Why do clinical trials of Xpert MTB/RIF fail to show an effect on patient relevant outcomes?
T. H. Boyles
Int J Tuberc Lung Dis 2017; 21(3): 249-250
THE WORLD HEALTH ORGANIZATION (WHO) recommends that Xpert MTB/RIF (Cepheid, Sunnyvale, CA, USA) should be used instead of conventional microscopy, culture and drug susceptibility testing (DST) as the initial diagnostic test in adults with suspected multidrug-resistant tuberculosis (MDR-TB) or human immunodeficiency virus (HIV)-associated TB (strong recommendation, high quality evidence). This advice is based primarily on a systematic review and meta-analysis of diagnostic accuracy studies, which found a pooled sensitivity of 88% (95% credibility interval [CrI] 83–92) and pooled specificity of 98% (95% CrI 97–99) when Xpert replaces microscopy as the initial diagnostic test. This equates to a positive likelihood ratio (LR+) of 44 and a negative LR (LR- ) of 0.12.
There have now been eight trials evaluating the impact of Xpert on patient-relevant outcomes such as morbidity and mortality, and all have shown no benefit. This not only calls the WHO guidance into question, it also raises the question as to why a test with seemingly impressive diagnostic accuracy should fail in impact trials. A number of theories have been advanced, including deficiencies in trial design and trial conduct and the weaknesses of the health systems in which the trials were conducted. Application of the threshold approach to clinical decision making may also be helpful in solving this