Dexamethasone adjunctive treatment for tuberculous meningitis

Claims for improved survival from systemic corticosteroids in diverse conditions: an umbrella review

“…On further scrutiny, we identified 14 different trials that had over 100 deaths. Table 1 lists randomized trials of systemic corticosteroids where mortality is an outcome and where at least 100 deaths were recorded during the follow-up [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21]. These 14 independent trials have addressed 10 medical indications, including infectious diseases (bacterial meningitis, tuberculous meningitis, sepsis ⁄ septic shock), malignancies (operable breast cancer, metastatic breast cancer, preterminal cancer), head trauma, myocardial infarction, antenatal corticosteroids in preterm labour, and postnatal corticosteroids in preterm infants.…”

“…In at least six of the 14 trials [5,8,9,11,16,21], the authors have focused their primary mortality analysis in either a subgroup of the patients randomized or in a time point that censors some of the deaths that were recorded during the trial. In five of these six subset analyses, the authors concluded that there were nominally statistically significant (P < 0AE05) differences in mortality between the steroid and control arms (in favour of steroids in four trials [9,11,16,21], against steroids in one trial [8]).…”

“…In five of these six subset analyses, the authors concluded that there were nominally statistically significant (P < 0AE05) differences in mortality between the steroid and control arms (in favour of steroids in four trials [9,11,16,21], against steroids in one trial [8]). However, only one of these five trials also shows a nominally statistically significant difference in mortality when all randomized patients are considered and all recorded deaths are counted.…”

“…Two other trials [11,16] claimed significant benefits with steroids only in subsets of patients apparently predefined as the key group of interest: the subgroup of corticotropin test non-responders (a test result that suggests adrenal insufficiency) in a trial of septic shock [11], and the subgroup of CSFculture-positive patients in a trial of tuberculous meningitis [16]. Yet another trial [9] claimed significant benefits with steroids for tuberculous meningitis when deaths were censored at 9 months follow-up, presumably the time of the pre-specified primary analysis, while the Kaplan-Meier plots tend to converge again after this point and there is no significant difference based on the shown published data to 11 months (relative risk 0AE81, P = 0AE052 based on our calculations).…”

“…The other four favourable subset analyses for tuberculous meningitis (two trials), sepsis and antenatal corticosteroids [9,11,16,21] are either claimed to be a priori or it is unclear when they were conceived. Yet, in the absence of preregistered protocols, it is tempting to question whether they were indeed clearly intended upfront as primary analyses.…”