Mycobacterium tuberculosis bloodstream infection prevalence, diagnosis, and mortality risk in seriously ill adults with HIV: a systematic review and meta-analysis of individual patient data

Background

The clinical and epidemiological significance of HIV-associated Mycobacterium tuberculosis blood stream infection (MTB-BSI) is incompletely understood. We hypothesised that MTB-BSI prevalence has been underestimated, that MTB-BSI independently predicts death, and sputum Xpert has suboptimal diagnostic yield for MTB-BSI.

Methods

We conducted a systematic review and individual patient data (IPD) meta-analysis of studies performing routine mycobacterial blood-culture (TBBC). Harmonised inclusion criteria were applied to IPD: age ≥13 years, HIV-positivity, available CD4 count, a valid mycobacterial blood culture result (excluding patients with missing data from lost or contaminated blood cultures), and meeting WHO definitions for suspected tuberculosis (presence of screening symptom). Predicted probabilities of MTB-BSI from mixed-effects modelling were used to estimate prevalence. Estimates of diagnostic yield of sputum (Xpert or culture if Xpert unavailable) and urine-lipoarabinomannan for MTB-BSI were obtained by two-level random-effect meta-analysis, mortality hazard of MTB-BSI by mixed-effect Cox proportional-hazard modelling, and effect of treatment delay with propensity-score analysis. PROSPERO registration: CRD42016050022·

Findings

We identified 23 data sets for inclusion (20 published and 3 unpublished at time of search), and obtained data from 20, representing 96% of eligible IPD. 5751 patients met inclusion criteria. Predicted probability of MTB-BSI was 45% (95%CI 38-52%) for danger-sign positive tuberculosis inpatients with cohort median CD4 count of 76 cells/L. Diagnostic yield of sputum was 77% (95%CI 63–87%), rising to 89% (95%CI 80-94%) when combined with urine-lipoarabinomannan testing. Presence of MTB-BSI compared to absence in patients with HIV-associated tuberculosis increased hazard of death before 30-days (aHR 2·5, 95%CI 2·1–3·1) but the effect waned after 30 days (aHR 1.25 95% 0.84-2.49). In a propensity-score matched cohort of participants with HIV-associated tuberculosis (n=630), mortality increased with anti-tuberculosis treatment delay >4 days in the subgroup with MTB-BSI (OR 3·2, 95%CI 1·1–10·3) but not convincingly in the overall cohort (OR 1.25 95% CI 0.68 – 2.25).

Interpretation

In critically-ill adults with HIV-tuberculosis, MTB-BSI is a frequent manifestation of tuberculosis and strongly predicts mortality within 30 days. Better diagnostic yield in patients with MTB-BSI can be achieved by parallel use of sputum-Xpert and urine-LAM. Anti-tuberculosis treatment delay may increase mortality hazard.