Table 1:
Baseline Characteristics of VSG vs. RYGB Patients - Pre and Post Matchinga
| Pre-Match | Post-Match | |||||
|---|---|---|---|---|---|---|
| Variable | RYGB % (n=1,167) | VSG % (n=1,398) | Standardized Differenceb | RYGB % (n=922) | VSG % (n=1,111) | Standardized Differenceb |
| Age Group | ||||||
| 18–39 y | 14% | 19% | 16% | 16% | ||
| 40–49y | 32% | 32% | 0.14 | 34% | 34% | 0.00 |
| 50–59y | 40% | 36% | 40% | 40% | ||
| 60–64y | 14% | 13% | 10% | 10% | ||
| Female Sex | 65% | 65% | 0.00 | 69% | 64% | −0.11 |
| Non-Hispanic White Race/Ethnicityc | 55% | 52% | −0.04 | 50% | 52% | −0.05 |
| Educationc | ||||||
| < HS | 1% | 1% | 1% | 1% | ||
| HS Graduate | 27% | 27% | 0.02 | 28% | 27% | 0.04 |
| Some College | 59% | 59% | 60% | 59% | ||
| College Graduate | 12% | 12% | 10% | 13% | ||
| Missing | 1% | 1% | 1% | 1% | ||
| Region | ||||||
| West | 26% | 20% | 22% | 20% | ||
| South | 42% | 52% | 0.15 | 47% | 51% | 0.10 |
| Midwest | 22% | 19% | 20% | 20% | ||
| Northeast | 10% | 10% | 11% | 9% | ||
| Time Period | ||||||
| 2010–12 | 47% | 23% | 0.52 | 27% | 26% | 0.02 |
| 2013–16 | 53% | 77% | 73% | 74% | ||
| BMI (kg/m2) Category | ||||||
| 30–39.9 | 18% | 22% | 20% | 22% | ||
| 40–49.9 | 43% | 46% | 0.18 | 47% | 45% | 0.05 |
| 50–59.9 | 20% | 18% | 19% | 19% | ||
| ≥60 | 3% | 3% | 2% | 3% | ||
| Non-specific obesity | 15% | 10% | 11% | 10% | ||
| Missing | 1% | 1% | 1% | 1% | ||
| ACG Score ≥3c | 25% | 26% | 0.01 | 21% | 23% | 0.06 |
| Hypertensionc | 69% | 66% | −0.06 | 66% | 65% | −0.02 |
| Cardiovascular Diseasec | 12% | 12% | −0.02 | 8% | 11% | 0.10 |
| Mental Illnessc | 16% | 15% | −0.03 | 14% | 15% | 0.02 |
| Acute DM Complicationsc | 18% | 17% | −0.04 | 9% | 9% | 0.00 |
| Chronic Microvascular Complicationsc | 21% | 16% | −0.13 | 11% | 11% | 0.00 |
| Baseline Insulin Usersc | 38% | 26% | −0.26 | 23% | 23% | 0.00 |
| Baseline Metformin Only Usersc | 28% | 37% | 0.19 | 40% | 40% | 0.00 |
We conducted coarsened exact matching on key predictors of diabetes remission: age group, baseline insulin use, baseline total diabetes medication burden (tertile of total MDD in the 6 months prior to surgery), and presence or absence of acute diabetes complications and chronic microvascular complications in the 6 months prior to surgery. Additionally, we exact matched on duration of enrollment after surgery to minimize bias introduced by differential loss to follow-up, and tertile of a propensity score for undergoing VSG vs. RYGB, based on sex, race, education, region, year group, BMI category, ACG score, and comorbidities.
Standardized differences are the difference in means between the intervention and control groups divided by the SD of the difference in means. Lower values indicate greater similarity between VSG and RYGB, and values <0.2 indicate minimal differences between groups.
for complete descriptions of how we constructed baseline variables, please refer to the methods section.