| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
*Pharmacy Department and Department of Anesthesia, Intensive Care, and Pain Medicine, Repatriation General Hospital, Daw Park, Australia; and Royal Danish School of Pharmacy, Copenhagen, Denmark
Address correspondence and reprint requests to Greg Roberts, Pharmacy Department, Repatriation General Hospital, Daws Rd., Daw Park SA 5041, Australia. Address e-mail to greg.roberts{at}rgh.sa.gov.au.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Despite the large amount of research performed in this area, it is only recently that expert consensus regarding the selection of patients for prophylaxis of POV has been published (5). Prevention of PON and POV is important, not only in relation to the quality of care afforded to each patient, but also in monetary terms.
A number of studies have identified risk factors for POV in various surgical populations (6–9). These studies have often used regression analysis to construct risk scoring equations that incorporate various risk factors, although not all groups have identified the same risk factors as being significant contributors to POV. This may reflect the different populations that have been studied. The commonly identified risk factors that have consistently contributed to POV are: female gender, nonsmoking status, history of POV or motion sickness, extended duration of anesthesia, postoperative opioid use, and age.
As the number of risk factors increases, so does the chance of POV. It is possible that in some instances the type of anesthesia and surgical procedure may also contribute.
The above-mentioned studies have not specifically examined the dose-response relationship between opioids and vomiting. Most clinicians would intuitively assume that a patient receiving a large dose of postoperative opioid is more likely to vomit than a similar patient who receives less.
We examined the effect of these known risk factors on our patient group, with a focus on the relationship between vomiting and opioid use in the 48 h postoperatively. Although well recognized, the relationship between opioids and POV, and the degree of influence it bears, has not been well explored. In particular, many of the equations developed through logistic regression predicting risk of POV have only included opioid use as a dichotomous variable, with no acknowledgment of any dose-response relationship.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
2 days who did not receive perioperative antiemetic prophylaxis were eligible. The approach to analgesia for any given patient was at the discretion of the anesthesiologist. Those patients not using epidural analgesia or PCA were given pain relief with a combination of IV and oral medication, on an "as required" basis. Patients already receiving drugs with antiemetic properties, including corticosteroids, were excluded.
An episode of POV was defined as vomiting or retching over any 2-min period. The severity or duration of nausea was not recorded, only if it was present or not, as determined by the patient. Patients who vomited were automatically included as having experienced nausea at that point. Patients routinely received postoperative rescue antiemetics if they vomited, or experienced 10 min of debilitating nausea. In the first instance, they received 10 mg of IV metoclopramide, followed 10 min later by 4 mg of IV ondansetron if the nausea and vomiting were still not controlled.
Nausea and vomiting episodes were recorded 0.5, 1, 2, 4, 8, 12, 24, and 48 h postoperatively. Opioid doses, both intra- and postoperative, were recorded for the 0- to 24-h and 24- to 48-h periods postoperatively. All opioid doses, regardless of route of administration or type of opioid, were converted to the equianalgesic dose of IV morphine for comparative purposes, using the values in Table 1 (10,11). These values were predetermined on current available literature and the clinical expertise of the participating anesthesiologists. Fentanyl was used for all epidurals and was considered equipotent via epidural or IV route. One milligram of IV morphine was considered to be equianalgesic with 10 µg of spinal morphine.
|
Parametric data were compared using a Student’s t-test. Kaplan-Meier plots were used to examine the incidence of POV over time. Cox regression analysis was used to examine variables influencing POV. The significance level was set at 5%.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
|
|
|
Cox regression analysis included gender, history of POV or motion sickness, smoking, duration of anesthesia, age, and opioid dose, and revealed only opioid use (P = 0.025) and female gender (P = 0.038) as factors influencing POV. The influence of history of POV or motion sickness, smoking, duration of anesthesia, or age were not significant in this relatively small group.
Use of PCA or epidural analgesia were markers for large-dose opioid use in the first 24 h (91.5 and 83.2 mg of morphine or equivalent for PCA and epidural analgesia, respectively, versus 17.5 mg for non-users, P < 0.001). This was associated with more frequent POV and PON (Table 4). The majority of POV had occurred by 12 h in the PCA group whereas the majority of POV for epidural patients was seen in the 12- to 24-h period (Fig. 1). Patients not using PCA or epidural analgesia experienced less POV and PON (P < 0.001 for both). Seven patients in this group required no opioid analgesia and did not experience PON or POV. Those patients who experienced POV and PON in the 24- to 48-h period postoperatively had significantly larger opioid use during this period than those who did not (P < 0.01 for both).
|
Patients were divided into quartiles according to opioid dose to further examine the relationship between opioid dose and POV in the first 24 h postoperatively. The time course of POV for the 4 morphine dose quartiles is shown in Figure 2 (P = 0.05). There was a strong logarithmic dose-response relationship with POV (r2 = 0.98, P < 0.01), as well as PON (r2 = 0.98, P = 0.01, Fig. 3). The relationship between POV and morphine dose in the 0- to 24-h postoperative period was described by:
|
|
|
|
When patients receiving opioids via the spinal or epidural route were removed from analysis, this relationship remained largely intact, although the dose-response relationship with POV in this subgroup was better suited to a linear relationship (r2 = 0.99, P < 0.01 for linear, r2 = 0.82, P = 0.09 for logarithmic, n = 145). PON remained best correlated to a logarithmic relationship (r2 = 0.99, P < 0.01 for logarithmic versus r2 = 0.88, P = 0.07 for linear).
For given types of surgery there was marked variability in opioid use. In the first 24 h postoperatively, knee arthroplasties had a median morphine dose of 70 mg (range 7–134, n = 33) whereas hip arthroplasties had a median morphine dose of 47 mg (range 4–296, n = 23).
The highest at-risk groups for POV were women receiving PCA or epidural analgesia with or without a history of POV (70% and 54%, respectively, PON rates 100% for both groups) and then men receiving PCA or epidural analgesia with a previous history of POV (35%, PON rate 71%).
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The relationship between POV and morphine dose in the 0- to 24-hour postoperative period was described by:
|
|
This means that for each halving of the opioid dose in this 24-hour period, there was an absolute reduction of 6.0% in POV, assuming all other factors remained constant. Regardless of whether the dose is halved from 200 to 100 mg, 100 to 50 mg, or 50 to 25 mg, the absolute reduction in POV was constant at 6.0%. The relative reduction in these groups, however, was 14.9%, 17.5%, and 21.2%, respectively, giving greater relative returns for dose reductions in those patients in the smaller dosing range. If a patient had not already experienced POV or PON in the first 24 hours postoperatively, there was only a very small likelihood they would experience it after 24 hours.
PCA or epidural analgesia use was targeted toward patients undergoing surgical procedures with known large postoperative analgesic requirements, such as hip and knee arthroplasties. Not surprisingly, these types of analgesia were markers for large-dose opioid use and hence also predictive for POV. There was much variation in opioid use, however, even within various surgical subgroups receiving PCA or epidural analgesia. Accurate prediction of postoperative opioid use, perhaps as part of a risk scoring assessment to determine prophylactic antiemetic use, would be very difficult.
The postoperative opioid dose seems to have a large part in determining the likelihood of POV or PON. The strength of the dose-response relationship with POV has been largely unrecognized, and this may have led to the inability of risk scoring equations to accurately predict POV (12). Given the apparent strength of the opioid-POV relationship, the inclusion of opioid use as a dichotomous variable rather than a continuous variable may undermine the ability of these risk scoring equations to predict POV. Unfortunately, the ability to predict a patient’s opioid dose is poor, even for a given type of surgery, as seen with the large dose range in both hip and knee arthroplasties in this study. This source of variability may have to be something clinicians have to accept when attempting to assess a patient’s POV risk. Exploration of opioid-sparing strategies for the postoperative period seems appropriate, and patients likely to have large postoperative opioid requirements should be targeted for prophylactic antiemetic strategies.
It is important to note that although these data indicate a strong relationship between opioid use and POV, they do not represent definitive proof of this relationship, the population studied in this series being relatively small. The possibility exists that postoperative opioid use may be a marker for a further variable that is the true cause of increased POV, such as duration of surgery, or a specific type of surgery. Consistent with the belief that postoperative opioid dose is a primary driver of POV, however, Buvanendran et al. (13) found a reduction in POV from 26% to 6% in 2 identical groups undergoing knee arthroplasty when postoperative opioid was reduced by the perioperative administration of rofecoxib. It remains to be seen if other associated variables, such as the type of surgery, may be a marker for opioid use and hence POV. This needs further investigation in populations large enough to determine these possibilities.
The opioid dose was the most dominant influence in determining POV in this patient group, followed by female gender. The small population studied was likely to have lacked sufficient power to determine other known associated risk factors for POV. The addition of various known risk factors resulted in increased rates of POV, consistent with the findings of others (14). The two highest at-risk groups for POV were women receiving PCA or epidural analgesia with or without a history of POV (70% and 54%, respectively). Men receiving PCA or epidural analgesia with a history of POV also experienced a frequent rate of POV (35%).
There were two practical problems we faced in obtaining a history of POV. First, some patients had not previously experienced anesthesia, and, second, the occasional elderly patient described POV after the use of ether as an anesthetic—hardly a fair comparison with the anesthetics used today. This decreased the value of a history of POV as a reliable predictor. We also experienced difficulty in ascertaining a clear history of motion sickness in our elderly patients.
Propofol has been shown to possess dose-related antiemetic activity (15,16). This patient group received a small dose of propofol as part of the induction process, with just five patients receiving continuing propofol as a continuing part of the anesthetic procedure. The effect of propofol in this patient group remains unknown, because the numbers were too small for meaningful analysis. Given the small doses used, however, the effect on POV was unlikely to be significant.
There is inevitably some degree of subjectivity in determining equianalgesic potency within the opioid group and across various routes of administration. The equivalency tables used for this study were determined before any patient data were collected and are based on currently available literature. The duration and potency of spinal or epidural opioids is greatly enhanced compared with the IV route, and exact comparisons of analgesic potency are difficult. When patients receiving spinal or epidural analgesia were excluded, however, the dose-response relationship remained intact. Although conversions to equivalent doses of IV morphine were made on the basis of analgesic potency, this may not be reflective of their differing emetogenic potencies.
The incidence of POV and PON found in this study is likely to be an underestimate of the true incidence for two reasons. First, patients who received antiemetic prophylaxis were excluded from study. For ethical reasons, we allowed the normal practice of empirical prophylaxis, as judged by the anesthesiologist, to continue, and hence were unable to study an entire surgical population uninfluenced by perioperative antiemetic prophylaxis. These excluded patients (n = 72) contained a number of higher-risk patients who would likely have increased the rate of POV and PON had they been eligible for inclusion. Compared with our study group, this group had a significantly larger proportion of orthopedic patients, women, patients with a history of POV, and fewer smokers. Exclusion of these patients would make identification of the higher-risk groups less likely. Second, patients who experienced PON, but not yet POV, were subsequently given rescue antiemetic therapy, which may have prevented them from progressing to POV. This may have generated a falsely small incidence of POV, without affecting the incidence of PON.
It is possible that accounting for the dose-response relationship between postoperative opioids and POV may lead to further refinement in risk scoring algorithms. The ability to predict a patient’s postoperative opioid requirement is likely to be poor, however, and this may continue to hamper accurate prediction of which patients should be targeted for preoperative antiemetics.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Coburn, O. Kunitz, C. C. Apfel, M. Hein, M. Fries, and R. Rossaint Incidence of postoperative nausea and emetic episodes after xenon anaesthesia compared with propofol-based anaesthesia Br. J. Anaesth., June 1, 2008; 100(6): 787 - 791. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. J. Gan, T. A. Meyer, C. C. Apfel, F. Chung, P. J. Davis, A. S. Habib, V. D. Hooper, A. L. Kovac, P. Kranke, P. Myles, et al. Society for Ambulatory Anesthesia Guidelines for the Management of Postoperative Nausea and Vomiting Anesth. Analg., December 1, 2007; 105(6): 1615 - 1628. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. K. Kim, S. B. Nam, M. J. Cho, and Y.-S. Shin Epidural naloxone reduces postoperative nausea and vomiting in patients receiving epidural sufentanil for postoperative analgesia Br. J. Anaesth., August 1, 2007; 99(2): 270 - 275. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. J. Gan Risk factors for postoperative nausea and vomiting. Anesth. Analg., June 1, 2006; 102(6): 1884 - 1898. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Redmond and P. Glass Opiate-Induced Nausea and Vomiting: What Is the Challenge? Anesth. Analg., November 1, 2005; 101(5): 1341 - 1342. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
, n = 64), epidural analgesia (
, n = 26), neither (•, n = 101). P < 0.0001.
