To the Editors: We concur with Chien et al1 who conclude that colonization studies should include measurement of density. We also prefer to use culture to determine if a volunteer is carriage positive because we are concerned, like them, that real-time quantitative polymerase chain reaction (qPCR) could give false positives through the detection of dead bacteria. We have developed an experimental human pneumococcal carriage (EHPC) model2,3 and have examined colonization dynamics in the adult nasopharynx. The authors state that at a qPCR carriage density of ≤105 colony forming units (CFU)/mL the detection of pneumococcal carriage by culture was significantly less sensitive than qPCR. In our experiment, however, we do not freeze the nasopharyngeal sample and observe a much improved sensitivity of culture. We first tested our culture detection method ex vivo by spiking nasal washes with predetermined amounts of pneumococcus. Using this method, we were able to detect a lowest carriage density of 101 CFU/mL, as measured by culture and confirmed by qPCR. We have not had any carriage positive volunteers in our EHPC model with a carriage density exceeding 105 CFU/mL and we have been able to detect down to 101 CFU/mL by culture, without including an enrichment step as Chien et al did. The differences we see in the limit of detection between our study and the Chien study are likely because we do not freeze our samples before processing. Although STGG media (skim milk, tryptone, glucose, glycerol) has been shown to be as good as direct plating4, there is often some bacterial loss after freezing, which is very important in low density samples. We also have to take into account the differences between our sampling populations and sampling methods. Using a reduction in, or elimination of, carriage density and/or duration as an endpoint could be a useful mechanism to test a vaccine, but we would like to add that the detection of low density carriage is crucial. We suggest that culture remains an accurate and sensitive method of detection in fresh samples whereas qPCR offers substantial advantage when samples must be frozen. Jenna F. Gritzfeld, MSc Stephen B. Gordon, MD Respiratory Infection Group Liverpool School of Tropical Medicine Liverpool, United Kingdom Amelieke Cremers, MSc Laboratory of Pediatric Infectious Diseases Radboud University Nijmegen Medical Centre Nijmegen, The Netherlands