, Pietro Serra , for the Italian Society of Internal Medicine ( SIMI ) Study

12 www.thelancetinfection.com Vol 11 January 2011 immediately. Moreover, caution should be taken when results of MRSA cultures are interpreted. In the absence of quantitative cultures, only bacteraemic episodes, or positive cultures from sites that are normally sterile, would be defi nite evidence for infection. Pseudomonas aeruginosa was even slightly higher in communityacquired than in hospital-acquired pneumonia (9·7% vs 7·1%), and Gram-negative enterobacteriaceae were more frequent in HCAP than in hospital-acquired pneumonia (32·1 vs 16·7). The unusually high rates of P aeruginosa and Gramnegative enteric bacilli in communityacquired pneumonia add to our reservations about the validity of the microbiological data. The investigators do not provide data for resistance patterns of P aeruginosa and Gram-negative enteric bacilli; we cannot therefore know the true rate of multidrug-resistant pathogens, although they claim to have identifi ed an excessive rate of multidrug resistance in patients meeting the defi nition for HCAP. Venditti and colleagues try to convince us that HCAP is diff erent from community-acquired pneumonia with just 22 patients (11 with MRSA, two with P aeruginosa, nine with pneumonia, both epidemiologically and microbiologically.