|
Title: |
|
Authors:
|
|
Abstract: Background: Pulmonary tuberculosis in the context of HIV infection produces diverse radiographic appearances that shift according to the degree of immunological compromise. In this Nigerian cohort, we sought to quantify how CD4 T-cell depletion shapes the pattern and severity of chest radiograph abnormalities using a structured numerical scoring approach. Methods: Between March 2019 and February 2020, we prospectively enrolled 242 consecutive HIV-positive adults with suspected pulmonary TB at DELSUTH, Oghara, Nigeria. Two consultant radiologists, blinded to all clinical and laboratory data, independently scored each poster anterior chest radiograph using the Chest Radiographic Severity Score (CRSS), a 0-20 ordinal instrument adapted from the Chest Radiograph Reading and Recording System. We examined the correlation between CD4 count and CRSS total score using Spearman rank correlation, compared scores across CD4 strata with Kruskal-Wallis testing, evaluated discriminatory performance with ROC analysis, and adjusted for confounders using multivariable linear regression. Results: A robust inverse gradient emerged: mean CRSS total score declined from 11.1 ± 2.8 in participants with CD4 <100 cells/µL to 8.8 ± 2.4 (CD4 100-199), 6.9 ± 2.0 (CD4 200-349), and 4.5 ± 1.5 (CD4 ≥350). The Spearman correlation between CD4 count and CRSS total score was rho = -0.754 (95% CI: -0.804, -0.689; p < 0.01). Those with severe immunosuppression predominantly exhibited pleural effusion, intrathoracic lymphadenopathy, and miliary shadowing, while participants with preserved immunity showed cavitation. Agreement between the two independent readers was substantial (Cohen weighted kappa = 0.903; ICC = 0.972). The CRSS total score distinguished CD4 <200 from CD4 ≥200 cells/µL with an AUC of 0.862 (95% CI: 0.815-0.905). Conclusion: Among Nigerian adults with HIV-associated pulmonary TB, CD4 count demonstrated a strong inverse relationship with radiographic severity as captured by the CRSS. DOI: http://dx.doi.org/10.51505/ijmshr.2026.10404 |
|
PDF Download |