Metagenomic next-generation sequencing to detect Pneumocystis jirovecii pneumonia in critically ill, HIV-negative children: a retrospective multicenter study

TL;DR

This study evaluates metagenomic sequencing to detect Pneumocystis jirovecii pneumonia in HIV-negative children and identifies factors that distinguish infection from colonization.

Contribution

The study introduces a potential method using mNGS read thresholds to differentiate P. jirovecii infection from colonization in non-HIV pediatric patients.

Findings

Higher mNGS read counts and serum C-reactive protein levels were associated with P. jirovecii pneumonia versus colonization.

A threshold of 556 reads showed 77.6% sensitivity and 100% specificity for distinguishing infection from colonization.

Younger age, lower T-cell counts, and primary immunodeficiency were linked to higher mortality in P. jirovecii pneumonia cases.

Abstract

Metagenomic next-generation sequencing (mNGS) plays a critical role in the rapid detection of infectious pathogens. We aimed to analyze the clinical characteristics of Pneumocystis jirovecii infection in children without HIV infection and to evaluate the performance of mNGS in distinguishing P. jirovecii colonization from true infection.

A multicenter, retrospective analysis was conducted on critically ill, non-HIV-infected pediatric patients who tested positive for P. jirovecii via mNGS analysis of bronchoalveolar lavage fluid (BALF). Group differences were assessed using Mann–Whitney U-tests (for continuous data) and chi-square tests (for categorical data). Discriminatory performance was evaluated by calculating the area under the receiver operating characteristic curve.

A total of 59 HIV-negative children (age range: 2 months to 14 years) from four children’s hospitals were included and classified into two groups based on P. jirovecii status: P. jirovecii pneumonia (PCP; n = 51) and P. jirovecii colonization (PCC; n = 8). Compared with the PCC group, the PCP group had significantly higher serum C-reactive protein levels and median P. jirovecii read counts in mNGS (both P < 0.05). The optimal threshold value for discriminating P. jirovecii infection from colonization appeared to be 556 reads (sensitivity, 77.6%; specificity, 100.0%). Eighteen patients (35.3%) in the PCP group died. Compared with survivors, these patients were significantly younger, had lower T-cell subset counts (CD3+, CD4+, and CD8+), and a higher prevalence of primary immunodeficiency (all P < 0.05).

BALF mNGS analysis may have utility for differentiating between colonization and infection by P. jirovecii, warranting further investigation.

The online version contains supplementary material available at 10.1186/s12890-026-04163-9.

Figures