Unfavorable outcomes to second-line tuberculosis therapy among HIV-infected versus HIV-uninfected patients in sub-Saharan Africa: A systematic review and meta-analysis

by Dumessa Edessa, Mekonnen Sisay, Yadeta Dessie

Background

Drug resistance is a key obstacle to the global target set to end tuberculosis by 2030. Clinical complexities in drug-resistant tuberculosis and HIV-infection co-management could worsen outcomes of second-line anti-tuberculosis drugs. A comprehensive estimate for risks of unsuccessful outcomes to second-line tuberculosis therapy in HIV-infected versus HIV-uninfected patients is mandatory to address such aspects in segments of the target set. Therefore, this meta-analysis was aimed to estimate the pooled risk ratios of unfavorable outcomes to second-line tuberculosis therapy between HIV-infected and HIV-uninfected patients in sub-Saharan Africa.

Methods

We conducted a literature search from PubMed/MEDLINE, EMBASE, SCOPUS and Google Scholar. We screened the retrieved records by titles and abstracts. Finally, we assessed eligibility and quality of full-text articles for the records retained by employing appraisal checklist of the Joanna Briggs Institute. We analyzed the data extracted from the included studies by using Review Manager Software, version 5.3 and presented our findings in forest and funnel plots. Protocol for this study was registered on PROSPERO (ID: CRD42020160473).

Results

A total of 19 studies with 1,766 from 4,481 HIV-infected and 1,164 from 3,820 HIV-uninfected patients had unfavorable outcomes. The risk ratios we estimated between HIV-infected and HIV-uninfected drug-resistant tuberculosis patients were 1.18 (95% CI: 1.07–1.30; I² = 48%; P = 0.01) for the overall unfavorable outcome; 1.50 (95% CI: 1.30–1.74) for death; 0.66 (95% CI: 0.38–1.13) for treatment failure; and 0.82 (95% CI: 0.74–0.92) for loss from treatment. Variable increased risks of unfavorable outcomes estimated for subgroups with significance in mixed-age patients (RR: 1.22; 95% CI: 1.10–1.36) and eastern region of sub-Saharan Africa (RR: 1.47; 95% CI: 1.23–1.75).

Conclusions

We found a higher risk of unfavorable treatment outcome in drug-resistant tuberculosis patients with death highly worsening in HIV-infected than in those HIV-uninfected patients. The risks for the unfavorable outcomes were significantly higher in mixed-age patients and in the eastern region of sub-Saharan Africa. Therefore, special strategies that reduce the risks of death should be discovered and implemented for HIV and drug-resistant tuberculosis co-infected patients on second-line tuberculosis therapy with optimal integration of the two programs in the eastern region of sub-Saharan Africa.