Back to Annual Meeting Program

Background: In the US more blacks than whites are severely obese (26% vs. 15%) and suffer from hypertension (40% vs. 27%). Prior studies examining the influence of race on bariatric surgery have been from single-institution or small cohorts. This is the first study to examine disparities in national patterns of utilization and the influence of ethnicity on outcomes after Roux-en-Y gastric bypass surgery (RYGB) from the large multi-institutional prospective database for the American Society for Bariatric and Metabolic Surgery.

Methods: All research-consented white, black or Hispanic patients undergoing RYGB between 6/2007 and 10/2011 and eligible for one year of follow-up were included. Other races were excluded. Descriptive statistics were used for demographic information. Multivariate logistic and normal regression models examined relationships between race and outcomes, controlling for age, gender, baseline BMI and comorbid conditions. Races were compared using a t-test for continuous variables and Pearson chi-square test for categorical variables. Reported p-values were adjusted for the false discovery rate (FDR) to control for multiple testing.

Results: The racial distribution of the 135,262 study patients was 79% white, 12% black, and 9% Hispanic. Among the blacks undergoing RYGB only 15% were male whereas 22% of the white and Hispanic patients were men. Compared to whites, blacks were younger (42.8 ± 10.6 vs. 46.3 ± 11.6 yrs.), heavier BMI (50.2 ± 9.2 vs. 47.6 ± 8.0 kg/m² and more often hypertensive (58% vs. 53%) at baseline. Although mortality rates within 30 days were equivalent for all races (0.23 - 0.26%), serious adverse events were higher for blacks (3.65%) versus whites (3.19%) and Hispanics (2.01%). At 1 year, mean BMI decreased markedly to 35.0 ± 7.5 for blacks, 31.6 ± 6.73 for whites and 32.6 ± 7.0 kg/m² for Hispanics. However, the percentage decrease in BMI from baseline was lower for blacks (-30%) compared to whites (-34%) and Hispanics (-32%). Similarly, hypertension decreased from 57% to 37% (blacks), 53% to 27% (whites) and 42% to 29% (Hispanics) but, the percentage decline was less for blacks (-35%) versus whites (-49%) and Hispanics (-50%). Resolution of diabetes also demonstrated a similar pattern for blacks (59%) versus whites (65%) and Hispanics (61%). Racial differences in outcomes for weight loss and major comorbid conditions persisted after adjustment for baseline characteristics (p values and odds ratios are in Table1).

Conclusions: Race exerts a significant influence on outcomes after RYGB. Despite lower efficacy in blacks, overall benefits from RYGB were significant. Given the higher prevalence of obesity in blacks, bariatric surgery is underutilized by this group (particularly males). Higher baseline BMI and more frequent hypertension in blacks indicate need for earlier surgical intervention.
Effect of ethnicity on outcomes at 1 year

	Black vs. Hispanic		Black vs. White		Hispanic vs. White
Outcomes	P-value	PPM (95% CI)	P-value	PPM (95% CI	P-value	PPM (95% CI
BMI 1	<.0001	1.23 (1.04, 1.42)	<.0001	1.77 (1.65, 1.89)	<.0001	0.54 (0.38, 0.71)
Excess Body Weight (kg)	<.0001	2.00 (1.59, 2.42)	0.0155	0.50 (0.14, 0.87)	<.0001	-1.50 (-1.73, -1.28)
Outcomes	P-value	OR (95% CI)	P-value	OR (95% CI)	P-value	OR (95% CI)
GERD	0.7875	1.05 (0.91, 1.20)	0.0095	0.87 (0.79, 0.96)	0.0056	0.83 (0.74, 0.93)
Diabetes	0.8914	0.98 (0.82, 1.17)	0.0258	1.15 (1.03, 1.28)	0.0661	1.17 (1.02, 1.35)
Hypertension	<.0001	1.66 (1.44, 1.90)	<.0001	1.69 (1.57, 1.83)	0.7339	1.02 (0.90, 1.16)
Obstructive Sleep Apnea Syndrome	0.8914	1.02 (0.86, 1.21)	0.9332	.00 (0.89, 1.11)	0.7339	0.97 (0.84, 1.13)

Outcomes are fit with a generalized linear model controlling, sex, age, current tobacco use, prior medical history, and current BMI unless noted. Reported p-values were adjusted with FDR. ¹ The covariate BMI was replaced with baseline BMI PPM: Predicted population marginal mean difference.