Abstract
Background
Obesity poses unique risks in patients with advanced liver fibrosis; however, given surgical risks of bariatric surgery in cirrhosis treatment recommendations are currently limited to lifestyle interventions. This study seeks to inform a potential treatment gap by describing the safety and efficacy of pharmacologic weight loss in patients with advanced liver disease.
Methods
A retrospective chart review of the electronic medical record was conducted for all patients in the Scripps Health system from 2005 to 2017 with established advanced liver fibrosis that were prescribed medications associated with weight loss. The primary outcome was safety as defined by the model for end‐stage liver disease (MELD) score. Secondary outcomes included total body weight loss, reasons for medication discontinuation, medication adverse events, and hospitalization before and after medication initiation.
Results
Thirty‐eight patients and 63 prescriptions were included in the final analysis. The most frequently prescribed medication associated with weight loss was metformin (63%, n = 24) followed by a GLP‐1 agonist (39%, n = 15). There was no significant effect of weight‐loss medication on MELD score (p > 0.18) or number of hospitalizations when adjusting for subject (p > 0.26). There was a significant adjusted mean weight loss of 2.2 kg (p < 0.02) following prescription of a medication associated with weight loss. The Federal Drug Administration‐approved anti‐obesity medications as a group resulted in a significant adjusted weight loss of 7.22 kg (p < 0.013). In a linear mixed‐effects model accounting for subjects, weight loss was not independently associated with a change in MELD (t[51] = −1.972, p > 0.05).
Conclusion
Pharmacologic weight loss in patients with advanced liver fibrosis appears feasible based on preliminary safety and efficacy outcomes in this study. Future prospective studies are warranted to evaluate a potential significant treatment gap in the management of obesity in this vulnerable population.
Keywords: cirrhosis, liver disease, medications, obesity, weight loss
1. INTRODUCTION
Obesity in the general population has been linked to multiple liver‐related morbidities including primary liver malignancy and nonalcoholic steatohepatitis with progressive fibrosis leading to cirrhosis.1, 2 The risk of hepatocellular carcinoma is nearly doubled in patients with body mass index (BMI) ≥30 kg/m2 compared to BMI <25 kg/m2 independent of other known risk factors.3, 4 It's not surprising then that in patients with underlying chronic liver disease obesity has been shown to be an independent risk factor of fibrosis progression and clinical decompensation.5, 6 In a cohort of patients with cirrhosis of mixed etiology, a one unit increase in BMI independently predicted a 6% increased risk of clinical decompensation over an average 5 year follow‐up period. 7
In patients with decompensated liver cirrhosis, liver transplantation is often the only intervention capable of extending life expectancy and improving quality of life. 8 Unfortunately, elevated BMI, especially 40 kg/m2 and above, has been shown to result in up to 41% increased risk of death posttransplant due to infectious complications and cancer. 9 For these reasons liver transplantation is a relative contraindication in patients with class 3 obesity. 10
The benefits of weight loss in patients with obesity are well‐established and extend to patients with cirrhosis and liver‐related morbidities with additional unique considerations as discussed above.11, 12, 13, 14 Unfortunately, treatment of obesity in this patient population is limited by increased operative risks of bariatric surgery, 15 limited data for endoscopic bariatric procedures, and the complexity of nutritional management and weight loss in patients with concurrent sarcopenia, catabolic chronic disease, and elevated BMI. Current guideline recommendations for weight loss in patients with cirrhosis are limited to lifestyle interventions under the supervision of a dietician. 10 Bariatric surgery is contraindicated in patients with decompensated cirrhosis and not routinely recommended in those with compensated cirrhosis or advanced fibrosis.
To date there have been no recommendations or data published promoting pharmacologic management of obesity in patients with cirrhosis, representing a potential treatment gap. The American Association of Clinical Endocrinologists and American College of Endocrinology 2016 clinical practice guidelines suggest an overall favorable safety profile of weight‐loss medications in patients with liver disease and provide guidance for dosing adjustments. Ultimately, however, the guidelines simply advise caution with the use of weight‐loss medications in patients with hepatic impairment and avoidance in patients with severe hepatic impairment defined as Child‐Pugh score >9. 16 Indeed, the use of weight‐loss medications in patients with cirrhosis is likely limited in practice due to the lack of safety data and complexity of the true nutritional status in this patient population. The purpose of this study is to provide preliminary data on the safety of pharmacologic agents typically prescribed for weight loss in patients with advanced liver disease, as well as the potential efficacy.
2. MATERIALS AND METHODS
2.1. Design
A retrospective chart review of the electronic medical record (EMR) was conducted for all patients in the Scripps health system over a 12‐year period between 2005 and 2017. Scripps Health is a health system in San Diego county consisting of 5 hospitals and over 20 ambulatory care sites caring for over 1 million patients annually with support by various multispecialty groups. This study was approved and waiver of authorization for research granted by the Scripps Health Institutional Review Board.
2.2. Sample
Potential subjects were identified by an EMR query for adults over the age of 18 years old with a diagnosis of cirrhosis or advanced liver fibrosis based on ICD‐10 coding and a documented prescription of a weight‐loss medication during the study period. Weight‐loss medications as defined in this study consist of medications associated with weight loss and Federal Drug Administration (FDA)‐approved anti‐obesity medications. Multiple prescriptions were considered for the same patient only for different weight‐loss medications. Only the initial prescription event was considered when consecutive prescriptions were encountered for the same medication in a single patient.
Inclusion criteria of the study were confirmation of cirrhosis based on liver biopsy or hepatology subspecialty evaluation in the EMR and confirmation of a weight‐loss medication in the EMR that was started following the diagnosis of cirrhosis. From initial screening, n = 75 patients met inclusion criteria (Figure 1). Exclusion criteria were patients with neither model for end‐stage liver disease (MELD) labs nor weights documented in the EMR and for prescriptions of weight‐loss medication that were started after liver transplant. Specifically, weight‐loss prescriptions were excluded from the primary safety outcome analysis if there was a concurrent warfarin prescription during the weight‐loss medication prescription period. Criteria for exclusion from the secondary outcome analysis of weight loss were concurrent prescriptions of furosemide, spironolactone, bumetanide, or metolazone during the prescription period of the weight‐loss medication or a prescription period of less than 30 days (Figure 2). Obesity status was not considered in the inclusion or exclusion criteria since the study's primary outcome was related to medication safety in patients with advanced liver disease. Because pharmacologic management of obesity in patients with advanced liver disease is not common in clinic practice as described above, many of the patients were anticipated to receive the weight‐loss medications as defined in this study for indications other than obesity.
FIGURE 1.
Flow diagram depicting study sample following inclusion and exclusion criteria
FIGURE 2.
Flow diagram depicting prescription inclusion and exclusion criteria for study analyses
A total of n = 38 patients and n = 63 prescriptions met inclusion and exclusion criteria for the retrospective analysis (Figures 1 and 2).
2.3. Data collection
The primary outcome was safety as defined by change in MELD score before and 1 year after initiation of weight‐loss medication. Secondary outcomes included hospitalizations 1 year before and 1 year after medication initiation, reasons for medication discontinuation, medication adverse events, and weight loss at 1 year after initiation of weight‐loss medication.
All MELD scores were calculated using the revised 2016 Organ Procurement and Transplantation Network and the United Network for Organ Sharing equation. Premedication measurements of MELD score and weight were retrieved at time of medication start or the closest measurement within the previous year. Post‐medication measurements of MELD score and weight were retrieved at time of medication end or closest measurement to 1 year after medication start, whichever came sooner, to mitigate potential changes related to the natural history of cirrhosis. The number of hospitalizations outcome was collected for the year prior to and year following medication start. Child's Pugh Class (CPC) was retrieved when available from hepatology consultation reports within the year preceding start of weight‐loss medication. If the documentation reported decompensated cirrhosis but did not specify CPC this was collected as well.
2.4. Statistical analyses
Patient characteristics, medications prescribed, and reasons for medication discontinuation were descriptively analyzed as percentages of the sample for categorical variables, or as medians and interquartile ranges for continuous variables since the data were not normally distributed. The outcomes of changes in MELD scores, weight, and hospitalizations were compared pre‐ and post‐medication for each prescription on record and adjusted to account for individuals that had multiple weight‐loss medication prescriptions through linear mixed‐effect models. Therefore, all outcomes described in terms of change (i.e., weight, MELD change) are derived from regression coefficients and are interpreted as expected average weight change to account for the subject effect in the mixed effects model. Time between measurements was tested as a fixed effect for each outcome, and each mixed effects model included subject as a random effect. Follow‐up mixed effects models were run for each outcome to determine whether any weight‐loss medication class was predictive of a change in the outcome over time.
Medication classes used in these models included the FDA‐approved anti‐obesity medications GLP‐1 receptor agonists, phentermine, and lorcaserin as well as medications associated with weight loss including metformin, SGLT‐2 inhibitors, topiramate, and bupropion. FDA‐approved anti‐obesity medications were included in the model as a group post‐hoc due to low sample sizes of lorcaserin and phentermine. Adjusted means (estimated marginal means) of outcome measures were reported for any significant predictors. Associations of change in weight with MELD score were also tested in a mixed effects model with change in MELD as the dependent and change in weight as the independent variable. Change in MELD scores, changes in weight, and the occurrence of any adverse events were also run as dependent variables in linear mixed‐effects models for MELD and weight, and in a generalized logistic mixed effects model for adverse events with CPC using a dummy‐coded predictor for CPC A versus all other classes. Exploratory analyses of predictors of change in MELD scores and change in weight were independently conducted through a linear mixed effects model with percent weight change, sex, ethnicity, age at diagnosis, etiology of liver disease, diuretics, bariatric surgery, history of transplant, and known weight‐gain medication use. The model was reduced for variables that reached significance in the full model, and only included weight change and alcoholic liver disease etiology in the final model. The statistical software R v 3.5.3 and GraphPad Prism 8 were used to conduct all analyses. A two‐tailed p‐value of 0.05 or less was considered statistically significant.
3. RESULTS
Thirty‐eight patients and 63 prescriptions were included in the final analysis. Patient characteristics are shown in Table 1. The median BMI of 33.11 (IQR = 11.26) was consistent with class I obesity. Over 65% (n = 41) of patients had a BMI of 30.0 or above. The majority of cirrhosis cases were secondary to Non‐alcoholic steatohepatitis (NASH) (55%, n = 21). The most frequently prescribed medication associated with weight loss was metformin (63%, n = 24) followed by a GLP‐1 agonist (39%, n = 15) (see Table 2). Collectively, FDA‐approved anti‐obesity medications (GLP‐1 agonists, lorcaserin, and phentermine) made up 50% of all prescriptions and were presented collectively in efficacy analyses due to sample size limitations. Naltrexone and topiramate were not included in efficacy analyses due to small sample size in the present study. Half of the patients (n = 19) were also on a medication known to cause weight gain. The most common weight‐gain medication was a beta‐blocker which was prescribed in 29% (n = 11) of patients (Table 2).
TABLE 1.
Patient characteristics are described as either mean (SD) for continuous, normally distributed data, median (IQR) for continuous, skewed data, or as a percentage (n) of total sample for categorical data
| Patient characteristics | |
|---|---|
| Mean (SD) | |
| Age | 55.05 (2.85) |
| Baseline weight (kg) | 91.96 (22.10) |
| Median (IQR) | |
| BMI at medication start | 33.11 (11.26) |
| % (n) | |
| Sex (male) | 42% (16) |
| Ethnicity (Hispanic) | 18% (7) |
| Diabetes | 84% (32) |
| Etiology of cirrhosis | |
| NASH | 55% (21) |
| HCV | 32% (12) |
| EtOH | 16% (6) |
| Other (HBV, PBC, AIH) | 11% (4) |
| Diuretic use >3 months | 24% (9) |
| Coumadin | 3% (1) |
| Prior bariatric surgery | 11% (4) |
| Received transplant | 13% (5) |
| Cancer diagnosis (HCC) | 13% (5) |
Abbreviations: AIH, Autoimmune hepatitis; BMI, body mass index; EtOH, Ethanol; HBV, hepatitis B; HCC, hepatocellular carcinoma; HCV, hepatitis C; IQR, Interquartile range; NASH, Non‐alcoholic steatohepatitis; PBC, Primary biliary cholangitis.
TABLE 2.
Prescription characteristics are shown for all weight‐gain medications and broken down into percentage (n) of patients with a prescription of each class of weight‐gain medication
| Weight‐gain medications | % (n) of patients |
|---|---|
| Beta‐blockers | 29% (11) |
| Insulin | 21% (8) |
| Sulfonylurea/TZD | 18% (7) |
| Anti‐psychotics | 8% (3) |
| Prednisone (>3 months) | 3% (1) |
| Gabapentin | 3% (1) |
| Any weight‐gain medication | 50% (19) |
Abbreviation: TZD, thiazolidinedione.
TABLE 3.
Prescription characteristics are shown for all weight‐negative medications, with each category further broken down into percentage (n) of total sample of prescriptions in each class as well as median (and range) of doses, duration of medication prescriptions, and interval follow‐up of MELD and weight in days
| Weight‐loss medication | % (n) | Total dose median [range] | Median time on treatment (days) [range] | Median time to MELD follow‐Up (days) [range] | Median time to weight follow‐Up (days) [range] |
|---|---|---|---|---|---|
| Metformin | 63% (24) | 1000 [500–2000] (mg/day) | 1153 [172–2462] | 363 [47–622] | 335 [140–463] |
| GLP‐1 agonists | 39% (15) | ‐ | 708 [36–1586] | 353 [47–794] | 296.5 [64–439] |
| Liraglutide | ‐ | 1.2 [1.2–2.4] (mg/day) | 616.5 [36–1586] | 353 [91–794] | 305 [64–366] |
| Dulaglutide | ‐ | 0.75 [0.25–1.5] (mg/week) | 567 [198–929] | 327 [47–546] | 140 [113–364] |
| Exenatide | ‐ | 15 [10–20] (mcg/day) a | 875 [249–1166] | 332.5 [228–441] | 332 [228–439] |
| SGLT‐2 inhibitors | 24% (9) | ‐ | 357 [12–1314] | 363 [47–454] | 326 [94–737] |
| Canagliflozin | ‐ | 300 [100–300] (mg/day) | 357 [95–1314] | 395 [94–454] | 326 [94–737] |
| Dapagliflozin | ‐ | 5 (mg/day) | 12 | b | c |
| Empagliflozin | ‐ | 10 [10–25] (mg/day) | 519 [13–708] | 47 | 252 [140–364] |
| Bupropion | 16% (6) | 225 [100–300] (mg/day) | 926.5 [346–3509] | 256 [244–349] | 356.5 [258–618] |
| Topiramate | 8% (3) | 50 [25–75] (mg/day) | 1011 [750–1011] | 431 [391–589] | 397 [376–500] |
| Phentermine | 8% (3) | 16 [15–30] (mg/day) | 135 [29–146] | 215 [213–217] | 135 [29–151] |
| Naltrexone | 5% (2) | 27 [4–50] (mg/day) | 734.5 [102–1367] | 455 [101–809] | 232.5 [101–364] |
| Lorcaserin | 3% (1) | 20 (mg/day) | 275 | 317 | 330 |
Abbreviation: MELD, Model for End‐Stage Liver Disease.
A 2 mg/week dose was considered equivalent to 10 mcg/day.
No MELD labs meeting data collection criteria.
Excluded from weight loss analysis due to prescription period <30 days.
CPC at the time of medication start was identified in 82% (n = 52) of prescriptions. Among all prescriptions, the severity of liver disease at time of medication start was identified as CPC A in 59% (n = 37), CPC B in 14% (n = 9), CPC C in 3% (n = 2), and decompensated cirrhosis in 6% (n = 4). Supporting the finding of predominantly compensated liver disease in this study cohort was also a median MELD of 7 Interquartile range (IQR = 4) and median number of hospitalizations of zero (IQR = 0) prior to medication start.
With regards to safety outcomes, there was no significant overall effect of weight‐negative medications on MELD score or number of hospitalizations when adjusting for subject (t[52] = 1.593, p > 0.11; t[35.9] = −1.121, p > 0.26, respectively). There were additionally no significant effects of any particular medication class tested (all FDA‐approved weight‐loss medications, metformin, SGLT‐2 inhibitors, bupropion; see Table 4). All FDA‐approved weight‐loss medications were presented collectively since phentermine and lorcaserin were infrequently prescribed in the present study; however, no conclusions were changed when GLP‐1 agonists were analyzed as a standalone group, which also had nonsignificant changes in MELD and hospitalizations (p's > 0.4). Of all weight‐loss medications, 40% (n = 25) were discontinued. Reasons for discontinuation are shown in Figure 3. Adverse effects constituted the reason for discontinuation in 32% (n = 8) of prescriptions with one event requiring hospitalization for diabetic ketoacidosis and acute pancreatitis attributed to SGLT‐2 initiation in a patient with decompensated alcoholic liver disease (see Table 6). Thirty‐two percent (n = 8) of patients discontinued medications after undergoing successful liver transplantation. CPC did not significantly predict incidence of adverse events (>0.5, data not shown).
TABLE 4.
Patient‐adjusted mixed effects models show the effects of classes of weight‐negative medications on the outcomes of MELD score change and change in hospitalizations before and after prescription
| MELD change | |||
|---|---|---|---|
| Average MELD change a | SE | p | |
| Approved anti‐obesity med. | −0.317 | 2.212 | 0.887 |
| Metformin | 1.123 | 2.052 | 0.587 |
| SGLT‐2 inhibitor | −1.119 | 2.441 | 0.649 |
| Bupropion | 0.530 | 2.829 | 0.852 |
| Hospitalizations change | |||
|---|---|---|---|
| Average change in number of hospitalizations a | SE | p | |
| Approved anti‐obesity med. | −0.217 | 0.486 | 0.659 |
| Metformin | 0.121 | 0.446 | 0.788 |
| SGLT‐2 inhibitor | −0.264 | 0.504 | 0.604 |
| Bupropion | 0.182 | 0.662 | 0.784 |
Notes: Regression coefficients (b), their SE, and p‐values presented are derived from linear mixed effects models adjusting for subject as a random effect.
Expected average change based on regression coefficient from a mixed effects model adjusting for subject as a random effect.
FIGURE 3.
Pie chart representing reasons for discontinuation of weight‐negative medications
TABLE 6.
Adverse events that were reported as reason for discontinuation are listed by weight‐negative medication class and include overview of adverse events
| Medication class (n) | Adverse event |
|---|---|
| GLP‐1 agonist (2) | Decreased appetite; nausea |
| SGLT‐2 (3) | Vulvovaginal itching; acute pancreatitis a ; urinary frequency |
| Metformin (2) | GI side effects; headache |
| Lorcaserin (1) | Cognitive dysfunction |
Required hospitalization.
With regards to secondary outcomes, there was a significant adjusted mean weight loss of 2.2 kg (t[37.3] = −2.46, p < 0.02) following prescription of a weight‐loss medications (Table 5, Figure 4), and this was not significantly predicted by the time between weight measurements (p > 0.2; data not shown). Each of the weight‐loss medications demonstrated a trend toward expected weight loss in the model but none were statistically significant. The FDA‐approved anti‐obesity medication as a group, however, resulted in a significant adjusted weight loss of 7.22 kg, on average (p < 0.013) (Table 5). In a linear mixed‐effects model accounting for subjects, weight loss was not significantly associated with a change in MELD (t[51] = −1.972, p > 0.05). However, when weight change is included in an exploratory mixed‐effects linear model with other potential clinical predictors, the reduced model indicates that weight change inversely (β = −0.201 ± 0.080, p < 0.04) and alcoholic liver disease (β = 4.101 ± 0.1.737, p < 0.03) positively predict change in MELD score following weight‐loss medication use. Child's Pugh scores were not predictive of either change in MELD or change in weight following weight‐loss medications (t[19.85] = −1.398, p > 0.17; t[41.02] = −0.198, p > 0.84).
TABLE 5.
Patient‐adjusted mixed effects models show the overall change in weight (kg) among all weight‐negative medications, and the effects of classes of weight‐negative medications on weight loss before and after prescription
| Weight change (kg) | |||||
|---|---|---|---|---|---|
| Median wt pre‐med | Median wt post‐med | Average wt change a | SE | p‐value | |
| All weight‐negative med | 92.53 | 87.09 | −2.23 | 0.91 | 0.0186 |
| By medication type | |||||
| FDA‐approved anti‐obesity medications | 92.31 | 87.77 | −7.22 | 2.79 | 0.013 |
| GLP‐1 agonists | 91.63 | 87.32 | ‐ | ‐ | ‐ |
| Phentermine | 92.99 | 85.27 | ‐ | ‐ | ‐ |
| Lorcaserin b | 104.78 | 100.70 | ‐ | ‐ | ‐ |
| Metformin | 91.17 | 81.19 | −4.14 | 2.58 | 0.115 |
| SGLT‐2 inhibitor | 123.83 | 113.40 | −4.15 | 2.91 | 0.162 |
| Bupropion | 100.24 | 94.57 | −4.96 | 3.61 | 0.1751 |
| Topiramate c | 92.08 | 88.00 | ‐ | ‐ | ‐ |
| Naltrexone c | 72.12 | 78.70 | ‐ | ‐ | ‐ |
Note: Median weight (kg) pre‐and post‐prescription are shown. Regression coefficients (b), their SE, and p‐values presented are derived from linear mixed effects models adjusting for subject as a random effect.
Abbreviation: FDA, Federal Drug Administration.
Expected average weight change (kg) based on regression coefficient from a mixed effects model adjusting for subject as a random effect.
Lorcaserin median weights represent a single sample.
Topiramate and Naltrexone not included in mixed effects model for medication types due to low sample size (n = 3 and 2, respectively).
FIGURE 4.
Spider plot of unadjusted pre‐ and post‐medication weights per prescription
4. DISCUSSION
Treatment of obesity in patients with cirrhosis is a common clinical challenge with significant potential implications on morbidity and mortality. Current guideline recommendations are limited in patients with liver disease due to a lack of data, representing a true unmet need. While pharmacologic treatment can be two to three times more effective at achieving weight loss compared to placebo, 17 there is a scarcity of information regarding use of pharmacologic agents for the treatment of obesity in patients with advanced liver fibrosis and cirrhosis.
This study provided novel retrospective data on the safety and efficacy of medications associated with weight loss, including FDA‐approved anti‐obesity medications, in a patient population with advanced liver disease. Study results did not show any concerning safety signals such as increased MELD scores or hospitalizations after median follow‐up of 1 year after initiation of a medication associated with weight loss. While weight loss significantly predicted increase in MELD in an exploratory mixed‐effects linear model, the effect was small and weight loss was otherwise not independently associated with a significant increase in MELD score. Therefore, despite the high frequency of protein‐calorie malnutrition in cirrhotic patients, associated risks of increased mortality, and clinical complications of sarcopenic obesity,18, 19, 20 preliminary data on pharmacologic weight loss in patients with mixed compensated and decompensated cirrhosis do not suggest significant exacerbation of this pathophysiology. These findings should be confirmed in a clinical trial setting and certainly cannot be extrapolated to patients with decompensated, high MELD cirrhosis.
As a secondary endpoint, medication discontinuation was common at 40%. Approximately one‐third of discontinuations were due to successful liver transplantation. Medication discontinuation due to cost or insurance coverage occurred in 8% of cases, a relatively low occurrence in the setting of weight‐loss pharmacology likely driven by inclusion of metformin in the study. While 32% of discontinuations were due to a variety of reported medication adverse effects, only one was severe enough to result in hospitalization. This event occurred in a patient with CPC B alcohol‐related cirrhosis and active alcohol use admitted for diabetic ketoacidosis and acute pancreatitis. The patient had a prior episode of alcohol‐induced pancreatitis; however, dapagliflozin had been initiated one week prior to presentation and was therefore presumed to be the etiology of his acute pancreatitis. While this adverse effect has been reported for SGLT‐2 inhibitors and underwent investigation by the United States Food and Drug Administration in 2016, a recent cohort study and meta‐analysis did not show increased risk of acute pancreatitis with this class of medications.21, 22, 23 Furthermore, dapagliflozin has been shown previously to be well‐tolerated in hepatic impairment including CPC C cirrhosis. 24 Overall, based on the results of the primary and secondary endpoints of safety and at least 23% of prescriptions occurring in the setting of non‐CPC A cirrhosis, commonly prescribed medications associated with weight loss appear well‐tolerated in patients with advanced hepatic fibrosis.
Secondary end‐points of efficacy demonstrated an overall modest weight loss at a median of approximately 2.4% total body weight when all weight‐loss medications were analyzed as a group. The prescription numbers in this study were likely too small to make definitive conclusions regarding the efficacy of individual medications. Regardless, the results seem to encourage the use of topiramate, bupropion, SGLT‐2 inhibitors, and metformin for their respective indications when weight loss is desired in patients with underlying advanced liver fibrosis.
When only FDA‐approved anti‐obesity medications were evaluated, a statistically significant median weight loss of 7.8% was observed, which is generally consistent with the published efficacy of these medications in patients without cirrhosis. 17 These results seem to extend established efficacy of the FDA‐approved anti‐obesity medications to patients with both obesity and cirrhosis. It's worth noting that since the design of this study, lorcaserin was voluntarily withdrawn from the market following an FDA request in February 2020 due to increased incidence of cancer in the CAMELLIA‐TIMI 61 study. 25 Phentermine use additionally may be limited given increasing prevalence of NAFLD with associated cardiovascular disease and hypertension which may serve as relative or absolute contraindications. Therefore, GLP‐1 agonists continue to have the most appeal in this patient population as they have been shown to have weight‐independent improvements in NASH and liver fibrosis. 26
One of the main limitations of this study was the inability to discriminate between lean mass, fat mass, and fluid losses which is especially relevant in a population at high risk of sarcopenia and development of ascites. Based on patient MELD scores and low rate of hospitalizations; however, it appears likely that a large portion of the patient population did not have significant volume overload. Additionally, the effects of volume overload were mitigated by excluding patients on diuretics from the efficacy analysis. The study also lacked a control group or confirmation of patient adherence to prescribed medications. Finally, there was no prescription data on three FDA‐approved anti‐obesity medications: orlistat, phentermine‐topiramate, and bupropion‐naltrexone. Based on the prescription pattern in this study and given that 84% of patients had comorbid diabetes, the primary indication for the majority of prescriptions was likely diabetes management. Bupropion and naltrexone constituted the only prescriptions seen in patients without diabetes. The strengths include a complete medical record with frequent patient care management by hepatology and obesity medicine specialists.
In summary, this study did not demonstrate any clear contraindications to the use of medications associated with weight loss in patients with obesity and cirrhosis. The weight loss effect of these medications appears to be maintained without apparent worsening of MELD, increased decompensation events, or serious medication intolerance even in the setting of decompensated cirrhosis. Based on the findings above, GLP‐1 agonists should be further investigated as strong candidates for the treatment of obesity or diabetes in patients with excess weight and underlying liver disease. Phentermine similarly appears safe and efficacious as an anti‐obesity medication in this population when no contraindications are present. Finally, medications that might contribute to weight loss, specifically bupropion, topiramate, SGLT‐2 inhibitors, and metformin do not appear detrimental in patients with advanced liver disease and when indicated may deserve preferential use in the setting of excess weight. Given limited treatment options in patients with advanced liver disease and the various liver‐related and nonliver‐related obesity‐driven morbidities, further research is warranted to confirm these findings and help elucidate a potential treatment gap in this vulnerable population.
AUTHOR CONTRIBUTIONS
All authors: study concept and design; Ali Y. Fakhreddine acquisition of data; all authors: analysis and interpretation of data; Ali Y. Fakhreddine and Samantha Bagsic drafting of the manuscript; all authors: critical revision of the manuscript for important intellectual content; Catherine T. Frenette and Ken Fujioka study supervision).
ACKNOWLEDGEMENTS
This research was supported in part by the National Center for Advancing Translational Sciences of the National Institutes of Health under Award number UL1TR002550(PI: Topol). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Waiver of informed consent for authorization to use and/or disclose protected health information for publication per institutional protocol was obtained.
References
REFERENCES
- 1. Welzelz TM , Graubard BI, Zeuzem S, El‐Serag HB, Davila JA, Mcglynn KA. Metabolic syndrome increases the risk of primary liver cancer in the United States: a study in the SEER‐Medicare database. Hepatology. 2011;54(2):463‐471. 10.1002/hep.24397 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Calle EE, Rodriguez C, Walker‐Thurmond K, Thun MJ. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med. 2003;348(17):1625‐1638. 10.1056/NEJMoa021423 [DOI] [PubMed] [Google Scholar]
- 3. Shen Y, Risch H, Lu L, et al. Risk factors for hepatocellular carcinoma (HCC) in the northeast of the United States: results of a case–control study. Cancer Causes Control. 2020;31(4):321‐332. 10.1007/s10552-020-01277-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Larsson SC, Wolk A. Overweight, obesity and risk of liver cancer: a meta‐analysis of cohort studies. Br J Canc. 2007;97(7):1005‐1008. 10.1038/sj.bjc.6603932 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Raynard B, Balian A, Fallik D, et al. Risk factors of fibrosis in alcohol‐induced liver disease. Hepatology. 2002;35(3):635‐638. 10.1053/jhep.2002.31782 [DOI] [PubMed] [Google Scholar]
- 6. Ortiz V, Berenguer M, Rayon JM, Carrasco D, Berenguer J. Contribution of obesity to hepatitis C‐related fibrosis progression. Am J Gastroenterol. 2002;97(9):2408‐2414. 10.1111/j.1572-0241.2002.05995.x [DOI] [PubMed] [Google Scholar]
- 7. Watt KD. Reducing the load: the evolution and management of obesity and nonalcoholic steatohepatitis before liver transplantation. Liver Transpl. 2012;18(S2):S52‐S58. 10.1002/lt.23515 [DOI] [PubMed] [Google Scholar]
- 8. Duffy JP, Kao K, Ko CY, et al. Long‐term patient outcome and quality of life after liver transplantation: analysis of 20‐year survivors. Ann Surg. 2010;252(4):652‐659. 10.1097/SLA.0b013e3181f5f23a [DOI] [PubMed] [Google Scholar]
- 9. Dick AAS, Spitzer AL, Seifert CF, et al. Liver transplantation at the extremes of the body mass index. Liver Transpl. 2009;15(8):968‐977. 10.1002/lt.21785 [DOI] [PubMed] [Google Scholar]
- 10. Martin P, Dimartini A, Feng S, Brown R, Fallon M. Evaluation for liver transplantation in adults: 2013 practice guideline by the American Association for the Study of Liver Diseases and the American society of Transplantation. Hepatology. 2014;59(3):1144‐1165. 10.1002/hep.26972 [DOI] [PubMed] [Google Scholar]
- 11. Miñambres I, Rubio MA, de Hollanda A, et al. Outcomes of bariatric surgery in patients with cirrhosis. Obes Surg. 2019;29(2):585‐592. 10.1007/s11695-018-3562-8 [DOI] [PubMed] [Google Scholar]
- 12. Abdalla Salman M, Maurice Sabri Mikhail H, Nafea MA, et al. Impact of laparoscopic sleeve gastrectomy on fibrosis stage in patients with child‐A NASH‐related cirrhosis. Surg Endosc;1:3. 10.1007/s00464-020-07498-4 [DOI] [PubMed] [Google Scholar]
- 13. Vilar‐Gomez E, Martinez‐Perez Y, Calzadilla‐Bertot L, et al. Weight loss through lifestyle modification significantly reduces features of nonalcoholic steatohepatitis. Gastroenterology. 2015;149(2):367‐378. 10.1053/j.gastro.2015.04.005 [DOI] [PubMed] [Google Scholar]
- 14. Wing RR, Bolin P, Brancati FL, et al. Cardiovascular effects of intensive lifestyle intervention in type 2 diabetes. N Engl J Med. 2013;369(2):145‐154. 10.1056/NEJMoa1212914 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Mosko JD, Nguyen GC. Increased perioperative mortality following bariatric surgery among patients with cirrhosis. Clin Gastroenterol Hepatol. 2011;9(10):897‐901. 10.1016/j.cgh.2011.07.007 [DOI] [PubMed] [Google Scholar]
- 16. Garvey WT, Mechanick JI, Brett EM, et al. American association of clinical endocrinologists and American college of endocrinology comprehensive clinical practice guidelines for medical care of patients with obesity. Endocr Pract. 2016;22(aop):1‐203. 10.4158/EP161365 [DOI] [PubMed] [Google Scholar]
- 17. Yanovski SZ, Yanovski JA. Long‐term drug treatment for obesity: a systematic and clinical review. J Am Med Assoc. 2014;311(1):74‐86. 10.1001/jama.2013.281361 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Sinclair M, Gow PJ, Grossmann M, Angus PW. Review article: sarcopenia in cirrhosis ‐ aetiology, implications and potential therapeutic interventions. Aliment Pharmacol Ther. 2016;43(7):765‐777. 10.1111/apt.13549 [DOI] [PubMed] [Google Scholar]
- 19. Schiavo L, Busetto L, Cesaretti M, Zelber‐Sagi S, Deutsch L, Iannelli A. Nutritional issues in patients with obesity and cirrhosis. World J Gastroenterol. 2018;24(30):3330‐3346. 10.3748/wjg.v24.i30.3330 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Montano‐Loza AJ, Angulo P, Meza‐Junco J, et al. Sarcopenic obesity and myosteatosis are associated with higher mortality in patients with cirrhosis. J Cachexia Sarcopenia Muscle. 2016;7(2):126‐135. 10.1002/jcsm.12039 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Sujanani SM, Elfishawi MM, Zarghamravanbaksh P, Castillo FJC, Reich DM. Dapagliflozin‐induced acute pancreatitis: a case report and review of literature. Case Rep Endocrinol. 2020;2020:6724504. 10.1155/2020/6724504 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Ueda P, Svanström H, Melbye M, et al. Sodium glucose cotransporter 2 inhibitors and risk of serious adverse events: nationwide register based cohort study. BMJ. 2018;363. 10.1136/bmj.k4365 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Tang H, Yang K, Li X, Song Y, Han J. Pancreatic safety of sodium‐glucose cotransporter 2 inhibitors in patients with type 2 diabetes mellitus: a systematic review and meta‐analysis. Pharmacoepidemiol Drug Saf. 2020;29(2):161‐172. 10.1002/pds.4943 [DOI] [PubMed] [Google Scholar]
- 24. Kasichayanula S, Liu X, Zhang W, Pfister M, LaCreta FP, Boulton DW. Influence of hepatic impairment on the pharmacokinetics and safety profile of dapagliflozin: an open‐label, parallel‐group, single‐dose study. Clin Therapeut. 2011;33(11):1798‐1808. 10.1016/j.clinthera.2011.09.011 [DOI] [PubMed] [Google Scholar]
- 25. Bohula EA, Wiviott SD, McGuire DK, et al. Cardiovascular safety of lorcaserin in overweight or obese patients. N Engl J Med. 2018;379(12):1107‐1117. 10.1056/NEJMoa1808721 [DOI] [PubMed] [Google Scholar]
- 26. Armstrong MJ, Gaunt P, Aithal GP, et al. Liraglutide safety and efficacy in patients with non‐alcoholic steatohepatitis (LEAN): a multicentre, double‐blind, randomised, placebo‐controlled phase 2 study. Lancet. 2016;387(10019):679‐690. 10.1016/S0140-6736(15)00803-X [DOI] [PubMed] [Google Scholar]