Impact of common surgical and anaesthetic strategies for pain control after elective gastrointestinal surgery: cohort study

Introduction: Poor pain control after surgery is associated with chronic pain and opioid dependence. This study aimed to determine the incidence of patient reported pain control failure following gastrointestinal surgery, and to evaluate the impact of common preventative surgical and anaesthetic strategies. Methods: Data were extracted from an electronic health record that linked real-time, ward-based pain scores with prescribing data. Adults undergoing major elective gastrointestinal surgery in 2011-18 were included. The primary endpoint was early pain control failure (≥1 instances of moderate or severe pain on postoperative days 0-2). Secondary outcomes were late (postoperative days 3-5) and persistent (both early and late) pain control failure. Results: Of 2238 patients, half underwent planned open surgery (50.3%, 1126/2238). Patient controlled analgesia (PCA) was initially used in 49.7% (1113/2238) and epidural in 35.0% (784/2238). Early (54%, 1211/2238), late (33.7%, 755/2238), and persistent (24.9%, 557/2238) pain control failures occurred frequently. In multivariable analyses, minimally invasive surgery was associated with fewer early, late, and persistent pain control failures than open surgery. There was no association between initial epidural analgesia and early or persistent pain control failure, but there was an association with increased late failure (OR 1.37, 95% CI 1.08-1.73, p=0.009). Of patients with initial epidural analgesia


Introduction
The most common complication experienced by the 260 million people 1 who undergo surgery worldwide each year is postoperative pain 2,3 .Modern practice has shifted from dismissing postoperative pain as an inevitable sequelae of surgery to recognising that perioperative teams should diligently manage acute pain 4 .However, the James Lind Alliance has identified a significant unmet need for research to inform strategies to reduce postoperative pain 5 .A key challenge in postoperative pain control is balancing the risks of undertreatment against the potential complications of excess use of analgesics, particularly opiates 6 .Undertreatment of acute pain is associated with both delayed postoperative recovery and risk of chronic postsurgical pain (CPSP) 7 .In the short-term, both poor pain control and excessive opiate use can contribute to complications including pneumonia, ileus, delirium, and immobility leading to venous thromboembolism 3,8,9 .In the longer-term, undertreatment of acute pain and excessive inhospital opiates are associated with chronic opiate use, contributing to the ongoing opioid crisis 7,[10][11][12] .Surgical strategies to reduce pain include minimally invasive surgery, whilst anaesthetic strategies include patient controlled analgesia (PCA), and local and regional analgesia techniques such as epidural analgesia.A large proportion of research and innovation in acute pain is in intensive care unit patients or is based in the USA.There is little high-quality data on pain control for gastrointestinal surgery patients who are cared for on general postoperative wards.The little data that does exist suggests that pain control is suboptimal in these patients 13,14 .
The Queen Elizabeth Hospital Birmingham's electronic health record captures patient pain scores at the point of care and is paired with electronic prescribing.The aim of this study was to use this quality, detailed data to determine the incidence of poor pain control following gastrointestinal surgery, and to evaluate the impact of key surgical and anaesthetic strategies that aim to improve pain control.

Data source
In 2005 the Prescribing Information and Communications System (PICS) electronic health record was introduced at the Queen Elizabeth Hospital Birmingham with builtin ward-based drug prescribing functions.During drug rounds nursing staff use PICS to record the time that drugs are administered to patients.Since June 2011, nursing staff have also used PICS to record patient reported pain scores at the point of care.Initially, pain scores were recorded using a 4-point verbal rating scale (0 -no pain at rest or on movement; 1 -no pain at rest, mild pain on movement; 2 -mild pain at rest, moderate pain on movement; 3 -continuous pain at rest, severe pain on movement).A decision was made by the hospital to switch to an 11-point verbal rating scale (rated 0 to 10) in December 2015.Hospital guidelines require pain scores to be recorded alongside each pulse or blood pressure measurement, as well as prior to each administration of analgesia.Consequently, in a postoperative ward setting pain should be measured a minimum of four times daily.The highest pain score for each postoperative day was extracted from the database.

Inclusion criteria
The PICS system was used to identify all adult patients (age 16 years and above) who underwent elective small bowel or colorectal resection, reversal of stoma, or stoma formation from 1 October 2011 to 10 November 2018.To ensure that primary endpoint data was available, only patients with a hospital length of stay of two days or longer were included.Patients were excluded if they were transferred directly from the operating theatre to the intensive care unit, as pain management protocols and nursing staff levels for patients on the intensive care unit are substantially different to those for wardbased patients.The day of surgery was defined as postoperative day zero.

Analgesic interventions
Unless contraindicated, all patients were prescribed regular paracetamol and regular non-steroidal antiinflammatory drugs (NSAID), such as ibuprofen 400mg three times daily.In addition, patients received one of three main analgesia strategies: • Epidural: typically 0.125% l-Bupivacaine and 2mcg/ ml Fentanyl in a 500ml bag run at 1-15ml/hour, with rate titrated for optimal pain control under the supervision of the acute pain service.The acute pain service was delivered by acute pain specialist nurses and consultant anaesthetists with an interest in pain management.If following titration it was not possible to ensure good pain control with epidural, patients were converted to PCA.
• Other: regular weak opiates (30-60mg codeine phosphate four times daily or 50-100mg tramadol four times daily), with Oramorph as required for breakthrough pain.For escalation, patients were prescribed strong opiates (e.g.intravenous morphine sulfate) as required.
Based on the drugs administered postoperative days 0-5, patients were classified as having received initial epidural, PCA, or other analgesia strategy.Secondary insertion of epidural (i.e. after the patient had left the operating theatre) was very rare.Therefore, if a patient received both epidural and PCA within that period, they were recorded has having initially received an epidural strategy, with subsequent conversion to PCA.

Primary outcome measure
The primary outcome measure was early pain control failure.Each patients' highest recorded pain score was extracted from PICS for each of the first five postoperative days.On the 4-point VRS, pain control failure was defined as at least one patient recorded score of 2 or greater, and on the 11-point VRS it was defined as a score of 5 or greater.These cut-offs were based on Royal College of Anaesthetists guidance 15 .
Early pain control failure was defined as at least one episode of moderate or severe pain on postoperative days 0-2.Late pain control failure was similarly defined based on pain measurements recorded on postoperative days 3-5.Persistent pain control failure was defined as experiencing both early and late pain control failure.

Explanatory variables
Use of preoperative chronic pain medication (i.e.gabapentin, pregabalin, amitryptiline, morphine sulfate slow release, buprenorphine patches, fentanyl patches, oxycontin) was extracted from the PICS system.Postoperative use of weak (codeine, dihydrocodeine, tramadol) and strong opiates (tramadol, oxynorm, oramorph, morphine injection) on postoperative days 0-5 was ascertained from PICS.Patients were recorded as having received no opiates, weak opiates only, or strong opiates.Electronic operation notes were reviewed to ascertain the site of operation (small bowel, colon, rectum), whether there was a resection of bowel, and the operative approach.Approach was classified as planned open surgery, completed minimally invasive (laparoscopic or robotic) surgery, or minimally invasive converted to open surgery.

Statistical analysis
Testing between categorical demographics and outcome groups was performed with the Chi-squared test.

Study approval
This study was registered as clinical audit at University Hospitals Birmingham NHS Foundation Trust (CARM-11953).

Analgesia strategies
The initial analgesia strategy was most frequently PCA (49.7%, 1113/2238).Of the 784 patients who initially received an epidural, 308 were subsequently converted to a PCA (39.3%, 308/784, Supplementary Table 1).Epidurals were more frequently used than PCA in patients who underwent planned open surgery (Supplementary Figure 3) and colorectal resection (Table 1 2).

Predictors of pain control failure
In multivariable analyses, female sex, pre-existing use of chronic pain medications, and rectal resection were found to be independent predictors of early pain control failure, whereas increasing age and minimally invasive surgery were associated with reduced risk (Table 3).An initial epidural analgesia strategy was not associated with early pain control failure (OR 1.09, 95% CI 0.88-1.36,p=0.423).
Similar patient and procedure factors were found to be associated with late and persistent pain control failures (Tables 4 and 5).In both analyses, female sex and pre-existing use of chronic pain were independently associated with increased pain control failure, and increasing age and minimally invasive surgery were independently associated with reduced pain control failure.

Discussion
Pain after major elective gastrointestinal surgery was common and not adequately controlled with current anaesthetic strategies, creating a cycle of opiate use.The only strategy that was associated with reduced pain was minimally invasive surgery, although its implementation was limited to 40-60% of cases, both within this study and across Europe 16 .Early pain failure was found to be associated with a 3.6 times greater odds of late pain failure, indicating that salvage is challenging once pain occurs.Given that revolutionary new pain medications are unlikely to come to the market in the near future, greater adoption of minimally invasive (laparoscopic, robotic, trans-orifice) surgical strategies is likely to offer the greatest and earliest benefit to patients by preventing severe postoperative pain.Treatments aimed at down-staging disease will support uptake of minimally invasive techniques by permitting smaller operations.
Epidural analgesia was found to have no clear advantage over PCA, even after adjusting for procedure type.
The shortcomings of epidurals were evident, as around 40% of patients were subsequently switched PCA.This is consistent with small randomised trials in laparoscopic surgery, that found epidurals to have limited or no superiority for pain control compared to PCA [17][18][19] .The disadvantages of epidurals include restricting patients' movement, slower recovery, and complications 20 .This pragmatic observational study tested real-world application of epidurals and PCA (stage 4 IDEAL study 21 ) over a seven-year period, with patients treated by a large number of anaesthetists.PCA may be preferred even in major open surgery, as it is less time-consuming than epidural catheter placement and avoids rare but catastrophic complications of epidural haematoma and infection.
This large cohort study benefited from being based on electronic health record data, with patient reported pain measurements recorded at the point of care by nursing staff trained in pain assessment.Although the frequency and timing of pain outcome assessment was not strictly protocolised, it is likely that by measuring pain levels Weak opiate 164 (  prior to administration of analgesia, peak pain levels were reliably recorded.The visual analogue scale used was a single dimension instrument and did not capture more global patient satisfaction or quality of life.A pain related VAS is frequently used for pain assessment in clinical studies as it offers high inter-rater correlation and validity, but it remains a uni-dimensional, subjective instrument 22,23 .Although this study was based on data from a single centre, its generalisability is increased by reflecting the practice of a large number of trainee and consultant anaesthetists and surgeons who participated in the management of patients at this large tertiary centre over a 7-year period. Opioid prescribing has significantly increased in UK general practice in the past two decades 24 and there is an increasing focus on the relationship between postoperative pain management and chronic opiate usage10, 11.This study was not able to address chronic opiate usage, as the PICS database has no access to community opiate prescribing and usage data.We did not capture data on measures including wound based local anaesthetic infusion catheters or transversus abdominis plane (TAP) blocks.Implementation of these multimodal interventions is highly variable 25 , and they were not in common practice within the hospital during the study period.Most studies testing these interventions have been small, single practitioner trials, providing little generalisable evidence [26][27][28][29] .
The study was conducted during a period of integration of minimally invasive surgical techniques in to routine practice.Greater adoption of these strategies may offer the best means of safely improving postoperative pain control.Future research should focus on strategies to promote its widespread adoption.This may include testing treatments aimed at down-staging disease, for example testing local excision of rectal cancer, or targeted chemotherapy down-staging of colonic cancer.Increasing the dissemination of laparoscopy, introducing cheaper robotics to reduce training times, and technological advancements that support the development of trans-orifice techniques in specialist centres can ensure that less invasive surgery is available for more patients in the future.
Early pain control failureOver half the patients experienced early pain control failure (54.1%, 1211/2238).Early failures occurred more frequently in patients with an initial epidural analgesia strategy than those with an initial PCA strategy (58.3% versus 53.2% respectively).Early pain control failure was associated with strong opiate use on univariable analysis (

Table 1 : Patient demographics stratified by initial analgesia strategy
*formation/reversal †Analgesia strategies not including an epidural or PCA MIS: minimally invasive surgery; PCA: patient controlled analgesia; SD: standard deviation

Table 3 : Univariable and multivariable models for early pain control failure
*formation/reversal †Analgesia strategies not including an epidural or PCA MIS: minimally invasive surgery; PCA: patient controlled analgesia; SD: standard deviation

Table 4 : Univariable and multivariable models for late pain control failure
*formation/reversal †Analgesia strategies not including an epidural or PCA CI: confidence interval; MIS: minimally invasive surgery; OR: odds ratio PCA: patient controlled analgesia

Table 5 : Univariable and multivariable models for persistent pain control failure
*formation/reversal †Analgesia strategies not including an epidural or PCA CI: confidence interval; MIS: minimally invasive surgery; OR: odds ratio PCA: patient controlled analgesia