Postoperative nausea and vomiting (PONV) is a critical issue for both healthcare professionals and patients, especially in this era of ambulatory surgery. Although mortality is rare, PONV can result in serious complications  and has been shown to be a leading cause of unanticipated hospital admissions [2,3]. PONV also can significantly increase costs associated with surgery, particularly hospital costs related to increased recovery room time . For example, time of discharge of ambulatory surgical patients receiving general anaesthesia who experienced PONV was delayed 25% when compared with patients not experiencing PONV . Furthermore, results of a patient preference study of potential postoperative outcomes suggest that avoiding vomiting is more important to patients than avoiding incisional pain and avoiding nausea is nearly as important as avoiding pain .
Although published evidence suggests that universal PONV prophylaxis is not cost-effective, a recent evidence-based review of clinical trials and case series demonstrated that patients who are at moderate-to-high risk for PONV are likely to benefit . The review also identified primary risk factors for PONV in adults including female gender, smoking status, a history of motion sickness or PONV, the use of volatile anaesthetics or nitrous oxide, administration of intraoperative or postoperative opioids, longer surgical duration and certain types of surgical procedures. While the risk of PONV can be significantly reduced by modifying baseline factors and introducing other clinical measures, many patients are still likely to benefit from prophylactic antiemetic therapy. Recent recommendations include the use of monotherapy in patients at moderate risk of PONV and combination therapy with two or three agents in patients at high risk .
Among drugs available for PONV prophylaxis, selective serotonin (5-hydroxytryptamine type 3, 5-HT3) receptor antagonists are considered as a first choice due to their comparative safety and lack of serious side-effects . Although these agents are similar as a class in efficacy, the timing of administration and dose influence their value in this setting. The current dose recommendation for the 5-HT3 receptor antagonist granisetron in the prevention of PONV is 1 mg administered before induction of anaesthesia or immediately prior to the end of anaesthesia. Mikawa and colleagues  have demonstrated, however, that granisetron at doses of 5, 10 and 20 μg kg−1 (approximately 0.3, 0.6 and 1.2 mg, respectively, based on patient weights in each treatment group) administered immediately after the induction of anaesthesia were equally effective in preventing PONV. Similarly, doses of 1.0 and 3.0 mg were found to be equally effective in preventing PONV when administered just before induction . Thus, doses of granisetron in excess of 0.3 mg appear to offer no additional benefit when administered at induction. On the other hand, lower doses of 0.1 mg  and 2 μg kg−1 (approximately 0.1 mg, based on patient weight in the treatment group)  administered just prior to or immediately after induction were ineffective in preventing PONV.
Taken together, these studies suggest that 0.3 mg granisetron is the lowest dose with demonstrated efficacy in preventing PONV when administered at induction of anaesthesia. However, 0.1 mg granisetron has been shown to effectively treat existing PONV , while previous prevention studies have given the drug at induction rather than the end of surgery [6,7]. These findings raise the possibility that granisetron in doses of <0.3 mg may be effective if administered later than induction. To identify which granisetron dose should be further evaluated in a randomized, controlled trial, an exploratory pilot study compared the efficacy of 0.1, 0.2 and 0.3 mg doses of granisetron in preventing PONV when administered just prior to the end of surgery in patients at moderate risk of developing PONV. Given the exploratory nature of the research, no testing of the statistical significance of any differences identified was planned.
This randomized, double-blind, parallel-group, placebo-controlled, pilot study was conducted at 11 study centres in the USA. Approval was acquired from the Institutional Review Board at each participating centre, and written informed consent was obtained from each patient. Female patients aged 18 through 64 yr were eligible for enrollment if they were scheduled to undergo elective open abdominal hysterectomy requiring general anaesthesia and had an ASA physical status of I-III. Concomitant procedures were limited to salpingectomy, oophorectomy and lysis of adhesions. Patients were excluded from the study if they
- were scheduled to receive propofol for induction and/or maintenance of anaesthesia;
- had known hypersensitivity to granisetron, or any agent that was part of the anaesthesia regimen, or other medications to be administered relative to surgery;
- were pregnant or breast feeding;
- had experienced retching/vomiting or moderate or severe nausea in the 24 h prior to anaesthesia or suffered from chronic nausea and/or vomiting;
- had received an antiemetic drug or a drug with antiemetic properties within the 24 h prior to anaesthesia or were scheduled to receive such medications during surgery or within the 24 h following surgery;
- had a body mass index ≥34;
- had a condition that required chronic opioid use.
Patients were randomized at the time of surgery to one of four treatment groups: granisetron 0.1, 0.2 or 0.3 mg or placebo (0.9% saline solution). Treatment assignment was based on computer-generated randomization lists prepared by the study sponsor prior to the study. Separate randomization lists were generated for each study site using a prespecified block size. Patient randomization numbers were assigned sequentially at each site in the order in which patients were enrolled. The randomization list at each site was provided only to the unblinded pharmacist who prepared the injection. The pharmacist was not involved in any other part of the study. To maintain blinding, study medications were presented by the pharmacist to the investigator as identical 5 mL solutions for injection in 10 mL syringes.
On the day of the scheduled surgery, a baseline physical examination was completed, a urine pregnancy test was administered to patients of childbearing potential and a standardized anaesthesia regimen was followed. Premedication, if desired, was with intravenous (i.v.) midazolam 1 mg. General anaesthesia was induced with thiopental 4-5 mg kg−1 i.v. and maintained with isoflurane titrated between 0.6% and 2.5% and nitrous oxide ≥66%. Intubation and neuromuscular blockage was facilitated with rocuronium. Total intraoperative fentanyl doses were between 2 and 10 μg kg−1 i.v. The study medication was administered as a 30 s i.v. injection approximately 15 min prior to the end of surgery (time of closure). At the end of surgery, neuromuscular blockade was reversed by neostigmine 0.07 plus glycopyrolate 0.01 mg kg−1 i.v. The i.v. morphine (1-4 mg doses) was provided as needed for the management of postoperative pain. All patients were under continuous medical supervision for at least 24 h postoperatively.
The time of each vomiting episode and the time and intensity of each nausea episode were recorded for 24 h following administration of study medication. A vomiting episode was defined as either vomiting (expulsion of stomach contents) or retching (an involuntary attempt to vomit but not productive of stomach contents). Vomiting and/or retching separated by <5 min were recorded as a single episode. The intensity of each nausea episode was graded as mild (discomfort noticed but no disruption of anticipated normal activity), moderate (discomfort sufficient to reduce or affect anticipated normal activity) or severe (inability to work or perform anticipated normal daily activity). Rescue medication was administered at the discretion of the investigator according to customary and usual clinical practice. If rescue medication was used during the study, both the time of administration and type of medication used were recorded.
Adverse events were assessed and recorded by the investigator throughout the observation period. Vital signs were recorded as part of the baseline physical examination, 24 h after dosing, and when clinically indicated.
The primary efficacy end-point was the proportion of patients with no vomiting in the 0-6 h interval following administration of study medication. Secondary efficacy parameters included the proportion of patients with no vomiting in the 0-24 h interval following administration of study medication; proportions with no moderate or severe nausea and with no rescue medication during the 0-6 and 0-24 h intervals; time to first use of rescue medication; and the proportion of patients with total control (no moderate or severe nausea, no vomiting and no rescue medication) in the 0-6 h interval. Additional post hoc analyses compared groups with regard to average times to first vomiting episode in patients having vomiting and to first moderate or severe nausea event in those with nausea.
As this was an exploratory pilot study, no inferential statistics were planned and results are summarized using descriptive statistics. It was expected that 30 patients per treatment group would be sufficient to observe a monotonic dose-response relationship among granisetron doses and placebo, as well as provide information concerning response rates. Up to 150 patients were to be enrolled in the study in order to ensure a minimum of 120 evaluable patients, with at least 30 evaluable patients expected in each of the three parallel granisetron dose groups and the placebo group.
In all, 140 patients scheduled for abdominal hysterectomy were enrolled in the study between 12 March and 16 August 2002 at 11 study centres in the USA. Of these, 121 patients were randomized, treated with study medication and included in efficacy and safety analyses. One patient in the granisetron 0.1 mg treatment group was withdrawn from the study prematurely due to additional surgery requiring repeat anaesthesia approximately 2 h after the original surgery was completed.
Patient characteristics and other baseline characteristics are summarized in Table 1. For some variables such as history of motion sickness or smoking, randomization did not result in equal distributions across treatment groups. The most frequently administered rescue medications were ondansetron (n = 64) and promethazine (n = 41); no imbalances in types of medication administered across groups were noted.
The proportion of patients with no vomiting episode in the 0-6 h interval after administration of study medication, the primary efficacy end-point, was >90% in each granisetron treatment group compared to 77% in the placebo group (Table 2). As expected in this pilot study, 95% confidence intervals (CI) were relatively wide. No dose-response relationship between the use of granisetron and the prevention of postoperative vomiting was observed, with all three ascending doses of granisetron showing a trend to improvement relative to placebo. Also at 0-6 h and compared with placebo-treated patients, more patients in each granisetron group had no moderate or severe nausea, fewer required rescue medication and more achieved total control. The mean time to first use of rescue medication was >3 times longer in each granisetron group compared to placebo. In the subgroups of patients experiencing vomiting or moderate/severe nausea, these events were delayed in granisetron-treated patients. At 0-24 h trends for improvements were observed for granisetron-treated patients compared to placebo with respect to proportions of patients with no moderate or severe nausea and no rescue medication (Table 2).
Proportions of patients reporting at least one adverse event were similar across treatment groups and as follows: placebo (90%), granisetron 0.1 mg (84%), granisetron 0.2 mg (78%) and granisetron 0.3 mg (88%). The most common adverse events were gastrointestinal disorders, headache and injection-related reactions. Adverse events reported in ≥10% of patients in any treatment group are summarized in Table 3. Most adverse events were judged by investigators to be unrelated to study medication and mild or moderate in intensity. No serious drug-related adverse events were reported, and no patient was withdrawn from the study due to adverse events. No clinically meaningful changes in vital signs were seen during the study.
This pilot study is the first randomized, prospective, controlled trial to examine the dose-response relationship between granisetron administered just prior to the end of surgery and the prevention of PONV. In the first 6 h following administration, 14-19% more patients treated with granisetron (0.1, 0.2 and 0.3 mg) than with placebo had no vomiting, and more than 90% of patients in each granisetron group had no vomiting. The timing as well as the occurrence of vomiting appears to have been influenced by granisetron. In patients with vomiting, the average time to the first such episode was ≥8.5 h in each granisetron group, more than twice the average of 3.9 h in the placebo group. As would be expected given the findings concerning vomiting, fewer patients in each granisetron group experienced moderate or severe nausea during the 0-6 h interval compared to placebo. Among patients experiencing moderate or severe nausea, the time to the first nausea episode was longer in those treated with granisetron. Further supporting granisetron's positive effect on PONV, 20-30% more patients treated with granisetron than placebo had total control during the first 6 h following administration.
The use of rescue medication probably impacted rates of vomiting and nausea in both the active treatment and placebo groups in the first 6 h following surgery, but several findings suggest that the effect was greater in placebo-treated patients. Rescue medication was required by fewer patients receiving granisetron (approximately one-third vs. 60%, respectively), and the mean time to administration of rescue medication was >12 h in all granisetron groups compared to 4 h in the placebo group. Types of rescue medication used were relatively consistent across treatment groups making it unlikely that this variable affected overall rates of and/or times to episodes of either vomiting or nausea.
This pilot study has not resolved the issue of whether the beneficial effect of granisetron extends into the 6-24 h time period. The duration of efficacy of 0.1-0.3 mg doses of granisetron in preventing PONV may be similar to that of other selective serotonin receptor antagonists [9,10]. Gan and colleagues  reported that ondansetron (4 mg) was effective in preventing postoperative nausea only in the first 6 h after anaesthesia. Paxton and colleagues  found ondansetron (4 mg) was effective in reducing nausea only in the first 4 h post-anaesthesia. We found that the proportion of patients in the placebo group with no vomiting remained relatively unchanged between the 0-6 and 0-24 h assessments, while proportions of patients receiving granisetron with no vomiting declined. Whether the 0.1-0.3 mg doses of granisetron have beneficial effects over the full 24 h requires further study. Unmeasured factors including differences in amounts of narcotic, levels of pain, times to first oral intake and times to initiation of ambulatory activities also may affect the incidence of PONV over the longer term.
In addition to addressing the issue of the longer-term effect of granisetron, future research should include larger numbers of patients to overcome several shortcomings of the present pilot study. First, a larger study would provide the power needed to evaluate whether a dose-response effect exists for granisetron. Second, a larger sample would be expected to yield more even distributions of baseline characteristics across treatment groups and would allow statistical adjustments to be made for uneven distributions. Finally, data from a larger sample of patients would allow the effects of rescue therapy use on other outcomes to be more carefully evaluated. Based on previous research by Wilson and colleagues , we expected approximately half of placebo-treated patients to have no vomiting in the 0-6 h interval; however, 77% of patients in our placebo group actually had no vomiting. This unexpectedly high rate in our study reflects, at least in part, the early use of rescue medication by many placebo-treated patients; Wilson and colleagues  did not report mean time to rescue medication. In general, rescue medication use complicates efficacy assessments in studies of antiemetics, but ethical standards require that such agents be made available at the discretion of the investigator.
More comprehensive research based on larger samples of patients also might evaluate potential cost savings associated with granisetron therapy. Studies could prospectively evaluate the cost-effectiveness in the 0-6 and 6-24 h intervals of granisetron monotherapy and of combination therapies that include both low-dose granisetron and agents from other therapeutic classes. This pilot study found granisetron at doses of 0.1, 0.2 and 0.3 mg administered at the end of surgery suggested a trend of improved efficacy compare to placebo in preventing PONV in the first 6 h after abdominal hysterectomy. A dose-response relationship was not demonstrated. A larger comparative trial is underway to confirm the lowest effective dose of granisetron.
This research was supported by Roche Laboratories Inc, Nutley, New Jersey. The authors thank Douglas Eberhardt, MS, MBA for statistical support; Mike Andria, PharmD, Ko-Chin Khoo, MS, Leena Patel-Shah, PharmD, Anna Chan, PharmD and Kelly Celuch, PharmD for overall study management, design and data analysis; and Jane G. Murphy, PhD and Linda E. Whetter, PhD for editorial assistance.
The members of the Granisetron Dose-ranging Study Group comprised the following: Brigham and Women's Hospital: John Abbatematteo, RN, Joseph Garfield, MD, Robert Knapp, MD, Naila Moghul, MD, James Philip, MD and Lee Silk, MD; Duke North Hospital: Terrance W. Breen, MD, Holly Ann Muir, MD and Adeyemi Olufolabi, MD; Magee Women's Hospital: Derek Davis, MD and Helene Finegold, MD; Memorial City Hospital: John Gerard Baerenstecher, MD; University of Alabama at Birmingham: Gwendolyn Boyd, MD, Pankaj Desai, MD, James Hunter, MD, Holly Richter, MD, Douglas Barry Shaw, MD and Diana Wilhite, RN; University of Kansas Medical Center: Karl E. Becker, Jr, MD, Kirk T. Benson, MD, Jack Galen Bray, MD, Anna Breiberg, RN, Terry L. Chaffee, MD, Deanna K. Fox, MD, Joyce Goldstein, MD, Hiroshi Goto, MD, Peter G. Hild, MD, James D. Kindscher, MD, Grace H. Shih, MD, Stephen D. Tarver, MD, Ronald L. Torline, MD and Gregory K. Unruh, MD; University of Texas Medical School: Ralf Gebhard, MD, Jennifer R. Greger, MD, Maria Matuszczak, MD, Didier A. Sciard, MD and Navtej Singh Tung, MD; Wake Forest University Baptist Medical Center: Terrence Bogard, MD, Laura Dean, MD, David Dewan, MD, James Eisenach, MD, Lynnette C. Harris, RN, CCRC, David Hood, MD, Kenneth Nelson, MD, Medge Owen, MD, Peter H. Pan, MD, Vernon Ross, MD, Philip E. Scuderi, MD and John Thomas, MD.
1. Gan TJ, Meyer T, Apfel CC et al
. Consensus guidelines for managing postoperative nausea and vomiting
. Anesth Analg
2. Fortier J, Chung F, Su J. Unanticipated admission after ambulatory surgery - a prospective study. Can J Anaesth
3. Gold BS, Kitz DS, Lecky JH, Neuhaus JM. Unanticipated admission to the hospital following ambulatory surgery. JAMA
4. Chung F, Mezei G. Factors contributing to a prolonged stay after ambulatory surgery. Anesth Analg
5. Macario A, Weinger M, Carney S, Kim A. Which clinical anesthesia outcomes are important to avoid? The perspective of patients. Anesth Analg
6. Mikawa K, Takao Y, Nishina K, Shiga M, Maekawa N, Obara H. Optimal dose of granisetron
for prophylaxis against postoperative emesis after gynecological surgery. Anesth Analg
7. Wilson AJ, Diemunsch P, Lindeque BG et al
. Single-dose i.v. granisetron
in the prevention
of postoperative nausea and vomiting
. Br J Anaesth
8. Taylor AM, Rosen M, Diemunsch PA, Thorin D, Houweling PL. A double-blind, parallel-group, placebo-controlled, dose-ranging, multicenter study of intravenous granisetron
in the treatment of postoperative nausea and vomiting
in patients undergoing surgery with general anesthesia. J Clin Anesth
9. Gan TJ, Ginsberg B, Grant AP, Glass PS. Double-blind, randomized comparison of ondansetron and intraoperative propofol to prevent postoperative nausea and vomiting
10. Paxton LD, McKay AC, Mirakhur RK. Prevention
of nausea and vomiting after day case gynaecological
laparoscopy. A comparison of ondansetron, droperidol, metoclopramide and placebo. Anaesthesia
1For the Granisetron Dose-ranging Study Group (see Appendix).