Postoperative instructions: good compliance but is the advice sound?

In 1972, Tom Ogg contacted a group of 100 day-case patients and found that 31% travelled home without a responsible escort [1]; 9% of car-owning patients drove themselves home, while 39% had driven within 12 h and 73% within 24 h of their operation. He also reported the now infamous patient who not only drove home but then went on to drive a bus-load of passengers on a 95-mile journey on the same day as his operation. In addition to these violations, 6% of patients drank alcohol within 24 h [1]. A smaller survey in 1978 showed 23% of car-driving patients drove within 24 h of day surgery, while 20% consumed alcohol over a similar time frame [2]. Another investigation from the United States revealed 19% driving and 8% drinking alcohol within 24 h [3]. This current edition of Anaesthesia reports the first large evaluation of patients' compliance with instructions given after day-case surgery. Correa and colleagues from the Toronto Western Hospital [4] telephoned 750 day-case patients 24 h after their operation. Four hundred and sixty-one of the patients had received general anaesthesia, 254 underwent monitored anaesthesia care (or conscious sedation) while 35 had regional anaesthesia. Prior to discharge, all had been instructed not to drink alcohol or drive for 24 h and were advised to have a responsible adult stay with them overnight. All patients went home with an escort, but 4% did not have a companion with them overnight. Of the regular drinkers of alcohol, 1.8% indulged their habit within the 24-h period, while 4% of car owners drove. Although the current results were obtained in Canada, it is likely that a similar pattern would be observed in the United Kingdom. Should we be concerned by these findings? Certainly there seems to be an improving trend over the years in patient compliance with postoperative instructions, perhaps reflecting the amount of time devoted to passing on this information. Written instructions have previously been shown to increase patient compliance compared with verbal advice alone [2]; all of the patients in the current study received both verbal and written guidelines. In addition to forgetting or misunderstanding instructions, patients may deliberately disregard them and there is some evidence that a few choose to do so [2]. Patients may be more inclined to ignore instructions if they fail to appreciate the risks involved, or if they consider the advice to be inappropriate. It would also be interesting to know how often surgery is cancelled on the proposed day of operation because a patient admits that they are unable to comply with postoperative requirements. Most day units, mine included, would not proceed with anaesthesia if the patient could not arrange for a suitable escort or was intending to drive home. Such cancellations sometimes occur, although their frequency is seldom reported. We need to be confident that such a dramatic course of action is justifiable. The hospital environment is more sheltered than the outside world and sending patients home with an escort seems wise, but little is known about the value of a care provider beyond that time. A patient requiring help is likely to take advantage of an assistant who is nearby, but may well manage safely in their absence. An incapacitated patient may be detected by a partner sleeping in the same bed or room, but would almost certainly remain undiscovered by a friend sleeping elsewhere within the house, yet both would be acceptable escorts! Such dramatic events are exceedingly rare and research could usefully examine how great a role escorts actually play in ensuring patient safety, rather than convenience. There is similarly little evidence regarding alcohol consumption after anaesthesia. In a volunteer study, both thiopentone and alcohol independently impaired a variety of psychomotor functions [5]. However, only body sway and the patient's feeling of lightheadedness increased to a greater degree in volunteers consuming the equivalent of 2½ pints of beer 4 h after an induction dose of thiopentone compared to those who had not received a prior anaesthetic [5]. In a similar study design, the effects of alcohol consumption were not enhanced by the previous administration of sedative doses of midazolam or midazolam–fentanyl [6, 7]. There is little evidence concerning other anaesthetics. Although alcohol may potentiate the depressant effects of residual anaesthesia, there is no evidence that this occurs to a dangerous degree (especially after sedation). It may be more logical to advise patients to reduce their alcohol consumption rather than avoiding it completely. Many anaesthetists have applied such a policy to themselves, with good effect. Driving is the activity most likely to cause serious harm to the patient (and others), so meaningful advice is therefore most important. ‘Official’ advice is remarkably lacking. The Driver and Vehicle Licensing Agency (DVLA) offer no guidance on fitness to drive after anaesthesia, as it is ‘only a temporary condition’ (personal communication). Insurance companies advise against driving, but defer the details to the responsible doctor or hospital. Most day units advise patients not to drive for 24 h, but some suggest 48 h. This advice is little different to that advocated almost 30 years ago [1], despite significant improvements in the available drugs. The suggested limits are conveniently round numbers, implying an arbitrary element to the original guidelines. Is this advice still appropriate and to what extent is it evidence based? In 1983, Herbert and colleagues showed that choice reaction time and the patient's assessment of co-ordination were still impaired to some extent on the second postoperative day after elective hernia repair [8]. Anaesthesia consisted of diazepam premedication, induction with thiopentone and maintenance with halothane – nitrous oxide; hardly a modern ‘day-case’ anaesthetic. Numerous investigators have shown that similar measures of psychomotor performance return to pre-operative values within 2–5 h following anaesthesia with propofol, isoflurane, sevoflurane or desflurane [9-14]. These basic tests may not accurately reflect driving ability, however. Some investigators have used sophisticated driving simulators to assess postoperative recovery, but they have generally looked at anaesthetic agents in isolation. Korttila and colleagues recommended no driving for 24 h after a single anaesthetic dose of either methohexitone or thiopentone. This advice was based on the severity of the impairment recorded at 8 h, when testing ended [15]. Somewhat better performance was observed with the short-acting induction agents available at that time, propanidid and Althesin, with simulated driving returning to normal within 8 h or less [15]. Of some concern was the observation that the patient's perception of their driving ability improved more rapidly than their actual performance [15]. Using similar methodology, it was concluded that driving should not be attempted for at least 7 h after a brief (< 5 min) period of anaesthesia induced and maintained with either halothane or enflurane [16]. Surprisingly little work has been done on driving skills after modern ‘day-case’ anaesthetics. Psychomotor function relevant to driving was found to be minimally impaired at a blood propofol concentration of 0.2 µg.ml−1[17]. This degree of impairment was found to be comparable to that produced by a blood alcohol concentration of 20 mg.100 ml−1[17, 18], the maximum legal limit for driving in Sweden and one-quarter of the current UK driving limit. After a 1-h anaesthetic maintained at a constant propofol target of 6 µg.ml−1, the blood propofol concentration should have declined to 0.2 µg.ml−1 within less than 4 h (simulation based on current Diprifusor®). No similar work appears to have been done with modern inhaled anaesthetics. Even more rapid recovery would be expected after sedation, a technique used in 36% of the patients contacted by Correa et al. Sedative doses of fentanyl, midazolam or propofol, given alone or in combination, produced similar depressant effects to moderate amounts of alcohol [19], although impairment was no longer evident within 2–3 h. Simulated driving was unaffected within minutes of receiving Entonox for analgesia during sigmoidoscopy [20]. Should advice on driving be modified? The recent data would suggest that this is a possibility, especially after sedation, but the available evidence is severely limited. There is a big step between simulated and actual driving. In addition, most studies have not accurately reflected typical anaesthetic practice, neither have they taken into account the effects of analgesics, sleep deprivation due to pre-operative anxiety and the discomfort of the operation. Rational advice on the hazardous effects of other depressant drugs has been obtained from accident statistics [21]. There are few reported accidents from patients driving after anaesthesia, but the numbers so doing are relatively small and therefore unreliable. Unless more patients disregard current guidelines, we cannot be sure that this is a safe practice, yet we cannot risk altering our advice unless we know the alternative to be safe. Perhaps we should be grateful that so many patients comply with the somewhat arbitrary guidelines they are currently issued.