Guidelines for Perioperative Care for Emergency Laparotomy Enhanced Recovery After Surgery (ERAS) Society Recommendations: Part 1—Preoperative: Diagnosis, Rapid Assessment and Optimization

ERAS protocols are designed to minimize the physiological impact and stress response of the surgical insult. For patients who require emergency laparotomy the insult and physiological derangement driven by inflammation, surgical stress and decompensation, are already occurring prior to the surgery. Resuscitation must go hand in hand with diagnostic interventions and preparation for surgery. Optimization consists of two parts: (1) patient optimization and (2) system optimization regarding availability of facilities and efficient care pathways [45]. Physiological derangements at presentation include a marked stress response, gut dysfunction, insulin resistance, fluid shifts, SIRS and sepsis with varying degrees of organ dysfunction [12, 30]. Emergency surgical patients may be hypovolemic with a potential critical impact on renal function and circulation. These derangements require early recognition and management with a sense of urgency. There is little evidence supporting delay for optimization prior to surgery in this patient category [1, 3, 17, 29, 46]. Some cohort studies have used standardized perioperative care protocols including screening and management of sepsis in line with the Surviving Sepsis Campaign [46] such as initial circulatory and respiratory stabilization, early goal-directed fluid therapy, thorough and invasive monitoring of vital parameters, and minimization of surgical delay. These studies have shown a reduction in mortality [1, 3, 6, 29]. Preoperative goal-directed fluid therapy was used in two of the multimodal cohort studies which showed significant reduction in mortality [3, 29]. Another small study used a goal-directed fluid optimization protocol in the preoperative holding room in patients with perforated peptic ulcer and showed reduced length of stay and mortality compared with a usual management control group [47]. Lactate guided resuscitation of patients with septic shock has been shown to reduce mortality [46] and may be beneficial in patients undergoing emergency laparotomy. The issue is not whether to delay for optimization but rather that staff competent in the management of significant physiological derangement must be involved at the earliest possible stage.

The pathophysiological abnormalities vary depending on the patient’s underlying health and co-morbidities, metabolic and immune status [48], and the underlying disorder and the duration of injury before presentation [27, 30, 36]. Patients who undergo emergency laparotomy represent only a fraction of the total volume of emergency general surgical cases but constitute the overwhelming majority of cases resulting in mortality and morbidity [32]. Use of physiological track and trigger systems [49] such as Early Warning Scores have been found to be highly predictive for severity of outcome including ICU admission and mortality in emergency surgical patients [50]. Scoring systems found to be predictive in emergency surgical patients include the Acute Physiological and Chronic Health Evaluation (APACHE II), Physiological and Operative Severity Score for the enUmeration of Mortality (POSSUM), Portsmouth-POSSUM (P-POSSUM), Modified Early Warning Score (MEWS) and National Early Warning Score (NEWS) (UK) [51]. In particular deteriorating early warning scores, in comparison to stable or improving scores, are highly predictive of mortality [52].

Recommendations:

The presence of sepsis should be considered in all patients undergoing emergency surgery at presentation. One large prospective study found an incidence of over 20% of sepsis or septic shock in patients presenting to an emergency general surgical service [53]. Upregulation of the inflammatory response as occurs with SIRS and sepsis is a major contributor to death; one major study found an increased hazard ratio of death in emergency surgical patients of 1.9 for those with SIRS, and 6.7 for patients with septic shock [54]. A study of 360,000 general surgical patients from the NSQIP database found that the presence of any comorbidity increased the risk of sepsis and septic shock six-fold, and increased 30-day mortality 22-fold [55]. The three major risk factors for sepsis and septic shock were age > 60 years, comorbidity and emergency surgery. The authors commented that these patients would benefit from mandatory sepsis screening in order not to miss the window of early intervention in which the septic source must be eliminated, and physiologic derangements corrected. The presence of hypotension secondary to sepsis has a particularly poor outcome [55‐58], with one large study of perforated peptic ulcer showing a 6% increased odds of 90-day mortality on adjusted risk analysis per hour delay to surgery in patients with preoperative hypotension [57]. Clinicians should have a high index of suspicion of sepsis when assessing emergency surgery patients.

The Sepsis 3 guidelines [59] recommend the use of quick Sepsis-related Organ Failure Assessment (qSOFA) as a screening tool to identify patients who are at risk of developing sepsis and septic shock. A positive qSOFA score should prompt further investigation for organ dysfunction, to initiate or escalate therapy as appropriate, and to consider referral to critical care or increase the frequency of monitoring. The qSOFA score may have limitations for emergency surgery patients and the use of the EWS to identify deterioration due to sepsis outperformed SIRS and qSOFA score in one large study [51]. In a review of sepsis-screening tools for surgery, it was noted that signs of sepsis in surgical patients may be diffuse and there is no perfect screening tool—what is clear is that when screening tools are used there is increased recognition of sepsis [60].

Once sepsis is suspected clinically, validated management algorithms should be completed with a sense of urgency [46, 61, 62]. These algorithms all include the empiric administration of broad-spectrum antibiotics (after relevant cultures have been obtained when possible), and cardiovascular resuscitation using intravenous fluids titrated to clinical endpoints. Specific antibiotic choice should be guided by local protocols in line with antimicrobial stewardship. Further evaluation and escalation should also follow these algorithms (Fig. 1). The NELA 2019 report found that there was major room for improvement in speed and urgency of management with only 19% of patients with suspected sepsis receiving antibiotics within the first hour [17]. Studies have shown an association between early risk scoring, active management and a reduction in mortality [1, 17]. Blood lactate has been used as a marker of risk [63], and in monitoring of response to resuscitation in line with the Surviving Sepsis guidelines [46, 63].

Recommendations:

Delay to surgical intervention can be due to any or all of the following: delayed diagnosis, preoperative therapeutic interventions/optimization or logistics. However, delay for patients undergoing emergency laparotomy can lead to increased mortality. In a Danish National cohort of perforated peptic ulcers, there was a 2.4% decreased probability of survival for every hour delay from hospital admission [65]. A large UK study of perforated peptic ulcer found that for patients in shock there was an increase of 6% in risk-adjusted odds of death for every hour laparotomy was delayed after admission [57]. A small Japanese study of perforation with septic shock found no patients survived to 60 days if surgery was delayed more than 6 h [66]. In another Danish cohort of all high-risk emergency laparotomy patients including those with obstruction as well as perforation, no statistically significant adjusted association between delay and surgical mortality at 90 days was found [67]. In a study from the NELA database, specifically focusing on small bowel obstruction, increased mortality was found for patients delayed more than 72 h [68]. In a multi-center study of septic patients in German ICUs when source control was needed, a delay of 6 h or more was associated with increased mortality [69]. Multimodal perioperative cohort intervention studies in emergency laparotomy have all had a surgical delay target of < 6 h from time of decision to operate to surgery, and have been associated with improved outcome [1, 3, 6, 29]. However, none of these studies have analyzed the impact of delay on survival in patients with perforation versus patients with obstruction. While these two clinical entities comprise the majority of indications for emergency laparotomy, the initial pathophysiology may be quite different. Patients with perforations often present with sepsis and the data are very clear that there should be minimal delay to surgical intervention, while the perioperative pathophysiology of patients with obstruction is poorly researched and the impact of delay and resuscitation less clear [3, 57, 65, 66, 68, 70]. Standards for time to surgery have been set in the UK by the Royal College of Surgeons [70, 71] and monitored by NELA and have been shown to be achievable at the 80% level [17].

Recommendations:

Risk assessment has become an important tool to support clinical assessment in management of patients undergoing emergency laparotomy [70]. Risk scoring was promoted in the first Higher Risk Surgical patient guidelines in the UK [71], as many laparotomy patients were not receiving care appropriate to their risk, such as planned admission postoperatively to an ICU. Clinical teams, inexperienced in management of emergency laparotomy, underestimated the potential for poor outcome. A risk score facilitates communication amongst clinical teams about priorities and pathways including for patient transfer, prompts involvement of highly experienced staff and helps direct discussion with the patient and family. There are a number of surgical or disease-specific risk prediction tools [76‐80]. Some, such as the Portsmouth-POSSUM (P-POSSUM) score [77], were developed many years ago for retrospective comparison of observed and expected outcomes when all variables are known. There is some concern regarding over-interpretation of individual patient preoperative prediction when these scores are used prospectively, and some variables must be estimated. Risk prediction scores give a population risk based on a risk model, scores can over and under-estimate risk for individual patients and clinicians must be cautious in using these scores for prognosticating patient outcomes. An example is a patient with a perforated peptic ulcer, who is acutely unwell with markedly deranged physiology and a very high-risk score, but who may benefit from rapid relatively simple surgery.

The large number of patients in the NELA database and the American College of Surgeons NSQIP database have allowed development of specific risk tools for patients undergoing emergency laparotomy which more consistently predict the actual risk for high-risk patients [78, 79]. A recent review comparing the NELA tool with ACS-NSQIP, P-POSSUM and APACHE II found the NELA score showed the highest discrimination with an area under the curve (AUC) of 0.83, ACS-NSQIP had an AUC of 0.80 [81]. When a risk score was calculated retrospectively on patients in the NELA dataset who had not been risk scored preoperatively or at the end of surgery, those patients were less likely to have potentially protective perioperative interventions such as planned postoperative critical care admission, and had poorer outcomes than a risk matched cohort who had prospective risk scoring performed [17]. Having a risk score available preoperatively may facilitate communication and planning for patient care. An ethnographic study of surgical teams, particularly those who did not routinely manage emergency surgical patients, found that use of a score enhanced inter-professional communication and decision making [82].

Most risk prediction tools do not directly adjust for certain relatively uncommon co-morbidities and some specific acute abdominal pathologies (other than cancer) which probably impact significantly on the survivability of an emergency laparotomy. Examples include a patient with a severe neurological condition or a patient with bowel infarction. When applying a risk prediction tool to an individual patient it is prudent to consider whether there are additional risk-altering variables present that have not been taken into account in any calculation. Additionally, these models are derived from patients who underwent surgery and do not include those evaluated for surgery but declined secondary to prohibitive risk.

Recommendation:

Age and frailty are not the same. Frailty is defined as an age-related state of decreased physiologic reserve [83]. All the large studies show that age alone is significantly associated with poor outcomes for emergency laparotomy [14]. In NELA mortality for patients over 80 years old in particular, remains very high at 17% at 30 days and 22% at 90 days [17]. Clearly the risk for these patients is so high that if emergency surgery is to be performed meticulous delivery of all evidence-based pathway components is essential. An ERAS approach has been shown to reduce mortality in patients over the age of 70 [1, 6]. A systematic review found an ERAS approach to be beneficial for older patients undergoing emergency surgery [7]. Many of these patients will also be frail, resulting in a lack of resilience in the face of a physiological insult [84‐87], and a validated frailty assessment [70, 83, 88, 89] should be performed if possible acknowledging the limitations in the acute environment. Frail patients and those with cognitive impairment have a higher risk of mortality and morbidity which may not be captured by the commonly used surgical risk scores [70, 81, 90]. In a study of outcomes at 12 months in older patients after emergency laparotomy, the strongest predictors of mortality were frailty and increased American Society of Anesthesiologists (ASA) status [91]. Involvement of a physician specialized in the care of older adults to co-manage these patients, and/or the use of targeted interventions should occur as soon as possible and is associated with better outcomes. The strongest evidence for comprehensive geriatric assessment exists for patients with hip fractures [92‐95], but a recent paper shows postoperative geriatrician review was associated with reduced mortality in patients over the age of 65 years undergoing emergency laparotomy [26, 96]. Another recent study using a proactive approach for patients over 65 years presenting for emergency general surgery with integration of a geriatric assessment team, optimization of evidence-based elder-friendly practices, promotion of patient-oriented rehabilitation, and early discharge planning found a significant reduction in mortality, length of stay and discharge to a higher level of care [95]. Proactive management of these frail patients may also decrease the costs of care [97, 98]. At present the evidence indicates that most older emergency laparotomy patients are not reliably assessed for frailty nor co-managed with a care of the elderly team [17].

Compared with elective surgical patients undergoing a comparable intra-abdominal procedure, emergency patients are at increased risk of venous thromboembolism (VTE) [131, 132]. Patients should be assessed for risk with a validated tool at admission, and VTE prophylaxis (mechanical and/or pharmacologic) should be initiated as soon as possible even if surgery is planned [131, 133, 134]. If pharmaceutical prophylaxis is not an option, mechanical prophylaxis should be used. The patient should be reassessed at regular intervals pre and postoperatively [131].

Recommendation:

Patients with an acute abdomen often require strong analgesia. Pain should be assessed, and an appropriate intervention should be made using multimodal analgesic titration and by minimizing the amount of opioid used to achieve effective analgesia. Opioids increase the risk of a patient being over-sedated, hypo-ventilating and even aspirating, so appropriate monitoring should be performed. The addition of benzodiazepines or other sedative agents compound these risks and should be avoided, and can increase the risk of postoperative delirium in older patients [106]. The use of preoperative nerve blocks such as transversus abdominis plane (TAP) blocks prior to surgery do not address the peritoneal and visceral pain of an acute abdomen. Therefore, opioids are often necessary in addition to other multimodal agents. Non-steroidal anti-inflammatory drugs (NSAIDs) are best avoided, due to the high risk of acute kidney injury (AKI) in this population, until postoperatively when renal function has normalized [44, 135, 136].

Recommendations:

Perioperative glucose control is important to maintain body homeostasis and reduce downstream complications [137]. Hyperglycemia during and after surgery is common, occurring in 20–40% of elective surgical patients, particularly in diabetic patients or those with impaired glucose tolerance. Elevated blood glucose impairs neutrophil function and can cause overproduction of inflammatory mediators, reactive oxygen species and free fatty acids causing direct cellular damage, vascular endothelial changes and immune dysfunction. Substantial evidence indicates that correction of hyperglycemia with insulin administration reduces hospital complications and decreases mortality in general surgery patients [137‐139]. The stress response drives insulin resistance at a time when patients are likely to have poor oral calorie intake and omit their insulin or diabetic tablets for fear of hypoglycemia. ERAS pathways for elective surgery try to mitigate this insulin resistance by using components such as oral carbohydrate loading and regional anesthesia [44]. However, this is usually not feasible in emergency laparotomy patients.

A proactive approach to avoid both hyper and hypoglycemia should be taken in emergency laparotomy patients. Pre-operative blood glucose levels should be controlled in a similar range to critical care patients—between 144–180 mg/dL (8–10 mmol/L). Tight control of blood sugar has been relaxed since the first tight glycemic control ICU studies, the incidence of complications appears not to be significantly altered when allowing blood glucose to be 180 mg/dL (10 mmol/L) [140] but with a reduction in hypoglycemic neurological complications [141]. Most patients will be taking minimal calories by mouth and be receiving intravenous resuscitation and ongoing maintenance fluid with balanced crystalloid infusions which contain no glucose. Hypoglycemia should be treated with an intravenous 50% dextrose bolus and appropriate follow-up dextrose administration, again according to local hospital policy. It is unclear for hyperglycemia, whether a basal-bolus of insulin [138] is beneficial compared with a standard sliding scale in the emergency surgical population. Judicious use of a variable rate insulin infusion (sliding scale) regimen should be utilized according to local hospital policy and attention given to plasma potassium levels that can be lowered by insulin administration. The ongoing management of glucose control is outlined in Part 2 of this Guideline. An HbA1c taken on admission is useful in guiding whether a patient has long-term glycemic control issues and may aid decision making on clinical intervention.

Electrolyte disturbance is common in this group of patients due to high fluid shifts and external losses of body fluids. Hypokalemia, hypomagnesemia and hypophosphatemia are risk factors for cardiac dysrhythmias, particularly atrial fibrillation which is particularly common in this patient group due to age and preexisting atrial fibrillation, fluid shifts, electrolyte imbalance and septic shock needing vasopressor infusions. Attempts should be made to correct low potassium, phosphate and magnesium using intravenous repletion with appropriate monitoring and according to local policy to reduce the risk of atrial fibrillation [142].

Recommendations:

An oral carbohydrate drink given preoperatively is a recommendation in most other ERAS Society Guidelines to reduce dehydration and improve insulin sensitivity by giving a carbohydrate load 2–4 h prior to surgery. Emergency laparotomy patients are already under physiological stress and giving carbohydrates in this setting may elevate glucose levels further with no effect on insulin sensitivity. We could not identify any studies on the use or benefit of carbohydrate loading in emergency general surgery. The increased risk of gastric stasis, intra-abdominal pathology, preoperative use of opioids and generalized practice of using preoperative nasogastric tubes and avoiding oral intake prior to surgery meant we extrapolated evidence of potential harm and this group could not recommend the use of carbohydrate loading [44].

Recommendation:

Nasogastric tubes (NGTs) have been traditionally used in emergency surgery to reduce gastric distension and drain gastric contents. The use of nasogastric tubes in elective colorectal surgical patients is declining as the evidence base has shown an increase in complications such as respiratory infections and pharyngolaryngitis as well as patient discomfort and delay to feeding, [143, 144] with no change in morbidity or mortality. [145]. The use in the emergency setting is very different with a risk–benefit ratio depending on the clinical circumstances and cause of abdominal pathology and patient factors. Patients may have pathology causing gastric distension and high gastric fluid volumes, and decompression may be beneficial and reduce the risk of aspiration at induction of anesthesia. This risk benefit is different in the preoperative and postoperative setting. We therefore discuss the postoperative use and continuation in part 2 of this guideline.

Recommendations:

Patient education is a central pillar of elective enhanced recovery pathways, benefits include reduced pain and anxiety [146‐148]. There is less time for education or explanation of complex surgery in the emergency setting, although handouts can be given to patients and families to read before or after surgery. In very high-risk patients, surgery should not be undertaken without discussion about ceilings of care, even though this is challenging in the acute situation. Objective mortality scores can support conversations and should be used in combination with other assessments such as frailty scores [70, 149]. Shared decision making (SDM) and personalization of care is especially challenging in a patient in pain and with acute physiological disturbance from abdominal pathology [150]. Additionally, there is less time to develop the clinician-patient rapport/relationship which SDM relies upon [151]. Scenario planning and the use of decision aids may support SDM, helping to move detailed, complex conversations toward more rapidly understood and patient-centered information [152]. Using the BRAN methodology (‘benefits, risks, alternatives, do nothing) or best/worst case scenario may support a clear structured standardized approach [149, 150, 153]. Discussions should not just be about life or death, but loss of independence, quality of life and other important factors to patients, such as long-term stoma formation [154]. There is guidance available to surgical teams to help manage these situations [155] although the complexity and acuity of the situation means that eliciting patient preference and achieving goal- concordant care is challenging [156]. The goal should be to achieve active joint and realistic decision making before surgery. For all patients and families, satisfaction with emergency abdominal surgery is associated with receiving sufficient information about the risks and benefits of surgery [20, 149, 157], and it is feasible to collect patient reported outcome measures from patients who have undergone emergency laparotomy [158].

Recommendation: