Effects of Anticholinergics on Postoperative Vomiting, Recovery, and Hospital Stay in Children Undergoing Tonsillectomy with or without Adenoidectomy 

After we received institutional review board approval and written informed parental consent, we studied children ages 3 to 16 yr who were classified as American Society of Anesthesiologists physical status I or II and undergoing tonsillectomy with or without adenoidectomy. Patients who regularly received any medication were excluded from the study. Patients were assigned randomly to one of the two groups (atropine or glycopyrrolate) using a computer-generated random number. Midazolam (0.5 mg/kg) was given orally as a premedicant only to those children whose parents expressed concern for their children's separation anxiety. General anesthesia was induced with 70% nitrous oxide, 30% oxygen, and increasing concentrations of halothane by mask. After an intravenous catheter was inserted, 0.5 mg/kg atracurium and 70 [micro sign]g/kg morphine were administered to facilitate endotracheal intubation. No additional atracurium or morphine was given. The depth of anesthesia was controlled and fluids were administered to maintain hemodynamic parameters within 20% of baseline during the surgical procedure. The study plan included treatment of any decrease in heart rate greater than 20% with 10 [micro sign]g/kg atropine and exclusion of the patient from the study. Ventilation was controlled in all patients with end-tidal carbon dioxide maintained between 30 and 40 mmHg. Intravenous fluid management included administration of lactated Ringer's solution. The fluid deficit was calculated to be replaced over 3 h and maintenance fluid was calculated according to patients' body weights. Blood loss was replaced with 3 ml lactated Ringer's solution per 1 ml blood loss.

After surgery was complete, the neuromuscular blockade was antagonized in all patients with neostigmine (60 [micro sign]g/kg, maximum of 5 mg) and glycopyrrolate (10 [micro sign]g/kg) or atropine (15 [micro sign]g/kg). Heart rate was recorded before and 5 min after administration of the study drug. The oral and nasopharyngeal cavities were suctioned before extubation. As a standard practice at our institution, a nasoorogastric tube was not passed in any patient. After the patients emerged from the anesthetic, an anesthesia research assistant, unaware of which study drug was used, assessed patient recovery and recorded the incidence of postoperative emesis and the number of patients who required postoperative antiemetic therapy and analgesics. The time from extubation until the patient met specific discharge criteria from the recovery room and the hospital was also recorded from the nursing notes. The criteria for discharge from the postanesthesia care unit (PACU) included alertness, ability to maintain a patent airway without assistance, stable vital signs, minimal bleeding, return of muscle strength, and satisfactory pain management. The criteria for discharge from the hospital were stable vital signs, temperature within normal range, unobstructed respiration, no vomiting within 30 min before discharge, adequate pain control, no active bleeding, and ability to drink clear liquids. The observer monitored each patient in the PACU and the ambulatory surgical center (ASC) to note the incidence of emesis (not retching) and telephoned each child's parents 24 h later to determine the incidence of emesis after discharge.

An episode of vomiting was defined as expulsion of stomach contents through the mouth, nose, or both. A single episode of vomiting or any number of episodes during a 24-h period constituted an episode of vomiting. A dose of antiemetic was given to patients after two or more episodes or at the request of a parent after one episode. The type of antiemetic medication was left to the discretion of the treating physician and included promethazine, trimethobenzamide hydrochloride, droperidol, or ondansetron.

The sample size was calculated using the Sigmastat statistical software program network version 1.01 (Jandel Scientific Software, San Rafael, CA), using alpha 0.05 and power 0.8 to detect a 30% difference in the incidence of emesis between the two groups, anticipating a 70% incidence of emesis in children undergoing tonsillectomy under general anesthesia. This yielded a sample size of 84 patients. To account for the possibility that the incidence of vomiting between groups might be smaller than expected, we enrolled a total of 93 patients. Data were analyzed using the Fisher exact test and chi-squared analysis when applicable and the Mann-Whitney rank sum test for nonparametric equivalents. Data are expressed as the mean +/− SD. P values <0.05 were considered significant.

Ninety-three children (42 girls and 51 boys) were enrolled in the study and 90 of them completed it. Two patients (one from each group) were excluded from the study because of bleeding in the PACU that necessitated a return to the operating room. One patient in the glycopyrrolate group who required intraoperative atropine for bradycardia was also excluded. Although these protocol deviations should not have resulted in patient exclusion, postoperative data regarding the incidences of vomiting were not collected in these patients. Table 1shows demographic characteristics and the duration of surgery.

(Table 2) shows the duration of PACU and ASC stay and the incidence of emesis. There was no significant difference in the discharge time from the PACU and ASC between the two groups. One patient from the atropine group and four patients from the glycopyrrolate group (P = 0.36) required overnight admission because of multiple episodes of vomiting and were excluded for the purpose of assessing ASC stay. The incidence of vomiting within 24 h after operation was reduced significantly in the atropine group compared with the glycopyrrolate group (P < 0.05). Of the 38 patients in the glycopyrrolate group and 24 patients in the atropine group who experienced emesis; 90%(n = 34) and 92%(n = 22), respectively, experienced their first episodes of emesis in the hospital. There was no statistical difference between the atropine or glycopyrrolate groups in the number of patients treated for postoperative emesis and pain (Table 3). The patients who vomited in this study, compared with those who did not, stayed longer in the PACU (31 +/− 12 min vs. 21 +/− 9 min; P = 0.008) and the ASC (384 +/− 167 min vs. 294 +/− 109 min; P = 0.004). All the patients in both the groups who had postoperative narcotics had emesis in the hospital or at home. There was no significant difference in change of heart rate 5 min after administration of atropine or glycopyrrolate (10 +/− 10 beats/min vs. 9 +/− 5 beats/min; P = 0.81).

The data from our study suggest that in children undergoing tonsillectomy and adenoidectomy or tonsillectomy alone, administration of atropine compared with glycopyrrolate to reverse neuromuscular blockade leads to a significantly lower incidence or postoperative vomiting. Unlike Salmenpera et al., [3]we confined our observations to emesis because of the difficulty of assessing nausea in small children. This antiemetic effect of atropine is consistent with the ability of atropine to cross the blood-brain barrier and to block afferent cholinergic pathways involved in emesis. [6]Glycopyrrolate, on the other hand, is a quarternary compound and does not significantly cross the blood-brain barrier. [7] 

The incidence of emesis in our study is similar to that reported by Litman et al. [8]Although a significantly lower incidence of emesis (20-33%) was reported in ambulatory pediatric patients, [9,10]the difference may have resulted from the type of surgery (tonsillectomy and adenoidectomy) and anesthetic management, especially neostigmine to reverse neuromuscular blockade and the use of nitrous oxide and morphine in all patients. In addition, we did not suction gastric contents at the end of the procedure, which may have increased the possibility of postoperative emesis caused by gastric distention from positive pressure ventilation and blood collecting in the stomach. [11,12]As reported by Schreiner et al., [13]it could also be attributed, in part, to the requirement that our patients drink clear liquids before discharge. Similar to the findings of Gold et al., [14]who reported unanticipated admission to the hospital because of postoperative nausea and vomiting, 8% of children who had emesis in our study needed unplanned admission to the hospital because of excessive emesis.

Anticholinergics (atropine and glycopyrrolate) are administered routinely to counteract the side effects of anticholinesterase drugs used to reverse neuromuscular blockade by nondepolarizing muscle relaxants. Watcha et al. [15]showed that patients who received a glycopyrrolate-neostigmine combination to antagonize neuromuscular blockade had a higher incidence of emesis compared with those who received atropine and edrophonium. These results are consistent with our data and could have resulted from the antiemetic effect of atropine, the emetic properties of neostigmine, and the inability of glycopyrrolate to cross the blood-brain barrier to provide an antiemetic effect.

We expected to find significantly shorter hospital stays in the group of patients with a lower incidence of emesis, but we observed no difference between the two groups. This could be explained in part by the lack of a strict protocol to treat postoperative emesis. Patients received different antiemetics as prescribed by their physicians. Although Schreiner et al. [13]found that drinking and retaining clear fluids is not a necessary requirement for hospital discharge, our ASC protocol, which requires that patients drink clear liquids before they are discharged, could have masked differences in discharge times.

The main limitation of this study is that the doses of atropine, glycopyrrolate, and neostigmine used were not selected on the basis of dose-response curves. We did not establish equipotent doses of atropine and glycopyrrolate. That we found no significant difference in the percentage change in heart rate 5 min after administration of atropine and glycopyrrolate in our study indicates that we were working with approximately equipotent doses of atropine and glycopyrrolate in terms of heart rate. This can also be debated because glycopyrrolate and atropine have differing time courses of action, and it is possible that we might find a difference with different measurement intervals. Furthermore, equipotency in terms of on heart rate may not represent equipotency in terms of gastrointestinal effects. Because we do not have relative shapes of dose-response curves for heart rate and gastrointestinal effects, a different dose comparison might result in an identical incidence of nausea and vomiting. The neostigmine dose was at the upper limit of that typically used, but it was standardized for both groups.

In conclusion, the use of atropine and neostigmine, compared with glycopyrrolate and neostigmine, to antagonize neuromuscular blockade is associated with a lesser incidence of postoperative emesis.

The authors thank Lisa Thompson, pharmacy technician, for assistance during the study and Ronald A. Gabel, M.D., for critically reviewing the manuscript.