This is the authors' abstract. We don't add key points for this paper.
(See the article by Sax et al., on pages 197–202.) Pneumocystis pneumonia (PCP) has historically been a leading cause of morbidity and mortality in individuals with HIV infection. PCP still causes substantial morbidity in other immunosuppressed populations, including transplant recipients and patients receiving prolonged courses of high-dose steroids, especially when anti-pneumocystis prophylaxis is not administered. With the advent of combination antiretroviral therapy (ART) for patients with HIV infection, the incidence of this disease has decreased in this patient population [1, 2]. PCP still occurs in individuals without access to ART, in persons not aware of their HIV status, and in those not able or willing to take ART. PCP remains the leading serious opportunistic infection [3, 4]. Over the past thirty years, the diagnostic approach for PCP has improved dramatically. No longer does diagnosis depend on obtaining lung tissue by open or even transbronchial biopsy: bronchoalveolar lavage fluid or induced sputum is adequate for identification of the organisms. Because pneumocystis cannot be cultured, identification requires direct staining; the cumbersome methenamine silver and the difficult-to-read Giemsa or Diff-Quik have been replaced by a highly specific, highly sensitive immunofluorescent stain in most laboratories. Although a number of groups have developed nucleic acid amplification tests that are very sensitive for detecting pneumocystis in respiratory secretions, they lack specificity in distinguishing colonization from disease. There is no commercially available kit in the United States for diagnosis of PCP. For many years, a holy grail for investigators has been the development of a serologic test to diagnose PCP. Unfortunately, results of polymerase chain reaction (PCR)–based assays using peripheral blood samples are rarely positive. Other markers, such as blood S-adenosylmethionine, LDH, and detection of blood KL-6, have not demonstrated performance characteristics to suggest that they would be useful diagnostically [5–8]. Is the serum (1->3)-β-D-glucan test the answer to this quest for a reliable serologic test for PCP? The peripheral measurement of (1->3)-β-D-glucan is based on detection of β-glucan, a polysaccharide present in the wall of Pneumocystis cysts and in other fungi [9, 10]. β-glucan triggers an innate immune response, and it can be detected both in bronchoalveolar lavage fluid and in blood [9, 11]. Previous work, primarily in retrospective case series, has reported a sensitivity of β-glucan ranging from 90% to 100% and a specificity of 88%–96% [12,13]. In the current issue, Sax et al [14] report on the use of plasma β-glucan for diagnosing PCP in a cohort of 282 HIV-infected individuals enrolled in ACTG A5164, a study of early ART for acute opportunistic infections. Of the 252 participants in this cohort with analyzable β-glucan levels, 173 had PCP, making this the largest study of use of β-glucan for diagnosis of PCP. The study investigators used the Fungitell Assay (Associates of Cape Cod) with a positive threshold of ≥80 pg/mL. They found that individuals with PCP had a higher β-glucan level and that the test had a sensitivity of 92% and specificity of 65% with the threshold that they selected. Detection of β-glucan was not affected by treatment and did not seem to correlate with disease severity as measured by use of concomitant corticosteroids. β-glucan levels overlapped between those with and without PCP, and results of the test were also positive in other fungal infections. What do these results mean for clinicians caring for patients with suspected PCP? Can we send a plasma sample for β-glucan and avoid the use of bronchoscopy? Testing for β-glucan is an attractive option, because it can be done on peripheral blood samples, it is cheaper than other testing modalities, and it is not heavily dependent on highly trained personnel. β-glucan has several major disadvantages. First, β-glucan is positive with certain other fungi; thus, results are nonspecific. This could be a particular problem in areas where, for instance, pulmonary histoplasmosis is common. Second, the assay is not available on site at many hospitals. Thus, the use of this test could lead to inappropriate therapy, because a substantial number of patients might be treated for PCP even though another pathogen was the causative agent. This study also has some methodological challenges. First, the PCP group included a large number of individuals with concomitant oral candidiasis, which might have influenced results. Second, Pneumocystis colonization (detection of the organism or its DNA in patients without acute pneumonia) is common in pulmonary secretions of HIV-infected individuals with low CD4 cell counts who either have no pulmonary disease or who have disease due to a process other than PCP. The authors did set a threshold to distinguish colonization from disease, and that may in fact be a valid way of distinguishing colonization from disease. Clinicians need to be aware that a low positive β-glucan does not have the same diagnostic significance as a high value. A recent study that included 8 HIV-uninfected patients with Pneumocystis colonization found that β-glucan levels were <80 pg/mL in most of these patients, but 2 patients had minimally positive levels [15]. Lastly, this study does not include HIV-uninfected patients in whom the burden of pneumocystis may be lower, suggesting that the β-glucan levels in patients with true PCP may be lower than those reported here, and different thresholds may be necessary. Serum β-glucan provides an interesting window on the biology of fungal pneumonia and the presence of circulating fungal cell wall components. A serologic test for PCP would be a welcomed addition to our ability to diagnose PCP. However, this serum β-glucan test will not yet allow us to lay down the bronchoscope. Potential conflicts of interest.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Morris et al. (2011) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: