This article is more than 5 years old. Information may no longer be current.
Sputum monitoring may not predict pulmonary TB progression
Click Here to Manage Email Alerts
Smears and cultures of sputum were unable to predict the failure or relapse of pulmonary tuberculosis, according to meta-analysis results.
Researchers from Washington and California reviewed a group of 28 randomized controlled trials, cohort and case-control studies. In all cases, patients were previously untreated for pulmonary TB and received a standardized medication regimen with rifampicin in the initial phase. Fifteen studies met inclusion criteria.
At 2 months, sputum smears and mycobacterial cultures had 24% (95% CI, 12%-42% for six studies) and 40% (95% CI, 25%-56% for four studies) sensitivity for predicting relapse. Corresponding specificities were 85% (95% CI, 72%-90%) for sputum smears and 85% (95% CI, 77%-91%) for mycobacterial cultures. For failure, sputum smears were 57% (95% CI, 41%-73% for seven studies) sensitive and 81% (95% CI, 72%-87%) specific.
These findings have already spurred WHO to rescind its recommendation to use sputum exams at end of the second treatment month to determine whether further treatment is needed, according to the analysis.
The researchers said they were surprised that the evidence supporting such a fundamental and important component of TB control as sputum monitoring was not robust.
“The widespread use of this test and its uncertain utility suggest a critical need to appraise tests and strategies for identifying patients at risk for poor outcome that are suitable for resource-limited settings,” the researchers wrote. “The use of a sputum sample to predict poor outcome and decide on an extension of treatment should be questioned until further studies have addressed these issues.”
Horne DJ. Lancet Infect Dis. 2010;10:387-394.