Treatment of severe acute respiratory syndrome (SARS) is experimental, and the effectiveness of ribavirin–steroid therapy is unclear. Forty SARS patients with progressive disease after ribavirin treatment and 1.5 g of pulsed methylprednisolone were given either convalescent plasma (n = 19) or further pulsed methylprednisolone (n = 21) in a retrospective non‐randomised study. Good clinical outcome was defined as discharge by day 22 following the onset of symptoms. Convalescent plasma was obtained from recovered patients after informed consent. Patients in the plasma group had a shorter hospital stay (p 0.001) and lower mortality (p 0.049) than the comparator group. No immediate adverse effects were observed following plasma infusion.