The Safety and Efficacy of Prophylactic Ondansetron in Patients Undergoing Modified Radical Mastectomy

Abstract

We aimed to evaluate the antiemetic efficacy, safety, and clinical utility of prophylactic ondansetron administered at the end of the surgery for the prevention of postoperative nausea and vomiting (PONV) in a homogenous population of 54 women undergoing modified radical mastectomy (MRM). A standard general anesthetic and perioperative analgesic technique were used. After surgery, patients received either saline placebo or ondansetron 4 mg IV. Episodes of PONV, as well as rescue antiemetic requirements, were recorded for the first 24 h after surgery. The 24-h incidence of PONV (33.3% vs 81.5%; P = 0.0010) was significantly lower in the ondansetron group. The severity of PONV, evaluated by the number of emetic episodes per patient (1.59 ± 1.90 vs 0.29 ± 0.66; P = 0.0029), and the rescue antiemetic requirement (59.2% vs 14.8%; P = 0.0019) was significantly lower, in the ondansetron group. Patient satisfaction scores and number needed to prevent PONV (2.07) were significantly better and therapeutically more favorable in the ondansetron group. The incidence of adverse events such as headache, dizziness, and increased liver enzyme levels (number needed to harm = ∞) was similar in both groups. Administered at the end of the surgery in adult female patients undergoing general anesthesia for MRM, ondansetron 4 mg is effective and safe in preventing PONV. We recommend the clinical practice of routine prophylactic ondansetron to prevent PONV after MRM, as it significantly improves perioperative patient satisfaction and outcome.

Implications 

We evaluated the antiemetic efficacy, safety, and routine use of prophylactic ondansetron, a “gold standard” antiemetic, in women undergoing radical breast surgery who were at a high risk of postoperative vomiting. We analyzed more meaningful “true” and “therapeutic” outcome measures, and we conclude that prophylactic ondansetron is safe and effective and that its routine use is justified.

Postoperative nausea and vomiting (PONV) are common and distressing symptoms after surgery performed under general anesthesia. Ondansetron, a 5-HT₃ antagonist, is the “gold standard” antiemetic because of its safety and efficacy compared with alternatives. Most studies regarding ondansetron and PONV have reported only surrogate measures, such as the incidence of PONV and the number of emetic episodes per patient, rather than the more clinically meaningful “true” (nonsurrogate) outcome measures, such as patient satisfaction, duration of hospital stay, and incidences of unanticipated hospital admissions (¹). Moreover, a recent meta-analysis has challenged the practice of “routine” prophylactic ondansetron to prevent PONV based on risk (3% headache and 3% increased postoperative liver enzymes) and benefit data (²). Breast surgery is associated with frequent PONV, with an incidence as high as 60% (³). Female patients more often experience PONV that is more severe than that experienced by male patients (^5,6). Moreover, the phase of the menstrual cycle influences the incidence of PONV (⁷). No clinical trial has evaluated the efficacy, safety, and clinical utility of routine prophylactic ondansetron administered at the end of the surgery in a homogenous adult female population undergoing modified radical mastectomy (MRM). We aimed to assess the safety and efficacy of ondansetron with true and surrogate outcome measures, in addition to outcome measures that help physicians to assess the clinical utility of a drug, such as the number needed to prevent (NNTP) and the number needed to harm (NNTH). We also tried to correlate PONV after MRM with postchemotherapy nausea and vomiting (PCNV) and with phases of the menstrual cycle.

Methods

With institutional review board approval and informed patient consent, we studied 54 ASA physical status I or II adult female patients undergoing MRM for carcinoma of the breast. Patients who had morbid obesity, who had gastrointestinal diseases with gastric outlet obstructions, or who had received drugs with an antiemetic effect within 24 h before anesthesia were excluded. However, patients with previous PONV and motion sickness were not excluded. Patients were allocated to receive, in a randomized, double-blinded manner, a single dose of either ondansetron 4 mg (4 mL) IV or saline at an equal volume (4 mL) as a placebo at the end of the surgery.

All patients were premedicated with oral diazepam 0.2 mg/kg the night before surgery and 2 h before the induction of anesthesia. With standard monitoring, anesthesia was induced with thiopental 5 mg/kg IV, and vecuronium 0.1 mg/kg IV was used to facilitate tracheal intubation. Meperidine 1 mg/kg IV was administered as an intraoperative analgesic. Anesthesia was maintained with 66% nitrous oxide and 1%–2% isoflurane (inspired concentration) in oxygen. Supplements of vecuronium and meperidine were administered if required. At the end of the operation, the study drug was administered. Residual neuromuscular block was reversed with neostigmine 50 μg/kg and glycopyrrolate 10 μg/kg IV. The stomach was suctioned, and the trachea was extubated. Patients were transported to the postanesthesia care unit.

Postoperative analgesia was provided by diclofenac sodium 75 mg IM for moderate pain and by tramadol 25 mg IV for severe pain. We recorded all episodes of nausea and vomiting in the first 24 h postoperatively at the intervals of 0–2 h, 2–6 h, and 6–24 h. All patients were evaluated by using a PONV numeric scoring system (0 = no nausea or vomiting, 1 = nausea alone, 2 = vomiting once, 3 = vomiting two or more times in 30 min). A PONV score of 3 or persistent nausea (>2 h) was defined as severe PONV and was treated with 150 μg/kg IV metoclopramide as a rescue antiemetic. If metoclopramide failed to control emesis, ondansetron 4 mg was used as the rescue antiemetic of second choice. After the first 24 h postoperatively, patients were asked to assess their satisfaction throughout the first 24 h by using a verbal numeric scoring system (0 = not satisfied to 10 = fully satisfied).

Patients received preoperative neoadjuvant chemotherapy with cyclophosphomide 600 mg/m², adriamycin (doxorubicin) 40 mg/m², and 5-flurouracil 600–750 mg/m². In patients with ischemic heart disease, epirubicin 60–70 mg/m² was used instead of doxorubicin. All patients underwent six cycles of chemotherapy. Surveillance for PCNV was not conducted objectively. We simply asked the patients who received neoadjuvant chemotherapy whether they had PCNV (“yes” or “no”) and whether it was severe enough to require antiemetics (“yes” or “no”).

Adverse events throughout the first 24 h postoperatively were also recorded. On the fifth or sixth postoperative day, blood samples were taken from all patients to compare aspartate aminotranaminase (AST), alanine aminotranaminase (ALT), and alkaline phosphatase (APT) levels with corresponding levels from the immediate preoperative period.

Prestudy power analysis determined a sample size of 25 for each group to have an 80% chance (β = 0.20) of detecting a 40% reduction in PONV from a basal incidence of >50% at the 95% confidence interval limits (α = 0.05) (⁴). Two sample t-tests and Mann-Whitney U-tests were used to compare the age, weight, duration of anesthesia, perioperative analgesia requirements, time of first PONV, patient satisfaction scores, and preoperative and postoperative liver enzyme levels. ASA physical status, incidence of PONV at various intervals, requirements of rescue antiemetic therapy, and correlation of PONV with PCNV and phases of menstrual cycle were analyzed between the groups by using χ² tests with Yates’ continuity correction and Fisher’s exact tests wherever appropriate. The positive NNTP PONV indicates how many patients had to be exposed to ondansetron to prevent PONV in one patient who would have had PONV had she received placebo. Mathematically, NNTP is equivalent to the reciprocal of the absolute risk reduction. The NNTP and NNTH were calculated as the reciprocals of absolute risk reductions of incidence of PONV and side effect (increased postoperative liver enzymes), respectively, for patients exposed to prophylactic ondansetron. P values <0.05 were considered statistically significant, and data are presented as mean ± SD unless otherwise specified.

Results

There were no significant differences between the groups with regard to age, weight, ASA physical status, duration of anesthesia, and intraoperative and postoperative analgesic requirements (Table 1). The 24-h incidence of PONV and the incidence of PONV at 0–2 h and 2–6 h were significantly less frequent in patients who received prophylactic ondansetron (Table 2). The incidence of PONV at 6–24 h was comparable between the groups (Table 2).

Table 1: Demographic and Clinical Data

Table 2: Incidence and Severity of PONV and Therapeutic and Nonsurrogate Outcome Measures

The number of emetic episodes per patient and the requirement for rescue antiemetics were significantly lower in the ondansetron group (Table 2). Patients in the ondansetron group remained PONV-free longer after surgery than patients in the control group (Table 2). No differences in the incidence of other adverse events, such as headache and dizziness, were observed between the groups (Table 3).

Table 3: Correlation of PONV with PCNV and Menstrual Phase

The NNTP for ondansetron was 2.07, which indicates that ondansetron prevented PONV in one of every two patients exposed to ondansetron (Table 2). The NNTH for increased liver enzymes was infinity (∞), which indicates that ondansetron did not produce any extra harm compared with placebo. Patients in the ondansetron group reported superior satisfaction throughout the first 24 h postoperatively compared with those in the control group (Table 2).

There was no correlation between PONV and PCNV. Although the incidence of PONV in those who had PCNV in the ondansetron group (50%) was lower than that in the control group (80%), this did not attain statistical significance (Table 3). The incidence of PONV in the ondansetron group in patients at the postovulatory phase of the menstrual cycle was significantly lower than that in the placebo group. Although the incidence of PONV in the ondansetron group in menstruating, postmenopausal, and amenorrhic patients was less than that in placebo group patients, it did not attain statistical significance (Table 3).

The incidence of PONV in patients who did not receive prophylactic ondansetron was higher (81.48%) than we had assumed at the beginning of the study. We performed a post hoc power analysis to validate our study results. Our study has >90% (93.7%) power (1 − β) to detect a relative risk of 0.41 (0.23–0.72) and a prevented fraction of 59.1% (28.2–76.7) with 95% confidence interval limits (α = 0.05).

Discussion

In our study, the incidence of PONV in the placebo group was 81.5%, which is higher than that previously reported after breast surgery (^8,9). Less frequent incidences of PONV in previous studies (^8,9) may be due to the use of atropine for premedication (¹⁰) and for reversal of neuromuscular blockade, along with neostigmine (¹¹). It has been reported that reversal of neuromuscular block with atropine and neostigmine is associated with less PONV compared with glycopyrrolate and neostigmine (¹¹). By not requiring neostigmine for reversal of neuromuscular blockade, mivacurium (which was not available to us) has advantages over vecuronium in reducing PONV. However, the authors of a systemic review (¹³) reported that omitting antagonism of a neuromuscular block has a non-negligent risk of residual paralysis, even with short-acting neuromuscular blocking drugs (NNTH 30). These factors explain the difference in the incidence of PONV between our patients and those in previous studies (^8,9). Our groups were similar with respect to patient characteristics (age, sex, previous PONV, phase of menstrual cycle), type of surgery, anesthetics administered, and analgesics used perioperatively. Therefore, the difference in the incidences of PONV between the groups can be attributed to differences in the study drugs.

Ondansetron is effective in preventing PONV after a variety of surgical procedures, including high-risk PONV procedures (^14–19). However, the authors of one meta-analysis (²) have challenged the clinical utility of prophylactic ondansetron in preventing PONV based on efficacy and safety. However, validity of meta-analyses has been questioned (²⁰) because they are apt to reflect publication bias and nonuniformity of collected data, and they might magnify the problems of individual studies (²¹). Moreover, meta-analyses on PONV have the additional weakness that they depend on published studies that have reported only surrogate outcome measures (¹). The outcome measures that help in therapeutic decision making—such as NNTP, NNTH, and number needed to improve patient satisfaction (NNTS)—might be seriously misinterpreted if calculated in a meta-analysis from the pooled data of studies performed on different patient populations with different PONV risks, with different antiemetic doses and different anesthetic techniques, and for different reasons. One editorial has justified a cautious approach to adopting clinical practice guidelines based only on meta-analytic techniques and concludes that large randomized controlled trials remain the gold standard for determining the best choice among different therapeutic options (²¹).

We used a more clinically meaningful, nonsurrogate (true) outcome measure (patient satisfaction) and therapeutic outcome measures (NNTP and NNTH) to evaluate the efficacy, safety, and clinical utility of prophylactic ondansetron. Ondansetron was administered in the recommended dose of 4 mg (^23,24) at the end of the surgery, as it has been shown that prophylactic ondansetron is more effective when administered after the surgical procedure than before the induction of anesthesia (^4,22). Patient satisfaction during the postoperative experience was superior in the ondansetron group. To assess the cost-benefit of routine prophylactic ondansetron, we used the NNTS. The improvement in postoperative patient satisfaction with the prevention of PONV achieved with prophylactic ondansetron was 2 (NNTS 2). This indicates that every alternative patient who received prophylactic ondansetron had a higher level of satisfaction (patient satisfaction scores ≥7.5) than those who received the placebo (cost-to-benefit ratio 2:1). If we consider the wholesale acquisition cost of ondansetron 4 mg as $16, the direct cost associated with the improvement in the satisfaction of one of every two patients (NNTS 2) is $32. A study that assessed the monetary benefits of prophylactic ondansetron intervention using the “willingness to pay” method reported that more than half (80%) of the high-risk PONV patients undergoing outpatient laparoscopy were willing to pay an additional cost (>$50) to prevent PONV if they underwent a similar procedure in the future (²²). Ondansetron may be criticized for its higher acquisition cost. If we aim to improve the perioperative patient outcome, drug (ondansetron) cost containment is no longer an option; rather, it might be necessary to lower the direct and indirect cost of care to patient and hospital (global cost of care). An evaluation of three multicenter clinical trials has justified using prophylactic ondansetron in patients at a high risk of PONV, even at a higher cost, because of its benefits in terms of efficacy and safety (²³). The nonsurrogate outcome measures (duration of postanesthesia care unit stay and unanticipated hospital admission after discharge) were not studied to assess outcome, as all of our patients were in patients.

The positive NNTP for PONV, which indicates how many patients had to be exposed to ondansetron to prevent PONV in one of them who would have had PONV had she received placebo, is becoming widely used as a tool for therapeutic decision making. In their meta-analysis, Tramer et al. (²) showed that if the risk of PONV is very high (40%–80%), the NNTP with ondansetron is 5–6. In our trial, the NNTP for ondansetron was 2.07, which is far superior to the best NNTP of the meta-analysis (²). Although a meta-analysis can provide an overall estimate of therapeutic efficacy, it may obscure and dilute the differences among trials, resulting in misinterpretation of the benefits and risks of a drug. With our present knowledge, it is very difficult to predict which one of the two patients (NNTP 2) will benefit from prophylactic ondansetron. If we aim to achieve complete elimination of PONV, prediction of PONV will not be a matter of concern. Prophylactic ondansetron is a worthwhile means by which to achieve the goal of a PONV-free postoperative period. In the postoperative period, one patient in the control group (pre/post AST 30/53, ALT 28/64, APT 227/205) and one patient in ondansetron group (pre/post AST 32/52, ALT 42/70, APT 196/187) had an asymptomatic increase in postoperative AST and ALT levels. The patient in the control group who had increased enzyme levels did not require ondansetron as a rescue antiemetic at any time in the postoperative period. The mean liver enzyme levels in the preoperative and postoperative period were comparable between groups (Table 1). The NNTH for increased postoperative liver enzymes (↑LE) was infinity (∞), which indicates no extra harm to patients exposed to ondansetron. These two therapeutic outcome measures, NNTP (PONV) and NNTH (↑LE), demonstrate that prophylactic ondansetron administered at the end of MRM is effective and safe (Table 2).

The lack of significant PONV control with ondansetron 6–24 hours postoperatively may be due to its relatively short half-life, but the average operating time for MRM is approximately 2.5 hours (mean surgical time in this trial 152.40 ± 30.51 minutes). To derive the best benefit from prophylactic ondansetron, it should be administered at the end of the surgery (^4,22). More potent and longer-acting antiemetics with fewer side effects might produce a better outcome. Although granisetron has been shown to be effective in breast surgery (^8,9), in these trials, the type of surgical procedure was not homogenous, and atropine was used for premedication and reversal of neuromuscular block (^8,9), which might have significantly influenced the effect of granisetron on the incidence and severity of PONV.

We could not demonstrate any relationship between PONV and PCNV, as the number of patients who received chemotherapy in our study was inadequate. Although ondansetron reduced the incidence of PONV at different phases of the menstrual cycle, because of the small number of patients at each phase, we could not demonstrate a statistical difference, except a significant reduction in women in the postovulatory phase (Table 3). A study with more power might show a statistical significance.

Our results are in accordance with those of Tang et al. (²²), who reported that the prophylactic administration of ondansetron may offer advantages over administering a rescue antiemetic after PONV develops and reserving ondansetron for patients in whom the rescue antiemetic was unsuccessful. However, whether routine antiemetic prophylaxis in certain patients at high risk of PONV actually has a superior effect on “true” outcome measures, compared with timely treatment of PONV (²⁵), has not been adequately studied. A study that addressed this controversial issue after outpatient surgery concluded that prophylactic ondansetron is no better than early symptomatic treatment (²⁶); however, there are many serious concerns about this study that might explain its failure to report a benefit with prophylactic ondansetron over the early symptomatic treatment. First, ondansetron, with a short elimination half-life of 2.8 ± 0.6 hours, has been reported to be more effective if administered at the end of the surgery (^4,22). With the duration of anesthesia >60 minutes and ondansetron administered at anesthetic induction (²⁶), one would expect a reduced efficiency of prophylactic ondansetron due to loss of effective antiemetic duration in the intraoperative period (wasted antiemesis). Second, the true outcome measure, patient satisfaction, is not merely an all-or-none (“yes” or “no”) outcome; it is more subjective. More accurate presentation of patient satisfaction scores (such as mean ± SD) in that study (²⁶) might have unmasked the real clinical advantage of prophylactic ondansetron. Third, although the study reported significantly greater patient satisfaction in the prophylactic group compared with the treatment group, especially in a subgroup of patients at high risk of PONV undergoing more emetogenic surgical procedures (²⁶), by choosing NNTS to assess the clinical significance of this statistical difference, the authors reported that the NNTS was 25 and the direct cost to benefit a patient with prophylactic ondansetron was $400. This is more than 10 times higher than what we observed ($32). The lack of a clinical benefit of prophylaxis in the study (²⁶) is probably due to inclusion of all subgroups of patients, with varying risks of PONV, in the calculation of the NNTS, instead of stratifying the NNTS to the subgroups, which would dilute and obscure the benefit (statistically and clinically greater patient satisfaction) of prophylactic ondansetron in a high-risk PONV subgroup undergoing highly emetogenic procedures with that of other subgroups with a relatively lower risk of PONV.

Routine ondansetron prophylaxis may not completely eliminate PONV after MRM in women—although it is a worthwhile goal—but it improves the true outcome measure, patient satisfaction, and has a favorable influence on therapeutic outcome measures such as NNTP, NNTH, and NNTS. For adult female patients undergoing MRM, which carries a high risk of PONV, after assessing the risks and benefits of using clinically and therapeutically more meaningful outcome measures, we justify and recommend using routine prophylactic ondansetron 4 mg at the end of the surgical procedure until a better alternative is found.