Skip to main content
. 2012 Nov 14;35(12):2690–2697. doi: 10.2337/dc12-0312

Figure 1.

Figure 1

Search results. *Articles that were not a peer-reviewed, published randomized clinical trial or observational study were not double-counted (among articles that did not specify the proportion [numerator and denominator] of subjects seroprotected by diabetes status [or data available to allow the calculation of the proportions] or the odds ratio from a multivariate analysis that included diabetes as a predictor variable) so that the total would equal 15. †The number of articles does not total 13 because one article met two inclusion criteria.