Impact of Diabetes on Weight Loss Outcomes After Bariatric Surgery

Experience From 5-Year Follow-Up of Michigan Bariatric Surgery Cohort

Yingying Luo; Raad A. Haddad; Mehmet Selman Ontan; Abdel Wahab Jalal Eldin; Mohammed Abu-Rumaileh; Matheos Yosef; Shokoufeh Khalatbari; Oliver Varban; Andrew Kraftson; Nazanene H. Esfandiari; Elif A. Oral

Disclosures

Clin Endocrinol. 2023;99(3):285-295.

Abstract and Introduction

Abstract

Objective: For patients with obesity and diabetes, bariatric surgery can lead to the remission of both diseases. However, the possible impact of diabetes on the magnitude of weight loss outcomes after bariatric surgery has not been precisely quantified.

Research Design and Methods: Data from Michigan Bariatric Surgery Cohort (MI-BASiC) was extracted to examine the effect of baseline diabetes on weight loss outcomes. Consecutive patients older than 18 years of age undergoing gastric bypass (GB) or sleeve gastrectomy (SG) for obesity at University of Michigan between January 2008 and November 2013 were included. Repeated measures analysis was used to determine if diabetes was a predictor of weight loss outcomes over 5 years postsurgery.

Results: Out of the 714 included patients, 380 patients underwent GB [mean BMI 47.3 ± 0.4 kg/m², diabetes 149 (39.2%)] and 334 SG [mean BMI 49.9 ± 0.5 kg/m², diabetes 108 (32.3%)]. Multivariable repeated measures analysis showed, after adjusting for covariates, that individuals with diabetes had a significantly lower percentage of total (p = .0023) and excess weight loss (p = .0212) compared to individuals without diabetes.

Conclusions: Our data demonstrate that patients with diabetes undergoing bariatric surgery would experience less weight loss than patients without diabetes.

Introduction

Obesity and its related consequences have become a major healthcare burden in recent decades. The prevalence of obesity was 39.8% among US adults in 2015–2016, with a speculated prevalence of 51% of the US population by 2030.^[1–3] Obesity-related comorbidities are well recognised and impose a heavy economic burden.^[4] It is known that obese individuals have a greater risk of metabolic diseases and increased all-cause mortality.^[5,6] Thus, prevention and treatment of obesity have become essential healthcare priorities.

A comprehensive treatment approach for obesity can lead to remission or prevention of progression and its associated consequences. Among existing approaches, bariatric surgery has been shown to achieve the most favourable outcomes of weight loss, prevention, and remission of comorbidities.^[7–9]

We previously published our data on outcomes comparing the two surgery procedures from our cohort, MI-BASiC (Michigan Bariatric Surgery Cohort), at years 2 and 4.^[10,11] Similar to other studies, comparison between gastric bypass (GB) and sleeve gastrectomy (SG) has shown more favourable outcomes for GB in terms of weight loss and achieving remission rates of diabetes, hypertension, and dyslipidemia.^[11–14] However, we found that weight loss outcomes could vary substantially among individuals and sought to understand the impact of baseline comorbidities on the magnitude of postoperative weight loss. While a few papers have supported the negative impact of baseline diabetes on ultimate weight loss,^[15–17] we specifically explored to quantify the effect of diabetes on weight loss outcomes after bariatric surgery on a larger study group. We believe that this study will offer new insight into the relationship between diabetes, obesity, and bariatric surgery.

1 2 3 4 5

Next Section

Table 1. Summary for definitions of some of the used measures.
Measure	Description
TWL	Presurgery weight − weight at any point
TWL%	TWL divided by presurgery weight
EW	Initial weight − ideal weight (BMI at 25 kg/m²)
EWL%	Total weight loss divided by excess weight
WR	Last recorded weight − minimum recorded weight
WR%	(Last recorded weight − minimum recorded weight) × 100/(preoperative weight − minimum recorded weight)
Primary nonresponse	EWL% never reaching 50%
Secondary nonresponse	Having significant WR after reaching an EWL% of 50% or more
Successful weight loss	Reaching an EWL% of 50% or more and avoiding secondary nonresponse during the 5-year follow-up period

Abbreviations: EW, excess weight; EWL%, excess weight loss percent; TWL%, total weight loss percent; TWL, total weight loss; WR%, weight regain percent; WR, weight regain.

Table 2. Repeated measure univariable analysis ^a of achieving BMI < 30 kg/m ² or achieving 50% or more excess weight loss.
Parameter	Achieving either BMI < 30kg/m² or 50% or more excess weight loss			Achieving BMI < 30 kg/m²			Achieving 50% or more excess weight loss
Parameter	OR	95% CI	p Value	OR	95% CI	p Value	OR	95% CI	p Value
Age (years)	0.976	0.964, 0.989	.0002	0.993	0.979, 1.007	.3009	0.976	0.964, 0.989	.0002
Sex (male vs. female)	0.842	0.600, 1.180	.3178	0.567	0.373, 0.862	.0079	0.846	0.603, 1.186	.3306
Baseline DM (no vs. yes)	1.527	1.147, 2.033	.0038	1.458	1.049, 2.028	.0249	1.531	1.151, 2.037	.0035
Baseline HTN (no vs. yes)	1.473	1.116, 1.946	.0064	1.631	1.195, 2.227	.0021	1.475	1.116, 1.946	.0063
Baseline dyslipidemia (no vs. yes)	1.104	0.819, 1.488	.5160	1.205	0.853, 1.698	.2897	1.111	0.825, 1.499	.4864
Baseline weight (kg)	0.989	0.984, 0.994	<.0001	0.956	0.949, 0.964	<.0001	0.990	0.984, 0.995	<.0001
Baseline SBP (mmHg)	0.996	0.989, 1.004	.3124	0.990	0.981, 0.998	.0189	0.997	0.989, 1.004	.3574
Baseline HbA_1c (%)	0.844	0.740, 0.962	.0113	0.799	0.672, 0.950	.0114	0.843	0.739, 0.961	.0108
Baseline Hb (g/L)	1.051	0.935, 1.181	.4063	1.011	0.887, 1.152	.8716	1.052	0.936, 1.182	.3971
Baseline ALT (>35 vs. <35 U/L)	0.932	0.695, 1.251	.6405	0.819	0.585, 1.145	.2419	0.938	0.699, 1.258	.6695
Baseline TG (mg/dL)	0.999	0.997, 1.001	.3742	0.999	0.997, 1.002	.5542	0.999	0.997, 1.001	.3585

Abbreviations: ALT, alanine aminotransferase; CI, confidence interval; DM, diabetes; Hb, haemoglobin; HbA1c, haemoglobin A1c; HTN, hypertension; OR, odds ratio; SBP, systolic blood pressure; TG, triglyceride.
^aFollow-up time and surgery type were adjusted.

Table 3. Summary of repeated measure univariable and multivariable linear model for predicting percentage of total weight loss: Impact of covariates.
	Univariable analysis			Multivariable analysis
	Beta	95% CI	p Value	Beta	95% CI	p Value
Baseline DM (yes vs. no)	−3.174	−4.423, −1.925	<.0001	−2.008	−3.295, −0.720	.0023
Baseline HTN (yes vs. no)	−2.325	−3.543, −1.107	.0002
Baseline dyslipidemia (yes vs. no)	−2.435	−3.759, −1.110	.0003
Baseline weight (kg)	0.035	0.012, 0.057	.0024	0.026	0.000, 0.051	.0469
Baseline SBP (mmHg)	0.030	−0.003, 0.063	.0782	0.037	0.004, 0.069	.0264
Baseline TG (mg/dL)	−0.004	−0.014, 0.005	.3761
Baseline Hb (g/L)	−0.024	−0.541, 0.494	.9288
Baseline ALT (>35 vs. <35 U/L)	−1.145	−2.451, 0.161	.0857
Time						<.0001
Surgery type (GB vs. SG)				5.914	4.702, 7.127	<.0001
Surgery type by time						.0112
GB vs. SG at time = 2				7.347	5.775, 8.919	<.0001
GB vs. SG at time = 3				7.834	6.160, 9.507	<.0001
GB vs. SG at time = 4				7.297	5.597, 8.997	<.0001
GB vs. SG at time = 5				7.640	5.856, 9.423	<.0001
Age (years)	−0.208	−0.262, −0.154	<.0001	−0.166	−0.227, −0.106	<.0001
Sex (male vs. female)	−1.575	−3.074, −0.077	.0394	−1.546	−3.134, 0.043	.0564

Abbreviations: ALT, alanine aminotransferase; CI, confidence interval; DM, diabetes; GB, gastric bypass; Hb, haemoglobin; HTN, hypertension; SBP, systolic blood pressure; SG, sleeve gastrectomy; TG, triglyceride.

Table 4. Summary of repeated measure univariable and multivariable linear model of percentage of excess weight loss versus covariates.
	Univariable analysis			Multivariable analysis
	Beta	95% CI	p Value	Beta	95% CI	p Value
Baseline DM (yes vs. no)	−6.111	−9.115, −3.107	<.0001	−3.498	−6.437, −0.523	.0212
Baseline HTN (yes vs. no)	−5.188	−8.102, −2.273	.0005
Baseline dyslipidemia (yes vs. no)	−2.743	−5.927, 0.441	.0912
Baseline weight (kg)	−0.226	−0.277, −0.175	<.0001	−0.283	−0.339, −0.226	<.0001
Baseline SBP (mmHg)	−0.066	−0.145, 0.012	.0983
Baseline TG (mg/dL)	−0.005	−0.027, 0.018	.6670
Baseline Hb (g/L)	0.769	−0.466, 2.004	.2221
Baseline ALT (>35 vs. <35 U/L)	−2.825	−5.944, 0.293	.0757
Time						<.0001
Surgery type (GB vs. SG)				13.764	10.958, 16.570	<.0001
Surgery type by time						.0585
Age (years)	−0.222	−0.355, −0.089	.0011	−0.355	−0.492, −0.218	<.0001
Sex (male vs. female)	−4.258	−7.834, −0.681	.0197	4.567	0.880, 8.254	.0153

Table 5. Summary of repeated measure multivariable linear model for predicting percentage of total weight loss: Impact of covariates.
	Beta	95% CI	p Value
HbA1c (time varying)	−1.536	−1.970, −1.102	<.0001
Baseline weight (kg)	0.035	0.006, 0.064	.0176
Baseline SBP (mmHg)	0.031	−0.008, 0.069	.1163
Time			<.0001
Surgery type (GB vs. SG)	5.594	4.161, 7.026	<.0001
Surgery type by time			.61
GB vs. SG at time = 2	0.726		ns
GB vs. SG at time = 3	1.143		ns
GB vs. SG at time = 4	0.6165		ns
GB vs. SG at time = 5	0.346	−1.643, 2.3345	ns
Age (years)	−0.174	−0.244, −0.104	<.0001
Sex (male vs. female)	−1.360	−3.204, 0.483	.1477