Subclinical Hypothyroidism and Type 2 Diabetes: A Systematic Review and Meta-Analysis

Background Abundant evidence suggests an association between subclinical hypothyroidism (SCH) and type 2 diabetes mellitus (T2DM), but small sample sizes and inconclusive data in the literature complicate this assertion. Objective We measured the prevalence of SCH in T2DM population, and investigated whether T2DM increase the risk of SCH and whether SCH was associated with diabetic complications. METHODS We conducted a meta-analysis using PubMed, EMBASE, Web of Science, Wan Fang, CNKI and VIP databases for literature search. We obtained studies published between January 1, 1980 to December 1, 2014. The studies were selected to evaluate the prevalence of SCH in T2DM subjects, compare the prevalence of SCH in T2DM subjects with those nondiabetics, and investigate whether diabetic complications were more prevalent in SCH than those who were euthyroid. Fixed and random effects meta-analysis models were used, and the outcome was presented as a pooled prevalence with 95% confidence interval (95% CI) or a summary odds ratio (OR) with 95% CI. RESULTS Through literature search, 36 articles met the inclusion criteria and these articles contained a total of 61 studies. Funnel plots and Egger's tests showed no publication bias in our studies, except for the pooled prevalence of SCH in T2DM (P = 0.08) and OR for SCH in T2DM (P = 0.04). Trim and fill method was used to correct the results and five potential missing data were replaced respectively. The adjusted pooled prevalence of SCH in T2DM patients was 10.2%, meanwhile, T2DM was associated with a 1.93-fold increase in risk of SCH (95% CI: 1.66, 2.24). Furthermore, SCH might affect the development of diabetic complications with an overall OR of 1.74 (95% CI: 1.34, 2.28) for diabetic nephropathy, 1.42 (95% CI: 1.21, 1.67) for diabetic retinopathy, 1.85 (95% CI: 1.35, 2.54) for peripheral arterial disease, and 1.87 (95% CI: 1.06, 3.28) for diabetic peripheral neuropathy. Conclusions T2DM patients are more likely to have SCH when compared with healthy population and SCH may be associated with increased diabetic complications. It is necessary to screen thyroid function in patients with T2DM, and appropriate individualized treatments in addition to thyroid function test should be given to T2DM patients with SCH as well.