Xpert test failed to affect mortality in TB
BOSTON — Replacing smear microscopy with the Xpert MTB/RIF test did not decrease mortality among people with suspected tuberculosis in South Africa, according to study results presented at the 2014 Conference on Retroviruses and Opportunistic Infections.
“In the laboratory, the Xpert test (Cepheid) outperforms traditional microscopy, but we wanted to know if it makes a difference to patients in a real-world setting,” Alison Grant, MD, PhD, professor of international health at the London School of Tropical Medicine and Hygiene, said during her presentation. “South Africa has rolled out Xpert nationwide and embedded within the rollout was our trial.”
Grant and colleagues randomly assigned 20 laboratories in four provinces to immediate implementation of Xpert testing or standard smear microscopy, deferring Xpert test implementation. Patients with suspected TB presenting at two health clinics served by each laboratory were identified by the clinic staff and invited to participate in the study. Study staff collected data on factors associated with TB and mortality risk, and vital status after 6-month follow-up was determined by interview or the National Vital Statistics Report. The researchers estimated the effect of the Xpert test on mortality risk and conducted an adjusted analysis to control for baseline, individual factors.
From June to November 2012, there were 4,972 people who presented with suspected TB. Among them, 4,656 were eligible and included in the analysis. Seventy-six percent of patients knew their HIV status, and 62% of those reported being HIV-positive. The median self-reported CD4 count was 311 cells/mcL, and 33% of the patients reported ever receiving antiretroviral therapy.
Among the 4,656 total participants, 99% had a known vital status at 6 months. There were a reported 207 deaths. In the Xpert test intervention arm, the 6-month mortality risk was 3.9% compared with 5% in the smear microscopy control arm (RR=0.86; 95% CI, 0.56-1.28). There was no difference in mortality between the arms after controlling for age, sex, BMI, HIV status and imbalanced factors at baseline (adjusted RR=1.10, 95% CI, 0.75-1.62).
Grant said Xpert confirmed a diagnosis of TB in approximately 50% more people than microscopy. However, it made no difference in the number of people who started TB treatment in the 6-month follow-up. About 16% of people who had a positive TB test result did not start treatment, and this was not different between the arms.
For those who did not know their HIV status, the 6-month mortality risk was 2.4 times higher compared with those reporting negative status (adjusted RR=2.41; 95% CI, 1.47-3.98). For those with HIV who reported not receiving ART, the 6-month mortality risk was 3.3 times higher compared with those reporting negative status (RR=3.32; 95% CI, 2.03-5.41).
“Using Xpert, more people had rapid confirmation of a TB diagnosis and that’s important because it allows nurses to start correct TB treatment promptly,” Grant said. “Our study also underlines that people who are tested for TB need to know their HIV status and that we need stronger systems to ensure that everybody with a positive TB test result starts treatment promptly.”
For more information:
Churchyard G. #95. Presented at: CROI 2014; March 3-6, 2014; Boston.