Genotyping samples of
P. jiroveci provides a method for exploring epidemiologic issues. For example, one study examined the possibility that the low incidence of PCP in African HIV-infected persons might be due to the presence or absence of certain strains of
P. jiroveci. However, samples of
P. jiroveci from Zimbabwe, Brazil, the United States, and the United Kingdom have exhibited no major differences in genotypes (
47,
48). Another example is a study in which genotyping at four different genetic loci was used to compare isolates of
P. jiroveci collected before (1968–1981) and after (1982 to present) the beginning of the AIDS pandemic
(47). Pre- and postpandemic samples were the same except for a single base polymorphism (in the mitochondrial large subunit rRNA gene) found in the pre-pandemic samples only. These data show that the large increase in incidence of PCP was not accompanied by a shift in the kinds or frequencies of strains of
P. jiroveci.Strain analysis has also led to observations that are difficult to reconcile with the traditional view of the relationship between P. jiroveci and humans. The traditional theory holds that clinically important infection results from reactivation of a latent infection that was acquired during childhood. While infection of young children appears to be common, latent P. jiroveci has not been directly observed in healthy adults. In addition, indirect evidence is difficult to reconcile with lifelong latency.
The latency issue is important for several reasons. Under the reactivation of latent infection theory, little rationale exists for instituting measures to minimize the risk of infection during adulthood because this infection has already occurred. On the other hand, person-to-person transmission of the disease would have important public heath implications for medical centers that treat HIV-infected patients or other immunocompromised persons (
42–
44,
49–
52). Furthermore, transmission from patients who are undergoing treatment for PCP might enhance the opportunity for drug resistance to arise. By contrast, the generation of drug resistance would be less of a concern if most or all infections were due to transmission from an immunocompetent person, such as a young child's mother, or another child (i.e, someone who is not being treated for PCP). Under these conditions, drug-resistant strains, if they arose, would not spread very effectively.
PCP develops in infants infected with HIV perinatally, suggesting that
P. jiroveci was present in these infants’ environments early in their lives
(53). Evidence of
P. jiroveci has also been found in some victims of sudden infant death syndrome (SIDS)
(54). In normal, healthy children, serologic data have long indicated that infection of young children is common. Most children develop anti-
Pneumocystis antibodies early in life, and the prevalence of these antibodies appears to increase with age (
48,
55). Recently,
P. jiroveci has been linked to clinical illness in normal, healthy infants
(51).
P. jiroveci DNA was identified in nasopharyngeal aspirates obtained during episodes of mild respiratory infection in 24 (32%) of 74 infants. Seroconversion developed by 20 months of age in 67 (85%) of 79 infants who remained in the study and occurred in the absence of any symptoms of disease in 14 (18%). These reports confirm previous ones showing infection of children (
1,
3,
4). Young children may be a reservoir of infectious
P. jiroveci in the community.
Although infection of children seems common, little evidence exists for lifelong latency. Using PCR, Wakefield found no evidence of
P. jiroveci in bronchoalveolar lavage fluid from 10 healthy persons
(56). Peters replicated this result in postmortem lung tissue from 15 immunocompetent adults (
56,
57). (The techniques used to detect
P. jiroveci have found it in HIV-negative adults but only those with other health problems [
58].) Studies on recurrent PCP have shown that different
P. jiroveci genotypes are present during different PCP episodes in patients with repeat episodes of PCP, a result suggestive of infection proximal to the time of disease (
42-
44). Recent infections of adults are also suggested by the high frequency of mutations that cause changes in the sequence of the DHPS gene, the enzyme associated with sulfonamide resistance in other pathogens (
59–
61). These mutations have not been detected in patients in whom PCP occurred at a time before the widespread use of sulfonamides to treat and prevent it
(62) but are common in today's patients, even in those with no known exposure to sulfonamides (
61,
63). Mutant DHPS genes have been found in a variety of
P. jiroveci genetic backgrounds, suggesting that selection for DHPS mutations is an ongoing process
(64).
An alternative approach to exploring the importance of latency is employing population genetics and epidemiology to test the following hypothesis. If lifelong latency is important, adult patients who reside far from their birthplace should have the strain of
P. jiroveci common in their place of birth, not in their place of residence. Data pertaining to this hypothesis are now available
(64). The strains infecting adult patients were more similar to those common in their place of residence than their place of birth, suggesting that infections had been recently acquired, rather than carried since early childhood.
Latent P. jiroveci have not been found in healthy adults, but proving that they do not exist is practically impossible. A single organism anywhere in the body could be sufficient to maintain a latent infection. Therefore, the possibility of latency remains. However, latent infections may be transitory, and humans who have eliminated the microbe may be subject to reinfection. The observations described above seem more consistent with this “transient colonization” scenario than with lifelong latency.