Routine use of dexamethasone for postoperative nausea and vomiting: the case for

Dexamethasone is a synthetic steroid known for its anti-inflammatory properties, but is now increasingly used as an antiemetic in surgical patients. It has been used as an antiemetic in patients receiving chemotherapy for more than three decades and was first studied in surgical patients in the early nineties 1. Although the precise mechanism of action is not known, it has been suggested that the antiemetic effect could be due to the inhibition of prostaglandins, prevention of serotonin release in the gut, reduction in neural 5-hydroxytryptophan levels or release of endorphins 2. The effectiveness of dexamethasone as a peri-operative antiemetic has been well substantiated. Carlisle and Stevenson 3, in their Cochrane Review, calculated a risk ratio for dexamethasone of 0.48 (95% CI 0.43–0.54) for prevention of postoperative nausea and vomiting (PONV). This was similar to ondansetron (RR 0.56 (95% CI 0.50–0.62)) and marginally better than cyclizine (0.67 (0.56–0.79)) and droperidol (0.62 (0.58–0.67)). They also confirmed the findings of the IMPACT Group 4 that, when combined with other antiemetics, dexamethasone had an additive effect and the magnitude of the effect of combined antiemetic agents was equal to the product of the risk ratio for each agent. These findings were similar to a previous meta-analysis of 17 trials consisting of 1946 patients by Henzi et al. 2, which also found that dexamethasone had a more pronounced effect on late emesis (6–24 h). In addition to its antiemetic efficacy, dexamethasone has been shown to have other beneficial effects. De Oliveira et al. 5 analysed 24 randomised controlled trials (n = 2751) published between 1997 and 2010 and concluded that an intermediate dose (0.11–0.2 mg. kg−1) of dexamethasone conferred significant benefits in early as well as late pain. It was also shown to have a significant morphine-sparing effect. The differences in pain scores and morphine consumption between patients receiving dexamethasone and controls were small, which might call into question their clinical significance. On the other hand, one can argue that an antiemetic with analgesic properties able to provide lower pain scores at lower doses of painkillers should be an obvious first choice in surgical patients. Dexamethasone has also been shown to reduce the incidence of sore throat. Thomas and Beevi 6 showed that dexamethasone reduced both the incidence (by 36%) and the severity of sore throat in patients undergoing tracheal intubation. These findings were confirmed in a recent randomised controlled trial 7. It has also been shown that use of dexamethasone results in a quicker attainment of discharge criteria, a shorter hospital stay and a better quality of recovery 8. A single dose of dexamethasone resulted in earlier oral intake after tonsillectomy 9 and reduced swelling after dental extractions 10. On the basis of its many benefits in surgical patients, it would be logical to suggest that dexamethasone is routinely used as a peri-operative antiemetic. The caveat, however, is that the harm caused by dexamethasone should not outweigh the benefits, and its safety profile must be comparable with alternative antiemetics. Increased risk of bleeding, particularly in patients having tonsillectomy, increased risk of postoperative infection, impairment of glucose homeostasis and psychiatric disturbances are some of the worrying adverse effects that have been associated with steroid therapy in the literature. In a recent well-conducted meta-analysis 11, the incidence of bleeding episodes and hospital admission for bleeding was found to be similar in patients who had received dexamethasone compared with those who did not receive the drug. However, operative re-intervention for bleeding was higher in the dexamethasone group. In subgroup analysis, this effect was confined to children and those who had concurrent administration of non-steroidal anti-inflammatory drugs (NSAIDs). Furthermore, when studies with high risk of bias were excluded, there was no clear association between steroid use and the incidence of operative re-intervention. A more recent randomised controlled trial 12 and another recent meta-analysis 13 did not show any increased risk of bleeding with dexamethasone. Another concern is the increased likelihood of postoperative infection. Percival et al. 14 used a matched case control design and found increased adjusted odds of 3 (95% CI 1.06 to 19.3, p = 0.035) of developing postoperative infection in patients receiving a single dose of 4–8 mg dexamethasone. In their meta-analysis 5, De Oliviera et al. also looked at the incidence of wound infection and found no difference in patients receiving dexamethasone compared with placebo. As a worst-case scenario, they estimated the possibility of increased incidence of infection to be 1.2% in patients receiving intermediate (0.11–0.2 mg.kg−1) doses of dexamethasone. Dexamethasone is known to cause hyperglycaemia in major neurosurgical procedures 15. However, randomised studies examining changes in blood glucose after a single dose have failed to demonstrate significant impairment of blood glucose homeostasis in both non-diabetic and diabetic patients 16, 17. Although there was a statistically significant rise in blood glucose levels in non-diabetic patients, the peak levels reached (8 ± 1 mmol.l−1) were within the recommended peri-operative glycaemic range. The data, however, suggest that obese patients with impaired glucose tolerance may be at particular risk 18, 19. Steroid therapy is known to cause a range of psychiatric disturbances including labile mood, hallucinations, sleep disturbances, suicidal thoughts and aggravation of schizophrenia 20. These manifestations are usually seen with higher doses and prolonged therapy. There are no reliable data on psychiatric manifestations following a single dose of dexamethasone and as such it is difficult to comment on the harm that might result from its use as a peri-operative antiemetic. There are ample data on the many benefits of dexamethasone in dental, ENT, laparoscopic and day surgery procedures. In addition, it has a long duration of action with consequent effectiveness in late emesis. Although it is not the cheapest of the commonly used agents (4 mg ondansetron £0.11; 3.3 mg dexamethasone £0.45; 50 mg cyclizine £0.68; 2.5 mg droperidol £2.22 (personal communication with my hospital pharmacy)), it is reasonably priced and it would be interesting to see how it compares with others in an up-to-date cost-effectiveness analysis. We do not have reliable data to estimate the number of patients who might be harmed by use of dexamethasone as an antiemetic. As a worst-case scenario based on the data from the two meta-analyses 5, 11, it is possible that 1–3 patients in every hundred might be coming to harm from bleeding or postoperative infection. Whether more patients come to harm from complications of diabetes, psychiatric disturbances, exacerbation of peptic ulcer disease and immunosuppression from a single dose of dexamethasone is unknown. It is worth noting that other alternative agents have an excellent safety profile and severe adverse effects such as prolongation of QT interval and extrapyramidal manifestations are rare. The optimal single dose of dexamethasone is unknown. It appears that 0.1–0.2 mg.kg−1 dexamethasone phosphate (4 mg dexamethasone phosphate equates to 3.3 mg dexamethasone) achieves the optimal benefit-risk ratio and as such it may be prudent to restrict dexamethasone administration to within this dose range. Further studies should examine the dose range that can provide the best risk-benefit ratio and also report on all possible adverse effects mentioned above. In the meantime, I suggest that these postoperative complications are closely monitored in all patients receiving dexamethasone, and that caution is exercised when combining dexamethasone with NSAIDs in surgical procedures where postoperative bleeding is an issue, and in patients with diabetes (particularly those with high BMI), peptic ulcer disease and a known psychiatric history. Although some may argue against the use of dexamethasone 21, citing the theoretical potential for harm, undue weight ascribed to anecdote and to evidence extrapolated from steroid therapy in general, high-dose steroids and prolonged therapy may mean that many patients are deprived of benefits of single-dose peri-operative dexamethasone. If we can exclude the minority of high-risk individuals, the routine use of dexamethasone in others would benefit a large, low-risk surgical population. No external funding and no competing interests declared.