Hypothetical Pneumocystis jirovecii transmission from immunocompetent carriers to infant.

To the Editor

The recent dispatch article by Rivero et al. reports the transmission of Pneumocystis jirovecii from immunocompetent grandparents to their granddaughter (). The authors’ conclusion was based on 2 facts: the grandparents were carriers but neither the parents nor the child’s brother was a carrier, and the P. jirovecii genotype observed in the grandparents was identical to that found in the infant. In our opinion, the data provided by the authors do not support the conclusion that transmission has occurred. First, the 2 markers used for typing show a small number of alleles and thus provide low discrimination among isolates (). Consequently, the P. jirovecii isolates present in the grandparents and in the infant may have been epidemiologically unrelated. Second, the frequency of occurrence of the different genotypes obtained was not investigated. The presence of the same genotype in the grandparents and in the infant may result from a high frequency of this genotype in the geographic area where the family lived. In fact, the use of a validated typing method and the analysis of unlinked control patients have proven necessary in other studies to demonstrate transmission of P. jirovecii (–). We believe that the reported transmission event remains a hypothesis.

We appreciate the comments from Hauser and colleagues () regarding our article that reported a case of Pneumocystis jirovecii transmission from colonized grandparents to their infant granddaughter (). We agree with the authors that the 2 markers used for typing, which are described in our article, present a small number of alleles and thus provide low discrimination between isolates. However, these typing methods have been validated and have proven useful for molecular epidemiologic applications in P. jirovecii colonization studies (,). Unfortunately, other typing methods that can identify a high number of alleles, such as the sequence analysis of the internal transcribed spacer no. 1 and 2 gene regions, could not be used in our study because a low amplification rate has been observed for these regions when such methods are used to study colonized subjects (). On the other hand, the multitarget single-strand conformation polymorphism method has been used only in patients with Pneumocystis pneumonia, and its usefulness for epidemiologic studies in colonized subjects has not been proven (). For our study, we think that genotyping analysis of the mtLSU rRNA gene together with the dihydropteroate synthase (DHPS) gene provided sufficient epidemiologic information because this strategy allows identification of 24 different combinations of genotypes. However, no typing method is able to demonstrate interhuman P. jirovecii transmission conclusively because a common environmental source of infection cannot be ruled out in any case. Therefore, as we noted in our article, “We cannot exclude the possibility that the cases described were infected by the same environmental source,” and we only hypothesized that “the infant was infected by P. jirovecii through close contact with her grandparents.” However, we continue to think that the airborne transmission of P. jirovecii from the grandfather to the grandmother and the infant is the most probable explanation based on genotype data. Also, all persons in close contact with the infant were studied, and only her grandparents were colonized by P. jirovecii. Future research is needed to assess the importance of colonized subjects in the P. jirovecii transmission to susceptible hosts.