Twenty-eight studies (n=2,382) were included in the meta-analyses. Mean study quality was 2.2. Fourteen studies had a short follow-up and fourteen studies had a long follow-up.
When all studies were pooled, risk of glucocorticoid-related adverse events was 150 per 100 patient-years (95% CI 132 to 169).
In patients with rheumatoid arthritis (14 studies, n=796), risk of glucocorticoid-related adverse events was 43 per 100 patient-years (95% CI 30 to 55).
In patients with polymyalgia rheumatica (four studies, n=167), risk of glucocorticoid-related adverse events was 80 per 100 patient-years (95% CI 15 to146).
In patients with inflammatory bowel disease (10 studies, n=1,419), risk of glucocorticoid-related adverse events was 555 per 100 patient-years (95% CI 391 to 718).
The most frequently reported adverse events were psychological and behavioural adverse events (such as minor mood disturbances). Gastrointestinal adverse events (such as dyspepsia and dysphagia) were also commonly reported. Higher rates of adverse events were reported in high-quality studies with a short follow-up, particularly in studies of patients with inflammatory bowel disease. Subgroup analysis of low and medium doses did not show dose dependency for any adverse event.