Cancellation of elective operations on the day of intended surgery in a Hong Kong hospital: point prevalence and reasons.

OBJECTIVE: To determine the point prevalence of elective surgical case cancellations and the reasons. DESIGN: Cross-sectional study. SETTING: Teaching hospital, Hong Kong. PATIENTS; Operating theatre records of elective surgery cancellations from 1 January 2009 to 31 December 2009 were retrospectively reviewed. MAIN OUTCOME MEASURES. Cancellation of scheduled elective surgery on the day of surgery and the corresponding reasons. RESULTS: Of 6234 cases scheduled, 476 were cancelled, which yielded a point prevalence of 7.6%, with a 95% confidence interval of 7.0-8.3%. The highest number of cancellations occurred in patients scheduled for major general surgical procedures (n=94, 20%), major urological procedures (n=64, 13%), major orthopaedic surgery (n=38, 8%), and ultra-major cardiothoracic surgery (n=29, 6%). The most common category for cancellation was facility (73%), followed by work-up (17%), patient (10%), and surgeon (1%). No available operating room time due to overrun of the previous surgery was the most common reason for case cancellation (n=310). Compared to general surgery, the odds of no available operating time was significantly less in orthopaedics (odds ratio=0.26; 95% confidence interval, 0.17-0.39), otolaryngology (0.25; 0.13-0.46), neurosurgery (0.36; 0.16-0.70), paediatrics (0.53; 0.31-0.87), gynaecology (0.18; 0.11-0.29), ophthalmology (0.19; 0.07-0.41), and dentistry (0.10; 0.00-0.60). CONCLUSIONS. Case cancellations were mainly due to facility factors, such as no operating room time being available. The odds of having no operating room time available varied between surgical specialties.

Utility of a preoperative assessment clinic in a tertiary care hospital.

OBJECTIVES: To assess the utilisation rate of a preoperative assessment clinic and its impact on length of stay and discharge destinations. DESIGN: Retrospective case series with internal comparisons. SETTING: A tertiary hospital in Hong Kong. PATIENTS: All medical records of elective surgical admissions to a hospital in Hong Kong from April to June 2008 were retrieved. Medical records of patients who did not attend the preoperative assessment clinic were further reviewed by surgeons to assess if the patients could have been referred to the clinic. MAIN OUTCOME MEASURES: Total length of stay, preoperative and postoperative length of stay, and the discharge destinations of the patients attending and not attending the clinic were compared. RESULTS. In all, 640 patients underwent elective operations, of whom 22 (3%) patients were seen in the preoperative assessment clinic. In patients who had a major operation, the mean (standard deviation) total length of stays for clinic attenders and non-attenders were: 5.2 (3.6) versus 13.2 (18.8) days (P<0.001). The respective figures for preoperative and postoperative length of stay were: 1.3 (2.3) versus 4.5 (8.9) days (P=0.001), and 3.9 (2.9) versus 8.7 (14.5) days (P<0.001). For patients who had an intermediate operation, the respective mean (standard deviation) length of hospital stays were 2.4 (2.0) versus 7.3 (13.9) days (P=0.002) and the figures for postoperative length of stays were 1.3 (0.5) versus 4.5 (9.3) days (P=0.001). Surgeons had classified 108 (17%) of the cases as possible preoperative assessment clinic users. Among the latter, 71 (66%) had no special reason to stay in the hospital. The discharge destination was not associated with the use of preoperative assessment clinic for patients having major (Chi squared=0.18, P=0.912) or intermediate (Chi squared=0.34, P=0.468) operations. CONCLUSION: Successful implementation of preoperative assessment clinic service requires close collaboration between surgeons, anaesthetists, clinicians, and also the re-engineering of health service delivery.

Real-time ultrasound-guided paramedian epidural access: evaluation of a novel in-plane technique.

BACKGROUND: Current methods of locating the epidural space rely on surface anatomical landmarks and loss-of-resistance (LOR). We are not aware of any data describing real-time ultrasound (US)-guided epidural access in adults. METHODS: We evaluated the feasibility of performing real-time US-guided paramedian epidural access with the epidural needle inserted in the plane of the US beam in 15 adults who were undergoing groin or lower limb surgery under an epidural or combined spinal-epidural anaesthesia. RESULTS: The epidural space was successfully identified in 14 of 15 (93.3%) patients in 1 (1-3) attempt using the technique described. There was a failure to locate the epidural space in one elderly man. In 8 of 15 (53.3%) patients, studied neuraxial changes, that is, anterior displacement of the posterior dura and widening of the posterior epidural space, were seen immediately after entry of the Tuohy needle and expulsion of the pressurized saline from the LOR syringe into the epidural space at the level of needle insertion. Compression of the thecal sac was also seen in two of these patients. There were no inadvertent dural punctures or complications directly related to the technique described. Anaesthesia adequate for surgery developed in all patients after the initial spinal or epidural injection and recovery from the epidural or spinal anaesthesia was also uneventful. CONCLUSIONS: We have demonstrated the successful use of real-time US guidance in combination with LOR to saline for paramedian epidural access with the epidural needle inserted in the plane of the US beam.

Surgeons' attitudes and perception of an acute pain service.

OBJECTIVE: To evaluate the attitude and perception of surgeons about postoperative pain management, and an anaesthesiologist-based acute pain service. DESIGN: Questionnaire survey. SETTING: Tertiary university teaching hospital, Hong Kong. PARTICIPANTS: All surgical staff members (specialists and trainees) of the Departments of Surgery, Orthopaedics and Traumatology, and Obstetrics and Gynaecology. MAIN OUTCOME MEASURES: Opinions on postoperative pain management, different pain management modalities, and services provided by the acute pain service. RESULTS: Of the 147 questionnaires, 104 (71%) were returned. The majority (97%) agreed that effective pain control improves patient recovery and 88% believed that anaesthetists should be involved in postoperative pain management. Overall, 85% of the respondents were satisfied with the acute pain service. However, about one third of them wanted to maintain an active role in postoperative pain management and only 54% thought that the acute pain service has a significant impact on patient outcomes. In addition, only 10% of surgeons agreed that patients receiving acute pain service intervention would be discharged earlier. The respondents also thought that, compared to intravenous patient-controlled analgesia, epidural analgesia required more nursing care and was less cost-effective. Areas of the acute pain service warranting improvement included: education of surgeons on postoperative pain and its management (92%), communication (74%), and referral systems (80%). CONCLUSION: The majority of surgeons were satisfied with the acute pain service and agreed that anaesthetists should be involved in postoperative pain management. However, a proportion wanted to maintain an active role in postoperative pain management.

Characteristics, management process, and outcome of patients suffering in-hospital cardiopulmonary arrests in a teaching hospital in Hong Kong.

OBJECTIVES: To examine the demographics, process indicators of adult in-hospital cardiopulmonary arrest resuscitation, and outcomes in a teaching hospital in Hong Kong. DESIGN: Retrospective study. SETTING: A university-affiliated tertiary referral hospital with 997 acute adult beds in Hong Kong. PATIENTS: Those who suffered a cardiopulmonary resuscitation event, as documented in retrieved records of all in-patients during the inclusive period January 2002 to December 2005. RESULTS: There were 531 resuscitation events; the mean (standard deviation) age of the corresponding patients was 70.7 (15.4) years. Most (83%) occurred in non-monitored areas and most (97%) were cardiopulmonary arrests. The predominant initial rhythm was asystole (52%); only 8% of patients had ventricular tachycardia/fibrillation. All the resuscitations were initiated by on-site first responders. The median times from collapse to arrival of the resuscitation team, to defibrillation, to administration of adrenaline, and to intubation were: 5 (interquartile range, 2-6) minutes, 5 (1-7) minutes, 5 (3-10) minutes, and 9 (5-13) minutes, respectively. The overall hospital survival (discharge) rate was 5%. The survival rate was higher among patients in monitored areas (9 vs 4%, P=0.046), among patients with isolated respiratory arrests (61 vs 3%, P<0.001), primary ventricular tachycardia/fibrillation arrests (13 vs 4%, P<0.001), shorter interval times from collapse to medication (1.5 vs 5 min, P=0.013), and longer interval times to intubation (12 vs 8 min, P=0.013). CONCLUSION: Hospital survival after in-hospital cardiopulmonary arrests was poor. Possible strategies to improve survival include shorten time interval to defibrillation, and provision of more monitored beds.

Ultrasound-guided sciatic nerve block: description of a new approach at the subgluteal space.

Sciatic nerve block is frequently used for anaesthesia or analgesia during orthopaedic foot surgery and there are several different approaches to the sciatic nerve. This report describes a new approach to the sciatic nerve using ultrasound. Local anesthetic was injected into the 'subgluteal space' under ultrasound guidance which was effective in producing sciatic nerve block in a small series of five patients. The anatomy, sonographic features, technique of identifying the subgluteal space, and potential advantages of this approach to the sciatic nerve are discussed.

Performance of an oxygen delivery device for weaning potentially infectious critically ill patients.

Oxygen delivery via a heat and moisture exchange filter with an attached T-shaped reservoir satisfies infection control requirements of high efficiency bacterial and viral filtration and low gas flows. In order to assess the performance of such a device in critically ill patients being weaned from mechanical ventilation, we simulated 16 patients using a human patient simulator, measuring fractional inspired oxygen and carbon dioxide concentrations and work of breathing at three oxygen flow rates. Oxygen concentration was dependent on peak inspiratory flow rate, tidal volume and oxygen flow rate. Rebreathing, as indicated by inspired carbon dioxide concentration, was greatest at high respiratory rates and low tidal volumes. Imposed inspiratory work of breathing was relatively high (mean 0.88 J.l(-1)[SD 0.30]). We conclude that this method of oxygen delivery is only suitable for patients in whom rapid extubation is anticipated.

Early extubation after transthoracic oesophagectomy.

OBJECTIVES: To assess patient outcome following transthoracic (Ivor-Lewis) oesophagectomy and the effects of epidural analgesia and early extubation compared with overnight sedation and ventilation. DESIGN: Retrospective study. SETTING: University teaching hospital, Hong Kong. SUBJECTS AND METHODS: A retrospective review of patients undergoing oesophagectomy during two periods, 1990 to 1994 (n=65) and 1995 to 1998 (n=83), was completed. In the latter period, factors associated with early extubation were also evaluated. RESULTS: Between 1990 and 1994, only three (4.6%) of 65 patients were extubated early compared with 34 (41.0%) of 83 patients between 1995 and 1998 (P<0.001). Comparing these two periods, there were no differences in respiratory complications or hospital mortality. In the period 1995 to 1998, more patients who were extubated early had received epidural analgesia (85% versus 41%, P<0.001). There were no differences between the early and late extubation groups in terms of respiratory complications and hospital mortality. Patients extubated early had shorter stays in the intensive care unit (1 versus 2 days, P=0.005). Epidural analgesia was an independent factor associated with early extubation (odds ratio=9.4; 95% confidence interval, 2.8-31.2). CONCLUSION: After transthoracic oesophagectomy, early extubation is safe and can lead to a shorter stay in the intensive care unit. Epidural analgesia appears to facilitate early extubation.

Acute pain services in Hong Kong: facilities, volume, and quality.

Acute pain services in public hospitals in Hong Kong were studied. Audit data on the volume and quality of acute pain services were collected prospectively from 1997 to 1999, and data on related facilities were collected in 2000. About 20% of patients undergoing a major operation received an acute pain service; of these, 78.6% were satisfied with the treatment provided. In 2000, 86% (18/21) of hospitals providing anaesthetic services were running an acute pain service. Staffing was better in hospitals providing a high volume of acute pain services, ranging from a full-time specialist anaesthesiologist assisted by a half-time trainee to a half-time specialist assisted by a full- or half-time trainee. However, only four hospitals were staffed with pain nurses. In total, 57% of patients received intravenous patient-controlled analgesia and 32% epidural analgesia. The mean duration of acute pain service treatment was 3.1 days. Currently anaesthesiologist-based acute pain services take care of a limited number of patients. To expand the coverage, there should be a move towards an anaesthesiologist-led, pain nurse-based, acute pain service. The present shortage of pain nurses should be addressed.

Local anaesthesia outside the operating room.

An increasing number of minor surgical procedures are performed under local anaesthesia in clinical settings outside the operating room, where monitoring and resuscitation equipment--as well as personnel skilled in resuscitation--may not be readily available. Serious adverse effects and even fatalities may result from the use of local anaesthetic agents, arising from a variety of causes such as systemic toxicity, allergy, vasovagal syncope, and reaction to additives present in the local anaesthetic. This article briefly reviews the pharmacology of local anaesthetic agents, and describes various techniques commonly used for local anaesthesia, with special emphasis on safety. Clinical features of toxicity, and its differential diagnosis and management, are also discussed.

Anaesthetic clinical indicators in public hospitals providing anaesthetic care in Hong Kong: prospective study.

OBJECTIVES: To assess the quality of anaesthetic services as defined in the six anaesthetic clinical indicators against preset standards and to identify risk factors for adverse events in the recovery room. DESIGN: Prospective study. SETTING: All public hospitals providing anaesthetic care in Hong Kong. PATIENTS: Eighteen thousand, seven hundred and fifty-nine patients receiving elective or emergency anaesthesia administered by anaesthetists from June 1998 to July 1998. MAIN OUTCOME MEASURES: Patient demographics, American Society of Anesthesiologists status, category and nature of operation, presence of preoperative anaesthetic visit in ward, type of anaesthesia, reasons for a recovery room stay of more than a 2-hour duration, intubation to relieve respiratory distress in the recovery room, presence of hypothermia in the recovery room for operations lasting more than 2 hours, and dental or ocular injuries attributable to anaesthesia. RESULTS: There are two major findings from this study. Firstly, a high incidence of hypothermia in the recovery room was reported. Secondly, a greater risk of prolonged stay in the recovery room was identified for patients older than 65 years, major operations, and anaesthetic techniques using combined general and regional anaesthesia. CONCLUSION: The six anaesthetic clinical indicators reflected the provision of anaesthetic care in public hospitals in Hong Kong. Good compliance to the preset standard of the anaesthetic clinical indicators was achieved during the study period.

Does patient position during liver surgery influence the risk of venous air embolism?

BACKGROUND: It is generally believed that positioning of the patient in a head-down tilt (Trendelenberg position) decreases the likelihood of a venous air embolism during liver resection. METHODS: The physiological effect of variation in horizontal attitude on central and hepatic venous pressure was measured in 10 patients during liver surgery. Hemodynamic indices were recorded with the operating table in the horizontal, 20 degrees head-up and 20 degrees head-down positions. RESULTS: There was no demonstrable pressure gradient between the hepatic and central venous levels in any of the positions. The absolute pressures did, however, vary in a predictable way, being highest in the head-down and lowest during head-up tilt. However, on no occasion was a negative intraluminal pressure recorded. CONCLUSION: The effect on venous pressures caused by the change in patient positioning alone during liver surgery does not affect the risk of venous air embolism.

Thoracic paravertebral block for management of pain associated with multiple fractured ribs in patients with concomitant lumbar spinal trauma.

BACKGROUND AND OBJECTIVES: The need for continual neurological assessment in patients with lumbar spinal injury poses a challenge for effective management of pain associated with multiple fractured ribs. Two cases are presented to illustrate the benefits of using thoracic paravertebral block to control the pain of multiple fractured ribs without compromising the ongoing neurological assessment. CASE REPORT: Thoracic paravertebral block was used in 2 patients with concomitant multiple fractured ribs and lumbar spinal injury. Case 2 also had a head injury and there was moderate coagulopathy. The thoracic paravertebral catheter was placed in the upper thoracic region and radiological imaging was used to delineate spread before the injection of relatively small volumes (10 to 15 mL) of local anesthetic. In case 1, the thoracic paravertebral block produced ipsilateral segmental thoracic anesthesia, providing excellent pain relief for the fractured ribs. It also spared the lumbar and sacral nerve roots, preserving neurological function in the lower extremities and bladder sensation. In case 2, effective analgesia without systemic sedation and opioids resulted in the patient regaining consciousness, which allowed continuous assessment of central and peripheral neurological function. CONCLUSION: Thoracic paravertebral block is an option for managing pain associated with multiple fractured ribs in the presence of concomitant lumbar spinal injury requiring continual neurological assessment.

A prospective survey of patients after cessation of patient-controlled analgesia.

UNLABELLED: Cessation of IV patient-controlled analgesia (PCA) in the postoperative period is often an arbitrary clinical decision. We conducted a prospective survey of patients 24 h after cessation of IV PCA morphine to determine whether they wished to be restarted on PCA, and to evaluate factors affecting this decision. One hundred and fifteen patients were surveyed over a 3-mo period. Thirty-eight patients (33%) wished to restart PCA. The most common reason was the expectation that IV PCA would be more effective. Age, sex, type of surgery, duration of PCA use, side effects, pain scores, and reasons for cessation of PCA did not affect the decision. The reasons given by those who did not wish to restart PCA were minimal pain (51.9%), inconvenient PCA machine (15.6%), ineffective analgesia by IV PCA (11.7%), side effects during PCA (11.7%), and wishing to tolerate pain (7.8%). PCA morphine consumption in the 24-h period before cessation of PCA (mean [SD]) was larger in patients wishing to restart PCA than in those who did not (21.1 [14.8] mg vs 15.1 [15.1] mg; P < 0.05). In conclusion, the clinical decisions to cease IV PCA do not predict patient acceptance of and satisfaction with the decision and with subsequent pain treatment. Morphine consumption may predict a patient's acceptance of ceasing PCA. IMPLICATIONS: We surveyed patients 24 h after cessation of IV patient-controlled analgesia (PCA). This survey revealed that the usual clinical reasons to stop IV PCA might not be the most appropriate. Patients have different reasons why they wish to restart or not restart IV PCA. The cessation of PCA should be individualized.

Randomised double-blind comparison of fentanyl, mivacurium or placebo to facilitate laryngeal mask airway insertion.

In a double-blind randomised study, we compared conditions during insertion of the laryngeal mask airway in 150 patients who received either fentanyl 1 microg.kg-1, mivacurium 0.04 mg.kg-1 or normal saline, before induction of anaesthesia with propofol 2 mg.kg-1. Insertion conditions, including mouth opening, swallowing, gagging or coughing, head or limb movement and ease of insertion, were each graded using a three-point scale. The median (interquartile range) summed insertion scores were more favourable with the use of fentanyl [8.0 (7.0-9.0)] and mivacurium [7.5 (6.8-8.3)] than with normal saline [9.0 (7.8-10.3); p < 0.01]. Fentanyl and mivacurium decreased swallowing and head or limb movement, and mivacurium improved mouth opening. Insertion conditions were similar between fentanyl and mivacurium, while both prolonged apnoea. Fentanyl and mivacurium are equally effective in facilitating insertion of the laryngeal mask airway following anaesthetic induction with propofol.

Anesthesia in a patient with undiagnosed salicylate poisoning presenting as intraabdominal sepsis.

An 81-year-old woman with unintentional salicylate intoxication presented with features of sepsis, abdominal pain, and tenderness. Laparotomy was performed to rule out acute cholecystitis. Anesthesia was complicated by severe hypercarbia despite hyperventilation, and progressive cardiovascular and neurologic deterioration postoperatively. The adverse neurologic, respiratory, and hepatic effects of abdominal surgery and general anesthesia probably potentiated salicylate toxicity and increased patient morbidity. Anesthesiologists should be aware of the protean manifestations of salicylate poisoning and consider it as a cause of "medical abdomen."

The use of mini-dose suxamethonium to facilitate the insertion of a laryngeal mask airway.

The use of mini-dose suxamethonium to facilitate the insertion of a laryngeal mask airway was investigated. Sixty patients were assigned randomly in a double-blind manner to receive 0.9% sodium chloride or suxamethonium 0.1 mg.kg-1 intravenously, following intravenous induction with propofol 2.5 mg.kg-1. The laryngeal mask was inserted after the first attempt in 87% of patients. Mini-dose suxamethonium improved the correct positioning of the laryngeal mask during the first attempt (93 vs. 67%, p < 0.02), decreased the incidence of swallowing (p < 0.001), gagging (p < 0.001) and head or limb movement (p < 0.05). Laryngeal mask insertion was graded as easy in 93% of patients who had mini-dose suxamethonium, compared with 60% in the placebo group (p < 0.01). The duration of apnoea between the two groups was not significantly different (0.54 vs. 0.61 min, p = 0. 46). The total dose of propofol needed to insert the laryngeal mask was lower in the suxamethonium group (2.57 vs. 3.25 mg.kg-1, p < 0. 01) and was associated with less hypotension (p < 0.05). Fasciculation (17%) and mild myalgia (23%) were common despite the small dose of suxamethonium used. In conclusion, mini-dose suxamethonium facilitates laryngeal mask insertion. Myalgia is common and the technique is not recommended for patients who are prone to suxamethonium myalgia.

Low dose mivacurium is less effective than succinylcholine in electroconvulsive therapy.

PURPOSE: To compare the efficacy of low dose (LD) mivacurium (0.08 mg.kg-1) with LD succinylcholine (0.5 mg.kg-1) in modifying seizure activity during electroconvulsive therapy (ECT). Partial muscle relaxation is used in ECT to prevent violent muscle contractions. Current practice is to use LD succinylcholine which has several undesirable side effects. METHOD: Sixteen depressed, but otherwise healthy, patients, aged 27-67 yr were studied. In a randomized, double-blind, cross-over study, either LD mivacurium or LD succinylcholine was given at consecutive ECTs 120 and 30 sec respectively before inducing ECT. Neuromuscular blockade following mivacurium was not reversed. Seizure modification was scored--0 = no seizure activity, 1 = over-modified, 2 = desired level, 3 = under-modified, 4 = unmodified. Duration of seizures, time to first breath and adequate ventilation, ability to protrude tongue and sustain hand grip for five seconds were recorded. Paired t-tests and Wilcoxon matched pairs test were used to compare data. P < 0.05 was considered significant. RESULTS: Seizure modification was better (mean (range)) after succinylcholine 2.06(1-3) than after mivacurium 2.56(2-4) (P < 0.05). Mivacurium was unsatisfactory in eight cases compared with two cases after succinylcholine. The study was terminated early because of unsatisfactory seizure control. Clinical assessments of recovery from both relaxants were similar. CONCLUSION: Low dose mivacurium is unsuitable for use in ECT.

Anatomy of the normal brachial plexus revealed by sonography and the role of sonographic guidance in anesthesia of the brachial plexus.

OBJECTIVE: The purpose of this study is to describe the sonographic appearance of the normal brachial plexus and to evaluate the use of imaging guidance for brachial plexus anesthesia. SUBJECTS AND METHODS: Twenty adults requiring upper limb surgery underwent sonography with a high-frequency transducer to identify the brachial plexus at the interscalene (n = 16) and supraclavicular (n = 4) regions. Catheters for brachial plexus anesthesia were placed using sonographic guidance and evaluated using radiography and CT after injection of contrast material. The success of the neural blockade and surgical anesthesia was assessed. General anesthesia was chosen preoperatively or used for supplementation if blockade was incomplete. RESULTS: The brachial plexus appeared as three discrete rounded hypoechoic nodules between the scalenus anterior and medius muscles on transverse sonography at the lower cervical (C6) region, representing the trunks in sagittal oblique section. A cluster of hypoechoic nodules corresponding to the divisions was seen cephalad to the subclavian artery on sagittal scans of the supraclavicular region. Radiography was used to verify correct catheter placement; the brachial plexus sheath appeared as a tubular area of contrast material in the neck that was well circumscribed in patients who had supraclavicular injections of contrast material and corrugated in those who had interscalene injections. CT scans showed contrast material spread around the scalenus muscles in most patients with interscalene injections. Successful neural blockade at 20 min and postoperative analgesia were achieved in all patients. Surgical anesthesia was achieved in nine of 15 patients. Five patients chose general anesthesia before surgery and therefore did not have surgical anesthesia assessed. CONCLUSION: High-resolution sonography can show normal brachial plexus anatomy and facilitate catheter-based brachial plexus anesthesia without complications.

Single-dose tropisetron for preventing postoperative nausea and vomiting after breast surgery.

UNLABELLED: In this randomized, double-blind, placebo-controlled study, we compared the efficacy of tropisetron 5 mg with tropisetron 2 mg for the prevention of postoperative nausea and vomiting (PONV) after breast surgery. One hundred forty-eight female patients were randomized to receive either tropisetron 5 mg (n = 49), tropisetron 2 mg (n = 49), or saline (n = 50) before the induction of anesthesia with thiopental and morphine. Anesthesia was maintained with nitrous oxide and isoflurane. Postoperative analgesia was provided by patient-controlled analgesia with i.v. morphine. The incidence of PONV, the pain score, and the analgesic requirement were recorded for 48 h. There was no difference among groups in patient characteristics, risk factors for PONV, morphine consumption, or side effects. During the first 6 h postoperatively, the incidence of PONV after tropisetron 2 mg and 5 mg were similar and were superior to placebo (P < 0.001). After 6 h, the incidence of PONV increased significantly in patients who had received tropisetron 2 mg (P = 0.01) and was greater than that in patients who had received tropisetron 5 mg (P = 0.001). We conclude that single-dose tropisetron 5 mg is more effective than tropisetron 2 mg in the prevention of PONV after breast surgery. IMPLICATIONS: Breast surgery is associated with a high incidence of postoperative nausea and vomiting. A single dose of i.v. tropisetron 5 mg is well tolerated and decreases the number of vomiting and nausea episodes after surgery.