Abdominal sepsis and abdominal compartment-syndrome
Emma Barrow and Iain D. Anderson
Abdominal sepsis accounts for many of the serious and all too often fatal emergency conditions that the general surgeon is called upon to treat. It may arise primarily from conditions as seemingly simple and routine as appendicitis, to more complicated and serious conditions such as diverticulitis and perforated peptic ulcer disease. Alternatively it may arise as a consequence of complications of surgery, particularly intestinal anastomotic leakage. While, with experience and a few basic principles, even complex cases of abdominal sepsis can be relatively straightforward to manage, not infrequently the diagnosis is obscured and delayed. As a consequence treatment is often complicated and prolonged. Some 80–90% of general surgical deaths follow emergency admission,1 and much of this morbidity and mortality expresses itself through the processes of abdominal sepsis. Delay to operation is the single most common reason for adverse outcomes and anastomotic leak the single commonest cited complication in fatal cases. These will be recurring themes in this chapter as even with optimal treatment of abdominal sepsis, multiple organ failure (MOF) may ensue. MOF is more likely and often fatal when treatment is slow or suboptimal. In the intensive care unit (ICU), abdominal sepsis constitutes a substantial proportion of the patients with MOF. Given its frequency and severity, a sound understanding of abdominal sepsis must be integral to every general surgeon's professional armamentarium.
This chapter will address the diagnosis and management of abdominal sepsis, including the complex patient in the ICU, where sequelae such as abdominal compartment syndrome, management of the open abdomen and intestinal fistulation can create particular difficulties. The reader is referred to Chapter 16 for a description of the intensive care management of the surgical patient, Chapters 6, Chapters 8, 9, 10 and 13 for more detail on specific causal conditions, and to Chapter 17 for surgical nutrition.
Pathophysiology of sepsis
The systemic inflammatory response syndrome (SIRS) is a clinically defined response (Box 18.1) to a variety of insults including trauma, burns, pancreatitis, tissue ischaemia and inflammatory bowel disease. When the cause of SIRS is a proven or suspected infection, it is termed sepsis. The normal response to infection serves to localise and control bacterial invasion. This occurs through the chemotaxis of neutrophils and macrophages, which in turn release inflammatory mediators. When this inflammatory response becomes generalised, sepsis results. This is characterised by systemic vasodilation and resultant hypotension, increased vascular permeability leading to fluid exudate, and microcirculatory dysfunction with decreased capillary flow. These factors ultimately result in tissue hypoxia. Once triggered, the downward spiral of severe sepsis is believed to be independent of the precipitating infectious insult.
Box 18.1 Sepsis definitions13
Systemic inflammatory response syndrome (SIRS)
SIRS is defined by the presence of two or more of the following clinical findings:
SIRS plus a documented or suspected infection
Sepsis plus clinical evidence of organ dysfunction:
Sepsis with acute circulatory failure, despite adequate volume resuscitation in the absence of other causes of hypotension
Multiple organ dysfunction syndrome (MODS)
The theory that sepsis is due to an exaggerated, uncontrolled inflammatory response has now been shown to be overly simplistic. There is no single mediator, system or pathway that drives the pathophysiology of SIRS and sepsis. The predominant theories can be summarised as follows:
The slippery slope of sepsis
SIRS and sepsis occur commonly amongst surgical patients on the ward, although they are usually resolved by appropriate treatment of the underlying problem. Sometimes the physiological derangement persists and, particularly when it does so beyond 48 hours, outcome is worsened as progression to organ dysfunction is much more likely. Organ dysfunction denotes severe sepsis and the next step is organ failure, which typically carries a mortality of the order of 40%. Mortality, in general, increases with the number of organ systems affected and with the severity of physiological disturbance at onset.8 While the onset of SIRS does not accurately predict development of sepsis or the multiple organ dysfunction syndrome (MODS), the progression from SIRS to severe sepsis is associated with increasing risk of multiple organ failure.9 Approximately 30% of septic patients develop at least one organ dysfunction.10 Therefore, timely recognition of SIRS, particularly if persistent, alerts the clinician to a potentially deteriorating situation at a time when prompt intervention may yet avert catastrophe. By the time the patient with abdominal sepsis has developed shock, the mortality increases from less than 10% to greater than 50%.11,12
The identification and active management of SIRS and early organ dysfunction is therefore an important first step as, once organ dysfunction is under way, the patient is on a slippery slope that can lead them rapidly downwards despite the best treatment. It is essential that the surgeon appreciates that these processes often start insidiously on the ward and that early detection is vital, as management is most successful at this stage. While there are objective criteria that define organ dysfunction, clinical findings are useful pointers. Hypoxia, oliguria, hypotension, deranged liver function tests or clotting, thrombocytopenia, acidosis and confusion are some of the plethora of signs that indicate that a severe systemic derangement is in process.
The importance of detecting the often subtle signs of abdominal sepsis at the earliest stage cannot be overemphasised and while the rate with which organ dysfunction develops in individual patients will vary, the requirement for rapid identification and treatment remains key.9
Treatment strategies in sepsis
SIRS, sepsis and their sequelae are recognised and defined according to a number of clinical criteria;13 these are summarised in Box 18.1. Whilst these specific criteria are not intended as a substitute for clinical acumen, their use facilitates early identification and appropriate treatment of sepsis by even the most junior member of the surgical team. SIRS and sepsis can be adequately managed on the ward, providing there is response to treatment. However, the deteriorating patient or those with severe sepsis are more appropriately transferred to a critical care environment, where invasive arterial and central venous pressure monitoring will guide resuscitation. The benefits of managing such high-risk surgical patients with early critical care input are well recognised.14,15
The Surviving Sepsis Campaign
In 2008, informed by the results of a number of clinical trials, an international campaign was launched,16 with the intention of improving outcomes in severe sepsis and septic shock by standardising care. The emphasis of the campaign was timely identification and treatment of patients with severe sepsis, using goal-directed strategies (the rationale for this is detailed inChapter 16). Evidence-based guidelines were published in 2004, split into ‘bundles’ of care to be accomplished within certain time frames (Box 18.2). A total of 165 sites participated in the campaign, submitting bundle compliance and outcome data on 15 022 patients with severe sepsis. Despite incomplete compliance, a significant reduction in unadjusted hospital mortality (37% to 31% over the 2-year study period) was identified in those centres participating in the campaign.17
Box 18.2 Surviving Sepsis Campaign bundles16
Sepsis resuscitation bundle
To be accomplished within the first 6 hours of identification of severe sepsis:
Sepsis management bundle
To be accomplished within the first 24 hours of identification of severe sepsis:
The goal-directed treatment bundles developed by the Surviving Sepsis Campaign are to be recommended as a standard of care. Their use in the timely identification and management of patients with severe sepsis has been shown to reduce mortality.17
Although effective management of patients with severe sepsis may entail complex investigations and procedures, the results of these manoeuvres are often suboptimal or even lethal without adequate prior resuscitation. A systematic approach such as that described in the Care of the Critically Ill Surgical Patient (CCrISP) course18 has much to recommend it, as it provides a common management structure for problems of any type or severity (Fig. 18.1). Having a structured approach in times of crisis facilitates speed and may also be important in reducing the likelihood of management errors. It certainly provides a common language and transparency that lets other health professionals understand interventions more easily. With complex abdominal sepsis, a team approach is required: firstly, because help will often be needed from radiology, anaesthesia and intensive care; and secondly, because the illness will often run a prolonged time course of days or weeks and hence many doctors will be involved.
FIGURE 18.1 The CCrISP system of assessment. Reproduced from Anderson ID. Assessing the critically ill surgical patient. In: Anderson ID (ed.) Care of the critically ill surgical patient. London: Arnold, 1999; pp. 7–15. © Hodder Arnold. Reproduced by permission of Hodder Education.
Patients with abdominal sepsis will present with some degree of instability and CCrISP advocates rapid immediate management following ABC principles of assessment with simultaneous correction of life-threatening conditions and initiation of high-flow oxygen therapy, fluid resuscitation and basic monitoring as required. Although some patients will deteriorate catastrophically and require immediate intensive care support, simple resuscitation will more commonly buy sufficient time for a more thorough full assessment to be carried out. This aims to determine the cause and severity of any problem and to exclude other conditions that would prove deleterious if left untreated. It also includes a thorough appraisal of the patient's notes and charts.
As the clinical manifestations of abdominal sepsis can be subtle and varied (Box 18.3), a high index of suspicion, combined with anticipation of potential complications, is essential. Complications can usually be anticipated from the surgical condition in question, any operation recently carried out and knowledge of comorbid conditions. Frequently, the range of possible diagnoses is large (Box 18.4) and the initial diagnostic net must be cast wide before drawing it in rapidly with the assistance of selective investigations. Reaching a provisional diagnosis and management plan rapidly is important as outcome worsens with delay and deterioration.
Box 18.3 General manifestations of abdominal sepsis in the ward or HDU patient
Pyrexia or hypothermia
Poor peripheral perfusion
Leucocytosis or neutropenia
Box 18.4 Some possible differential diagnoses in patients presenting with abdominal sepsis on the ward (this depends on presenting features)
Sepsis of other origin (urine, line, chest, etc.)
Cardiac (ischaemia, infarction, dysrhythmias, failure)
Cerebral (toxic confusion, ischaemia)
Pulmonary (atelectasis, collapse, infection, pulmonary embolism)
Other non-septic abdominal complications (e.g. ileus, bleeding)
Patients should improve after clinical interventions. Failure to progress, or signs of deterioration, suggest a new problem or an unresolved one. The same systematic CCrISP approach forms the basis of ongoing assessment of the critically ill or at-risk patient on the critical care unit or ward. As repeated complications and setbacks are likely in complex cases, the surgeon must be prepared for a long campaign as compared to a single battle, and be prepared to take a leading role in ongoing management.
Antimicrobial therapy in abdominal sepsis
Definitive management of sepsis requires eradication of the source of infection. However, the role of antimicrobial therapy is also vital.19 When sepsis is suspected, blood cultures, urine, wound swabs and sputum should be submitted for urgent Gram staining and culture, with all sources of sepsis considered. Cultures from the main source of sepsis are several times more likely to be positive (75% vs. 18%)20 than blood cultures, but both are important in the critically ill patient. Once cultures are taken, best-guess antibiotic therapy should begin immediately as delay may influence outcome.17 The role of cultures is to enable the antibiotics to be changed successfully if the patient fails to respond. The choice of antibiotic will be influenced by the clinical circumstances to cover the expected range of infecting organisms. Early combination antibiotic therapy yields significantly improved survival compared with single-agent use in septic shock.21 The route of administration must ensure adequate plasma levels and the drugs should penetrate adequately into the tissues. Intravenous infusion is usually necessary. Whenever there is doubt concerning the optimal choice of antibiotics, the advice of a medical microbiologist should be urgently sought. For most abdominal sepsis, coverage of Gram-negative and anaerobic bacteria will be necessary. With biliary sepsis, approximately 15% of cases will involve streptococci species that are resistant to cephalosporins, so the addition of a modern penicillin is a common approach. With postoperative hospital acquired infection, cover against a broader and more resistant spectrum of organisms will be needed.19 Fungal infection (usually Candida species) is not uncommon in complex abdominal sepsis requiring ICU care and often antifungal therapy will be required.
When severe sepsis is identified, blood cultures should be taken, and broad-spectrum antibiotics administered within 1 hour. As part of a management strategy in severe sepsis and septic shock, this has been shown to reduce mortality.17
Combination antibiotic therapy should be used in preference to monotherapy in severe sepsis and septic shock, as it is associated with a reduction in mortality.21
Imaging in abdominal sepsis
Various imaging techniques may be employed to localise an infective focus (see also Chapter 5). Computed tomography (CT), usually with gastrointestinal and/or intravenous contrast enhancement, can provide excellent information in thoracic, abdominal and pelvic sepsis. Most surgical patients can be stabilised sufficiently for safe scanning to take place and the assistance that CT gives in terms of accurate diagnosis and selective therapeutic intervention should not be underestimated. CT is excellent at primary diagnosis and at least as useful in the complex or postoperative patient where clinical examination is more difficult.22 Contrast can be usefully inserted up drains or down stomas when needed. Comparison with previous scans is important and the input of a senior, specialist radiologist will increase the accuracy of the report. In emergency cases, the surgeon should ideally be present at the scan so that decisions about any interventional radiological procedure can be made jointly.
It should not be considered that CT or any other diagnostic test is perfect. Interference from infusions, drains and metallic prostheses may reduce image quality. Intravenous contrast use is often contraindicated in acute renal failure, although gastrointestinal contrast can still be used to advantage. Even in expert hands, there is a small but significant rate of missed diagnoses. When emergency scans are interpreted by trainees, the rate is probably higher.
The chest radiograph remains an integral part of patient assessment and ultrasound has the advantages of being portable, harmless and repeatable. The greatest utility of ultrasound probably lies with the assessment of biliary and renal pathology and the monitoring of identified collections. However, it is limited by operator dependency, and a negative scan will offer little reassurance when the clinical picture is concerning. When a focus of subacute sepsis cannot be identified radiologically, isotopic methods such as labelled white cell scanning using indium-111 may help.
Early source control in abdominal sepsis
Source control describes the physical measures taken to control an infective focus. This includes the drainage of collections, debridement of necrotic tissue and definitive surgical procedures to correct the anatomical abnormality. Whilst it is intuitive that early source control will improve outcomes in abdominal sepsis, there is a relative paucity of data to support this, and obvious ethical considerations in performing prospective randomised trials. Delay to source control significantly increases mortality in septic shock,23 and there are clear advantages of expedient source control before progression to septic shock occurs.17,24
In a complex system such as a hospital, it is all too easy for multiple small delays to add up. Managing the multidisciplinary team to achieve prompt and timely intervention is a considerable skill, which requires active and continued leadership from the surgeon. The Royal College of Surgeons of England and the Department of Health have issued timelines regarding the urgency of source control in sepsis (Box 18.5), which are advocated as a standard of care.
Box 18.5 Timelines for source control in sepsis14
Patients with sepsis require immediate broad-spectrum antibiotics with fluid resuscitation and source control.
Control of the source of sepsis by surgery or other means should be immediate and under way within 3 hours
Control of the source of sepsis should be performed within 6 hours of the onset of deterioration
Control of the source of sepsis should be performed within 18 hours
Expedient control of the septic focus is of utmost importance in the management of severe sepsis. Neither overly prolonged resuscitation nor observation should delay this.14
Aims of treatment in abdominal sepsis
The management of abdominal sepsis in the emergency surgical admission is that of the underlying disease, as covered elsewhere in this book. The aim is to deal not only with the sepsis, but also to deal definitively with the underlying disease process or cause. In all patients, pus must be drained and dead tissue removed with specimens submitted for urgent microbiology. While localised collections can be drained percutaneously, generalised peritonitis remains an indication for laparotomy. Exceptions to this rule include primary spontaneous bacterial peritonitis and acute pancreatitis. There is an increasing use of laparoscopy in the management of specific surgical conditions causing abdominal sepsis (discussed further in Chapters 6, Chapters 8, 9, 10). However, it must be remembered that a laparoscopic approach may not allow sufficient access to adequately debride and drain the septic focus, and that the physiological sequelae of a pneumoperitoneum may be poorly tolerated in those with septic shock.
Whether treatment is radiological or surgical, adequate preparation is essential (see also Chapter 16). Coagulopathy must be excluded or corrected beforehand and blood for transfusion should be available. Although drainage may be an essential prerequisite to resolution of sepsis, it is not infrequent for bacteraemia, as a consequence of the intervention, to cause a temporary deterioration in the patient's condition. Indeed, a bacteraemia may represent the ‘second hit’ that precipitates MODS. Such circumstances should be anticipated and an appropriate level of post-procedure care arranged beforehand.
With more complex cases and with older and sicker patients, treatment may need to be tempered in order to be survivable. However, there is a balance to be struck as inadequate treatment will condemn the patient to recurrent sepsis and likely death. The opinion of senior colleagues and specialist centres can be invaluable.
Image-guided percutanous drainage of both spontaneous and postoperative intra-abdominal collections has reported success rates of 70–90%.25,26 The importance of adequate systemic support for the patient having imaging or image-guided intervention should nowadays be self-evident. Surgeons have a role to play in this, which may involve the assistance of anaesthetists. Percutaneous procedures may be less invasive than open surgery but will only be effective if good drainage is achieved. Many percutaneous drains are narrow and inadequate when infected fluid is viscous or contains necrotic tissue. Larger or multiple drains may be more effective and daily flushing can help. When radiological drains are placed for abdominal sepsis, the responsibility lies with the surgical team to ensure that the patient's condition improves as expected. Failure to respond to radiological intervention is another indication for laparotomy.
When open surgery is performed for sepsis, the procedure will vary according to the underlying pathology. Nonetheless, general principles apply in that the most straightforward adequate procedure is often preferable to a complex and time-consuming operation. The aim is to improve the patient's condition sufficiently without risking further complication. There is a current trend towards primary bowel resection and anastomosis in the acute setting, but this should be avoided in unstable patients, the presence of significant comorbidity, or heavy contamination. Generous saline lavage is recommended on completion of the procedure but there is nothing to be gained either by removing fibrinous debris piecemeal or by postoperative lavage.19 Delayed skin wound closure may be preferable to primary suture, or wounds may be left to close by secondary intention if sepsis is substantial.
Obtaining informed consent for treatment will involve both patient and relatives. The potential severity of the situation should not be underestimated and the possibility of death, stoma creation, the need for intensive care treatment and further surgery will usually need to be discussed explicitly. The average mortality for an emergency laparotomy is around 15%, and increases with age and physiological disturbance.14 Complementing clinical assessment with an objective determinant of risk from a scoring system (such as P-Possum) is valuable to help focus efforts.14,27 See also Chapter 15.
Often the early phase of postoperative care will be delivered on the surgical high-dependency unit (HDU) or ICU. Although we, as surgeons, may sometimes feel uncomfortable on ICU, there is a range of skills that we are usually best placed to deliver, and close cooperation between surgeon and intensivist is essential (Box 18.6).
Box 18.6 Surgeon's role on the ICU
Daily surgical input to:
Intraperitoneal abscesses with safe access routes should be drained percutaneously. This intervention carries high success and low recurrence rates.26
Abdominal sepsis on the ICU
Not infrequently, the surgeon is also involved with the assessment and management of patients already on the ICU who develop recurrent abdominal sepsis. The outcome of patients with abdominal sepsis who require ICU treatment depends on age, comorbidities, source of sepsis and organ dysfunction. Control of the source of sepsis is essential for survival among patients with MOF. When sepsis is eradicated successfully from these patients by surgery over 60% survive, whereas survival is close to zero if significant abdominal sepsis continues.28
The principal causes of recurrent abdominal sepsis in the ICU are shown in Box 18.7. The reader is referred to Chapters 6, Chapters 8, 9 and 10 for the management of each organ-specific condition. Leaking anastomosis remains the commonest single cause and should be actively suspected in all ‘at-risk patients’ who are critically ill or deteriorating. Enterotomy during difficult surgery is not uncommon, occurring in up to 20% of patients,29 and these can also leak in the postoperative period. Gastrostomy or other tubes inserted into the gut occasionally leak as well and this is more likely when tissue healing is poor.
Box 18.7 The principal causes of recurrent abdominal sepsis in the ICU
Leaked anastomosis or enterotomy
Leaking gastrostomies and other tubes
Abscesses or collections
Dead or ischaemic gut
Clostridium difficile-associated pseudomembranous colitis
Acute massive gastric dilatation
Continuing sepsis from ‘common’ peritonitis (perforation of peptic stress ulcer or diverticulum)
Postoperative small-bowel ileus usually resolves within days, regardless of the extent of bowel handling.30 Drugs, especially opiates, or electrolyte abnormalities (hypokalaemia, uraemia) may delay resolution, but failure to progress may also indicate ongoing retroperitoneal or abdominal pathology. Ileus may be difficult to distinguish from adhesive obstruction, and contrast studies may clarify the situation (see also Chapter 5). Adhesive obstruction frequently resolves but refractory cases occasionally require laparotomy. However, in a hostile abdomen, such as found in abdominal sepsis, considerable caution should be exercised in subjecting the patient to further surgery.31 Whereas a 5- to 7-day period of non-operative treatment might be acceptable in the presence of straightforward adhesion obstruction (and on occasions perhaps even longer if nutrition is supported), one should be prepared to wait for considerably longer when faced with a hostile abdomen, provided specific indications for surgery are not present. These indications include a known point of unrelieved complete obstruction or where there is a known septic focus.
Assessment On The ICU
Assessing patients on the ICU is difficult for several reasons. Firstly, the patients are often complex yet unfamiliar, perhaps having been treated previously by other surgeons. Secondly, sedated, postoperative patients in organ failure show their abdominal sepsis in different ways to a new emergency admission with peritonitis. Abdominal signs are unlikely to be evident unless gross (e.g. flank cellulitis, bowel contents in a drain, necrotic stoma), and the diagnosis of recurrent abdominal sepsis is often made from deterioration in vital organ function (seeChapter 16) and suspicion based on previous treatment and imaging. Again, contrast-enhanced CT is the gold-standard investigation and of great value, but occasionally patients will be too unwell to travel to the scanner. For them, exploratory laparotomy may still occasionally be necessary. Interpreting CT images in the recently operated abdomen is not easy. With expert help, the CT will not only confirm the diagnosis but can also potentially identify areas where there is no evidence of inflammation. In a difficult re-operative procedure with adhesions, that knowledge can save time and reduce the risk of surgical damage to other organs. Percutaneous drainage has a similar role here as in primary inflammation and the same caveats apply.
The importance of surgeons making their own thorough assessment cannot be overestimated as the stakes are high. The surgeon should be satisfied that the diagnosis is secure and that surgery is the best course of action. Part of that process will be engaging in detailed discussion with the intensivist to weigh up alternative diagnoses and sources of sepsis, and the risks and benefits of intervention at this point in time. Often this is not clear-cut as patients may have multiple potential sources (e.g. simultaneous pneumonia and abdominal sepsis) or other complicating factors, usually revolving around comorbid diseases or other ICU treatments. That said, any significant abdominal sepsis will require treatment.
Whilst the risks of surgery in the ICU population are often self-evident, these patients can sometimes be described as being ‘too sick not to have an operation’. Clear indications for life-saving surgery include generalised peritonitis, multiple collections and presence of dead tissue. Patients with deteriorating organ function and a strong suspicion of abdominal pathology also remain a significant group in whom laparotomy may be necessary. In some patients it will be clear that there is no realistic prospect of survival, either of the required operation or, more commonly, of the inevitably prolonged ICU course thereafter. It is important that both intensivist and surgeon counsel the family if care is to be limited.
Re-operating in abdominal sepsis
Re-operating in abdominal sepsis is always difficult: the timing of the most recent operation will influence the degree of difficulty and hence the risk of future complications. Beyond 72 hours adhesions will make surgery increasingly difficult and the risk of bowel damage increases: often the adhesions are stronger than the bowel. Entry to the abdomen can be difficult and an extension of the previous midline incision may help. Adhesions are generally most dense around any site of inflammation, as well as a recent incision. Here, having a preoperative CT scan comes into its own, guiding the surgeon as to the relative necessity to dissect in any difficult area. While a full laparotomy is ideal, prolonged dissection of adhesions in an area not thought from CT to contain inflammation or a collection may not be merited.
The surgeon must deal with the sepsis as thoroughly as possible but also simply and quickly. Prolonged and complex procedures will likely lead to a systemic deterioration in the patient and the prospect of bowel anastomoses healing under these adverse circumstances is not as high as one would like. The ability of the patient to withstand further surgery for further complications is very much lower next time around,28 and the surgeon should see the present operation as the best opportunity for salvaging the patient. Intestinal reconstruction can be attempted when the patient is well and recovered, some months later.
Hence, in general terms the simplest and safest procedure will be best. This holds true especially in the patient who already has incipient or established organ failure: drain sepsis, debride any necrotic tissue and exteriorise any leaking bowel or anastomoses as stomas. Most controversy relates to the management of enteric and colorectal anastomotic leaks. While in a well patient, with a small leak and minimal contamination, preservation of a repaired anastomosis with proximal defunctioning and local drainage may be appropriate, it is all too often foolhardy in the critically ill. Preserving the anastomosis might work but there is a significant risk of further peritonitis and likely death as further acute surgery will seldom be successful. A first salvage operation in an ICU patient will successfully eradicate sepsis over 40% of the time but a second operation carries a success rate of only 25% and a third operation only 7%.28This is reflected in the rate of survival.
In difficult cases, the surgeon must be ready to employ alternative strategies, although it remains essential to drain significant collections of pus or enteric content and to debride any necrotic tissue. In the multiply operated septic abdomen, it may not be possible to take down and exteriorise the prime source of sepsis because of dense adhesions or the anatomical location (oesophagus, duodenum). For inaccessible pelvic sepsis arising from distal small bowel or colon, it may be possible to identify and exteriorise a proximal loop of jejunum without entering and damaging the matted pelvic loops other than to achieve necessary drainage. This will usually relieve the sepsis but at the price of a high-output stoma and prolonged intravenous feeding. For oesophagogastric or duodenal sepsis, all that may be possible is to drain collections and leave large tube or sump drains beside the leaking anastomosis or other septic focus. Placing a second tube in the hole to create a controlled fistula is also of merit. Proximal intestinal contents or secretions can sometimes be diverted away by tube gastrostomy or other techniques, although each has its complications. A further option in difficult situations or resistant cases is to gain entry to locules of pus or enteric content, then to leave the abdomen open as a laparostomy (Fig. 18.2). Further pus or enteric content will usually find its way to the surface, assisted by subsequent manual lavage on the ICU or in theatre as necessary.
FIGURE 18.2 Laparostomy with polyglactin mesh, gastrostomy, jejunostomy, cholecystostomy and drains.
In addition to their role in draining proximal gut secretions, gastrostomy and enterostomy tubes can be usefully placed to facilitate future, and often prolonged, enteral feeding. The same tube can serve both roles: drainage initially, then feeding as intestinal function returns, provided there is no persisting distal bowel leak. These laparotomies are often oozy and contaminated, and many surgeons would leave large drains (24Ch tube or sump) in the subphrenic spaces and pelvis at the end of the procedure. This is particularly so for certain deep cavities (e.g. psoas abscess) or when further leakage is likely or inevitable. Large Foley catheters can be used to intubate inaccessible bowel to create a controlled fistula (typically the duodenum) and it is often advisable to place an additional large drain just outside the bowel. There is a particular role for local lavage in pancreatic necrosis (see Chapter 8) but, otherwise, continuing lavage down abdominal drains in the postoperative period is of no proven benefit.
Outcome from surgery for abdominal sepsis in the ICU patient depends on many factors. Age and eradication of sepsis have been discussed but numbers of failed organs, comorbid conditions and the underlying surgical disease are also important. Overall, 30–40% will survive,28 but a further factor that is associated with survival is the early response to the first operation in the ICU. If the patient has improved clinically within 48 hours then survival is much better (80%) than if the patient does not improve in this time frame (10%). Given the cost of intensive care, attempts have been made using a variety of scoring systems to define numerically those patients with negligible chance of survival (see Chapter 15). However, due to the heterogeneity of the patient group and the nature of scoring systems in general, it is not possible to use them for decision-making in individual patients. Clinical judgment is most important and the scoring systems remain primarily tools for audit and research.
Damage control laparotomy
Damage control laparotomy (DCL) is a concept that has expanded from its initial role in trauma surgery (see also Chapter 13). Trauma patients become hypothermic, acidotic and coagulopathic, and surgery becomes unsurvivable. Rapid, immediately life-saving surgery (‘staple, pack and go’) is carried out and the patient returned to ICU for warming and resuscitation, with more definitive surgery deferred for 24–48 hours once coagulopathy has been corrected and homeostasis returned towards normal. In abdominal sepsis it holds true that the first laparotomy carries the best chance for salvage.28 However, in severely unstable patients it can be better to make an active decision to quickly drain pus, remove dead tissue, stop bleeding by packing and close overtly leaking bowel with staples (without resection) before terminating the operation. Indications for this include haemodynamic instability, massive haemorrhage, coagulopathy, abdominal compartment syndrome and acute mesenteric ischaemia.32 Packs should generally be removed as early as possible once clotting is restored and usually the next day. The small bowel becomes adherent remarkably quickly and can be damaged as packs are removed. Packs should be removed cautiously under direct vision with saline irrigation and gentle finger separation.
Second-look or planned re-laparotomy
This term refers to the planned re-exploration of the abdomen, planned at the previous procedure in order to re-operate before any clinical deterioration occurs. The term distinguishes it from ‘laparotomy on demand’, which is now the more common approach and where the abdomen is only explored when a new problem is diagnosed. After DCL, a second look is obviously required to complete the necessary definitive procedures, but the term is more commonly applied to ‘looking again’ after laparotomy for intestinal ischaemia. The extent of intestinal ischaemia may not be fully evident at the first operation and, particularly if the bowel has been re-anastomosed, looking again at 48–72 hours, depending on progress, can identify further ischaemia before the patient deteriorates. In some cases of intestinal ischaemia where there may be doubt as to the extent, it may be preferable to resect and staple off the bowel ends, forcing the surgeon to re-explore 48 hours or so later, rather than gamble on anastomosing bowel that might be subclinically ischaemic and subsequently break down at a later date.
Repeated planned re-laparotomies have been used aggressively for abdominal sepsis with MOF in both Europe and North America for some years. In this method of treatment, the abdomen is typically re-operated upon every 24–48 hours for several days to wash out the peritoneal cavity and remove any ongoing sepsis. However, a recent randomised trial has shown that this approach is associated with significantly more laparotomies and prolonged ICU stay compared to a ‘re-laparotomy on demand’ approach.33 Although on occasions there may be a case for a single planned second-look laparotomy in patients with severe faecal peritonitis, even this is debatable. This should not, however, divert the surgeon from maintaining a low threshold for early laparotomy on demand when indicated, as delaying necessary surgery worsens outcome.28
In patients undergoing surgery for severe secondary peritonitis, re-laparotomy on demand is preferable to ‘planned re-laparotomy’.33
Leaving the abdomen open (laparostomy)
In adverse circumstances, the abdomen is occasionally left open as a last resort in allowing pus and enteric contents to drain. More often it is left open in preference to closing it with undue tension. Abdominal compartment syndrome (see below) may also lead to an open abdomen. Leaving the abdomen open in appropriate circumstances can avoid or reduce septic, enteric and wound complications. It also allows improved systemic function, faster weaning from the ventilator and can facilitate early enteral nutrition. However, it is deforming, increases evaporative losses, and exposes the bowel to risk of damage and fistulation. The open abdomen is particularly challenging to nurse, and comfortable, secure, wound care may be difficult to achieve.34 Prolonged healing is usually required with a risk of late herniation. These deleterious effects are significant enough to make one caution against unnecessary use of laparostomy.19 It is undoubtedly a valuable and life-saving technique when needed but it is also a significant future burden to the patient in its own right.
In closing the abdomen during laparotomy for abdominal sepsis, the surgeon should close the fascia conventionally but avoid tension. Tension sutures have little to recommend them and their continued use is not supported. If bowel distension, oedema, haemorrhage (or packing) combine to make closure impossible, then there are several options open to the surgeon, depending on a number of factors, including the likely time course of recovery. In general, each end of the wound is closed conventionally to the point of reasonable tension and the central defect left open. For short-lived oedema (e.g. after aortic aneurysm repair or traumatic haemorrhage) the defect can be covered with abdominal packing, which is changed daily until such time as the defect can be closed or a longer-term solution implemented (see below). Early dressing changes should be carried out by the surgical team. The small bowel will become adherent to any gauze dressings, and gentle separation will be required to avoid injury.
A double-sandwich dressing of semipermeable adhesive dressing with moist gauze between the layers of dressing protects the bowel with less adhesion formation than with standard gauze packs. There are various commercial plastic sheets that can be used. Alternatively, a version of the Bogota bag can be used (Fig. 18.3). In this technique, a sterile 3-litre intravenous fluid bag is slit open and sutured to the fascia, covering and protecting the bowel and providing it with a clean, moist environment. Again, as the oedema subsides, usually within 72 hours or so, the Bogota bag or double-sandwich dressing can be removed and the abdomen either closed or a longer-term technique instituted.
FIGURE 18.3 Bogota bag.
In recurrent sepsis, where recovery is likely to be slow, prosthetic mesh can be used to restrain the viscera. In abdominal sepsis, absorbable polyglactin meshes are preferred to non-absorbable polypropylene meshes, as there is less likelihood of chronic mesh infection and fistulation to underlying bowel.35 Caution is also required in the use of bioprosthetic collagen meshes on account of complications and cost.35
Whichever technique is employed to close the abdomen, the later development of an incisional hernia is almost inevitable, but might be less so with the non-absorbable mesh. If the contamination is particularly severe such that a further laparotomy and lavage is planned in 24–48 hours, then mesh placement can be deferred until then. Likewise if oozing requires gauze packing then mesh placement can be effected when the gauze is removed. The mesh and bowels must still be kept moist and protected, and a double-sandwich dressing achieves this admirably. If there is a lot of effluent or tissue fluid, then use of low-grade suction can help provide control. As the wound shrinks, these bulky dressings can be changed to a large fistula bag, even if there is no fistula, as again it provides a non-adherent moist environment.
Use of commercial vacuum dressings has become widespread as they make wound management considerably easier, and provide a ready-made technique for the non-expert unit. Early cohort studies suggested that vacuum-assisted closure was well tolerated in the open abdomen, with intestinal fistulation rates of 5%.36 However, these findings have not been borne out by the only randomised trial on vacuum-assisted closure in the open abdomen, which compared it to polyglactin absorbable mesh. Whilst not statistically significant, the rate of fistulation in the vacuum-assisted closure arm was 21%.37 Controversy regarding the rate of intestinal fistulation continues,34,38 while the results of a national audit coordinated by the National Instititute of Health and Clinical Excellence (NICE) are awaited. Until the intestinal fistulation rate has been clarified, particular caution is required on the use of vacuum dressings in the presence of suture or staple lines, repaired serosal tears or enterotomies.
Abdominal compartment syndrome
Normally, intra-abdominal pressure is low (< 10 mmHg), but it is increasingly recognised that it can be raised in abdominal sepsis (and other acute abdominal conditions, including trauma and pancreatitis) to the significant detriment of the patient. The condition is recognised with increasing frequency and is probably not as uncommon in our practice as previously thought.39 As abdominal pressure rises, venous return is impaired and with it cardiac output. The tense abdomen can cause pulmonary compromise, oliguria, mesenteric ischaemia and even raised intracranial pressure. These features constitute the abdominal compartment syndrome (ACS) – similar in some ways to a tension pneumothorax, within the abdomen.40
ACS is caused by multiple factors that exist after certain operations, typically those for peritonitis, abdominal aneurysm and abdominal trauma. Tissue oedema results from the combined effects of tissue injury, intravenous fluid infusion and leaky capillaries, while bowel distension and haematomas also contribute. ACS is seen most commonly after fascial closure but it can also occur in the abdomen that has been left open and packed, especially if there is ongoing haemorrhage. Operations likely to cause ACS include those involving significant haemorrhage, retroperitoneal or intestinal oedema, bowel distension, aortic clamping, hypothermia, massive transfusion or prolonged surgery. These circumstances should heighten the surgeon's awareness of ACS as a potential complication. ACS is not restricted to emergency surgery and prolonged elective surgery with a scarred and rigid abdominal wall may also lead to this condition. The anaesthetist may signal an unacceptable rise in the ventilatory pressure as the abdomen is closed but more commonly ACS develops on the critical care unit, some 12–30 hours after surgery. Oligo/anuria and raised ventilatory pressures are the usual presenting features.
Intra-abdominal pressure (IAP) is measured via the bladder following instillation of 25 mL of normal saline. The IAP is measured through the aspiration port of the catheter tubing using a transducer. The transducer should be zeroed at the level of the mid axillary line, and IAP measured in the supine position at end expiration.41 Standardised definitions for ACS were developed in 2006 by an international consensus group41 (Box 18.8).
Box 18.8 Definitions of abdominal compartment syndrome41
Intra-abdominal pressure (IAP)
Steady-state pressure in the abdominal cavity
Between 5 and 7 mmHg in critically unwell adults
Abdominal perfusion pressure (APP)
APP = MAP − IAP
Intra-abdominal hypertension (IAH)
Sustained or repeated pathological elevation in IAP ≥ 12 mmHg
Abdominal compartment syndrome (ACS)
Sustained IAP > 20 mmHg (with or without an APP < 60 mmHg) that is associated with new organ dysfunction or failure
In intra-abdominal hypertension (IAH) and ACS, a number of medical treatment options are of benefit in reducing IAP.42 Abdominal wall compliance can be improved with adequate sedation, analgesia and neuromuscular blockade. Intraluminal contents should be evacuated with nasogastric and rectal decompression and the use of prokinetic agents. Abdominal fluid collections should be aspirated. Positive fluid balance can be corrected with fluid restriction, diuretics or dialysis/ultrafiltration. However, if pressures above 20 mmHg persist despite these measures, or organ dysfunction worsens, then the abdomen will need to be decompressed and left open using the techniques discussed above.
Intestinal fistulas pose their own particular problems and can greatly complicate patient management. They contribute to sepsis, malnutrition, fluid and electrolyte imbalances, difficulties in wound care, as well as posing an enormous psychological challenge to the patient.
A fistula is defined as an abnormal communication between two epithelial lined surfaces. There are many types and exhaustive description is beyond the remit of this chapter. The great majority of those seen within the context of severe abdominal sepsis follow surgery and result from anastomotic leakage or an inadvertent enterotomy, either overlooked or unsuccessfully repaired.
An intestinal fistula, from somewhere in the intestinal tract to the laparotomy wound, will occur occasionally in every gastrointestinal surgeon's practice, arising most commonly as a result of anastomotic leakage or bowel damage. While a few will show signs of severe sepsis, more often than not there is a period of apparent ileus, wound infection and clinical stagnation. When the wound ruptures or is opened to treat the infection, the enteric or faecal nature of the contents will become apparent. This may not be convincing to begin with as the enteric flow is usually preceded by a volume of pus and blood, as with most postoperative wound infections. Postoperative fistulas may also occur through drains or along recent drain sites, to the vaginal vault (especially in those who have undergone previous hysterectomy) and occasionally to the rectum or other parts of the gut.
However, fistulas are more likely to occur after emergency surgery, usually carried out for another postoperative complication, commonly sepsis, obstruction or bleeding. In these operations, the inherent difficulty of the procedure, brought about by adhesions, bowel distension and softened tissues, makes further bowel damage a real possibility. If that damage is not, or cannot, be repaired effectively, if there is persisting obstruction of the intestine postoperatively, a postoperative phlegmon or an open abdomen, then the likelihood of fistulation escalates considerably. Postoperatively, the combination of small-bowel obstruction and an undrained abscess is also likely to result in a fistula as the obstructed bowel eventually softens and gives way at, or into, the collection. Thus it will be evident that repairing a leaking anastomosis once the patient is septic and local tissues oedematous and friable is all too often doomed to failure. Bowel exposed in an open abdomen will also invariably be subject to some degree of trauma unless the dressings or appliances are handled and changed expertly. Again, if there is distal obstruction, a local and exposed suture line or local trauma, a fistula will often occur. Perhaps 10–20% of ‘laparostomies’ will fistulate and this sometimes happens even with expert care.
A significant number of intestinal fistulas heal spontaneously, although they will cause misery and morbidity while they do so. The factors that contribute to persistence of a fistula are shown in Box 18.9.
Box 18.9 Factors contributing to the occurrence and persistence of postoperative fistulas
Inadvertent enterotomy (repaired or missed)
New anastomosis in unfavourable circumstances
Persisting abscess or phlegmon causing obstruction
Distal obstruction (including constipation)
Disconnected bowel ends
High output from fistula
The best-known approach to fistula care is that described at the specialist fistula unit in Salford (UK), and known by the acronym SNAP (Sepsis, Nutrition, Anatomy, Procedure). The most pressing effects of a fistula relate to intra-abdominal sepsis. There can also be wound sepsis, usually in the form of cellulitis, although occasionally local necrosis can occur. Once the necessary resuscitation is in hand, an early priority is CT of the abdomen, preferably with both gut and intravenous contrast. The frequency of intra-abdominal abscess formation is high, in the region of 66%,43 and often these are not evident clinically. Without diagnosis and control of the sepsis, the prognosis is bleak. Management strategies follow the principles outlined above.
An integral part of early management is wound care and control of the fistula effluent. When there is a small fistula orifice in a drain site or through part of a wound, the effluent can be controlled with a stoma bag. At the other extreme, when bowel contents are leaking into a laparostomy, management may be very difficult. Large fistula bags are available; the largest may be needed for a new laparostomy. The author's unit often uses a sandwich dressing initially, as described above, with low-grade suction attached to soft rubber catheters placed near the fistula site. The same suction technique can be placed inside a fistula bag when appropriate in order to reduce the frequency of bag leakage.
Important ancillary techniques include reducing fistula output by avoiding enteral feeding and reducing gastric acid output with proton-pump inhibitors. Codeine and loperamide can subsequently be used to contribute to this control of gut secretions. Octreotide will also reduce gut secretion in some patients but, in the author's experience, its effect is inferior to the steps already described. It is perhaps most useful in pancreatic fistulas and upper small-bowel fistulas.
Immediate fluid management will have been considered as the patient was resuscitated but attention will have to turn rapidly to the optimal means of nutrition. The deleterious effect that enteral nutrition often has on output and wound care, particularly in the early stages, has been noted above. Additionally, enteral feeding may also ‘feed’ the abdominal sepsis, if there is complexity to the fistula with a further unrecognised proximal hole in the bowel. There is often partial obstruction associated with the sepsis and with most fistulas it is often better and more reliable to resort to parenteral nutrition while the patient stabilises and the situations with regard to sepsis, wound care and fistula anatomy each become clearer. Indeed, it is often better to see parenteral nutrition as the ongoing mainstay of nutritional support in intestinal fistulation with abdominal sepsis. In some circumstances, enteral nutrition can take over some, or even all, of the role, but this will usually take time and is only possible in selected patients.
Fistulas can be categorised as high or low output (high > 500 mL per 24 hours). This cut-off represents a level above which the fistula is likely to have a significant effect on fluid balance and nutrition. Some proximal fistulas have outputs of one or more litres per day and fluid balance will be challenging in its own right. Senior surgical staff will need to commit to adequate supervision of the recording of the various inputs and outputs, particularly in the initial unstable and complicated phase of care. Fistula output will often reduce with time and, when possible, some healing begins to occur. A typical anastomotic wound fistula will be a side hole and if there is free distal flow, no sepsis, no bowel disease and no distal obstruction (or constipation), then healing may occur. It is unknown whether restricted oral intake helps, but it is intuitive to minimise the flow through the hole to encourage healing. As healing occurs, usually after 2–4 weeks, oral intake can be increased. If there is no healing by 6 weeks and no apparently reversible factor (e.g. abscess, constipation), then many would assume the fistula will not heal and allow more liberal oral intake.
In the ICU setting, once sepsis is excluded or controlled and the open abdomen, if present, is starting to granulate, oral or enteral feeding can begin cautiously, maintaining a careful watch for recurrent sepsis. The nutritional effect of this will depend on the length and condition of gut available for absorption and it will often take time for enteral feeding to be tolerated. During this phase, combined enteral and parenteral feeding is used.
It will be clear from the above that further surgery is usually only advocated to deal with abdominal sepsis. Further operations at this stage to try and correct the fistula often cause more harm than good through loss of bowel (risking short-bowel syndrome) or the creation of yet more fistulas. If surgery is undertaken then the approach follows that described above for abdominal sepsis. Surgery for fistulas is best deferred until the patient is well physically, nutritionally and emotionally, and only after the anatomy has been defined radiologically. This is often some months later and in complex cases is a highly specialist undertaking.
Chronic Abdominal Sepsis
Intestinal fistulas often settle down and can run a course of chronic abdominal sepsis where there is low-grade inflammation but no organ failure. This can be punctuated with further acute episodes and it may be difficult to be certain if the sepsis comes from the feeding line, urinary catheter, chest, wounds or abdomen. Unusual bacteria or fungi can be involved and occasionally less common sites become infected (e.g. myocardium or heart valves secondary to feeding lines). During this period there is a continuing requirement for close daily monitoring with senior surgical input, often for several months, if the patient is to recover. Ongoing parenteral nutrition is often integral to this and, again, the help of a specialist unit can be invaluable. In these units, the support of specialist nursing staff in maintaining sepsis-free nutrition and providing expert fistula care is fundamental to success. The overall recovery or otherwise of the patient will be indicated by the clinical and biochemical picture combined. A falling white count and C-reactive protein with a steady rise in the serum albumin, together with increasing general well-being, are all good signs. Further surgery should probably be deferred for at least 6 months after the last laparotomy in order to allow the abdomen to become less hostile, and prolapse of stomas or fistulas is a useful indicator that re-peritonealisation has occurred. Restorative surgery is still usually prolonged and difficult, and reconstructing the abdominal wall after laparostomy has its own difficulties, which may benefit from help from a plastic surgeon.