British Journal of Anaesthesia 1992; 69 (Suppl. 1): 20S-23S

THE STUDY OF POSTOPERATIVE NAUSEA AND VOMITING K. KORTTILA

PATIENTS, PROCEDURES AND ANAESTHESIA

Several factors may influence PONY and these must be controlled when studying PONY. Background factors Table I lists some background factors which have been shown to influence the incidence and severity of PONY. The frequency of PONY in children is the same in both boys and girls and is about twice that in adults [41] until it decreases after puberty [36]. PONY is more common in women than in men, a difference that is thought to be hormonal in origin [8]. Diamond, Bailey and McPhee [12] suggested that the effect of an antiemetic may be related to the gender of the patient. They suggested that there is a dose-dependent pharmacodynamic difference in the antiemetic action of metoclopramide, and that a dose of 50 mg in females is equivalent to 20 mg in males. A previous history of motion sickness and also emesis after previous anaesthetics, contribute to a higher incidence of PONY [20, 25]. It is obvious that administration of antiemetic drugs before surgery may decrease the incidence or severity of PONY [26,27]. If women of childbearing age are studied,

the phase of the menstrual cycle should be controlled [5, 15]. With advancing age, in excess of 70 yr, it appears that gender does not influence PONY [33]. Factors related to operation and anaesthesia Table II lists some factors related to operation and administration of anaesthesia which have been shown to influence PONY. It is important that before conclusions are drawn from clinical data (e.g. on the efficacy of a new antiemetic drug) these surgical and anaesthesia related variables are controlled. The incidence of PONY may differ in surface operations compared with intra-abdominal operations or ophthalmic surgery [22]. Surgery which involves stretching the gut wall may contribute to PONY whereas the presence of a nasogastric tube may decrease the frequency of symptoms [22,38]. Premedication with opioids may stimulate the vomiting centre and cause nausea and vomiting [33], whereas premedication with atropine [33] and hyoscine [4] may prevent PONY. The incidence of PONY may depend on the anaesthetic induction agent used. Anaesthesia induced with etomidate has been associated with more PONY compared with thiopentone [29] and it has been suggested that propofol possesses antiemetic actions [24]. The incidence of PONY may be as great as 80 % after ether anaesthesia [19,22] whereas a low incidence of PONY has been noted after propofol TABLE I. Backgroundfactors which may modify PONV and which must be controlled in clinical studies Patient age [36,41] and sex [7, 12] History of motion sickness [20, 25] History of PONY after previous anaesthesia [20, 25] Administration of antiemetics before operation [26, 27] Phase of menstrual cycle [5, 15]

TABLE II. Factors related to operation and anaesthesia which may modify PONV Type and duration of operation [10, 12,38] Type of premedication [4,33] Type of induction agent [24,41] Type of maintenance agent [19,28] Reversal of muscle relaxation [21] Postoperative pain and its treatment [3, 32] Movement of patient [20,31]

(Br. J. Anaesth. 1992; 69 (SuppI.1): 20S-23S) KEY WORDS

Measurement techniques. Vomiting: nausea.

KARl KORTTILA, M.D., PH.D., Department of Anaesthesia, Women's Hospital, University of Helsinki, Haartmaninkatu 2, SF-00290 Helsinki, Finland.

Downloaded from http://bja.oxfordjournals.org/ at RMIT University Library on July 2, 2015

Clinical investigators have been interested in factors related to postoperative nausea and vomiting (PONY) and in its assessment for several decades [10,11,13,22]. It was reported as early as 1956 that the type of anaesthesia or operation, and also the administration of antiemetics was likely to influence the incidence and severity of PONY [8, 10,22]. Recently it has become evident that among the many factors which are associated with PONY are included the experience of the anaesthetist [16] and the phase of the menstrual cycle [15]. As several factors may modify PONY and new drugs continue to be developed, it is important, therefore, that studies evaluating PONY are designed carefully. This review will present some studies on factors related to PONY and on the assessment of PONY. It is hoped that the methodology discussed may help investigators in designing and carrying out studies on PONY and may also help clinicians in reading the literature on PONY.

ASSESSMENT OF PONY anaesthesia [28]. One study [21] suggested that the incidence and severity of PONY were lower if neostigmine was omitted and muscle power allowed to recover spontaneously compared with patients who received neostigmine. It has been suggested also that postoperative pain may be a major cause of nausea [31,32]. Movement of the patient from one bed to another or even transport to the ward may cause nausea and vomiting [20,31]. ASSESSMENT OF NAUSEA AND VOMITING

Overall assessment of symptoms I t is necessary to define methods which categorize all symptoms over a defined period rather than listing the presence or absence of each of the symptoms. Knapp and Beecher [22] have termed nausea slight, moderate or severe and retching or vomiting slight if it occurred once or twice, moderate if there were three to five episodes and severe if it occurred more than five times in the first 4 h after operation. Bellville and co-workers [7] used a scoring system for nausea, retching and vomiting which provided a pooled numeric score for the 24-h period after operation and the overall score was then rated as none, mild, moderate or severe. Tigerstedt, Salmela and Aromaa [39] used a simple method where they rated nausea as none, mild or severe and vomiting as none, mild (one to two times) or severe (three or more times). Bailey and colleagues [4] defined nausea as mild if it lasted less than 2 hand severe if it lasted more than 2 h. In several studies [15,17,25-27], we have evaluated PONY at five intervals during the first 24 h after operation: 0-2 h, 2-6 h, 6-12 h, 12-18 hand 18-24 h. At the end of each interval, a trained nurse assessed whether retching or vomiting had occurred and asked the patients if they felt nauseated. The results were scored as none, nausea, retching and

vomiting. If the patients had all of the symptoms (nausea, retching and vomiting) they were categorized as having had vomiting. If the postoperative period is not classified into several intervals it is still useful to separate the time spent in the postanaesthesia care unit from that spent in the ward. Bodner and White [9] used a subjective rating scale (from 0 = no nausea to 11 = as bad as possible) and a visual analogue scale (100 mm: no nausea to as bad as possible) for nausea. They also classified retching and vomiting together as emetic episodes and counted the total number of such episodes. If an investigator does not want to study prevention of PONY but only treatment, it is necessary to define when treatment is given. In one study [9], the investigators waited until the patient had either nausea for as long as 10 min or at least two emetic episodes (i.e. retched or vomited at least twice) before antiemetic drugs were administered.

Rescue medication and outcome of PONV The outcome which should be included in studies on PONY is the need to give an antiemetic or the need to give rescue medication if new drugs are being studied. The design of a study should define when to administer an antiemetic or rescue medication, that is to describe the criteria for administration of these drugs. For this purpose we have usually [15,17,25-27] used droperidol both as a definitive antiemetic and also the first rescue medication for prolonged nausea (lasting 10 min) or for vomiting. Bodner and White [9] used two rescue medications. The first one was administered for nausea lasting 30 min and the second for persistent nausea after failure of the first rescue medication to treat symptoms. The numbers of patients needing anti emetics or rescue antiemetics are good endpoints for statistical analysis. Transient PONY is very distressing to the patient and every effort should be made by the anaesthetist to decrease this problem. Rarely, persistent nausea together with prolonged vomiting may lead to dehydration and interfere with nutrition and oral drug therapy. In day surgery, PONY may lead to delay in discharging patients home or unanticipated admission to hospital [23, 28, 40]. Thus in day surgery the assessment of home readiness and also the number of patients admitted to hospital after day surgery are important endpoints. However, studies using such endpoints would require very large numbers of patients. In inpatients, forceful vomiting has resulted in dehiscence of abdominal wounds and even rupture of the oesophagus [6]. These complications are as rare as aspiration of stomach contents into the lungs and it would not be possible to use these complications as outcome measures in studies of PONY. Other adverse effects such as sedation, headache, facial flushing, constipation, visual disturbances and changes in biochemical variables obviously need to be studied with all new drugs. One example of the importance of including comparison of side effects is a recent study on the efficacy of ephedrine in the prevention of postoperative nausea and vomiting

Downloaded from http://bja.oxfordjournals.org/ at RMIT University Library on July 2, 2015

Interpretation of symptoms It is important to define precisely the meanings of nausea, retching and vomiting. Knapp and Beecher [22] gave excellent definitions 35 years ago of nausea and emesis, and these are still useful today. Nausea is a subjective sensation which should be evaluated by the patient, not by the observer. The feeling is best described as the desire to vomit without indulging in expulsive muscular movements. When nausea becomes severe, the secretion of saliva is increased and is associated with vasomotor disturbances and sweating. The feature that distinguishes retching from vomiting is production of even the smallest amount of stomach contents. When no stomach contents are expelled, the expulsive efforts are classified as retching. Retching is usually indicative of an empty stomach and is generally as unpleasant for the patient as vomiting. With both retching and vomiting, pharyngeal muscles relax, the soft palate is elevated, the diaphragm descends and there are spasmodic contractions of respiratory chest muscles and abdominal wall muscles. Retching and vomiting may also be grouped together under the common term "emetic episode".

21S

22S TABLE

BRITISH JOURNAL OF ANAESTHESIA III. Factors involved in the design of a study of PONV

Variables to be controlled are listed in tables I and II. These should not be different between groups. Drugs should be administered in a double-blind, random fashion. If different patient groups/anaesthesias are used, stratified randomization may be required. The main endpoint of efficacy should be patients with no PONY. Nausea should be rated separately from retching and vomiting. Nausea may be rated as yes or no, as none, mild or severe, on a verbal (0-10 choices) or visual analogue (0-100 mm) scale. Emetic episodes (i.e. retching and vomiting) may be assessed together as the number of emetic episodes or separately as retching and vomiting. Adverse effects (sedation, headache) should be compared. In day surgery, outcome variables, such as time to oral intake, time to discharge home and the number of unanticipated admissions to hospital are important secondary (or even primary) endpoints.

[35]. In patients undergoing gynaecological laparoscopy, both ephedrine and droperidol were effective in preventing PONY compared with saline but patients given ephedrine were less sedated and could be discharged home earlier than patients given droperidol [35]. The study of PONY should take into account several factors including patient variables, type of surgery and anaesthesia, and should include assessment of the adverse effects of anaesthetic techniques or drugs studied. One example of the design of a study and of assessment of PONY is given in table III. STATISTICS AND POWER CONSIDERATIONS

Because of the nature of the endpoints of efficacy obtained when assessing PONY, non-parametric statistics (e.g. Fisher's exact probability, chi-square or Kruskal-Wallis tests) should be used for comparison between different treatments. The main endpoint of efficacy should be the number of patients with no nausea, retching or vomiting. A second comparison may be the number of patients with nausea only. As discussed above, the severity of nausea and also the number of emetic episodes may also be secondary endpoints for comparison. However, it is recommended that the secondary endpoints are first well defined and after that comparisons are made between numbers of patients. It is also of utmost importance that power analysis is applied in negative studies before concluding that there is no difference between two treatments in the prevention or treatment of PONY. An example of probably inadequate group sizes has been published recently [37]. The authors [37] concluded that nitrous oxide had no effect on PONY in patients undergoing day-case laparoscopy. If the sample size is too small there is an increased risk of a false negative finding. What is the power of a test? The required size of a clinical trial is related to a (significance) and ~ (power) and the size of the

REFERENCES

1. Aleong J, Bartlett DE. Improved graphs for calculating sample size when comparing two independent binomial distributions. Biometrics 1979; 35: 875-881. 2. Altman DG. Statistics and ethics in medical research III. How large a sample? British Medical Journal 1980; 281: 1336-1338. 3. Andersen R, Krogh K. Pain as a major cause of postoperative nausea. Canadian Anaesthetists Society Journal 1976; 23: 366-369. 4. Bailey PL, Steisand JB, Pace NL, Bubbers SJM, East KA, Mulder S, Stanley TH. Transdermal scopolamine reduces nausea and vomltmg after outpatient laparoscopy. Anesthesiology 1990; 72: 977-980. 5. Beattie WS, Lindblad T, Buckley DN, Forrest JB. The incidence of postoperative nausea and vomiting is influenced by the day of menstrual cycle. Canadian Journal of Anasthesia 1991; 38: 298-302. 6. Bellville JW. Postanesthetic nausea and vomiting. Anesthesiology 1961; 22: 773-780. 7. Bellville JW, Bross IDJ, Howland WS. A method for the clinical evaluation of antiemetic agents. Anesthesiology 1959; 20: 753-760. 8. Bellville JW, Bross IDJ, Howlans WS. Postoperative nausea and vomiting. IV. Factors related to postoperative nausea and vomiting. Anesthesiology 1960; 6: 186-193. 9. Bodner M, White PF. Antiemetic efficacy of ondansetron after outpatient laparoscopy. Anesthesia and Analgesia 1991; 73: 250-254. 10. Bonica 11, Crepps W, Monk B, Bennet B. Postanesthetic nausea, retching and vomiting. Evaluation of cyclizine suppositories for treatment. Anesthesiology 1958; 19: 532-540. 11. Dent S, Ramachandra V, Stephen RC. Postoperative vomiting: incidence, analysis and therapeutic measures in 3000 patients. Anesthesiology 1955; 16: 564-572. 12. Diamond MJ, Bailey D, McPhee A. The gender-dependent pharmacodynamic difference in the antiemetic action of metoclopramide is a dose-related phenomenon. Canadian Journal of Anaesthesia 1988; 35: S65. 13. Didier EP, Barila TG, Slocum HC, Lindgren VV, McCawley EL. An evaluation of antiemetic drugs in the control of postoperative vomiting. Surgical Forum 1954; 5: 707-712. 14. Fleiss JL. Statistical Methods for Rates and Proportions, 2nd Edn. New York: John Wiley and Sons, 1981; 258-280. 15. Honkavaara P, Lehtinen A-M, Hovorka J, Korttila K. Nausea and vomiting after gynaecologicallaparoscopy depends upon the phase of the menstrual cycle. Canadian Journal of Anaesthesia 1991; 38: 876-879.

Downloaded from http://bja.oxfordjournals.org/ at RMIT University Library on July 2, 2015

Non-parametric statistics should be used to compare different treatments and a large enough sample size should be studied before concluding that there is no difference between treatments.

difference between the success criteria of the two treatment groups [2, 30]. The power may be calculated retrospectively to assess the chance that a study had of detecting a significant and clinically relevant difference. The power may also be used prospectively to calculate a suitable sample size. An estimate of the incidence of, for example nausea and vomiting, should usually be available from previous studies or from pilot studies. For example, in two negative studies on PONY, Hovorka, Korttila and Erkola [16, 18] used power analysis to indicate they had an 80 % chance of finding an effect (at P < 0.05) that stomach emptying decreased the incidence of PONY by 20 % (from 70 % to 50 %) in 101 patients [16] and that nitrous oxide increased PONY by 20 % in 150 patients (from 50% to 70%) [18]. Usually a power of 80-90 % is recommended (80 to 90 % chance of detection) and the significance should be at least < 0.05. Some authors [1,30] have published graphs to estimate the sample size and power but accurate tables for sample size per group for a twotailed test on proportions [14] have also proved useful [16, 18,34] and are recommended.

ASSESSMENT OF PONY

29.

30. 31.

32. 33. 34.

35.

36. 37.

38.

39.

40.

41.

surgery. Acta Anaesthesiologica Scandinavica 1991; 34: 400-403. Korttila K, Tammisto T, Aromaa U. Comparison of etomidate in combination with fentanyl or diazepam, with thiopentone as an induction agent for general anaesthesia. British Journal of Anaesthesia 1979; 51: 1151-1157. Mould RF. Clinical trial design in cancer. Clinical Radiology 1979; 30: 371-381. Muir Warner MA, Offord KP, Buck CF, Harper IV, Kunkel SE. Role of nitrous oxide and other factors in postoperative nausea and vomiting: a randomised and blinded prospective study. Anesthesiology 1987; 66: 513-518. Parkhouse J. The cure for postoperative vomiting. British Journal of Anaesthesia 1963; 35: 189-193. Purkis IE. Factors that influence postoperative vomiting. Canadian Anaesthetists Society Journal 1964; 11: 335-353. Rosen MA, Roizen MF, Eger EI, Glass RH, Martin M, Dandekar PV, Dailey PA, Litt L. The effect of nitrous oxide on in vitro fertilization success rate. Anesthesiology 1987; 67: 42-44. Rothenberg DM, Parnass SM, Litwack K, McCarthy RJ, Newman LM. Efficacy of ephedrine in the prevention of postoperative nausea and vomiting. Anesthesia and Analgesia 1991; 72: 58-61. Rowley MP, Brown TCK. Postoperative vomiting in children. Anaesthesia and Intensive Care 1982; 10: 309-313. Sengupta P, Plantevin O. Nitrous oxide and day case laparoscopy: Effects on nausea, vomiting, and return to normal activity. British Journal of Anaesthesia 1988; 60: 570-573. Smessaert S, Schehr CA, Artusio JF jr. Nausea and vomiting in the immediate postanesthetic period. Journal of American Medical Association 1959; 170: 118-121. Tigerstedt I, Salmela L, Aromaa U. Double-blind comparison of transdermal scopolamine, droperidol and placebo against postoperative nausea and vomltmg. Acta Anaesthesiologica Scandinavica 1988; 32: 454--457. Valanne J, Korttila K. Effect of a small dose of droperidol on nausea, vomiting and recovery after outpatient enflurane anaesthesia. Acta Anaesthesiologica Scandinavica 1985; 29: 359-362. Vance JP, Neill RS, Norris W. The incidence and aetiology of postoperative nausea and vomiting in a plastic surgical unit. British Journal of Plastic Surgery 1973; 26: 336-339.

n,

Downloaded from http://bja.oxfordjournals.org/ at RMIT University Library on July 2, 2015

16. Hovorka J, Korttila K, Erkola O. Nitrous oxide does not increase nausea and vomiting following gynaecological laparoscopy. Canadian Journal of Anaesthesia 1989; 36: 145-148. 17. Hovorka ], KorttiIa K, Erkola O. The experience of the person ventilating the lungs does influence postoperative nausea and vomiting. Acta Anaesthesiologica Scandinavica 1990; 34: 203-205. 18. Hovorka J, Korttila K, Erkola O. Gastric aspiration at the end of anaesthesia does not decrease postoperative nausea and vomiting. Anaesthesia and Intensive Care 1990; 18: 58-61. 19. Jenkins LC, Lahay D. Central mechanisms of vomiting related to catecholamine response: anaesthetic implications. Canadian Anaesthetists Society Journal 1971; 18: 434--441. 20. Kamath B, Curran J, Hawkey C, Beattie A, Gorbutt N, Guiblin H, Kong A. Anaesthesia, movement and emesis. British Journal of Anaesthesia 1990; 64: 728-730. 21. King ML, Milazkiewicz R, Carli F, Deacock AR. Influence of neostigmine on postoperative vomiting. British Journal of Anaesthesia 1988; 61: 403-406. 22. Knapp MR, Beecher HK. Postanesthetic nausea, vomiting and retching. Journal of American Medical Association 1956; 160: 376-385. 23. Korttila K. Practical discharge criteria. Problems in Anesthesia 1988; 2: 144-151. 24. Korttila K. Propofol-an update. Current Opinion in Anesthesiology 1990; 3: 559-563. 25. Korttila K, Hovorka J, Erkola O. Nitrous oxide does not increase the incidence of nausea and vomiting after isoflurane anesthesia. Anesthesia and Analgesia 1987; 66: 761-765. 26. KorttiIa K, Kauste A, Auvinen J. Comparison of domperidone, droperidol and metoclopramide in the prevention and treatment of emetic sequelae after balanced general anesthesia. Anesthesia and Analgesia 1979; 58: 396-400. 27. Korttila K, Kauste A, Tuominen M, Salo H. Droperidol prevents and treats nausea and vomiting after enflurane anaesthesia. European Journal of Anaesthesiology 1985; 2: 379-385. 28. Korttila K, Ostman PL, Faure E, Apfelbaum JL, Prunskis J, Ekdawi M, Roizen MF. Randomized comparison of recovery after propofol-nitrous oxide versus thiopental-isofluranenitrous oxide anaesthesia in patients undergoing ambulatory

23S

The study of postoperative nausea and vomiting.

British Journal of Anaesthesia 1992; 69 (Suppl. 1): 20S-23S THE STUDY OF POSTOPERATIVE NAUSEA AND VOMITING K. KORTTILA PATIENTS, PROCEDURES AND ANAE...
432KB Sizes 0 Downloads 0 Views