Fourteen observational studies (3,801 patients, range 36 to 1,025), were selected. Twelve studies scored 7 or more on the nine-point Newcastle-Ottawa Scale. Four studies were classed as high quality, eight were medium quality and two were low quality.
Patients of Caucasian background had significantly great percentage of excess weight loss (MD -8.36%, 95% CI -10.79 to -5.93; 13 studies). There was evidence of significant heterogeneity (Ι²=67%). Subgroup analyses by type of surgery showed similar results (MD -8.39%, 95% CI -11.38 to -5.40 in 10 studies for malabsortive surgery and MD -8.46, 95% CI -12.95 to -3.97 in four studies for restrictive surgery).
Remission of diabetes was more frequent in African American than in Caucasian patients but the difference between the groups was not statistically significant (OR 1.41, 95% CI 0.56 to 3.52; three studies). There was evidence of significant heterogeneity (Ι²=59%).
Results of sensitivity analyses were reported. Some evidence of publication bias was found but sensitivity analyses suggested that this did not significantly affect the pooled estimates.