The choice of anaesthetic agent is usually driven by a combination of availability, anaesthetic practitioner preference and economics. This paper aims to provide information about the induction and recovery profile of different anaesthetic regimens and patient preferences for future anaesthesia.
Newer anaesthetic agents have contributed significantly to the expansion of day-surgery because they are safe and generally allow swift recovery and return to street fitness. However, there are differences between them in terms of administration methods, side-effect profile, and acquisition cost [1-4]. Promotion of individual agents may focus on supposed reductions in cost, improved throughput of day case load or increased patient acceptability. However, these claims have not been investigated using head-to-head trials, as these are not required as part of the drug licensing process. Anaesthetists use the best evidence available but that evidence is often of poor quality , and generally of insufficient quality to support anaesthetists in choices between anaesthetic agents. It is not clear as to how relevant varying clinical profiles are to patients and to what extent anaesthetists should be selecting agents based on these parameters.
We carried out a pragmatic, head-to-head randomized controlled trial (RCT) of four principal anaesthetic regimens used in adult day-surgery. We compared relative induction and pre- and post-discharge recovery characteristics, patient throughput and patient preferences when anaesthesia with propofol induction and maintenance (target-controlled infusion), or propofol induction with isoflurane/N2O, or sevoflurane/N2O maintenance, or sevoflurane/N2O alone was used in adult day case patients. This paper represents part of the data presented in an HTA (Health Technology Assessment) monograph  and an economic paper published elsewhere in which the clinical data were only partially presented . However, this paper presents purely clinical information with the full 7-day post-discharge data, patient preferences for future anaesthesia and all the adverse event data.
The study formed part of the Cost Effectiveness Study in Anaesthesia (CESA). This paper describes clinical differences in anaesthetic induction events, postoperative nausea and vomiting (PONV) rates, recovery characteristics with throughput timings and future anaesthetic preferences for 1158 adult patients randomized to one of four possible anaesthetic induction and maintenance agents. Full details of the methods used may be found in the Health Technology Assessment report , so only an overview is given here. The economic data from this trial can be found elsewhere .
CESA had complete Ethics Committee approval and all patients gave written informed consent. The ICMJE registration number is ISRCTN87609400. The sample population was recruited from consecutive day-surgery patients at Wirral Hospitals NHS Trust and Central Manchester and Manchester Children's University Hospitals NHS Trust (CMMCHT). The patients were scheduled for day case general, gynaecological or orthopaedic surgery between October 1999 and January 2001.
To be enrolled in the study, the patients needed to be assessed as fit for day case anaesthesia using the hospital's routine day case protocol. Factors that precluded day case anaesthesia were a history of an allergic or other serious adverse experience with anaesthesia; severe cardiovascular, respiratory, metabolic and central nervous system disease or anticipated airway management problems. Additional exclusion criteria were termination of pregnancy, use of sedative premedication or anticipated use of succinylcholine as part of the anaesthetic.
Patients were assigned randomly to one of the four study groups:
- Total intravenous (i.v.) anaesthesia utilizing target-controlled infusion induction and maintenance with propofol (Group P/P);
- I.v. propofol induction with isoflurane/N2O maintenance (Group P/I);
- I.v. propofol induction with sevoflurane/N2O maintenance (Group P/S);
- Inhalational sevoflurane/N2O induction and maintenance (Group S/S).
All patients underwent routine monitoring, which consisted of continuous electrocardiogram, automatic non-invasive blood pressure and finger or ear pulse oximetry recording. Gaseous monitoring consisted of inspired and expired oxygen, carbon dioxide and volatile agent concentrations.
All patients were preoxygenated via a clear, polythene facemask. Laryngeal masks used were size 3 or 4 for females and size 4 or 5 for males and endotracheal tubes were size 7.5 or 8.0 for females and size 8.5 or 9.0 for males. The doses of i.v. and inhalational anaesthetic agents were at the discretion of the anaesthetist. The fresh gas flow was reduced in all patient groups to 4 L after 5 min and 2 L after a further 10 min. Patients received between 50% and 70% nitrous oxide with oxygen during maintenance of anaesthesia; however, nitrous oxide was not used at any stage in the propofol TIVA (total intravenous anaesthesia) group. The use of local anaesthetic infiltration or regional blocks was encouraged. If adjunct analgesia was required during anaesthesia, this was provided by i.v. fentanyl, alfentanil or remifentanil. Patients given intraoperative morphine were excluded from the study. Prophylatic anti-emetics were not employed.
The primary outcome measure was PONV, which was recorded using a four-point scale both in the recovery room and in the postoperative ward. Adverse events during induction were noted (pain on injection, excitatory movement, laryngospasm, breath holding and coughing). Postoperatively, the recovery room nurse judged the patient's mental state on awakening as ‘alert and awake', ‘drowsy' or ‘agitated and distressed'. The patients were able to leave the anaesthetic recovery room when they were awake, protecting their airway and obeying commands. Return to the community was allowed when ambulatory, taking fluids and not bleeding or reporting excessive pain.
We recorded time spent in each part of the anaesthetic process (induction, maintenance, recovery room, discharge) and the number of patients admitted for overnight stay, together with the reason for admission.
We gave patients a 7-day diary to complete. Patients were asked to record their experience of nausea, vomiting, difficulty with concentration and forgetfulness each day post-discharge starting from the day of surgery (day 0) using 10 cm visual analogue scales that ranged from ‘no symptom', at zero, to ‘worse symptom possible', at 10 cm. The ratings from the visual analogue scale (VAS) were converted to indicate ‘presence' (VAS marked at 1 cm of more) or ‘absence' (VAS marked at less than 1 cm) of a symptom. Patients were also asked to report when they felt they resumed their normal day-to-day activities.
Patient preferences for future procedures.
Patients were interviewed by telephone around 7 days postoperatively using a structured questionnaire. During the interview they were asked about their views of the anaesthetic process and their future preferences for anaesthetic induction technique should they require anaesthesia again in the future.
The study was powered to detect a reduction in PONV from 20% to 10% at 80% power using a two-tailed significance test at the 1% level of significance. This required 330 patients in each treatment arm. This stringent significance level was chosen because it was anticipated that there would be multiple comparisons among the four treatment arms .
Computer-generated pseudorandom numbers were used to allocate patients to one of the four arms. Block randomization with randomly varied block size was used to stratify for gender and hospital site. Randomization was carried out after recruitment into the study, and codes were held by the study office. Trained research staff enrolled participants and informed the anaesthetist of the participant's study arm allocation.
Due to the obvious differences in the induction techniques, it was not possible to blind the anaesthetists to the treatment allocation. The data analysts were masked to treatment allocation until after the analysis of trial results.
All data were collected prospectively by trained research staff. The patients' age, sex and type of surgery were recorded, as was the duration of anaesthetic induction, maintenance, recovery and time to discharge (Table 1). The patients were discharged from the anaesthetic recovery room when they were awake, protecting their airway and obeying commands. Return to the community was allowed when the patient was ambulant, taking fluids and not bleeding or reporting excessive pain.
Analysis was planned to be by intention to treat. The data were analysed using SASTM 6.12 (SAS Institute, Cary, NC, USA) and SPSSTM 10 (SPSS Inc., Chicago, IL, USA). Categorical data were analysed using χ2-tests; Fisher's exact test was used when expected frequencies were less than five. Anaesthetic time data and time to resume normal activities were analysed using an unpaired t-test.
Patients were recruited between October 1999 and January 2001. Consent was sought from 1548 eligible patients. Over half (226/425) of those who did not want to take part in the trial refused because they were not willing to risk a 1 : 4 chance of undergoing a volatile induction. Only 5 (1%) were unwilling to accept the chance of an i.v. induction. ‘No reason' or ‘I don't like the idea' was given by 106 (25%) people, and 32 (8%) had no time to be involved.
We achieved a 73% recruitment rate, and 1158 patients were randomized to take part in the trial. Ninety-five patients (8%) were withdrawn from the trial after randomization but prior to anaesthesia commencing, 32 due to the operation being cancelled (34%), 27 (28%) withdrawn by the anaesthetist, 15 (16%) due to protocol violations and 15 (16%) due to withdrawal of consent. These 95 patients were withdrawn from the analysis as no meaningful analysis could be completed. Protocol violations reflected the administration of the drugs morphine and/or anti-emetics, in the light of new clinical decisions about the nature of the necessary surgery. These drugs would render the end-point measurement of PONV meaningless in the context of the study.
Data in Table 1 show no significant differences in patients' age, sex, ASA grade or type of surgery between the study arms among those remaining in the study.
Table 2 gives the incidence of adverse events during induction. Fifty-eight patients reported pain during injection of propofol. Compared to the propofol groups, more patients displayed excitatory movements and breath holding during induction with sevoflurane. Hiccup was more common in the propofol groups.
On awakening in the recovery room, assessment of the patient's mental state in the anaesthetic recovery room showed that most patients were reported to be alert and awake. There was no significant difference between groups in the number of patients described as ‘alert and awake', ‘agitated and distressed' or ‘drowsy' (Table 3).
PONV in the anaesthetic recovery room and in the postoperative ward was scored using a four-point scale, 0 = absence, 1 = nausea only, 2 = one emetic episode and 3 = multiple emetic episodes (Table 3). In both areas, using ‘at least one episode of PONV' as a measure, more emetic episodes were recorded in the sevoflurane group.
The mean time required for induction was longer in the sevoflurane group. Maintenance and recovery time was similar in all groups. There was no difference between groups in the time spent on the postoperative ward before discharging home (Table 4).
Ninety-four patients (9%) were admitted to the hospital overnight during this study (Table 5). Twenty-three of these patients (2.16%) were admitted for control of pain or PONV.
The return rate for the diaries was 778 (73%) and 73% (n = 571) of this sample were women.
No statistical difference (χ2, P < 0.01) in reported pain was found between the four randomization groups for each day over the 7-day postoperative episode. No statistical difference in reported difficulties in concentrating or forgetfulness was found between the four groups over the 7-day postoperative episode. Patients reported that it took a mean of 6.5 days (95% CI: 6.0-7.0, n = 693) to resume their normal day-to-day routine or return to work. There was no significant difference between the four groups (Table 6).
Patient preferences for future anaesthesia
Table 7 describes patient preferences for future anaesthetic induction.
This study indicates that there are transient differences in outcome between the four regimens investigated. Sevoflurane alone is associated with increased PONV, both in the recovery room and on the ward. This difference persists beyond discharge from hospital but disappears by 48 h. There was also an increased incidence of adverse events at induction experienced with sevoflurane alone. There was no difference in patients' mental state on awakening in the recovery room. There was no difference in recovery time, time to discharge or admissions to hospital overnight for anaesthetic reasons between groups. Patients who had sevoflurane only were more likely to recall an unpleasant induction experience and least likely to want the same induction method again. Women completed 73% of the returned diaries, which closely reflects the percentage of women taking part in the study. Men and women appear to be equally likely to complete and return a postoperative diary.
Our main conclusions are that sevoflurane/sevoflurane is not a preferred regimen for day case surgery in adults both when pre-discharge PONV is used as the primary outcome measure and when taking patient preferences into consideration. In addition, this study shows marked differences in PONV with different regimens but with no difference in the length of stay; hence the differences in PONV rate did not change patient throughput when measured as time to home discharge or overnight admission rates. This corroborates other studies that show that home discharge and overnight admission after day-surgery is affected by social, surgical and organizational factors, rather than the choice of anaesthetic regimen .
Despite marked differences in pre-discharge PONV with different regimens, differences in PONV rates had disappeared 48 h after discharge. From our results, it is clear that patients were not basing their anaesthetic preferences on either pre- or post-discharge PONV rates, but on the induction method. Patients were generally able to recall the type of induction agent they were given and they had clear views about the anaesthetic process, expressing preferences for i.v. induction. It is not always possible to incorporate preferences into anaesthetic practice. However, patients undergoing day case surgery are generally fit and healthy. Discussing their preferences for alternative methods of anaesthesia prior to surgery may lead patients to feel more involved in their care and to be better prepared for what to expect. There seems no reason as to why patient preferences cannot be incorporated, particularly as the PONV rate is highest in the least preferred option. This is much more straightforward than in other therapeutic areas where patient preferences can be in direct opposition to clinical priorities, such as management of atrial fibrillation with warfarin .
Limitations of the study
The CESA project was designed in 1997. The four regimens were selected as they represented current and emerging anaesthetic practice patterns . A literature review at that time showed no clear evidence that nitrous oxide has any impact on clinical parameters . Since the trial was carried out, no new agents or technologies have emerged, so these comparators retain clinical relevance.
The data analysts were masked to the allocated anaesthetic regimen until analysis was complete. However, the anaesthetists, researchers and patients were not masked to the allocated anaesthetic regimen, although, at follow-up, 13% of patients did not correctly recall the induction method used. This may be because all patients were pre-oxygenated and had an i.v. cannula inserted before induction. The inability to mask participants to the treatment regimen could have biased treatment effects if preconceived ideas about the relative merits of the techniques caused research workers to score outcomes subjectively.
Despite the trial protocol excluding the use of prophylactic anti-emetics, the PONV incidence in this trial is at the low end of the reported range [2,10]. Our data lend support to the notion that prophylactic anti-emetics may be ineffective. The direction of the treatment effect in favour of propofol/propofol is supported by the literature. Meta-analyses of the evidence generally suggest that the use of propofol reduces the risk of PONV, with odds ratios of 4.24 for sevoflurane compared with propofol induction  and 0.267 for propofol compared with inhalational agents for induction .
Problems with patient recruitment meant the sample size target was not met (81% of estimated target), although this trial is larger than other published studies in this area . In addition, the trial was powered to test for differences in the incidence of PONV between treatment groups rather than equivalence. The absence of statistically significant differences in the PONV rate between the three regimens that used propofol for induction could be due to insufficient observations rather than equivalence between the three propofol regimens.
Multiple outcomes were tested in this RCT. The analysis used a low level of statistical significance (1%) to reduce the impact of multiple testing. In addition, the key variables tested were defined a priori in a detailed analysis plan, before data analysis, reducing the chance of spurious results from ‘dredging' the data.
The eligibility criteria and willingness of patients to participate in a trial may limit the relevance of the results to patients outside the trial. However, the day case procedures studied in this trial broadly reflect the case-mix of UK day case surgery, indicating that the eligibility criteria used in this study were unlikely to give an atypical trial population. There was no evidence of differences in demographic or clinical characteristics between trial participants and non-participants.
Our study emphasizes the importance of 7-day follow-up when measuring post-anaesthesia sequelae. We suggest that inhalational induction of anaesthesia with sevoflurane is not ideal for adult day case anaesthesia. Approximately one in five patients who underwent inhalational induction of anaesthesia found this an unpleasant experience. Patient preferences for anaesthetic regimen appear to relate more closely to the method of induction than to the observed differences in PONV rate. This study provides practice-based evidence to inform the debate in this key area of the acute sector at a time when effective expansion of day-surgery services is an important priority.
This study was funded by a grant from the National Co-ordinating Centre for Health Technology Assessment (UK). The views expressed in this paper are those of the authors and not of the funding bodies.
1. Lerman J, Sikich N, Kleinman S, Yentis S. The pharmacology of sevoflurane in adults and children. Anesthesiology
2. Fish WH, Hobbs AJ, Daniels MV. Comparison of sevoflurane and total intravenous anaesthesia
for daycase urological surgery. Anaesthesia
3. Smith I, Thwaites AJ. Target-controlled propofol vs. sevoflurane: a double-blind, randomised comparison in day-case anaesthesia. Anaesthesia
4. Gupta A, Stierer T, Zuckerman R, Sakima N, Parker SD, Fleischer LA. Comparison of recovery profile after ambulatory anesthesia with propofol, isoflurane, sevoflurane and desflurane: a systematic review. Anesth Analg
5. Myles PS, Bain DL, Johnson F, McMahon R. Is anaesthesia evidence-based? A survey of anaesthetic practice. Br J Anaesth
6. Elliott RA, Payne K, Moore JK et al.
. Which anaesthetic agents and techniques are most cost effective for day surgery? Health Technol Assess
7. Elliott RA, Payne K, Moore JK et al.
. Clinical and economic choices in anaesthesia for day surgery: a prospective randomised controlled trial. Anaesthesia
8. Protheroe J, Fahey T, Montgomery AA, Peters TJ, Smeeth L. The impact of patients' preferences on the treatment of atrial fibrillation: observational study of patient based decision analysis Commentary: patients, preferences, and evidence. BMJ
9. Pollard BJ, Elliott RA, Moore EW. Anaesthetic agents in adult day case surgery. Eur J Anaesthesiol
10. Tramer M, Moore A, McQuay H. Propofol anaesthesia and postoperative nausea and vomiting
: quantitative systematic review of randomized controlled studies. Br J Anaesth
11. Joo HS, Perks WJ. Sevoflurane versus propofol for anesthetic induction: a meta-analysis. Anesth Analg
12. Sneyd JR, Carr A, Byrom WD, Bilski AJT. A meta-analysis of nausea and vomiting following maintenance of anaesthesia with propofol or inhalational agents. Eur J Anaesthesiol