Developing an Evidence-Based Guideline: Prophylaxis of Post-Operative Nausea and Vomiting

Tim Crichton, Michael Edmonds

Supervisor: Dr. Malcolm Pradhan

Faculty of Medicine, University of Adelaide

Abstract

Post-operative nausea and vomiting (PONV) represents a significant problem in anaesthetic procedures, particularly in the day-surgery patients. It is the most common adverse event in the specialty of anaesthesia and is the most commonly stated fear of patients prior to elective surgery. Despite this, there is no current general consensus regarding the prophylactic management of PONV. This study aims to examine the patient, operation and anaesthetic factors associated with development of PONV, and use the best available evidence to construct an influence diagram and decision model which will predict the risk of PONV, and determine the optimal management in a particular patient. This may then be further extended to develop guidelines for future patient management. This study involved extensive literature review, construction of a qualitative and quantitative model representing the associations of these predictive factors with PONV, and statistical analysis of the predictive accuracy of the model. The final predictive model predicts a general population rate of PONV of 33.4% with no anti-emetics administered, and generated decision thresholds for the most cost-effective choice of anti-emetic. Sensitivity analysis shows that the patient risks are the most significant factors in the development of PONV, and that anaesthetic choice has relatively little effect. Statistical analysis using data from a Royal Adelaide Hospital Day Surgery Unit database indicates that the model produced has a higher error rate than statistical models produced from the data, but has a higher sensitivity to predicting the occurrence of PONV, and will result in less readmissions due to PONV.

Introduction

Post-operative nausea and vomiting (PONV) was first recorded in 1848, just 18 months after the introduction of anaesthesia into Britain 1 . There have been numerous studies into the problem over the years, many suggested aetiologies and varying different remedies, including meat-free diets, administration of glucose and insulin, iodine, black coffee, iced water, mustard leaves, olive oil, wine, opium, ginger and acupuncture 1-3 . More recently the focus has shifted to the contributing factors associated with PONV, and the prophylactic use of pharmacological agents. There have been significant advances in the anaesthetic techniques and agents used, reducing the incidence of PONV from above 80% under ether or cyclopropane, down to 30% with newer less emetic agents. The importance of PONV in prolonging patient morbidity following surgery has been further highlighted by the increase in day surgery over the last two decades where such an event can cause a delay in discharge, unanticipated readmission and disruption of the smooth management of ambulatory surgery patients.

In 1995 the Quality in Australian Health Care Study identified post-operative nausea and vomiting as the most common adverse event in the speciality of anaesthesia 4 . It is estimated to cost over $1.2m per year in unanticipated hospital admissions following day surgery alone, which represents but a small percentage of the incidence of PONV. A large proportion of the cases of PONV do not result in delayed discharge or require readmission, but costs are felt in other areas, such as visits to general practitioners, slower recovery and loss of productivity by the patient. In our study we have decided to define PONV as any episode of nausea, retching or vomiting in the 24 hours immediately following an operation, and the focus of this current study is the area of day surgery.

The cost of PONV can not be quantified simply in monetary terms, as the patient's wellbeing and comfort must also be considered. Kovac (1996 5 ) stated that 38% of patients experiencing PONV perceived it to be more debilitating than the effects of the surgery itself, and a recent survey of patients has shown that it is the single most commonly stated fear preceding elective surgery. Nausea alone can also limit patient activity, and severe nausea can be as, or more, distressing than vomiting or retching. There are also medical and surgical consequences of PONV. Vomiting may lead to dehydration, interference with nutrition and oral drug therapy or dehiscence of abdominal wounds 6 . It may also result in tension on suture lines, increased bleeding under skin flaps, venous hypertension and increased risk of pulmonary aspiration of vomitus 7, 8 . PONV has been found to be the cause of 17% of unanticipated readmissions following day surgery 9 , which reflects a large cost in itself as well as much patient discomfort and continuing morbidity.

Despite the large amount of research performed in this area there is currently no general consensus or guideline about the management and prophylaxis of PONV. Prevention of nausea and vomiting is important, not only in monetary value, but also in relation to the quality of care afforded to each patient. Routine anti-emetic prophylaxis, however, has not been indicated since only 30% of patients experience post-operative emetic sequelae 8, 10 , and of these, many cases are only transient nausea, or only one or two bouts of vomiting or retching. Many of the anti-emetic agents also have a spectrum of associated adverse effects, potentially causing further discomfort for the patient 8 . If anti-emetic prophylaxis was to be used routinely, then in patients with a low risk of developing PONV, the risk of the adverse effects has been suggested to outweigh the potential benefits, and the cost of administering the agents would amount to more than the cost of readmissions if no prophylaxis was given. On the other hand, use of anti-emetic agents in patients at high risk of developing PONV could greatly reduce the incidence of PONV, reducing cost, patient discomfort and medical complications, while increasing the quality of care. It is necessary, therefore, to identify the subset of the population that will be at a high risk of developing PONV, and would benefit from prophylactic anti-emetics.

Aim

The aim of our study is to construct an evidence based guideline in the area of PONV. To do this we had to determine which factors contribute to the aetiology of PONV, and how these may be used to predict the risk of PONV for an individual patient. We would then construct an evidence-based decision model representing the interaction of the different factors to produce post-operative nausea and vomiting. This model may then be used as a decision support tool or evidence based guideline. The term 'evidence-based' may be defined as "the conscientious, explicit and judicious use of best current evidence in making decisions" (Sackett et al, 1997 11 ). Evidence used would primarily include studies involving randomised trials, combined with expert opinion from the Royal Adelaide Hospital (RAH) Department of Anaesthesia and Intensive Care. The decisions that we planned to cover were: (a) should a prophylactic anti-emetic agent be given, in relation to an individual patient, and (b) which anti-emetic agent should be given, if any. We hypothesised that this decision support model could then be used to maximise the cost-effective use of prophylactic agents, as well as the quality of care provided to the patient.

Method

Software for influence diagrams and decision models

To construct our models, suitable tools for the construction and implementation of influence diagrams and decision models had to be obtained. The software packages suggested by our supervisor for such purposes included Hugin v5.2 Lite , GeNIe (a Graphical Network Interface) v1.0 , and Netica v1.12 . Hugin and Netica were readily obtainable as limited shareware versions from their relative web-sites at http://www.hugin.dk and http://www.norsys.com, and these versions were found to be suitable for our intended applications. Familiarisation with the process of influence diagram and decision model construction, as well as their use in medical decision making was necessary, and this was available through tutorials in the individual decision model programs and various medical decision making reference texts 11-15 .

Literary review

An extensive literary review was necessary to locate the best current knowledge regarding the aetiology, physiology, risk factors and treatments of PONV. Our collection of literature was compiled from numerous sources, primarily the results of searches conducted in Medline, the Cochrane library and the Internet, as well as suggested reference texts supplied by the Department of Anaesthesia and Intensive Care, Royal Adelaide Hospital (RAH). These consisted of a total of over 60 references, dating from 1998 back to 1957. These provided a comprehensive list of all factors thought to be associated with the development of PONV. Using this information an initial qualitative model was constructed.

Qualitative model

To be consistent with previous studies, and for simplicity, all of the suggested factors were categorised as contributing to patient, anaesthetic or operative risks. This association is represented in the influence diagram by an arc between the nodes of these factors. These three main categories combine to give an overall risk of PONV.

The risk of developing PONV influences the decision of whether an anti-emetic prophylaxis should be given, and which one would be the optimal choice. The anti-emetic agents included in this study were Metoclopramide, Droperidol and Tropisetron, as these agents are thought to represent the major classes, as well as what is currently in use in the setting of the RAH. The potential outcomes were also included in the initial model, represented by the value nodes of cost and patient utility. Due to time constraints, in the final model we did not include patient utility, but focused on the costs. The cost includes that of the prophylactic agent, including no prophylaxis, and the cost of readmission given that the response rate of the various prophylactic agents is considerably less than 100%.

In a similar process a mechanistic model was constructed based on current knowledge regarding the physiological and causal processes involved in the development of PONV. The complexity of the complete mechanistic model required that we also construct a simplified version for ease of use.

Quantitative model

Further reading and expert opinion from Professor W. Runciman and Dr Guy Ludbrook (RAH Department of Anaesthetics and Intensive Care) led to refinement of the model and then those factors remaining were quantified. Quantification involved a further literature review and consultation with Dr Ludbrook, locating original evidence from studies and meta-analyses. The evidence from these articles were then assessed for relevance and validity. This included the study design, the patient demographics of the study, the sample size, and various confounding factors such as operation types or anaesthetic techniques. Through this process the studies that gave the best evidence were sorted from studies that contained too many factors that could have biased the results. Comparison between studies, and a discussion of the varying results with Dr Ludbrook determined which evidence from the studies was the best to use in our setting.

The evidence from the studies came in the form of percentages, odds ratios and number-needed-to-treat, and these were standardised and used in our decision model. A number of factors were not substantiated by study evidence, or had evidence that showed they were not an important factor, and these factors were then removed from the model. There was insufficient data at the physiological level of PONV to quantify the mechanistic model.

The influence of each individual factor was weighted on the impact it would have on PONV. This was achieved by weighting the influence of each of the major categories of patient, operative and anaesthetic risk on the risk of PONV, and weighting the influence of each of the factors on the categorical risk they fell into. A 'noisy or' function was used to determine the influences that each of the patient factors had on overall patient risk.

The cost and efficacy of the different anti-emetic agents were determined and entered into the model, along with the costs arising from each option and potential outcome. The resulting model was then capable of predicting the risk of PONV as well as calculating the costs of each of the treatment options for that risk.

Statistical analyses

A one-way sensitivity analysis was performed on the completed model to determine the relative sensitivity of the model to each of the individual factors, and a further sensitivity analysis was performed on the three main risk categories.

A database containing information about the patients, procedures, anaesthetics and subsequent recovery and follow-up in the RAH day surgery unit (DSU) between 1989-93 was also obtained from the Department of Anaesthesia and Intensive Care. This database consisted of 6000 day surgery patients, of which 5000 were randomly chosen as a 'training set' and the other 1000 as a 'test set'. Models were built using the training set including factors from the database of age, sex, operative risk, opioid administration, use of propofol or volatile anaesthetic agents, use of N2O and use of Neostigmine. This information was used to build a logistic regression model and a classification tree (see appendix 1 for models). This information was also used to 'learn' some prior probabilities into our model, from which the decision was removed to produce a belief network that would predict the incidence of PONV. This model and the original model without any learnt parameters were then compared with the statistical models for predictive accuracy and sensitivity to detecting PONV using the test set. These statistical analyses using the database were performed by Dr Pradhan using a statistics package called S-Plus.

Results

A number of different models representing the development of PONV resulted from our research. Figure 1 indicates all of those factors suggested within the literature to be associated with PONV, regardless of the validity of such relationships. Figure 2 represents the final predictive model and includes only those factors found to be of qualitative and quantitative significance to the problem. The postoperative factors were also excluded from this model because it was assumed that these would remain unknown at the time the decision regarding prophylaxis was made. The reasons for exclusion of each of the factors is indicated in Table 1. The model describes the associations between the various factors and the outcome. The factors are represented by chance nodes, which includes information about their prior probabilities or prevalence. These then influence the conditional probabilities of the main categorical nodes. The resulting risk of PONV influences the decision to be made and the predicted cost

Figure 3 shows the complete mechanistic model, indicating the hypothesised causal relationships between each of the factors and the final outcome of PONV. Figure 4 represents the simplified mechanistic model. Nearly all factors associated with PONV were found to act through the emetic centre in the brainstem, with pathways through the chemoreceptor trigger zone (CTZ), visceral afferent fibres through the sympathetic and vagal nerves, higher cortical pathways or by altering function of the vestibular apparatus. This model could not be quantified due to the lack of evidence examining the aetiology of PONV at a physiological level, and as such this model remained qualitative.

(See Appendix 3 for full size diagrams)

(See appendix 3 for full size diagrams)

Table 1. Reasons for excluding factors

Table 2 shows the prior probabilities used for each of the factors included in our model, and figure 5 shows the results of the sensitivity analysis that was performed, indicating the importance of knowing about each factor with regards to the predicted cost of treatment per patient. Evidence used to determine these figures and results came from a variety of different studies and articles 2, 9, 16-36 (see appendix 2 for further details).

The anti-emetic agents used in this model were Metoclopramide, Droperidol and a 5HT3-antagonist, Tropisetron. The costs of each of the agents were obtained from the RAH Department of Anaesthesia and Intensive Care, and these were used in a function estimating the value of the cost of treatment node. These costs of prophylaxis are shown in table 3, along with the efficacy rate in preventing PONV of each of the agents as well as the references these rates were based upon.

Our model indicated a population risk of 33.4% when no anti-emetic prophylaxis is given. The maximum risk estimated by the model is 85.4% for a patient with all the risk factors (obese, female under 16 years who does not smoke, has a past history of PONV, and is undergoing a high-risk operation using volatile anaesthetic agents, combined with nitrous oxide, opioids and neostigmine to reverse the muscle blockade). The lowest possible risk is estimated at 1.4% for a

Table 2. Prior probabilities of factors

patient with none of the risk factors (non-obese male, over 80 years old, who smokes and has no past history of PONV, undergoing a low risk operation using propofol, no nitrous oxide, neostigmine or opioids). In the calculation of cost of readmission, a rate of 20% of patients with PONV being readmitted was used, at an average cost of $500 for the period of the readmission. A readmission rate of 0.1% in patients who were predicted to be at low risk for PONV was also included.

Analysis of the data from the decision model indicates there are decision thresholds at which the most cost-effective choice of prophylaxis changes. As the risk of PONV increases, the choice of agent changes from the low-efficacy, cheaper Metoclopramide to the mid-range efficacy and cost Droperidol, and for those at high risk the most effective, but also more expensive, Tropisetron. These decision thresholds are at a 17% and 36% risk of PONV(fig 6). These results also show that the option of giving no treatment is dominated throughout the range of PONV risk by the option to administer some form of prophylaxis. The cost per patient will therefore be less if anti-emetic prophylaxis is given when compared to giving no prophylaxis at all. This cost is a function of the cost and efficacy of the anti-emetics and the probability and cost of readmission.

An important result of this model is that it shows that, compared to patient and operative risks, the anaesthetic factors play a relatively minor role in contributing to PONV, and therefore to the

cost that will be incurred (fig 7). Most of the risk derives from the patient characteristics, with the operation type playing a smaller role.

The logistic regression model produced from the database had 4827 degrees of freedom. The partial t-values suggest that all factors apart from Propofol are important in this model (see appendix). The predictive accuracy of the model on the test set produced a 9% error rate (table 4), all of which were misclassifications of the positive cases.

The classification tree produced much the same results (table 5). Using a pruned version of the tree to reduce complexity, the error rate was 9%. Again, all were misclassifications of the positive cases. In both these models positive cases were not predicted to minimise the error rate.

The belief network, however, attempted to diagnose positive cases, and therefore had a higher error rate, but this higher sensitivity to detecting PONV would result in fewer admissions. The results for the original belief net (table 6) showed an error rate of 32%. After 'learning', some parameters changed slightly, but due to the amount of missing data in the database, the parameters in the intermediate nodes (patient, anaesthetic and operative risks) did not change significantly. The learned belief network produced similar results to the initial model (table 7), with an error rate of 31%.

Discussion

Our model produced a population risk of 33.4% when no anti-emetic is given, which is concurrent with most recent research, as well as with expert opinion. As well as this, the results of the sensitivity analysis showed the variation due to each factor was in concordance with the literature, validating the figures that we used. The sensitivity analyses showed that the most important factors in predicting PONV were a past history of PONV and gender, and more interestingly that, of the main risk categories, most of the risk derives from the patient, followed by the operation, and the risks due to the anaesthetics contributes the least. This indicates that the decisions made by the anaesthetist will have little impact on the risk of the patient developing PONV, and it is more the fixed patient characteristics that will determine the risk of PONV.

The choice of anti-emetic treatment for the varying ranges of PONV risk is also an important outcome of this model. Firstly, the model shows that the choice of giving no prophylaxis is dominated throughout by the option of administration of prophylaxis, mainly due to the low cost of Metoclopramide. This low cost means that even at the lowest risk of PONV, the benefits of giving Metoclopramide outweigh the minimal cost of the drug. The varying costs and efficacies of the different anti-emetic agents has resulted in a stratification of suggested use, with the most effective, yet most expensive, agent, Tropisetron, being reserved for those patients at the higher risk of developing PONV. It must be remembered that this model does not take into account the side-effect profile of each drug, and the additional costs and patient discomfort that this could result in. This may alter patient preferences and highlights that monetary cost is not the only consideration to be made in prophylaxis administration.

The statistical models proved to have a much lower error rate in predictive accuracy than our model, but also had a much lower sensitivity to predicting PONV. Using a cost of prophylaxis between the costs of Droperidol and Tropisetron, and assuming a rate of readmission in patients with PONV of 20% at a cost of $500, usage of the statistical models as a policy would result in a 70% higher cost than the decision model. This is because these models did not predict the occurrence of PONV in order to minimise the error rate of the model, and hence no prophylaxis would ever be given, resulting in a higher rate of readmission. The analysis also showed very little difference between the test results of the belief networks, with and without learning, and this is mainly due to the amount of missing data in the database. To improve this learning process a different method that handles missing data better, such as an expectation maximisation algorithm, should be used. This comparison of predictive accuracy with the statistical models shows that our model is less accurate in predicting PONV in this test set, but also shows that it is much more sensitive to detecting PONV, and will result in fewer readmissions, as well as increased patient comfort as more patients will receive anti-emetics and will experience less PONV.

The model we constructed attempted to represent the clinical situation as closely as possible. To achieve this a number of assumptions were made so that our model remained usable, avoided unnecessary complexity, yet was still true to the clinical setting. One of these was the assumption that the different factors were not inter-related or acting in synergy to influence PONV risk, despite it being suggested for example, that age and operation types are related. Other assumptions that we made included the division of some of the factors into discrete state groups, when in reality there exists a spectrum of variable states (ie past history of PONV, smoking history, and operation risk). Another assumption was that Ondansetron and Tropisetron, the 5HT3-antagonists, have the same level of efficacy but differ only in cost. This was made because almost all of the research has been done on Ondansetron, which is the prototype of its class, and Tropisetron is the agent that is used in the RAH. Many assumptions were also made in the quantitative stage of the model construction, such as 20% of patients with PONV are readmitted and the cost of an average readmission is $500. Such assumptions were made throughout the construction of the whole model using best available evidence and expert opinion.

In obtaining the best available evidence, there were many instances where data to support our model was lacking. Our general impression of the research conducted in the area of PONV is that much of the data that is available is relatively out of date. This is mostly due to the changes that have occurred with relation to anaesthetics and anti-emetics used, as well as advances in procedural techniques in operations. As such, many of the studies were excluded from being incorporated into our model. There also seemed to be a lack of data for some of the commonly associated factors of PONV, such as gastroparesis and patient anxiety. For these factors, review of the original referenced studies revealed that many authors had simply included such factors through a process of 'ritual inclusion' by reiterating what others had said previously. As such, we refrained from using any of these factors in our model, but they may well pose a question for future research.

This research study, has shown that decision models have great potential in the field of medicine. A major advantage of the model type used is the ability to easily make adjustments and additions to the model as new information becomes available. This would include new evidence about influencing factors or new factors, new antiemetics, drug combinations changes in drug pricing, or other advances in patient management. The inclusion of patient utility into the model, in terms of patient discomfort resulting from PONV itself, or from the side-effects of antiemetics, may further increase the value of the decision model and add reality to its representation of a true clinical situation. The model may also be customised for other hospitals by adjusting the individual patient, operation and anaesthetic factors prior probability values. Another possibility is to integrate the model into a simple computer program or into a web-site, making it available to health care professionals either as a tool or to simply introduce the decision model theory and its utility in health care.

If more information becomes available following research into the physiological aetiology and mechanisms of nausea and vomiting, it may be possible to integrate the mechanistic model with the decision model. This may be of great value since the actions of the specific antiemetics may be used to target the specific mechanism underlying each case of PONV, and therefore increase their overall efficacy. For example, Metoclopramide in general is the least effective antiemetic, but in the case of a patient vomiting due to the presence of blood in the stomach, it may have a greater efficacy than other anti-emetics due to its action in promoting gastric emptying, clearing the emetic stimulus from the stomach. This is an area where much research is yet to be conducted, but is potentially very beneficial.

In the near future it is hoped that this model will undergo prospective studying to test its predictive accuracy in reality, as well as the cost-saving potential by judicious use of prophylactic anti-emetics. It is hoped that this will lead to the introduction of a guideline for prophylaxis of PONV based on this evidence-based model. As well as this there is a lot of potential for dealing with many other decisions using this type of model. An extension of this model may include the decision of which anaesthetic agent should be used to minimize the risk of PONV. This type of model can also extend out to many other decisions in the health care setting.

Our research is but a small step in the direction of dealing with the problem of PONV, but some of the results are a great leap forward in the future management of the problem. Our model includes all factors that have been found to be of significance in the development of PONV, and their relative importance in predicting its incidence. The recommendations that arise from our model relate mainly to the prophylactic management of PONV. The option of no prophylaxis should not be considered, as this will lead to a higher rate of readmission and cost, as well as greater patient discomfort. Patients at a risk of less than 17% of developing PONV using our model should receive Metoclopramide, those with a risk between 17-36% should receive Droperidol, and those at more than 36% should receive Tropisetron. The accuracy and clinical relevance of our model can be improved by the inclusion of patient utility, other significant factors and new evidence as it becomes available. There are also many areas in which this study can be extended to add reality and increase its usefulness. The advantage of using a decision model to map this problem is that it is easily updated, and can always be kept up to date with the latest developments in the area. We hope that this decision model may the starting step in the process of improving management of PONV.

Acknowledgments

Many thanks to our supervisor, Dr M. Pradhan, and to Professor W. Runciman and Dr G. Ludbrook from the RAH Department of Anaesthesia and Intensive Care. Many thanks also to Dr G. Osborne for providing us with the database, and to Dr A. Bowie for providing us with some starting references.

References

1. Andrews P. Physiology of Nausea and Vomiting. British Journal of Anaesthesia. 1992;69:2s-19s.

2. Burtles R, Peckett B. Postoperative Vomiting: some factors afecting its incidence. British Journal of Anaesthesia. 1957;29:114-123.

3. Rowbotham D. Current Management of Postoperative Nausea and Vomiting. British Journal of Anaesthesia. 1992;69:46s-59s.

4. Runciman W. Anaesthetists and the "Quality in Australian Health Care Study". Australian society of anaesthetists Newsletter. 1997:15-17.

5. Kovac A, Pearman M, Khalil S, et al. Odansetron prevents postoperative emesis in male outpatients. Journal of Clinical Anesthesia. 1996;8:644-651.

6. Korttila K. The Study of Postoperative Nausea and Vomiting. British Journal of Anaesthesia. 1992;69:20s-23s.

7. Watcha M, White P. Postoperative Nausea and Vomiting: its etiology, treatment and prevention. Anesthesiology. 1992;77:162-184.

8. Watcha M, White P. Nausea and vomiting, pharmacology and clinical uses of antiemetic drugs. In: Prys-Roberts C, Brown BJ, eds. International Practice of Anaesthesia. Oxford: Butterworth-Heinemann; 1996:2/131/1-2/131/15.

9. Gold B, Kitz D, Lecky J, J N. Unanticipated Admission to the Hospital Following Ambulatory Surgery. Journal of the American Medical Association. 1989;262:3008-3010.

10. White P, Watcha M. Are New Drugs Cost-effective for Patients Undergoing Ambulatory Surgery? Anesthesiology. 1993;78:2-5.

11. Sackett D, Richardson W, Rosenberg W, Haynes R. Evidence-based Medicine. In: Richardson P, ed. New York: Churchill Livingstone; 1997:2.

12. Nease R, Owens D. Use of Influence Diagrams to Structure Medical Decisions. Medical Decision Making. 1997;17:263-275.

13. Owens D, Shachter R, Nease R. Representation and Analysis of Medical Decision Problems with Influence Diagrams. Medical Decision Making. 1997;17:241-262.

14. Owens D, Nease R. A Normative Analytic Framework for Development of Practice Guidelines for Specific Clinical Populations. Medical Decision Making. 1997;17:409-426.

15. Sox H, Blatt M, Hiigins M, Marton K. Expected Value Decision Making. . Medical Decision Making. Stoneham: Butterworth Publishers; 1988:147-166.

16. Cohen M, Duncan P, DeBoer D, Tweed W. The Postoperative Interview: assessing risk factors for nausea and vomiting. Anesthesia and Analgesia. 1994;78:7-16.

17. Pataky A, Kitz D, Andrew R, Lecky J. Nausea and Vomiting following Ambulatory Surgery: Are all procedures created equal? Anesthesia and Analgesia. 1998;67:s163.

18. Tramer M, Moore A, McQuay H. Meta-analytic comparison of prophylactic antiemetic efficacy for postoperative nausea and vomiting: propofol anaesthesia vs omitting nitrous oxide vs i.v. anaesthesia with propofol. British Journal of Anaesthesia. 1997;78:256-259.

19. Toner C, Broomhead C, Littlejohn I, et al. Prediction of postoperative nausea and vomiting using a logistic regression model. British Journal of Anaesthesia. 1996;76:347-351.

20. Carpenter R, Caplan R, Brown D, Stephenson C, Wu R. Incidence and Risk Factors for Side Effects of Spinal Anesthesia. Anesthesiology. 1992;76:906-916.

21. King M, Milazkiewicz R, Carli F, Deacock A. Influence of Neostigmine on Postoperative Vomiting. British Journal of Anaesthesia. 1988;61:403-406.

22. Andersen R, Krohg K. Pain as a Major Cause of Postoperative Nausea. Canadian Anaesthetists' Society Journal. 1976;23:366-369.

23. Melnick M, Johnson L. Effects of eliminating nitrous oxide in outpatient anesthesia. Anesthesiology. 1987;67:982-984.

24. Bellville J, Bross I, Howland W. Postoperative nausea and vomiting IV: Factors related to postoperative nausea and vomiting. Anesthesiology. 1960;21:186-193.

25. Apfel C, Greim C, Haubitz I, et al. A risk score to predict the probability of postoperative vomiting in adults. Acta Anaesthesiologica Scandinavica. 1998;42:495-501.

26. Apfel C, Greim C, Haubitz I, et al. The discriminating power of a risk score for postoperative vomiting in adults undergoing various types of surgery. Acta Anaesthesiologica Scandinavica. 1998;42:502-509.

27. Bonica J, Crepps W, Monk B, Bennet B. Postanesthetic Nausea, Retching and Vomiting. Evaluation of cyclizine (marezine) suppositories for treatment. Anesthesiology. 1958;19:532-540.

28. Smessaert A, Schehr C, Artusio J. Nausea and Vomiting in the immediate post anesthetic period. Journal of the American Medical Association. 1959;170:2072-2076.

29. Koivuranta M, Laara E, Snare L, Alahunta S. A survey of postoperative nausea and vomiting. Anaesthesia. 1997;52:443-449.

30. Palazzo M, Evans R. Logistic regression analysis of fixed patient factors for postoperative sickness: a model for risk assessment. British Journal of Anaesthesia. 1993;70:135-140.

31. Larsson S, Lundberg D. A prospective study of postoperative nausea and vomiting with special regard to incidence and relations to patient characteristics, anesthetic routines and surgical procedures. Acta Anaesthesiologica Scandinavica. 1995;39:539-545.

32. Oddby-Muhrbeck E, Jakobsson J, Andersson L, Askergren J. Postoperative nausea and vomiting. Acta Anaesthesiologica Scandinavica. 1994;38:52-56.

33. Apfel C, Rauch S, Goepfert C, Sefrin P, Roewer N. The impact of smoking on postoperative vomiting. Anesthesiology. 1997;87:A25.

34. Gratz I, Allen E, Afshar M, Joslyn A, Buxbaum J, Prilliman B. The effect of the menstrual cycle on the incidence of emesis and efficacy of Ondansetron. Anesthesia and Analgesia. 1996;83:565-569.

35. Tramer M, Moore A, McQuay H. Omitting nitrous oxide in general anaesthesia: meta-analysis of intraoperative awareness and postoperative emesis in randomised controlled trials. British Journal of Anaesthesia. 1996;76:186-193.

36. Tramer M, Moore A, McQuay H. Propofol anaesthesia and postoperative nausea and vomiting: quantitative systematic review of randomised controlled studies. British Journal of Anaesthesia. 1997;78:247-255.

Home > Research > Pre-op

© 2004 The University of Adelaide Last Modified Wednesday, 28-Jun-2006 11:32:05 CST  Health Informatics	Copyright | Privacy | Disclaimer