Background: Quantitative PCR to detect Pneumocystis jirovecii (Pj) is a new tool for the diagnosis of Pneumocystis jirovecii pneumonia (PJP). The yield of this technique, in cases of low fungal burden, when the standard technique using immunofluorescence (IF) is negative, needs to be evaluated. Methods: We retrospectively reviewed the charts of all patients with a positive PCR but negative IF test (PCR+/IF-) in bronchoalveolar lavage (BAL) fluid performed over one year. We used an algorithm based on underlying immunosuppression, clinical picture, thoracic CT scan appearances, existence of an alternative diagnosis and the patient's outcome on treatment. Using this, each case was classified as probable PJP, possible PJP or colonization. Results: Among the 416 BAL performed, 48 (12%) were PCR+/IF- and 43 patients were analyzed. Patients were mostly male (56%) with a median age of 60 years. Thirty-five (84%) were immunocompromised: 4 (9%) HIV-infected patients, 26 (60%) with hematologic or solid organ cancer, 3 (7%) were renal transplant recipients. Seven (16%) were classified as probable PPJ and 9 (21%) as possible PJP. Patients with a probable or possible PJP were more frequently admitted to the ICU (P<0.02) and had higher risk of death (P<0.01) when compared to those with colonization. Median PCR levels were very low and were not different between PJP or colonized patients (P=0.23). Conclusions: Among patients with a positive Pj PCR in BAL but with negative IF, only 37% had probable or possible PJP and PCR could not discriminate PJP from colonization.