Customize your JAMA Network experience by selecting one or more topics from the list below.
Lassen K, Soop M, Nygren J, et al. Consensus Review of Optimal Perioperative Care in Colorectal Surgery: Enhanced Recovery After Surgery (ERAS) Group Recommendations. Arch Surg. 2009;144(10):961–969. doi:10.1001/archsurg.2009.170
To describe a consensus review of optimal perioperative care in colorectal surgery and to provide consensus recommendations for each item of an evidence-based protocol for optimal perioperative care.
For every item of the perioperative treatment pathway, available English-language literature has been examined.
Particular attention was paid to meta-analyses, randomized controlled trials, and systematic reviews.
A consensus recommendation for each protocol item was reached after critical appraisal of the literature by the group.
For most protocol items, recommendations are based on good-quality trials or meta-analyses of such trials.
The Enhanced Recovery After Surgery (ERAS) Group presents a comprehensive evidence-based consensus review of perioperative care for colorectal surgery. It is based on the evidence available for each element of the multimodal perioperative care pathway.
Fast-track or enhanced-recovery programs integrate a range of perioperative interventions proven to maintain physiological function and facilitate postoperative recovery, especially after elective colonic resections in dedicated centers.1-4 The Enhanced Recovery After Surgery (ERAS) Group has achieved similar results in general surgical departments using an evidence-based care platform.5,6
We present an updated and expanded consensus review of perioperative care for colorectal surgery based on the evidence available for each element of the multimodal pathway.
The MEDLINE database was searched up to December 31, 2007 (3 exceptions were made: 2 meta-analyses and an editorial published in 2008), and the ERAS protocol6 from 2005 was updated. Recommendations were evaluated according to the system developed by the Centre for Evidence Based Medicine, Oxford, England.7 Those based on at least 2 good-quality randomized controlled trials (RCTs) (in patients undergoing gastrointestinal surgery) or 1 meta-analysis of RCTs with homogeneity were designated as grade A. Other recommendations were designated as consensus recommendations based on the best available evidence. The evidence is presented in the text and the recommendations are summarized in the Table.
Explicit preoperative information can facilitate postoperative recovery and pain control, particularly in patients exhibiting denial and anxiety.8,9 A clear explanation of expectations during hospitalization facilitates adherence to the care pathway and allows early recovery and discharge.10,11 At this first encounter, the patient should also be given a clear role with specific tasks, including targets for postoperative food intake, oral nutritional supplements, and mobilization.12,13
Mechanical bowel preparation can cause dehydration and fluid and electrolyte abnormalities, particularly in elderly patients.14 Two recent large, multicenter RCTs15,16 confirm the conclusions of earlier meta-analyses17-19 that bowel preparation is not beneficial in elective colonic surgery, and 2 smaller recent RCTs suggest that it increases the risk for anastomotic leak.20,21 Bucher et al20 included only left-sided colonic resections and demonstrated increased morbidity after routine bowel cleansing. Bowel preparation may be necessary in selected patients who require intraoperative colonoscopy. For colonic surgery, data indicate that bowel preparation is stressful and prolongs postoperative ileus.22
A 2005 Cochrane analysis23 included 231 low anterior resections without finding an increased leak rate in those without bowel preparation. A recent RCT that included a substantial proportion of ultralow rectal anastomoses24 reported that bowel preparation protects against anastomotic leaks requiring reoperations. There was, however, increased cardiovascular mortality in the group receiving bowel preparation. Further trials are needed to establish the optimal routine for very low rectal resections. Nevertheless, logic dictates that the bowel distal to the stoma should be cleansed if a diverting stoma is constructed to protect the anastomosis.
Although fasting from midnight has been standard practice to avoid pulmonary aspiration in elective surgery, a review has found no evidence to support this.25 Equally, a Cochrane review26 of 22 RCTs in adult patients provides robust evidence that reducing the preoperative fasting period for clear fluids to 2 hours does not increase complications. National Anaesthesia Societies now recommend intake of clear fluids until 2 hours before induction of anesthesia as well as a 6-hour fast for solid food.27-30 Obese and even morbidly obese patients have the same gastric emptying characteristics as lean patients.31,32 Diabetic patients with neuropathy may have delayed gastric emptying, possibly increasing the risk of regurgitation and aspiration.33 Patients with uncomplicated type 2 diabetes mellitus can have normal gastric emptying, and a study of preoperative carbohydrate loading did not find increased aspiration rates in such patients.34
Having patients undergo surgery in a metabolically fed state can be achieved by provision of a clear carbohydrate-rich beverage before midnight and 2 to 3 hours before surgery. This reduces preoperative thirst, hunger, and anxiety26,35 and postoperative insulin resistance.36 Patients in a more anabolic state have less postoperative nitrogen and protein losses37,38 as well as better-maintained lean body mass39 and muscle strength.40 Data from RCTs indicate accelerated recovery and shorter hospital stay in patients receiving preoperative carbohydrate loading in colorectal surgery.41,42
Adverse effects from long-acting premedication such as opioids, long-acting sedatives, and hypnotics hamper recovery (eg, immediate ability to drink and mobilization after surgery), leading to prolonged length of stay.43 Short-acting anxiolytics do not prolong recovery or length of stay.44
Meta-analyses have shown subcutaneous low-dose unfractionated heparin regimens to be effective in reducing deep vein thrombosis, pulmonary embolism, and mortality in patients undergoing colorectal surgery.45-48 Meta-analyses comparing low-molecular-weight heparin (LMWH) with unfractionated heparin have shown no difference in efficacy47,48 or associated bleeding risks.49,50 The LMWH is preferable because of its once-daily dosage and a lower risk of heparin-induced thrombocytopenia.50-52
Although antiplatelet drugs and intravenous dextran are less effective for prophylaxis of deep vein thrombosis and in reducing mortality, they can be as effective for the prevention of pulmonary embolism.48,53 Their adverse effect profiles53,54 make them advisable only in high-risk patients when LMWH and unfractionated heparin are contraindicated.
The safety of continuing LMWH and continuous epidural analgesia is debatable. In the United States, higher doses of LMWH are used twice daily and may account for the greater numbers of epidural hematomas reported.55 Prophylactic doses of LMWH should be given no later than 12 hours prior to insertion and removal of an epidural catheter.56,57 Although concomitant use of nonsteroidal anti-inflammatory drugs and LMWH is considered safe, a potential link with epidural hematoma is debated. Care should be taken with other factors affecting coagulation, and alternative thromboprophylaxis (such as thromboembolism-deterrent stockings) should be used when appropriate.
The use of prophylactic antibiotics effective against both aerobes and anaerobes can minimize infectious complications in colorectal surgery,58 with the first dose being administered about 1 hour prior to skin incision.59 A single dose is as effective as multidose regimens, but further doses should be given in prolonged cases (>3 hours).58 The optimal combination of antibiotics is not established, but a second-generation cephalosporin and metronidazole are suggested. New generations of antibiotics should be reserved for infectious complications.
There is no evidence to direct the choice of the optimal anesthetic method for colorectal procedures. However, it is rational to use short-acting agents (propofol, remifentanil hydrochloride)60 instead of long-acting intravenous opioids (morphine sulfate, morphine hydrochloride, fentanyl citrate), thereby allowing proactive recovery to start soon after surgery. Short-acting inhalational anesthesia is a reasonable alternative to total intravenous anesthesia. There is no evidence that intraoperative epidural analgesia improves postoperative outcome in colorectal procedures, but its use reduces the dose of general anesthetic agents. For colonic surgery, the epidural catheter is best placed at the midthoracic level (T7/8) to achieve both analgesia and sympathetic blockade, preventing gut paralysis.61 If activated before commencement of surgery, it blocks stress hormone release and attenuates postoperative insulin resistance.62 The catheter is inserted in the awake patient to avoid neurological complications. Intraoperatively, the block can be maintained by continuous infusion of local anesthetic (eg, bupivacaine hydrochloride, 0.1%-0.25%, or ropivacaine hydrochloride, 0.2%) plus a low-dose opiate (eg, 2.0-μg/mL fentanyl citrate or 0.5- to 1.0-μg/mL sufentanil citrate) at 4 to 10 mL/h. Epidural opioids in small doses act synergistically with epidural local anesthetics in providing analgesia,63 without major systemic effects.64-66 Addition of epinephrine (1.5- to 2.0-μg/mL) to the thoracic epidural infusion improves analgesia.67-69
Patient experience suggests that postoperative nausea and vomiting can be more stressful than pain.70-73 Risk factors include being female and having nonsmoking status, history of motion sickness (or postoperative nausea and vomiting), and postoperative administration of opioids.74,75 Individuals at moderate risk (2 factors) should receive prophylaxis with dexamethasone sodium phosphate at induction or serotonin receptor antagonist at the end of surgery.76 High-risk individuals (3 factors) should receive general anesthesia with propofol and remifentanil as well as 4 to 8 mg of dexamethasone sodium phosphate at the beginning of surgery, supplemented with serotonin receptor antagonists or droperidol76 or with 25 to 50 mg of metoclopramide hydrochloride 30 to 60 minutes before the end of surgery.77
The most recent meta-analysis78 confirms that significant improvements in short-term outcomes are achievable by laparoscopy-assisted colonic resection as a single intervention. This was associated with significant reductions in short-term wound morbidity, time to first bowel movement, and discharge from the hospital.
The potential of combining laparoscopy and enhanced-recovery care has been evaluated in only 2 small trials randomizing patients to either laparoscopy-assisted or open surgery within an established enhanced-recovery protocol.79,80 In the setting of a long-established and efficient enhanced-recovery protocol, no further improvement in short-term outcome was seen by adding laparoscopy (median postoperative length of stay of 2 days in both groups).79 The second study had longer hospitalizations, and here a reduction in postoperative stay was seen in the laparoscopy-assisted group as compared with the group undergoing open surgery (3.5 vs 6 days, respectively).80 Further investigation will hopefully more clearly evaluate the full potential of combining laparoscopy and enhanced-recovery care.81
Some RCTs suggest that transverse or curved incisions cause less pain and pulmonary dysfunction than vertical incisions following abdominal procedures,82,83 while others have found no advantage of transverse incisions.84,85 A recent Cochrane review86 of RCTs comparing midline with transverse incisions for abdominal surgery confirms that although analgesic use and pulmonary compromise may be reduced with transverse or oblique incisions, complication rates and recovery times are the same as with midline incisions. Hence, while incision length affects patient recovery,87 the choice of incision for abdominal surgery still remains the preference of the surgeon.
A meta-analysis88 in 1995 showed that routine nasogastric decompression should be avoided after colorectal surgery since fever, atelectasis, and pneumonia are reduced in patients without a nasogastric tube. A recent Cochrane meta-analysis89 of 33 trials with more than 5000 patients confirmed this and also found earlier return of bowel function in patients when nasogastric decompression was avoided. Gastroesophageal reflux is increased during laparotomy if nasogastric tubes are inserted,90 and there is no rationale for routine insertion of a nasogastric tube during elective colorectal surgery, except to evacuate air that may have entered the stomach during ventilation by facial mask prior to endotracheal intubation. Nasogastric tubes placed during surgery should be removed before reversal of anesthesia.
Several RCTs have demonstrated that preservation of normothermia by using an upper-body forced-air heating cover reduces wound infections,91,92 cardiac complications,92-94 bleeding, and transfusion requirements.92,95 Extending systemic warming to 2 hours before and after surgery had additional benefits.96
It has been standard practice in recent years to infuse volumes of intravenous fluids substantially in excess of actual perioperative losses.97 Traditional perioperative intravenous fluid regimens in abdominal surgery can lead to patients receiving 3.5 to 7 L of fluid on the day of surgery and more than 3 L/d for the following 3 to 4 days, leading to a 3- to 6-kg weight gain.98,99 Such regimens can delay the return of normal gastrointestinal function,98 impair wound or anastomotic healing, and affect tissue oxygenation, leading to prolonged hospitalization.99,100 Several trials have compared restrictive and liberal fluid or sodium regimens.98-102 The results are not uniform and comparison is difficult as administered volumes and electrolytes in both arms differed substantially, reflecting nonuniform standard practice.
However, evidence does suggest that avoidance of overload and restricting fluid intake to that which will maintain balance, guided by body weight, may significantly reduce postoperative complications and shorten hospital stay and should therefore be recommended.98,100 The best way to limit postoperative intravenous fluid administration is to stop intravenous infusions and return to oral fluids early, which should be feasible on the first postoperative day.1 Patients with epidural anesthesia experiencing hypotension due to vasodilation and relative intravascular hypovolemia, which is traditionally treated with fluid loading, can be treated with the judicious use of a vasopressor.103
Intraoperative transesophageal Doppler monitoring helps titrate fluids in relation to cardiac output and may be useful in high-risk patients. Four RCTs104-107 and a meta-analysis108 with patients undergoing major bowel surgery found that when intraoperative fluid administration was guided by transesophageal Doppler monitoring, there was a better ejection fraction, better oxygenation, and fewer postoperative complications. Although patients in these trials were not treated according to enhanced-recovery protocols, it seems that transesophageal Doppler monitoring enables optimization of intravascular volume and tissue perfusion in major abdominal surgery. In low-risk patients undergoing surgery of moderate magnitude, flow-guided therapy may not be warranted. High-grade evidence regarding the optimal regimen in terms of timing, type of fluid, and risk stratification is currently lacking.
Meta-analyses109,110 have demonstrated that the use of drains after colonic surgery does not reduce the incidence or severity of anastomotic leaks or other complications. Drainage of the pelvic cavity for 24 hours following low anterior resection is supported by the Dutch total mesorectal excision trial,111 although this remains to be proven in RCTs specifically designed to answer this question.
A recent meta-analysis112 of RCTs concluded that suprapubic catheterization is more acceptable to patients and reduces morbidity compared with urethral catheterization. Most trials have been undertaken in patients requiring 4 to 7 days of urinary drainage. The risk of urinary retention after only 24 hours of catheterization is low after colonic resection above the peritoneal reflection during epidural analgesia.113 Therefore, the advantages of suprapubic over urethral catheterization are probably small for colonic surgery, while the benefits are significant for pelvic surgery with longer catheterization times.
Prevention of postoperative ileus, a major cause of delayed discharge after abdominal surgery, is a key objective of enhanced-recovery protocols. While no current prokinetic agent is effective in attenuating or treating postoperative ileus, several other interventions have been successful. Midthoracic epidural analgesia61 as compared with intravenous opioid analgesia is highly efficient at preventing postoperative ileus.65,114 Fluid overloading during101 and after98 surgery impairs gastrointestinal function and should be avoided. Oral magnesium oxide has been demonstrated to promote postoperative bowel function in a double-blinded RCT in abdominal hysterectomy115 and in reports from a well-established enhanced-recovery program in colonic resection.1,116 Laparoscopy-assisted colonic resection also leads to faster return of bowel function as well as resumption of an oral diet compared with open surgery.78 Oral alvimopan, a μ-opioid receptor antagonist approved for clinical use in postoperative ileus, accelerates gastrointestinal recovery and reduces the duration of hospitalization in patients undergoing colonic resection compared with postoperative intravenous opioid analgesia.117
Meta-analyses have shown that optimal analgesia is achieved by continuous epidural local anesthetic with or without opioids for 2 to 3 days postoperatively in both open64,114 and laparoscopic118 surgery. Analgesia based on intravenous opioids does not provide the same efficient analgesia114 and has fewer beneficial effects on surgical stress responses compared with epidural local anesthetic techniques. While it is possible to achieve almost the same pain scores with patient-controlled analgesia at rest compared with epidural analgesia, this is at the expense of patients remaining sedated and in bed. Some RCTs114,119 have demonstrated that continuous epidural local anesthetic techniques reduce pulmonary morbidity but not other types of morbidity, hospital stay, or convalescence.
There are some concerns about the risk of anastomotic complications after epidural analgesia for colonic resection.114,120,121 Perfusion of the splanchnic area after establishment of the epidural block is probably more closely associated with changes in mean arterial pressure than with changes in cardiac output.122 Therefore, vasopressors to maintain pressure should be considered. In the case of cardiac insufficiency, an adequate preload and positive inotropes are mandatory to improve colonic blood flow. Low-dose norepinephrine and dobutamine hydrochloride are probably not harmful for splanchnic perfusion.123-127 The unanswered questions are the acceptable range of blood pressure in individual patients and the duration for which vasopressors should be used.120
Avoidance of opioids and their adverse effects is the goal after removal of the epidural catheter, and nonsteroidal anti-inflammatory drugs have been shown to be opioid sparing128 and to provide efficient analgesia during this period.1,129 Nabumetone is a widely used nonsteroidal anti-inflammatory drug that does not affect bleeding time and may be a safer choice in patients with epidurals.130
The RCTs of early enteral or oral feeding vs “nil by mouth” conclude that there is no advantage of keeping patients fasted after elective gastrointestinal resection.131-133 Early feeding reduced both the risk of infection and the length of hospital stay and was not associated with an increased risk of anastomotic dehiscence. However, the risk of vomiting increased in patients fed early, and in the absence of multimodal anti-ileus therapy, early feeding was associated with bloating, impaired pulmonary function, and delayed mobilization.134,135
For malnourished patients, there is a clear advantage of prescribing postoperative oral nutritional supplements for 8 weeks in terms of recovery of nutritional status, protein economy, and quality of life.136 Positive clinical outcomes from oral nutrition supplements have also been documented in studies of patients undergoing elective surgery who are not screened for malnutrition.137,138 In enhanced-recovery programs, oral nutritional supplements have been used successfully on the day prior to operation and for at least the first 4 postoperative days to achieve recommended intakes of energy and protein.1,139,140 When used in combination, preoperative oral carbohydrate loading, epidural analgesia, and early enteral nutrition have been shown to result in nitrogen equilibrium without concomitant hyperglycemia.141
Bed rest not only increases insulin resistance and muscle loss but also decreases muscle strength, pulmonary function, and tissue oxygenation.142 Additionally, there is an increased risk of thromboembolism. Effective pain relief using ambulatory thoracic epidural analgesia is a key adjuvant measure to encourage postoperative mobilization. A prescheduled care plan should list daily goals for mobilization, and a patient diary for out-of-bed activities is helpful. It is essential that the patient is nursed in an environment that encourages early mobilization (food and television removed from the bedroom) and one that maintains the patient's independence (ordinary ward or level 1 facility). The aim is for patients to be out of bed for 2 hours on the day of surgery and for 6 hours per day until discharge. Abdominal drains and urinary catheters hinder mobilization and should be avoided whenever possible.
A systematic audit is mandatory to determine clinical outcome and to establish the successful implementation of the care protocol. Distinguishing between unsuccessful implementation and lack of desired effect from an implemented protocol is vital if results are short of desired quality standards. Comparison with other centers using similar protocols via identical tools of registration and identical definitions of key factors is needed.
This article outlines the recommendations of the ERAS Group for clinical perioperative care of patients undergoing elective colorectal surgery, based on the best available evidence. However, neither evidence nor protocol is sufficient to ensure evidence-based care. Evidence dictates care only to a very limited extent,143 and an evidence-based protocol alone is insufficient to ensure change.144 We echo the words of Urbach and Baxter: “the immediate challenge to improving the quality of surgical care is not discovering new knowledge, but rather how to integrate what we already know into practice.”145
Correspondence: Kristoffer Lassen, MD, PhD, Department of Gastrointestinal Surgery, University Hospital Northern Norway, 9038 Tromsø, Norway (firstname.lastname@example.org).
Accepted for Publication: October 21, 2008.
Author Contributions:Study concept and design: Lassen, Soop, Nygren, von Meyenfeldt, Fearon, Revhaug, Ljungqvist, Lobo, and Dejong. Acquisition of data: Lassen, Soop, Cox, Hendry, von Meyenfeldt, Norderval, and Dejong. Analysis and interpretation of data: Lassen, Nygren, Hendry, Spies, Fearon, Norderval, Ljungqvist, and Dejong. Drafting of the manuscript: Lassen, Soop, Nygren, Cox, Hendry, Fearon, Lobo, and Dejong. Critical revision of the manuscript for important intellectual content: Lassen, Soop, Spies, von Meyenfeldt, Fearon, Revhaug, Norderval, Ljungqvist, Lobo, and Dejong. Obtained funding: von Meyenfeldt, Revhaug, Ljungqvist, and Dejong. Administrative, technical, and material support: Lassen, Soop, Cox, Spies, von Meyenfeldt, Fearon, Revhaug, Ljungqvist, Lobo, and Dejong. Study supervision: Lassen, Soop, Nygren, Hendry, von Meyenfeldt, Revhaug, and Dejong.
Enhanced Recovery After Surgery (ERAS) Group Members: Kristoffer Lassen, MD, PhD, Arthur Revhaug, MD, PhD, Stig Norderval, MD, PhD, University Hospital Northern Norway, Tromsø, Norway; Mattias Soop, MD, PhD, University of Auckland, Grafton, Auckland, New Zealand; Jonas Nygren, MD, PhD, Jonathan Hausel, MD, Ersta Hospital, Stockholm, Sweden; P. Boris W. Cox, MD, Maarten F. von Meyenfeldt, MD, PhD, Cornelis H. C. Dejong, MD, PhD, José Maessen, BSc, Ronald M. van Dam, MD, Maastricht University Medical Centre, Maastricht, the Netherlands; Paul O. Hendry, MBChB, MRCS, Kenneth C. H. Fearon, MD, FRCS, Royal Infirmary of Edinburgh, Edinburgh, Scotland; Claudia Spies, MD, PhD, Charité Universitaetsmedizin Berlin, Berlin, Germany; Olle Ljungqvist, MD, PhD, Karolinska University Hospital Huddinge, Stockholm, Sweden; Dileep N. Lobo, DM, FRCS, Nottingham Digestive Diseases Centre Biomedical Research Unit, Nottingham University Hospitals, Queen's Medical Centre, Nottingham, England; Robin Kennedy, MD, St Mark's Hospital, London, England.
Financial Disclosure: Dr Ljungqvist is the owner of a patent for a preoperative carbohydrate-rich drink licensed to Danone/Nutricia, which produces and markets a drink based on this patent.
Funding/Support: This work was supported by Fresenius Kabi, which has been sponsoring the ERAS Group with an unrestricted grant since 2006.
Role of the Sponsor: Fresenius Kabi (or any other commercial company) has not participated in the research work, the discussions, the writing of the manuscript, or the decision to publish the work.
Create a personal account or sign in to: