SARS infection control.

Sir

The Research letter by W H Seto and colleagues (May 3, p 1519) about infection control for severe acute respiratory syndrome (SARS) begs a few questions.

First, were any of the individuals with a cough among the 11 index patients? Cough was a symptom in eight of ten patients reported by Tsang and colleagues and was a clinical feature in only 57% of the 156 patients reported by Lee and co-workers. Readers cannot and should not assume in a study with few patients that the sample is representative of the population. Patients who do not cough obviously do not generate aerosol. A small non-representative sample of patients with SARS who happen not to have a cough would give a false impression that wearing surgical masks alone offers adequate protection against aerosol when no aerosol exists. Whereas coughing produces large droplets that settle rapidly in the absence of air currents, it also generates particles less than 1 μm in diameter, suggesting that surgical masks are not an absolute means of protection against respiratory pathogens.

Second, were any of the 11 index patients intubated or admitted to the intensive care unit (ICU)? If none of the 11 patients were admitted to the ICU, then this group might not be representative of all patients with SARS, of whom up to 23% were admitted to an ICU in one report.

Finally, none of the affected staff seem to be ICU staff, possibly because none of the index patients was admitted to the ICU. However, many of the health-care workers infected elsewhere have been ICU staff, suggesting that again the sample studied by Seto and colleagues is not representative. We would like a more thorough breakdown of the roles of the 241 non-infected staff.

The results of studies that involve limited numbers of participants should always be viewed as preliminary.