Pretreatment Time to Detection of Mycobacterium tuberculosis in Liquid Culture Is Associated with Relapse after Therapy

LETTER

The most important outcome of tuberculosis (TB) treatment is nonrelapsing cure. Identification of predictors of relapse is useful for evaluating new drugs and treating patients. Relapse has been associated with a greater pretreatment sputum bacillary load, measured by sputum smear grade and numbers of CFU in quantitative cultures (1, 7). A more recent microbiological measure of the bacillary load is time to detection (TTD), which is the time elapsed from inoculation of liquid culture until the detection of growth of Mycobacterium tuberculosis. In studies in Uganda and South Africa, a shorter pretreatment TTD was correlated with greater numbers of CFU, and patients who relapsed or had positive 1- and 2-month sputum cultures had a shorter baseline TTD than those who did not (2, 3, 5). Weiner also reported a similar association of a shorter mean TTD during the first 2 months of therapy among patients with treatment failure (6). We analyzed data from a recent trial of shortening treatment of pulmonary TB, performed in Brazil, the Philippines, and Uganda (4), to further assess the relationship between the baseline TTD and relapse.

Patients in this secondary analysis participated in a phase 3 trial of TB treatment shortening (clinicaltrials.gov identifier NCT00130247) (4) which enrolled non-HIV-infected adults with culture-confirmed noncavitary pulmonary TB and negative sputum cultures after 2 months of treatment. The primary endpoint was relapse, defined as recurrent disease with a positive sputum culture that matched the initial M. tuberculosis isolate by IS6110 DNA fingerprinting. The TTD was measured prior to treatment using the Bactec 460 TB system.

Three hundred ninety-two of the 394 subjects randomized in the original trial were included in this substudy. Two subjects with relapse due to exogenous reinfection were excluded. Sixty-one percent were male, the mean age was 31 years, the baseline median body mass index was 20.1 kg/m², and 64% were sputum smear positive. Fifteen (4%) subjects were not included in the analysis of baseline TTD: 8 (2%) due to missing data and 7 (2%, including 1 relapse subject) due to culture contamination. Fifteen of the sixteen subjects who relapsed in the original trial were included in this analysis; twelve (80%) of these were in the shorter 4-month treatment arm. The median TTD among the subjects who relapsed was 5.0 days (interquartile ratio [IQR], 3.5 to 8.5), compared with 9.6 days (IQR, 5.5 to 14.5) among the 362 subjects that did not relapse (P = 0.01, Mann-Whitney-Wilcoxon).

In this randomized clinical trial, we found that the median pretreatment TTD in liquid culture was shorter among patients who relapsed than among those that did not relapse. Our data confirm the findings of Hesseling with a South African cohort (3) and lend support to the use of the TTD as an early biomarker of treatment response. While this study utilized the older Bactec 460 system, the newer, widely implemented MGIT 960 system allows automated detection of mycobacterial growth and recording of the TTD. Further studies will be needed to assess the association between TTD and relapse using the MGIT 960 system and also to accumulate more data in order to determine whether TTD can predict relapse with a given sensitivity and specificity. TTD offers the potential for a widely accessible and easily measured biomarker of disease response and relapse that is available early in the course of treatment.