Health and Social Factors Associated with Adverse Treatment Outcomes Among People with Multidrug-Resistant Tuberculosis in Sierra Leone: A National, Retrospective Cohort Study

Background: Multidrug-resistant Tuberculosis (MDR-TB) is a global health emergency. In Sierra Leone, a high TB burden country, we evaluated health and social factors associated with adverse MDR-TB treatment outcomes.  Methods: This national, retrospective cohort study recruited all people notified with MDR-TB to the Sierra Leone National TB Program between April 2017 and September 2019, and with follow up to May 2021. MDR-TB was defined programmatically as rifampicin or rifampicin/isoniazid resistance. Treatment was with the 2017 WHO-recommended short (9–11 month) or long (18-24 month) regimens, both of which contained injectable aminoglycosides. An imputed multivariable logistic regression model examined associations of programmatic sociodemographic and clinical data with WHO-defined adverse treatment outcomes (death, treatment failure, loss-to-follow-up). A supplementary unimputed logistic regression model examined the association of direct susceptibility testing (DST) results with adverse outcome.  Findings: Of notified MDR-TB cases, 365/370 (99%) were eligible to participate. Median age was 35 years (interquartile range 26-45), 263/365 (72%) were male, 71/365 (19%) HIV-positive, and 127/365 (35%) severely underweight (body mass index <16.5 units). DST was available for 315/365 (86%) participants. Treatment was started by 341/365 (93%), 317/341 (93%) and 24/341 (7%) with the short vs long regimen respectively. Overall, 267/365 (73%) had treatment success, of whom 254/317 (80%) and 13/24 (54%) were on the short vs long regimen; 95/365 (26%) had an adverse outcome, and 3/365 (1%) were still on treatment. Age 45-64 years (adjusted odds ratio [aOR]=2.3, 95%CI=1.1-4.8), untreated HIV (aOR=11, 95%CI=2.7-43), severe underweight (aOR=4.4, 95%CI=2.0-9.4), chronic renal failure (aOR=4.0, 95%CI=1.1-16), chronic lung disease (aOR=2.2, 95%CI=1.1-4.4), and long regimen (aOR=7.6, 95%CI=2.7-21) were associated with adverse outcome. Being HIV positive and on antiretroviral therapy was not associated with adverse outcome (aOR=0.93, 95%CI=0.37-2.4). Of people with MDR-TB and untreated HIV, 14/20 (70%) died, 8/14 (57%) of whom received no treatment. The supplementary unimputed regression model showed that, after adjusting for regimen received, prothionamide resistance was associated with adverse outcome (aOR=2.9, 95%CI=1.3-6.3).  Interpretation: MDR-TB treatment success rates in Sierra Leone approached WHO targets (73% vs ≥75%). The short regimen was associated with higher success rates. Adverse outcome rates among working-age people and those with severe underweight, untreated HIV, chronic lung disease, and renal failure, suggest a role for integrated TB, HIV, and non-communicable disease services alongside nutritional and socioeconomic support for people with MDR-TB in Sierra Leone. The association of prothionamide resistance with adverse outcome appears to support the WHO conditional recommendation to consider alternative regimens in such cases or in areas with high prevalence of prothionamide resistance.  Funding: Wellcome Trust (209075/Z/17/Z), Joint Global Health Trials (MR/V004832/1) Declaration of Interest: None to declare. Ethical Approval: Ethical approval was obtained from the Ethics Review and Scientific Committee of the Ministry of Health and Sanitation, Government of Sierra Leone.