Bariatric surgery leads to clinically significant weight loss and improvements in cardiometabolic risk factors (1–3). However, population-based evidence evaluating the degree of improvements in cardiometabolic outcomes among those receiving bariatric surgery is limited. Using the National Health and Nutrition Examination Survey (NHANES) 2015–2018, we examined cardiometabolic risk factors among individuals who had undergone bariatric surgery, those eligible for but not receiving bariatric surgery, and normal-weight adults.
This study included adults aged ≥18 years who responded to bariatric surgery questions during the NHANES 2015–2018 cycles. The NHANES uses a stratified multistage probability method to sample the nationally representative U.S. population (https://www.cdc.gov/nchs/nhanes/index.htm). We analyzed six cardiometabolic measures—systolic blood pressure (SBP), diastolic blood pressure (DBP), hemoglobin A1c (HbA1c), total cholesterol (TC), HDL cholesterol (HDL-C), and hs-CRP—measured in a mobile examination center.
We used survey design–adjusted descriptive statistics to characterize the study population into three groups: 1) individuals with normal weight (BMI 18.5–24.9 kg/m2), 2) individuals reporting receipt of bariatric surgery, and 3) individuals medically eligible for bariatric surgery but reporting they had not received it. Surgery eligibility criteria include BMI ≥40 kg/m2 or BMI ≥35 kg/m2 and one or more obesity-related comorbidity (4). Rao-Scott χ2 tests were used to compare the study group characteristics. For the main analyses, propensity score weighting (5) was used to minimize selection bias of receiving bariatric surgery using potential confounding factors given in Table 1 (except current BMI). We then fitted general linear models to compare levels of cardiometabolic outcomes between groups. All analyses were conducted with SAS 9.4 and considered an adjusted P < 0.05 for multiple comparisons to be significant. This study was deemed exempt from review by the University of Florida Institutional Review Board because we used deidentified, publicly available data.
Propensity score–weighted comparisons of cardiometabolic measures, NHANES 2015–2018a
| Outcomes . | Normal weight . | Had bariatric surgery . | Eligible but did not have bariatric surgery . |
|---|---|---|---|
| SBP (mmHg) | 119.4 (116.9–121.8) | 121.8 (117.0–126.6) | 127.9 (126.0–129.8) |
| Difference | Ref. | 2.42 (−2.54 to 7.39) | 8.56 (4.99–12.13) |
| P value | — | 0.315 | <0.001 |
| Adjusted P valueb | — | 0.344 | 0.002 |
| DBP (mmHg) | 69.2 (68.3–70.0) | 69.3 (65.9–72.7) | 71.9 (70.7–73.1) |
| Difference | Ref. | 0.13 (−3.23 to 3.48) | 2.72 (1.78–3.66) |
| P value | — | 0.937 | <0.001 |
| Adjusted P valueb | — | 0.937 | 0.002 |
| HbA1c (%) | 5.5 (5.4–5.6) | 5.8 (5.5–6.2) | 6.3 (6.1–6.4) |
| Difference | Ref. | 0.37 (0.04–0.70) | 0.80 (0.62–0.99) |
| P value | — | 0.032 | <0.001 |
| Adjusted P valueb | — | 0.055 | 0.002 |
| TC (mg/dL) | 196.3 (189.7–202.8) | 187.5 (174.4–200.6) | 193.9 (187.0–199.0) |
| Difference | Ref. | −8.76 (−19.45 to 1.93) | −3.17 (−9.38 to 3.03) |
| P value | — | 0.101 | 0.293 |
| Adjusted P valueb | — | 0.135 | 0.344 |
| HDL-C (mg/dL) | 65.2 (62.9–67.5) | 55.1 (48.1–62.1) | 49.9 (48.1–51.6) |
| Difference | Ref. | −10.09 (−17.31 to −2.87) | −15.34 (−17.26 to −13.41) |
| P value | — | 0.009 | <0.001 |
| Adjusted P valueb | — | 0.018 | 0.002 |
| hs-CRP (mg/dL) | 2.4 (1.8–2.9) | 4.2 (2.5–5.8) | 6.9 (5.9–7.9) |
| Difference | Ref. | 1.81 (−0.02 to 3.63) | 4.69 (3.18–6.20) |
| P value | — | 0.052 | <0.001 |
| Adjusted P valueb | — | 0.078 | 0.002 |
| Outcomes . | Normal weight . | Had bariatric surgery . | Eligible but did not have bariatric surgery . |
|---|---|---|---|
| SBP (mmHg) | 119.4 (116.9–121.8) | 121.8 (117.0–126.6) | 127.9 (126.0–129.8) |
| Difference | Ref. | 2.42 (−2.54 to 7.39) | 8.56 (4.99–12.13) |
| P value | — | 0.315 | <0.001 |
| Adjusted P valueb | — | 0.344 | 0.002 |
| DBP (mmHg) | 69.2 (68.3–70.0) | 69.3 (65.9–72.7) | 71.9 (70.7–73.1) |
| Difference | Ref. | 0.13 (−3.23 to 3.48) | 2.72 (1.78–3.66) |
| P value | — | 0.937 | <0.001 |
| Adjusted P valueb | — | 0.937 | 0.002 |
| HbA1c (%) | 5.5 (5.4–5.6) | 5.8 (5.5–6.2) | 6.3 (6.1–6.4) |
| Difference | Ref. | 0.37 (0.04–0.70) | 0.80 (0.62–0.99) |
| P value | — | 0.032 | <0.001 |
| Adjusted P valueb | — | 0.055 | 0.002 |
| TC (mg/dL) | 196.3 (189.7–202.8) | 187.5 (174.4–200.6) | 193.9 (187.0–199.0) |
| Difference | Ref. | −8.76 (−19.45 to 1.93) | −3.17 (−9.38 to 3.03) |
| P value | — | 0.101 | 0.293 |
| Adjusted P valueb | — | 0.135 | 0.344 |
| HDL-C (mg/dL) | 65.2 (62.9–67.5) | 55.1 (48.1–62.1) | 49.9 (48.1–51.6) |
| Difference | Ref. | −10.09 (−17.31 to −2.87) | −15.34 (−17.26 to −13.41) |
| P value | — | 0.009 | <0.001 |
| Adjusted P valueb | — | 0.018 | 0.002 |
| hs-CRP (mg/dL) | 2.4 (1.8–2.9) | 4.2 (2.5–5.8) | 6.9 (5.9–7.9) |
| Difference | Ref. | 1.81 (−0.02 to 3.63) | 4.69 (3.18–6.20) |
| P value | — | 0.052 | <0.001 |
| Adjusted P valueb | — | 0.078 | 0.002 |
Data are estimates (95% CI) unless otherwise indicated. Ref., reference.
Propensity score weighting adjusted for age, sex, race, education, employment, family income, health insurance type, general health status, current smoking, and the number of comorbidities.
Adjusted for multiple comparisons using the Benjamin-Hochberg method.
Of 6,274 participants (mean age 49.8 years, 55.4% women, 64.9% White), 132 (2.1%) reported having bariatric surgery and 2,698 (43.0%) were eligible for bariatric surgery. Compared with normal-weight individuals, those receiving bariatric surgery were more likely to be older, female, White, and highly educated and to have higher family income. After the propensity score weighting, there were no significant differences in these characteristics.
Despite significantly lower BMI among individuals with normal weight (22.3 kg/m2) relative to individuals receiving bariatric surgery (34.9 kg/m2), SBP, DBP, HbA1c, TC, and hs-CRP were not significantly different between groups (Table 1). Individuals who were eligible for bariatric surgery but did not have it had significantly higher levels of SBP, DBP, HbA1c, and hs-CRP and lower HDL-C compared with normal-weight individuals. As sensitivity checks, we tested the robustness of the main findings by including individual characteristics as covariates and excluding individuals diagnosed with diabetes or heart diseases from the analytic sample. Results remained consistent across groups.
This population-based study included a weighted sample size of 3.6 million adults who reported receiving bariatric surgery. No statistically significant differences in numerous cardiometabolic risk factors were observed between normal-weight and bariatric surgery groups, despite those having received bariatric surgery having BMI values in the obesity range. Moreover, cardiometabolic risk factors in individuals reporting having undergone bariatric surgery were significantly different versus individuals eligible but not having undergone surgery, suggesting that even in the absence of achieving a normal-weight BMI following bariatric surgery, the cardiometabolic risk factor profile appears to normalize substantially.
Study limitations included the cross-sectional design, small sample size for bariatric surgery, and self-reported measures. Although propensity weighting helps to account for potential confounders, lack of baseline information before or at the time of surgery made it difficult to have well-matched comparison groups (5). Long-term follow-up studies with larger samples will be necessary to confirm the effect of bariatric surgery on cardiometabolic health benefits and potential harms (3).
In summary, this population-based study adds to the growing body of evidence suggesting that bariatric surgery can meaningfully improve a number of cardiometabolic risk factors to a degree roughly equivalent to those factors in normal-weight individuals despite residual adiposity.
Article Information
Funding. There was no external funder for this study. M.I.C. is supported by the National Institutes of Health National Heart, Lung, and Blood Institute (K01HL141535). D.J.L. was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (K01DK115632) and the University of Florida Clinical and Translational Science Institute (UL1TR001427).
The funders had no role in the design and conduct of the study, analysis and interpretation of the data, or preparation, review, or approval of the manuscript.
Duality of Interest. A.S.K. receives research support (drug and placebo) from AstraZeneca Pharmaceuticals and serves as a consultant for Novo Nordisk, Orexigen, Vivus Pharmaceuticals, and WW but does not accept personal or professional income for these activities. M.I.C. has served as a paid consultant for WW and an unpaid consultant for Novo Nordisk. No other potential conflicts of interest relevant to this article were reported.
Author Contribution. Y.-R.H., D.J.L., and M.I.C. conceived the research. Y.-R.H. analyzed the data and wrote the manuscript. A.S.K., C.J.-M., D.J.L., and M.I.C. assisted with the study design and contributed to the interpretation of the results. All authors contributed critical intellectual content and made important revisions to the manuscript. Y.-R.H. and M.I.C. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
