A Novel Tool to Detect Treatment Failures Early in the course of Antituberculosis treatment: Old Wine in New Bottle?

Abstract

Background. Despite being a suboptimal test, sputum smear microcopy is widely used in many TB-endemic countries to assess a patient's response to anti-tuberculosis treatment (ATT). Mycobacterial growth indicator tube (MGIT), with better sensitivity can detect Mycobacterium tuberculosis (M. tb) even at low bacillary loads. With varying results on the utility of time to detect (TTD) M. tb in MGIT culture, we sought to assess whether TTD M. tb can predict response to ATT. Methods. We enrolled new drug-sensitive adult (>18 years) pulmonary TB cases initiating ATT in Chennai and Pune, India from January 2014 to 2017. Participants were prospectively evaluated at week 2, 4, 8, and month 5 and 6 for sputum culture by MGIT. Response to ATT was defined as cure (two or more consecutive cultures negative for M. tb with no subsequent positive cultures) or failure (any two sputum cultures positive for M. tb or one culture/ smear positive with symptoms suggestive of PTB during the last two months of ATT). Cox-proportional hazards model was used to measure the association between TTD of M. tb on MGIT and treatment response. Kaplan-Meier survival curves for time to positive test by those failing treatment and cured were estimated. Results. Treatment outcome was available for 442 adults; [256 (58%) men and 202 (46%) diabetics]. At the end of ATT, 407 (92%) were cured and 35 (8%) failed treatment. After 2 months of ATT, a shorter TTD (<14 days) was observed in a significant number of treatment failures than treatment responders [69% vs. 42%, P = 0.01]. With ATT, TTD M. tb steadily increased in both the groups (Fig 1) with gross difference in TTD between the two groups (Fig 2). Also after 8 weeks of ATT, M. tb was detected much earlier among treatment failures than among cured [(median days (IQR): 15.1 (9.2-23.1) vs. 18.2 (16.0-25.1), P < 0.01]. Cox-regression model showed that while on ATT, lesser the days to detect M. tb on MGIT, higher was the risk to treatment failure [up to 7 days: adjusted HR 16.07 (95% CI 8.5-30.5) and 8 to 14 days: adjusted HR 11.56 (95% CI 7.2-18.6)]. Conclusion. Our study shows that while on ATT, lesser time to detect M. tb on MGIT culture has a higher risk of treatment failure. Positive MGIT cultures at the end of two months of ATT can be considered as an early indicator of treatment non-responders. (Figure Presented) .