The long-term safety of topical corticosteroids in atopic dermatitis: A systematic review.

BACKGROUND

Topical corticosteroids (TCS) are a first-line treatment for eczema, but there are concerns about their safety when used long-term.

OBJECTIVES

To systematically review adverse effects associated with longer-term use of TCS for eczema.

METHODS

Randomised controlled trials (RCTs), cohort and case-control studies reporting adverse effects of TCS (comparators: no TCS treatment, other topicals) in patients with eczema were identified. Included studies had greater than one year of follow-up, minimum cohort size of 50 participants, or minimum 50 per arm for RCTs. Evidence was GRADE-assessed. Prospero registration CRD42021286413.

RESULTS

We found seven studies (two randomised, five observational); two RCTs (n = 2570, including 1288 receiving TCS), two cohort (all received TCS n = 148) and three case-control studies (cases n = 10 322, controls n = 12 201). Evidence from two RCTS (n = 2570, children, three and five years' duration) comparing TCS to topical calcineurin inhibitors found intermittent TCS use probably results in little to no difference in risk of growth abnormalities, non-skin infections, impaired vaccine response and lymphoma/non lymphoma malignancies. The five-year RCT reported only one episode of skin atrophy (n = 1213 TCS arm; mild/moderate potency), suggesting TCS use probably results in little to no difference in skin thinning when used intermittently to treat flares. No cases of clinical adrenal insufficiency were reported in 75 patients using mild/moderate TCS in the three-year RCT. Small associations between TCS and type-2 diabetes and lymphoma were identified in two case-control studies compared to no TCS, but the evidence is very uncertain. No long-term studies concerning topical steroid withdrawal or eye problems were identified.

CONCLUSION

This review provides some reassuring data on growth and skin thinning when TCS are used intermittently for up to 5 years, but many knowledge gaps remain.