Background: A credible body of epidemiologic evidence documents elevated bladder cancer risk among patients with type 2 diabetes, but risk among those with type 1 diabetes (T1DM) has received less attention. We sought to identify and quantitatively synthesize epidemiologic data on the association between T1DM and incident bladder cancer.
Methods: We searched Medline, Embase, Cochrane, and Web of Science through June 5, 2018, to identify non-redundant English language reports of original human data estimating the association between T1DM and incident bladder cancer addressing age as potential confounder. Additional citations were identified through hand searching efforts including a Google Scholar search, contacting authors, and retrieving citations in selected studies. Candidate reports were critically appraised by two independent reviewers, with conflicts resolved by consensus. Harmonized estimates of the T1DM-bladder cancer association were extracted from qualifying reports and synthesized by random effects meta-analysis.
Findings: Among 7,673 reports identified in the search, 59 warranted critical full text review, and 4 described 7 non-redundant studies satisfying inclusion criteria (N=362 T1DM patients). Estimate of the overall summary odds ratio association was 1.07 (95% CI: 0.87-1.32; I2: 69.1%, p heterogeneity= 0.001; 7 studies). Gender-specific summary estimates were 1.40 (95% CI: 0.95-2.05; I2: 73.9%, p heterogeneity= 0.002; 6 studies) for females, and 0.95 (95% CI: 0.71-1.26; I2: 61.4%, p heterogeneity= 0.024; 6 studies) for males.
Interpretation: Available data reveal a modest positive association between incident bladder cancer and T1DM that does not achieve statistical significance. Etiologic interpretation is precluded by indicators of heterogeneity and design limitations of contributing studies, primarily confounding by smoking and error in measurement of diabetes in referent groups.
H.B. Oskoui: None. D.D. Bogumil: None. L. Kysh: None. M. Barrett: None. K. Siegmund: None. V.K. Cortessis: None.