Reply 1

Abstract

Dr Chang et al. rightly point to the differences in the various clinical, radiological and prognostic features of Pneumocystis pneumonia (PCP) in patients with or without the acquired immune deficiency syndrome. PCP is more acute in HIV-negative patients, has different radiological findings, and carries a worse prognosis. Mortality in HIV-negative patients with PCP has been reported to be 17.2–52.9%, which is much higher than in patients with acquired immune deficiency syndrome and PCP. Considering the acute presentation and the high mortality rates associated with PCP among HIV-negative patients, it is likely that a more timely diagnosis would improve the prognosis. However, the yield of direct staining in respiratory secretions is unacceptably low in HIV-negative patients because of a much lower burden of organisms.1 In addition, a diagnosis is often established only after performance of a lung biopsy—a potentially dangerous procedure for immunocompromised patients who often have thrombocytopenia, impending respiratory failure and significant comorbid conditions. These difficulties are illustrated in the case described by Chang et al., as non-invasive microbiological studies were inconclusive, and a lung biopsy was eventually required. A polymerase chain reaction (PCR)-based assay for the detection of Pneumocystis jirovecii DNA may lead to a more rapid diagnosis of PCP among patients without acquired immune deficiency syndrome. We have published recently our experience with a ‘home-brewed’ PCR assay for the detection of P. jirovecii DNA based on three different genes. The sensitivity of such a test was superior to direct staining of bronchoalveolar lavage.2 On the other hand, the use of PCR-based assays carries the risk of a false-positive diagnosis of PCP in patients who are colonized, but not infected, with P. jirovecii. A number of other studies that investigated the use of PCR in the diagnosis of PCP have been published in recent years as well.3-5 Because current PCR assays are not standardized and use different PCR techniques (e.g. quantitative, conventional vs real-time PCR), the reported accuracy of the various assays differs. In nearly all PCR-based assays published in recent years, however, the negative predictive value and the sensitivity are extremely high.3-5 Although reported specificity and positive predictive value differ widely, they are high in most studies and can be improved by using higher cut-off values when quantitative PCR is used. PCP in HIV-negative patients runs a rapid course, and is associated with a high mortality rate. It also frequently yields negative results when direct staining is used. We, therefore, argue that a validated PCR assay should be used for all immunocompromised, HIV-negative patients who are suspected of having PCP on epidemiological, clinical and radiological grounds. A negative test essentially rules out the diagnosis within hours, and a positive test is more sensitive than ‘traditional’ direct staining. A prospective trial designed to test this hypothesis is warranted.