Sunteți pe pagina 1din 11

Journal of Infection (2014) 68, S83eS93

Prolonged and recurrent fevers in children

Gary S. Marshall*

University of Louisville School of Medicine, 571 S. Floyd St., Suite 321, Louisville, KY 40202, USA

Accepted 20 September 2013

Available online 10 October 2013

KEYWORDS Summary Some children referred for prolonged fever are actually not having elevated tem-
Fever of unknown peratures; the approach here requires dissection of the history and correction of health mis-
origin; perceptions. Others have well-documented fevers associated with clinical, laboratory, or
Periodic fever; epidemiologic findings that should point to a specific diagnosis. “Fever-of-Unknown-Origin”
Occult infection; (FUO) is the clinical scenario of daily fever for 14 days that defies explanation after a careful
Autoinflammatory history, physical examination, and basic laboratory tests. The diagnostic approach requires a
diseases; meticulous fever diary, serial clinical and laboratory evaluations, vigilance for the appearance
PFAPA syndrome of new signs and symptoms, and targeted investigations; the pace of the work-up is deter-
mined by the severity of the illness. Approximately half of children with FUO will have a
self-limited illness and will never have a specific diagnosis made; the other half will ultimately
be found to have, in order, infectious, inflammatory, or neoplastic conditions. Irregular, inter-
mittent, recurrent fevers in the well-appearing child are likely to be sequential viral illnesses.
Monogenic autoinflammatory diseases should be considered in those who do not fit the picture
of recurrent infections and who do not have hallmarks of immune deficiency. Stereotypical
febrile illnesses that recur with clockwork periodicity should raise the possibilities of cyclic
neutropenia, if the cycle is approximately 21 days, or periodic fever, aphthous stomatitis,
pharyngitis, and adenitis (PFAPA) syndrome, the most common periodic fever in childhood.
ª 2013 The British Infection Association. Published by Elsevier Ltd. All rights reserved.

Introduction axilla, tympanic membrane, temporal arterydvary in the de-

gree (no pun intended) to which they approximate core body
temperature.3 Technique also matters; for example, oral tem-
In 1871, Wunderlich established 98.6  F (range, 97.2e99.5)
peratures are only accurate in those who can keep their
(37.0  C, range 36.2e37.5), as the normal human body
mouths closed and the thermometer under their tongues.
temperature, suggesting that temperatures above 100.4  F
Fever, or elevated core body temperature, is one of the
(38.0  C) were “suspicious”.1 Contemporary authors have
most reliable indicators that something is wrong, yet all of
argued that, given the wide variation between and within in-
the above make it difficult to define exactly what fever is.
dividuals, the concept of a “normal” body temperature should
Most pediatricians would agree that inserting a glass or
be abandoned.2 Sites of measurementdrectum, mouth,

* Tel.: þ1 (502) 852 3774; fax: þ1 (502) 852 3939.

E-mail address:

0163-4453/$36 ª 2013 The British Infection Association. Published by Elsevier Ltd. All rights reserved.
S84 G.S. Marshall

digital thermometer into the rectum is the most accurate expected for a self-limited viral illness, i.e., beyond a
way to measure core body temperature. This is feasible, week or so.
however, only up to about 3 years of age. Oral tempera-
tures, if properly done, are a close-enough approximation
to core body temperature; this technique becomes feasible Not-Fever-of-Unknown-Origin
around 5 years of age. Between 3 and 5, one may have to
settle for axillary temperatures (most physicians add 1  F Some such patients will actually not be having fever, at least
[0.6  C] to approximate core body temperature). The not according to any medical definition, despite the com-
Brighton Collaboration, a research network focused on plaintdthis might be termed Not-Fever-of-Unknown-Origin,
vaccine safety, defines fever as “the endogenous elevation to distinguish it from (true) Fever-of-Unknown-Origin, as
of at least one measured body temperature of 38  C,”4 ir- below. In evaluating patients with the complaint of pro-
respective of device used, anatomic site, age, or environ- longed fever, providers should be aware that elevated tem-
mental conditions. peratures are seen with ovulation and after meals, exercise,
For purposes of this discussion, undifferentiated fever and even chewing gum. Moreover, there is a normal diurnal
refers to the situation where a previously healthy child pre- variation in temperature, with lows in the morning and highs
sents with fever as the chief complaint, without signs or in the afternoondin some healthy adults, temperatures
symptoms of a specific clinical illness (e.g., upper respira- during the day may fluctuate by as much as 2.4  F
tory tract infection, viral exanthem, etc.). A distinction is (1.3  C).2 There are usually clues to the diagnosis of Not-Fe-
drawn between undifferentiated fever and fever without ver-of-Unknown-Origin6: the child appears healthy and has a
source, which refers to the special situation of an acute fe- normal physical exam; the weight is stable and growth and
ver in an otherwise well-appearing child under 3 years of development are normal. The parents may have misconcep-
age, where there is well-documented risk for occult serious tions about health and illnessdin particular, what tempera-
bacterial infection.5 tures constitute true fever (parents may say, for example,
Fig. 1 shows a conceptual framework for categorizing that 99  F [37.2  C] is a fever because the child’s tempera-
patients who present with undifferentiated fever. tures “normally run low”). Complicated medical histories,
behavioral problems, and multiple school absences for sub-
jective complaints may point to a vulnerable child. Family
Prolonged fever stressors and fear of malignancy may be present. The
approach should in carefully documenting the “fever” his-
In prolonged fever, the patient comes to medical attention tory, probing for underlying causes, correcting misconcep-
because the illness has lasted beyond what might be tions, and reassuring parents.


Prolonged Recurrent

Not-Fever-of- Fever-of-Not- Fever-of- Intermittent

Periodic Fever
Unknown-Origin Unknown-Origin Unknown-Origin Fever

Patient not Diagnosis Sequential

Daily fever Cyclic
truly having readily self-limited
≥14 days neutropenia
fevers achievable illnesses
presentation PFAPA
Infectious inflammatory
of common syndrome
presentation Inflammatory
of unusual



No diagnosis

Figure 1 Algorithm for undifferentiated fevers in children (See text for explanation).
Prolonged and recurrent fevers in children S85

Fever-of-Not-Unknown-Origin the classic 1961 paper by Petersdorf and Beeson, which re-
ported a series of 100 adults with illness greater than 3
Other patients will present with bona-fide prolonged fever, weeks, fever over 101  F (38.3  C) “on several occasions”,
but with diagnoses that become apparent after a careful and no diagnosis after 1 week in the hospital (interesting
history, physical exam, and relatively simple, non-invasive historical note: in the title of the paper, the “U” means
testsdthis might be termed Fever-of-Not-Unknown-Origin. “Unexplained”, and in the running header, it means “Unde-
Making the diagnosis here does not necessarily require termined”).7 This paper set the stage for thinking about 3
engagement of a sub-specialist. Many times the cause of fe- main categories of diagnoses in adults with FUO, namely in-
ver is an uncommon presentation of a common illness, like fectious (39% of cases), neoplastic (21%), and inflammatory
bacterial sinusitis without much in the way of nasal (19%), in addition to a miscellaneous category (21%). How-
drainage (but, perhaps, with bifrontal headache) or pneu- ever, the definition of FUO used in this paper lacks speci-
monia without cough or chest pain (but with, perhaps, ficity (what does “on several occasions” mean?) and is
increased voice transmission and dullness to percussion). unworkable today, as hospital admissions for diagnostic
The key to making these diagnoses is thinking outside the purposes may not be covered by insurance.
box and selecting targeted diagnostic tests. Between 1972 and 1998, 5 important case series of
Other times Fever-of-Not-Unknown-Origin is caused by children with FUO were published.8e12 As can be seen in
a common presentation of an uncommon disease. Thus, Table 1, some of the definitions used in those papers suffer
for example, prolonged fever accompanied by any combi- from the same problems as the Petersdorf and Beeson defi-
nation of rash, conjunctivitis, changes in the oral mucosa nition. A reasonable definition of FUO for clinical purposes
or peripheral extremities, or cervical lymphadenopathy would be similar to the one employed by Jacobs and
should prompt evaluation for Kawasaki disease, including Schutzeda core body temperature of at least 100.5  F
measurement of acute phase reactants and echocardio- (38.1  C) measured at least once a day for at least 14
graphic visualization of the coronary arteries. In endemic consecutive days. The fever curve looks something like
areas, cough and infiltrate on chest X-ray, especially if hi- that shown in Fig. 2, Panel A.
lar or mediastinal lymphadenopathy is present, should Our understanding of the final diagnoses in patients with
prompt serological testing for histoplasmosis. In any geog- FUO derive from these case series and others, which are
raphy, the same findings should prompt tuberculin skin generated largely by academicians at tertiary care referral
testing and investigation into possible epidemiologic links. centers. This introduces an important bias, as referral
Bruising, bleeding, splenomegaly, and/or lymphadenopa- patterns differ from one locale to another. Thus, for
thy should lead to an examination of the peripheral example, final diagnoses among children referred to an
smear, looking for blasts. If it’s summertime and the pa- academic medical center with an active, open-door outpa-
tient has leukopenia, thrombocytopenia and elevated he- tient consultation service might be different from those
patic transaminases, consideration should be given to a reported from a similar center that sees only inpatients, or
serologic work-up and empiric treatment for a tick- that has restrictions on who is seen in clinic. Likewise,
borne rickettsiosis. History of an abnormal heart value, referrals may be heavily influenced by the local insurance
a changing murmur, and embolic phenomena should market, as there may be barriers to referral.
lead to blood cultures, echocardiography, and consider- Fig. 3 shows the final diagnoses among children reported in
ation of subacute bacterial endocarditis. You get the the aforementioned pediatric case series, stratified by those
picture. reported in the 1970s and those reported in the 1990s (there
have been no similarly large case series from the U.S. pub-
lished in the last 15 years). One thing is evident: for
Fever-of-Unknown-Origin children in whom a specific diagnosis is ultimately made, the
“Big 3” diagnostic categoriesdinfectious, inflammatory, and
The designation of Fever-of-Unknown-Origin (FUO) is neoplastic, in that orderdcontinue to dominate. However,
reserved for those children who are truly having fever and the proportion of all children who never receive a specific diag-
in whom no diagnosis is apparent after the above consider- nosis has increased; in fact, this is now arguably the most com-
ations. The concept of FUO as an entity was crystallized in mon outcome of referral for FUO. What is going on?

Table 1 Defining fever of unknown origin in children.

Reference (year) Temperature () Frequency Duration
F C Method
McClung (1972) 102.1 38.9 Rectal Multiple occasions Outpatients: 3 weeks
Inpatients: 1 week
Pizzo (1975) 101.4 38.6 Rectal 5 occasions 2 weeks
Lohr (1977) 101.1 38.4 Not specified Multiple occasions Outpatients: 3 weeks
Inpatients: 1 week
Steele (1991) 100.5 38.1 Oral or rectal 2 occasions per week 3 weeksa
Jacobs (1998) 100.5 38.1 Core temperature 1 occasion every day 14 days
Normal urinalysis and chest X-ray required.
S86 G.S. Marshall

1970s 1990s


Infectious Inflammatory Neoplastic Other No diagnosis

Figure 3 Causes of FUO in children. The graph compares

final diagnoses among children evaluated for fever of unknown
origin in the 1970s (N Z 240) and the 1990s (N Z 255). Data
from references.8e12

the 1990s (Fig. 4)dits not that children are no longer getting
these infections, perhaps, its just that they are being diag-
nosed and treated by those who do not write papers on
FUO. A similar trend has been seen among adult patients.13
Other things have changed over time. New infections
that cause prolonged fever have appeared on the radar
screen; examples include cat scratch disease, Lyme dis-
ease, and ehrlichiosis (Fig. 4). Diagnoses like mononucleosis
may have become more prominent in the 1990s due to the
widespread availability of EpsteineBarr virus serology at
tertiary care centers.
Fig. 5 shows bacterial diagnoses among children with
FUO reported from developed and developing countries.
Differences reflect both local epidemiologydnote, for
example, the occurrence of brucellosis and typhoid fever
in developing countriesdas well as the availability of diag-
nostic modalitiesdthis may explain, for example, why ab-
scesses more commonly appear in case series from
developing countries, reflecting perhaps the fact that imag-
ing studies are not readily available outside of referral cen-
Figure 2 Fever patterns. Different fever curves suggest
ters. One can only speculate why osteomyelitis is more
different etiologies. Relentless daily fevers (Panel A) suggest
commonly reported in developed countries.
an underlying infectious, rheumatologic, or malignant condi-
Among the chronic inflammatory and autoimmune disor-
tion, although in many children such fevers are ultimately
ders that can present as FUO are inflammatory bowel
self-limited and no definitive diagnosis is made. Intermittent,
disease, juvenile idiopathic arthritis, systemic lupus
irregularly spaced episodes of fever of varying duration (Panel
B) suggest an autoinflammatory disease, unless the patient is
simply having recurrent viral illnesses. High fevers of abrupt 1970s 1990s
Upper respiratory tract
onset that occur with clockwork periodicity (Panel C) are Lower respiratory tract
classic for PFAPA syndrome; cyclic neutropenia should be Urinary tract
considered if the periodicity is approximately 21 days. Central nervous system
In all likelihood, children with FUO caused by identifiable
underlying diseases are being diagnosed outside of academic
centers, and thus never make their way into published case Cat scratch disease
series. Thus, for example, children with inflammatory or Tularemia
autoimmune disorders are screened by available serological Lyme disease
tests and then, when the diagnosis is suspected, are referred Ehrlichiosis
to rheumatologists. Likewise, the availability of detailed Other
imaging studies (solid tumors), flow cytometry (hematologic 0 10 20 30 40
malignancies), and sophisticated chemistries (neuroblas-
toma) results in straightaway referral to the oncologist,
suspected diagnosis in hand. Notice that infections of the Figure 4 Infectious causes of FUO in the U.S. The graph
respiratory tract, tuberculosis, and endocarditis, which compares final infectious disease diagnoses among children
accounted for a substantial proportion of infectious disease evaluated for fever of unknown origin in the 1970s (N Z 240)
diagnoses in the 1970s, were not reported in the series from and the 1990s (N Z 255). Data from references.8e12
Prolonged and recurrent fevers in children S87

Mycoplasma Developed Developing of nonspecific viral illnesses, blended together in the par-
Rickettsiosis ents’ mind. On the other hand, the daily variation in tem-
Lyme disease perature may itself provide the clue; for example,
Abscess temperatures in patients with JIA may be subnormal in
Endocarditis the mornings and high in the afternoons. The diagnosis
Brucellosis may be stumbled upon in a detailed review of systems.
Typhoid fever Periorbital edema?dthink primary EBV infection. Loose
Septicemia stools?dinflammatory bowel disease. Palpitations?d
Cat scratch disease hyperthyroidism.
Pyelonephritis From an infectious disease point of view, the questions
Tuberculosis range from the sublime to the ridiculous. New kitten in the
Urinary tract house?dthink cat scratch disease. Recent travel?dtyphoid
Osteomyelitis fever. Unpasteurized goat’s milk?dbrucellosis. Jumped on
a beaver dam?dblastomycosis. Exposed to a discarded
0 10 20 30
sheep placenta?dQ fever. You get the picture.
Percent In adults, an evidence-based, structured approach to
Figure 5 Bacterial infections in children with FUO in FUO has been offered, beginning with routine studies such
developed and developing countries. The graph compares as a complete blood count (CBC), peripheral smear,
final bacterial diagnoses among children evaluated for fever comprehensive metabolic panel (CMP), lactate dehydroge-
of unknown origin in developed (N Z 153) and developing nase (LDH), urinalysis, blood cultures (3 specimens sepa-
(N Z 338) countries, from a systematic review of papers pub- rated in time from different sites), anti-nuclear antibody
lished between 1972 and 2008. Data from references.45 (ANA), rheumatoid factor, human immunodeficiency virus
(HIV) antibody, chest X-ray (CXR), and selected serol-
ogies.13 If these tests are unrevealing, the algorithm
erythematosis (SLE), rheumatic fever, Wegener’s granulo- proceeds, in order, to discontinuation of nonessential med-
matosis, sarcoidosis, histiocytic lymphohistiocytosis, and ications (drug fever); abdominal CT or technetium-99 nu-
Behcet disease. Malignancies include leukemia, lymphoma, clear medicine scan (abscess, lymphoma, solid tumor);
neuroblastoma, hepatoma, and soft tissue sarcomas. There application of the Duke criteria (endocarditis); Doppler ul-
are a whole host of miscellaneous causes of FUO, including trasound of the lower extremities (deep vein thrombosis);
drug fever, dysautonomia, diabetes insipidus, ectodermal and consideration of temporal artery biopsy (temporal
dysplasia, hyperthyroidism, pulmonary embolus, and hema- arteritis), liver biopsy (various diagnoses), and laparoscopy
toma. In addition, clinicians must be aware of the possibil- (various diagnoses). Studies have also found bone marrow
ities of Munchausen syndrome by proxy and factitious fever. biopsy16 and positron emission tomography (PET)17 to be
In patients with FUO, the severity of the illness (magni- useful.
tude and intensity of fever and associated “red flags” such In pediatrics, because the underlying causes of FUO
as night sweats, weight loss, and localizing findings) differ, the approach also differs. It begins with a complete
dictates the pace of the work-up.14,15 The approach for history and physical exam, and the results of any previous
well-appearing children without “red flags” is given in work-up are cataloged. The basic laboratory studies shown
Fig. 6. The initial visit includes a careful history, with in Fig. 6 are done, along with targeted studies if there are
particular attention paid to delineating the fever pattern historical findings, physical abnormalities, or epidemiologic
and dissecting the presenting illness. Occasionally, it be- circumstances suggesting the possibility of these diagnoses
comes obvious that the FUO is nothing more than a series (Table 2). The parents are asked to maintain a fever and

First Visit Second Visit Third Visit

Initial Hx and PE Interval Hx and PE Interval Hx and PE

Previous lab results CBC with smear CBC with smear
CBC with smear CMP CMP
U/A and UCx

Targeted studies Targeted studies Targeted studies

Fever/symptom Fever/symptom
diary diary

Figure 6 Diagnostic approach to FUO (See text for explanation). Abbreviations: BCx: blood culture; CBC: complete blood count;
CMP: comprehensive metabolic panel; CRP: C-reactive protein; CXR: chest X-ray; ESR: erythrocyte sedimentation rate; Hx: history;
PE: physical examination; U/A: urinalysis; UCx: urine culture.
S88 G.S. Marshall

Table 2 Targeted studies in children with FUO.

Test Diagnosis Reason for suspicion
Infectious diseases
Tuberculin skin test, IGRA Tuberculosis Exposure to active case, homeless shelter, travel, immigration
Histoplasma serology Histoplasmosis Residence in Ohio Valley, bird exposures
EBV serology EBV infection Fatigue, lymphadenopathy, cytopenias, elevated transaminases
Bartonella serology Cat scratch disease Exposure to kittens, chronic granulomatous papule
Brucella serology Brucellosis Consumption of unpasteurized milk products
Toxoplasma serology Toxoplasmosis Exposure to cats, consumption of poorly cooked meat
Francesella serology Tularemia Tick bite with eschar, rabbit hunting
HIV Ab/Ag or PCR HIV infection Sexually active, mononucleosis-like illness, cytopenias
Stool culture for Salmonella Typhoid fever Travel to developing country, hepatosplenomegaly, rose spots
Antinuclear Ab, RF, C3, C4, Various rheumatologic Rash, arthritis, cytopenias, renal dysfunction,
slit lamp exam diseases acute phase reaction
ASO, anti-DNAse-B Ab Rheumatic fever Migratory polyarthritis, erythema marginatum, heart murmur
Flow cytometry, LDH, UA, Leukemia/lymphoma, Fatigue, bruising, bleeding, lymphadenopathy,
bone marrow aspirate various malignancies weight loss, cytopenias
Catecholamine screen, Neuroblastoma Opsoclonusemyoclonus, diarrhea
mIBG scan
CT, PET Solid tumors/lymphoma Localizing symptoms
Thyroid function profile Hyperthyroidism Nervousness, irritability, tachycardia, tremors
Sinus CT Sinusitis Headache, congestion
Echocardiogram Endocarditis Fatigue, murmur, splenomegaly, hematuria, thrombocytopenia
Endoscopy IBD Loose stools, weight loss, acute phase reaction, anemia
Abdominal U/S or CT Abdominal abscess Abdominal discomfort
Bone scan Osteomyelitis Musculoskeletal complaints
Abbreviations: Ab: antibody; Ag: antigen; ASO: anti-streptolysin-O; C3: complement component 3; C4: complement component 4; CT:
computed tomography; EBV: EpsteineBarr virus; HIV: human immunodeficiency virus; IBD: inflammatory bowel disease; IGRA: interferon
gamma release assay; LDH: lactate dehydrogenase; mIBG: (radio-ionated) meta-iodobenzylguanidine; PCR: polymerase chain reaction;
PET: positron emission tomography; RF: rheumatoid factor; U/S: ultrasound; UA: uric acid.
Data modified from references.14,15

symptom diary and the patient is brought back in a week, or visits to primary care providers, immediate care centers
earlier if anything changes. Unwarranted antimicrobial and/or emergency rooms; nonspecific diagnoses have
therapy is avoided. been made and antimicrobial therapy may have been
At the second visit, the interval history and test results prescribed.
are reviewed, the physical examination and basic labora-
tory studies are repeated, and targeted studies are Intermittent fever
reconsidered. New symptoms may prompt a new line of
investigation. For example, increasing lymphadenopathy Kids get sick all the time. The average child under 2, for
might prompt a lymph node biopsy; loose stools might example, has 6 or 7 respiratory illness per year (most of
prompt endoscopy; or rash, anemia and thrombocytopenia which are viral), unless they attend day care or have
might prompt an ANA screen. Trends in acute phase school-aged siblings, in which case they may has as many
reactants such as the erythrocyte sedimentation rate as 10.18 That’s almost one a month. In as much as some of
(ESR) and C-reactive protein (CRP) can be very informa- these illnesses are accompanied by fever, and other ill-
tiveddownward trends mean whatever is causing the nesses are acquired, histories of recurrent and unpredict-
fever may be resolving, whereas upward trends may able febrile illnesses are quite common. How does one
mean something is being missed. The process is reiterated differentiate the child who is having sequential viral ill-
at subsequent weekly visits until one of two things nesses from the one whose fevers indicate something
happens: a definitive diagnosis is made, or the fever more ominous?
resolves. A fever and symptom diary is one of the most useful tools
to employ here. Particular attention should be paid to
variation in associated symptoms from episode-to-episode.
Recurrent fever For example, one episode with sore throat, followed by an
episode associated with nausea, followed by one with red
Patients with recurrent fever present with a history of eyes and a transient rash is suggestive of recurrent viral
multiple episodes of fever. There have usually been many illnesses, particularly if one or another family member is
Prolonged and recurrent fevers in children S89

also sick when the patient is sick. Children with recurrent, It is unusual for a single infectious disease to present
self-limited viral illnesses will appear well and have a with recurrent fever over a long period of time.20 One sce-
normal physical examination, or may have findings that nario where this might occur is in the patient with a deep-
arise downstream of viral respiratory infection, such as seated infectiondoccult abdominal abscess, let’s say, or
otitis media with effusion, post-nasal drainage, and osteomyelitisdthat is partially treated by empiric courses
“allergic shiners”. Normal basic screening laboratory of antibiotics (all the more reason not to use antimicrobials
studies, including a CBC, CMP, and ESR, can go a long way without a definitive diagnosis). Paroxysmal fever cycles
in reassuring parents that there is no serious underlying occur every 2e3 days in malaria, and the diagnosis is sus-
disease. In those with a family history of immune defi- pected based on travel history and associated findings
ciency, failure to thrive, and/or a history of treatment with such as hemolysis, jaundice, and hepatosplenomegaly. Re-
intravenous antimicrobials, immune deficiency should be lapsing fever is caused by infection with Borrelia recurren-
considered.19 tis (louse-borne, seen mostly in Africa and associated with

Activation by Activation by K+ efflux,

internal and external reactive oxygen species,
danger signals phagosomes, inert substances
Mevalonate kinase
NF-κB *Cryopyrin




gene Pro-IL-1β


Figure 7 Pathogenesis of inflammasomopathies. Internal (e.g., damage-associated molecular patterns) or external (e.g.,
pathogen-associated molecular patterns, acting through Toll-like receptors) danger signals activate gene transcription in leuko-
cytes via NF-kB. This results in the production of pro-IL-1b, cryopyrin (also known as NALP3 and NLRP3), procaspase-1, and the
adaptor molecule ASC, among other proteins. Danger signals also activate cryopyrin, which then associates with procaspase-1
and ASC to form a macromolecular complex called the inflammasome, in essence a molecular organelle whose purpose is to
generate proinflammatory cytokines. Other danger or stress signals (including Kþ efflux, reactive oxygen species, and phagosome
rupture), as well as the presence of irritants like gout crystals and asbestos fibers, activate the inflammasome, which cleaves
procaspase-1 into caspase-1. This in turn cleaves pro-IL-1b into IL-1b, a potent proinflammatory cytokine, which is released
from the cell (other cytokines like IL-18 also may be produced). The end result is inflammation. Autoinflammatory diseases
(Table 3) can result from altered function of components of the inflammasome itself (so-called intrinsic inflammasomopathies)
or regulatory elements external to the inflammasome (so-called extrinsic inflammasomopathies). Intrinsic inflammasomopathies
include the cryopyrin-associated periodic syndromesdfamilial cold autoinflammatory syndrome, Muckle-Wells syndrome, and
neonatal-onset multisystem inflammatory disease. These result from activating, gain-of-function mutations in cryopyrin (*).
Extrinsic inflammasomopathies include familial Mediterranean fever, which results from gain-of-proinflammatory function or
a loss-of-anti-inflammatory function mutations in pyrin, a regulatory protein that interacts with the inflammasome. Likewise,
hyper-IgD syndrome results from the loss of mevalonate kinase activity, which, through a variety of signaling molecules, leads
to activation of caspase-1. The end result in each of these disorders is an increase in inflammation. Key: Solid arrows represent
activation or assembly. Dashed arrows represent inhibition or modulation. Abbreviations: ASC: apoptosis-associated speck-like pro-
tein with a capsase-recruitment domain; ER: endoplasmic reticulum; IL-1b: interleukin-1b; Kþ: potassium ion; NALP3: NACHT
(nucleotide-binding), leucine-rich repeat, and pyrin-binding domains-containing protein 3; NF-kB: nuclear factor kappa-light-
chain-enhancer of activated B cells; NLRP3: nucleotide-binding, leucine rich repeat, and pyrin domains-containing protein 3;
pro-IL-1b: precursor protein to IL-1b. Adapted from references.46,47
S90 G.S. Marshall

poverty and crowding) or other Borrelia species such as disorders, with well-defined inheritance patterns and
Borrelia hermsii (tick-borne, distributed worldwide). The ethnic predilections. Second, many have onset early in
disease is characterized by paroxysms of fever associated life. Third, attacks occur with variable frequency and dura-
with chills, sweats, musculoskeletal complaints, hepatos- tion (Fig. 2, Panel B); for some diseases, like neonatal onset
plenomegaly and jaundice; initial relapses occur every multi-system inflammatory disease (NOMID), symptoms are
5e7 days but quickly become milder and less frequent. continuous, albeit with exacerbations and remissions.
Other infections that can present with recurrent fever pat- Fourth, associated signs and symptoms such as rash, seros-
terns include brucellosis, leptospirosis, and rat-bite fever. itis, splenomegaly, and arthritis are common. Finally, some
Autoinflammatory diseases should be considered in of these diseases result in long-term sequelae such as
children with recurrent fevers who do not fit the picture amyloidosis.
of recurrent viral infections and do not have classic
indicators of immunodeficiency. These diseases are char- Periodic fever
acterized by abnormally increased inflammation mediated
by cells and molecules of the innate immune system (as While the above autoinflammatory diseases are often
opposed to autoantibodies and autoreactive T-cells, as is referred to as periodic fevers, the variable frequency and
seen in autoimmune diseases like SLE), occurring in persons duration of attacks and the irregular periodicity stand in
with a genetic predisposition.21 Whether they result from sharp contrast to what is seen in cyclic neutropenia and
intrinsic or extrinsic perturbations of the inflammasome the periodic fever, aphthous stomatitis, pharyngitis, and
(Fig. 7), the expression of mutated receptors (Fig. 8), or adenitis (PFAPA, pronounced “fap-ah”) syndrome. These
other mechanisms, autoinflammatory diseases have one syndromes are marked by stereotypical episodes that recur
thing in common: increased expression of inflammatory cy- with clockwork periodicity (it should be noted that some
tokines by leukocytes, leading to unchecked inflammatory studies point to clockwork periodicity in hyper-IgD syn-
responses. drome22). In cyclic neutropenia, blood neutrophil counts
Table 3 lists features of some of the autoinflammatory reach a nadir every 21 days, resulting in fever, malaise,
diseases, most of which are rare. Several things should be mouth ulcers, and bacterial infections.23 This condition is
noted. First, these are for the most part monogenic inherited in autosomal dominant fashion and is caused by

Unfolded protein response
Ligand-independent signaling
Inhibition of apoptosis





Figure 8 Pathogenesis of TRAPS. TNF, also known a cachectin, is an inflammatory cytokine produced by macrophages, CD4-
positive lymphocytes, and natural killer cells. TNFR1 is expressed on most cell types. Engagement of membrane-bound TNFR1
with TNF activates a number of signaling pathwaysdsome of these lead to apoptosis, others, as in the case of immune cells,
lead to transcription of proinflammatory genes. Leukocytes in TRAPS patients express both wild-type TNFR1 and mTNFR1. mTNFR1
misfolds and traffics inefficiently, accumulating in the ER and activating the unfolded protein response, which increases suscep-
tibility of the cell to inflammatory stimuli. Aggregated mTNFR1 also may induce proinflammatory responses independent of engage-
ment with TNF (so-called ligand-independent signaling). Wild-type TNFR1 traffics to the cell surface, engages TNF, and calls for an
inflammatory response through mechanisms such as activation of NF-kB. Cleavage of membrane-bound TNFR1 releases a soluble
receptor that neutralizes TNF in the environment, thus keeping the inflammatory stimulus in check. However, mTNFR1 that
does happen to make it to the cell surface is inefficiently cleaved, effectively increasing the pool of free TNF available to engage
membrane-bound wild-type TNFR1; this leads to increased proinflammatory signaling. The end result of all of these processes is an
increase in inflammation. Abbreviations: ER: endoplasmic reticulum; NF-kB: nuclear factor kappa-light-chain-enhancer of acti-
vated B cells; TNFR1: tumor necrosis factor receptor type 1; mTNFR1: mutant tumor necrosis factor receptor type 1. Adapted
from references.46,48
Prolonged and recurrent fevers in children S91

mutations in the neutrophil elastase gene (ELANE), which Some have questioned whether PFAPA syndrome is
leads to reduced production and accelerated apoptosis of actually a distinct clinical entity, likening it to what used
myeloid progenitor cells in the bone marrow (the mecha- to be called “recurrent tonsillitis”.29 Support for this view
nisms by which the neutrophil counts cycle so precisely comes from the fact that tonsillectomy appears to be cura-
have not been elucidated). The diagnosis can be made by tive in a majority of patients.30,31 However, several featur-
gene testing and by serial (Mondays, Wednesdays, and Fri- esdmost notably the clockwork periodicity and aphthous
days) CBCs bracketing an illness episode (the neutropenia ulcersdare distinctly unusual in recurrent tonsillitis. Suf-
reaches a nadir before the fever and other symptoms fice it to say that when you’ve seen a PFAPA syndrome pa-
become prominent). Patients usually respond well to treat- tient, you know it.
ment with granulocyte colony-stimulating factor. There is tantalizing evidence that PFAPA syndrome is an
PFAPA syndrome, first described in 1987,24 is the most autoinflammatory disease. Serum inflammatory mediators
common periodic fever in children. Distinctive features are elevated during attacks, and IL-1 blockade with
are listed in Table 4, and an exemplary fever curve is shown anikinra (an IL-1 receptor antagonist) aborts episodes.32
in Fig. 2, Panel C. The typical patient has onset of symp- Genes associated with complement, inflammasome, and
toms at around 3 years of age.25 Episodes occur every interferon activity are upregulated during attacks, but
4e6 weeks and last 3e4 days, with maximum temperatures gene expression during asymptomatic intervals is similar
as high as 104.5  F (40.3  C). Constitutional symptoms to that in healthy controls. Interestingly, the overall gene
and malaise are common. Localizing symptoms such as expression profile during PFAPA syndrome attacks differs
abdominal pain, diarrhea, arthralgia, and rash have been from that seen during attacks of hereditary fever syn-
reported in case series26; however, since there is no widely dromes. As familial occurrence of PFAPA syndrome has
accepted research case definition of PFAPA syndrome, it is been described,33e37 there is the possibility that a heritable
difficult to know if all of these series actually report on the basis may be discovered.
same patients. CRP27 (but, curiously, not procalcitonin28) is That being said, comprehensive DNA sequencing for the
elevated during PFAPA episodes. monogenic autoinflammatory fever syndromes is already

Table 3 Autoinflammatory diseases.a

Feature Inflammasomopathies Protein
Intrinsic (Cryopyrin-associated) Extrinsic folding disorder


Synonyms FCUS e CINCA e MKD Hibernian fever
Inheritance Autosomal Autosomal Sporadic Autosomal recessive Autosomal Autosomal
pattern dominant dominant recessive dominant
Ethnicity European European Any Mediterranean European European
Age at onset <1 year <20 years <1 year <20 years <1 year <20 years
Frequency Variable Variable Continuous Variable 2e4 weeks Variable
of attacks
Duration 1e2 days 2e3 days Continuous 1e3 days 3e7 days >7 days
of attacks
Clinical findings Rash Rash Rash Serositis Rash Rash
Conjunctivitis Conjunctivitis Meningitis Splenomegaly Adenopathy Arthritis
Headache Deafness Arthropathy Erysipeloid Serositis Conjunctivitis
Nausea Deafness erythema Vomiting Splenomegaly
Adenopathy Diarrhea
Hepatomegaly Arthralgia
Splenomegaly Headache
Amyloidosis No Yes No Yes No Yes
Treatment Anti-IL-1 Anti-IL-1 Anti-IL-1 Colchicine Anti-IL-1 Anti-IL-1
modalities Anti-TNF Anti-TNF
Abbreviations. CINCA: chronic infantile neurological, cutaneous and articular syndrome; FCAS: familial cold autoinflammatory syn-
drome; FUCS: familial cold urticaria syndrome; FMF: familial Mediterranean fever; HIGDS: hyper-IgD syndrome; IL-1: interleukin-1;
MKD: mevalonate kinase deficiency; MWS: Muckle-Wells syndrome; NOMID: neonatal onset multi-system inflammatory disease; TNF: tu-
mor necrosis factor; TRAPS: tumor necrosis factor receptor-associated periodic syndrome.
Some autoinflammatory diseases result from mechanisms other than inflammasome dysregulation and altered protein folding. For
example, Crohn’s disease and Blau syndrome result from abnormal activation of NF-kB. Other autoinflammatory diseases not listed in
the table include PAPA (pyogenic arthritis, pyoderma gangrenosum and acne) and DIRA (deficiency of the interleukin-1-receptor
Data from references.49e51
S92 G.S. Marshall

fever episodes by 9 years of age; while 9 continued to have

Table 4 Distinctive features of PFAPA syndrome.
episodes, the episodes became shorter and more spread out
d Fever episodes are stereotypical, circumspect, unprovoked, in time, and most of them had experienced remission long
and recur with clockwork periodicity periods of time.40 Steroids are very effective in aborting
d Identifiable prodrome is common
PFAPA episodes.30 In general, 1e2 mg/kg of prednisone is
d There is evidence of upper respiratory tract inflammation
given as soon as it is clear that an episode is beginning
d Rash and arthritis are absent
d Acute phase reactants are elevated during episodes and (there is good evidence that 0.5 mg/kg is just as effec-
normal between episodes tive41). In general, one dose is sufficient, although addi-
d The child is failing to fail-to-thrive tional doses may be given over the next 2 or 3 days if
d Episodes are aborted by steroid therapy fevers persist. Caution needs to be exercised that the fever
d Episodes resolve after tonsillectomy illness being treated is a PFAPA episode and not an intercur-
d The syndrome resolves by adolescence rent infection. As mentioned earlier, tonsillectomy is often
d There are no long-term sequelae curative, but, given the associated morbidity and risks, this
Data from references.25,40,44,52,53 option should be reserved for children whose episodes are
severely disruptive to the child and family (as well as those
children who have other indications for tonsillectomy).42,43
commercially available.38 The question is, which children The bottom line is that most families can live with PFAPA
with recurrent fevers should be tested? Findings of a posi- syndrome until the child “grows out of it”. As summarized
tive family history, thoracic pain, abdominal pain, diarrhea, by Sarah Long, “.upbeat parents and patients with PFAPA
vomiting, rash and arthralgia favor monogenic disorders syndrome leave our offices with their disability, relieved
rather than PFAPA syndrome; exudative pharyngitis and that we recognize the constellation and predict ultimate
aphthous ulcers favor PFAPA syndrome.22 In 2008, Gattorno good health, satisfied with the notion of an upregulated
and colleagues39 derived and validated a diagnostic score immune system, and armed with a prescription for predni-
that predicts which children with PFAPA-like symptoms sone should an episode land on a birthday party or trip to
are likely to test positive for one of the heritable periodic Disney World”.44
syndromes (Table 5). This score may be useful for guiding
diagnostic work-ups in the clinical setting. Disclosures
The prognosis in PFAPA syndrome is good. In a longitu-
dinal study, 50 of 59 patients had complete resolution of
The author has no potential conflicts of interest to disclose.
This work was funded by in-kind contributions from the
Table 5 Gaslini score.a University of Louisville.
Variable Codeb Coefficient Value
Age at Months X 0.067 Z a
Abdominal 0 Z Never 1.494 b
1. Wunderlich CA, Seguin E. Medical thermometry and human
pain 2 Z Sometimes
temperature. New York, NY: William Wood & Company; 1871.
or often 2. Mackowiak PA, Wasserman SS, Levine MM. A critical appraisal
3 Z Always of 98.6  F, the upper limit of the normal body temperature,
Aphthous 0 Z Never 1.504 c and other legacies of Carl Reinhold August Wunderlich. J Am
ulcers 1 Z Sometimes Med Assoc 1992;268(12):1578e80.
or often 3. Rideout ME, First LR. Fever: measuring and managing a
2 Z Always sizzling symptom. Contemp Pediatr 2001;18. 42e50, 53e4.
Thoracic 0 Z Absent 1.958 d 4. Marcy SM, Kohl KS, Dagan R, Nalin D, Blum M, Jones MC, et al.
pain 1 Z Present Fever as an adverse event following immunization: case defi-
Diarrhea 0 Z Never 0.901 e nition and guidelines of data collection, analysis, and presen-
tation. Vaccine 2004;22:551e6.
1 Z Sometimes
5. Baraff LJ. Management of infants and young children with fe-
2 Z Often ver without source. Pediatr Ann 2008;37(10):673e9.
3 Z Always 6. Kleiman MB. The complaint of persistent fever: recognition
Family 0 Z Negative 1.503 f and management of pseudo fever of unknown origin. Pediatr
history 1 Z Positive Clin North Am 1982;29(1):201e8.
For each variable, multiply the code number (e.g., 0, 1, 2, etc.) 7. Petersdorf RG, Beeson PB. Fever of unexplained origin: report
by the corresponding coefficient to obtain the value. Add the on 100 cases. Medicine 1961;40:1e30.
values (a þ b þ c þ d þ e þ f) to obtain the total score. In a 8. McClung HJ. Prolonged fever of unknown origin in children.
validation set, a score >1.32 identified children who were pos- Am J Dis Child 1972;124:544e50.
itive for heritable periodic fever syndromes with 87% sensitivity 9. Pizzo PA, Lovejoy FH, Smith DH. Prolonged fever in children:
and 72% specificity). review of 100 cases. Pediatrics 1975;55(4):468e73.
Named for the Giannina Gaslini Institute in Genoa, Italy, the 10. Lohr JA, Hendley JO. Prolonged fever of unknown origin: a re-
principal investigators’ home institution. cord of experiences with 54 childhood patients. Clin Pediatr
Codes apply during each episode of fever and not during in- 1977;16(9):768e73.
terval periods. 11. Steele RW, Jones SM, Lowe BA, Glasier CM. Usefulness of
scanning procedures for diagnosis of fever of unknown origin
Data from Gattorno et al.39
in children. J Pediatr 1991;119(4):526e30.
Prolonged and recurrent fevers in children S93

12. Jacobs RF, Schutze GE. Bartonella henselae as a cause of pro- 34. Valenzuela PM, Majerson D, Tapia JL, Talesnik E. Syndrome of
longed fever and fever of unknown origin in children. Clin periodic fever, aphthous stomatitis, pharyngitis, and adenitis
Infect Dis 1998;26:80e4. (PFAPA) in siblings. Clin Rheumatol 2009;28:1235e7.
13. Mourad O, Palda V, Detsky AS. A comprehensive evidence- 35. Cochard M, Clet J, Le L, Pillet P, Onrubia X, Gueron T, et al.
based approach to fever of unknown origin. Arch Intern Med PFAPA syndrome is not a sporadic disease. Rheumatology
2003;163:545e51. 2010;49:1984e7.
14. Long SS. Distinguishing among prolonged, recurrent, and peri- 36. Anton-Martı́n P, Oritz Movilla R, Guillen Martin S, Allende LM,
odic fever syndromes: approach of a pediatric infectious dis- Cuesta Rubio MT, Lopez Gonzalez MF, et al. PFAPA syndrome
eases subspecialist. Pediatr Clin North Am 2005;52:811e35. in siblings: is there a genetic background? Eur J Pediatr
15. Tolan RW. Fever of unknown origin: a diagnostic approach to 2011;170:1563e8.
this vexing problem. Clin Pediatr 2010;49(3):207e13. 37. Adachi M, Watanabe A, Nishiyama A, Oyazato Y, Kamioka I,
16. Hot A, Jaisson I, Girard C, French M, Durand DV, Rousset H, et al. Murase M, et al. Familial cases of periodic fever with aphthous
Yield of bone marrow examination in diagnosing the source of fe- stomatitis, pharyngitis, and cervical adenitis syndrome. J Pe-
ver of unknown origin. Arch Intern Med 2009;169(21):2018e23. diatr 2011;158(1):155e9.
17. Qiu L, Chen Y. The role of 18F-FDG PET or PET/CT in the 38. Periodic fever syndromes panel (7 genes). Gene Dx Web site.
detection of fever of unknown origin. Eur J Radiol 2012. [accessed periodic-fever-syndromes-panel-7-genes [accessed 29.09.13].
29.09.13]. 39. Gattorno M, Sormani MP, D’Osualdo A, Pelagatti MA, Caroli F,
18. Denny FW, Collier AM, Henderson FW. Acute respiratory infec- Federici S, et al. A diagnostic score for molecular analysis of
tions in day care. Rev Infect Dis 1986;8(4):527e32. hereditary autoinflammatory syndromes with periodic fever
19. Subbarayan A, Colarusso G, Hughes SM, Gennery AR, Slatter M, in children. Arthritis Rheum 2008;58(6):1823e32.
Cant AJ, et al. Clinical features that identify children with pri- 40. Wurster VM, Carlucci JG, Feder HM, Edwards KE. Long-term
mary immunodeficiency diseases. Pediatrics 2011;127(5): follow-up of children with periodic fever, aphthous stomati-
810e6. tis, pharyngitis, and cervical adenitis syndrome. J Pediatr
20. Knockaert DC. Recurrent fevers of unknown origin. Infect Dis 2011;159(6):958e64.
Clin North Am 2007;21:1189e211. 41. Yazgan H, Gu € Uzun L. Compar-
€ltekin E, Yazicilar O, Sagun OF,
21. Kastner DL, Aksentijevich I, Goldbach-Mansky R. Autoinflam- ison of conventional and low dose steroid in the treatment of
matory disease reloaded: a clinical perspective. Cell 2010; PFAPA syndrome: preliminary study. Int J Pediatr Otorhinolar-
140:784e90. yngol 2012.
22. Gattorno M, Caorsi R, Meini A, Cattalini M, Federici S, [accessed 31.08.12].
Zulian F, et al. Differentiating PFAPA syndrome from mono- 42. Burton MJ, Pollard AJ, Ramsden JD. Tonsillectomy for peri-
genic periodic fevers. Pediatrics 2009;124(4):e721e8. odic fever, aphthous stomatitis, pharyngitis and cervical
23. Dale DC, Welte K. Cyclic and chronic neutropenia. Cancer adenitis syndrome (PFAPA). Cochrane Database Syst Rev
Treat Res 2011;157:97e108. 2010;(9).
24. Marshall GS, Edwards KM, Butler J, Lawton AR. Syndrome of Art.No.:CD008669.
periodic fever, pharyngitis, and aphthous stomatitis. J Pediatr 43. Baugh RF, Archer SM, Mitchell RB, Rosenfeld RM, Armin R,
1987;110(1):43e6. Burns JJ, et al. Clinical practice guideline: tonsillectomy
25. Thomas KT, Feder HM, Lawton AR, Edwards KM. Periodic fever in children. Otolaryngol Head Neck Surg 2011;144(4):
syndrome in children. J Pediatr 1999;135(1):15e21. S1e30.
26. Vigo G, Zulian F. Periodic fevers with aphthous stomatitis, 44. Long SS. Syndrome of periodic fever, aphthous stomatitis,
pharyngitis, and adenitis (PFAPA). Autoimmun Rev 2012. pharyngitis, and adenitis (PFAPA)dwhat it isn’t. What is it? [accessed J Pediatr 1999;135(1):1e5.
29.09.13]. 45. Chow A, Robinson JL. Fever of unknown origin in children: a
27. Førsvoll JA, Øymar K. C-reactive protein in the periodic fever, systematic review. World J Pediatr 2011;7(1):5e10.
aphthous stomatitis, pharyngitis and cervical adenitis (PFAPA) 46. Masters SL, Simon A, Aksentijevich I, Kastner DL. Horror in-
syndrome. Acta Paediatr 2007;96:1670e3. flammaticus: the molecular pathophysiology of autoinflam-
28. Yazgan H, Keleş E, Yazgan Z, Gebeşçe A, Demirdöven M. C-reac- matory disease. Ann Rev Immunol 2009;27:621e68.
tive protein and procalcitonin during febrile attacks in PFAPA 47. Strowig T, Henao-Mejia J, Elinav E, Flavell R. Inflammasomes
syndrome. Int J Pediatr Otorhinolaryngol 2012;76:1145e7. in health and disease. Nature 2012;481:278e86.
29. Hornibrook J. More euphemisms for tonsillitis. Otolaryngol 48. Kimberley FC, Lobito AA, Siegel RM, Screaton GR. Falling into
Head Neck Surg 2011;144:649. TRAPSdreceptor misfolding in the TNF receptor 1-associated
30. Peridis S, Pilgrim G, Koudoumnakis E, Athanasopoulos I, periodic fever syndrome. Arthritis Res Therapy 2007;9:217.
Houlakis M, Parpounas K. PFAPA syndrome in children: a
meta-analysis on surgical versus medical treatment. Int J Pe- 49. Kastner DL. Hereditary periodic fever syndromes. Hematol
diatr Otorhinolaryngol 2010;74:1203e8. 2005:74e81.
31. Garavello W, Pignataro L, Gaini L, Torretta S, Somigliana E, 50. Bodar EJ, Drenth JPH, van der Meer JWM, Simon A. Dysregu-
Gaini R. Tonsillectomy in children with periodic fever with lation of innate immunity: hereditary periodic fever syn-
aphthous stomatitis, pharyngitis, and adenitis syndrome. J dromes. Br J Hematol 2008;144:279e302.
Pediatr 2011;159(1):138e42. 51. Lachmann HJ. Clinical immunology review series: an approach
32. Stojanov S, Lapidus S, Chitkara P, Feder H, Salazar JC, to the patient with a periodic fever syndrome. Clin Exper Im-
Fleisher TA, et al. Periodic fever, aphthous stomatitis, phar- munol 2011;165:301e9.
yngitis, and adenitis (PFAPA) is a disorder of innate immunity 52. Marshall GS, Edwards KM, Butler J, Lawton AR. Syndrome of
and Th1 activation responsive to IL-1 blockade. Proc Natl periodic fever, pharyngitis, and aphthous stomatitis. J Pediatr
Acad Sci U S A 2011;108(17):7148e53. 1987;110(1):43e6.
33. Sampaio ICRM, Rodrigo MJ, Marques JGDPM. Two siblings with 53. Feder HM, Salazar JC. A clinical review of 105 patients with
periodic fever, aphthous stomatitis, pharyngitis, adenitis PFAPA (a periodic fever syndrome). Acta Paediatrica 2010;
(PFAPA) syndrome. Pediatr Infect Dis J 2009;28(3):254e5. 99:178e84.