The Effect of Remifentanil or Fentanyl on Postoperative Vomiting and Pain in Children Undergoing Strabismus Surgery

Abstract

Postoperative vomiting (POV) after strabismus surgery in children results in discomfort and prolonged hospital stays. Opioids increase the incidence of POV. Remifentanil has a context-sensitive half-life of 3 to 4 min, and how this short half-life influences POV in those patients is unknown. We conducted a prospective, double-blinded study in 81 ASA status I or II children from 2 to 12 yr of age undergoing elective strabismus surgery under general anesthesia. Patients were randomized to receive either remifentanil (bolus 1 μg/kg; infusion 0.1–0.2 μg · kg⁻¹ · min⁻¹) or fentanyl (2 μg/kg, and 1 μg/kg every 45 min). POV episodes were recorded for 25 h. Pain scores were obtained by using an objective pain scale for 60 min during recovery. The number of patients who experienced POV did not differ significantly between groups (49% vs 48%). However, in the Remifentanil group, POV episodes were significantly less frequent (0.95 vs 2.2 episodes). In contrast, fentanyl was associated with lower pain scores during the first 30 min of recovery. We conclude that children undergoing strabismus surgery under balanced anesthesia with remifentanil, compared with fentanyl, showed less frequent POV. However, early postoperative analgesia was better with fentanyl.

Children undergoing strabismus surgery frequently experience postoperative vomiting (POV). The incidence ranges from 15% to 75%, depending on the anesthetic technique ⁽¹⁾, antiemetic treatment ⁽²⁾, length and specifics of the surgery ⁽³⁾, and patient-related risk factors ⁽⁴⁾. Postoperative nausea and vomiting contributes to patients’ discomfort ⁽⁵⁾ and unanticipated hospital admissions after day-case surgery ⁽⁶⁾.

The use of perioperative opioids increases the incidence of POV significantly ^(7,8). In addition, the amount of opioid administered seems to affect the incidence ⁽⁹⁾. In children undergoing strabismus surgery, the use of IV morphine given on the induction of anesthesia increased the incidence of POV from 19% to 71% in comparison to ketorolac ⁽¹⁰⁾.

Remifentanil has been recently introduced into anesthetic practice. Its methyl ester linkage makes the drug susceptible to metabolism by nonspecific blood and tissue esterases, with a context-sensitive half-life of 3 to 4 min ⁽¹¹⁾. Whether the short half-life of remifentanil influences the incidence or the time course of POV in comparison with longer-acting opioids in children undergoing strabismus surgery is unknown. Therefore, we performed a prospective, randomized, double-blinded study in children undergoing strabismus surgery under general anesthesia to investigate the effect of IV remifentanil on POV and postoperative pain in comparison to fentanyl.

Methods

After obtaining approval from the hospital’s institutional review board, 81 ASA status I and II children aged 2 to 12 yr (42 boys and 39 girls) who were scheduled for elective strabismus surgery under general anesthesia were enrolled in a prospective, double-blinded study. Written informed consent was obtained from the parents. Patients with a history of POV or motion sickness or who were at risk for malignant hyperthermia were excluded from the study. Patients were randomized by lottery to receive either remifentanil or fentanyl during anesthesia.

All children were medicated with oral midazolam 0.4 mg/kg, up to a maximal dose of 10 mg, before surgery. After IV access was obtained, and after placement of electrocardiogram, noninvasive arterial blood pressure and oxygen saturation monitors, atropine 10 μg/kg was given IV. The doses of remifentanil and fentanyl were elected on the basis of studies by Song et al. ⁽¹²⁾ and Guy et al. ⁽¹³⁾. Remifentanil was given IV as a bolus of 1 μg/kg over 2 min to achieve stable hemodynamics during tracheal intubation and was followed by a continuous infusion of 0.2 μg · kg⁻¹ · min⁻¹. The infusion rate was decreased to 0.1 μg · kg⁻¹ · min⁻¹ after the incision was made, was kept constant throughout the surgery, and was discontinued at the closure of the conjunctiva. In the Fentanyl group, patients received a bolus of 2 μg/kg at induction, followed by 1 μg/kg every 45 min in the course of the operation. After the bolus of remifentanil or fentanyl was given, anesthesia was induced via mask induction with sevoflurane in oxygen at an inspired concentration of 6% until loss of consciousness. Patients were then paralyzed with 0.6 mg/kg of rocuronium, and tracheal intubation was performed. After the induction was complete, rectal acetaminophen 10 mg/kg was given. Anesthesia was maintained with sevoflurane at an end-tidal concentration of 2.5% in air/oxygen. The preoperative fluid deficit was calculated and replaced. After the conjunctiva was closed, the volatile anesthetic was discontinued. No antiemetic prophylaxis was given. Before emergence, muscle relaxation was reversed if no sustained tetanus was present. With the patient spontaneously breathing 100% oxygen and with adequate airway reflexes present, the trachea was extubated, and the patient was transferred to the recovery area.

Patients were discharged to their rooms after 60 min of monitoring in the recovery area if their room air saturation was more than 96%, vital signs were stable, and a calm and cooperative mental status was achieved. It is surgical practice at this institution to hospitalize patients after strabismus surgery for >24 h. In their rooms, oral fluid was allowed 2 h after the completion of surgery upon the child’s request. If tolerated, the diet was advanced to solids. If vomiting occurred, 40–70 mg of dimenhydrinate was given rectally according to the patient’s body weight. The dose was repeated after 2 h if vomiting still persisted. Postoperative pain therapy consisted of rectal acetaminophen 10 mg/kg or 0.4% oxybuprocaine eye drops ⁽¹⁴⁾ for pain scores >3 and 5, respectively (see below).

The nurse taking care of the patient documented vomiting over 25 h after surgery. Parents were instructed to call for a nurse immediately if vomiting occurred, in case the nurse was not present. Vomiting was defined as expulsion of gastric contents. A new episode was recorded if the patient did not vomit for a 10-min interval. Pain was recorded at 15, 30, 45, and 60 min in the recovery area by using an objective pediatric pain scale ⁽¹⁵⁾¹ consisting of zero- to two-point scores for five criteria, including blood pressure, crying, movement, agitation, and verbal responses or body language. For a pain score >3, a second dose of 10 mg/kg of acetaminophen was given rectally. With a pain score >5, the patient received 0.4% oxybuprocaine eye drops additionally. Patients and their parents, as well as the person obtaining the pain score and recording POV, were blinded to the administered opioid.

Similarity of treatment groups was analyzed by comparing odds and means of patients’ demographic data, anesthesia and surgical time, and pain scores. The dependence of POV on age, sex, the number of operated muscles, anesthesia and surgical times, and treatment, as well as the interaction between age and sex, was analyzed by logistic regression. The interaction effect investigates the sex difference in age dependency of POV incidence.

Occurrence of POV was the primary outcome variable of this trial. In a study in children undergoing strabismus surgery, fentanyl (0.1 μg/kg) was associated with an approximately 50% more frequent incidence of POV compared with placebo (16). Assuming a similar incidence of POV in the Remifentanil group as after placebo, power analysis demonstrated that 37 patients would be required in each treatment group to achieve a significance level of 5% and a power of 80%. Forty patients were thus randomized to each treatment group.

The mean number of POV episodes was estimated by maximal likelihood under the assumption that the number of episodes was Poisson distributed. A logistic regression model was fitted from the occurrence of POV to group and logarithm of hour after surgery, allowing for lower POV levels during the night. The probability of POV was explained by a logistic regression including several factors. We assumed an exponential decay of POV probability over time; this was modeled by including the logarithm of time in the model, thus assuming a Hill function. Furthermore, we estimated the change points for a reduced POV probability. Treatment groups could differ in initial POV and rate of POV decrease. Because the distribution of pain scores was positively skewed, we analyzed their logarithms, i.e., computed geometric means. An exponential trend was assumed with Score 0 as an asymptote in regressing on the logarithm of time. Both the intercept and slope of the two regression lines could be different for the two treatment groups.

Results

Patient demographics, number of corrected muscles, anesthesia, and surgical time did not differ between groups (Table 1). Sustained tetanus >5 s was present in all patients at the end of the surgery; therefore, muscle relaxation was not reversed in any patient. All patients were discharged without complications from the recovery area after 60 min of emergence.

Table 1: Patient Demographics

The incidence of POV and the use of rescue medication were similar in both groups (Remifentanil, 49%; Fentanyl, 48%). The logistic regression of POV incidence to known predictors of POV (age, sex, duration of surgery and anesthesia, and number of operated muscles), treatment, and the interaction between age or sex revealed no significant effects (smallest P = 0.45, for sex). However, the probability of experiencing a POV episode was 2.3 (95% confidence interval, 1.6 to 3.5) times higher after fentanyl compared with remifentanil (2.2 vs 0.95 episodes). In the 25 h after the surgery, the relative frequency of POV decreased over time (P = 0.0002) in both groups and was reduced during the night (P < 0.0001) with change points 9:10 pm and 7:35 am. Initially, the occurrence of POV episodes was significantly more frequent in the Fentanyl group (P = 0.004), but nighttime reductions and decrease over time did not differ significantly between the groups (P = 0.8 and P = 0.09, respectively) (Fig. 1).

Figure 1: Probability of postoperative vomiting (POV) in the 25 h after anesthesia. Multiple logistic regression (lines) with two change points from the occurrence of POV to the logarithm of time, a nighttime indicator, treatment group, and time/group interaction. Dots indicate the relative frequency of patients with POV in hourly intervals after the end of anesthesia. POV decreased over time (P = 0.0002) and was reduced during the night (P < 0.0001). Treatment groups had different POV levels initially (P = 0.004).

Mean pain score (P < 0.0001) and rate of decrease (P = 0.0004) differed significantly between treatment groups (P < 0.0001 for intercept and P = 0.002 for rate). Patients in the Remifentanil group had higher pain scores initially (4.7 vs 2.0 in the Fentanyl group) and showed a steeper decline of pain scores over time (e.g., to 2.6 vs 1.9 at 15 min in the Fentanyl group). The confidence intervals for the geometric mean pain scores predicted by the analysis of covariance model (Fig. 2) do not overlap at 0 and 15 min. We conclude that the differences pertain for only approximately 15 to 30 min.

Figure 2: Pain scores (objective pain scale) of patients on arrival in the recovery room over time. Predicted values from the analysis of covariance of log(score + 1) × group, log(time + 1), and group/time interaction are connected by bold lines. Patients in the Remifentanil group had significantly higher pain scores initially (solid bold line) compared with patients in the Fentanyl group (broken bold line). Thin lines indicate the 95% confidence limits.

Discussion

This study investigated the effect of remifentanil on POV and postoperative pain in children undergoing strabismus surgery after general anesthesia compared with fentanyl. Previous studies investigating the incidence of POV after general anesthesia with remifentanil have yielded conflicting results ^(17,18). Our study demonstrates that the intraoperative use of remifentanil in children undergoing strabismus surgery is associated with a significantly decreased frequency of POV episodes compared with fentanyl. However, pain scores after remifentanil were higher during the first 30 minutes of recovery.

Several studies have investigated the effect of remifentanil on POV in children. Pinsker and Carroll ⁽¹⁹⁾ studied the incidence of POV after dental restoration in children under general anesthesia with desflurane, comparing a remifentanil group (0.2 μg · kg⁻¹ · min⁻¹) with an opioid-free treatment group. The incidence of POV in patients having received remifentanil was not increased. Even though no comparison with a longer-acting opioid was performed in this study, the authors attributed the infrequent incidence of POV in the Remifentanil group to the short half-life of the drug. Rusch et al. ⁽¹⁷⁾ compared total IV anesthesia with remifentanil and propofol with a purely volatile anesthetic regimen in children undergoing strabismus surgery. Patients in the Remifentanil group received a remifentanil bolus of 0.5 μg/kg followed by an infusion of 0.05 μg · kg⁻¹ · min⁻¹ in combination with propofol; patients in the Volatile group were anesthetized with sevoflurane and nitrous oxide. The authors found a significantly smaller incidence of POV in the Remifentanil/Propofol group. Although the results of this study could be attributed to a proemetic effect of nitrous oxide (and in the Volatile group to an antiemetic effect of propofol as well), it seems that remifentanil does not dramatically increase POV. Davis et al. ⁽¹⁸⁾ compared remifentanil (bolus of 1 μg/kg followed by an infusion of 1 μg · kg⁻¹ · min⁻¹) with propofol, alfentanil, or isoflurane in children undergoing strabismus surgery. They could not demonstrate a difference in POV among the four therapeutic groups. Their results suggest that there is no difference between an opioid-based anesthetic regimen with short-acting opioids (remifentanil or alfentanil) and anesthetic regimens avoiding the use of opioids (isoflurane or propofol). However, the use of IV fentanyl (0.1 μg/kg) given during the induction of anesthesia in children undergoing strabismus surgery was associated with a fourfold increase in POV episodes as compared with ketorolac or placebo ⁽¹⁶⁾. All these studies suggest that remifentanil might have a less pronounced proemetic effect compared with the longer-acting opioid fentanyl.

In our study, the use of remifentanil was associated with a significant decrease in the number of POV episodes. Although the number of children who developed POV was similar in both groups, the patients in the Remifentanil group developed on average 0.95 POV episodes during the 25-hour study period, compared with 2.2 episodes in the Fentanyl group. However, the effect of this finding on clinical outcome variables such as time to discharge or number of unexpected hospital readmissions could not be addressed in this study, because all patients remained in the hospital for at least 24 hours after discharge from the recovery area.

To investigate the time course of POV, a logistic regression model was used. The probability of experiencing a POV episode was initially increased in the Fentanyl group. Nevertheless, the time course and the reduction of the probability to experience a POV episode decreased in both groups (Fig. 1). A relationship between the time course of POV and the short half-life of remifentanil could not be demonstrated from the data of this study.

It is interesting to note that the model reveals an almost complete elimination of POV during nighttime in both groups. There are currently no studies available investigating the course of POV during rest at nighttime in children. During sleep, proemetic factors such as stimulation of the vestibular system by body movement, pain, or diet are absent. Also, a change of optical axis, which might affect POV after strabismus surgery, is eliminated during sleep ⁽²⁰⁾. The results of this study suggest a circadian distribution of POV episodes, with an almost complete arrest during night hours.

Patients in the Remifentanil group showed higher pain scores after extubation. Because patients were treated with 0.4% oxybuprocaine eye drops and rectal acetaminophen according to their pain scores, no differences between groups could be found after 30 minutes. Therefore, the used analgesic therapy seemed to be effective to control pain after strabismus surgery.

In conclusion, the data presented in this study demonstrate that the number of POV episodes after strabismus surgery in children is smaller with use of IV remifentanil compared with fentanyl. No relationship between the time course of POV and the short half-life of remifentanil could be demonstrated. Finally, the infrequent incidence of POV during nighttime suggests an antiemetic effect of sleep compared with an awake state in postoperative patients.