Prophylaxis for Vomiting by Children After Tonsillectomy

Vomiting after surgery may lead to dehydration, metabolic derangements, seizures, delayed discharges from hospital, and unscheduled admissions to hospital ^[1-3]. Children undergoing tonsillectomy are at particular risk, with as many as 71%-73% of children vomiting after general anesthesia for tonsillectomy. Anesthesiologists are searching for cost-effective techniques that will minimize this problem ^[3,4].

Dexamethasone and perphenazine are two inexpensive antiemetics that decrease vomiting after tonsillectomy in children ^[3,5]. Dexamethasone, which decreases vomiting after chemotherapy ^[6], has a terminal half-life of approximately three hours and a duration of up to 72 hours ^[7]. The mechanism of antiemetic action of dexamethasone is unknown. Perphenazine is a phenothiazine with moderate anticholinergic effects and weak to moderate sedative effects. Compared with ondansetron, perphenazine has similar antiemetic effects and fewer side effects among adults after hysterectomy ^[8]. We compared the effects of dexamethasone and perphenazine on vomiting after tonsillectomy in children.

Methods

Healthy children aged 2-12 yr undergoing elective tonsillectomy or adenotonsillectomy were enrolled in this randomized, stratified, blocked, double-blind study after approval of our ethics committee. Within the balanced study design, blocks of five patients were stratified according to use of premedication and induction of anesthesia technique so that a similar number of patients in each study group received premedication and each of the two available induction techniques. The children were randomized to receive dexamethasone or perphenazine in a double-blind fashion. Exclusion criteria included an allergy to a study drug, history of sleep apnea, or ASA physical status III or greater.

Patients were permitted to drink clear fluids up to 3 h before anesthesia, but they did not ingest solid food on the day of surgery. Those children requiring premedication received midazolam 0.5 mg/kg (maximal dose 15 mg) orally 20-30 min before induction of anesthesia. General anesthesia was induced by inhalation with N₂ O and halothane or intravenously (IV) with propofol 2.5-3.5 mg/kg. Mivacurium 0.25 mg/kg was administered if a muscle relaxant was indicated. The study drug, 150 micro g/kg dexamethasone (maximal dose 8 mg) or 70 micro g/kg perphenazine (maximal dose 5 mg), was administered IV in a double-blind fashion immediately after the induction of anesthesia. Anesthesia was maintained with 70% N₂ O and 0.75%-2.0% halothane. Midazolam 50 micro g/kg (maximal dose 3 mg) IV was given if the child did not receive premedication. [Midazolam was given to all patients because it has been shown to have antiemetic effect in children undergoing tonsillectomy ^[9].] All patients received codeine 1.5 mg/kg intramuscularly in the right thigh before the commencement of surgery. Intraoperative IV fluid was lactated Ringer's solution at standard rates; that is, half the deficit during the first hour, plus maintenance fluids, in addition to replacement of 4 times the estimated blood loss with crystalloids.

Management of postoperative emesis, pain, and fluids was standardized. Vomiting was defined as the forceful expulsion of liquid gastric contents. Retching and nausea were not considered vomiting. Nursing staff and parents were aware of this definition of vomiting. Nursing staff recorded vomiting in the charts, which is mandatory at our institution. Patients who vomited twice in the hospital received dimenhydrinate 1 mg/kg IV, whereas those who vomited four times received ondansetron 0.1 mg/kg IV. The parents were contacted 24 h after surgery by a research assistant, who asked the parents the number of times their child vomited after discharge from the hospital and what the parents did in response to the vomiting. Pain in the postanesthesia recovery room (PAR) was treated with morphine 50 micro g/kg IV, whereas subsequent pain was treated with acetaminophen and/or codeine in the daycare surgical unit (DCSU) and at home. IV fluid was lactated Ringer's solution at twice maintenance rates in the PAR and DCSU. Patients were encouraged, but not coerced, to drink clear fluids in the DCSU before discharge. Patients were discharged after a minimal 4-h stay in the DCSU.

Normally distributed data, such as patient age, were compared by using one-way analysis of variance, whereas binomial data, such as the incidence of vomiting, were compared by using chi squared analysis or Fisher's exact test, whichever was appropriate. Logistic regression analysis was used to determine whether factors such as study drug, age, weight, in-hospital analgesics, premedication, muscle relaxant use, operative time, estimated intraoperative blood loss, and induction technique were significant predictors of vomiting. Data are presented as mean +/- SD. Sample size was determined by assuming that the acceptable difference in vomiting is 20%. The alpha error was set at 0.05 (two-sided), and type II error was set at 0.20. The projected sample size was 115 patients per group after due consideration for potential dropouts.

Results

Over 6 months, 230 patients were enrolled in the study. Four patients were subsequently excluded because of major protocol violations. The groups were almost identical after stratification and blocking of patients (Table 1), and they had similar demographic characteristics (Table 2). Mivacurium was administered to 83 patients in the dexamethasone group and to 86 patients in the perphenazine group. The groups had almost identical lengths of stay in the PAR (43 vs 44 min, dexamethasone versus perphenazine) and the DCSU (234 vs 239 min, dexamethasone versus perphenazine). Analgesic use in the hospital was similar. In the dexamethasone group, 9 patients received morphine in the PAR and 105 patients received codeine in the DCSU, whereas in the perphenazine group, 8 patients received morphine in the PAR and 104 patients received codeine in the DCSU.

Forty percent of the patients vomited. The episodes of vomiting per patient ranged from 0 to 18, with two patients in each group having 10 or more episodes of vomiting. Perphenazine significantly reduced the incidence of in-hospital vomiting, but the incidence of out-of-hospital vomiting was almost identical (Table 3). Overall, the perphenazine group had a lower incidence of postoperative vomiting compared with the dexamethasone-treated group (P = 0.038) using logistic regression analysis (Table 3). By logistic regression analysis, both sex and induction technique were significant predictors of postoperative vomiting (P < 0.05, 35% vs 46%, male versus female patients, and IV versus inhalation induction). No other factors were identified as predictors of vomiting by logistic regression analysis.

Fifteen patients, nine of whom were in the dexamethasone group, were admitted or readmitted to the hospital after surgery. Most of these patients were admitted for surgical reasons, such as bleeding or refusing to drink, but two patients in each group required in-hospital management of emesis. Patients who vomited twice while they were in the hospital received a rescue antiemetic, dimenhydrinate, which was given to 15 patients in the dexamethasone group and to 8 patients in the perphenazine group for in-hospital vomiting (P = 0.1). Patient sex and induction technique did not affect the requirement for rescue antiemetic as determined using logistic regression analysis. Premedication did not alter the incidence of vomiting, with 39% of the patients who received premedication vomiting, whereas the incidence of vomiting among the remaining patients was 41%. No adverse events possibly attributable to dexamethasone or perphenazine, such as extrapyramidal symptoms and allergic reactions, were observed or reported by the nursing staff and/or parents in the current study.

Discussion

Compared with dexamethasone, perphenazine more effectively decreased the incidence of in-hospital vomiting by children after tonsillectomy. This increased antiemetic efficacy did not extend into the posthospital discharge period, in which dexamethasone and perphenazine seemed to have almost identical efficacy.

Many investigators have studied postoperative nausea and vomiting. Most of the studies involving children, like the current study, have reviewed vomiting in isolation, because it is difficult to assess nausea in this age group. Postoperative vomiting may be decreased by a variety of means, including prophylactic antiemetic drugs (antidopaminergic agents, gastric emptying agents, steroid administration, and sedatives), avoidance of food and water in the early post-operative period, avoidance of emetic drugs such as narcotics, and decreased length of surgery ^[10]. In the current study, all of these potential confounding factors were controlled within the study design.

This investigation used a blocked, stratified design to permit both control and flexibility. Patients were enrolled into this study before the parents and attending anesthesiologists had come to a decision as to the premedication, the anesthesia induction technique, and the need for muscle relaxants. The blocked and stratified approach leads to a balanced design, which simplifies intergroup comparisons, whereas flexibility in choice of premedication, induction, and muscle relaxant use optimizes clinical management.

Adverse effects with a single therapeutic dose of dexamethasone or perphenzine are extremely rare and generally minor. In Desilva et al.'s study of adults ^[8], only the patients who received perphenazine were free of adverse side effects. Perphenazine is a phenothiazine, a class of drugs that has a high therapeutic index and "are remarkably safe agents" ^[11]. There are no reported adverse side effects associated with the use of a single dose of dexamethasone, and there is only one report of dyskinesia after a single dose of perphenazine ^[12]. Also, there were no adverse events noted during this or previous studies ^[3,5].

Both perphenazine and dexamethasone decrease the incidence of vomiting by children after tonsillectomy compared with placebo ^[3,5]. Perphenazine decreased vomiting from 60% (placebo group) to 40%, and dexamethasone decreased vomiting from 72% to 40%. Thus, the current incidence of vomiting after tonsillectomy was similar to that observed in previous investigations. Although both of the study drugs decrease the incidence of vomiting, the incidence of emesis after prophylactic treatment is still excessive in this study. Ideally, the incidence of vomiting would be less than the 33% rate noted after perphenazine. Also, a decrease in the use of rescue antiemetics and the number of episodes of vomiting is desirable.

Both of the prophylactic antiemetics used in the current study are relatively inexpensive. The dexamethasone dose for a 20-kg child (3 mg) costs $0.15 Canadian, whereas perphenazine costs $0.81 Canadian for the treatment of a 20-kg patient. This is markedly less costly than a similar dose of an alternate antiemetic, ondansetron, which costs $12.90 Canadian for the 3-mg dose needed for a 20-kg child. Direct drug costs are not the only costs associated with vomiting. Treatment of emesis by nursing staff and family, delayed hospital discharge, and unscheduled admissions all add to the cost of care.

The doses of perphenazine and dexamethasone are extrapolated from the doses used in adult investigations. There are, to the best of our knowledge, no dose-response curves available for the antiemetic effects of these drugs in perioperative settings. Although the doses used in the current study have been shown to be effective, they are not necessarily the optimal dose.

In addition to the study intervention, two factors (sex and induction technique) were identified as significant predictors of postoperative vomiting by using logistic regression analysis. It was surprising to observe that male patients vomited less frequently than female patients after tonsillectomy. Further investigations need to determine whether this was a spurious result or a genuine observation. The beneficial effect of IV induction with propofol has been noted before and was not unexpected ^[3].

In conclusion, perphenazine more effectively decreased in-hospital vomiting after tonsillectomy in children within a daycare surgical setting, compared with dexamethasone. Perphenazine and dexamethasone seem to have similar effects on vomiting after discharge from hospital.