The severe acute respiratory syndrome (SARS) pandemic caught the world by surprise in 2003 and spread rapidly within a relatively short period of time. Hence, randomised placebocontrolled clinical trials on the treatment of SARS were not possible. Our understanding was obtained from observational, cohort studies, case series and reports. Nevertheless, such information is useful in providing clinical management guidelines and directing future research in case SARS recurs. Early in the pandemic, a combination of ribavirin and corticosteroids was adopted as the standard treatment in Hong Kong, Canada and elsewhere because of the apparent good results of the first few patients. Subsequent reports showed that ribavirin was associated with a high rate of toxicity and lacked in vitro antiviral effect on SARS-coronavirus (SAR-CoV). The timing and dosage regimens of steroid in the treatment of SARS are controversial. Pulse methylprednisolone 250 to 500 mg/day for 3 to 6 days has been reported to have some efficacy in a subset of patients with “critical SARS,” ie, critically ill SARS patients with deteriorating radiographic consolidation, increasing oxygen requirement with PaO2< 10 kPa or SpO2< 90% on air, and respiratory distress (rate of 30/min). Prolonged therapy with high-dose steroids, in the absence of an effective antimicrobial agent, could predispose patients to complications such as disseminated fungal infection, and avascular necrosis. Kaletra (400 mg ritonavir and 100 mg lopinavir), a protease inhibitor used in the treatment of human immunodeficiency virus infection, may be considered for early treatment of SARS patients, preferably in a randomised double-blind placebocontrolled clinical trial setting. Interferon (IFN) is not recommended as standard therapy in SARS. However, there are enough data on in vitro activity of IFN preparations and a few clinical studies for these products to support a controlled trial if SARS recurs. Many other experimental treatments have been tried in an uncontrolled manner, and they should not be recommended as standard therapy.