Research Article: Treatment initiation among persons diagnosed with drug resistant tuberculosis in Johannesburg, South Africa

Date Published: July 26, 2017

Publisher: Public Library of Science

Author(s): Denise Evans, Kathryn Schnippel, Caroline Govathson, Tembeka Sineke, Andrew Black, Lawrence Long, Rebecca Berhanu, Sydney Rosen, Pere-Joan Cardona.


In South Africa, roughly half of the drug-resistant TB cases diagnosed are reported to have been started on treatment. We determined the proportion of persons diagnosed with rifampicin resistant (RR-) TB who initiated treatment in Johannesburg after the introduction of decentralized RR-TB care in 2011.

We retrospectively matched adult patients diagnosed with laboratory-confirmed RR-TB in Johannesburg from 07/2011-06/2012 with records of patients initiating RR-TB treatment at one of the city’s four public sector treatment sites (one centralized, three decentralized). Patients were followed from date of diagnosis until the earliest of RR-TB treatment initiation, death, or 6 months’ follow-up. We report diagnostic methods and outcomes, proportions initiating treatment, and median time from diagnosis to treatment initiation.

594 patients were enrolled (median age 34 (IQR 29–42), 287 (48.3%) female). Diagnosis was by GenoType MTBDRplus (Hain-Life-Science) line probe assay (LPA) (281, 47.3%), Xpert MTB/RIF (Cepheid) (258, 43.4%), or phenotypic drug susceptibility testing (DST) (30, 5.1%) with 25 (4.2%) missing a diagnosis method. 320 patients (53.8%) had multi-drug resistant TB, 158 (26.6%) rifampicin resistant TB by Xpert MTB/RIF, 102 (17.2%) rifampicin mono-resistance, and 14 (2.4%) extensively drug-resistant TB. 256/594 (43.0%) patients initiated treatment, representing 70.7% of those who were referred for treatment (362/594). 338/594 patients (57.0%) did not initiate treatment, including 104 (17.5%) who died before treatment was started. The median time from sputum collection to treatment initiation was 33 days (IQR 12–52).

Despite decentralized RR-TB treatment, fewer than half the patients diagnosed in Johannesburg initiated appropriate treatment. Offering treatment at decentralized sites alone is not sufficient; improvements in linking patients diagnosed with RR-TB to effective treatment is essential.

Partial Text

In 2015, a global total of 132 120 cases of multi-drug resistant tuberculosis (MDR-TB) and rifampicin resistant TB (RR-TB) were notified to the World Health Organization (WHO). This represented 23% of the estimated 580 000 cases of MDR/RR-TB cases worldwide demonstrating a major diagnostic gap [1].

We enrolled 594 patients in the study. As described in Table 1, they had a median (IQR) age of 34 (29–42) years, and 48.3% were female. Most were diagnosed by LPA (281, 47.3%) or Xpert MTB/RIF (258, 43.4%), with a few by phenotypic DST (30, 5.1%) or by unknown (missing) diagnostic method (25, 4.2%). 320 patients (53.8%) had MDR-TB, 158 (26.6%) had rifampicin resistant TB by Xpert MTB/RIF with no additional drug susceptibility results available, 102 (17.2%) had rifampicin mono-resistance, and 14 (2.4%) had XDR TB.

In this study, fewer than half (43.0%) the patients diagnosed with DR-TB in the City of Johannesburg in 2011–2012 initiated appropriate treatment within six months of diagnosis despite the quadrupling of facilities offering DR-TB treatment under the new decentralization framework–from one inpatient facility to one inpatient plus three outpatient facilities. The proportion of patients who initiated treatment within six months (43.0%) is similar to national estimates reported for South Africa in the 2013 WHO Global TB report (6 494 cases started on MDR-TB treatment/15 419 cases of laboratory confirmed MDR-TB for 2012; 42%) [22]. Pre-treatment loss to follow-up among drug sensitive patients in Africa ranges from 6 to 38%, and is similar among patients with drug resistant TB [23–26]. We showed high pre-treatment loss of 17.8% and many patients (17.5%) died before DR-TB treatment could be initiated. Only 3 out of 5 patients diagnosed with DR-TB (60.9%) could be traced and referred for treatment. Although decentralization and the implementation of Xpert MTB/RIF for the diagnosis of TB and rifampicin resistance has resulted in a significant reduction in time to treatment initiation as reported in previous studies [26], these programmatic changes did not solve the problem of loss of patients between diagnosis and treatment initiation seen in our cohort. This initial loss from care has contributed to lack of improvement in treatment outcomes [27]. This is an area where additional health system strengthening is required.

Despite these limitations, our data indicate very high rates of failure to initiate appropriate RR-TB treatment for patients diagnosed with drug-resistant TB in the City of Johannesburg. Though it is surely a step in the right direction, offering treatment at decentralized sites alone is not sufficient; improvements in linking patients diagnosed with RR-TB to effective treatment remains a high priority.




0 0 vote
Article Rating
Notify of
Inline Feedbacks
View all comments