cost per quality-adjusted life-year gained.
151
NEJM.org
(Table S1 in the Supplementary Appendix).8-10,13,17 lesions outside the colon that were discovered
However, the benefit of sigmoidoscopy is limited incidentally.
to cancer in the distal colon (rectum, sigmoid,
and descending colon), for which the reduction Blood-Based Tests
in mortality was reported to be 46%.17 The FDA has approved a blood-based colorectal
Colonoscopy cancer screening test that detects circulating
Case-control and prospective cohort studies have methylated SEPT9 DNA. Data from studies evalu-
estimated cancer mortality to be 68 to 88% ating the colorectal cancer mortality benefit of
lower among persons who undergo screening blood-based screening are lacking. In a prospec-
colonoscopy than among those who do not tive study conducted in a screening population,
(Table S1 in the Supplementary Appendix).12,17-19 in which colonoscopy was used as the reference
A meta-analysis of observational studies showed standard, the presence of circulating methylated
that despite a 68% lower mortality overall, no SEPT9 DNA was shown to have a sensitivity of
significant mortality benefit from colonoscopy 48.2% (95% CI, 32.4 to 63.6), a specificity of
was seen with respect to cancer in the proximal 91.5% (95% CI, 89.7 to 93.1), a positive predic-
colon.17 This discrepancy may be explained by tive value of 5.2% (95% CI, 3.5 to 7.5), and a
the quality of colonoscopy (i.e., incomplete colo- negative predictive value of 99.5% (95% CI, 99.2
noscopy, poorer bowel preparation, or more dif- to 99.6) for detection of colorectal cancer.14
ficult polyp removal in the proximal colon) or
differences in the biologic characteristics of prox- When to Start and Stop Screening
imal and distal colorectal cancer. The mechanism The U.S. Preventive Services Task Force used com-
underlying the majority of colorectal cancers is parative effectiveness modeling to examine dif-
chromosomal instability with early mutations ferent ages at which to initiate screening. Start-
in APC followed by KRAS and late mutations in ing screening at 45 years of age instead of 50
P53.20 Sessile serrated polyps are more commonly years could increase life-years and reduce cancer
associated with early KRAS or BRAF mutations mortality but could also increase the potential
that may lead to chromosomal or microsatellite harms due to the increased burden of colonos-
instability than are adenomatous polyps.21 Ses- copy; for this reason, the recommendations are
sile serrated polyps are often flat or minimally to begin screening at 50 years of age in patients
raised, making them more difficult than other at average risk for colorectal cancer.23 Although
polyps to detect by colonoscopy, and are more the risk of colorectal cancer increases with age,
common in the proximal colon, which may con- the competing risk of death from other diseases
tribute to the higher risk of proximal cancer and the risk of serious complications from colo-
than distal cancer after colonoscopy.22 noscopy also increase with age.24,25 Several na-
tional organizations recommend that screening
Radiographic Tests for patients between 76 and 85 years of age
Published data from studies evaluating the effect should be tailored on the basis of the presence
of CT colonography on colorectal cancer inci- of coexisting illnesses and that screening should
dence and mortality are lacking. The reported be stopped after patients reach 85 years of age.23,26
sensitivity and specificity of CT colonography to A microsimulation model suggested that the
detect adenomas 1 cm in diameter or larger have intensity of prior screening and the individual
ranged from 66.7 to 93.5% and from 86.0 to risk of colorectal cancer should also be consid-
97.9%, respectively.13 Because sessile serrated ered in determining the age at which to stop
polyps are flat, CT colonography is inferior to screening. Patients without a notable coexisting
colonoscopy for detection of these polyps.22 A illness who are at average or higher risk for
polyp 6 mm in diameter or larger typically colorectal cancer and have had no prior screen-
prompts a referral for optical colonoscopy, al- ing would be expected to benefit from screening
though the most useful size cutoff is controver- into their 80s.27
sial. Disadvantages of CT colonography include
exposure to radiation and associated concerns Adherence
regarding radiation-induced cancers and the po- The percentage of U.S. residents with up-to-date
tential need for additional testing and care for screening for colorectal cancer has not increased

43 January 12, 2017

Clinical Pr actice
appreciably since 2010 and remains at approxi- implementing screening strategies. The quality of
mately 60%.28 Currently, the percentage of U.S. a screening program should be measured by its
adults undergoing colonoscopy screening greatly ability to identify patients who are due for screen-
exceeds the percentage screened by fecal occult ing, provide access to screening, assess adherence
blood testing, and less than 1% undergo flexible to the screening test and to follow-up colonos-
sigmoidoscopy.29 Barriers to screening include copy if a noncolonoscopy screening test is posi-
costs, lack of knowledge of colorectal cancer and tive, document test outcomes and disseminate
screening, underappreciation of the effect or se- accurate follow-up recommendations, identify
verity of colorectal cancer, fatalism, and a per- patients with a negative test to follow them for
ceived lack of importance or fear of screening repeat screening at the appropriate intervals, and
tests.30 Costs remain a barrier despite the man- provide timely surgery for cancers. The rate of
date in the Affordable Care Act that health plans adenoma detection (the percentage of patients in
cover colorectal cancer screening with no patient whom precancerous polyps are detected during
cost-sharing, because Medicare and other insur- screening colonoscopy) differs substantially among
ers impose a cost-sharing requirement when a endoscopists and may be used as a measure of
colonoscopy is performed to evaluate a positive the ability of screening to prevent colorectal
screening test or when a screening colonoscopy cancer.36 A retrospective study showed that for
becomes a therapeutic procedure with the inclu- every 1% increase in the rate of adenoma detec-
sion of polypectomy.31 tion, there is a 3% decrease in the rate of cancer
Various interventions used in randomized, developing after colonoscopy.37
controlled trials have been shown to increase pa-
tient participation in screening; such interventions Harms and Cost-Effectiveness
include sending patients invitations from their
The harms of noncolonoscopy screening tests are
primary care provider, sending reminder letters
low; however, all strategies require colonoscopy
and making telephone calls, and mailing fecal as follow-up to a positive test. As a result, the
occult blood test kits to patients homes. Theprogrammatic harms of screening are propor-
most successful programs use patient navigators
tional to the number of colonoscopies and in
to reduce logistic barriers, address cultural issues,
particular polypectomies that are performed over
and encourage participants to undergo screen- the lifetime of the screened population. The tech-
ing; the use of patient navigators is especially
nical report from the Cancer Intervention and
important in underserved populations.32,33 Surveillance Modeling Network (CISNET) Colorec-
The National Colorectal Cancer Roundtable has
tal Cancer Working Group estimated the number
established a goal of 80% adherence to colorec-
of complications (perforations, gastrointestinal
tal cancer screening by the year 2018. Kaiser bleeding, nausea and vomiting, ileus, dehydration,
Permanente has implemented a comprehensive abdominal pain, myocardial infarction, angina,
strategy focused on FIT screening, with colonos-
arrhythmias, congestive heart failure, respira-
copy performed as follow-up to a positive test,
tory arrest, syncope, hypotension, or shock) in a
and has reached and maintained the goal of 80%population of 1000 persons screened between 50
adherence through four rounds of screening.34 and 75 years of age to be 14 to 15 with colonos-
copy at 10-year intervals, 9 to 12 with flexible
Adherence to screening tests varies among strat-
sigmoidoscopy at 5-year intervals, 9 to 10 with
egies, and preference of strategy varies by race
and ethnic group; white participants more com-FIT-DNA at 3-year intervals, and 10 to 11 with
annual FIT.38
monly prefer colonoscopy, and nonwhite partici-
pants tend to prefer fecal testing.30,35 To achieve
Cost-effectiveness models more than a decade
the highest level of adherence to colorectal can-
ago suggested that programmatic screening with
cer screening, it may be best to provide partici-
fecal occult blood testing, flexible sigmoidoscopy,
pants a choice, because the best strategy is the
or colonoscopy reduced colorectal cancer mor-
one that they will adhere to consistently. tality at a cost society was willing to pay.39 Find-
ings from a more recent analysis suggest that
Quality of Screening given the high expense of late-stage cancer treat-
Maximizing the benefit of colorectal cancer ment with biologic agents, FIT screening may be
screening requires a programmatic approach to cost-saving, with reductions in both cancer mor-

44

Table 2. U.S. Guideline Recommendations for Screening and Screening Intervals to Reduce Mortality from Colorectal Cancer in Patients at
Average Risk.
U.S. Preventive National Comprehensive Multi-Society American College

Services Task Force Cancer Network Task Force of Gastroenterology
Strategy (2016)23* (2015)43 (2008)50 (2009)51
Sensitive guaiac FOBT Annually Recommended Recommended Recommended
(annually) (annually) (annually)
FIT Annually Recommended Recommended Recommended
(annually) (annually) (annually)
Stool DNA test Annually or every Not recommended Recommended Recommended
3 yr (interval (every 3 yr)
unknown)
Flexible sigmoidoscopy Every 5 yr Recommended Recommended Recommended
(every 5 yr) (every 5 yr) (every 5 yr)
Flexible sigmoidoscopy Every 10 yr, with Not recommended Not recommended Not recommended
plus FIT annual FIT or
sensitive FOBT
Colonoscopy Every 10 yr Recommended Recommended Preferred
(every 10 yr) (every 10 yr) (every 10 yr)
CT colonography Every 5 yr Not recommended Recommended Recommended
(every 5 yr) (every 5 yr)
Circulating methylated Not specified Unavailable for Unavailable for Unavailable for
SEPT9 DNA guideline guideline guideline
* No recommended strategy was provided.

The Multi-Society Task Force included the American Gastroenterological Association, the American Society for Gastrointestinal Endoscopy,
the American College of Gastroenterology, the American Cancer Society, and the American College of Radiology.
Sensitivity for detection of colorectal cancer is higher than 70%.
The screening interval is for multitarget FIT-DNA.
Stool-based testing may be added at year 3.
Colonoscopy was identified as the preferred strategy.
tality and overall costs even if one includes A r e a s of Uncer ta in t y

program support costs to increase screening
uptake.40 Moreover, screening with colonosco- Data from completed randomized, controlled
py or FIT is more effective in reducing cancer trials of screening colonoscopy are lacking, al-
mortality and less expensive than screening with though several studies are under way. The Colo-
FIT-DNA41 or CT colonography.42 From an eco- noscopy versus Fecal Immunochemical Test in
nomic perspective, FIT-DNA or CT colonography Reducing Mortality from Colorectal Cancer
should not be recommended unless screening (CONFIRM) trial (ClinicalTrials.gov number,
with FIT, sigmoidoscopy, or colonoscopy has NCT01239082) is a randomized comparison of
been declined. one-time colonoscopy with annual FIT plus colo-
noscopy as follow-up to a positive test, to exam-
Patients at Elevated Risk for Colorectal ine colorectal cancer incidence and mortality
Cancer over 10 years.44 A similar trial comparing colo-
Earlier and more frequent screening is recom- noscopy with FIT is being conducted in Spain
mended for patients at higher risk (Table S2 in (COLONPREV).45 Two additional European stud-
the Supplementary Appendix). Patients with a ies are comparing screening colonoscopy with
first-degree relative in whom colorectal cancer no screening (the Nordic-European Initiative on
developed before 60 years of age should undergo Colorectal Cancer [NordICC])4647 or with FIT or
colonoscopy at 40 years of age or an age 10 years no screening (Screening of Swedish Colons
younger than the relatives age when cancer de- [SCREESCO], NCT02078804) with respect to mor-
veloped, whichever is earlier.52 tality from colorectal cancer.
154 45
n engl j med 376;2 nejm.org January 12, 2017

Clinical Pr actice
Additional factors that might influence colorec- 50 and 75 years of age, with tailored screening for
tal screening strategies include race, lifestyle fac- those between 76 and 85 years of age.23
tors, or aspirin use. For example, among black
men and women, the rates of death from colorec- C onclusions a nd
tal cancer are 28.4 and 18.9 per 100,000 popula- R ec om mendat ions
tion, respectively; among white men and women,
the corresponding rates are 18.7 and 13.2 per Although the patient described in the vignette is
100,000 population.48 Obesity, tobacco smoking, 79 years of age, she has not previously under-
low physical activity, high intake of alcohol, high gone screening for colorectal cancer. Because of
intake of red or processed meat, and low intake her limited coexisting illnesses, she is expected
of fruits and vegetables are associated with in- to derive an overall benefit from a first screening
creased risk of colorectal cancer, and regular use for colorectal cancer, and thus I would recom-
of aspirin has been associated with reduced risk. mend screening for this patient. I would explore
However, none of these factors are currently her reasons for not previously pursuing screen-
used to differentiate screening strategy, age of ing and review with her the benefits and harms
screening initiation, or surveillance intervals.49 of different strategies. Because it is not feasible
to summarize the entire menu of options dur-
ing the span of a routine health maintenance
Guidel ine s
visit, I would initially focus the discussion on
Several national organizations have published colonoscopy or on annual FIT, followed by
guidelines on strategies to reduce colorectal colonoscopy if the test was positive, and engage
cancer mortality, including the National Com- the patient in shared decision making. If she
prehensive Cancer Network,43 the U.S. Multi-Society declined colonoscopy and FIT, I would discuss
Task Force,50 and the American College of Gas- additional screening options, with the under-
troenterology.51 Whereas these organizations rec- standing that insurance coverage, cost-sharing,
ommend certain screening strategies over others and other barriers may affect the feasibility of
(Table 2), the 2016 U.S. Preventive Services Task some options.
Force recommendations do not support any spe- Dr. Inadomi reports receiving consulting fees from ChemImage.
cific testing strategy or strategies over others, but No other potential conflict of interest relevant to this article was
reported.
rather highlight the importance of screening pa- Disclosure forms provided by the author are available with the
tients at average risk for colorectal cancer between full text of this article at NEJM.org.
References
1. United States Cancer Statistics. 1999- Rasmussen M. Randomized study of bien- SH, et al. Multitarget stool DNA testing
2013 Incidence and mortality data. Atlan- nial screening with a faecal occult blood for colorectal-cancer screening. N Engl J
ta: Centers for Disease Control and Pre- test: results after nine screening rounds. Med 2014;370:1287-97.
vention, 2015 (https://nccd.cdc.gov/uscs/). Scand J Gastroenterol 2004;39:846-51. 12. Nishihara R, Wu K, Lochhead P, et al.
2. Lee JK, Liles EG, Bent S, Levin TR, 7. Lindholm E, Brevinge H, Haglind E. Long-term colorectal-cancer incidence and
Corley DA. Accuracy of fecal immuno- Survival benefit in a randomized clinical mortality after lower endoscopy. N Engl J
chemical tests for colorectal cancer: sys- trial of faecal occult blood screening for Med 2013;369:1095-105.
tematic review and meta-analysis. Ann colorectal cancer. Br J Surg 2008;95:1029- 13. Lin JS, Piper MA, Perdue LA, et al.
Intern Med 2014;160:171. 36. Screening for colorectal cancer: an updated
3. Faivre J, Dancourt V, Lejeune C, et al. 8. Atkin WS, Edwards R, Kralj-Hans I, systematic review for the U.S. Preventive
Reduction in colorectal cancer mortality et al. Once-only flexible sigmoidoscopy Services Task Force. Portland, OR: Kaiser
by fecal occult blood screening in a French screening in prevention of colorectal can- Permanente Research Affiliates Evidence-
controlled study. Gastroenterology 2004; cer: a multicentre randomised controlled based Practice Center, October 2015.
126:1674-80. trial. Lancet 2010;375:1624-33. (AHRQ publication no. 14-05203-EF-1)
4. Shaukat A, Mongin SJ, Geisser MS, 9. Schoen RE, Pinsky PF, Weissfeld JL, (https://www.uspreventiveservicestaskforce
et al. Long-term mortality after screening et al. Colorectal-cancer incidence and mor- .org/Home/GetFile/1/685/coloncandraftes
for colorectal cancer. N Engl J Med 2013; tality with screening flexible sigmoidos- 135/pdf).
369:1106-14. copy. N Engl J Med 2012;366:2345-57. 14. Church TR, Wandell M, Lofton-Day C,
5. Scholefield JH, Moss SM, Mangham 10. Holme , Lberg M, Kalager M, et al. et al. Prospective evaluation of methylated
CM, Whynes DK, Hardcastle JD. Notting- Effect of flexible sigmoidoscopy screen- SEPT9 in plasma for detection of asymp-
ham trial of faecal occult blood testing ing on colorectal cancer incidence and mor- tomatic colorectal cancer. Gut 2014;63:
for colorectal cancer: a 20-year follow-up. tality: a randomized clinical trial. JAMA 317-25.
Gut 2012;61:1036-40. 2014;312:606-15. 15. Hoff G, Grotmol T, Skovlund E,
6. Kronborg O, Jrgensen OD, Fenger C, 11. Imperiale TF, Ransohoff DF, Itzkowitz Bretthauer M. Risk of colorectal cancer

46

Clinical Pr actice
seven years after flexible sigmoidoscopy 28. Sabatino SA, White MC, Thompson Brenner H. Cost-effectiveness of colorec-
screening: randomised controlled trial. TD, Klabunde CN. Cancer screening test tal cancer screening. Epidemiol Rev 2011;
BMJ 2009;338:b1846. use United States, 2013. MMWR Morb 33:88-100.
16. Segnan N, Armaroli P, Bonelli L, et al. Mortal Wkly Rep 2015;64:464-8. 41. Ladabaum U, Mannalithara A. Com-
Once-only sigmoidoscopy in colorectal 29. Vital signs: colorectal cancer screen- parative effectiveness and cost effective-
cancer screening: follow-up findings of ing test use United States, 2012. ness of a multitarget stool DNA test to
the Italian Randomized Controlled Trial MMWR Morb Mortal Wkly Rep 2013;62: screen for colorectal neoplasia. Gastroen-
SCORE. J Natl Cancer Inst 2011;103: 881-8. terology 2016;151(3):427-439.e6.
1310-22. 30. Brenner AT, Ko LK, Janz N, Gupta S, 42. Vijan S, Hwang I, Inadomi J, et al. The
17. Brenner H, Stock C, Hoffmeister M. Inadomi J. Race/ethnicity and primary cost-effectiveness of CT colonography in
Effect of screening sigmoidoscopy and language: health beliefs about colorectal screening for colorectal neoplasia. Am J
screening colonoscopy on colorectal can- cancer screening in a diverse, low-income Gastroenterol 2007;102:380-90.
cer incidence and mortality: systematic population. J Health Care Poor Under- 43. Provenzale D, Jasperson K, Ahnen DJ,
review and meta-analysis of randomised served 2015;26:824-38. et al. Colorectal Cancer Screening, version
controlled trials and observational stud- 31. Doubeni CA, Corley DA, Zauber AG. 1.2015. J Natl Compr Canc Netw 2015;13:
ies. BMJ 2014;348:g2467. Colorectal cancer health disparities and 959-68.
18. Kahi CJ, Imperiale TF, Juliar BE, Rex the role of US law and health policy. Gas- 44. Robertson DJ, Kaminski MF, Brett-
DK. Effect of screening colonoscopy on troenterology 2016;150:1052-5. hauer M. Effectiveness, training and qual-
colorectal cancer incidence and mortality. 32. Myers RE, Bittner-Fagan H, Daskala- ity assurance of colonoscopy screening for
Clin Gastroenterol Hepatol 2009;7:770-5. kis C, et al. A randomized controlled trial colorectal cancer. Gut 2015;64:982-90.
19. Manser CN, Bachmann LM, Brunner of a tailored navigation and a standard 45. Castells A, Quintero E. Program-
J, Hunold F, Bauerfeind P, Marbet UA. intervention in colorectal cancer screen- matic screening for colorectal cancer:
Colonoscopy screening markedly reduces ing. Cancer Epidemiol Biomarkers Prev the COLONPREV study. Dig Dis Sci 2015;
the occurrence of colon carcinomas and 2013;22:109-17. 60:672-80.
carcinoma-related death: a closed cohort 33. Ritvo PG, Myers RE, Paszat LF, et al. 46. Kaminski MF, Bretthauer M, Zauber
study. Gastrointest Endosc 2012;76:110-7. Personal navigation increases colorectal AG, et al. The NordICC Study: rationale
20. Fearon ER, Vogelstein B. A genetic cancer screening uptake. Cancer Epide- and design of a randomized trial on colo-
model for colorectal tumorigenesis. Cell miol Biomarkers Prev 2015;24:506-11. noscopy screening for colorectal cancer.
1990;61:759-67. 34. Jensen CD, Corley DA, Quinn VP, et al. Endoscopy 2012;44:695-702.
21. OBrien MJ, Zhao Q, Yang S. Colorec- Fecal immunochemical test program per- 47. Bretthauer M, Kaminski MF, Lberg
tal serrated pathway cancers and precur- formance over 4 rounds of annual screen- M, et al. Population-based colonoscopy
sors. Histopathology 2015;66:49-65. ing: a retrospective cohort study. Ann In- screening for colorectal cancer: a random-
22. IJspeert JE, Tutein Nolthenius CJ, Kui- tern Med 2016;164:456-63. ized clinical trial. JAMA Intern Med 2016;
pers EJ, et al. CT-colonography vs. colo- 35. Inadomi JM, Vijan S, Janz NK, et al. 176:894-902.
noscopy for detection of high-risk sessile Adherence to colorectal cancer screening: 48. Siegel RL, Miller KD, Jemal A. Cancer
serrated polyps. Am J Gastroenterol 2016; a randomized clinical trial of competing statistics, 2015. CA Cancer J Clin 2015;65:
111:516-22. strategies. Arch Intern Med 2012;172:575- 5-29.
23. U.S. Preventive Services Task Force, 82. 49. Cancer facts & figures 2016. Atlanta:
Bibbins-Domingo K, Grossman DC, et al. 36. Barclay RL, Vicari JJ, Doughty AS, American Cancer Society, 2016 (http://www
Screening for colorectal cancer: US Pre- Johanson JF, Greenlaw RL. Colonoscopic .cancer.org/acs/groups/content/@research/
ventive Services Task Force recommenda- withdrawal times and adenoma detection documents/document/acspc-047079.pdf).
tion statement. JAMA 2016;315:2564-75. during screening colonoscopy. N Engl J 50. Levin B, Lieberman DA, McFarland B,
24. Rutter CM, Johnson E, Miglioretti DL, Med 2006;355:2533-41. et al. Screening and surveillance for the
Mandelson MT, Inadomi J, Buist DS. Ad- 37. Corley DA, Jensen CD, Marks AR, et al. early detection of colorectal cancer and
verse events after screening and follow-up Adenoma detection rate and risk of adenomatous polyps, 2008: a joint guide-
colonoscopy. Cancer Causes Control 2012; colorectal cancer and death. N Engl J Med line from the American Cancer Society,
23:289-96. 2014;370:1298-306. the US Multi-Society Task Force on Colo-
25. Day LW, Kwon A, Inadomi JM, Walter 38. Zauber A, Knudsen A, Rutter CM, rectal Cancer, and the American College
LC, Somsouk M. Adverse events in older Lansdorp-Vogelaar I, Kuntz KM. Evaluat- of Radiology. Gastroenterology 2008;134:
patients undergoing colonoscopy: a sys- ing the benefits and harms of colorectal 1570-95.
tematic review and meta-analysis. Gastro- cancer screening strategies: a collabora- 51. Rex DK, Johnson DA, Anderson JC,
intest Endosc 2011;74:885-96. tive modeling approach. Rockville, MD: Schoenfeld PS, Burke CA, Inadomi JM.
26. Qaseem A, Denberg TD, Hopkins RH Agency for Healthcare Research and American College of Gastroenterology
Jr, et al. Screening for colorectal cancer: Quality, October 2015. (AHRQ publica- guidelines for colorectal cancer screening
a guidance statement from the American tion no. 14-05203-EF-2) (https://www 2009 [corrected]. Am J Gastroenterol 2009;
College of Physicians. Ann Intern Med .uspreventiveservicestaskforce.org/Home/ 104:739-50.
2012;156:378-86. GetFile/1/16540/cisnet-draft-modeling 52. Schoen RE, Razzak A, Yu KJ, et al. In-
27. van Hees F, Saini SD, Lansdorp-Voge- -report/pdf). cidence and mortality of colorectal cancer
laar I, et al. Personalizing colonoscopy 39. Pignone M, Saha S, Hoerger T, Man- in individuals with a family history of
screening for elderly individuals based delblatt J. Cost-effectiveness analyses of colorectal cancer. Gastroenterology 2015;
on screening history, cancer risk, and colorectal cancer screening: a systematic 149(6):1438-1445.e1.
comorbidity status could increase cost review for the U.S. Preventive Services Task Copyright 2017 Massachusetts Medical Society.
effectiveness. Gastroenterology 2015;149: Force. Ann Intern Med 2002;137:96-104.
1425-37. 40. Lansdorp-Vogelaar I, Knudsen AB,

47

Clinical Practice
Heart Failure with Preserved

Ejection Fraction
Margaret M. Redfield, M.D.
This Journal feature begins with a case vignette highlighting a common clinical problem. Evidence
supporting various strategies is then presented, followed by a review of formal guidelines, when they exist.
The article ends with the authors clinical recommendations.
From the Department of Cardiovascular A 73-year-old woman with a history of dyspnea on exertion presents for a follow-up
Diseases, Mayo Clinic, Rochester, MN. visit after hospitalization for acute worsening of dyspnea and orthopnea. On admission
Address reprint requests to Dr. Redfield
at Guggenheim 9, Mayo Clinic, 200 First to the hospital, the patient had atrial fibrillation with a ventricular rate of 120 beats
St. SW, Rochester, MN 55905, or at per minute, and chest radiography revealed pulmonary venous hypertension. Despite
redfield.margaret@mayo.edu. anticoagulation, rate control with a beta-blocker, and administration of loop diuret-
N Engl J Med 2016;375:1868-77. ics during the hospitalization, she continues to have fatigue and exertional dyspnea.
DOI: 10.1056/NEJMcp1511175 On physical examination, the body-mass index (BMI; the weight in kilograms divided
Copyright 2016 Massachusetts Medical Society.
by the square of the height in meters) is 39, pulse 76 beats per minute, and blood pres-
sure 160/70 mm Hg. There is jugular venous distention and lower-extremity edema but
no third heart sound, murmurs, or rales. The serum creatinine level is 1.4 mg per
deciliter (124 mol per liter), estimated glomerular filtration rate (GFR) 37 ml per min-
ute per 1.73 m2 of body-surface area, and N-terminal probrain natriuretic peptide
(NT-proBNP) level 300 pg per milliliter (age-specific and sex-specific normal range,
An audio version 10 to 218 pg per milliliter). Echocardiography reveals an ejection fraction of 70%, a
of this article normal left ventricular cavity dimension and wall thickness, and left atrial enlarge-
is available at ment. Doppler echocardiography shows elevated left atrial pressure (E/e ratio, 22)
NEJM.org
and an estimated pulmonary-artery systolic pressure of 52 mm Hg. How should this
patients condition be managed?
The Cl inic a l Probl em
E
pidemiologic studies indicate that up to 50% of patients with
heart failure have a preserved ejection fraction, and this proportion has in-
creased over time.1 In observational studies, rates of hospitalization and death
among patients who have heart failure with a preserved ejection fraction approach
those among patients who have heart failure with a reduced ejection fraction,1 but
in clinical-trial populations, outcomes are better in patients who have heart failure
with a preserved ejection fraction.2 Death from noncardiovascular causes is more
common in patients who have heart failure with a preserved ejection fraction than
in those with a reduced ejection fraction.3,4
Ventricular diastolic dysfunction (impaired relaxation and increased diastolic stiff-
ness) that is present at rest or induced by stress (from exercise, tachycardia, or hyper-
tension) is a central perturbation in heart failure with a preserved ejection frac-
tion.1,5-9 Although the ejection fraction is normal at rest, the ejection fraction does not
increase appropriately with stress,1 and other measures of systolic function are abnor-
mal.10 Endothelial dysfunction, arterial stiffening, and increased ventricular systolic
1868 48
n engl j med 375;19 nejm.org November 10, 2016

Clinical Pr actice
Key Clinical Points
Heart Failure with Preserved Ejection Fraction

In patients who have signs and symptoms of heart failure but a preserved ejection fraction, objective
evidence of abnormal cardiac structure and function should be confirmed by means of echocardiography,
electrocardiography, chest radiography, and measurement of natriuretic peptide levels.
Natriuretic peptide levels may be normal in patients who have heart failure with a preserved ejection
fraction, particularly in obese patients or those with symptoms only on exertion.
Right heart catheterization may be required in patients in whom there is indeterminate noninvasive
testing or evidence of pulmonary hypertension.
Medications that improve outcomes in patients who have heart failure with a reduced ejection fraction
have not been shown to be of benefit in those who have heart failure with a preserved ejection fraction.
Treatment of heart failure with a preserved ejection fraction should include diuretics for volume
overload, treatment for cardiovascular and noncardiovascular coexisting conditions, aerobic exercise
training to increase exercise tolerance, education regarding self-care, and disease management
programs for patients with refractory symptoms or frequent hospitalizations for heart failure.
stiffness are also common and may result in previous hospitalizations for or complicated by
heightened sensitivity to changes in load; this sen- heart failure. In some patients, heart failure mani-
sitivity manifests as rapid-onset pulmonary edema fests as unexplained exertional dyspnea. In such
with increases in load and excessive hypotension patients, differentiating heart failure from noncar-
with decreases in load.1 Exercise performance is diac dyspnea or deconditioning can be challeng-
impaired owing to impaired chronotropic, vaso- ing. In patients with suspected heart failure, com-
dilatory, and ventricular diastolic and systolic re- prehensive Doppler echocardiography should be
serve functions and impaired oxygen uptake and performed.
utilization in the peripheral muscles.5,11,12
The fundamental pathophysiological pertur- Echocardiographic Findings and Natriuretic
bation leading to heart failure with a preserved Peptide Levels
ejection fraction remains incompletely defined, but In observational studies and clinical trials, the
traditionally it has been attributed to hypertensive value used to define a preserved ejection frac-
left ventricular remodeling1 (Fig. 1). Systemic mi- tion has ranged from 40 to 55%, but current guide-
crovascular endothelial inflammation related to lines recommend a partition value of 50%.17,18 An
coexisting conditions has been proposed as an ejection fraction of 40 to 49% is a gray area.17
additional mechanism leading to myocardial in- Patients who previously had an ejection fraction
flammation and fibrosis, increases in oxidative of less than 40% but in whom the ejection frac-
stress, and alterations in cardiomyocyte signal- tion increased with therapy for heart failure are
ing pathways. These alterations promote cardio- considered to have recovered heart failure with
myocyte remodeling and dysfunction (Fig. 1)13,14 a reduced ejection fraction. In these patients,
as well as microvascular dysfunction and rarefac- medications for heart failure that have a proven
tion in cardiac15,16 and skeletal11,12 muscle (Fig. 1). benefit in patients with a reduced ejection frac-
tion should be continued.
If the ejection fraction is preserved, evidence
S t r ategie s a nd E v idence
of altered cardiac structure and function should
Diagnosis and Evaluation be sought to provide further objective evidence
Since signs and symptoms of heart failure are non- of heart failure (Fig. 2). The size of the left ven-
specific, clinicians should maintain a high index of tricular cavity is usually normal. Evidence of left
suspicion for heart failure in patients with risk ventricular hypertrophy (Fig. 2) is common but
factors, but they also should consider alternative or absent in many patients.8,19 Doppler echocardio-
contributing diagnoses (Fig. 2). The clinical his- graphic evidence of diastolic dysfunction (slowed
tory should include ascertainment of reduced ventricular relaxation and increased diastolic stiff-
symptoms in response to diuretic therapy and ness or elevated left atrial pressure) is common

49 November 10, 2016 1869

A Traditional Model B Emerging Model

Systemic hypertension
Proinflammatory coexisting conditions
Vascular dysfunction
Concentric hypertrophy
Systemic microvascular endothelial inflammation
Left Ventricle Fibrosis
Diastolic dysfunction
Left atrial hypertension
Increases in oxidative stress Microvascular dysfunction

Muscle inflammation
Remodeling Decreases in NOcyclic GMP and rarefaction
Left Atrium Diastolic dysfunction signaling
Systolic dysfunction
Pulmonary hypertension
Atrial fibrillation
Myofiber stiffness
Fibrosis
Cardiomyocyte hypertrophy
Remodeling
Right Ventricle Diastolic dysfunction
Right atrial hypertension Global cardiac remodeling and dysfunction

Impaired coronary flow reserve
Impaired oxygen delivery, uptake, and utilization
Remodeling in skeletal muscle
Right Atrium Diastolic dysfunction
Figure 1. Traditional and Emerging Pathophysiological Models of Heart Failure with Preserved Ejection Fraction.
Most patients who have heart failure with a preserved ejection fraction have a history of hypertension. In the traditional pathophysiolog-
ical model, pressure overload leads to concentric left ventricular hypertrophic and fibrotic remodeling and diastolic dysfunction. Ulti-
mately, the left ventricular diastolic dysfunction leads to left atrial hypertension and remodeling, pulmonary venous hypertension, and
right ventricular and atrial remodeling and dysfunction. Atrial fibrillation is common because of the chronic left atrial hypertension and
subsequent structural and electrical remodeling. In the emerging model, proinflammatory cardiovascular and noncardiovascular coex-
isting conditions (e.g., hypertension, obesity, diabetes, the metabolic syndrome, lung disease, smoking, and iron deficiency) lead to sys-
temic microvascular endothelial inflammation, global cardiac and skeletal-muscle inflammation, and subsequent fibrosis. These condi-
tions also lead to increases in oxidative stress that limit nitric oxidecyclic guanosine monophosphate (NOcyclic GMP)protein kinase
G signaling, promoting global cardiomyocyte hypertrophy and intrinsic myofiber stiffness. Finally, coronary microvascular inflammation
results in microvascular dysfunction and rarefaction with reduced microvascular density and coronary flow reserve. Similar changes oc-
cur in the skeletal-muscle vasculature with reduced oxygen delivery and utilization.
(Fig. 2).8,9,20 However, diastolic dysfunction also Atrial fibrillation is very common and may
may be present in patients who do not have heart precede, present concurrently with, or occur sub-
failure21 and absent in patients who have received sequent to the onset of heart failure with a pre-
aggressive treatment for heart failure or those served ejection fraction.24 Radiographic evidence
with predominantly exertional symptoms.6,7 The of heart failure (Fig. 2) is common in patients
left atrium is usually enlarged. Pulmonary-artery who present with acute heart failure, but radio-
systolic pressure, estimated by means of Doppler graphic evidence of heart failure is not necessar-
echocardiography, is often elevated (>35 mm Hg).22 ily present in patients who are in stable condi-
Right ventricular systolic dysfunction is present tion. Ventricular wall stress and thus circulating
in 20 to 30% of patients, often in association with levels of natriuretic peptides are lower in patients
atrial fibrillation.23 Atrial remodeling can lead to who have heart failure with a preserved ejection
annular dilatation and functional mitral and tri- fraction than in patients who have heart failure
cuspid regurgitation, but primary valvular disease with a reduced ejection fraction.25 Levels of na-
should be ruled out. triuretic peptides may be normal in up to 30%
1870 50
n engl j med 375;19 nejm.org November 10, 2016

Clinical Pr actice
Symptoms and signs of heart failure

Risk factors for heart failure:
>60 yr of age
Hypertension
Proinflammatory coexisting conditions
Previous hospitalization for heart failure
Cardiac imaging: ejection fraction 50%
Consider alternative or contributing

causes
Objective findings that support heart failure

(the more positive features, the greater the likelihood of heart failure with preserved ejection fraction)
Electrocardiography Doppler echocardiography Doppler echocardiography, continued

Left ventricular hypertrophy Cardiac remodeling Abnormal relaxation
Left atrial enlargement Relative wall thickness >0.42 e <8 (septal)
Atrial fibrillation Left ventricular mass >95 g/m2 Supportive findings (in absence
Chest radiography (current or past) in women or >115 g/m2 in men of other causes)
Cardiomegaly Left atrial volume >34 ml/m2 Pulmonary-artery systolic
Pulmonary venous hypertension Elevated left atrial pressure pressure >35 mm Hg
Interstitial or alveolar edema E/A 2.0 Right ventricular enlargement
Pleural effusion E/e ratio 15 (using septal e) or systolic dysfunction
Natriuretic peptide assay E-wave deceleration time 140 msec
BNP >100 pg/ml Decrease in E/A by 0.5 and to <1.0
NT-proBNP >400 pg/ml with Valsalva maneuver
Consider specialized testing in selected patients

Stress imaging or coronary angiography Cardiopulmonary exercise testing
Angina or cardiovascular risk factors Quantify severity of functional limitation
Right heart catheterization Exclude pulmonary limitation
Diagnosis in indeterminate cases Detect chronotropic incompetence
Define severity and mechanism of pulmonary Detect exaggerated hypertensive response
hypertension Technetium-99m pyrophosphate or technetium-99m
Exercise hemodynamic testing: to detect 3,3-diphosphono-1,2-propanedicarboxylic acid
abnormal exercise but normal resting scintigraphy
hemodynamic status Exclude suspected transthyretin cardiac
amyloidosis
Figure 2. Systematic Approach to the Diagnosis of Heart Failure with Preserved Ejection Fraction.
Ejection fraction values of 40 to 49% are considered to be a gray area but are probably consistent with heart failure
and a preserved ejection fraction in some patients. Common alternative or contributing conditions that should be
considered are cardiac ischemia due to epicardial coronary disease, lung disease, or pulmonary arterial hyperten-
sion that is unrelated to heart failure. Hypertrophic or infiltrative cardiomyopathy, pericardial disease, or uncorrect-
ed primary valve disease should be ruled out. In the outpatient setting, heart failure is less likely in patients with a
brain natriuretic peptide (BNP) level of less than 35 pg per milliliter or an N-terminal probrain natriuretic peptide
(NT-proBNP) level of less than 125 pg per milliliter than in patients with higher levels.17 However, patients who have
heart failure with a preserved ejection fraction can have normal natriuretic peptide levels, particularly if they are
obese or have only exertional symptoms. E/A denotes ratio of E wave to A wave, E/e ratio of early mitral inflow ve-
locity (E) to early diastolic mitral annular velocity detected by tissue Doppler imaging (e).
of patients who have heart failure with a pre- the more likely it is that the patient has heart
served ejection fraction,26 particularly in those failure (Fig. 2). However, some elderly patients28,29
who are obese27 or have purely exertional symp- or patients who have atrial fibrillation30 without
toms.6 The higher the natriuretic peptide level, heart failure may have natriuretic peptide levels

51 November 10, 2016 1871

that are similar to those of patients with heart suspected. Scintigraphy with specific radioactive
failure. tracers can also assist in the recognition of
transthyretin cardiac amyloidosis34 and should
Specialized Testing in Selected Patients be considered in older patients with increased
Specific cardiac conditions that can cause heart ventricular-wall thickness (12 mm) on echocar-
failure when a preserved ejection fraction is pres- diography.35
ent (e.g., pericardial disease and hypertrophic or Renal artery stenosis should be considered in
infiltrative cardiomyopathies) must be considered patients with risk factors for this condition (e.g.,
in the differential diagnosis in patients who have renal dysfunction or peripheral vascular disease)
heart failure with a preserved ejection fraction and a history of recurrent acute episodes of heart
(Fig. 2). Epicardial coronary atherosclerosis can failure with a preserved ejection fraction.36 In pa-
account for symptoms of heart failure with exer- tients who have a normal or only mildly elevated
tional dyspnea or angina, but angina is also com- creatinine level, the requirement for a high dose
mon in patients who do not have coronary dis- of a diuretic should prompt further evaluation of
ease.31 In most patients with coronary disease, the renal function (e.g., measurement of the cystatin C
coronary disease is of insufficient severity to ac- level).
count for the severity of heart failure, but it is a
risk factor for future coronary events and death.31 Treatment
Stress testing, coronary angiography, or both Since no therapy has been shown to improve out-
should be performed if the patient has symptoms comes in patients who have heart failure with a
of or risk factors for coronary artery disease and preserved ejection fraction, current therapy (Fig. 3)
is a candidate for anti-ischemic medications or includes the relief of volume overload (when pres-
revascularization. Standard exercise stress testing ent), treatment of coexisting conditions, addition-
provides information about functional limitation al strategies that may increase exercise tolerance
and about the possibility of chronotropic incom- or reduce symptoms, and strategies to manage
petence or exaggerated hypertensive response to chronic disease and prevent hospitalizations.
exercise. Cardiopulmonary exercise testing can
be useful to rule out noncardiac limitations to Trials of Therapies to Improve Outcomes
exercise such as poor effort, deconditioning, and Individually or in a meta-analysis, three random-
pulmonary disease. Pulmonary-artery catheteriza- ized trials of angiotensin antagonists (angiotensin-
tion with or without exercise may be needed to convertingenzyme [ACE] inhibitors or angio-
establish the diagnosis in patients in whom the tensin-receptor antagonists) involving patients
findings of noninvasive studies are indeterminate who had heart failure with a preserved ejection
or to document the severity and mechanism of fraction did not show significant effects of these
pulmonary hypertension when pulmonary-artery agents on composite end points of all-cause or
systolic pressure estimated with Doppler echocar- cardiovascular mortality and hospitalizations for
diography is significantly elevated (>50 mm Hg). heart failure.37 The mineralocorticoid-receptor an-
Pulmonary hypertension in heart failure is due tagonist spironolactone did not reduce rates of the
to pulmonary venous hypertension and sometimes primary composite outcome of death from cardio-
modest increases (2 to 4 Wood units) in pulmo- vascular causes, aborted cardiac arrest, or hospi-
nary vascular resistance22; higher values should talization for heart failure in these patients.38
spur evaluation of other causes contributing to Spironolactone reduced the rate of hospitalization
pulmonary hypertension. Large V waves (twice for heart failure but not the rate of death from
the mean pulmonary arterial wedge pressure any cause or hospitalization for any cause, and it
value and >25 mm Hg) in the pulmonary arterial increased the rate of renal dysfunction and hyper-
wedge pressure wave forms at rest or with stress kalemia. Analyses that were limited to patients
(in the absence of marked mitral regurgitation) who were enrolled in centers in the Americas
indicate reduced left atrial compliance, a hemo- (which had higher event rates) showed beneficial
dynamic hallmark of this condition.32,33 effects of spironolactone on the composite pri-
Cardiac magnetic resonance imaging may be mary end point, but these post hoc analyses must
useful if infiltrative cardiomyopathy (amyloidosis) be interpreted with caution.39The effect of beta-
or inflammatory cardiomyopathy (sarcoidosis) is blockers in patients with heart failure and a pre-

52 November 10, 2016

Clinical Pr actice
Volume overload
Yes No
Diuretics
Evaluate and manage underlying cardiovascular diseases and coexisting conditions
Elevated cardio- Lung disease

Hypertension Coronary disease Atrial fibrillation Obesity Kidney disease
vascular risk Sleep apnea
Diuretics Statins according Statins Rate control Behavioral ACE or ARB Therapy according
ACE or ARB to guidelines Medical therapy Anticoagulation strategies (for hypertension) to guidelines
(if patient has Consider revas- according to Pharmacotherapy
chronic kidney cularization risk scores Surgery
disease)
Other agents
according to
side effects and
effectiveness
Consider rhythm control

for persistent symptoms
Education regarding heart failure and self-care

Aerobic exercise training
Recent
Persistent symptoms
hospitalization
Consider disease management Referral to clinical trials of agents and devices for heart
program for heart failure failure with preserved ejection fraction
Consider pulmonary-artery
pressureguided
management
Figure 3. Treatment Algorithm for Heart Failure with Preserved Ejection Fraction.
ACE denotes angiotensin-convertingenzyme inhibitor, and ARB angiotensin-receptor blocker.
served ejection fraction has not been evaluated beta-blockers in the treatment of patients who
in an adequately powered study, and the limited have heart failure with a preserved ejection frac-
available data are conflicting.40-43 tion should be limited to patients who have al-
Thus, the use of angiotensin antagonists and ternative indications for their use. The use of
n engl j med 375;19 nejm.org 53

November 10, 2016 1873

spironolactone in patients who have heart failure hypertension and concomitant kidney disease
with a preserved ejection fraction remains con- should receive an angiotensin antagonist, re-
troversial. gardless of their race or diabetic status44 (Fig. 3).
In patients who do not have concomitant kidney
Treatment of Volume Overload disease, a thiazide-like diuretic, angiotensin an-
Diuretics, which should be used for relief of symp- tagonist, or calcium-channel blocker for non-
toms in patients with volume overload, should be blacks and a thiazide-like diuretic or calcium-
adjusted according to the patients body weight, channel blocker for blacks are appropriate for
symptoms, and electrolyte status. Intermittent use initial management.44 Aggressive use of vasodi-
of a thiazide-like diuretic such as metolazone, lators may lead to unacceptable side effects in
administered before the dose of a loop diuretic, patients with heart failure with a preserved ejec-
may be helpful in outpatients with volume over- tion fraction. The choice of additional agents to
load that is refractory to higher doses of loop achieve blood-pressure control should be guided
diuretics. However, the use of this agent calls for by the presence of coexisting conditions, the pa-
careful monitoring because of the risk of hypo- tients ability to receive the agent without adverse
kalemia, hyponatremia, and worsening renal effects, and the effect of the agent on blood
function. Persistent diuretic resistance may result pressure.
from impaired diuretic absorption, necessitating Patients should be treated with statins accord-
intravenous administration of loop diuretics. ing to the usual criteria. Observational studies,
Although the evidence base is limited,17,18 so- including a propensity-scorematched analysis,46
dium restriction (to 2 g per day) may be helpful have shown lower mortality among patients with
in patients who are prone to volume overload. At heart failure with a preserved ejection fraction
a minimum, high-sodium diets (>6 g per day) and who have received statins than among those who
rapid fluctuations in sodium intake should be have not received statins, but it remains unclear
avoided.17,18 whether this association is causal.
Patients with coronary artery disease should
Treatment of Coexisting Conditions receive medical therapies according to current
Data to guide treatment of coexisting conditions guidelines.47 Limited (and potentially confound-
and risk factors specifically in patients with heart ed) observational data in patients who have
failure and a preserved ejection fraction are very heart failure with a preserved ejection fraction
limited. Hypertension can exacerbate heart failure and coronary disease have suggested better out-
and predispose patients to other adverse outcomes among those who have undergone com-
comes.18 The Eighth Joint National Committee plete revascularization than among those who
guidelines do not include a specific blood- have not.31 Revascularization can be considered
pressure target for persons with heart failure. for symptom relief in patients who are otherwise
However, they recommend target blood pres- eligible for this procedure and who have clini-
sures of less than 150/90 mm Hg in persons cally significant angina or in whom clinically
who are 60 years of age or older in the general significant ischemia is evident and thought to
population44 and of less than 140/90 mm Hg in contribute to dyspnea as an angina equivalent.18
persons with kidney disease (estimated GFR, Atrial fibrillation should be managed accord-
<60 ml per minute per 1.73 m2 of body-surface ing to current guidelines, which recommend
area or >30 mg of albumin per gram of creatinine, rate control and anticoagulation initially, and a
regardless of diabetic status) and for persons with trial of rhythm control should be considered if
diabetes, regardless of age. A recent trial showed symptoms persist despite adequate rate con-
that lower rates of cardiovascular events and death trol.17,18,48 Patients may be most likely to benefit
were associated with blood-pressure targets lower from rhythm control if the symptoms of heart
than those recommended by current guidelines, failure started or worsened after the onset of
but the trial did not enroll patients with heart atrial fibrillation.
failure.45 Obesity may contribute to exercise intoler-
Most patients with heart failure and hyper- ance. In a small randomized trial, intentional
tension will require a diuretic. All patients with weight loss significantly increased exercise toler-


Clinical Pr actice
ance but did not increase a heart failurespe- The effect of remote-monitoring strategies is
cific quality-of-life score in obese patients who unclear. However, a randomized trial of pulmo-
had heart failure with a preserved ejection frac- nary-artery pressureguided management in pa-
tion.49 To increase exercise tolerance, weight loss tients with heart failure showed that this strat-
in obese patients (BMI, 35) with heart failure egy reduced hospitalizations for heart failure in
should be considered.17 patients with a reduced or a preserved ejection
Lung disease and disordered breathing dur- fraction.52
ing sleep are common comorbid conditions in
patients with heart failure, provoke symptoms A r e a s of Uncer ta in t y
(dyspnea and fatigue) that are similar to those of
heart failure, and may exacerbate hypertension Owing to positive findings in a phase 2 study,53
and heart failure. Thus, aggressive treatment of a large outcomes trial of a neprilysinangioten-
concomitant lung disease and sleep apnea ac- sin-receptor inhibitor (sacubitrilvalsartan) in
cording to current guidelines is reasonable. patients with heart failure and a preserved ejec-
tion fraction is ongoing (ClinicalTrials.gov num-
Other Therapies to Reduce Symptoms or Increase ber, NCT01920711). Information from ongoing
Exercise Tolerance phase 2, randomized trials of a variety of other
Nitrates are often prescribed for patients who drugs and medical devices in patients with heart
have heart failure and a preserved ejection frac- failure and a preserved ejection fraction is
tion. However, a randomized, placebo-controlled needed.54
trial of isosorbide mononitrate did not show in- The incidence of ventricular arrhythmias and
creases in submaximal exercise capacity or qual- the role of implantable defibrillators are unknown.
ity-of-life scores in these patients.50 The most appropriate strategies for the treatment
In small studies, exercise training has consis- of hypertension, obesity, diabetes, atrial fibrilla-
tently been shown to produce clinically meaning- tion, iron deficiency, anemia, and coronary dis-
ful increases in exercise capacity and a reduction ease in patients with heart failure and a pre-
in symptoms.49,51 Cardiac rehabilitation programs served ejection fraction have not been defined.
are reimbursed by U.S. government payers for
patients who have heart failure with a reduced Guidel ine s
ejection fraction but not for those with a preserved
ejection fraction. Clinicians should recommend a Recently updated guidelines for the management
daily target of 30 minutes of aerobic exercise of heart failure with a preserved ejection fraction
tailored to the abilities and resources particular are available.17,18 The recommendations in this ar-
to each patient and should monitor compliance ticle are largely consistent with those guidelines.
and address barriers to exercise training in on-
going follow-up.17,18 C onclusions a nd
R ec om mendat ions
Disease Management
All patients with heart failure should receive edu- The patient in the vignette has heart failure with
cation regarding self-care. Self-care includes moni- a preserved ejection fraction, exacerbated by, but
toring of weight and symptoms, adjustment of probably predating, the onset of atrial fibrilla-
doses of diuretics, compliance with dietary re- tion. The dose of diuretics should be increased to
strictions, use of medications, exercise, and regu- reduce the patients clinical congestion. Given her
lar follow-up. hypertension and renal dysfunction, an angioten-
In patients with refractory symptoms or fre- sin antagonist should be added and other agents
quent hospitalizations for heart failure, referral to used as needed to achieve a blood pressure of less
a disease management program should be con- than 140/90 mm Hg. She should receive education
sidered. In patients who do not have a response regarding self-care for heart failure. Anticoagula-
to aggressive management, a palliative care pro- tion should be continued. If symptoms persist, a
gram for symptom management and assistance trial of rhythm control should be considered.
in end-of-life planning should be considered.18 The patients atherosclerotic risk and the

55 November 10, 2016 1875

presence of coronary disease should be assessed formed about clinical trials of therapeutic strate-
to guide the use of statins and other treatments gies for heart failure with a preserved ejection
for coronary disease. Evaluation for sleep apnea fraction.
may also be reasonable, given her obesity, fa-
tigue, hypertension, and atrial fibrillation. Once Dr. Redfield reports serving as a member of a scientific com-
mittee for Novartis (unpaid) and receiving consulting fees from
her condition is stable, exercise and weight-loss Merck, Eli Lilly, and Actelion, fees for serving on a data and
programs should be commenced. Persistent symp- safety monitoring board for Covaria, and grant support from St.
toms or recurrent hospitalizations should prompt Jude Medical and Medtronic. No other potential conflict of in-
terest relevant to this article was reported.
referral to a disease management program for Disclosure forms provided by the authors are available with
patients with heart failure. She should be in- the full text of this article at NEJM.org.
References
1. Gladden JD, Linke WA, Redfield MM. et al. Mechanisms of exercise intolerance Lam CS, Redfield MM, Nishimura RA.
Heart failure with preserved ejection frac- in heart failure with preserved ejection Diastolic relaxation and compliance re-
tion. Pflugers Arch 2014;466:1037-53. fraction: the role of abnormal peripheral serve during dynamic exercise in heart
2. Campbell RT, Jhund PS, Castagno D, oxygen extraction. Circ Heart Fail 2015;8: failure with preserved ejection fraction.
Hawkins NM, Petrie MC, McMurray JJ. 286-94. Heart 2011;97:964-9.
What have we learned about patients with 12. Haykowsky MJ, Tomczak CR, Scott 21. Redfield MM, Jacobsen SJ, Burnett JC
heart failure and preserved ejection frac- JM, Paterson DI, Kitzman DW. Determi- Jr, Mahoney DW, Bailey KR, Rodeheffer
tion from DIG-PEF, CHARM-preserved, nants of exercise intolerance in patients RJ. Burden of systolic and diastolic ven-
and I-PRESERVE? J Am Coll Cardiol 2012; with heart failure and reduced or pre- tricular dysfunction in the community:
60:2349-56. served ejection fraction. J Appl Physiol appreciating the scope of the heart failure
3. Chan MM, Lam CS. How do patients 2015;119:739-44. epidemic. JAMA 2003;289:194-202.
with heart failure with preserved ejection 13. Paulus WJ, Tschpe C. A novel para- 22. Thenappan T, Prins KW, Cogswell R,
fraction die? Eur J Heart Fail 2013;15:604- digm for heart failure with preserved ejec- Shah SJ. Pulmonary hypertension second-
13. tion fraction: comorbidities drive myocar- ary to heart failure with preserved ejec-
4. Lee DS, Gona P, Albano I, et al. A sys- dial dysfunction and remodeling through tion fraction. Can J Cardiol 2015;31:430-
tematic assessment of causes of death af- coronary microvascular endothelial in- 9.
ter heart failure onset in the community: flammation. J Am Coll Cardiol 2013;62: 23. Chatterjee NA, Steiner J, Lewis GD. It
impact of age at death, time period, and 263-71. is time to look at heart failure with pre-
left ventricular systolic dysfunction. Circ 14. Shah SJ, Kitzman DW, Borlaug BA, et served ejection fraction from the right
Heart Fail 2011;4:36-43. al. Phenotype-specific treatment of heart side. Circulation 2014;130:2272-7.
5. Borlaug BA. Mechanisms of exercise failure with preserved ejection fraction: 24. Zakeri R, Chamberlain AM, Roger VL,
intolerance in heart failure with preserved a multiorgan roadmap. Circulation 2016; Redfield MM. Temporal relationship and
ejection fraction. Circ J 2014;78:20-32. 134:73-90. prognostic significance of atrial fibrilla-
6. Borlaug BA, Nishimura RA, Sorajja P, 15. Mohammed SF, Hussain S, Mirzoyev tion in heart failure patients with pre-
Lam CS, Redfield MM. Exercise hemody- SA, Edwards WD, Maleszewski JJ, Red- served ejection fraction: a community-
namics enhance diagnosis of early heart field MM. Coronary microvascular rar- based study. Circulation 2013;128:1085-93.
failure with preserved ejection fraction. efaction and myocardial fibrosis in heart 25. Iwanaga Y, Nishi I, Furuichi S, et al.
Circ Heart Fail 2010;3:588-95. failure with preserved ejection fraction. B-type natriuretic peptide strongly re-
7. Franssen C, Paulus WJ. Normal rest- Circulation 2015;131:550-9. flects diastolic wall stress in patients with
ing pulmonary artery wedge pressure: a 16. Mohammed SF, Majure DT, Redfield chronic heart failure: comparison be-
diagnostic trap for heart failure with pre- MM. Zooming in on the microvasculature tween systolic and diastolic heart failure.
served ejection fraction. Eur J Heart Fail in heart failure with preserved ejection J Am Coll Cardiol 2006;47:742-8.
2015;17:132-4. fraction. Circ Heart Fail 2016;9(7): 26. Anjan VY, Loftus TM, Burke MA, et al.
8. Lam CS, Roger VL, Rodeheffer RJ, et e003272. Prevalence, clinical phenotype, and out-
al. Cardiac structure and ventricular-vas- 17. Ponikowski P, Voors AA, Anker SD, et comes associated with normal B-type na-
cular function in persons with heart fail- al. 2016 ESC guidelines for the diagnosis triuretic peptide levels in heart failure
ure and preserved ejection fraction from and treatment of acute and chronic heart with preserved ejection fraction. Am J
Olmsted County, Minnesota. Circulation failure. Eur Heart J 2016;37:2129-200. Cardiol 2012;110:870-6.
2007;115:1982-90. 18. Yancy CW, Jessup M, Bozkurt B, et al. 27. Bishu K, Deswal A, Chen HH, et al.
9. Zile MR, Baicu CF, Gaasch WH. Dia- 2013 ACCF/AHA guideline for the man- Biomarkers in acutely decompensated
stolic heart failure abnormalities in agement of heart failure: a report of the heart failure with preserved or reduced
active relaxation and passive stiffness of American College of Cardiology Founda- ejection fraction. Am Heart J 2012;164(5):
the left ventricle. N Engl J Med 2004;350: tion/American Heart Association Task 763-770.e3.
1953-9. Force on Practice Guidelines. Circulation 28. Costello-Boerrigter LC, Boerrigter G,
10. Borlaug BA, Lam CS, Roger VL, Rode- 2013;128(16):e240-327. Redfield MM, et al. Amino-terminal pro-
heffer RJ, Redfield MM. Contractility and 19. Zile MR, Gottdiener JS, Hetzel SJ, et B-type natriuretic peptide and B-type na-
ventricular systolic stiffening in hyper- al. Prevalence and significance of altera- triuretic peptide in the general commu-
tensive heart disease: insights into the tions in cardiac structure and function in nity: determinants and detection of left
pathogenesis of heart failure with pre- patients with heart failure and a pre- ventricular dysfunction. J Am Coll Cardiol
served ejection fraction. J Am Coll Cardiol served ejection fraction. Circulation 2011; 2006;47:345-53.
2009;54:410-8. 124:2491-501. 29. Redfield MM, Rodeheffer RJ, Jacob-
11. Dhakal BP, Malhotra R, Murphy RM, 20. Borlaug BA, Jaber WA, Ommen SR, sen SJ, Mahoney DW, Bailey KR, Burnett


Clinical Pr actice
JC Jr. Plasma brain natriuretic peptide atic review and meta-analysis. J Card Fail 46. Alehagen U, Benson L, Edner M,
concentration: impact of age and gender. 2010;16:260-7. Dahlstrm U, Lund LH. Association be-
J Am Coll Cardiol 2002;40:976-82. 38. Pfeffer MA, Pitt B, McKinlay SM. Spi- tween use of statins and mortality in pa-
30. Richards M, Di Somma S, Mueller C, ronolactone for heart failure with pre- tients with heart failure and ejection
et al. Atrial fibrillation impairs the diag- served ejection fraction. N Engl J Med fraction of 50. Circ Heart Fail 2015;8:
nostic performance of cardiac natriuretic 2014;371:181-2. 862-70.
peptides in dyspneic patients: results 39. Pfeffer MA, Claggett B, Assmann SF, 47. Ohman EM. Chronic stable angina.
from the BACH Study (Biomarkers in et al. Regional variation in patients and N Engl J Med 2016;374:1167-76.
ACute Heart Failure). JACC Heart Fail outcomes in the Treatment of Preserved 48. January CT, Wann LS, Alpert JS, et al.
2013;1:192-9. Cardiac Function Heart Failure With an 2014 AHA/ACC/HRS guideline for the
31. Hwang SJ, Melenovsky V, Borlaug BA. Aldosterone Antagonist (TOPCAT) trial. management of patients with atrial fibril-
Implications of coronary artery disease in Circulation 2015;131:34-42. lation: a report of the American College of
heart failure with preserved ejection frac- 40. Conraads VM, Metra M, Kamp O, et Cardiology/American Heart Association
tion. J Am Coll Cardiol 2014;63:2817-27. al. Effects of the long-term administra- Task Force on practice guidelines and the
32. Pichard AD, Diaz R, Marchant E, tion of nebivolol on the clinical symp- Heart Rhythm Society. Circulation 2014;
Casanegra P. Large V waves in the pulmo- toms, exercise capacity, and left ventricu- 130(23):e199-267.
nary capillary wedge pressure tracing lar function of patients with diastolic 49. Kitzman DW, Brubaker P, Morgan T,
without mitral regurgitation: the influ- dysfunction: results of the ELANDD et al. Effect of caloric restriction or aero-
ence of the pressure/volume relationship study. Eur J Heart Fail 2012;14:219-25. bic exercise training on peak oxygen con-
on the V wave size. Clin Cardiol 1983;6: 41. Liu F, Chen Y, Feng X, Teng Z, Yuan Y, sumption and quality of life in obese old-
534-41. Bin J. Effects of beta-blockers on heart er patients with heart failure with
33. Rossi A, Gheorghiade M, Triposkiadis failure with preserved ejection fraction: a preserved ejection fraction: a randomized
F, Solomon SD, Pieske B, Butler J. Left meta-analysis. PLoS One 2014;9(3): clinical trial. JAMA 2016;315:36-46.
atrium in heart failure with preserved ejec- e90555. 50. Redfield MM, Anstrom KJ, Levine JA,
tion fraction: structure, function, and sig- 42. van Veldhuisen DJ, Cohen-Solal A, et al. Isosorbide mononitrate in heart fail-
nificance. Circ Heart Fail 2014;7:1042-9. Bhm M, et al. Beta-blockade with nebivo- ure with preserved ejection fraction.
34. Gillmore JD, Maurer MS, Falk RH, et lol in elderly heart failure patients with N Engl J Med 2015;373:2314-24.
al. Nonbiopsy diagnosis of cardiac trans- impaired and preserved left ventricular 51. Pandey A, Parashar A, Kumbhani DJ,
thyretin amyloidosis. Circulation 2016; ejection fraction: data from SENIORS et al. Exercise training in patients with
133:2404-12. (Study of Effects of Nebivolol Intervention heart failure and preserved ejection frac-
35. Gonzlez-Lpez E, Gallego-Delgado on Outcomes and Rehospitalization in tion: meta-analysis of randomized con-
M, Guzzo-Merello G, et al. Wild-type Seniors With Heart Failure). J Am Coll trol trials. Circ Heart Fail 2015;8:33-40.
transthyretin amyloidosis as a cause of Cardiol 2009;53:2150-8. 52. Adamson PB, Abraham WT, Bourge
heart failure with preserved ejection frac- 43. Yamamoto K, Origasa H, Hori M. Ef- RC, et al. Wireless pulmonary artery pres-
tion. Eur Heart J 2015;36:2585-94. fects of carvedilol on heart failure with sure monitoring guides management to
36. Anderson JL, Halperin JL, Albert NM, preserved ejection fraction: the Japanese reduce decompensation in heart failure
et al. Management of patients with pe- Diastolic Heart Failure Study (J-DHF). Eur with preserved ejection fraction. Circ
ripheral artery disease (compilation of J Heart Fail 2013;15:110-8. Heart Fail 2014;7:935-44.
2005 and 2011 ACCF/AHA guideline rec- 44. James PA, Oparil S, Carter BL, et al. 53. Solomon SD, Zile M, Pieske B, et al.
ommendations): a report of the American 2014 Evidence-based guideline for the The angiotensin receptor neprilysin in-
College of Cardiology Foundation/Ameri- management of high blood pressure in hibitor LCZ696 in heart failure with pre-
can Heart Association Task Force on Prac- adults: report from the panel members served ejection fraction: a phase 2 double-
tice Guidelines. Circulation 2013;127: appointed to the Eighth Joint National blind randomised controlled trial. Lancet
1425-43. Committee (JNC 8). JAMA 2014;311:507- 2012;380:1387-95.
37. Shah RV, Desai AS, Givertz MM. The 20. 54. Nanayakkara S, Kaye DM. Manage-
effect of renin-angiotensin system inhibi- 45. Wright JT Jr, Williamson JD, Whelton ment of heart failure with preserved ejec-
tors on mortality and heart failure hospi- PK, et al. A randomized trial of intensive tion fraction: a review. Clin Ther 2015;37:
talization in patients with heart failure versus standard blood-pressure control. 2186-98.
and preserved ejection fraction: a system- N Engl J Med 2015;373:2103-16. Copyright 2016 Massachusetts Medical Society.
n engl j med 375;19 nejm.org November 10, 2016 1877

57

INTERACTIVE MEDICAL CASES

Each interactive case presents an evolving patient history and a series of questions and exercises designed to test
diagnostic and therapeutic skills. Immediate feedback is given on answers and treatment choices, along with the
opportunity to compare final scores with peers. Video, animation, and interactive content allows one to learn
more about mechanisms, diagnostic tests, and treatments.
58

UNDER PRESSURE
Play the Interactive Medical Case now. An internet connection is required.

The following text summarizes information provided in the Interactive Medical Case.
E.M. DeFilippis, M.D., A. Vaidya, M.D., S. Ramani, M.D., B.S., C. French, M.D., and J.M. Kirshenbaum, M.D.
C a se Pr e sen tat ion

A 61-year-old man presented to the emergency department with cough and shortness of breath. Earlier that
morning, he had had sudden onset of severe shortness of breath and tightness in his chest. The chest tightness
had resolved, but he continued to feel short of breath. He reported no fevers but did have a productive cough.
The patient also reported increasing shortness of breath when he was lying flat and waking from sleep, with the
sudden need to sit upright in order to breathe. To breathe comfortably, he had to sleep on two or three pillows.
The patient also described a 2-month history of worsening fatigue, along with swelling of the legs and an
unintentional weight loss of 6.5 kg. He used to walk about 5 km per day without difficulty, but now had chest
tightness and shortness of breath when climbing a few stairs or walking a few steps.
59

DISSECTING A CASE OF ABDOMINAL PAIN
Play the Interactive Medical Case now. An internet connection is required.

The following text summarizes information provided in the Interactive Medical Case.
J.Casey, M.D., A. Vaidya, M.D., N. Frank, M.D., J.A. Beckman, M.D., and A. Miller, M.D., Ph.D
C a se Pr e sen tat ion

A 43-year-old man with no notable medical history presented to the emergency department within 1 hour after
the abrupt onset of abdominal pain. The patient stated that he had been dressing for work when severe, crampy
abdominal pain occurred in the left upper quadrant. On a scale of 0 to 10, with 0 indicating no pain and 10 the
worst imaginable pain, the patient rated the pain at 10. After the onset of pain, the patient had nausea, several
episodes of nonbilious, nonbloody emesis, and flatus.
60

Cases in Primary Care

Încărcat de

Informații document

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Cases in Primary Care

Încărcat de

Drepturi de autor:

Formate disponibile

CASES IN

CASES IN PRIMARY CARE

800.843.6356 | f: 781.891.1995 | nejmgroup@mms.org

CASE RECORDS OF THE MASSACHUSETTS GENERAL HOSPITAL

CASE RECORDS OF THE MASSACHUSETTS GENERAL HOSPITAL

Back to Table of Contents

Case Records of the Massachusetts General Hospital

Founded by Richard C. Cabot

Case 10-2017: A 6-Month-Old Boy

Pr e sen tat ion of C a se

n engl j med 376;13 nejm.org5 March 30, 2017 1269

Back to Table of Contents

Table 1. Laboratory Data.*

1270 n engl j med6376;13 nejm.org March 30, 2017

Back to Table of Contents

Differ en t i a l Di agnosis overdose that leads to liver failure and gastroin-

n engl j med 376;13 7

Back to Table of Contents

1272 n engl j med 376;13 nejm.org March 30, 2017

Back to Table of Contents

n engl j med 376;13 nejm.org March 30, 2017 1273

Back to Table of Contents

20 min 25 min 30 min

35 min 40 min 45 min

1274 n engl j med 376;13 nejm.org March 30, 2017

Back to Table of Contents

Figure 1 (facing page). Technetium-99m Pertechnetate

treat a Meckels diverticulum that causes gastro-

n engl j med 376;13 nejm.org

Back to Table of Contents

mucous neck cells, parietal cells, and basal chief

1276 n engl j med 376;13 nejm.org March 30, 2017

Back to Table of Contents

n engl j med 376;13 nejm.org March 30, 2017 1277

Back to Table of Contents

Case Records of the Massachusetts General Hospital

Founded by Richard C. Cabot

Case 2-2017: An 18-Year-Old Woman

Pr e sen tat ion of C a se

Back to Table of Contents

n engl j med 376;3 nejm.org January 19, 2017 269

Back to Table of Contents

Table 1. Laboratory Data.*

Reference Range, On Presentation, Reference Range, On Presentation,

270 n engl j med 376;3 nejm.org January 19, 2017

Back to Table of Contents

Figure 1. Abdominal Ultrasound Images.

n engl j med 376;3 nejm.org January 19, 2017 271

Back to Table of Contents

Figure 2. Causes of Acute Liver Failure in Children and Adults.

272 n engl j med 376;3 nejm.org

Back to Table of Contents

n engl j med 376;3 nejm.org January 19, 2017 273

Back to Table of Contents

274 n engl j med 376;3 20

Back to Table of Contents

n engl j med 376;3 nejm.org January 19, 2017 275

Back to Table of Contents

Figure 3. Explanted Liver.

276 n engl j med 376;3 nejm.org January 19, 2017

Back to Table of Contents

tannish-brown to yellow (Fig. 3A). On micro- undergo transplantation.22 Wilsons disease is

n engl j med 376;3 nejm.org January 19, 2017 277

Back to Table of Contents