This article is accompanied by the following Invited Commentary:
Raeder J, Kranke P, Smith I. Free pre-operative clear fluids before day-surgery?. Challenging the dogma. Eur J Anaesthesiol 2018; 35:334–336.
Guidance on pre-operative fluids fasting policy continues to evolve. Current European guidelines encourage the intake of oral fluids up to 2 h before the induction of general anaesthesia.1 The stomach empties half the volume of ingested clear fluids in approximately 12 min, less than the time taken from transfer of a patient to induction of anaesthesia.2 Given this knowledge, combined with the difficulties of providing patients a true 2 h cut-off time for clear fluids, Torbay Hospital Day Unit (Torbay and South Devon NHS Foundation Trust, Devon) changed its policy in October 2014 from patients not being permitted to drink for two pre-operative hours to patients being encouraged to drink unrestricted clear fluids before transfer to theatre.
In February and March 2016, we used a curtailed questionnaire from the first Sprint National Anaesthesia Project3 to compare the experience of 80 patients on the day surgery pathway with that of 80 patients anaesthetised in May 2014 who had participated in Torbay Hospital's contribution to the Sprint National Anaesthesia Project. The results encouraged us to conduct this larger study.
In Torbay hospital, information is routinely collected from all patients who have undergone a surgical procedure on the day surgery pathway. Multiple aspects of each patient's peri-operative pathway are recorded, including: date of surgery, age, sex, American Society of Anesthesiologists (ASA) physical status, surgical specialty, anaesthetic type (including use of total intravenous or volatile anaesthesia, supraglottic airway or tracheal tube, spontaneous or intermittent positive pressure ventilation), surgical procedure(s), anaesthetic start time and time on entry to the postanaesthesia recovery ward. The duration of a procedure was timed from the start of anaesthesia until arrival in the recovery ward. Discharge or unexpected admission information is also collected for every patient. We do not collect data on BMI, smoking status, previous postoperative nausea and vomiting, nor susceptibility to motion sickness.
In addition, patient experience information is collated into this database. This involves a dedicated team of nurses carrying out a telephone follow-up service within approximately 24 h after discharge. Several domains of patient experience are assessed, including their general feeling (very good, good, reasonable, bad), experience of dizziness, nausea (mild, moderate, severe), drowsiness, pain (mild, moderate, severe), pain site, bleeding, bleeding site, and vomiting (once, two to three times, more than three times). Patients are also asked about their level of satisfaction (very good, good, reasonable, bad), and are given an opportunity to feedback general comments.
Patients were made aware of the fluids policy at pre-operative assessment, and there were signs next to a communal water dispenser in the waiting area encouraging patients to drink clear oral fluids. With regard to blinding, although all staff and patients were made aware of the change in policy, it was never intended as a means to reduce postoperative nausea and vomiting, nor promoted as such.
In April 2016, we analysed data collected for all day case pathway patients from November 2013 to February 2016 inclusive, following sedation, general anaesthesia, regional anaesthesia or their combination. The only exclusions were patients undergoing a termination of pregnancy, or patients undergoing community dental procedures, from whom patient experience data are not collected. This period included at least 12 months either side of the change in drinking policy, accommodating possible seasonal variations in patient characteristics and rates of postoperative nausea or vomiting. We did not perform a sample size calculation. When reviewing nausea and vomiting separately, we did not analyse patients with missing outcome data. We did not analyse the volumes or timing of ingested fluids. We did not seek ethics approval for an analysis of this anonymised, routinely collected data.
The primary outcomes were the rates of postoperative nausea or vomiting. Secondary outcomes were patient satisfaction, and the severity of nausea or vomiting (‘none or mild’ vs. ‘moderate or severe’).
We used the ‘cpt.meanvar’ function (Poisson distribution) in the R package [‘changepoint’ package (Killick [Lancaster, UK], Eckley, and Haynes 2014), available for R (http://www.R-project.org/) from the Comprehensive R Archive Network (CRAN) at http://CRAN.R-project.org/package=changepoint] ‘changepoint’ to determine whether there was a ‘point change’ or inflection in the rate of nausea or vomiting between November 2013 and February 2016 (including any crossover period).4 We compared such a ‘point change’ with the timing of our change in policy crossover period. The change in policy was first proposed at the end of October 2014, with official ratification in February 2015. Therefore, in directly comparing the patients subject to the liberal fluid policy with those subject to the restrictive fluid policy, we excluded a 4-month crossover period (November 2014 to February 2015 inclusive). We compared the total rates of nausea and vomiting either side of the change in policy with Pearson's χ2 test. We also adjusted analyses for patient age, sex, ASA physical status, mode of anaesthesia and airway management, surgical specialty and duration of surgery (logged anaesthetic start time to arrival in recovery). We further compared the rate of nausea and vomiting in the first 6 months of each period with the rate in the final 6 months of each period to identify continuous trends in each outcome, independent of our change in fluids policy. Rates were compared using Pearson's χ2 test.
The proportion of patients who provided patient experience information was 78% before the change in fluids policy and 74% after the change. When comparing baseline subgroups of nonrespondents, and taking account of any change in proportion of subgroups in respondents before and after the change in policy (Table 1), there was no significant difference in their proportion. We analysed rates of nausea from 11 490 patients and rates of vomiting from 11 474 patients. A ‘changepoint’ reduction in nausea was centred on 5 February 2015 (Fig. 1) and a changepoint reduction in vomiting centred on 7 January 2015 (Fig. 2), both of which were within our predefined crossover period.
After excluding patients from the 4-month crossover period, we analysed rates of postoperative nausea for 9916 patients, from a total of 9924 patients (data missing for eight patients) (Fig. 3): 5192 who could not drink within 2 h of surgery from before the change in fluid policy and 4724 encouraged to drink clear fluids up to the time of transfer to theatre. In these two patient populations, the incidence of nausea reduced from 270/5192 (5.2%) to 179/4724 (3.8%), a relative rate of 0.73 [95% confidence interval (CI), 0.61 to 0.88], P = 0.00074. The rates of ‘moderate or severe’ nausea were 68/5192 (1.3%) and 31/4724 (0.7%), a relative rate of 0.50 (95% CI, 0.33 to 0.76), P = 0.001. Figure 4 shows the rates of vomiting per month. Vomiting incidence was 146/5186 (2.8%) when fluids were restricted and 104/4716 (2.2%) when fluids were unrestricted, a relative rate of 0.78 (0.61 to 1.00), P = 0.053). The rates after excluding patients who had neither sedation nor general anaesthesia were nausea, 263/5088 (5.2%) vs. 175/4646 (3.8%), P = 0.0009; vomiting 143/5082 (2.8%) vs. 101/4638 (2.2%), P = 0.045.
Baseline data are presented in Table 1. On multivariate regression analysis, patients after colorectal surgery and women, particularly after gynaecology surgery, were more likely to be nauseated. On multivariate regression analysis, patients were more likely to vomit if they had longer anaesthesia or were undergoing urological, orthopaedic or colorectal procedures, or were women, particularly after gynaecology surgery. Multivariate regression analysis demonstrated a statistically significant reduction in both postoperative nausea and vomiting when fluids were unrestricted: nausea had an adjusted odds ratio (OR) (95% CI) of 0.26 (0.11 to 0.60), P = 0.0016, whereas vomiting had an adjusted OR of 0.52 (0.28 to 0.97), P = 0.039.
There were no differences in rates of nausea and vomiting between the first and second 6 months of each 12-month period. When fluids were restricted the rates of nausea were 141/2670 (5.3%) vs. 129/2522 (5.1%), P = 0.79 and the rates of vomiting were 66/2667 (2.5%) vs. 80/2520 (3.2%), P = 0.13. When fluids were unrestricted, the rates of nausea were 83/2381 (3.5%) vs. 96/2343 (4.1%), P = 0.27 and the rates of vomiting were 53/2379 (2.2%) vs. 51/2338 (2.2%), P = 0.91.
Patients categorised their experience as ‘Very good’ after 2331/5188 (44.9%) operations when drinking was restricted and after 2269/4697 (48.3%) operations when clear fluids were unrestricted, a relative rate (95% CI) of 1.08 (1.03 to 1.12), P = 0.0008.
We checked all reasons for unexpected admissions following day surgery. There were 116 unexpected admissions in total: 31 were for surgical complications or extended duration of surgery; three were due to syncope; seven were planned inpatient stays; 14 were due to various medical complications (none of which were respiratory related); seven were due to social reasons; 26 were due to pain; five were due to postoperative nausea and vomiting; 10 were due to bleeding; seven patients were unable to mobilise; and six patients were unable to urinate. There were no adverse events of pulmonary aspiration of gastric contents requiring patient admission.
Postoperative nausea and vomiting is a major cause of patient dissatisfaction after anaesthesia. Its causes are multifactorial and the subject of widespread review.5 In this large retrospective analysis of day surgery pathway patients, we identified distinct inflection points in mean rates of nausea and vomiting, both of which fell within the crossover period of our change in policy. The group who were subject to a policy of unrestricted clear fluids preoperatively experienced a relative reduction in the rates of postoperative nausea and vomiting compared with those limited to fluids up to 2 h preoperatively. Similar rates of nausea and vomiting between the first and second 6 months of each 12-month period suggest the lack of a continuous decline in such outcomes, independent of our change of policy. The combination of these analyses suggests that the liberal consumption of clear fluids before the induction of scheduled day case anaesthesia reduced the rates of postoperative nausea and vomiting.
We recognise that such an absolute reduction in postoperative nausea and vomiting is small, and that greater concerns may exist about the possibility of complications, such as pulmonary aspiration of gastric contents. As of October 2017, an additional 10 487 patients have undergone either a general anaesthetic, regional anaesthesia, sedation or their combination on our day surgery pathway. This represents a total of 15 214 patients since the change in policy. During 2017, one patient breathing via a supraglottic airway device had an aspiration event necessitating tracheal intubation and admission. This took place during an anterior cruciate ligament repair. The patient made a full, uneventful recovery. On review of the case, the patient had a BMI of 35 kg m−2. There was also evidence in his history that he may have had gastro-oesophageal reflux. Such an incidence is less than ‘normally’ reported rates of aspiration, which occurs in approximately 1 in 8000 elective general anaesthetics of ASA I and II patients6 (representing 95% of our day surgery pathway patients).
Our change in policy and guidance for patients was the main intervention in this study. Although we did not measure volume and timing of fluids for patients, we feel this is reflective of everyday practice. The option to drink clear fluids before surgery is not prescriptive and measuring and timing such fluids is not pragmatic in routine practice.
We demonstrated an improvement in patient satisfaction following the change in policy. Although this could be due to many factors, it is possible that the reduction in postoperative nausea and vomiting rates contributed to this outcome.
Variation in preoperative fluids intake has been the focus of a Cochrane review conducted in 2003.7 This review investigated the role of pre-operative fasting in the prevention of peri-operative complications. Several studies failed to demonstrate a difference in postoperative nausea and vomiting rates based on a shortened fluid fast vs. a standard fast.8–10 However, these studies only permitted pre-operative oral fluids up until 90 to 180 min pre-operatively. In addition, in view of the low numbers of patients included in each study, their power was too low to deliver conclusive results.
We feel that, given the large sample size involved and the collection of data across all months of the year for each period, we reduced the effects of unmeasured confounding variables as much as possible. Although comorbidities were unmeasured, ASA physical status could be viewed as a crude measure of this variable. This was adjusted within our multivariate regression analysis. A high proportion of our patients routinely receive an analgesic as premedication. This is generally paracetamol with the addition of ibuprofen if indicated. These drugs are preprinted on each patient's anaesthetic pre-assessment chart as a prompt, a practice that did not change during the study period. Based on a similar spend on antiemetic agents from our pharmacy department, we believe the use of antiemetic agents remained consistent throughout the study period. We were only able to collect data pertaining to anaesthetic type, rather than specific opioid use. This revealed a 78% use of total intravenous anaesthesia in both patient cohorts. The majority of such anaesthetics use propofol and alfentanil target controlled infusions, often obviating the need for additional opioids. We did not have data relating to smoking status, nor patient history of previous postoperative nausea and vomiting or motion sickness. However, a key component of our statistical analysis was a demonstration of a change point at a distinct point in a continuous series of patients that coincided with the crossover period of our change in fluids policy. The large numbers involved suggest that a distinct change point in other potential confounding variables (such as those described by Apfel et al.11) may have been less likely.
We hope that the open presentation of our data, including any weaknesses, may lead other centres to consider the potential benefits of unrestricted pre-operative clear fluids. To provide more conclusive evidence to this argument, further prospective studies would be of value. In the interim, however, we feel our findings may help to create a strong argument that a simple change in preoperative fluids policy not only has benefits for both patient satisfaction and theatre organisation, but also reduces the incidence of postoperative nausea and vomiting.
Acknowledgements relating to this article
Assistance with the study: we would like to thank Dr John Carlisle for his help and guidance with statistical analysis.
Financial support and sponsorship: none.
Conflicts of interest: none.
Presentation: preliminary audit presented at the British Association of Day Surgery Annual Scientific Meeting on 23 June 2016 in Manchester. Data for postoperative nausea and vomiting outcomes from the large retrospective analysis was included in an oral presentation at the British Association of Day Surgery Annual Scientific Meeting 22 to 23 June 2017 in Southport.
1. Smith I, Kranke P, Murat I, et al. Perioperative fasting in adults and children: guidelines from the European Society of Anaesthesiology. Eur J Anaesthesiol
2. Jolliffe DM. Practical gastric physiology. Continuing education in Anaesthesia. Crit Care Pain
3. Walker EMK, Bell M, Cook TM, et al. Patient reported outcome of adult perioperative anaesthesia in the United Kingdom: a cross-sectional observational study. Br J Anaesth
4. Killick R, Eckley IA. Changepoint: an R package for changepoint analysis. J Stat Softw
5. Pierre S, Whelan R. Nausea and vomiting after surgery. Continuing education in Anaesthesia. Crit Care Pain
6. Warner MA, Warner ME, Weber JG. Clinical significance of pulmonary aspiration during the perioperative period. Anesthesiology
7. Brady MC, Kinn S, Stuart P, et al. Preoperative fasting for adults to prevent perioperative complications. Cochrane Database Syst Rev
8. Gilbert SS, Easy WR, Fitch WW. The effect of preoperative oral fluids on morbidity following anaesthesia for minor surgery. Anaesthesia
9. Agarwal A, Chari P, Singh H. Fluid deprivation before operation. The effect of a small drink. Anaesthesia
10. Goodwin APL, Rowe WL, Ogg TW, et al. Oral fluids prior to day surgery. The effect of shortening the preoperative fluid fast on postoperative morbidity. Anaesthesia
11. Apfel CC, Heidrich FM, Jukar-Rao S, et al. Evidence-based analysis of risk factors for postoperative nausea and vomiting. Br J Anaesth