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

Dexamethasone is a synthetic adrenocortical steroid used widely in numerous specialities, including anaesthesia. The expanding number of procedures performed as day cases has led to increasing use of dexamethasone for reducing postoperative nausea and vomiting (PONV) 1. In addition to PONV, administration of dexamethasone is common for specific surgical procedures to reduce swelling and postoperative pain by reducing the inflammatory response 2. Dexamethasone is regarded by many as an ideal peri-operative agent: cheap; readily available; great for PONV; promotes appetite and a feeling of wellbeing; and associated with earlier discharge from day surgery units 3-5. Hence it is used widely; but should it be? Side-effects are generally thought of as being linked to long-term steroid use. What problems could possibly be caused by a single dose of dexamethasone at induction; surely the benefits outweigh any small risk of potential problems? Dexamethasone has very high glucocorticoid activity – 6–7 times greater than that of prednisolone 6, 7. Peak serum levels occur 2–12 hours after injection, and the drug is completely cleared within 3–5 days. However, cortisol levels are suppressed by 64–81% at 24 hours after injection, with most patients’ values returning to normal by one week 7. Dexamethasone clearance is inversely related to age. A single dose of dexamethasone induces whole body insulin resistance and alters cardiac glucose metabolism 8. This may result in hyperglycaemia in patients both with and without diabetes mellitus, which is associated with poor outcomes, particularly for cardiac surgery and subarachnoid haemorrhage 9-11. A meta-analysis identified that blood glucose is higher for 24 hours after administration but markedly higher in those with obesity and poor glycaemic control 12. A matched case control study of 63 patients concluded that use of dexamethasone for PONV may confer an increased risk of postoperative infection 13. Despite this, a recent meta-analysis failed to show an increase in the infection risk; however, the authors concluded that patients at higher risk of infection are often excluded from peri-operative studies on dexamethasone 10. The risk of infection following a single dose is unknown, but for larger doses there are effects on innate and acquired immunity, predisposing to infection 14. Infection from any pathogen can be difficult to detect due to dampening of febrile responses. Opportunistic or atypical organisms occur over 40 times more often in those receiving glucocorticoids 15. Dexamethasone is particularly contraindicated in systemic fungal infections and before the administration of live or attenuated vaccines because the response to these vaccines cannot be predicted. Even a single dose of 8 mg dexamethasone partially suppresses the hypothalamic pituitary axis for up to a week 16. The degree of suppression is greatest and most prolonged when given at night: 1 mg taken orally at night inhibits corticotrophin secretion for 24 hours 6. However, this suppression has minor effects and unknown clinical significance. In animal studies, a single dose of dexamethasone significantly reduced collagenisation, epithelisation and fibroblast content, implying negative effects on wound healing 17. In another animal model, the healing of tracheal anastomoses was significantly impaired by dexamethasone 18. Dexamethasone increases the risk of post-dural puncture headache after spinal anaesthesia for caesarean section when compared with placebo 19. Use of steroids may increase the risk of perforation in patients with peptic ulcers, diverticulitis, new anastomoses and non-specific ulcerative colitis. There are no data on the risk of gastric ulceration after a single dose; however, a study in rats demonstrated a 30% ulcer rate in animals receiving 1 mg.kg−1 intramucular dexamethasone 24 hours earlier, associated with an 83-87% reduction in prostaglandin action via inhibition of prostaglandin synthase and peroxidise, and a doubling in gastric acid secretion 20. Signs of perforation may also be masked 20. Patients taking long-term dexamethasone have a 4% risk of avascular necrosis 21. The risk is multifactorial but is higher with older age, low albumin, high lipid levels and the amount of dexamethasone exposure 22. The shortest use of dexamethasone causing avascular necrosis in case reports is one week; typically, onset of symptoms is after six months to three years’ exposure 23. The significance of one peri-operative dose is unknown. Use of corticosteroids in the healing phase of arthrodesis has been controversial for many years, particularly for spinal arthrodesis due to inhibition of bone graft incorporation 24. This syndrome is an oncological emergency and can be fatal. It occurs in haematological tumours that grow at a fast rate or have a large bulk. It is characterised by rapid onset of hyperkalaemia, hyperphosphataemia, lactic acidosis, hyperuricaemia and acute renal failure. It usually occurs after cytotoxic chemotherapy and dexamethasone is commonly used 25. Dexamethasone has a known lympholytic effect by arresting growth and inducing apoptosis in lymphocytes. There are case reports of tumour lysis syndrome following treatment with dexamethasone alone 26. Peri-operative death has also followed a single dose of dexamethasone causing tumour lysis syndrome in a 3-year-old undergoing adenotonsillectomy 27. It is therefore suggested that dexamethasone (and other corticosteroids) be avoided until a precise diagnosis is identified and to maintain a high index of suspicion. Dexamethasone is associated with neuropsychiatric side-effects ranging from mild symptoms of anxiety and irritability to mania and psychosis, even including suicide. Following introduction of steroid therapy in the 1950s, the reported frequency of psychiatric reactions has been as high as 5–10% 28. Critically, it remains poorly studied since the 1960s and reactions are substantially under-reported. Most current data exist from case reports, case series and review articles. There are varying reports of onset time, mostly within a few days to a week after taking the steroid 29, 30. This doesn't help quantification of risk for a single dose. Notable mania is more frequent then depression in approximately 75% of reported cases 31. The mechanism of psychiatric symptoms is multifactorial, including both direct and indirect effects on the brain. It is hypothesised that excessive levels of corticosteroids cause an increase in dopamine levels, which in turn result in symptoms of mania, psychosis and severe depression 32. Rome and Braceland performed a classic study in 1952 and reviewed the psychiatric effects of glucocorticoids 33. They identified four categories of psychiatric response to steroids (Table 1). Grade-1 effects are perceived as positive reasons to choose dexamethasone as a peri-operative anti-emetic. Symptoms reported less often include hypomania or mania, generalised anxiety, distractibility, agitation, insomnia, pressured speech, emotional lability and restlessness. These symptoms are probably more characteristic of grade-2 or -3 side-effects. At-risk groups were identified by Lewis and Smith in 1983 as women and those with systemic lupus erythematosus (SLE) and rheumatoid arthritis 28. The glucocorticoid side-effects are inversely related to serum albumin due to an increase in the amount of free, unbound steroid 34. There have been cases of suicide and homicide following use of dexamethasone. One author (AJH) has acted as an expert witness at an Inquest in a case of suicide following a day-case ENT procedure. The possible role of peri-operative dexamethasone was examined in depth, but on the balance of probabilities no causative link could be proved. No doubt it caused months or even years of worry for the anaesthetist involved. There are case reports of criminal activity related to acute steroid-induced psychosis. One related to a 26-year-old who was charged with the attempted murder of his fiancée following osteotomy of the maxilla for which he received peri-operative dexamethasone; he was found not guilty on the grounds of involuntary intoxication. His psychosis lasted 15 days and then subsided with no long-lasting effects 35. In a series of 79 patients with steroid-induced psychosis, 93% made a complete recovery, 4% had continued symptoms and 3% committed suicide 28. This phenomenon is not fully understood, but intravenous administration of dexamethasone to awake patients can lead to acute, but short-lived (3–45 seconds) perineal pain, irritation, burning or tingling. It is postulated that the phosphate ester in dexamethasone sodium phosphate plays a significant role. Females seem to be affected more than males 36. There are a number of case reports of patients who, after receiving dexamethasone 8 mg before induction, experience distressing effects, leading to the suggestion that it should only be given after induction of anaesthesia 37. These unpleasant effects can be eliminated by administration in 50 ml over 5–10 minutes 36. The argument that dexamethasone should be used because of the expense of ondansetron or granisetron is no longer valid. At a local hospital pharmacy in March 2013, antiemetics are priced as follows: ondansetron £0.16/4 mg; granisetron £0.30/1 mg; dexamethasone £0.48/4 mg; and cyclizine £0.68/50 mg (Drug Information, Hillingdon Hospital, UK, personal communication). Whilst those patients with a history of PONV stand to gain from a polypharmacy approach to antiemetics, there can be no economic justification in using dexamethasone alone. There is well proven clinical benefit for the use of peri-operative dexamethasone, confirmed in a recent meta-analysis of its anti-emetic properties. This all stands despite the withdrawal of multiple papers on PONV written by Fujii 38-40. Despite the obvious benefit in preventing PONV, dexamethasone can produce a wide variety of short- and long-term side-effects. Though neuropsychiatric reactions are reported in 10%, the true figure is unknown due to substantial under-reporting. Cognitive dysfunction is difficult to demonstrate, particularly when other agents including general anaesthesia have been given. It is appropriate to discuss steroid therapy as a course of treatment; in fact the British National Formulary recommends giving out a patient information leaflet, but the position for a single peri-operative dose is unclear 6. Day-case patients in particular may not be reviewed beyond the initial recovery period and any side-effects may occur after discharge, but before any follow-up appointment (if one is even to be had), leading to significant under-reporting. Mild neuropsychiatric effects may not even be reported at all unless sought after specifically. We are concerned that dexamethasone causes harm that goes unnoticed. Should we at least be informing patients of its potential effects? Without greater evidence of safety, we suggest avoiding steroids in certain known at-risk groups such as those with a personal or family psychiatric history, or immunological disorders such as SLE. With these patients excluded, more detailed studies of patients receiving dexamethasone are warranted. In addition, careful consideration before administration is required in other groups: oncology patients with bulky or highly proliferative tumours; diabetics; elderly orthopaedic patients; those undergoing surgery in which anastomotic breakdown is catastrophic; and those at particular risk from infection, particularly immunocompromised patients. Further, we recommend that all patients receiving dexamethasone should be warned of its potential side-effects. Steroids are powerful agents and should be used with caution. No external funding declared. AJH has acted as an expert witness as described.