MOAB0203 - Oral Abstract Session
Reduced treatment delays for drug-resistant TB/HIV co-infected patients with decentralized care and rapid Xpert MTB/RIF test in Khayelitsha, South Africa
Presented by Helen Cox (South Africa).
H. Cox1, J. Hughes2, S. Moyo3, J. Daniels3, G. van Cutsem4, V. Azevedo5, V. Cox3
1University of Cape Town, Medical Microbiology, Cape Town, South Africa, 2Medecins Sans Frontieres Khayelitsha, Khayelitsha, Cape Town, South Africa, 3Medecins Sans Frontieres Khayelitsha, Cape Town, South Africa, 4Medecins Sans Frontieres South Africa, Cape Town, South Africa, 5City of Cape Town, Health, Cape Town, South Africa
Background: More than 14,000 patients are diagnosed with
drug-resistant TB (DR-TB) in South Africa annually, with 65% estimated to be
HIV-infected. While DR-TB case detection is high, only 50% are started on second-line
TB treatment with delays often up to several months, leading to high mortality,
particularly for HIV-infected individuals, and ongoing transmission. The Xpert test,
with simultaneous TB and rifampicin-resistance diagnosis, has the potential to
increase the proportion of patients treated and reduce treatment delay,
particularly when combined with access to decentralized DR-TB management. The
Impact of Xpert on treatment delay and percentage treated were analysed in a
decentralized DR-TB programme.
Methods: Between 2007 and 2013, decentralized DR-TB
management (clinic-based and predominantly ambulatory) was implemented
progressively. Simultaneously, over 2007/2008, first-line drug susceptibility
testing (DST) moved from culture-based DST to line probe assay (LPA). Xpert was
introduced in late 2011. Data on sputum collection date and DR-TB treatment
initiation were collected routinely.
Results: Between 2007 and 2013 1,368 DR-TB patients
(rifampicin-resistance) were diagnosed through the decentralized programme (51%
female; 72% HIV-infected). Prior to decentralization (2003/2006) the median
time to treatment was 71 days (culture DST, IQR 22-120), declining to 50 (IQR
12-88) with LPA and some decentralization across 2007/08 (p< 0.0001). Further
decentralization in 2009, 2010 and 2011 led to median delays of 39 (IQR 11-67),
32 (IQR 12-52), and 27 (IQR 12-42) days respectively (LPA, p< 0.0001 for
trend). Xpert resulted in median delays of 13 (IQR 7-19) and 7 (IQR 3-11) days
for 2012 and 2013 (p< 0.0001 versus LPA).
While HIV infection was associated with longer
delays across 2009/11 (LPA, p=0.02); this difference disappeared in 2012/13 (Xpert).
Across 2009/11 (LPA), 94% of HIV negative patients initiated treatment,
compared to only 86% among HIV-infected (p=0.001). Corresponding figures for
2012-13 with Xpert were 100% and 89% respectively (p=0.001).
Conclusions: Decentralization of DR-TB management dramatically
improves treatment initiation and reduces treatment delay. Xpert further
improves time to treatment, particularly among HIV-infected patients, but the
proportion initiating treatment remains poorer than HIV negative, predominantly
due to early mortality. Nonetheless, these results would be expected to lead to
both reduced mortality and reduced community transmission.
Back to the Programme-at-a-Glance