Nomenclature of drug- resistant tuberculosis

www.thelancet.com/infection Vol 13 November 2013 917 identified studies overlooked by previous reviews. We thought that searching AIDSline was inessential because citations since 2001 are available in Medline. Conference abstracts often present different results to those in published studies and report quality indicators poorly. Google Scholar, because of the underpinning informatics processes, has poor specificity and cannot be searched replicably. Thus, neither were searched. We acknowledge that, as with all reviews, studies might have been overlooked and available evidence could limit our conclusions. Few original studies reported outcomes as time-to-event data, so a meta-analysis of hazard ratios was not feasible. The abstract Xi-Ran He and colleagues identifi ed is about immune reconstitution inflam matory syndrome (IRIS), which occurs in 10–25% of patients with HIV initiating antiretroviral therapy. Tuberculosis with IRIS accounts for 20% of cases. We did not include steroid trials for this disorder since we did not view tuberculosis with IRIS as an organ specifi c tuber culosis subtype. Use of steroids in this group needs separate investigation. Many trials were from low-income regions with higher prevalence of tuberculosis. However, more recent data are needed, particularly in those with pulmonary tuberculosis and tuberculosis alongside comorbidities, such as HIV, to make further inference.