Ethnic and Gender Differences in the Prevalence of Nonalcoholic Steatohepatitis Among Bariatric Surgery Patients

Abstract

Background & Aims:

Severe obesity is a risk factor for nonalcoholic steatohepatitis (NASH), particularly among ethnic minority groups. The aim of this analysis was to evaluate ethnic group differences in the prevalence of NASH in a large multiethnic sample of bariatric patients not previously well-described in the literature.

Methods:

A cross-sectional retrospective analysis of 611 liver biopsies among patients, who underwent bariatric surgery from 2001 to 2010, from one bariatric surgery practice in South Florida was conducted. Logistic regression analysis predicted the ethnic group-specific odds of having NASH.

Results:

Hispanics (43%) and non-Hispanic whites (NHW) (46%) were significantly more likely than non-Hispanic blacks (NHB) (21%) to have advanced steatosis (p = 0.002). Conversely, NHB were significantly more likely to have mild or moderate steatosis (48%) versus both NHW (34%) and Hispanics (36%). Women were over two and a half times as likely as men to have NASH (OR 2.66, 95% CI 1.68–4.24). NHB were almost twice as likely as NHW to have NASH (OR 1.86, 95% CI 1.17–2.94).

Conclusions:

The prevalence of NASH varies significantly by ethnicity and gender among bariatric surgery patients. These findings have important clinical implications for postoperative health maintenance as NASH carries a risk of progressive liver disease.

Introduction

The prevalence of obesity remains unacceptably high in the United States adult population and certain ethnic minority groups are disproportionately affected. Specifically, the 2011–2012 prevalence estimates of all grades of obesity (body mass index [BMI] ≥30 kg/m²) were 33.4% for non-Hispanic whites (NHW, 33.7% of women), 47.8% for non-Hispanic blacks (NHB, 56.7% of women), and 42.0% for Hispanics (43.3% of women).¹ Moreover, these ethnic group disparities widen as the degree of obesity increases. For example, prevalence estimates for grade 2 obesity (BMI ≥35 kg/m²) were 13.5% for NHW (15.5% for women), 23.4% for NHB (29.3% for women), and 16.0% for Hispanics (19.6% for women). For grade 3 obesity (BMI ≥40 kg/m²), Hispanics and NHW showed similar overall prevalence estimates (5.8% and 5.7%, respectively) compared to NHB (12.2%).¹ If these current national obesity trends continue, by 2030, 86.3% of adults will be overweight and 51.1% will be obese.²

The intersection of obesity and ethnic group disparities resonates to many of the chronic health conditions that are directly attributed to obesity. For example, both the incidence of cardiovascular disease (CVD) and type 2 diabetes (T2DM) and their precursor, the metabolic syndrome, disproportionately affect both NHB and Hispanics versus NHW.^3,4 Ethnic group differences have also been widely reported in nonalcoholic fatty liver disease (NAFLD), a condition widely considered to be the hepatic manifestation of the metabolic syndrome due to its strong association with obesity, insulin resistance, and dyslipidemia.^5,6 NAFLD represents a range of liver injury ranging from simple steatosis that involves a more benign course to nonalcoholic steatohepatitis (NASH), which may progress to advanced fibrosis and cirrhosis.⁵

Previous studies examining ethnic group differences in NAFLD and NASH have reported that Hispanics, but not blacks, are disproportionately affected.^7–14 For example, Younossi et al. reported that relative to Caucasians, the odds of having NASH were 72% higher for Hispanics and 48% lower for black individuals.¹⁵ Another study involving liver biopsies of bariatric surgery patients similarly reported a lower prevalence of NASH in black patients, although no significant differences were observed between Hispanics and non-Hispanics.¹⁶ The Dallas Heart Study reported that the prevalence of steatosis (via magnetic resonance spectroscopy) was highest in Hispanics (45%), followed by NHW (33%) and NHB (24%).⁷ Several other studies have consistently reported that NHB have a lower prevalence of elevated aminotransferase levels compared to NHW and Hispanic, despite having higher rates of obesity, particularly among women.^{8–14,17–19}

Those who have both the metabolic syndrome and NAFLD and NASH are at increased risk for overall mortality, cardiovascular mortality, and liver-related mortality compared with the general population.^20–23 Moreover, NAFLD/NASH-associated liver cirrhosis is currently the second most common indication for liver transplantation in obese patients.²⁴ The preponderance of studies exploring ethnic differences in NASH has relied on samples, in which Mexican Americans largely represent the Hispanic population. The epidemiology of NASH among bariatric surgery patients from more diverse ethnic backgrounds has not been previously well described in the literature. Therefore, the aim of this study was to further characterize the relationship between ethnicity, gender, and NASH among a bariatric surgery patient population, who originated predominantly from Central and South America, as well as the Caribbean Basin countries (e.g., Cuba, the Dominican Republic, and the West Indies).

Methods

A retrospective medical chart analysis was conducted for adults (N = 630) who underwent gastric bypass surgery and who underwent a liver biopsy during surgery in one South Florida practice with one surgeon between 2001 and 2010. The University of Miami Institutional Review Board approved this study.

Demographics (gender, race, and ethnicity), weight, BMI, prevalence of CVD, coronary artery disease (CAD), type 2 diabetes mellitus (T2DM), and preoperative complication information was collected before surgery.

Anthropometrics

Height and weight measurements were conducted during routine clinical visits by practice nursing staff. Weight was measured to the nearest 0.1 pound using a digital scale with the participants wearing light clothing and no shoes. Height was measured to the nearest 0.5 cm using an Accustat Genentech stadiometer. BMI was calculated as body weight in kilograms divided by height in meters squared (kg/m²).

Surgery criteria and procedure

All patients met the National Institute of Health criteria for bariatric surgery.²⁵ Thus, all patients had a BMI >35 kg/m² and at least one existing comorbidity (e.g., elevated blood pressure and hypercholesterolemia) or a BMI >40 kg/m². Patients underwent group and individual education with a practice interdisciplinary group about the potential surgical and nonsurgical options, outcomes, complications, and necessary lifestyle changes. All patients underwent presurgery psychological and nutritional evaluation and testing.

The laparoscopic approach was used in all but five patients.

Liver biopsy and histology procedure

Liver biopsies were done on all patients consecutively; thus, there was no selection bias. Furthermore, patients were excluded from bariatric surgery by standard screening instruments to exclude any patient identified as having significant alcohol abuse in the year before surgery.

Liver biopsies consisted of a single pass taken from the left lobe of the liver using a Tru-Cut needle under direct visualization during surgery. The average biopsy length was 18 ± 6 mm. Biopsy specimens were assessed by an experienced liver pathologist in a private pathology group at the hospital where all surgeries took place. The presence of NASH was determined according to the histological scoring criteria proposed by Kleiner et al., which comprise steatosis grade (percent parenchymal involvement in steatosis), lobular inflammation (number of inflammatory foci per field), hepatocellular ballooning (quantity of balloon cells), and stage of fibrosis (ranging from no fibrosis to cirrhosis).²⁶

Statistical analysis

Chi-squared tests were used to analyze differences in clinical and demographic categorical variables, as well as determine prevalence of each variable within race/ethnic groups. ANOVA was used to compare mean differences between race/ethnicity and continuous variables (BMI and age) using the PROC GLM procedure in SAS Version 9.3 (SAS. Institute, Inc., Cary, NC). A p-value of <0.05 was considered statistically significant. Two sets of logistic regression models were analyzed. In the first set, gender (male or female) and race/ethnicity (NHW, NHB, or Hispanic) were modeled separately as predictors of NASH diagnosis, with males and NHW as the reference groups and adjustment for age, BMI, and either race/ethnicity or gender as appropriate. The second set of models was stratified by race/ethnicity and gender and adjusted for age. NHW was the reference group for the second set. Adjusted odds ratios (AOR) were reported with corresponding 95% confidence intervals with an alpha set to 0.05.

Results

Sample characteristics

Demographic and clinical characteristics of the sample by race/ethnicity are illustrated in Table 1. Over a third (36%, n = 218) of the sample was Hispanic and 18% (n = 111) NHB. Females represented the majority of all categories (73.8%, n = 208 of NHW; 78.0%, n = 170 of Hispanics; 90.1%, n = 100 of NHB). Almost all (98.2%, n = 597) patients underwent gastric bypass surgery.

Table 1.

Demographic and Clinical Characteristics of a Sample of Patients who Underwent Bariatric Surgery from 2001 to 2010 Stratified by Ethnicity

	Total sample N = 611 N (%)	Non-Hispanic white n = 282 (46.2) n (%)	Hispanic n = 218 (35.7) n (%)	Non-Hispanic black n = 111 (18.2) n (%)	p^a
Age, years
Mean, SD	41.0 (12.1)	42.8 (11.8)	40.6 (13.0)	38.4 (10.6)	0.004
Gender
Male	133 (21.8)	74 (23.2)	48 (22.0)	11(9.9)	0.002
Female	478 (78.2)	208 (73.8)	170 (78.0)	100 (90.1)
Surgery type
RYGB	597 (98.2)	277 (98.6)	210 (97.2)	110 (99.1)	0.39
Gastric band	11 (1.8)	4 (1.4)	6 (2.8)	1 (0.9)
BMI, kg/m²
Mean, SD	47.2 (7.6)	47.2 (7.2)	46.4 (7.3)	49.0 (9.2)	0.02
BMI quartiles^b
≤42 kg/m²	196 (32.3)	94 (33.4)	81 (37.7)	21 (18.9)	0.03
42.1 kg/m²–46 kg/m²	140 (23.1)	59 (21.0)	50 (23.3)	31 (27.9)
46.1 kg/m²–51 kg/m²	127 (20.9)	57 (20.3)	42 (19.5)	28 (25.2)
51.1 kg/m²–111 kg/m²	144 (23.7)	71 (25.3)	42 (19.5)	31 (27.9)
Normal biopsy	169 (27.7)	86 (50.9)	61 (36.1)	22 (13.0)	0.002
Advanced steatosis
No	358 (58.6)	152 (53.9)	125 (57.3)	81 (73.0)	0.002
Yes	253 (41.4)	130 (46.1)	93 (42.7)	30 (21.0)
Mild/moderate steatosis
No	383 (62.7)	185 (65.6)	140 (64.2)	58 (52.3)	0.04
Yes	228 (37.3)	97 (34.4)	78 (35.8)	53 (47.6)

Chi-Squared tests used to compare proportions across ethnic groups; ANOVA used to compare means.

Quartiles were determined by body mass index (BMI) of patient sample.

Over 40% (41.4%, n = 253) of the entire sample had advanced steatosis. Hispanics (42.7%, n = 93) and NHW (46%, n = 130) were significantly more likely than NHB (21%, n = 30) to have advanced steatosis (p = 0.002). Conversely, NHB were significantly more likely to have mild or moderate steatosis (47.6%, n = 53) versus both NHW (34.4%, n = 97) and Hispanics (35.8%, n = 78). Significantly, more NHW had a normal/nondisease liver (50.9%, n = 86) than Hispanics (36.1%, n = 61) and NHB (13.0%, n = 13, p = 0.002).

Predictors of NASH and NAFLD

Logistic regression analyses results are illustrated in Table 2. NHB had almost twofold higher odds versus NHW to present with NASH (OR 1.86, 95% CI: 1.17–2.94). No significant differences were found when comparing Hispanics to NHW, but Hispanics had lower odds (AOR: 0.89, 95% CI: 0.60–1.30) versus NHB to present with NASH. Females had over double the odds of presenting with NASH compared to their male counterparts (AOR: 2.66, 95% CI: 1.68–4.24). Among women only, NHB were almost twice as likely as NHW to have NASH, and among men only, NHB were over twice as likely to have NASH versus NHW (Table 3).

Table 2.

Adjusted Odds of Nonalcoholic Steatohepatitis Among a Sample of Patients Who Underwent Bariatric Surgery from 2001 to 2010

	Nonalcoholic steatohepatitis AOR (95% CI)
Ethnicity^a
Non-Hispanic white	Reference
Non-Hispanic black	1.86 (1.17–2.94)
Hispanic	0.89 (0.60–1.30)
Gender^b
Male	Reference
Female	2.66 (1.68–4.24)

Adjusted for age, gender, and BMI.

Adjusted for age, ethnicity, and BMI.

AOR, adjusted odds ratios.

Table 3.

Adjusted Odds of Nonalcoholic Steatohepatitis Among a Sample of Patients Who Underwent Bariatric Surgery from 2001 to 2010 Stratified by Gender and Ethnicity

	Hispanics	Non-Hispanic blacks
	Reference group = Non-Hispanic whites
Nonalcoholic steatohepatitis
Among males	1.19 (0.48–2.98)	2.42 (0.62–9.45)
Among females	0.82 (0.54–1.24)	1.76 (1.10–2.87)

Adjusted for age and BMI.

Discussion

Our results show that advanced steatosis is a highly prevalent (41%) comorbidity among bariatric surgery patients, but varies significantly by ethnicity and gender. NHB bariatric surgery patients were two to two and a half times as likely to have NASH versus their NHW and Hispanic counterparts, and women had more than double the odds of presenting with NASH. Our findings are consistent with some previous literature with respect to elevated risk of NASH among women, but inconsistent with respect to race/ethnicity. Previous studies have indicated that NASH is independently associated with being Hispanic and inversely associated with being NHB, whereas we found a significantly elevated risk of NASH in NHB, but not Hispanic patients.¹⁵ Importantly, these studies diagnosed NASH by imaging and/or biochemical criteria rather than by histology, as we were able to do here. Our results have important clinical implications for postoperative health maintenance, as NASH carries a risk of progressive liver disease.

Although female gender has been associated with more advanced liver disease in some early series, other population-based studies show that men have a higher prevalence of NASH versus women.^27,28 Our biopsy results here showed that women were over two and half times as likely to have NASH versus men. Again, this may be due to the method of diagnosis, or the fact that for patients to be eligible for bariatric surgery, their BMI is already in the severely overweight range and they most likely have at least one related comorbidity (e.g., type 3 diabetes, prediabetes, heart disease); thus, they may be presenting with more advanced-stage liver disease as such.

Several etiological factors to explain these racial and ethnic NASH prevalence differences have been postulated and include a combination or interaction of lifestyle, genetic, and anthropometric/clinical characteristics.²⁷ However, the literature shows that genetic variants only account for a small amount of variation in hepatic steatosis among diverse ethnic groups. Instead, these genes may act as modifiers of the natural history of NAFLD and NASH when they are “turned on” by certain environmental factors, such as dietary intake and physical inactivity. For example, the literature has reported that diet composition alone has significant racial and ethnic differences, and that high carbohydrate consumption is associated with a greater prevalence of NASH, for example.²⁹ Studies consistently report that Hispanics reported higher energy and carbohydrate intakes and a lower fat intake versus NHB and NHW, while NHB reported a higher intake of candies or other sweets.³⁰ Population-based as well as clinical data show that sedentary individuals had a significantly higher prevalence of NAFLD independent of other risk factors.^10,30 NAFLD/NASH gender differences may be partially explained by differences in body fat distribution and sex hormone metabolism, while racial differences in body fat distribution and lipid metabolism may explain the insulin resistance paradox.

With an increasing immigrant population in the United States, it is more crucial than ever to study immigrant and minority health patterns to better understand ethnic group-specific disease patterns. One factor for clinicians beyond bariatric surgeons to consider while identifying those at risk for NAFLD/NASH is the striking increase in obesity occurring between first- and second-generation US immigrants due to acculturation.³¹ Studies have documented rapid acculturation of obesity-related behaviors and related comorbidities, including diet patterns (e.g., increased fast food consumption), smoking, and inactivity in US-born compared to foreign-born immigrants.^32,33 These studies have important implications for our study population in particular. The majority of the studies in the literature that have examined ethnic group differences in postbariatric surgery weight loss and comorbidity resolution have taken place in geographic areas, where NHB and Hispanics were the minority. In our study as well as other studies, we have published that patients were from a region where Hispanics are the major ethnical group, and we consistently find that Hispanics had similar health outcomes to NHW and better outcomes versus NHB.^34–36 While we did not follow these patients postoperatively in this analysis, our preoperative NASH prevalence estimates are consistent with these previous findings. Interestingly, Hispanics do not necessarily have to acculturate to the United States while living in South Florida; in fact, in many parts of the metro area, only Spanish is spoken. These patterns may partially explain why our results, particularly among Hispanics, are not consistent with other studies in the literature.

This study has limitations related to the retrospective design and the study population. With regard to the study population, all subjects were severely obese. Additional studies are needed to determine whether the findings in this study can be generalized to other groups of patients not undergoing bariatric surgery. Obtaining liver biopsies at the end of the operation could have increased the observed prevalence of hepatocellular ballooning, and perhaps the prevalence of the diagnosis of NASH. However, this effect should be equivalent among each ethnic group. Finally, data on comorbidities were not available for subsample of participants, which precluded statistical control for comorbid CVD, CAD, and T2DM in logistic regression models for the sample. While our study provides preliminary data on racial/ethnic and gender differences in NASH prevalence among bariatric surgery patients, future studies must control comorbidities that could impact statistical inferences about the degree of observed disparities.

The study also has several important strengths. The sample size was relatively large and included Hispanics and NHB originating from regions of the world not previously described, providing new insight into the relationship between ethnicity and NAFLD/NASH. Furthermore, histological data were available and were systematically reviewed by an experienced liver pathologist using a validated histological scoring system.

Conclusions

The prevalence of NASH varies significantly by ethnicity and gender among bariatric surgery patients. NHB have a higher prevalence of NASH versus Hispanics and NHW, and women are over two and half times as likely as men to have NASH. These findings have important implications for clinicians as the prevalence and degree of obesity and obesity-related comorbidities rise in the United States across all demographics, and specifically, among minority populations.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Ogden

, Carroll

, Kit

, Flegal

. Prevalence of childhood and adult obesity in the United States, 2011–2012. JAMA, 2014; 311:806–814.

Olshansky

, Passaro

, Hershow

, et al. A potential decline in life expectancy in the United States in the 21st century. N Engl J Med, 2005; 352:1138–1145.

Karter

, Ferrara

, Liu

, Moffet

, Ackerson

, Selby

. Ethnic disparities in diabetic complications in an insured population. JAMA, 2002; 287:2519–2527.

McWilliams

, Meara

, Zaslavsky

, Ayanian

. Differences in control of cardiovascular disease and diabetes by race, ethnicity, and education: U.S. trends from 1999 to 2006 and effects of medicare coverage. Ann Inter Med, 2009; 150:505–515.

Angulo

. Nonalcoholic fatty liver disease. N Engl J Med, 2002; 346:1221–1231.

Marchesini

, Brizi

, Bianchi

, et al. Nonalcoholic fatty liver disease: a feature of the metabolic syndrome. Diabetes, 2001; 50:1844–1850.

Browning

, Szczepaniak

, Dobbins

, et al. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology, 2004; 40:1387–1395.

Clark

, Brancati

, Diehl

. The prevalence and etiology of elevated aminotransferase levels in the United States. Am J Gastroenterol, 2003; 98:960–967.

Ioannou

, Boyko

, Lee

. The prevalence and predictors of elevated serum aminotransferase activity in the United States in 1999–2002. Am J Gastroenterol, 2006; 101:76–82.

10.

Lazo

, Hernaez

, Eberhardt

, et al. Prevalence of nonalcoholic fatty liver disease in the United States: the Third National Health and Nutrition Examination Survey, 1988–1994. Am J Epidemiol, 2013; 178:38–45.

11.

Schneider

, Lazo

, Selvin

, Clark

. Racial differences in nonalcoholic fatty liver disease in the U.S. population. Obesity (Silver Spring, Md.), 2014; 22:292–299.

12.

Smits

, Ioannou

, Boyko

, Utzschneider

. Non-alcoholic fatty liver disease as an independent manifestation of the metabolic syndrome: results of a US national survey in three ethnic groups. J Gastroenterol Hepatol, 2013; 28:664–670.

13.

Weston

, Leyden

, Murphy

, et al. Racial and ethnic distribution of nonalcoholic fatty liver in persons with newly diagnosed chronic liver disease. Hepatology (Baltimore, Md.), 2005; 41:372–379.

14.

Williams

, Stengel

, Asike

, et al. Prevalence of nonalcoholic fatty liver disease and nonalcoholic steatohepatitis among a largely middle-aged population utilizing ultrasound and liver biopsy: a prospective study. Gastroenterology, 2011; 140:124–131.

15.

Younossi

, Stepanova

, Negro

, et al. Nonalcoholic fatty liver disease in lean individuals in the United States. Medicine, 2012; 91:319–327.

16.

Kallwitz

, Guzman

, TenCate

, et al. The histologic spectrum of liver disease in African-American, non-Hispanic white, and Hispanic obesity surgery patients. Am J Gastroenterol, 2009; 104:64–69.

17.

Ruhl

, Everhart

. Determinants of the association of overweight with elevated serum alanine aminotransferase activity in the United States. Gastroenterology, 2003; 124:71–79.

18.

Schwimmer

, McGreal

, Deutsch

, Finegold

, Lavine

. Influence of gender, race, and ethnicity on suspected fatty liver in obese adolescents. Pediatrics, 2005; 115:e561–e565.

19.

Solga

, Clark

, Alkhuraishi

, et al. Race and comorbid factors predict nonalcoholic fatty liver disease histopathology in severely obese patients. Surg Obes Relat Dis, 2005; 1:6–11.

20.

Lazo

, Clark

. The epidemiology of nonalcoholic fatty liver disease: a global perspective. Semin Liver Dis, 2008; 28:339–350.

21.

Dunn

, Xu

, Wingard

, et al. Suspected nonalcoholic fatty liver disease and mortality risk in a population-based cohort study. Am J Gastroenterol, 2008; 103:2263–2271.

22.

Ong

, Pitts

, Younossi

. Increased overall mortality and liver-related mortality in non-alcoholic fatty liver disease. J Hepatol, 2008; 49:608–612.

23.

Younossi

, Otgonsuren

, Venkatesan

, Mishra

. In patients with non-alcoholic fatty liver disease, metabolically abnormal individuals are at a higher risk for mortality while metabolically normal individuals are not. Metabolism, 2013; 62:352–360.

24.

Nair

, Mason

, Eason

, Loss

, Perrillo

. Is obesity an independent risk factor for hepatocellular carcinoma in cirrhosis?. Hepatology, 2002; 36:150–155.

25.

Kleiner

, Brunt

, Van Natta

, Behling

, Contos

, Cummings

, et al. Nonalcoholic Steatohepatitis Clinical Research Network. Gastrointestinal surgery for severe obesity. NIH consensus development conference, March 25–27,1991. Nutrition, 1996; 12:397–404.

26.

Kleiner

, Brunt

, Van Natta

, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology, 2005; 41:1313–1321.

27.

Pan

, Fallon

. Gender and racial differences in nonalcoholic fatty liver disease. World J Hepatol, 2014; 6:274–283.

28.

Vernon

, Baranova

, Younossi

. Systematic review: the epidemiology and natural history of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis in adults. Aliment Pharmacol Ther, 2011; 34:274–285.

29.

Bambha

, Belt

, Abraham

, et al. Ethnicity and nonalcoholic fatty liver disease. Hepatology, 2012; 55:769–780.

30.

Krasnoff

, Painter

, Wallace

, Bass

, Merriman

. Health-related fitness and physical activity in patients with nonalcoholic fatty liver disease. Hepatology, 2008; 47:1158–1166.

31.

Gordon-Larsen

, Harris

, Ward

, Popkin

. Acculturation and overweight-related behaviors among Hispanic immigrants to the US: the National Longitudinal Study of Adolescent Health. Soc Sci Med, 2003; 57:2023–2034.

32.

Popkin

, Udry

. Adolescent obesity increases significantly in second and third generation U.S. immigrants: the National Longitudinal Study of Adolescent Health. J Nutr, 1998; 128:701–706.

33.

Winkleby

, Albright

, Howard-Pitney

, Lin

, Fortmann

. Hispanic/white differences in dietary fat intake among low educated adults and children. Prev Med, 1994; 23:465–473.

34.

de la Cruz-Munoz

, Messiah

, Arheart

, Lopez-Mitnik

, Lipshultz

, Livingstone

. Bariatric surgery significantly decreases the prevalence of type 2 diabetes mellitus and pre-diabetes among morbidly obese multiethnic adults: long-term results. J Am Coll Surg, 2011; 212:505–511; discussion 512–503.

35.

Khorgami

, Arheart

, Zhang

, Messiah

, de la Cruz-Munoz

. Effect of ethnicity on weight loss after bariatric surgery. Obes Surg, 2015; 25:769–776.

36.

Khorgami

, Zhang

, Messiah

, de la Cruz-Munoz

. Predictors of postoperative aftercare attrition among gastric bypass patients. Bariatr Surg Pract Patient Care, 2015; 10:79–83.