CONTEMPORARY CHEST RADIOGRAPHIC MANIFESTATIONS OF PULMONARY TUBERCULOSIS INFECTION WITH HIV/AIDS COINFECTION : A SYSTEMATIC REVIEW OF RANDOMIZED CONTROLLED TRIALS AND PRIMARY STUDIES

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Introduction: Tuberculosis (TB) remains the leading infectious cause of death among people living with HIV (PLHIV), with chest radiography (CXR) as the primary imaging modality in high-burden settings. Progressive CD4+ T-lymphocyte depletion disrupts granulomatous containment, producing "atypical" radiographic patterns including lymphadenopathy, lower-zone consolidation, miliary dissemination, and normal radiographs. This systematic review synthesises CXR manifestations of pulmonary TB in HIV/AIDS coinfection across fourteen outcome domains. Methods: Systematic review per PRISMA 2020 guidelines. Eligible designs included RCTs, cohort, case-control, cross-sectional, and observational primary studies. Risk of bias assessed with RoB 2, Newcastle-Ottawa Scale, JBI checklist, and QUADAS-2. Narrative synthesis with tabulated effect estimates performed. Results: Thirty-four primary studies (>55,000 participants) met inclusion criteria: 5 randomised/cluster-randomised trials and 29 observational studies. Cavitation was significantly less frequent with advanced immunosuppression: 32% vs 68% in CD4 <200 vs ≥200 cells/µL (P=.008); HIV independently reduced adjusted odds of cavities (aOR 0.34, 95% CI 0.13-0.85; P=.02). Consolidation: 47% vs 63% (P=.002; aOR 0.30 for any opacity). Lymphadenopathy more frequent at low CD4: 30% vs 7% (P=.01). Miliary pattern increased with immunosuppression: 64% vs 36% (P=.04). Normal/non-suggestive CXR occurred in 14-32% of bacteriologically confirmed cases. HIV was sole independent predictor of atypical radiographic appearance (OR 0.20 for typical pattern, 95% CI 0.13-0.31). Computer-aided detection sensitivity fell by 13.4 absolute percentage points in PLHIV for two of three algorithms (95% CI -21.1 to -6.9). Radiographic deterioration occurred in 53.4% of paradoxical TB-IRIS events; corticosteroids significantly improved CXR at weeks 2 (P=.002) and 4 (P=.02). Discussion: Radiographic phenotype in HIV/TB coinfection reflects quantitative immunological deficit. Cavitation requires intact granulomatous response; as CD4+ declines, adenopathic, disseminated, and radiographically silent phenotypes predominate. Molecular evidence confirms altered host immunity, not recent infection, drives this shift. Normal CXR does not exclude active TB in PLHIV; radiographic gatekeeping is unsafe. Computer-aided detection requires HIV-stratified calibration. CXR retains value for severity, monitoring, and prognosis but must be abandoned as rule-out test. Conclusion: CXR in HIV/TB coinfection shifts predictably away from cavitary post-primary disease toward adenopathic, miliary, and silent phenotypes with CD4 depletion, degrading diagnostic sensitivity. Every PLHIV investigated for TB should undergo molecular testing irrespective of CXR appearance. Radiological reports should state CD4 stratum and avoid "no evidence of active TB"; CAD thresholds require local HIV-stratified calibration.