Racial Differences After Bariatric Surgery: 24-Month Follow-up of a Randomized Controlled Trial for Post-operative Loss-of-Control Eating

Abstract

Background:

This prospective study examined and extended initial short-term findings regarding racial differences in post-bariatric surgery loss-of-control (LOC)-eating and weight-loss to longer-term outcomes through 24-month follow-ups.

Setting:

Academic Medical Center in the United States

Methods:

Participants were 140 patients (46.4% non-white) in a 3-month randomized controlled trial for LOC-eating performed about six months post-bariatric surgery. Participants were reassessed at 6-, 12-, 18-, and 24-months after treatment ended (about 33-months post-surgery). Doctoral assessors administered the Eating Disorder Examination-Bariatric Surgery Version interview to assess LOC-eating and eating-disorder psychopathology at 12- and 24-month follow-ups. The Beck Depression Inventory (BDI-II) was repeated and measured weight was obtained at all follow-ups.

Results:

White patients had significantly greater percent excess weight loss at all follow-ups than non-White patients (ps<.03). White patients reported significantly greater LOC-eating at 12- (p=.004) and 24-month (p=.02) follow-ups, and significantly greater eating-disorder psychopathology at 12-month follow-up (p<.03). Racial groups did not differ significantly in eating-disorder psychopathology at 24-month follow-up, or in BDI-II depression scores at any follow-ups.

Conclusions:

Our findings suggest that, among patients with LOC-eating post-bariatric surgery, non-White patients attain less percent excess weight loss than White patients but have comparable or better outcomes in LOC-eating, associated eating-disorder psychopathology, and depression over time.

Introduction

Bariatric surgery, the most effective treatment for severe obesity and medical comorbidities, results in substantial heterogeneity in outcomes ⁽¹⁾ that is further compounded by racial disparities ⁽²⁾. Emerging research suggests that bariatric surgery outcomes vary by race/ethnicity ^(2,3,4). For instance, Black patients have higher odds of serious complications, including death, and less weight loss after bariatric surgery than White patients ⁽⁵⁾. Little is known, however, about racial differences in psychosocial outcomes after bariatric surgery, including loss-of-control (LOC)-eating, the core feature of binge eating ⁽⁶⁾, and a consistent post-surgical predictor of poorer longer-term outcomes ^(7,8).

In a randomized controlled trial (RCT) delivering treatments for LOC-eating about six months post-surgery, Black and White patients did not differ significantly at the start of treatment in LOC-eating frequency, post-surgical onset of LOC-eating, eating-disorder psychopathology, depressive symptoms, or presurgical BMI, but Black patients had significantly less percent total weight loss (TWL) and excess weight loss (EWL) than White patients ⁽⁹⁾. Outcomes data from this RCT study performed after bariatric surgery were used for the present study to examine longer-term differences in outcomes by race. The study design and primary treatment outcomes previously reported ^(10,11) are briefly described here for context.

The original RCT evaluated guided-self-help treatments (cognitive-behavioral therapy and behavioral weight loss) and a control delivered approximately six-months postoperatively for reducing LOC-eating ^(10,11). Grilo and colleagues reported acute posttreatment ⁽¹⁰⁾ and their longer-term effects through 24-month follow-ups ⁽¹¹⁾ after having completed treatments (i.e., 27 months after baseline and approximately 33 months after bariatric surgery). Overall, significant reductions in LOC-eating frequency and weight losses were observed that did not differ significantly between the treatments at posttreatment ⁽¹⁰⁾. From posttreatment to the 24-month follow-up, no significant changes in LOC-eating frequency or secondary outcomes (eating-disorder psychopathology or depressive symptoms) were observed and weight increased substantially with no significant differences between treatment groups ⁽¹¹⁾. Although race neither predicted nor moderated treatment outcomes, analyses of treatment endpoints for categorical variables revealed non-White participants had higher rates attaining abstinence from LOC-eating but were less likely to achieve 5% weight loss than White participants ⁽¹⁰⁾.

The present study aimed to extend our initial short-term findings regarding racial differences in LOC-eating and weight-loss outcomes to longer-term outcomes through 24-months of prospective follow-up. Given the lack of treatment group differences on acute and longer-term outcomes in the main RCT, the data from the overall participant group with all conditions combined were used for the present study.

Methods

Participants

Methods were previously described in detail ^(10,11). Participants were 140 randomized patients recruited approximately six months following laparoscopic Roux-en-Y gastric bypass (RYGB) or sleeve gastrectomy (LSG) from a single bariatric center-of-excellence at an academic medical center. Eligibility criteria included age 18-65 years and recurrent (≥once weekly/past 28 days) LOC-eating, defined as feeling unable to stop/control an eating episode regardless of quantity. Exclusion criteria included taking medications known to effectively influence eating/weight, current substance dependence or severe psychiatric illness requiring acute care. All assessments were conducted independently from the bariatric center and team. The study received Institutional Review Board approval and all participants provided written informed consent.

At study baseline for this RCT, mean time since surgery was 6.4 (SD=1.5) months, %TWL was 20.3 (SD=7.9), and age was 45.6 (SD=10.9) years. 85.0% (n=119) were women, and racial ethnic composition was diverse. N=15 (10.7%) identified as Hispanic/Latino/a. For race, n=79 (56.4%) were White, n=44 (31.4%) Black, n=11 (7.9%) “Other”, n=3 (2.1%) Bi/Multiracial, n=2 (1.4%) American-Indian/Alaska-Native, and n=1 (0.7%) Native-Hawaiian/Other Pacific-Islander]. Most (n=123; 87.9%) had LSG and 17 (12.1%) had RYGB.

Diagnostic and Repeated Outcomes Measures

Eating Disorder Examination-Bariatric Surgery Version (EDE-BSV).

Doctoral assessors administered this interview to assess LOC-eating and eating-disorder psychopathology (reflected in Global score) at baseline, posttreatment, and 12- and 24-month follow-ups. The EDE-BSV was adapted for bariatric research and used in the Longitudinal Assessment Bariatric Surgery (LABS) study ^(12,13).

Weight/Height Variables.

Measured height and measured weight values were used to calculate BMI at all timepoints and to calculate percent total weight loss (%TWL) during treatment for posttreatment and after treatment for each of the follow-ups (from posttreatment to each follow-up). %TWL and %EWL from pre-surgery was also calculated.

Beck Depression Inventory (BDI-II) ⁽¹⁴⁾,

a well-established self-report measure of depression symptoms/levels, was completed at baseline, posttreatment, and 6-, 12-, 18-, and 24-month follow-ups.

Statistical Analyses

Racial group differences were examined in a series of univariate analyses of variance for continuous measures and chi-square tests for categorical variables. Group differences were also examined in a series of ANCOVA’s adjusting for %TWL from surgery to time of study enrollment about six-months post-surgery due to the variability in %TWL at study enrollment.

LOC-eating frequency, eating-disorder psychopathology (EDE-Global score), depression level (BDI-II total score), and weight variables (BMI, %TWL, and %EWL) were analyzed continuously. LOC-eating frequency was transformed prior to analysis as it did not conform to normality. LOC-eating was also analyzed categorically as proportion achieving abstinence from LOC-eating (zero episodes during the past 28 days) assessed by the EDE-BSV at posttreatment and at 12- and 24-month follow-ups). For partial eta-squared, an estimate of effect size, values are considered small at .01, medium at .06, and large at .14 ⁽¹⁵⁾.

Results

Study groups were comparable in size with 53.6% (n=75) classified as White (non-Hispanic) and 46.4% (n=65) classified as non-White. Study treatment assignment did not differ significantly by race (χ²=0.96, p=.620) and retention did not differ significantly by race for any follow-ups including posttreatment (p=.534), 6-month (p=.443), 12-month (p=.846), 18-month (p=.135), or the 24-month (p=.873) follow-ups. Racial groups differed significantly in age; White participants were significantly older (M=47.8, SD=10.6) than non-White participants (M=43.1, SD=10.9), F=6.88, p=.010; and gender with a greater proportion of women in the non-White category (n=62, 95.4% women) than the White category (n=57, 76.0%) χ²=10.26, p=.001.

Table 1 summarizes the clinical characteristics of White and non-White participants across follow-ups. BMI differed significantly by race at the 6-month (p= .030) and the 18-month (p=.041) follow-ups with small effects, but not at the 12-month (p=.052) or the 24-month (p=.060) follow-ups. Relative to non-White participants, White participants had significantly greater %EWL at each follow-up including the 6-, 12-, 18-, and 24-month follow-ups (all p-values less than .03) with small to medium effects. Figure 1 illustrates the heterogeneity in weight losses by race (bar graph) and a complementary schematic depicting the shape of the curve of the individual variability in %EWL from surgery to the study 24-month follow-up (about 33 months post-surgery). A similar pattern emerged for %TWL since surgery with small effects, except racial group differences in %TWL became nonsignificant at the 24-month follow-up (p=.069). The groups did not differ significantly in %TWL during the study at any follow-up (p-values >.15).

Table 1.

Clinical Variables by Racial Group at Post-Treatment and Follow-Up Assessments Through 24 Months After Loss-of-Control Eating Treatments Post-Bariatric Surgery

		Non-White n=65 M (SD)	White n=75 M (SD)	Test Statistic		Effect Size
				F	p-value
BMI
	Post-Treatment	37.47 (7.20)	34.78 (6.92)	4.35	.039	.04
	6-month F/U	37.55 (7.47)	34.70 (6.55)	4.83	.030	.04
	12-month F/U	37.89 (7.93)	35.16 (6.56)	3.85	.052	.03
	18-month F/U	39.50 (8.44)	36.35 (7.16)	4.28	.041	.04
	24-month F/U	39.72 (8.27)	36.90 (7.17)	3.61	.060	.03
%EWL from surgery
	Post-Treatment	−43.90 (20.31)	−55.81 (22.26)	9.30	.003	.07
	6-month F/U	−42.12 (23.99)	−53.63 (23.11)	6.97	.009	.06
	12-month F/U	−40.26 (22.30)	−50.48 (23.56)	5.39	.022	.05
	18-month F/U	−34.52 (22.75)	−45.64 (22.74)	6.34	.013	.06
	24-month F/U	−31.90 (25.55)	−42.05 (22.04)	4.95	.028	.04
%TWL from surgerya
	Post-Treatment	−19.81 (9.25)	−24.31 (9.79)	6.68	.011	.05
	6-month F/U	−19.10 (10.84)	−23.23 (10.39)	4.44	.037	.04
	12-month F/U	−17.89 (10.32)	−22.04 (10.94)	4.14	.044	.04
	18-month F/U	−15.57 (10.51)	−20.08 (10.82)	4.74	.032	.04
	24-month F/U	−14.52 (11.28)	−18.30 (10.24)	3.37	.069	.03
%TWL during studyb
	6-month F/U	1.16 (3.89)	0.81 (4.92)	0.18	.673	<.01
	12-month F/U	3.72 (5.84)	3.02 (7.23)	0.30	.585	<.01
	18-month F/U	6.22 (6.79)	5.83 (10.01)	0.05	.818	<.01
	24-month F/U	8.15 (9.00)	7.87 (10.50)	0.02	.885	<.01
EDE LOC-eating frequency (past month)†
	Post-Treatment	4.62 (8.79)	7.62 (11.20)	7.63	.007	.06
	12-month F/U	4.80 (13.17)	9.56 (14.86)	8.41	.004	.07
	24-month F/U	3.76 (6.63)	9.35 (16.44)	5.26	.024	.04
EDE Global Score
	Post-Treatment	1.43 (0.90)	1.85 (1.04)	5.57	.020	.04
	12-month F/U	1.29 (0.86)	1.68 (1.01)	4.96	.028	.04
	24-month F/U	1.22 (0.91)	1.56 (1.02)	3.53	.063	.03
BDI-II Score
	Post-Treatment	8.42 (9.19)	8.55 (10.14)	0.01	.941	<.01
	6-month F/U	8.34 (7.76)	9.05 (9.59)	0.19	.661	<.01
	12-month F/U	8.69 (9.26)	10.75 (10.95)	1.19	.277	.01
	18-month F/U	9.75 (9.24)	10.77 (11.06)	0.28	.598	<.01
	24-month F/U	9.26 (9.39)	11.67 (13.04)	1.23	.270	.01

Note. N=140. BMI=body mass index; LOC=loss-of-control eating; EDE Global=eating disorder examination (bariatric surgery version) global score; BDI-II=beck depression inventory second edition.

^a%TWL since surgery represents percent total weight loss from bariatric surgery to each assessment in the study including pre-surgery to study enrollment/pre-treatment (about six months post-surgery), pre-surgery to study post (about nine-months post-surgery), pre-surgery to the study six-month follow-up (about 15 months post-surgery), pre-surgery to the study 12-month follow-up (about 21 months post-surgery), pre-surgery to the study 18-month follow-up (about 27 months post-surgery), and pre-surgery to the 24-month follow-up (about 33 months post-surgery).

^b%TWL during study at post represents percent total weight loss from study enrollment (about six months-post-surgery) to study post (about nine-months post-surgery); %TWL during study at follow-up represents percent total weight from study post-treatment to each follow-up in the study.

For baseline data, refer to Grilo CM, Ivezaj V, Duffy AJ, Gueorguieva R. Randomized Controlled Trial of Treatments for Loss-of-Control Eating Following Bariatric Surgery. Obesity (Silver Spring) 2021;29: 689-697.

^†LOC-variable was log-transformed for analyses; however raw means and standard deviations are reported for context.

Figure 1. Individual variability in percent excess weight loss by race from pre-surgery to the study 24-month follow-up (about 33 months post-surgery).

White participants reported significantly greater frequency of LOC-eating episodes at the 12-month (p=.004) and 24-month (p=.024) follow-ups with small to medium effects. Figure 2 summarizes LOC-eating abstinence rates by race at each follow-up. Abstinence rates were significantly lower for White (24%) than non-White (46.2%) participants at the 12-month follow-up (χ²(1)=7.59,p=.006; phi=0.23), and at the 24-month follow-up (White: 29.3% vs non-White 46.2%),χ²(1)=4.22,p=.040; phi=.17. Significant group differences in EDE Global scores were observed with significantly higher scores for White than non-White participants at the 12-month follow-up (p=.028) with small effects, but not at the 24-month follow-up (p=.063). Finally, racial groups did not differ significantly for BDI-II scores (p-values >.270).

Figure 2. Percentage with LOC-eating abstinence by race at posttreatment, 12-month, and 24-month follow-up assessments.

*p<.05; **p<.01; ***p<.001

Discussion

Our findings suggest that, among patients with LOC-eating post-bariatric surgery, non-White patients attain less weight loss than White patients but have comparable or better outcomes in psychosocial functioning over time. Despite similar frequency of post-surgical LOC-eating episodes at study baseline, non-White participants reported fewer LOC-eating episodes, both continuously and categorically, than their non-White counterparts over time, with no significant differences in eating-disorder psychopathology or depression levels by the 24-month follow-up. Furthermore, weight loss during the post-surgical treatment study did not differ significantly between racial groups. The magnitude of racial differences in weight loss since surgery was not large but did attain statistical significance and was evident when examining individual variability between racial groups.

These findings of racial differences in bariatric surgery weight outcomes echo the extant bariatric literature ^(2,5), and parallel findings from the binge-eating disorder ⁽¹⁶⁾ literature, where Black patients show better binge-eating outcomes but are less likely to attain comparable weight loss than their White counterparts. Differences in weight loss outcomes by race may be due, in part, to other differences at the individual and/or structural levels. For instance, despite comparable levels of loss-of-control eating and eating-disorder psychopathology after bariatric surgery, Black patients were more likely to engage in other maladaptive eating patterns such as night eating and skipping meals including breakfast and dinner relative to their White counterparts ⁽⁹⁾. These reported differences in eating behavior may account for some differences in weight outcomes. In terms of structural level differences, the food environment, neighborhoods, and discrimination that disproportionally influence certain ethnic and racial groups may influence long-term weight outcomes ⁽¹⁷⁾. Overall, however, our findings add to the growing literature highlighting the benefits of bariatric surgery for minority groups, despite lower utilization of this treatment among such groups ^(18,19).

Strengths of the present study include the relatively diverse composition (nearly 50% did not identify as White), rigorous assessment methods, and good retention (83%) at 24-months of study follow-up (about 33-months post-surgery). Limitations include recruitment of a specific participant group with regular LOC-eating about six-months post-bariatric surgery. Thus, generalizability to individuals who do not have LOC-eating and who are not interested in research is limited. Another important limitation is the small sample size of various ethnic/racial groups. Due to the small number of participants within each racial/ethnic category (e.g., number of Latinx participants), participants from different ethnic/racial backgrounds were collapsed into a single “non-White” category. Future research with larger sample sizes of greater ethnic/racial variation are needed to understand better group differences in outcomes culturally-relevant treatment needs.

Conclusion

Despite differences in weight loss after bariatric surgery, non-White patients with LOC-eating after bariatric surgery appear to have comparable or better sustained outcomes following a brief 12-week treatment trial in psychosocial functioning including LOC-eating, eating-disorder psychopathology, and depression levels long-term.

Highlights: Study Importance Questions.

What is already known about this subject?

• Some bariatric surgery outcomes vary by race/ethnicity.

• Less is known about loss-of-control (LOC) eating and psychosocial outcomes by race/ethnicity.

What are the new findings in your manuscript?

• Among patients with LOC-eating post-bariatric surgery, non-White patients had less weight loss than White patients.

• Relative to White patients, non-White patients had comparable or better LOC-eating and psychosocial outcomes over time.

How might your results change the direction of research or the focus of clinical practice?

• Better understanding of variability in weight outcomes among racial/ethnic groups is needed.

• Barriers to bariatric surgery utilization among racial/ethnic minority groups should be examined given relatively comparable outcomes.

Data Sharing Statement:

De-identified data will be provided in response to reasonable written request to achieve goals in an approved written proposal.