Can we predict relapse in Graves' disease? Results from a systematic review and meta-analysis

Identification of pretreatment risk factors predicting relapse in patients with hyperthyroidism of Graves' disease after stopping anti-thyroid drugs (ATD) is decisive to guide therapeutic options. We performed a systematic search and meta-analysis to study predictors for relapse after stopping ATD in patients with Graves' disease. Based on a pre-specified protocol, we searched PubMed, EMBASE and Cochrane in July 2015 for case-control, controlled and randomized-controlled trials reporting risk factors for relapse after stopping ATD. The primary endpoint was relapse of disease until follow-up. PRISMA and SIGN statements were used for reviewing the data and assessing the quality of included trials. We included 54 trials with a total of 7595 participants. Most trials were small with moderate-to-high risk for bias. Ten trials were assessed only qualitatively (2227 patients), genomic data were reported in 13 trials (2178 patients) and 31 trials (4346 patients) were assessed quantitatively. In total, there were 3696 relapses in 7595 patients (48.7%). By using random-effects meta-analysis, orbitopathy, smoking, thyroid volume measured by sonography, goiter size, fT4, tT3, TRAb and TBII were significantly associated with relapse, whereas male vs female sex, age and initial tT4 level did not show significant associations. CONCLUSIONS: This analysis found several risk factors to predict relapse in Graves' disease, which can be combined in a risk score. Prospective studies should evaluate the prognostic accuracy of such a score to guide treatment decisions.