Clinical Gastroenterology and Hepatology 2018;-:-–-

1 59
2 60
3 61
4 62
5 AGA Clinical Practice Update: Surgical Risk Assessment and 63
6 64
7 Perioperative Management in Cirrhosis 65
8 66
Q4 Patrick G. Northup, Lawrence S. Friedman, and Patrick S. Kamath
9 67
10 68
11Q1 Division of Gastroenterology and Hepatology, University of Virginia, Charlottesville, Virginia 69
12 70
13 71
espite worldwide increases in the prevalence of hypertension also progresses with advancing cirrhosis,
14
15
D chronic liver disease due to viral hepatitis and
nonalcoholic steatohepatitis (NASH), patients with
thereby raising the risk of surgical bleeding complica-
tions involving the collateral circulation, especially in the
72
73
16 74
cirrhosis are living longer with more advanced disease splanchnic system in the thoracic and peritoneal cavities.
17 75
because of improved medical and surgical management.1 Splenic sequestration, in addition to a possible contri-
18 76
As a result, they are at risk for other diseases and mor- bution from thrombopoietin deficiency,5 leads to pe-
19 77
bidities that patients with cirrhosis might not have ripheral thrombocytopenia, which can interfere with
20 78
experienced in past decades. Patients with cirrhosis now primary hemostasis. Disruptions in secondary hemosta-
21 79
have a significant lifetime risk of hepatocellular carci- sis (coagulation) and fibrinolysis can lead to a propensity
22 80
noma (HCC) and obesity-related diseases, and they may to either bleeding or clotting, and the net effect can be
23 81
require emergent or elective surgical procedures other difficult to predict on the basis of traditional laboratory
24 82
than liver transplantation. Early reports of outcomes of testing.4 Current systems of deceased donor organ allo-
25 83
standard surgical procedures in patients with cirrhosis cation in the United States6 result in the selection of liver
26 84
described intolerable mortality rates as high as 67%.2 transplant recipients who are likely to have renal
27 85
Improved management and better selection of patients dysfunction, which complicates fluid management and
28 86
with cirrhosis have led to improved surgical survival leads to reduced renal excretion of drugs; hepatic protein
29 87
and the development of new risk prediction algorithms. synthetic dysfunction may also impair hepatic drug
30 88
Nevertheless, the evaluation and treatment of the patient metabolism. The result is a change in the pharmacoki-
31 89
with cirrhosis in whom an invasive surgical procedure is netic handling of medications. All of these pathophysio-
32 90
planned is not standardized, and there are no definitive logic alterations lead to a significant potential for
33 91
prospective trials to provide clarity to clinicians in morbidity and mortality related to surgery. Primary he-
34 92
assessing patients in the preoperative period and manag- patic resection, usually for malignancy, is particularly
35 93
ing them in the postoperative period. The modern litera- hazardous in this population, can acutely exacerbate all
36 94
ture on surgical risk stratification in cirrhosis patients of the aforementioned pathophysiologic changes due to a
37 95
consists of case reports, small series, and only a few further reduction in liver mass, and may thereby in-
38 96
serious attempts to stratify risk. This review summarizes crease the postoperative risk of rapidly progressive he-
39 97
the available data and recommendations based on expert patic failure.
40 98
opinion on how best to predict surgical outcomes and
41 99
optimize the condition of patients with cirrhosis who un- Predicting Surgical Risk in a Patient
42 100
dergo surgical procedures. Table 1 summarizes our
43 With Cirrhosis 101
recommendations.
44 102
45 Surgical procedures are classified broadly as those 103
46 Unique Pathophysiologic Risks in the that are lifesaving, such as cardiovascular or cancer 104
47 Patient With Cirrhosis surgery and emergency interventions, and those that are 105
48 carried out to improve a patient’s quality of life. The 106
49 The pathophysiology of chronic liver disease and challenge for the care team is to assess the surgical risks 107
50 portal hypertension predisposes the patient with 108
51 cirrhosis to complications related to surgical procedures. 109
Abbreviations used in this paper: ASA, American Society of Anesthesiol-
52 A comprehensive review of this pathophysiology is ogists; CTP, Child-Turcotte-Pugh; HCC, hepatocellular carcinoma; HVPG, 110
53 beyond the scope of this article,3,4 but several points hepatic vein pressure gradient; INR, international normalized ratio; MELD, 111
model for end-stage liver disease; NASH, nonalcoholic steatohepatitis;
54 warrant emphasis. The patient with cirrhosis generally NSAID, nonsteroidal anti-inflammatory drug; TIPS, transjugular intra- 112
55 has protein synthetic dysfunction that progresses as the hepatic portosystemic shunt. 113
56 liver disease advances and results in the malnutrition/ 114
© 2018 by the AGA Institute
57 sarcopenia syndrome, which hinders wound healing and 1542-3565/$36.00 115
58 physical recovery from a surgical procedure. Portal https://doi.org/10.1016/j.cgh.2018.09.043 116

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 2 Northup et al Clinical Gastroenterology and Hepatology Vol. -, No. -

117 Table 1. Recommendations for Preoperative Risk Assessment and Perioperative Management in Patients With Cirrhosis 175
118 176
Target audience Gastroenterologists, hepatologists, general surgeons, surgical subspecialists,
119 anesthesiologists, critical care physicians, and other clinicians seeing patients with 177
120 cirrhosis 178
121 Target population Patients with cirrhosis undergoing invasive non–liver transplant surgical procedures 179
122 Baseline factors common in patients with Significant malnutrition; portal hypertension, raising the risk of bleeding complications; 180
cirrhosis potentially contributing to thrombocytopenia, coagulation, and fibrinolytic abnormalities; renal dysfunction; impaired
123 181
surgical risk hepatic drug metabolism and elimination; increased susceptibility to infection; decreased
124 effective intravascular volume and susceptibility to acute renal injury; potential for a 182
125 thrombophilic state and predisposition to venous thromboembolism 183
126 Preoperative surgical mortality risk CTP score, MELD score, ASA Physical Status Classification, Mayo Postoperative Mortality 184
127 assessment tools Risk Calculator in Patients with Cirrhosis 185
Postoperative complications of surgical Hepatic decompensation and worsening of liver synthetic function; acute worsening of portal
128 186
procedures more common in or hypertension and associated complications, including ascites, hepatic encephalopathy,
129 unique to patients with cirrhosis and portal hypertensive bleeding; wound healing difficulties and dehiscence; pulmonary 187
130 complications—pleural effusions and pneumonia; other infections, including bacterial 188
131 peritonitis; increased risk of intraoperative and perioperative bleeding; multiple organ 189
132 failure 190
BPA BPA 1: Due to the profound effects of hepatic synthetic dysfunction and portal hypertension,
133 191
patients with cirrhosis are at increased risk of death after all invasive surgical procedures
134 compared with the healthy general population. It is not entirely clear how much the risk 192
135 attributed to cirrhosis is additive to traditional cardiopulmonary risk factors studied in the 193
136 general population. Patients with cirrhosis undergoing all but the most emergent surgical 194
137 procedures should be risk stratified and counseled on the magnitude of that risk. 195
BPA 2: In patients considered for surgery, use the CTP score (Child-Pugh class), MELD score,
138 196
Mayo Postoperative Mortality Risk Score, or another validated risk stratification system.
139 There is no single definitive risk stratification system to determine operative risk in all 197
140 patients with cirrhosis, and we recommend using multiple methods. 198
141 BPA 3: The type and anatomic site of the proposed surgical procedure is important in risk 199
142 stratification. The clinical teams must incorporate the surgical procedure itself into 200
discussions of risk with patients. Procedures associated with higher surgical risk include
143 201
hepatobiliary surgery, such as primary liver resection, other intra-abdominal procedures,
144 thoracic surgery, and cardiovascular procedures. Avoid elective cholecystectomy if 202
145 possible and if required, it should be carried out in centers with expertise in this population. 203
146 BPA 4: Surgical risk is continuous, and there are no absolute cutoff values for excluding 204
147 patients with cirrhosis from surgical procedures. The patient and the surgical and medical 205
teams caring for the patient must weigh the potential benefits and risks collaboratively.
148 206
Patients should be referred to a surgical team with experience in the care of patients with
149 cirrhosis and portal hypertension whenever possible. Patients with Child-Pugh class C 207
150 (CTP score >10) or a MELD score >20 pose a high risk of postoperative decompensation 208
151 and death. Avoid or delay until after liver transplantation, if possible, all but the most urgent 209
152 and life-saving procedures in this population. 210
BPA 5: TIPS is not routinely recommended before surgical procedures in patients with
153 211
cirrhosis and portal hypertension with abdominal collaterals. Small uncontrolled case
154 series have demonstrated the ability to decompress collateral vessels with TIPS 212
155 preoperatively in patients requiring deep pelvic and colonic resections; however, the 213
156 absolute benefit of TIPS over conservative management is not established. 214
157 BPA 6: The special case of primary liver resection in a patient with cirrhosis, usually for 215
malignancy, has been studied more thoroughly than other general surgical procedures.
158 216
Data support the safety profile of segmental liver resections in patients without clinically
159 significant portal hypertension, as indicated by an HVPG <10 mm Hg. A validated 217
160 surrogate measure of lack of clinically significant portal hypertension is the absence of 218
161 venous abdominal collaterals on imaging or of esophageal varices on endoscopy, 219
162 peripheral blood platelet count >100,000/mL, or hepatic transient elastography values of 220
<23kPa.
163 221
BPA 7: There are no established preoperative safety thresholds for common laboratory values
164 related to bleeding and clotting. The INR is not predictive of procedural bleeding risk in 222
165 patients with cirrhosis. We do not recommend protocol transfusions to a target INR in 223
166 patients with cirrhosis. In vitro studies suggest that a platelet count >50,000/mL is 224
167 adequate to generate thrombin and provide stable clot formation, and retrospective clinical 225
studies show an increased bleeding tendency in patients with a platelet count lower than
168 226
this threshold. The literature provides no evidence to support protocol transfusions, but
169 critically ill patients with plasma fibrinogen levels <100 mg/dL have more bleeding events. 227
170 Handle coagulation management on a case-by-case basis preferably using viscoelastic 228
171 testing directed therapy, and avoid needless transfusions or volume overload. 229
172 230
173 231
174 232

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 - 2018 Surgical Risk in Cirrhosis CPU 3

233 Table 1. Continued 291

234 292
BPA 8: Involve a skilled medical team with experience in treating patients with cirrhosis in the
235 postoperative management of patients with cirrhosis, and obtain early consultation to help 293
236 avoid progressive complications. Achieve optimal control of ascites, variceal bleeding risk, 294
237 and hepatic encephalopathy before surgery, if possible. Monitor renal and hepatic function 295
238 at least daily in the postoperative period. 296
BPA 9: Aggressively avoid exacerbations of portal hypertension and associated complications
239 297
in the postoperative period. Appropriate measures include close monitoring of renal
240 function and avoidance of volume excess or depletion. Excess volume can increase the 298
241 risk of variceal and other portal hypertensive bleeding. 299
242 BPA 10: Due to disturbed metabolism and elimination, patients with cirrhosis are at especially 300
243 high risk for medication-related complications. Generally, use opiates at lower doses and 301
with longer dosing intervals than in the general population. Use only short-acting
244 302
benzodiazepines. Avoidance of constipation should be a priority in the management of
245 these patients to minimize flares of hepatic encephalopathy in the postoperative period. In 303
246 patients who can take medications orally, rifaximin generally causes less bowel distension 304
247 than nonabsorbable disaccharides (lactulose). 305
248 BPA 11: Avoid medications that might be toxic in patients with portal hypertension and 306
cirrhosis. Do not use NSAIDs because they can impair renal blood flow. Patients with
249 307
cirrhosis, especially those with heavy alcohol use, can be susceptible to acetaminophen
250 toxicity at doses lower than those that may cause toxicity in the general population. Do not 308
251 prescribe combination opiate/acetaminophen pain relievers to take home after surgery 309
252 because the patient may be unaware of the presence of acetaminophen in various over- 310
253 the-counter products and unintentional acetaminophen overdose and hepatotoxicity can 311
result.
254 312
BPA 12: The gallbladder wall may appear thickened on imaging, which may lead to the
255 erroneous diagnosis of acute cholecystitis. A diagnosis of acute cholecystitis should be 313
256 made only in the appropriate clinical setting, usually in the presence of biliary pain. There is 314
257 a significant risk of complications after cholecystectomy in patients with cirrhosis. Avoid 315
258 elective cholecystectomy in a patient with confirmed cirrhosis. Refer patients who require 316
cholecystectomy to a surgical team with experience in invasive procedures in the cirrhotic
259 317
population.
260 BPA 13: Except for incarceration that cannot be manually reduced or suspected strangulation 318
261 with bowel ischemia or gangrene, abdominal hernia surgery should be avoided in the 319
262 patient with cirrhosis and ascites unless the ascites is completely controlled medically. 320
263 Wound dehiscence, peritonitis, and poor outcomes frequently occur when ascites recurs 321
after hernia surgery.
264 322
BPA 14: Patients who experience hepatic decompensation after surgery may become
265 candidates for liver transplantation. The MELD score cutoff value for selecting patients for a 323
266 liver transplant evaluation before elective surgery is not clear. We recommend preoperative 324
267 liver transplant evaluation when the predicted postoperative 3-month mortality rate is 325
268 greater than 15% as reflected by surgical risk stratification models or when the MELD 326
score is
15.
269 327
BPA 15: Centers with expertise in surgery in cirrhotic patients may perform bariatric surgery in
270 this population, but clinically significant portal hypertension is a contraindication to the 328
271 operation. In highly selected patients with obesity and decompensated cirrhosis 329
272 undergoing liver transplantation, a sleeve gastrectomy at the same time as liver 330
273 transplantation is an option. 331
274 332
275 NOTE. ASA, American Society of Anesthesiologists; BPA, best practice advice; CTP, Child-Turcotte-Pugh; HVPG, hepatic vein pressure gradient; INR, inter- 333
276 national normalized ratio; MELD, Model for End-Stage Liver Disease; NSAID, nonsteroidal anti-inflammatory drug; TIPS, transjugular intrahepatic portosystemic 334
277 shunt. 335
278 336
279 before deciding whether the procedure can be carried preoperatively as American Society of Anesthesiologists 337
280 out safely, should be delayed until the patient has un- (ASA) class V (see later) had a median survival of 2 days, 338
281 dergone liver transplantation, or is best avoided a 90% mortality rate at 14 days, and a 100% mortality 339
282 completely, especially in the high risk patient who may rate at 90 days. The risk of postoperative mortality and 340
283 not be a liver transplant candidate. Surgical and consul- complications is related to the degree of hepatic 341
284 tative teams must also consider the idea of futility in dysfunction, comorbidities, type of surgical procedure, 342
285 their decision-making process.7 If the surgical risk is and expertise of the management team. Of these factors, 343
286 much higher than the chance of a good outcome from the the severity of the liver disease appears to be the most 344
287 perspective of the patient, then the procedure should be important factor in determining mortality risk. The 345
288 considered futile and avoided regardless of the patient’s severity of liver disease has traditionally been assessed 346
289 degree of liver disease. For example, Teh et al8 showed in the literature by the Child-Turcotte-Pugh (CTP) score 347
290 that cirrhotic patients undergoing surgery categorized (Child-Pugh class)9 and the Model for End-Stage Liver 348

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 4 Northup et al Clinical Gastroenterology and Hepatology Vol. -, No. -

349 Disease (MELD) score. Table 2 summarizes the compo- following).12 The MELD score is less effective in 407
350 nents of the various severity scales that are used in describing the occasional patient with severe portal hy- 408
351 practice. pertension but preserved hepatic synthetic function; in 409
352 these patients, the CTP score may be a more useful tool 410
353 in those patients. 411
CTP Score and ASA Physical Status
354 The ASA physical classification system has been used 412
355 Classification in some form since the early 1940s as a gross measure of 413
356 the risk of administering anesthesia for surgical pro- 414
357 Historically, the CTP score has been used to stage the 415
cedures.13 The system classifies patients into illness
358 degree of liver dysfunction and severity of portal hy- 416
levels that correlate roughly with surgical outcomes in
359 pertension in patients with cirrhosis; this system has 417
some circumstances, but the ASA score is generally
360 largely withstood the test of time. Pre-MELD score era 418
thought of as a measure of comorbidity or disease
361 studies showed a good correlation between the Child- 419
severity rather than a direct predictor of outcome,
362 Pugh class (combined with other variables) and surgi- particularly in patients with cirrhosis.14 The classifica- 420
363 cal mortality.10 In general, patients with Child-Pugh class 421
tion system also suffers from poor interobserver vari-
364 A cirrhosis may undergo surgery in the absence of 422
ability and is somewhat subjective. Moreover, there is no
365 thrombocytopenia or clinically significant portal hyper- 423
category for moderate disease, only for mild and severe
366 tension; only selected patients with Child-Pugh class B 424
disease. Despite these weaknesses, the ASA classification
367 cirrhosis are elective surgical candidates; and patients 425
is still used widely for risk stratification before invasive
368 with Child-Pugh class C cirrhosis are not considered 426
procedures in the general surgical population, and it has
369 candidates for elective surgery. Recent analyses, how- 427
been studied as a predictive factor in case series of pa-
370 ever, show less of a correlation between the Child-Pugh tients with cirrhosis. 428
371 class and surgical outcomes, likely because of low 429
372 numbers of patients with Child-Pugh class C cirrhosis 430
included in modern series, in part due to avoidance of MELD Score
373 431
374 surgery in patients with decompensated cirrhosis.11 432
Nevertheless, the presence of ascites (a component of Initially developed to predict mortality after place-
375 433
the CTP score but not the MELD score), especially ment of a transjugular intrahepatic portosystemic shunt
376 434
massive ascites, correlates with poor surgical outcomes (TIPS),15 the MELD score was later introduced as a
377 435
in many analyses. Because of the somewhat subjective means of prioritizing organ allocation for liver trans-
378 436
nature of estimating the severity of ascites and enceph- plantation. The MELD score is calculated using the in-
379 437
alopathy, use of the CTP score for surgical risk stratifi- ternational normalized ratio (INR), serum bilirubin level,
380 438
cation has been largely supplanted by the MELD score, and serum creatinine level. Patients with a preoperative
381 439
which encompasses the objective liver function- MELD score <16 have lower postoperative mortality
382 440
dependent components of the CTP score that have rates than those with higher scores.16 One study
383 441
been found to be predictive of outcomes (see the demonstrated a linear increase in postoperative
384 442
385 443
386 Table 2. CTP Score, ASA Physical Status Classification, and MELD Score: Descriptions and Definitions 444
387 445
388 Predictive Model Formula/Description Reference 446
389 CTP score Encephalopathy grade: none (1), stage 1–2 (2), stage 3–4 (3) 9
447
390 Ascites level: absent (1), slight (2), moderate or severe (3) 448
391 Serum albumin (g/dL): >3.5 (1), 2.8–3.5 (2), <2.8 (3) 449
392 Total bilirubin (mg/dL): <2 (1), 2–3 (2), >3 (3) 450
393 INR: <1.7 (1), 1.7–2.3 (2), >2.3 (3) 451
Total points are calculated:
394 452
5–6: class A
395 7–9: class B 453
396 10–15: class C 454
397 ASA classification Class 1: normal healthy patient 13 455
398 Class 2: mild systemic disease 456
Class 3: severe systemic disease that is not life-threatening
399 Class 4: severe systemic disease that is a constant threat to life
457
400 Class 5: moribund and not expected to survive without the operation 458
401 Class 6: brain-dead patient/organ donor 459
402 MELD score ¼ 3.78  ln (serum bilirubin in mg/dL) þ 11.2  ln (INR) þ 9.57  ln (serum creatinine in mg/dL) þ 6.43 72 460
403 Final score is rounded to the nearest whole number 461
Patients on renal replacement therapy are assigned a serum creatinine value of 4.0 mg/dL
404 462
405 463
406 NOTE. ASA, American Society of Anesthesiologists; CTP, Child-Turcotte-Pugh; INR, international normalized ratio; MELD, Model for End-Stage Liver Disease. 464

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 - 2018 Surgical Risk in Cirrhosis CPU 5

465 mortality with an increase in MELD score. This large, Improvements in surgical technique and medical 523
466 single-center, retrospective study of patients with management have had a positive effect on surgical 524
467 cirrhosis who underwent cardiovascular, orthopedic, or mortality and morbidity rates. For example, a case series 525
468 major abdominal surgery showed that the mortality rate of patients with cirrhosis undergoing nonshunt cel- 526
469 increased steadily up to 90 days following major surgery iotomy published in 1984 showed a 30% mortality rate, 527
470 with increasing MELD scores.8 A higher ASA class and which correlated with the Child-Pugh class.21 By com- 528
471 age >70 years were also predictive of mortality. It is parison, a series published in 2010 from a center with 529
472 important to emphasize that the patients in this study extensive experience in surgery in patients with cirrhosis 530
473 were carefully selected for surgery after exclusion of reported a 7% mortality rate for abdominal procedures 531
474 significant comorbidity. Mortality could be predicted at 7 that was not well predicted by the CTP score but did 532
475 days, 30 days, 90 days, and 1 year, and in the long term, correlate with the MELD score and serum albumin 533
476 with a combination of the MELD score, ASA class, and level.11 The authors of the latter study suggested that use 534
477 age. Using these data, the postoperative mortality risk in of a laparoscopic approach was associated with a 535
478 patients with cirrhosis can be determined with use of an decreased rate of postoperative complications but 536
479 online calculator (Mayo Postoperative Surgical Risk acknowledged possible selection bias. There is also an 537
480 Score; https://www.mayoclinic.org/medical-professionals/ unquantifiable but significant contribution of case expe- 538
481 model-end-stage-liver-disease/postoperative-mortality- rience and multidisciplinary management in the 539
482 risk-patients-cirrhosis). One advantage of this calculator improved surgical outcomes in patients with cirrhosis 540
483 over the MELD score alone is that it includes age and ASA over the past decades. Several case series have suggested 541
484 class, which give the model more granularity and greater that, compared with more invasive approaches, laparo- 542
485 predictive value than MELD alone. The MELD-Na score scopic and other minimally invasive surgical techniques 543
486 has not been studied as a predictor of surgical outcomes may lead to favorable outcomes in patients with 544
487 in patients with cirrhosis. cirrhosis.22–26 The lack of concurrent risk-adjusted con- 545
488 trols and the small size of most reports make definitive, 546
489 evidence-based recommendations speculative at best. 547
490 Risks of Specific Surgical Procedures in Similarly, aside from strict avoidance of outdated tradi- 548
491 Patients With Cirrhosis tionally hepatotoxic inhalational anesthetic agents, no 549
492 studies have compared anesthesia techniques in patients 550
493 Many case series in the literature, some with his- with cirrhosis, so decisions regarding the choice of an 551
494 torical controls, address the characteristics of various anesthetic agent must be made on a case-by-case basis. 552
495 surgical procedures and associated mortality in patients As an example, spinal and epidural anesthesia are known 553
496 with cirrhosis. Among the different types of surgery, to reduce hepatic blood flow but there are no specific 554
497 hepatobiliary surgery is perceived to have the highest studies addressing anesthesia risk comparing spinal 555
498 risk of liver-related mortality and morbidity. Abdominal anesthesia to other modalities in cirrhosis patients. 556
499 surgery and possibly thoracic surgery in the presence of 557
500 portal hypertension are significant risk factors as well, Primary Hepatic Resection 558
501 but whether violation of the abdominal or thoracic 559
502 cavity is a specific risk factor in and of itself or whether Hepatic resection for HCC is typically considered only 560
503 these procedures increase surgical risk because they in patients with Barcelona Clinic Liver Cancer stage 561
504 are more invasive and disruptive of the splanchnic 0 (early) HCC.27 Stage 0 is defined as a tumor <2 cm in 562
505 vasculature and collaterals is unknown. Therefore, pa- diameter, excellent patient performance status, and 563
506 tients undergoing evaluation for surgery in the pres- normal liver function (Child-Pugh class A) without portal 564
507 ence of cirrhosis require careful selection and extensive hypertension. A surrogate marker for the lack of portal 565
508 discussion involving the patient and family before the hypertension in this population is a peripheral platelet 566
509 procedure. The urgency of the procedure, although not count >100,000/mL. Other studies have identified the 567
510 often a controllable variable, is an important predictor absence of clinically significant portal hypertension by an 568
511 of outcomes. In the general population, emergency hepatic venous pressure gradient (HVPG) <10 mm Hg28 569
512 surgery for any indication carries a significantly higher or the absence of venous collaterals on cross-sectional 570
513 risk for morbidity and mortality than nonemergent abdominal imaging along with the absence of esopha- 571
514 procedures.17 This trend has been shown in small series geal varices on upper endoscopy. The use of liver stiff- 572
515 of patients with cirrhosis as well.18,19 Based on the ness measurement (eg, by transient elastography) to 573
516 aforementioned small and retrospective studies, pa- guide decisions regarding hepatic surgical resection has 574
517 tients who undergo an emergency procedure should been suggested, but optimal cutoff values have not been 575
518 expect at least a 2-fold increased relative risk of validated. In one series, a liver stiffness measurement 576
519 morbidity and mortality that will not be directly rep- below 22 kPa was established as the optimal cutoff level 577
520 resented by MELD-based predictive models (although for selecting patients for surgery.29 A liver stiffness 578
521 some models do incorporate emergency surgery as a measurement of >22 kPa had greater sensitivity and 579
522 risk factor).20 higher negative as well as positive predictive value in 580

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 6 Northup et al Clinical Gastroenterology and Hepatology Vol. -, No. -

581 predicting patients likely to experience postoperative single large-volume paracentesis before a surgical pro- 639
582 hepatic decompensation compared with an HVPG >10 cedure does not provide adequate preoperative control 640
583 mm Hg. Much lower cutoff values, ranging from 12 to 15 of ascites. Medical control of clinically apparent ascites 641
584 kPa, have also been proposed.30 A MELD score <9 may with diuretics, as per practice guidelines, before a sur- 642
585 also be used to select patients for hepatic resection, gical hernia repair is required in a patient with decom- 643
586 because such patients have an excellent prognosis pensated cirrhosis to prevent recurrent ascites 644
587 following resection.31 postoperatively. Selected surgical candidates may benefit 645
588 from preemptive placement of a TIPS to control ascites 646
589 before surgery. Recurrent ascites leads to impaired 647
Cholecystectomy
590 wound healing and risks possible dehiscence. For this 648
591 reason, efforts should be undertaken to avoid recurrent 649
Routine abdominal imaging (ultrasonography and
592 ascites in the postoperative period and may include 650
other cross-sectional modalities) may show a thickened
593 frequent postoperative paracentesis and careful diuretic 651
gallbladder wall in patients with cirrhosis, particularly in
594 therapy while attempting to avoid renal injury. In 652
patients with ascites. The finding typically reflects
595 selected patients continuous pigtail catheter drainage of 653
fibrotic thickening associated with cirrhosis rather than
596 ascites postoperatively may prevent pressure on the 654
the edema seen in acute cholecystitis. In fact, in a
597 hernia repair. 655
cirrhotic patient with acute abdominal pain, a misdiag-
598 Several studies have examined risk factors for mor- 656
nosis of acute cholecystitis is not uncommon. In general,
599 tality after hernia surgery, but most relevant studies are 657
elective cholecystectomy should be delayed in patients
600 uncontrolled and include a heterogeneous population 658
with cirrhosis who are candidates for a liver trans-
601 (for example, elective and emergency cases). Both the 659
plantation. If urgent surgery must be carried out and the
602 MELD score12 and the Child-Pugh class combined with 660
diagnosis of acute cholecystitis is certain, patients should
603 the ASA class14 have been advocated for hernia surgery 661
be referred urgently to a center with experience in car-
604 risk stratification, but neither appears to be optimal in all 662
rying out cholecystectomy in patients with cirrhosis;
605 cases. 663
however, we do not recommend cholecystectomy in pa-
606 The emergent presentation of an incarcerated hernia 664
tients with Child-Pugh class C cirrhosis or refractory
607 is a special challenge in a patient with cirrhosis and is the 665
ascites. Such patients are also generally not candidates
608 most common indication for emergency surgery in this 666
for percutaneous cholecystostomy, unless they are criti-
609 population and a major risk factor for poor out- 667
cally ill, due to the presence of ascites and a high risk of
610 comes.17–19 Because of the significant increase in mor- 668
infections and procedural complications. In noncritically
611 tality and morbidity associated with emergency 669
ill patients, endoscopic transpapillary gallbladder
612 operations in patients with cirrhosis, some authors have 670
drainage may be carried out if technically feasible to
613 advocated elective hernia repair in patients with 671
relieve symptoms, but expertise in the technique is
614 cirrhosis to prevent an emergent presentation. A single- 672
required.32
615 center prospective case series without controls reported 673
Laparoscopic cholecystectomy has been associated
616 on patients listed for liver transplantation who presented 674
with a higher mortality rate in patients with cirrhosis
617 for elective umbilical hernia repair.25 Thirty patients 675
compared with noncirrhotic patients in a population-
618 eventually underwent elective surgery, and most had 676
based study in the United States.24 There is also an
619 Child-Pugh class B or C cirrhosis, with a median MELD 677
increased requirement for conversion of laparoscopic
620 score of 12. Two patients had prolonged hospital stays 678
cholecystectomy to open cholecystectomy in patients
621 due to complications, but the remainder did not experi- 679
with cirrhosis. In the setting of laparoscopic cholecys-
622 ence a direct surgical complication. The authors 680
tectomy, the MELD score correlates well with mortality
623 concluded that elective umbilical hernia repair is pref- 681
risk but not with the risk of complications from the
624 erable to emergency repair in patients with cirrhosis and 682
surgical procedure itself.33 Laparoscopic cholecystec-
625 ascites. The lack of controls in this series makes broad 683
tomy has not been compared with open cholecystectomy
626 conclusions difficult and the frequency of emergency 684
in a prospective study, but in a single retrospective study
627 incarceration in patients with cirrhosis with a stable 685
using historical controls, the laparoscopic approach
628 abdominal hernia is unknown. Surgical and medical 686
appeared to be preferable when technically feasible.34
629 teams should determine the safety and risk-benefit ratio 687
630 of elective hernia surgery carefully in this population. 688
631 Abdominal Wall Hernia Repair 689
632 Cardiovascular Surgery 690
633 Patients with cirrhosis are prone to abdominal her- 691
634 nias because of the presence of ascites and, to a lesser The need for cardiovascular surgery is increasing in 692
635 extent, abdominal organomegaly. Symptomatic hernias patients with cirrhosis as NASH, usually accompanied by 693
636 are common, especially in the setting of significant asci- components of the metabolic syndrome (which are also 694
637 tes, and control of ascites is key to successful hernia risk factors for cardiovascular disease), has become a 695
638 repair. Because the ascites inevitably recurs quickly, a leading cause of cirrhosis in the developed world.35 696

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 - 2018 Surgical Risk in Cirrhosis CPU 7

697 Cardiovascular surgery, especially for an urgent indica- centers in patients with compensated cirrhosis but 755
698 tion, can be life-extending, even in patients with signifi- without clinically significant portal hypertension. Even in 756
699 cant complications from cirrhosis. Cardiovascular skilled hands, the risk of complications is increased 757
700 operations, especially those requiring cardiac bypass and compared with that in the general population,53–56 758
701 extracorporeal circulation, pose unique physiologic although studies regarding risk stratification of 759
702 challenges for the patient, including hyperfibrinolysis, cirrhotic patients before surgery are lacking. However, 760
703 platelet activation, and the need for anticoagulation patients with decompensated cirrhosis or with clinically 761
704 during and frequently after the procedure. Nevertheless, significant portal hypertension are at highest risk for 762
705 studies are limited with regard to the optimal manage- adverse outcomes and are not candidates for bariatric 763
706 ment of patients with cirrhosis and these intraoperative procedures. Several centers have reported the outcomes 764
707 and postoperative issues.36,37 Case series and cohort of sleeve gastrectomy preformed at the time of liver 765
708 studies using historical controls show relatively high transplantation. Such a combined procedure in highly 766
709 morbidity and mortality rates in patients with cirrhosis selected patients is associated with durable weight loss 767
710 who undergo cardiac surgery.35,38,39 Most cardiac sur- and a lower risk of the metabolic syndrome57,58 after 768
711 gery teams use viscoelastic testing such as thromboe- liver transplantation. A sleeve gastrectomy, which in- 769
712 lastography or rotational thromboelastometry40 in the volves resection of the greater curvature of the stomach, 770
713 management of patients undergoing these operations, an does not add significant morbidity to the liver transplant 771
714 approach that seems clinically reasonable because the procedure and is not associated with intestinal malab- 772
715 technology has been tested extensively in persons with sorption postoperatively, which is a particular concern in 773
716 cirrhosis.41 the prevention of perioperative graft rejection that could 774
717 Several prospective and retrospective case series result from malabsorption of immunosuppression med- 775
718 without controls and with historical controls have eval- ications. Weight loss following the procedure is gradual, 776
719 uated the value of the MELD and CTP scores in predicting and endoscopic access to the biliary tract is preserved 777
720 outcomes after cardiac surgery, but the data are limited should the patient develop posttransplant biliary com- 778
721 by small numbers of patients, inclusion of only patients plications. A few centers have reported case series of 779
722 cleared for surgery, and a vast preponderance of patients Roux-en-Y gastric bypass59 and sleeve gastrectomy after 780
723 with compensated cirrhosis with a low MELD score.42–47 liver transplantation,60 but because of small numbers of 781
724 Overall, elective coronary artery bypass graft surgery is patients, we cannot make conclusions about safety and 782
725 associated with mortality rates ranging from 4% to 70%, efficacy. 783
726 which correlate with the MELD score and the Child-Pugh 784
727 class. In general, a MELD score
13.546 or a CTP score Other Surgical Procedures 785
728 >747 is considered a contraindication to cardiac surgery. 786
729 No specific studies have compared the various surgical Table 3 lists risk stratification tools used for various 787
730 approaches. A small study of patients undergoing aortic surgical procedures in the setting of cirrhosis. The data 788
731 valve replacement48 showed no significant difference in are derived primarily from small case series, and limited 789
732 mortality between transcatheter aortic valve replace- conclusions can be made on the accuracy of the predic- 790
733 ment and surgical aortic valve replacement. Coronary tive models for many of the operations. 791
734 angiography and cardiac catheterization are performed 792
735 routinely in patients with cirrhosis undergoing liver 793
736 transplant evaluation and appear to be relatively safe in
Preoperative and Perioperative 794
737 this population. Based on a single uncontrolled case se- Evaluation and Management of the 795
738 ries, transfusion of plasma to prophylactically “correct” Patient With Cirrhosis 796
739 the INR before the procedure cannot be recommended.49 797
740 There are sparse data on the use of antiplatelet agents Surgical procedures that are generally considered 798
741 after coronary stent placement and after cardiac surgery safe in the medical population without significant 799
742 in patients with cirrhosis, but available studies suggest comorbidities may pose great risk in the patient with 800
743 an increase in the risk of variceal and nonvariceal decompensated cirrhosis. Early involvement of an expe- 801
744 bleeding when these patients receive dual antiplatelet rienced multidisciplinary medical and surgical team is an 802
745 therapy.50,51 absolute key to success. Earnest and data-driven risk 803
746 discussions with both the patient and the surgical teams 804
747 Bariatric Surgery are a critical prerequisite for an informed decision about 805
748 surgical procedures in this population, and the discus- 806
749 Obesity is the most common risk factor for cirrhosis sion should begin as early in the process as possible. For 807
750 in developed countries, and NASH cirrhosis is among the patients undergoing elective or semielective procedures, 808
751 most common indications for listing patients for liver ascites should be controlled, variceal bleeding risk 809
752 transplantation. Weight loss through lifestyle changes reduced, and acute hepatic encephalopathy treated pre- 810
753 benefit patients with obesity and NASH but is difficult to operatively, with continuation of medical therapy to 811
754 achieve.52 Bariatric surgery may be carried out at expert prevent recurrent hepatic encephalopathy. With the 812

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 8 Northup et al Clinical Gastroenterology and Hepatology Vol. -, No. -

813 Table 3. Predictive Models Used in Patients With Cirrhosis Undergoing Various Surgical Procedures 871
814 872
Surgical Procedure Predictive Model
815 873
816 All surgical procedures MELD score (continuous risk),16 MELD score, age, ASA (continuous risk),8 MELD score <1412 874
817 Liver cancer resection HVPG <10 mm Hg,28 MELD score <9,31 liver transient elastography <22 kPa30 875
818 Cholecystectomy CTP score (continuous risk),33 Child-Pugh class AþB,83 Child-Pugh class A,74 MELD score <1514 876
819 Abdominal wall herniorrhaphy CTP score (continuous risk)14 and MELD score <1312 877
Coronary artery bypass grafting CTP score <847 or MELD score <13.546
820 Bariatric surgery Child-Pugh class A56
878
821 Lung cancer resection Child-Pugh class A84 879
822 Colonic resection MELD score <985 880
823 General orthopedic procedures MELD score (continuous risk)77 881
824 Lumbar spine surgery CTP score <686 882
Intracranial neurosurgery CTP score (continuous risk)87
825 Head and neck cancer resections Child-Pugh class A,80 MELD score <1081
883
826 884
827 885
828 NOTE. ASA, American Society of Anesthesiologists; CTP, Child-Turcotte-Pugh; HVPG, hepatic vein pressure gradient; MELD, Model for End-Stage Liver Disease. 886
829 887
830 aging of the population and the increasing prevalence of aspiration during induction of anesthesia. Therefore, in 888
831 obesity, NASH, and diabetes mellitus, the patient with patients with symptomatic ascites, large-volume total 889
832 cirrhosis is likely to have nonhepatic risk factors for abdominal paracentesis should be performed preopera- 890
833 surgical complications as well. No specific studies have tively, with intravenous administrative of albumin in a 891
834 examined whether nonhepatic comorbid risk factors dose of 6–8 g for each liter of ascitic fluid removed. Once 892
835 such as cardiopulmonary disease are additive or multi- oral medications are tolerated, daily antibiotic prophy- 893
836 plicative to the risk imparted by liver disease. Clinical laxis can be resumed in patients with a history of 894
837 experience suggests, however, that class III obesity and spontaneous bacterial peritonitis and patients with a low 895
838 cardiopulmonary disease are risk factors for morbidity ascitic total protein concentration, as per practice 896
839 and mortality independent of the severity of liver dis- guidelines. In patients who are not able to take oral 897
840 ease. Similarly, most of the major risk stratification medications, third-generation cephalosporins such as 898
841 scores do not include age, but the extremes of age are ceftriaxone, can be administered prophylactically in a 899
842 known to be independent risk factors for medical and dose of 1 g every 24 hours intravenously until oral intake 900
843 surgical complications. In fact, both the CTP and MELD of prophylactic antibiotics is possible again. Preoperative 901
844 scores are predictive of surgical outcome only in patients thoracentesis is only recommended in patients with 902
845 who are otherwise satisfactory candidates for surgery moderate to large pleural effusions that compromise 903
846 and have no other important comorbidities. Therefore, respiratory function. Administration of intravenous fluid 904
847 cardiology and pulmonary consultations are frequently and blood products should be limited perioperatively to 905
848 needed for these complex patients. avoid increasing extracellular volume, ascites, and portal 906
849 Elective operations in patients with cirrhosis should venous pressure.62 907
850 generally be avoided until the liver disease can be opti- Only a few case reports and uncontrolled case series 908
851 mized. Progressive decompensation is well described in have assessed the benefit of preemptive TIPS placement 909
852 the postoperative period in patients with previously fully before surgery to reduce the risk of collateral vessel and 910
853 compensated cirrhosis. For chronic viral hepatitis and intra-abdominal variceal bleeding intraoperatively.63–65 911
854 autoimmune hepatitis, strict adherence to antiviral and Although a decrease in the need for intraoperative 912
855 immunosuppressive therapy is required to avoid viral blood transfusions following TIPS placement is claimed, 913
856 resistance and relapse, respectively, and elective or semi- the available studies have been uncontrolled and the 914
857 elective surgery may be delayed until therapy is frequency of complications, such as hepatic encepha- 915
858 completed to avoid an interruption in treatment. Alter- lopathy, following TIPS placement is not reported. In 916
859 native routes of drug administration may be necessary contrast, a retrospective study of liver transplantation 917
860 such as the use of intravenous administration of corti- performed in patients who had undergone placement of 918
861 costeroids in patients with autoimmune hepatitis who a TIPS before transplantation compared with concurrent 919
862 are unable to take oral medications. controls who had not undergone preoperative TIPS 920
863 placement showed no difference in red blood cell or 921
864 Ascites plasma transfusion requirements between the 2 922
865 groups.66 Moreover, in a patient undergoing nontrans- 923
866 Ascites should be managed aggressively, as recom- plant surgery, a TIPS that is misplaced into the right 924
867 mended by practice guidelines,61 before surgical pro- atrium, inferior vena cava, or mesenteric veins may 925
868 cedures, if possible. Ascites can restrict pulmonary adversely affect future liver transplantation.67 Therefore, 926
869 function and delay perioperative recovery of lung func- owing to unproven efficacy, lack of published evidence, 927
870 tion. Patients with ascites are also at risk of pulmonary potentially high complication rates, and significant added 928

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 - 2018 Surgical Risk in Cirrhosis CPU 9

929 expense, preemptive TIPS placement before a surgical Low fibrinogen levels are associated with an 987
930 procedure cannot be recommended as a routine practice. increased bleeding risk,76 and levels <100 mg/dL are 988
931 associated with a risk of inhibiting clot formation in 989
932 Hemostasis and Coagulation patients with cirrhosis. Fibrinogen can be replaced with 990
933 low-volume cryoprecipitate transfusions, which do not 991
934 Protocol transfusion strategies based on preoperative significantly increase plasma volume. Although no 992
935 platelet counts and the INR are generally ineffective in controlled trials of the efficacy of fibrinogen transfusion 993
936 reducing intra- and postoperative bleeding. Patients with in patients with cirrhosis have been undertaken, it seems 994
937 cirrhosis are known to be in a state of “rebalance” in that physiologically reasonable to replete fibrinogen levels 995
938 deficits in procoagulant factors are offset by deficits in with infusions of cryoprecipitate before high-risk surgi- 996
939 anticoagulant proteins synthesized by the liver.4 There- cal procedures to allow the substrate for adequate clot 997
940 fore, they are not necessarily predisposed to severe formation. 998
941 bleeding.68 In fact, as liver disease advances, progressive Viscoelastic testing (thromboelastography and rota- 999
942 protein C deficiency leads to a thrombophilic state in tional thromboelastometry) to guide intraoperative 1000
943 some patients.69 These observations argue strongly transfusions is available at large centers and has been 1001
944 against preoperative protocol transfusions in most pa- shown to decrease the need for red blood cell and 1002
945 tients with cirrhosis. The INR is not predictive of plasma transfusions in the general surgical population77 1003
946 bleeding complications in patients with cirrhosis, and and specifically in patients with cirrhosis.78 If available, 1004
947 “prophylactic” preoperative fresh frozen plasma trans- viscoelastic testing should be considered before and 1005
948 fusions are not recommended.70 during surgical procedures in patients with cirrhosis to 1006
949 Although thrombocytopenia is common in patients help guide a rational transfusion strategy. Details about 1007
950 with cirrhosis, in vitro71 and retrospective cross-sectional these testing modalities are available elsewhere.40 1008
951 studies72 have shown that platelet counts above 1009
952 50,000/mL are adequate to allow clot formation in most Pain Control 1010
953 patients with cirrhosis. Prophylactic transfusions to raise 1011
954 the platelet count to higher levels are unlikely to be Because of altered drug metabolism and elimination, 1012
955 beneficial and may expose the patient to complications of patients with cirrhosis are at especially high risk of 1013
956 transfusions, volume overload, or unexpected thrombosis. medication-related toxicity. Opiates should generally be 1014
957 In recent years, thrombopoietin analogues (romiplostim) used at lower than standard doses and at longer dosing 1015
958 and receptor agonists (eltrombopag, avatrombopag, and intervals than in the general population. Although ran- 1016
959 lusutrombopag) have been developed for the treatment of domized controlled trials have not been performed in 1017
960 thrombocytopenia due to idiopathic thrombocytopenic patients with cirrhosis, hydromorphone and transdermal 1018
961 purpura. Eltrombopag is also labeled for use in patients fentanyl may be the preferred agents.79 Protein binding 1019
962 with cirrhosis due to chronic hepatitis C to allow and hepatic metabolism are the primary determinants of 1020
963 increasing platelet counts during antiviral therapy with benzodiazepine elimination, and only short-acting ben- 1021
964 interferon.73 Only avatrombopag and lusutrombopag are zodiazepines such as midazolam should be used. Tra- 1022
965 approved for the general thrombocytopenia related to madol, which has favorable hepatic metabolism,80 or 1023
966 liver disease. In patients with cirrhosis, avatrombopag and acetaminophen,81 in a dose of no more than 2 g daily, is 1024
967 lusutrombopag were effective in raising platelet counts also recommended despite the misconception that acet- 1025
968 above 50,000/mL in phase 3 clinical trials when given aminophen, even in small doses, is contraindicated in 1026
969 before invasive procedures compared with placebo.74,75 patients with cirrhosis. On the other hand, avoid routine 1027
970 Notably, the rates of bleeding events in both pivotal use of nonsteroidal anti-inflammatory drug (NSAIDs), 1028
971 studies were equivalent in the study drug and placebo which can impair renal blood flow in patients with 1029
972 arms, thus raising questions about the need for elevating cirrhosis.82 Similarly, oral combination products con- 1030
973 the platelet count before most invasive procedures. taining either acetaminophen or NSAIDs, especially when 1031
974 Moreover, thrombotic events have occurred with the use combined with a narcotic, should not be used and can 1032
975 of these agents in patients with liver disease, and the lead to accidental toxicity or overdose in a patient who is 1033
976 drugs should be used with caution. At this time, there is unaware that acetaminophen is a component of the 1034
977 no clear guidance in the literature on when to use these product. 1035
978 agents preoperatively in patients with cirrhosis, but it Avoid constipation in patients with cirrhosis to 1036
979 would seem reasonable to consider their use in patients minimize flares of hepatic encephalopathy in the post- 1037
980 undergoing an elective procedure associated with a high operative period. Polyethylene glycol solution may be 1038
981 bleeding risk if the baseline platelet count is less than used daily to maintain bowel regularity. In the absence of 1039
982 50,000/mL. Special attention should be paid to the po- comparative trials, rifaximin may be preferable in pa- 1040
983 tential for thrombotic events such as portal vein and other tients with hepatic encephalopathy who undergo pri- 1041
984 deep vein thromboses. Further clinical studies with these mary bowel surgery, because it generally causes less 1042
985 agents are ongoing. bowel distension than lactulose. 1043
986 1044

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 10 Northup et al Clinical Gastroenterology and Hepatology Vol. -, No. -

1045 Rescue Liver Transplantation 2. Aranha GV, Greenlee HB. Intra-abdominal surgery in patients 1103
1046 with advanced cirrhosis. Arch Surg 1986;121:275–277. 1104
1047 Before a surgical procedure, it is important to deter- 3. Garcia-Tsao G, Abraldes JG, Berzigotti A, et al. Portal hyper- 1105
1048 mine whether or not the patient will be a candidate for tensive bleeding in cirrhosis: Risk stratification, diagnosis, and 1106
management: 2016 practice guidance by the American Asso-
1049 liver transplantation in the future. In patients who, for 1107
ciation for the study of liver diseases. Hepatology 2017;
1050 whatever reason, are not potential candidates for liver 1108
65:310–335.
1051 transplantation, it is vital that this position be made clear 1109
4. Tripodi A, Mannucci PM. The coagulopathy of chronic liver
1052 before proceeding with surgery to avoid the need for the disease. N Engl J Med 2011;365:147–156.
1110
1053 patient’s family and the medical team to readdress this 5. Rios R, Sangro B, Herrero I, et al. The role of thrombopoietin in
1111
1054 issue if the patient’s condition deteriorates post- the thrombocytopenia of patients with liver cirrhosis. Am J 1112
1055 operatively. If the patient is a candidate for liver trans- Gastroenterol 2005;100:1311–1316. 1113
1056 plantation, a decision should be made as to whether the 6. Kim WR, Biggins SW, Kremers WK, et al. Hyponatremia and 1114
1057 planned surgery can be postponed until the patient has mortality among patients on the liver-transplant waiting list. 1115
1058 undergone liver transplantation. Elective operations such N Engl J Med 2008;359:1018–1026. 1116
1059 as orthopedic procedures may be best deferred despite 7. Fine RL. Futility, the Multiorganization Policy Statement, and the 1117
1060 uncertainty as to the length of time the patient will be on Schneiderman Response. Perspect Biol Med 2018;60:358–366. 1118
1061 the liver transplant waiting list. When a procedure 8. Teh SH, Nagorney DM, Stevens SR, et al. Risk factors for 1119
1062 cannot be delayed because of the adverse impact on the mortality after surgery in patients with cirrhosis. Gastroenter- 1120
1063 patient’s quality of life, the liver transplant evaluation ology 2007;132:1261–1269. 1121
1064 should be completed before the anticipated surgery, if 9. Pugh RN, Murray-Lyon IM, Dawson JL, et al. Transection of the 1122
1065 possible. Making a decision regarding liver transplant oesophagus for bleeding oesophageal varices. Br J Surg 1973; 1123
60:646–649.
1066 candidacy is not easy and depends in part on how 1124
10. Ziser A, Plevak DJ, Wiesner RH, et al. Morbidity and mortality in
1067 quickly the liver transplant evaluation can be completed 1125
cirrhotic patients undergoing anesthesia and surgery. Anesthe-
1068 if the patient deteriorates following surgery and how 1126
siology 1999;90:42–53.
1069 likely the patient will be to undergo transplantation in 1127
11. Telem DA, Schiano T, Goldstone R, et al. Factors that predict
1070 the near future. In general, if the postoperative mortality outcome of abdominal operations in patients with advanced
1128
1071 risk is calculated to be >15% at 3 months (using an cirrhosis. Clin Gastroenterol Hepatol 2010;8:451–457, quiz 1129
1072 appropriate risk calculator, as described previously) or if e458. 1130
1073 the MELD score is >15 (a score predictive of increased 12. Befeler AS, Palmer DE, Hoffman M, et al. The safety of intra- 1131
1074 mortality),16 preoperative evaluation for liver trans- abdominal surgery in patients with cirrhosis: model for end- 1132
1075 plantation is recommended. stage liver disease score is superior to Child-Turcotte-Pugh 1133
1076 classification in predicting outcome. Arch Surg 2005; 1134
1077 Conclusions 140:650–654, discussion 655. 1135
1078 13. Doyle DJ, Garmon EH. American Society of Anesthesiologists 1136
1079 Invasive surgery is frequently required in the patient
Classification (ASA Class). Treasure Island, FL: StatPearls, 2018. 1137
1080 with cirrhosis, in whom perioperative management is
14. Neeff H, Mariaskin D, Spangenberg HC, et al. Perioperative 1138
1081 challenging. Although morbidity and mortality rates
mortality after non-hepatic general surgery in patients with liver 1139
cirrhosis: an analysis of 138 operations in the 2000s using Child
1082 correlate with progression of portal hypertension in this 1140
and MELD scores. J Gastrointest Surg 2011;15:1–11.
1083 population, the MELD score and, to a lesser extent, CTP 1141
15. Malinchoc M, Kamath PS, Gordon FD, et al. A model to predict
1084 score (and Child-Pugh class) can provide some predic- 1142
poor survival in patients undergoing transjugular intrahepatic
1085 tion of risk in counseling patients and their families. A portosystemic shunts. Hepatology 2000;31:864–871.
1143
1086 multidisciplinary approach to these patients is critical for 16. Northup PG, Wanamaker RC, Lee VD, et al. Model for
1144
1087 optimizing all phases of perioperative care, including End-Stage Liver Disease (MELD) predicts nontransplant 1145
1088 planning, preoperative evaluation, intraoperative man- surgical mortality in patients with cirrhosis. Ann Surg 2005; 1146
1089 agement, and postoperative recovery. The data in the 242:244–251. 1147
1090 literature are limited in scope, usually lack risk-matched 17. Bohnen JD, Ramly EP, Sangji NF, et al. Perioperative risk factors 1148
1091 controls, and are based largely on single-center experi- impact outcomes in emergency versus nonemergency surgery 1149
1092 ences; therefore, evidence-based recommendations are
differently: time to separate our national risk-adjustment 1150
1093 mostly lacking. Study of larger cohorts with matched
models? J Trauma Acute Care Surg 2016;81:122–130. 1151
1094 controls and randomized controlled trials are needed to
18. Andraus W, Pinheiro RS, Lai Q, et al. Abdominal wall hernia in 1152
1095 improve our understanding and management of this
cirrhotic patients: emergency surgery results in higher morbidity 1153
1096 Q3 challenging population.
and mortality. BMC Surg 2015;15:65. 1154
1097 19. Odom SR, Gupta A, Talmor D, et al. Emergency hernia repair in 1155
cirrhotic patients with ascites. J Trauma Acute Care Surg 2013;
1098 1156
1099 References 75:404–409.
1157
1. Mokdad AA, Lopez AD, Shahraz S, et al. Liver cirrhosis mortality 20. Mullen MG, Michaels AD, Mehaffey JH, et al. Risk associated
1100 1158
in 187 countries between 1980 and 2010: a systematic analysis. with complications and mortality after urgent surgery vs elective
1101 BMC Med 2014;12:145. and emergency surgery: implications for defining “quality” and
1159
1102 1160

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 - 2018 Surgical Risk in Cirrhosis CPU 11

1161 reporting outcomes for urgent surgery. JAMA Surg 2017; prospective, randomized, double-blind, placebo-controlled, 1219
1162 152:768–774. multicenter trial. Hepatology 2007;46:1453–1463. 1220
1163 21. Garrison RN, Cryer HM, Howard DA, et al. Clarification of risk 37. Task Force on Patient Blood Management for Adult Cardiac 1221
1164 factors for abdominal operations in patients with hepatic Surgery of the European Association for Cardio-Thoracic Sur- 1222
cirrhosis. Ann Surg 1984;199:648–655. gery (EACTS) and the European Association of Cardiothoracic
1165 1223
22. Xu HW, Liu F, Li HY, et al. Outcomes following laparoscopic Anaesthesiology (EACTA), Boer C, et al. 2017 EACTS/EACTA
1166 1224
versus open major hepatectomy for hepatocellular carcinoma in Guidelines on patient blood management for adult cardiac sur-
1167 gery. J Cardiothorac Vasc Anesth 2018;32:88–120.
1225
patients with cirrhosis: a propensity score-matched analysis.
1168 Surg Endosc 2018;32:712–719. 38. Bizouarn P, Ausseur A, Desseigne P, et al. Early and late
1226
1169 23. Alshahrani AS, Gong GS, Yoo MW. Comparison of long-term outcome after elective cardiac surgery in patients with cirrhosis. 1227
1170 survival and immediate postoperative liver function after lapa- Ann Thorac Surg 1999;67:1334–1338. 1228
1171 roscopic and open distal gastrectomy for early gastric cancer 39. Chou AH, Chen TH, Chen CY, et al. Long-term outcome of 1229
1172 patients with liver cirrhosis. Gastric Cancer 2017;20:744–751. cardiac surgery in 1,040 liver cirrhosis patient - nationwide 1230
1173 24. Chmielecki DK, Hagopian EJ, Kuo YH, et al. Laparoscopic population-based cohort study. Circ J 2017;81:476–484. 1231
1174 cholecystectomy is the preferred approach in cirrhosis: a 40. Davis JPE, Northup PG, Caldwell SH, et al. Viscoelastic testing 1232
1175 nationwide, population-based study. HPB (Oxford) 2012; in liver disease. Ann Hepatol 2018;17:205–213. 1233
1176 14:848–853. 41. Fleming K, Redfern RE, March RL, et al. TEG-directed trans- 1234
1177 25. Eker HH, van Ramshorst GH, de Goede B, et al. A prospective fusion in complex cardiac surgery: impact on blood product 1235
1178 study on elective umbilical hernia repair in patients with liver usage. J Extra Corpor Technol 2017;49:283–290. 1236
1179 cirrhosis and ascites. Surgery 2011;150:542–546. 42. Arif R, Seppelt P, Schwill S, et al. Predictive risk factors for 1237
1180 26. Pei KY, Liu F, Zhang Y. A matched comparison of laparoscopic patients with cirrhosis undergoing heart surgery. Ann Thorac 1238
versus open inguinal hernia repair in patients with liver disease Surg 2012;94:1947–1952.
1181 1239
using propensity score matching. Hernia 2018;22:419–426. 43. Filsoufi F, Salzberg SP, Rahmanian PB, et al. Early and late
1182 1240
27. Llovet JM, Bru C, Bruix J. Prognosis of hepatocellular carci- outcome of cardiac surgery in patients with liver cirrhosis. Liver
1183 1241
noma: the BCLC staging classification. Semin Liver Dis 1999; Transpl 2007;13:990–995.
1184 19:329–338.
1242
44. Jacob KA, Hjortnaes J, Kranenburg G, et al. Mortality after
1185 28. Bruix J, Castells A, Bosch J, et al. Surgical resection of hepa- cardiac surgery in patients with liver cirrhosis classified by the
1243
1186 tocellular carcinoma in cirrhotic patients: prognostic value of Child-Pugh score. Interact Cardiovasc Thorac Surg 2015; 1244
1187 preoperative portal pressure. Gastroenterology 1996; 20:520–530. 1245
1188 111:1018–1022. 45. Macaron C, Hanouneh IA, Suman A, et al. Safety of cardiac 1246
1189 29. Rajakannu M, Coilly A, Adam R, et al. Prospective validation of surgery for patients with cirrhosis and Child-Pugh scores less 1247
1190 transient elastography for staging liver fibrosis in patients un- than 8. Clin Gastroenterol Hepatol 2012;10:535–539. 1248
1191 dergoing hepatectomy and liver transplantation. J Hepatol 2017 46. Thielmann M, Mechmet A, Neuhauser M, et al. Risk prediction 1249
1192 Aug 23 [E-pub ahead of print]. and outcomes in patients with liver cirrhosis undergoing open- 1250
1193 30. Llop E, Berzigotti A, Reig M, et al. Assessment of portal hy- heart surgery. Eur J Cardiothorac Surg 2010;38:592–599. 1251
1194 pertension by transient elastography in patients with compen- 47. Suman A, Barnes DS, Zein NN, et al. Predicting outcome after 1252
1195 sated cirrhosis and potentially resectable liver tumors. J Hepatol cardiac surgery in patients with cirrhosis: a comparison of Child- 1253
2012;56:103–108.
1196 Pugh and MELD scores. Clin Gastroenterol Hepatol 2004; 1254
31. Teh SH, Christein J, Donohue J, et al. Hepatic resection of 2:719–723.
1197 1255
hepatocellular carcinoma in patients with cirrhosis: Model of 48. Thakkar B, Patel A, Mohamad B, et al. Transcatheter aortic
1198 1256
End-Stage Liver Disease (MELD) score predicts perioperative valve replacement versus surgical aortic valve replacement in
1199 mortality. J Gastrointest Surg 2005;9:1207–1215, discussion
1257
patients with cirrhosis. Catheter Cardiovasc Interv 2016;
1200 1215. 87:955–962.
1258
1201 32. Itoi T, Coelho-Prabhu N, Baron TH. Endoscopic gallbladder 49. Townsend JC, Heard R, Powers ER, et al. Usefulness of inter-
1259
1202 drainage for management of acute cholecystitis. Gastrointest national normalized ratio to predict bleeding complications in 1260
1203 Endosc 2010;71:1038–1045. patients with end-stage liver disease who undergo cardiac 1261
1204 33. Bingener J, Cox D, Michalek J, et al. Can the MELD score catheterization. Am J Cardiol 2012;110:1062–1065. 1262
1205 predict perioperative morbidity for patients with liver cirrhosis 50. Krill T, Brown G, Weideman RA, et al. Patients with cirrhosis who 1263
1206 undergoing laparoscopic cholecystectomy? Am Surg 2008; have coronary artery disease treated with cardiac stents have 1264
1207 74:156–159. high rates of gastrointestinal bleeding, but no increased mor- 1265
1208 34. Poggio JL, Rowland CM, Gores GJ, et al. A comparison of tality. Aliment Pharmacol Ther 2017;46:183–192. 1266
1209 laparoscopic and open cholecystectomy in patients with 51. Russo MW, Pierson J, Narang T, et al. Coronary artery stents 1267
1210 compensated cirrhosis and symptomatic gallstone disease. and antiplatelet therapy in patients with cirrhosis. J Clin Gas- 1268
Surgery 2000;127:405–411.
1211 troenterol 2012;46:339–344. 1269
35. Shaheen AA, Kaplan GG, Hubbard JN, et al. Morbidity and
1212 52. Vilar-Gomez E, Martinez-Perez Y, Calzadilla-Bertot L, et al. 1270
mortality following coronary artery bypass graft surgery in pa- Weight loss through lifestyle modification significantly reduces
1213 1271
tients with cirrhosis: a population-based study. Liver Int 2009; features of nonalcoholic steatohepatitis. Gastroenterology 2015;
1214 29:1141–1151.
1272
149:367–378.e5, quiz e14–5.
1215 36. Lewis JH, Mortensen ME, Zweig S, et al. Efficacy and safety of
1273
53. Pestana L, Swain J, Dierkhising R, et al. Bariatric surgery in
1216 high-dose pravastatin in hypercholesterolemic patients with patients with cirrhosis with and without portal hypertension: a
1274
1217 well-compensated chronic liver disease: results of a single-center experience. Mayo Clin Proc 2015;90:209–215. 1275
1218 1276

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB

 12 Northup et al Clinical Gastroenterology and Hepatology Vol. -, No. -

1277 54. Hanipah ZN, Punchai S, McCullough A, et al. Bariatric surgery in chronic liver disease in the HALT-C trial. Clin Gastroenterol 1340
1278 patients with cirrhosis and portal hypertension. Obes Surg 2018; Hepatol 2010;8:877–883. 1341
1279 28:3431–3438. 73. Afdhal NH, Dusheiko GM, Giannini EG, et al. Eltrombopag in- 1342
1280 55. Mosko JD, Nguyen GC. Increased perioperative mortality creases platelet numbers in thrombocytopenic patients with 1343
1281 following bariatric surgery among patients with cirrhosis. Clin HCV infection and cirrhosis, allowing for effective antiviral 1344
1282 Gastroenterol Hepatol 2011;9:897–901. therapy. Gastroenterology 2014;146:442–452.e1. 1345
1283 56. Shimizu H, Phuong V, Maia M, et al. Bariatric surgery in patients 74. Terrault N, Chen YC, Izumi N, et al. Avatrombopag before pro- 1346
1284 with liver cirrhosis. Surg Obes Relat Dis 2013;9:1–6. cedures reduces need for platelet transfusion in patients with 1347
1285 57. Heimbach JK, Watt KD, Poterucha JJ, et al. Combined liver chronic liver disease and thrombocytopenia. Gastroenterology 1348
1286 transplantation and gastric sleeve resection for patients with 2018;155:705–718. 1349
1287 medically complicated obesity and end-stage liver disease. Am 75. Tateishi R, Seike M, Kudo M, et al. A randomized controlled trial 1350
1288 J Transplant 2013;13:363–368. of lusutrombopag in Japanese patients with chronic liver dis- 1351
1289 ease undergoing radiofrequency ablation. J Gastroenterol 2018 1352
58. Zamora-Valdes D, Watt KD, Kellogg TA, et al. Long-term out-
1290 Aug 13 [E-pub ahead of print]. 1353
comes of patients undergoing simultaneous liver transplantation
1291 1354
and sleeve gastrectomy. Hepatology 2018;68:485–495. 76. Drolz A, Horvatits T, Roedl K, et al. Coagulation parameters and
1292 1355
59. Al-Nowaylati AR, Al-Haddad BJ, Dorman RB, et al. Gastric major bleeding in critically ill patients with cirrhosis. Hepatology
1293 1356
bypass after liver transplantation. Liver Transpl 2013; 2016;64:556–568.
1294 1357
19:1324–1329. 77. Wikkelso A, Wetterslev J, Moller AM, et al. Thromboelastog-
1295 1358
60. Lin MY, Tavakol MM, Sarin A, et al. Safety and feasibility of raphy (TEG) or thromboelastometry (ROTEM) to monitor hae-
1296 1359
sleeve gastrectomy in morbidly obese patients following liver mostatic treatment versus usual care in adults or children with
1297 1360
transplantation. Surg Endosc 2013;27:81–85. bleeding. Cochrane Database Syst Rev 2016;(8):CD007871.
1298 1361
61. Runyon BA, Committee APG. Management of adult patients 78. De Pietri L, Bianchini M, Montalti R, et al. Thrombelastography-
1299 1362
with ascites due to cirrhosis: an update. Hepatology 2009; guided blood product use before invasive procedures in
1300 1363
49:2087–2107. cirrhosis with severe coagulopathy: a randomized, controlled
1301 1364
62. Lieberman FL, Reynolds TB. Plasma volume in cirrhosis of the trial. Hepatology 2016;63:566–573.
1302 1365
1303 liver: its relation of portal hypertension, ascites, and renal failure. 79. Tegeder I, Lotsch J, Geisslinger G. Pharmacokinetics of opioids 1366
1304 J Clin Invest 1967;46:1297–1308. in liver disease. Clin Pharmacokinet 1999;37:17–40. 1367
1305 63. Semiz-Oysu A, Moustafa T, Cho KJ. Transjugular intrahepatic 80. Kotb HI, Fouad IA, Fares KM, et al. Pharmacokinetics of oral 1368
1306 portosystemic shunt before cardiac surgery with cardiopulmo- tramadol in patients with liver cancer. J Opioid Manag 2008; 1369
1307 nary bypass in patients with cirrhosis and portal hypertension. 4:99–104. 1370
1308 Heart Lung Circ 2007;16:465–468. 81. Benson GD. Acetaminophen in chronic liver disease. Clin 1371
1309 64. Schlenker C, Johnson S, Trotter JF. Preoperative transjugular Pharmacol Ther 1983;33:95–101. 1372
1310 intrahepatic portosystemic shunt (TIPS) for cirrhotic patients 82. Laffi G, La Villa G, Pinzani M, et al. Arachidonic acid derivatives 1373
1311 undergoing abdominal and pelvic surgeries. Surg Endosc 2009; and renal function in liver cirrhosis. Semin Nephrol 1997; 1374
1312 23:1594–1598. 17:530–548. 1375
1313 65. Gil A, Martinez-Regueira F, Hernandez-Lizoain JL, et al. The role 83. Fernandes NF, Schwesinger WH, Hilsenbeck SG, et al. Lapa- 1376
1314 of transjugular intrahepatic portosystemic shunt before roscopic cholecystectomy and cirrhosis: a case-control study of 1377
1315 abdominal tumoral surgery in cirrhotic patients with portal hy- outcomes. Liver Transpl 2000;6:340–344. 1378
1316 pertension. Eur J Surg Oncol 2004;30:46–52. 84. Iwasaki A, Shirakusa T, Okabayashi K, et al. Lung cancer sur- 1379
1317 66. Guerrini GP, Pleguezuelo M, Maimone S, et al. Impact of tips gery in patients with liver cirrhosis. Ann Thorac Surg 2006; 1380
1318 preliver transplantation for the outcome posttransplantation. Am 82:1027–1032. 1381
1319 J Transplant 2009;9:192–200. 85. Lee JH, Yu CS, Lee JL, et al. Factors affecting the post- 1382
1320 67. Boyer TD, Haskal ZJ, American Association for the Study of operative morbidity and survival of patients with liver cirrhosis 1383
1321 Liver D. The role of transjugular intrahepatic portosystemic following colorectal cancer surgery. Int J Colorectal Dis 2017; 1384
1322 shunt (TIPS) in the management of portal hypertension: update 32:521–530. 1385
1323 2009. Hepatology 2010;51:306. 86. Liao JC, Chen WJ, Chen LH, et al. Complications associated 1386
1324 68. Tripodi A, Salerno F, Chantarangkul V, et al. Evidence of normal with instrumented lumbar surgery in patients with liver 1387
1325 thrombin generation in cirrhosis despite abnormal conventional cirrhosis: a matched cohort analysis. Spine J 2013;13: 1388
1326 coagulation tests. Hepatology 2005;41:553–558. 908–913. 1389
1327 1390
69. Tripodi A, Primignani M, Lemma L, et al. Evidence that low 87. Chen CC, Hsu PW, Lee ST, et al. Brain surgery in patients with
1328 1391
protein C contributes to the procoagulant imbalance in cirrhosis. liver cirrhosis. J Neurosurg 2012;117:348–353.
1329 1392
J Hepatol 2013;59:265–270.
1330 1393
70. Northup PG, Caldwell SH. Coagulation in liver disease: a guide
1331 Reprint Requests 1394
for the clinician. Clin Gastroenterol Hepatol 2013;11:1064–1074.
1332 Address requests for reprints to: Patrick G. Northup, MD, Division of Gastro- 1395
71. Tripodi A, Primignani M, Chantarangkul V, et al. Thrombin enterology and Hepatology, University of Virginia, JPA and Lee Street, MSB
1333 1396
generation in patients with cirrhosis: the role of platelets. Hep- 2145, Charlottesville, Virginia 22908–0708. e-mail: northup@virginia.edu; fax:
1334 (434) 244-9454. Q2 1397
atology 2006;44:440–445.
1335 1398
1336 72. Seeff LB, Everson GT, Morgan TR, et al. Complication rate of Conflict of interest 1399
1337 percutaneous liver biopsies among persons with advanced The authors disclose no conflicts.
1400
1338 1401
1339 1402

REV 5.5.0 DTD
YJCGH56114_proof
24 October 2018
2:19 pm
ce OB