Outcome of Pneumocystis jirovecii pneumonia diagnosed by polymerase chain reaction in patients without human immunodeficiency virus infection EMILIA HARDAK,1* AMI NEUBERGER,2* MORDECHAI YIGLA,1,4 GIDON BERGER,1 RENATO FINKELSTEIN,2,4 HANNAH SPRECHER3 AND ILANA OREN2,4 1
Division of Pulmonary Medicine, 2Infectious Diseases Unit and 3Microbiology Laboratory, Rambam Health Care Campus, Haifa, and 4Bruce Rappaport Faculty of Medicine, Technion, Israel Institute of Technology, Haifa, Israel
SUMMARY AT A GLANCE
Background and objective: Pneumonia caused by Pneumocystis jirovecii (PCP) in patients without human immunodeficiency virus (HIV) infection is associated with high mortality. The diagnosis of PCP at our institution is based on detection of DNA using a polymerase chain reaction (PCR) assay. The aim of this study was to describe the clinical manifestations, outcomes and factors associated with mortality due to PCP, as diagnosed by PCR, in patients without HIV infection. Methods: Over a 6-year period, all HIV-negative immunocompromised patients suspected of having an opportunistic pulmonary infection underwent diagnostic bronchoscopy. A multigene PCR assay that detects Pneumocystis jirovecii DNA was used for the diagnosis of PCP. Patients were considered to have PCP if they had underlying immunodeficiency, compatible signs and symptoms, abnormal radiological findings, and Pneumocystis jirovecii DNA was detected in a bronchoalveolar lavage fluid sample. Data was collected retrospectively. Results: PCP was diagnosed in 58 patients. The underlying conditions included haematological malignancies (60.3%), solid tumours (17.2%) and immunosuppressive treatment (22.4%). The most common clinical features in patients with PCP were fever (94.6%), dyspnoea (67.2%) and cough (36.2%). The overall in-hospital mortality was 17.2% (10/58). Mortality was associated with co-infections, high lactate dehydrogenase levels, female gender, and higher
Pneumocystis pneumonia (PCP) is associated with a high mortality in HIV-negative patients. In this study, mortality among these patients was lower than that reported previously; this may be associated with the use of a PCR-based assay for the early diagnosis of PCP.
Correspondence: Emilia Hardak, Rambam Health Care Campus, PO Box 9602, Haifa 31096, Israel. Email: [email protected]
rambam.health.gov.il *Emilia Hardak and Ami Neuberger contributed equally to this work and should be considered as first co-authors. Received 28 July 2011; invited to revise 30 August 2011, 26 November 2011; revised 8 November 2011, 1 December 2011; accepted 9 December 2011 (Associate Editor: Marcos Restrepo). © 2012 The Authors Respirology © 2012 Asian Pacific Society of Respirology
pneumonia severity index and acute physiology and chronic health evaluation III scores. Conclusions: In this study, the mortality of HIVnegative patients with PCP was low compared with previous reports. We hypothesize that this finding resulted from the increased sensitivity of a PCR-based assay, as compared with traditional methods, for the diagnosis of PCP in HIV-negative patients. Key words: diagnosis, immunocompromised, mortality, pneumocystis, pneumonia.
INTRODUCTION Pneumocystis jirovecii causes pneumocystis pneumonia (PCP) in immunocompromised patients. Although PCP is commonly associated with human immunodeficiency virus (HIV) infection and low circulating CD4+ T lymphocyte counts, it also occurs in patients with other causes of immunodeficiency, including treatment with corticosteroids, haematological and solid organ malignancies, stem cell transplantation, solid-organ transplantation, and autoimmune diseases.1 A number of studies have described the risk factors, clinical manifestations and outcomes of PCP in patients without HIV infection.2–13 In HIV-negative patients, PCP tends to follow a more acute course, and the mortality rate for patients with PCP has been reported to be 19.6–52.9%,2,5,7,8,10–13 Respirology (2012) 17, 681–686 doi: 10.1111/j.1440-1843.2012.02158.x
682 which is considerably higher than the reported mortality among HIV-infected patients with PCP.14 Most previous studies have used a variety of staining techniques for the diagnosis of PCP. However, such tests have a lower sensitivity for the diagnosis of PCP in patients without as compared with those with HIV infection. This might be explained by the decreased numbers of Pneumocystis organisms in bronchoalveolar lavage (BAL) fluid specimens from patients without HIV infection.15 The diagnosis of PCP at our institution is based on detection of DNA using a polymerase chain reaction (PCR) assay.16 When compared with traditional cytological stains, PCR detection of P. jirovecii DNA provides greater sensitivity for the detection of PCP.16–18 On the other hand, a PCR-based assay carries the risk of false-positive diagnosis of PCP in patients who are colonized, but not infected, with P. jirovecii.19 In the light of these changes in the diagnosis of PCP, we aimed to describe the clinical manifestations, outcomes and factors associated with mortality due to PCP, as diagnosed by PCR, in patients without HIV infection.
METHODS A retrospective, single-centre study was conducted in a 900-bed tertiary care medical centre. Over a 6-year period, from 2005 to 2010, all consecutive immunocompromised patients (excluding HIV-infected patients) suspected of having an opportunistic pulmonary infection (with PCP being one of the differential diagnoses) underwent bronchoscopy with BAL sampling to evaluate the aetiology of pulmonary infiltrates. Immunocompromised patients were defined as patients with underlying malignancies (haematological or solid tumours), transplant recipients (bone marrow or solid organ), patients receiving immunosuppressive treatment for autoimmune diseases and patients treated with high-dose corticosteroids (>15 mg/day for >2 months). The study was approved by the institutional ethics committee. Demographical and clinical details were retrieved from the hospital’s computerized record system. For each patient, underlying conditions, PCP prophylaxis, previous episodes of PCP, clinical symptoms and time from clinical onset to diagnosis were recorded. The respiratory status of the patients was defined as abnormal when either supplementary oxygen or respiratory support was required. To assess comorbid conditions and severity of illness, the Charlson comorbidity index and the pneumonia severity index were evaluated at admission. The acute physiology and chronic health evaluation III score was used to measure the severity of disease in patients requiring ventilatory support, and the duration of mechanical ventilation was recorded. The duration and treatment of PCP, and the presence and treatment of co-infections were also recorded. Clinical outcomes, in-hospital mortality and 1-year mortality were assessed. Laboratory processing of BAL samples included staining and culture for bacteria, fungi, Mycobacteria Respirology (2012) 17, 681–686
E Hardak et al.
and Legionella sp., and PCR assays for the detection of Aspergillus, Legionella and Pneumocystis DNA. Virological studies included viral culture, immunofluorescence assay for respiratory viruses (influenza viruses A and B, respiratory syncytial virus, parainfluenza viruses types 1, 2 and 3), and PCR assays for the detection of cytomegalovirus, herpes simplex virus, Epstein–Barr virus, varicella-zoster virus, respiratory syncytial virus, adenovirus, influenza virus and parainfluenza virus DNA. Stains for the detection of fungal elements, viral inclusion bodies in alveolar cells, and Giemsa and methenamine silver stains for the detection of Pneumocystis bodies were performed on all samples. All slides were examined by two experienced medical professionals. Bronchoscopy and BAL were performed according to existing guidelines.20 The beta-D-glucan test was not available at our centre. A PCR method for the detection of P. jirovecii DNA was developed at our institution, and showed a sensitivity of 74% and specificity of 95% for the diagnosis of PCP in immunocompromised patients without HIV infection.16 In order to reduce the risk of false-positive results, three distinct P. jirovecii genes were targeted: the mitochondrial large subunit ribosomal RNA gene, the major surface glycoprotein gene and the T1-T2 region of the large subunit ribosomal RNA gene (28S ribosomal RNA). Precautions against contamination of the PCR included the use of barrier methods, extraction and amplification of negative control samples. PCR results were considered positive if amplification of at least two of the three genes gave positive results. Patients were considered to have PCP if they had an underlying immunodeficiency known to be associated with PCP, clinical findings suggestive of a lower respiratory tract infection, new pulmonary infiltrates on chest radiography or computed tomography, and P. jirovecii DNA was detected in a BAL fluid sample with or without a positive direct staining result. Descriptive statistics, including means, standard deviation and frequencies, were calculated. Univariate binary logistic regression analysis was performed, with calculation of odds ratios, 95% confidence intervals and P-values. Two-tailed P-values ⱕ0.05 were considered statistically significant. Statistical analyses were performed using SPSS 18.0 software for Windows (SPSS Inc., Chicago, IL, USA).
RESULTS Over the 6-year study period, 58 immunocompromised patients with a negative serology test for HIV were diagnosed as having PCP. The baseline characteristics of these patients are presented in Table 1. There were 30 (52%) females and 28 (48%) males, with a mean age of 55.3 ⫾ 14.3 years. Most of the immunocompromised patients with PCP had haematological malignancies (60.3%). Eight patients, four of whom had severe graft-versus-host disease, were bone marrow transplant recipients. Ten patients (17.2%) had received chemotherapy for various solid tumours, and most patients with solid tumours (7/10) © 2012 The Authors Respirology © 2012 Asian Pacific Society of Respirology
Outcome of PCP in patients without HIV Table 1 Demographical details and underlying diseases for the immunocompromised patients Age (years), mean ⫾ SD Gender, males/females Underlying diseases Haematological malignancies, n (%) Acute myeloblastic leukaemia Acute lympoblastic leukaemia Chronic lympocytic leukaemia Lymphoma Multiple myeloma Solid tumours, n (%) Rectal carcinoma Breast carcinoma Glioblastoma Immunosuppression due to non-malignant diseases, n (%) Systemic lupus erythematosus Wegener’s granulomatosis Autoimmune hepatitis Renal transplantation Inflammatory bowel disease Bone marrow transplant recipients, n (%) Allogeneic Autologous Charlson comorbidity index, median (range) Pneumonia severity index, median (range) TMP/SMX prophylaxis, n (%)
55.9 ⫾ 14.3 28/30 35 (60.3) 8 4 4 17 2 10 (17.2) 1 4 5 13 (22.5) 1 1 2 8 1 8 (14) 6 2 4 (2–10) 94 (52–167) 3 (5.2)
SD, standard deviation; TMP/SMX, trimethoprim sulphamethoxazole.
had also received high-dose corticosteroids. Five of the patients with solid tumours had glioblastoma multiforme and were treated with high-dose corticosteroids. Thirteen patients (22.5%) received immunosuppressive, corticosteroid-based treatment for prevention of allograft rejection or for autoimmune disease. Three allogeneic bone marrow transplant recipients developed PCP while receiving trimethoprim-sulphamethoxazole prophylaxis of two double-strength tablets biweekly. None of the patients had a previous episode of PCP. The median Charlson comorbidity index was 4 (range 2–10). The median pneumonia severity index on admission was 94 (range 52–167). Eleven patients had respiratory failure requiring mechanical ventilation, and a further six patients were treated by non-invasive positive pressure ventilation. Overall, 17 patients (29%) required ventilatory support; five of these patients were hospitalized in the intensive care unit. The median time from admission to respiratory support was 2 days (range 1–12), and the median duration of mechanical ventilation was 7 days (range 1–180). Pneumothorax was not detected in any of the patients. The median acute physiology and chronic health evaluation III score on day 1 was 72 (range 33–102). The median time interval between the onset of symptoms and diagnosis was 11 days (range 2–30); it was shorter than 3 days in three patients and longer than 2 weeks in nine patients. The most common © 2012 The Authors Respirology © 2012 Asian Pacific Society of Respirology
clinical features in patients with PCP were fever (94.6%), dyspnoea (67.2%) and cough (36.2%). Chest computed tomography scans were performed in 52 patients; a further six patients only underwent plain chest radiography. All computed tomography scans were abnormal, with diffuse bilateral ground-glass attenuation being the most common finding (46/52, 88%). In computed tomography scans from a further six patients, abnormal findings included nodular infiltrates (n = 2), alveolar infiltrates (n = 2) and findings consistent with acute respiratory distress syndrome (n = 2). Giemsa and methenamine silver stains for the detection of Pneumocystis bodies were positive in 5/58 (8.6%) of the patients who were diagnosed with PCP by PCR. Co-pathogens were identified in BAL fluid specimens from 11 patients. Six patients were diagnosed with concomitant probable invasive pulmonary aspergillosis, as detected by the presence of galactomannan antigen or culture of Aspergillus in BAL fluid. Three patients were diagnosed with concomitant cytomegalovirus pneumonitis, as detected by positive results for cytomegalovirus DNA in BAL and blood specimens. One patient was diagnosed with concomitant herpes simplex virus pneumonitis, as detected by a positive result for herpes simplex virus type 1 DNA in BAL fluid, and one patient was diagnosed with both concomitant pulmonary aspergillosis and cytomegalovirus pneumonitis. Before PCR results were available, 31/58 (53.4%) patients with PCP were empirically treated with trimethoprim-sulphamethoxazole. In addition, patients were empirically treated with different combinations of anti-infective agents, including pipercillin-tazobactam (n = 15), imipenem (n = 10), levofloxacin (n = 29), azythromycin (n = 6), and voriconazole (n = 17). All patients were treated with trimethoprim-sulphamethoxazole after the diagnosis of PCP, and treatment with other anti-infective agents was withheld, except for patients with co-infections. The latter patients were treated with additional voriconazole, gancyclovir and acyclovir, according to the co-infection that was diagnosed. Anti-Pneumocystis treatment was administered for 21 days, with 23 patients receiving secondary prophylaxis thereafter. Adjuvant corticosteroid therapy (40 mg twice daily for 4 days) was administered to 44 patients with partial pressure of arterial oxygen