Objectives
Performance of chest X-ray (CXR) features for PCP diagnosis has been evaluated in small studies. We conducted a systematic review and meta-analysis to describe CXR changes in adults with HIV-associated laboratory-confirmed PCP, comparing these to non-PCP respiratory disease.
Methods
We searched databases for studies reporting CXR changes in people > 15 years old with HIV and laboratory-confirmed PCP and those with non-PCP respiratory disease. CXR features were grouped using consensus terms. Proportions were pooled and odds ratios (OR) generated using random effects meta-analysis, with subgroup analyses by CD4 count, study period, radiology review method, and study region.
Results
51 studies (with 1,821 PCP and 1,052 non-PCP cases) were included. Interstitial infiltrate (59%, 95% CI, 52-66%; 36 studies, n = 1,380, I2 85%) and ground-glass opacification (48%, 95% CI, 15-83%; 4 studies, n = 57, I2 86%) were common in PCP. Cystic lesions, central lymphadenopathy and pneumothorax were infrequent. Pleural effusion was rare in PCP (0% [95% CI, 0-2%]. Interstitial infiltrate (OR 2.3; 95% CI, 1.4-3.9; I2 = 60%); interstitial-alveolar infiltrate (OR 10.2 95% CI, 3.2-32.4; I2 = 0%); and diffuse CXR changes (OR 7.3; 95% CI, 2.7-20.2; I2 = 87%) were associated with PCP diagnosis. There was loss of association with alveolar infiltrate in African studies.
Conclusions
Diffuse CXR changes and interstitial-alveolar infiltrates indicate a higher likelihood of PCP. Pleural effusion, lymphadenopathy, or focal alveolar infiltrates suggest alternative causes. Findings could be incorporated into clinical algorithms to improve diagnosis of HIV-associated PCP.