A Dose Ranging Study of Dexamethasone for PreventingPatient-Controlled Analgesia-Related Nausea and Vomiting:A Comparison of Droperidol with SalineYi Lee, MD, Hsien-Yung Lai, MD, Pei-Chin Lin, MD, Youh-Sun Lin, MD,Shen-Jer Huang, MD, and Ming-Hwang Shyr, MD, PhD

From the Department of Anesthesiology, Buddhist Tzu-Chi Medical Center, Tzu-Chi University School of Medicine,Hualien, Taiwan

We designed this study to determine the minimumdose of dexamethasone for preventing nausea andvomiting associated with the use of morphine bypatient-controlled analgesia (PCA). Two hundred fortyfemale patients were randomly assigned to receivedexamethasone 2, 4, 8, or 12 mg IV immediately beforeinduction of anesthesia. Droperidol (0.1 mg/mL withmorphine 1 mg/mL in PCA pump) and saline wereused as controls. The complete response (no postopera-tive nausea and vomiting and no need for rescue anti-emetic for a 24-h postoperative period) rates for dexa-methasone 8 mg (72.2%) and 12 mg (78.9%) were

significantly more than for saline (42.9%) (P � 0.05).Patients who received dexamethasone 12 or 8 mg alsoreported higher patient satisfaction than those who re-ceived saline (P � 0.05). These results were as effectiveas adding droperidol 0.1 mg/mL to the morphine PCAwithout causing drowsiness, restlessness, or arrhyth-mias. Smaller doses of dexamethasone (4 or 2 mg) werenot effective for this propose. The results suggest thatdexamethasone 8 mg IV is the minimum effective dosefor the reduction of PCA morphine-related nausea andvomiting.

(Anesth Analg 2004;98:1066–71)

P atient-controlled analgesia (PCA) has been ex-tensively used for the treatment of postoperativepain. However, PCA-based opioid administra-

tion is associated with a frequent incidence of postop-erative nausea and vomiting (PONV). A systemic re-view suggested that the incidence of PONV with PCA-morphine is approximately 50% (1).

Droperidol is antiemetic even when small doses(�1 mg) are used (2). However, the effect is short-lived, suggesting that a repeated bolus of small dosesof droperidol should be used to achieve the best anti-emetic effect (2,3). Tramer and Walder (1) suggestedthat concomitant use of a small dose of droperidolwith morphine is an effective means to prevent nauseaand vomiting. Lamond et al. (4) demonstrated thatdroperidol 0.10 mg/mL with morphine 1 mg/mL in

the PCA container is the optimal dose for preventingPONV.

Dexamethasone has been effective in decreasingPONV after general anesthesia (5,6). Recently, we (7)have found that dexamethasone 8 mg administered IVimmediately before induction of anesthesia signifi-cantly decreases the incidence of PCA-morphine re-lated nausea and vomiting. Despite this antiemeticeffect, however, the optimal dose of dexamethasonefor this purpose has not been determined.

In this randomized, double-blind, placebo-controlledstudy, we evaluated the antiemetic effects of four dif-ferent doses of dexamethasone (2, 4, 8, and 12 mg) onPCA morphine-related nausea and vomiting. Droperi-dol (0.10 mg/mL) and saline were used as controls.

MethodsThe study was conducted with the appropriate insti-tutional approval, and written informed consent wasobtained from all patients. A total of 240 ASA phys-ical status I–II female patients scheduled for majororthopedic (total hip or knee replacement or correc-tive spinal procedures), general (laparotomy), or gy-necological surgery (abdominal total hysterectomy,

Accepted for publication October 23, 2003.Address correspondence to Ming-Hwang Shyr, MD, PhD, De-

partment of Anesthesiology, Buddhist Tzu-chi Medical Center,Tzu-chi University School of Medicine, No. 707, Sec 3, Chung-Yang Road, Hualien, 970, Taiwan. Address reprint requests to YiLee, MD, Department of Anesthesiology, Tzu-Chi Medical Cen-ter, Tzu-Chi University School of Medicine, no. 707, Sec 3,Chung-Yang Road, Hualien, Taiwan, ROC. Address e-mail todr.stone@msa.hinet.net or drleeyi@sinamail.com.

DOI: 10.1213/01.ANE.0000105875.05357.A0

©2004 by the International Anesthesia Research Society1066 Anesth Analg 2004;98:1066–71 0003-2999/04

myomectomy) under general anesthesia were enrolledin this randomized, double-blind, placebo-controlledstudy. Patients who had received any antiemetic med-ications in the 24 h before surgery or complained ofpreoperative nausea or vomiting were excluded. Indi-viduals with a history of immunosuppression, signifi-cant gastric reflux, peptic ulcer disease, diabetes mellitus,recent tuberculosis, Cushing’s syndrome, liver/renaldiseases, or for whom intubation during induction ofanesthesia was difficult were also deemed ineligible.Before the surgery, patients were taught to use the PCApump (Pain Management Provider; Abbott Laboratories,North Chicago, IL) and were informed that they couldrequest rescue antiemetics if necessary.

Patients were randomly assigned to 1 of 6 groups (n� 40 each) using a computer-generated random num-ber table. The study medication (3 mL), containing0.9% saline for the droperidol and saline-controlgroups and 2 mg, 4 mg, 8 mg, and 12 mg of dexameth-asone for the four dexamethasone groups, was admin-istered IV immediately before induction of anesthesia.Postoperatively, the dexamethasone and saline-control groups received PCA morphine 1 mg/mL. Thedroperidol-control group received the same concen-tration of morphine combined with droperidol0.1 mg/mL. All patients were blinded to the nature ofthe drug administered. The study medication prepa-rations were performed by a specially trained nurse-anesthetist who was not involved in any subsequentassessments.

The anesthesia was standardized for all patients.Glycopyrrolate 0.2 mg, fentanyl 2 �g/kg, and thiopen-tal 5 mg/kg were used for the induction of anesthesia.Tracheal intubation was facilitated by administrationof rocuronium 0.8 mg/kg. Anesthesia was maintainedby sevoflurane 2%–5% (inspired concentrations) and50% nitrous oxide in oxygen. Ventilation was mechan-ically controlled and adjusted to maintain end-tidalCO2 values between 4 and 5.3 kPa throughout thesurgery. Additional rocuronium was administered asrequired. All patients received ketorolac 30 mg IVapproximately 30 min before the end of surgery. Forreversal of residual muscle relaxation, the combina-tion of glycopyrrolate 0.6 mg and neostigmine 3 mgwere administered IV and the trachea was extubated.No opioids were given during the operation.

Postoperatively, all patients received an initial doseof morphine (0.05 mg/kg) immediately before initia-tion of PCA. The PCA device was programmed todeliver a 1-mL bolus of morphine solution (with orwithout addition of droperidol), with a 5-min lockoutinterval. No background infusion was used. Rescuemedication (metoclopramide 10 mg) was given IV ifthe patient experienced more than 15 min of nausea,had experienced an emetic episode, or if the patientrequested antiemetic medication. The treatment wasrepeated if necessary.

The incidence of PONV and the amount of mor-phine used were recorded by specially trained nurseanesthetists who were blinded as to the treatmentgroups at 2, 12, and 24 h after operation. Nausea wasmeasured using an 11-point numerical rating scalewith 0 � no nausea and 10 � nausea as bad as it canbe. A score of �8 was considered severe, 4 to 7 �moderate, and �3 � mild nausea. In this study, retch-ing and vomiting were grouped together under thecommon term “emetic episodes” (8). An emetic epi-sode was defined as vomiting/retching events occur-ring in rapid sequence within a 1-min period. If theinterval between 2 bouts of emesis exceeded 1 min,they were considered separate episodes. If there weremore than 4 episodes within the 24-h observationperiod, the emesis was considered severe (9).

Our primary efficacy end-point was complete re-sponse, defined as patients who stayed completelyfree from PONV and had no rescue antiemetic re-quirement during the first 24-h observation period.

Pain intensity was rated by the patients using an11-point numerical rating scale similar to that used fornausea, where 0 symbolized no pain and 10 repre-sented the worst pain imaginable. At the end of theobservation period, all patients would be askedwhether they were satisfied with the PCA procedure.

The satisfaction on PCA was also assessed using thesimilar 11-point numerical rating scale. A score of10 denoted the highest level of satisfaction, and pa-tient satisfaction was defined as high if the score was8 or more (10).

The severity of sedation was classified into 5 cate-gories: 0 � fully awake; 1 � drowsy, closed eyes; 2 �asleep, rousable; 3 � asleep, unrousable, answer totouch or pain; 4 � does not respond. Other side effectssuch as restlessness (felt nervous or jumpy), visualdisturbance, headache, or extrapyramidal symptomswere also considered at 2, 12, and 24 hpostoperatively.

Sample size calculation was performed before start-ing the trials by using a statistical power analysis.Based on an � error of 0.05 and � error of 0.20, 30patients were estimated to be needed in each group tohave a 90% chance with an error of 5% to detect adecrease in total PONV incidence from 60% to 40%after treatment (11). To compensate for patients notcompleting the study, we randomized 40 patients toeach group. Data were analyzed using one-way anal-ysis of variance with a linear contrast, �2 test withYates’ correction, and Mann-Whitney U-test as appro-priate. Data are expressed as mean values with sd ornumber and percentage. A P value �0.05 was consid-ered statistically significant. Commercial softwareSPSS 10.0 (SPSS Inc., Chicago, IL) for Windows wasused for data processing.

ANESTH ANALG PAIN MEDICINE LEE ET AL. 10672004;98:1066–71 DEXAMETHASONE FOR PCA-RELATED PONV

ResultsA total of 240 patients were recruited. Among thesepatients, 20 were later withdrawn for comparisons forthe following reasons: 2 patients in the droperidolgroup experienced excessive drowsiness and PCA hadto be discontinued; 2 patients underwent further sur-gery within 24 h; 6 patients had difficult intubation;and 10 patients had inadequate follow-up. The re-maining 220 patients completed this study. Patientcharacteristics, surgical types, duration of anesthesia,and numbers of risk factors for PONV (12), includingprior PONV, motion sickness, and nonsmokers, weresimilar in the 6 groups (Table 1).

The incidence of PONV, pain scores, morphine con-sumption, and sedation categories in each group in the0 to 2 h, 2 to 12 h, and 12 to 24 h postoperatively arereported in Table 2. The incidence of PONV in thedexamethasone 8 mg and 12 mg and droperidolgroups were significantly less frequent at all timesthan those of dexamethasone 2 mg and saline groups.No difference was found between dexamethasone8 mg and 12 mg and droperidol groups. Patients in the6 groups reported similar low pain scores for eachtime interval, and there was no significant differencein morphine consumption. The levels of sedation werenot different among the groups. Yet, 2 patients in thedroperidol group were asleep and only responded totouch (sedation category 3) and 4 patients experiencedrestlessness at 24 h postoperatively. At the end of thestudy period, no other side effects were observed inany of the groups.

Figure 1 shows the number of patients in mild (nau-sea score 1–3), moderate (nausea score 4–7) and severe(nausea score 8–10) nausea category and emesis epi-sodes in the 6 groups during the first 24 h observatoryperiod. The severity of nausea was no statisticallydifferent among groups. Yet, patients in the dexa-methasone 8 and 12 mg and droperidol groups re-ported a less frequent incidence of severe emesis

(emetic episodes �4; P � 0.046, 0.041, and 0.011 re-spectively) than did those in the saline group.

Table 3 demonstrates complete response rate (noPONV, no rescue antiemetic), the requirements of res-cue antiemetic, morphine consumption, and patientsatisfaction during the 24 h observation period. Pa-tients who received dexamethasone 8 and 12 mg anddroperidol produced a significantly higher completeresponse rate (P � 0.017, 0.002, and 0.001 respec-tively), less requirements for rescue antiemetics (P �0.011, 0.004 and 0.001 respectively), and higher satis-faction scores (P � 0.043, 0.035, and 0.001 respectively)than did those who received saline in the first 24 hafter anesthesia (Table 3). In addition, significantlymore patients treated with dexamethasone 8 or 12 mgor droperidol rated the experience as highly satisfac-tory (satisfaction score �8) in comparison to thosewho received placebo (P � 0.033, 0.019, and 0.004respectively). No significant differences were foundamong the dexamethasone 8 and 12 mg and droperi-dol groups. Dexamethasone 2 and 4 mg provided nobetter PONV protection than saline alone. The mor-phine consumption was not different among groups.

DiscussionMany published articles have suggested the use ofdexamethasone as a prophylactic antiemetic forPONV (6). Investigations have shown that the onsettime of dexamethasone on antiemesis is approxi-mately 2 h (5), and its biological half-life is 36 to 72 h(13). Thus, the delayed emesis (i.e., up to 24 h) is bettercontrolled with dexamethasone compared with classicantiemetics (6,14). Most patients receiving PCA-morphine vomited in the first 12–24 h postoperatively(15). Hence, in this study, dexamethasone was admin-istered immediately before the induction of anesthesiaand the data were collected in the first 24 h postoper-atively. We found that 57% of patients in the saline

Table 1. Clinical Characteristics, Surgical Types, and Duration of Anesthesia

Dexamethasone12 mg (n � 38)

Dexamethasone8 mg (n � 36)

Dexamethasone4 mg (n � 38)

Dexamethasone2 mg (n � 36)

Droperidol(n � 37)

Saline(n � 35)

Age (yr) 51 � 15 54 � 12 55 � 16 53 � 13 53 � 14 52 � 12Weight (kg) 51.4 � 7.4 52.5 � 6.8 52.8 � 7.3 53.2 � 7.4 51.7 � 8.3 53.6 � 7.6Risk factors

Previous PONV 9 (23.7) 7 (19.4) 6 (15.8) 9 (25.0) 10 (27.0) 8 (22.9)Motion sickness 18 (47.4) 16 (44.4) 19 (50.0) 20 (55.6) 19 (51.4) 15 (42.9)Nonsmoker 35 (92.1) 34 (94.4) 35 (92.1) 38 (88.4) 34 (91.9) 32 (88.6)

Surgical typeOrthopedic 12 (31.6) 11 (30.6) 12 (31.6) 10 (27.8) 13 (35.1) 10 (28.6)General 13 (34.2) 15 (41.7) 14 (36.8) 15 (41.7) 12 (32.4) 14 (40.0)Gynecologic 13 (34.2) 10 (27.8) 12 (31.6) 11 (30.6) 12 (32.4) 11 (31.4)

Duration of anesthesia (min) 143 � 29 138 � 31 148 � 23 150 � 33 145 � 35 135 � 28

Values are mean � sd or number (percentage). No significant differences among groups.PONV � postoperative nausea and vomiting.

1068 PAIN MEDICINE LEE ET AL. ANESTH ANALGDEXAMETHASONE FOR PCA-RELATED PONV 2004;98:1066–71

control group reported nausea and/or vomiting dur-ing the first 24 h after anesthesia. After receiving dexa-methasone 8 and 12 mg during induction, the inci-dence of PONV was reduced significantly to 27.8%and 21.1% respectively. Dexamethasone 2 and 4 mgwere ineffective for this purpose.

Study designs that analyze only one type of surgeryor restricted patient populations have recently been

criticized (16,17). In fact, large prospective investiga-tions have shown that the different incidences ofPONV are mainly caused by the associated risk factorsand less by the operation itself (17). Therefore, weused Apfel’s simplified risk score instead of selectingpatients undergoing just one type of surgery to iden-tify patients with an increased risk. In this study, thetypes of surgery and the number of risk factors weresimilar among groups (Table 1). We believe that thedifference in the six groups with respect to PONV wasdirectly related to the drug tested. We also found thatthe incidence of PONV in the droperidol and saline-controlled groups were comparable with previouslypublished data (1,4).

Droperidol is effective in reducing the incidence ofPONV when given concomitantly with a PCA devicewith morphine (1,4). However, the side effects in thepostoperative setting remains a concern even whenantiemetic treatment with droperidol is administeredin small doses (18–20). Lamond et al. (4) found thatthese previous patients were administered an initialloading bolus dose of droperidol as part of the anes-thetic technique, with the result that patients oftenreceived large doses of droperidol. Gan et al. (21)noted that administration of an initial loading dose ofdroperidol before addition of droperidol in PCAshould be avoided, as it resulted in more sedation.Lamond et al. (4) demonstrated that when given with-out an initial loading dose, a PCA bolus dose of

Table 2. Incidence of Postoperative Nausea and Vomiting, Morphine Consumption, Postoperative Pain Scores, andSedation Categories During Different Observatory Periods

Dexamethasone12 mg (n � 38)

Dexamethasone8 mg (n � 36)

Dexamethasone4 mg (n � 38)

Dexamethasone2 mg (n � 36)

Droperidol0.1 mg/mL

(n � 37)Saline

(n � 35)

0–2 hNausea 3 (7.9) 3 (8.3) 5 (13.2) 7 (19.4) 2 (5.4) 7 (20)Emesis 2 (5.3) 3 (8.3) 3 (7.9) 6 (16.7) 2 (5.4) 6 (17.1)Total 5 (13.2)*‡ 5 (13.9)* 8 (21.1) 13 (36.1)† 4 (10.8)* 13 (37.1)Morphine consumption (mg) 10 � 5 12 � 4 11 � 5 12 � 6 10 � 6 11 � 4Pain scores 2.5 � 1.2 2.6 � 0.8 2.3 � 1.1 2.4 � 1.0 2.5 � 0.9 2.6 � 1.0Sedation (0/1/2/3/4) 30/8/0/0/0 29/7/0/0/0 29/8/1/0/0 28/8/0/0/0 29/7/1/0/0 27/8/0/0/0

2–12 hNausea 2 (5.3) 3 (8.3) 5 (13.2) 8 (22.2) 2 (5.4) 7 (20)Emesis 2 (5.3) 2 (5.6) 4 (10.5) 6 (16.7) 2 (5.4) 7 (20)Total 4 (10.5)*‡ 5 (13.9)*‡ 9 (23.7) 15 (41.7)† 4 (10.8)* 14 (40)Morphine consumption (mg) 8 � 6 9 � 4 7 � 5 10 � 4 8 � 5 9 � 5Pain scores 2.2 � 0.9 2.4 � 1.3 2.0 � 1.2 2.1 � 1.1 2.0 � 0.8 2.2 � 1.2Sedation (0/1/2/3/4) 34/4/0/0/0 34/2/0/0/0 35/3/0/0/0 33/3/0/0/0 30/5/2/0/0 32/3/0/0/0

12–24 hNausea 2 (5.3) 1 (2.8) 3 (7.9) 5 (13.9) 1 (2.7) 5 (14.3)Emesis 1 (2.6) 2 (5.6) 3 (7.9) 5 (3.9) 2 (5.4) 5 (14.3)Total 3 (7.9)*‡ 3 (8.3)*‡ 6 (15.8) 10 (27.8)† 3 (8.1)* 10 (28.6)Morphine consumption (mg) 7 � 4 8 � 3 7 � 4 9 � 3 7 � 5 7 � 5Pain scores 1.9 � 1.1 1.8 � 0.8 1.8 � 0.9 2.0 � 0.8 1.9 � 0.9 1.8 � 0.8Sedation (0/1/2/3/4) 36/2/0/0/0 35/1/0/0/0 35/3/0/0/0 35/1/0/0/0 31/3/1/2/0 33/2/0/0/0

Values are mean � sd or number (percentage).Sedation categories: 0 � fully awake; 1 � drowsy, closed eyes; 2 � asleep, rousable; 3 � asleep, unrousable, answer to touch or pain; 4 � does not respond.P � 0.05 compared with * saline group, † droperidol group, and ‡ dexamethasone 2 mg group.

Figure 1. The number of patients having nausea and vomitingduring the first 24 h observatory period in patients receiving 12(D12), 8 (D8), 4 (D4), or 2 (D2) mg of dexamethasone, or droperidol(Droperidol) or saline (Saline) control. *P � 0.05.

ANESTH ANALG PAIN MEDICINE LEE ET AL. 10692004;98:1066–71 DEXAMETHASONE FOR PCA-RELATED PONV

droperidol of 0.1 mg/mL appears to provide the op-timal balance between antiemetic efficacy and an ac-ceptable incidence of side effects. In the current study,we found that a total dose of droperidol of between 2.0and 5.3 mg in 24 h was associated with an acceptablerate of PONV (18.9%), which was similar to dexameth-asone 8 and 12 mg. The levels of sedation were notdifferent among groups, although the small size of thegroups could also be the cause. Nevertheless, one insix patients (6 of 37 patients) in the droperidol groupreported side effects at 24 h after operation. Four ofthese patients had restlessness, a common droperidol-related side effect, after administration of 35, 36, 40,and 43 mg of morphine (3.5, 3.6, 4.0, and 4.3 mgdroperidol) respectively. None of these patients foundthe experience disturbing. However, PCA treatmenthad to be stopped in 2 patients after administration oftotal doses of 42 and 51 mg morphine (4.2 and 5.1 mgdroperidol) because of being asleep and only awak-ened by touch. We could not establish whether thesymptoms were caused by the droperidol, the mor-phine, or the combination of the two. However, Rob-erts et al. (19) found a significant increase in sedationassociated with PCA droperidol. Tramer and Walder(1) also noted that adverse events are likely when thedroperidol dosage exceeds 4–6 mg per day. Of note isthat these side effects were not reported in any of thedexamethasone-treated patients.

Droperidol has been used for the prophylaxis andtreatment of PONV for over 30 years (22). On Decem-ber 5, 2001, the US Food and Drug Administrationissued a “black box” warning regarding the risk ofcardiac arrhythmias during droperidol administration(23). However, White et al. (22) stated that there is noteven a single case report indicating that droperidol indoses used for the management of PONV has beenassociated with cardiac arrest or arrhythmias. Afterconducting an extensive literature research, we werealso unable to find a report of arrhythmias or cardiac

arrest associated with the use of droperidol with mor-phine PCA. In fact, continuing 12-lead electrocardio-gram (ECG) monitoring for all elective surgery pa-tients during the postoperative PCA period isclinically impractical. Thus, patients in the droperidolgroup in the current study were not monitored bycontinuous ECG in ward. Although no discerniblecardiac arrhythmia accompanying 0.1 mg/mLdroperidol in the PCA was found, further detailedinvestigation and a longer follow-up would be neededto prove this.

Baxendale et al. (24) reported that dexamethasonemight reduce pain intensity after tooth extraction. Inthis study, we found that both morphine consumptionand pain severity were similar in the dexamethasoneand saline control groups. These results indicate thatdexamethasone might not alter the intensity of painafter surgery, nor did it enhance the efficacy ofPCA-morphine.

Patient satisfaction should be an important consid-eration in antiemetic selection for PONV treatment.Fisher (25) indicated that patient satisfaction, ratherthan PONV, should be advocated as the main out-come. In this study, a higher satisfaction score wasreported for patients treated with dexamethasone 8 or12 mg or droperidol in comparison with those whoreceived saline. Further, more than half of the subjectsin these groups (dexamethasone 8/12 mg, 61.1/63.2%;droperidol, 70.3%) were highly satisfied (satisfactionscore � 8). With a similar pain score at all times in the6 groups, we believe that the degree of satisfactionwas mainly associated with the severity and incidenceof PONV during PCA. Hence, dexamethasone 8 or12 mg IV not only reduced the incidence of PCA-related PONV, but also provided a similar level ofpatient satisfaction as droperidol treatment.

In this study, patients who received dexamethasone8 or 12 mg or droperidol had less frequent incidence ofnausea than those who received saline (Table 2). How-ever, the differences were not statistically significant.

Table 3. Complete Response (no PONV, no Rescue Antimetic), Requirement for Rescue Antimetics, Pain Score andPatient Satisfaction in the First 24 h Postoperatively

Groups

Dexamethasone12 mg (n � 38)

Dexamethasone8 mg (n � 36)

Dexamethasone4 mg (n � 38)

Dexamethasone2 mg (n � 36)

Droperidol0.1 mg.ml�1

(n � 37)Saline

(n � 35)

Complete response 30 (78.9)*‡ 26 (72.2)*‡ 22 (57.9)† 15 (41.7)† 30 (81.1)* 15 (42.9)Rescue antiemetics 5 (13.2)*‡ 6 (16.7)*‡ 12 (31.6)† 17 (47.2)† 4 (10.8)* 16 (45.7)Morphine consumption 37.3 � 15.5 38.6 � 14.6 35.8 � 12.2 32.4 � 16.3 36.5 � 15.8 31.2 � 16.7Satisfaction score 7.5 � 1.6*‡ 7.4 � 1.5*‡ 6.9 � 1.4 6.6 � 1.3† 7.7 � 1.6* 6.1 � 2.3Satisfaction score � 8 24 (63.2)*‡ 22 (61.1)*‡ 17 (44.7)† 12 (33.3)† 26 (70.3)* 12 (34.3)

Values are mean � sd or number (percentage).P � 0.05 compared with * saline group, † droperidol group, and ‡ dexamethasone 2 mg group.Complete response � no postoperative nausea and vomiting, no rescue emetic.

1070 PAIN MEDICINE LEE ET AL. ANESTH ANALGDEXAMETHASONE FOR PCA-RELATED PONV 2004;98:1066–71

The a priori defined outcome in this study was the totalincidence of PONV and the sample size estimationassumed to be �60%. As the incidence of nausea in thesaline group was significantly less (25.7%) than thisestimation, the power of the study to detect suchdifferences was not sufficient. Thus, more patientswere needed to detect the same relative reduction innausea.

Although complications from corticosteroid use,such as delayed healing, additional wound infection,peptic-ulcer perforation, or adrenal suppression, areusually related to its long-term therapy, we avoidedits use in patients who were immunosuppressed orwho had peptic ulcer disease, diabetes, Cushing’s syn-drome, or recent tuberculosis to minimize the risk ofexacerbating their underlying diseases. In previousstudies, a single dose of dexamethasone was consid-ered safe (3–7,10). In this study, no apparent adverseeffects, such as peptic ulcer perforation, were noted.However, we do not know if a single dose of dexa-methasone would suppress adrenal function or haveclinical significance (for instance, if it would increasethe risk of wound infection) in healthy patients duringsurgical stress. This needs to be studied further.

In conclusion, this study has demonstrated that thepreinduction administration of dexamethasone 8 mgIV was the smallest effective dose for the reduction ofPCA morphine-related PONV. Moreover, this was aseffective as adding droperidol 0.1 mg/mL to the mor-phine PCA without causing drowsiness, restlessness,or arrhythmias.

References1. Tramer MR, Walder B. Efficacy and adverse effects of prophy-

lactic antiemetics during patient-controlled analgesia therapy: aquantitative systemic review. Anesth Analg 1999;88:1354–61.

2. Henzi I, Sonderegger J, Tramer MR. Efficacy, dose-response,and adverse effects of droperidol for prevention of postopera-tive nausea and vomiting. Can J Anaesth 2000;47:537–51.

3. Tramer MR. A rational approach to the control of postoperativenausea and vomiting: evidence from systemic reviews. Part 1.Efficacy and harm of antiemetic interventions, and methodolog-ical issues. Acta Anaesthesiol Scand 2001;45:4–13.

4. Lamond CT, Robinson DL, Boyd JD, Cashman JN. Addition ofdroperidol to morphine administered by the patient-controlledanalgesia method: what is the optimal dose? Eur J Anaesthesiol1998;15:304–9.

5. Wang JJ, Ho ST, Tzeng JI, Tang CS. The effect of timing ofdexamethasone administration on its efficacy as a prophylacticantiemetic for postoperative nausea and vomiting. AnesthAnalg 2000;91:136–9.

6. Henzi I, Walder B, Tramer MR. Dexamethasone in the preven-tion of postoperative nausea and vomiting: a quantitative sys-temic review. Anesth Analg 2000;90:186–94.

7. Lee Y, Lin YS, Chen YH. The effects of dexamethasone uponpatient-controlled-analgesia related nausea and vomiting. An-aesthesia 2002;57:705–9.

8. Korttila K. The study of postoperative nausea and vomiting. Br JAnaesth 1992;69:20S–3.

9. Wang JJ, Ho ST, Lee SC, et al. The use of dexamethasone forpreventing postoperative nausea and vomiting in females un-dergoing thyroidectomy: a dose-ranging study. Anesth Analg2000;91:1404–7.

10. Darkow T, Gora-Harper ML, Goulson DT, Record KE. Impact ofantiemetic selection on postoperative nausea and vomiting andpatient satisfaction. Pharmacology 2001;21:540–8.

11. Lerman J. Study design in clinical research: sample size estima-tion and power analysis. Can J Anaesth 1996;43:184–91.

12. Pierre S, Benais H, Pouymayou J. Apfel’s simplified score mayfavorably predict the risk of postoperative nausea and vomiting.Can J Anaesth 2002;49:237–42.

13. Haynes R. Adrenocortocotropic hormone: adrenocortical ste-roids and their synthetic analogs-inhibitors of the synthesis andactions of adrenocortical hormones. In: Goodman A, Gilman LS,Rall TW, Murad F, eds. The pharmacological basis of therapeu-tics. 8th ed. New York: Pergamon Press, 1990:1447–8.

14. Tavorath R, Hesketh PJ. Drug treatment of chemotherapy-induced delay emesis. Drugs 1996;52:639–48.

15. Woodhouse A, Mather LE. The effect of duration of dose deliv-ery with patient-controlled analgesia on the incidence of nauseaand vomiting after hysterectomy. Br J Clin Pharmacol 1998;45:57–62.

16. Kreisler NS, Spiekermann BF, Ascari CM, et al. Small-dosedroperidol effectively reduces nausea in a general surgical adultpatient population. Anesth Analg 2000;91:1256–61.

17. Apfel CC, Roewer N, Korttila K. How to study postoperativenausea and vomiting. Acta Anaesthesiol Scand 2002;46:921–8.

18. Woodhouse A, Mather LE. Nausea and vomiting in the postop-erative patient-controlled analgesia environment. Anaesthesia1997;52:770–5.

19. Roberts CJ, Millar JM, Goat VA. The anti-emetic effectiveness ofdroperidol during morphine patient-controlled analgesia. An-aesthesia 1995;50:559–62.

20. Barrow PM, Hughes DG, Redfern N, Urie J. Influence ofdroperidol on nausea and vomiting during patient-controlledanalgesia. Br J Anaesth 1994;72:460–1.

21. Gan TJ, Alexander R, Fennelly M, Rubin AP. Comparison ofdifferent methods of administering droperidol in patient-controlled analgesia in the prevention of postoperative nauseaand vomiting. Anesth Analg 1995;80:81–5.

22. White PF. Droperidol: a cost-effective antiemetic for over thirtyyears. Anesth Analg 2002;95:789–90.

23. Gan TJ, White PF, Scuderi PE, et al. FDA “black box” warningregarding use of droperidol for postoperative nausea andvomiting: is it justified? Anesthesiology 2002;97:287.

24. Baxendale BR, Vater M, Lavery KM. Dexamethasone reducespain and swelling following extraction of third molar teeth.Anaesthesia 1993;48:961–4.

25. Fisher DM. Surrogate end points, are they meaningful? Anes-thesiology 1994;81:795–6.

ANESTH ANALG PAIN MEDICINE LEE ET AL. 10712004;98:1066–71 DEXAMETHASONE FOR PCA-RELATED PONV