To the Editor
The use of dexamethasone in noncardiac outpatient surgery has been a topic of debate. In the most recent consensus, Rajan et al recommended a 4 mg dose based on a single study demonstrating noninferiority in postoperative site infection outcomes compared to placebo. This study included patients with a median glycated hemoglobin (HbA1c) level of 6.8%, while excluding those with poorly controlled diabetes (HbA1c >9%), as previous research has indicated a higher risk of postoperative site infection in such cases.
Jones et al published a meta-analysis where they found that the use of dexamethasone compared to the placebo group did not result in a higher risk of surgical site infection in diabetic patients (odds ratio [OR], −0.1; 95% confidence interval [CI], −0.64 to 0.44; P = .72) with low heterogeneity (I² = 25.92%). This study included in its quantitative analysis the trial by Corcoran et al, which Rajan cited as biased due to the baseline level of glycated hemoglobin (median = 6.8%). However, glycated hemoglobin levels above 6% are associated with an increased risk of infection. This is important to analyze the causal relationship between dexamethasone administration at a given dose in diabetic patients and the occurrence of surgical site infection.
Dieleman evaluated dexamethasone doses ranging from 1 mg/kg up to a maximum of 100 mg in noncardiac surgery, without finding an association between dexamethasone use and surgical site infection. However, only 38.4% of the patients were diabetic (18.7% in the dexamethasone group). Despite this, Corcoran et al, in another study, compared 4 mg and 8 mg doses of dexamethasone without finding evidence of an increased risk of infection in diabetic patients. Rajan et al might reconsider that the patients had an median glycated hemoglobin level of 6.4% [interquartile range 6.1–8.6], yet it is important to note the interquartile range, where 50% of the patients were between 6.1% and 8.6% of glycated hemoglobin, which refutes their argument. This indicates that none of Bradford Hill’s main criteria for causality are met, as there is no dose-response gradient (a higher dose of dexamethasone does not equate to a higher risk of surgical site infection in diabetics), nor the criterion of temporality, as the administration of dexamethasone during induction is not associated with surgical site infection but does offer greater analgesic benefit at doses as high as 48 mg intravenous (IV) without a higher risk of infection at the surgical site.
Given this situation, it is crucial to vindicate, as with Alan Turing, the fundamental role that dexamethasone plays at doses of up to 8 mg in the perioperative period, especially in the prevention of postoperative nausea and vomiting and in better pain control (Figure). By recognizing its clinical value, we avoid falling into the dishonor of its discredit and consequent disuse based on the prevention of a risk that may not exist, much like how Turing was unjustly marginalized despite his invaluable contributions.