Diagnostic accuracy of stool xpert MTB/RIF ultra assay for pulmonary tuberculosis among PLHIV in Kampala, Uganda

Abstract

Tuberculosis (TB) continues to be the leading cause of health complications and fatal outcomes among PLHIV. Current WHO-recommended strategies for diagnosing pulmonary tuberculosis among PLHIV faces challenges of poor or absolute lack of sputum in this population, hence, low sensitivity. Alternative samples are therefore being sought. We aimed to evaluate how accurately the stool Xpert MTB/RIF Ultra assay detects Mycobacterium tuberculosis in adult individuals living with HIV. Methods: This was a cross-sectional diagnostic accuracy study among 139 individuals with presumptive PTB, aged ≥18 years, male and female PLHIV, attending Mengo Hospital and Kiswa Health Centre III, Kampala Uganda. One stool sample and one routine spot sputum sample were collected from each participant, processed on the GeneXpert PCR System, brand name (Ultra; Cepheid, Sunnyvale,CA,USA). Between August 15, 2025, and October 15, 2025, 139 participants were enrolled (78 [55.7%] female and 62 [44.3%] men). PTB was confirmed using the molecular assay in 7 participants (5%). The sputum and stool Ultra tests demonstrated similar sensitivity and specificity. A higher positivity rate (57.1%) was observed in males aged 24–44 years relative to females in the corresponding age group. These results point to the potential role of stool Ultra as an additional diagnostic approach for tuberculosis in people with HIV. Although early morning sputum samples typically contain higher bacillary loads than spot samples, our study utilized spot sputum specimens as they were more practical for routine same-day collection at the study sites. Additionally, total CD4 count and HIV viral load measurements were not obtained due to financial and time constraints. These unmeasured variables may have influenced the diagnostic performance (sensitivity and specificity) of the stool Xpert MTB/RIF Ultra assay, representing an important limitation of our study. Future research should include early morning sputum samples and incorporate CD4 and viral load measurements to improve diagnostic accuracy and control for host-related confounders.