Gridlock from Diagnosis to Treatment of Multidrug-Resistant Tuberculosis (MDR-TB) in Tanzania: Illuminating Potential Factors for Possible Intervention

SETTINGS: Kibong'oto Infectious Diseases Hospital, Kilimanjaro, Tanzania OBJECTIVE: Characterise multidrug-resistant tuberculosis (MDR-TB)-treated cases during the scaling up of molecular diagnostics using Xpert MTB/RIF and GenoType MTBDRplus DESIGN: Retrospective cohort study RESULTS: A total of 223 MDR-TB patients were referred to the Kibong’oto Infectious Disease Hospital from January 2013 through December 2014. Four cities—Dar es Salaam, Mbeya, Mwanza, and Tanga—contributed 144 (65%) of referrals. Of the total referred patients, HIV coinfection was found in 92 (41%) and 180 (81%) had history of previous TB treatment. Molecular drug susceptibility testing (DST) contributed 201 (91%) of referrals and resulted in a shorter time from diagnosis to start of treatment, 30 days (95% confidence interval [CI], 26–37), compared to conventional phenotypic DST, 212 days (95% CI, 151–272; P<.001). Molecular DST found higher proportions of MDR-TB children and people living with HIV without prior treatment, 5 (12%) and 24 (56%), respectively, compared to those with previous treatment for TB, 4 (2%) and 68 (38%), respectively. The median CD4 count correspondingly was 131 cells/μl (IQR, 109–131) and 200 cells/μl (IQR, 94–337) for MDR-TB diagnosed by phenotypic and molecular diagnostics (P=.70). Despite the more rapid time to treatment initiation among patients diagnosed by molecular DST, treatment outcomes, including time to sputum culture conversion, did not differ compared to those diagnosed with conventional phenotypic DST. Regardless of the method of diagnosis, MDR-TB/HIV coinfected patients who died had lower CD4 counts (mean 86 ± 87 cells/μl) than survivors (mean 274 ± 224 cells/μl; P=.02). CONCLUSION: Molecular diagnostics appear to speedup the time to treatment initiation, but may not improve other treatment outcomes.