Tuberculosis patients admitted to intensive care units with acute respiratory failure had significantly better survival at 90 days after treatment with corticosteroids and anti-TB drugs, compared with patients not treated with the steroids, according to a retrospective study.
An adjusted inverse probability of treatment weighted analysis using propensity scores revealed corticosteroid use to be independently associated with a significantly reduced 90-day mortality rate (OR = 0.47; 95% CI, 0.22-0.98). This statistical approach was used because it reduces selection bias and other potential confounding factors in a way that a multivariate analysis cannot, wrote Ji Young Yang, MD, of Busan (South Korea) Paik Hospital and Inje University College of Medicine in Busan.
Zerbor/Thinkstock
Mortality rates were similar between the steroid-treated and non–steroid-treated groups (48.6% and 50%, respectively), and unadjusted 90-day mortality risk was not affected by steroid administration (odds ratio, 0.94; 95% CI, 0.46-1.92; P = .875), reported Dr. Yang and colleagues (Clin Infect Dis. 2016 Sep 8. doi: 10.1093/cid/ciw616).
The study involved the examination of records of 124 patients (mean age 62, 64% men) admitted to a single center over a 25-year period ending in 2014. Of these, 56.5% received corticosteroids, and 49.2% of the cohort died within 90 days.
The investigators acknowledged that their study was limited by various factors, including its small size, its use of data from a single center, and its lack of a standardized approach to steroid treatment.
“Further prospective randomized controlled trials will therefore be necessary to clarify the role of steroids in the management of these patients,” they wrote in their analysis. However, Dr. Yang and colleagues argued, in acute respiratory failure – a rare but dangerous complication in TB – “corticosteroids represent an attractive option because they can suppress cytokine expression and are effective in managing the inflammatory complications of extrapulmonary tuberculosis. Moreover, corticosteroids have been recently been shown to reduce mortality or treatment failure in patients with tuberculosis or severe pneumonia.”
Robert C. Hyzy, MD, director of the critical care medicine unit at the University of Michigan, Ann Arbor, said the findings “should be considered hypothesis generating.
“Clinicians should wait for prospective validation of this observation before considering the use of corticosteroids in hospitalized patients with tuberculosis,” he added.
Dr. Yang and colleagues disclosed no conflicts of interest or outside funding for their study.