Impact of the Use of Polymerase Chain Reaction for the Diagnosis and Management of Pneumocystis Jirovecii Pneumoniae in a Retrospective Cohort of Patients with Lymphoid Malignancies

Introduction: Pneumocystis jirovecii (Pj) is an opportunistic fungus and a common cause of pneumonia (PCP) in immunocompromised patient. The gold standard for the diagnosis is the direct exam by microscopy. Patients with hematological malignancies may have a small burden of Pj with a negative direct exam. A polymerase chain reaction assay (PCR) can be used for the diagnosis. Real-time qPCR methods have a good sensitivity (97-100%) and specificity (90-94%) for the PCP detection whereas direct examination has lower sensitivity. Currently, the diagnostic impact of a positive PCR with a negative direct examination for patients with lymphoid malignancies suspected of PCP is not well known.

Methods: We retrospectively analysed in this study all patients followed in the Department of Hematology of the Leon Berard Cancer Center (Lyon, France) between January 2003 and March 2014 who developed PCP. In at risk patients, PCP was considered when clinical signs and typical radiological features (e.g. ground glass opacities). Diagnosis was confirmed by the microscopic observation of Pj cysts and/or trophic forms on the bronchoalveolar lavage fluid (Group 1, gold standard) or a positive PCR with negative direct examination (Group 2).

Results: PCP was diagnosed in 68 patients with lymphoid malignancies (76% non-Hodgkin lymphoma, 15% Hodgkin lymphoma, 7% chronic lymphocytic leukemia and 2% myeloma). PCP was diagnosed during first-line therapy for 42 patients (62%). Seven breakthrough infections (10%) under sulfamethoxazole and trimethoprim (S/T) prophylaxis were observed. PCP was confirmed in 27 patients (40%) by a positive direct exam (Group 1) and 41 patients (60%) by PCR only (Group 2). The clinical and biological characteristics at diagnosis of PCP were not significantly different between Group 1 and 2 (fever, cough, lymphocyte count, neutrophil count, hemoglobin level, C-reactive protein level, lactate deshydogenase level). However, Group 1 presented more frequently an oxygen saturation of less than 92% (88.9% vs. 56.1%, P=0.007) and dyspnea (88.9% vs. 65.9%, P=0.045) as compared to Group 2. The CT-scan showed bilateral ground-glass opacities in 74% and 71% of Group 1 and 2, respectively. Median time between the hematological diagnosis (3.02 months vs. 4.7 months, P=0.54) or the last chemotherapy (20.5 days vs. 21 days, P=0.62) and PCP diagnosis were similar in both groups. The median time between the first day of hospitalization and the initiation of curative treatment by S/T was similar in the two groups (6 days [0-30] vs. 7 days [0-33], P=0.71) but Group 1 received corticosteroids more frequently as compared to Group 2 (74% vs. 49%, P=0.047). We observed a similar rate of co-infections between the two groups (Group 1, 44% vs. Group 2, 39%, P=0.80). After the initiation of S/T, the median time to fever resolution (2 days [1-21] vs. 1 days [0-8], P=0.006) and to normal oxygen saturation (6 days [2-28] vs. 3 days [0-11], P=0.002) was longer for Group 1 as compared to Group 2. Hence, the median time between S/T initiation and hospital discharge was longer for Group 1 (16 days [4-53] vs. 10 days [2-44], P=0.02). The rate of hospitalization in intensive care unit was higher for Group 1 than for Group 2 (59% vs. 24%, P=0.005). In Group 1 and 2, 5 (18.5%) and 3 (7.3%) patients died from PCP, respectively (P=0.25).

Conclusions: In our series, 60% of immunocompromised patients with lymphoid malignancies were diagnosed with PCP on the basis of a positive PCR and a negative direct examination. The presentation and clinical course of these patients are less severe than patients with a positive direct exam, possibly in relation with a lower fungal load. Quick improvement of clinical parameters was observed after curative S/T. These results suggest to promptly initiate a specific treatment for PCP in patients with a positive PCR and a negative direct examination.