The aim of this study was to estimate the timing of relapse, the prevalence of multiple relapses and the predictors of relapse in patients with giant cell arteritis (GCA).

PubMed, Embase and Cochrane databases were searched from inception till November, 30 2021. Outcome measures include cumulative relapse rate (CRR) of first relapse at year 1, 2, and 5 after treatment initiation, CRR of second and third relapse and predictors of relapse.

Thirty studies (2595 patients) were included for timing of relapse, 16 studies (1947 patients) for prevalence of multiple relapses and 40 studies (4213 patients) for predictors of relapse. One-year, 2-year and 5-year CRRs were 32% [95% confidence interval (CI) 22–43%], 44% [95% CI 31–59%], and 47% [95% CI 27–67%], respectively. The duration of scheduled glucocorticoid therapy was negatively associated with the 1-year CRR (P=0.03). CRR of second and third relapse were 30% [95% CI 21–40] and 17% [95% CI 8–33%], respectively. Female sex (OR 1.43) and large vessel involvement (OR 2.04) were predictors of relapse.

Relapse occurred in almost half of GCA patients mainly during the first two years after diagnosis. One in three patients had multiple relapses. The optimal glucocorticoid tapering schedule, which seeks a balance between the lowest relapse risk and the shortest glucocorticoid duration, needs to be determined in future studies. Longer scheduled glucocorticoid therapy or early introduction of glucocorticoid-sparing agents may be warranted in female patients and patients with large vessel involvement.