Management
of community-acquired peritonitisManagement
of community-acquired peritonitis
This text was published in Annales Francaises d'Anesthesie Reanimation 2001;20:368-73.
This conference was organized and conducted n compliance with the methodological rules established by the Anaes (Agence Nationale d'Accréditation et' d'Evaluation en Santé, the French national agency for health care practice and technology assessment), which has acknowledged the quality of the French version of the present text. The conclusions and recommendations presented in this document reflect work conducted independently by the consensus conference panel. The Anaes does not necessarily endorse the contents of this document. The letters in parentheses indicate the levels of proof deterrnined based on a literature review.
A : Double blind randomised trials (high level of proof).
B : A single randomised study or several prospective studies
C : Non-randomised studies with a concurrent control group or case series or consensus panels.
Older age, comorbidities, immunological compromise, undernutrition, and organ failures are independently associated with an adverse outcome (mortality). Institutionalisation and/or previous anti-microbial therapy are probably of adverse prognostic significance but have not been evaluated.
The Acute Physiology And Chronic Health Evaluation score II (APACHE II) and the Mannheim Peritonitis Index (MPI) provide values that are relevant for stratifying patients into subsets of similar disease severity. Furthermore, they predict overall mortality.
The Simplified Acute Severity Score (SAPS) has not been evaluated in community-acquired peritonitis.
Presence of faeces in the peritoneal fluid is associated with greater severity (mortality). Penetrating wounds of the gastrointestinal tract, perforation of the colon, extensive lesions, and lesions at other sites are independently associated with increased severity (postoperative infection).
The pathogenic role of Escherichia coli and anaerobes observed in clinical practice confirms experimental data. Although several studies suggest that presence of enterococci may be independently associated with postoperative infection, there is no compelling proof of an effect on mortality (Grade D).
Longer time to surgery, specific surgical conditions, and the quality of the surgical procedure deserve consideration, despite the absence of evaluations. Inappropriate first-line antimicrobial therapy is independently associated with increased severity (post-operative infection).
The intestinal flora is composed of a broad variety of bacterial species. The number of bacteria per gram of digestive tract content ranges from 102 or 103 at the stomach to 10l2 at the sigmoid colon. Most digestive tract bacteria are obligate anaerobes. The ratio of obligate anaerobes over aerobes varies from 10/1 at the small bowel to 1000/1 at the sigmoid colon.
The species recovered from patients with community-acquired peritonitis are digestive tract commensals that colonize the peritoneal fluid. Peritoneal fluid specimens usually contain multiple organisms: about two to four different species per specimen.
The most commonly recovered micro-organisms are as follows: among aerobes or aerotolerant anaerobes, E. coli (605'o to 70%) and Enterococcus spp (10% to 30%); among obligate anaerobes, members of the Bacteroides genus (with a marked predominance of the Bacteroides fragilis species, found in 20% to 45% of cultures) and members of the Clostridium genus (5% to 20%).
The other bacteria are Gram-negative rods belonging to the Klebsiella (10% to 20%), Enterobacter, and Proteus (5% to 10%) genera; Pseudomonas aeruginosa (10% to 20%); and Gram-positive cocci of the Staphylococcus, Streptococcus, and Peptostreptococcus genera. Candida albicans is found in 3% to 5% of cases.
Aerobic bacteria, most notably E. coli, act in synergy with obligate anaerobes in the pathogenesis of peritonitis.
The flora differs between peritonitis related to supracolic and infracolic lesions.
One or two pairs of blood specimens should be taken for culturing on aerobic and anaerobic media prior to initiation of antimicrobials (Grade E).
Collection of peritoneal fuid specimens (Grade D) should be performed under conditions that ensure survival of anaerobes: limited contact with air and immediate transport to the laboratory, at room temperature. Aerobes, obligate anaerobes, and Candida should be looked for. Antimicrobial susceptibility testing of predominant species is recommended (Grade A).
Microbiological studies are of little assistance for diagnosing peritonitis.
- However, presence of some micro-organisms (Candida, Enterococcus) in the peritoneal fluid and positive results of blood cultures can suggest a higher risk of severe outcomes. Furthermore, postoperative infectious complications may be significantly more common in patients whose microbiological studies show strains that are resistant to the antimicrobials used as first-line treatment.
The panel recommends the following microbiological studies:
- peritoneal fluid cultures to allow adjustment of the antimicrobial regimen according to the results of susceptibility testing on recovered organisms (Grade D);
- blood cultures in all situations; these cultures are indispensable, at least in the more serious forms (Grade E).
During the course of peritonitis, nothing is to be gained from culturing the drainage fluid. Conversely, in patients with a secondary deterioration in their condition, the panel recommends culturing specimens obtained by aspiration of collections under radiological guidance (Grade D).
From a public-health viewpoint, monitoring resistance profiles is a second reason for performing bacteriological studies in patients with community-acquired peritonitis.
Surgery is the mainstay of the management of community-acquired peritonitis. Surgery should take place as early as possible and should remove the source of the inoculum. Preoperative preparation of the patient involves correction of the main humoral disorders (serum potassium abnormalities in particular) and hemodynamic disturbances (volume replacement, vasopressors). Surgery should never be delayed by more than a few hours; prompt surgery is particularly important in patients with persistent shock despite intensive care (Grade D).
Either laparotomy or laparoscopy is used, according to the current health problem, underlying conditions, and experience of the surgeon. No controlled prospective studies of adequate quality have found evidence that either of these approaches is superior over the other in patients with community-acquired peritonitis.
The main indications for laparoscopy as the initial approach are perforation of a duodenal ulcer and peritonitis related to appendicitis. For laparotomy in adults, a long midline incision should be performed. In children, elective approaches are used in most cases. Specimens for microbiological studies should be collected routinely. Then, the entire peritoneal cavity should be explored.
When digestive tract perforation and/or necrosis is found, immediate removal of the abnormal area is recommended, although some lesions may warrant a more conservative procedure.
The management after bowel resection is not well standardized (Grade E): the high risk of dehiscence of a suture or anastomosis in a septic environment often prompts surgeons to prefer ostomies, particularly when surgery is performed late or when there is persistent shock.
In patients with appendicular lesions, appendectomy is performed as a rule, although simple drainage is used occasionally in patients with walled-off peritonitis; in children, laparoscopy may be the best approach.
In patients with peritonitis related to diverticular disease of the sigmoid colon, several procedures can be used: resection without anastomosis; resection with protected anastomosis; suture of the perforation, if needed with protection; or exteriorisation of the perforation (Grade E). The choice of the method depends mainly on local anatomic conditions.
In patients with colonic perforation related to a tumour, the usual procedure is immediate resection of the tumour and perforated area without anastomosis.
Trauma-related perforation of the colon and/or rectum requires, according to the extent of the lesion either protected suturing or segmental colectomy without anastomosis. Suturing or single-step resection and anastomosis can be considered only when surgery is needed in a patient with iatrogenic perforation during colonoscopy after colonic preparation.
Colonic perforation related to vascular disease requires excision of a bowel segment, without anastomosis. The length of bowel removed is determined according to the size of the ischaemic area.
Perforation of a duodenal ulcer should be treated by simple suturing, which can be performed by laparoscopy. In perforation of a gastric ulcer, excision of the ulcer should be performed also.
Excision of the stomach is required in patients with gastric perforation related to a tumour.
When the lesion is in the small bowel, single-step resection and anastomosis is warranted if surgery is performed promptly in a patient who is in good general condition and has no specific risk factors. A double ostomy is generally needed when the time to surgery exceeds 12 hours or the patient is in shock (Grade E).
Biliary peritonitis should lead to cholecystectomy, if needed with external biliary drainage. All forms of bile tract-to-digestive tract anastomosis should be avoided in a septic environment (Grade E).
Intra-operative irrigation of the peritoneal cavity is now performed routinely. A large amount of fluid should be used in order to decrease the risk of postoperative infection. Adding antimicrobials or antiseptics to the lavage fluid has not been proved effective. There is no compelling evidence that irrigating the peritoneal cavity during the postoperative period is beneficial.
Whether drainage is beneficial remains controversial, and no recent studies have evaluated the usefulness of routine drainage in patients with generalized peritonitis. Based on the rules of good clinical practice, the following can be recommended (Grade E):
- in purulent peritonitis with surgery within the first 12 hours, no drainage or simple drainage in contact with the treated causative lesion, particularly when there is reason to fear postoperative development of a fistula;
- in stercoral peritonitis and purulent peritonitis with a long time to surgery, routine drainage of the subphrenic areas, lateral gutters, and Douglas pouch is still used by a few groups of surgeons but has not been proved effective; not recommended formally as a rule of good clinical practice.
Conventional closure in layers is the rule, and addition of full-thickness stitches remains debated. Laparostomy is not indicated in patients with community-acquired peritonitis.
Few studies have specifically investigated community-acquired peritonitis; conversely, there is an abundance of publications on antimicrobial regimens used as adjunctive treatment for surgically-managed intra-abdominal infections. Only 12 trials provided a high level of proof.
The main obstacle to comparisons is the heterogeneity of these trials, particularly regarding characteristics related to underlying conditions. Furthermore, evaluation of severity was often inadequate. Only one meta-analysis has been performed.
A very large number of studies used clindamycin combined with an aminoglycoside as the reference treatment. Use of this combination in France has been marginal. Another combination suggested as the reference treatment is metronidazole with an aminoglycoside.
Although amoxicillin with clavulanic acid is widely used in France, it has been evaluated in only two trials (level II).
These trials investigated clindamycin + aminoglycoside, metronidazole + gentamicin, third-generation cephalosporin + metronidazole, ciprofloxacin + metronidazole, ampicillin-sulbactam, and carbapenems.
The results of these trials fail to demonstrate superiority of any of these treatments over the others.
These studies investigated ureidopenicillins, third-generation cephalosporins, cefamycins, oxacephems, monobactams, and fluoroquinolones, in addition to the compounds studied in level I trials.
Analysis of data from level II and III trials does not alter the conclusions above and provides no information on the possible role for the other compounds.
In conclusion, the clinical efficacy of antibiotherapy in intra-abdominal infections does not necessarily denote microbiological efficacy. An increase in disease severity may need to be factored into the choice of antibiotic regimen. Available information does not seem to identify a given antibiotic regimen as a reference standard for the management of community-acquired peritonitis.
Probabilistic antibiotherapy is started as soon as the diagnosis of peritonitis is made. Bacteria that should be considered routinely are enterobacteria, particularly E. coli, and anaerobes, particularly B. fragilis (Grade A). These bacteria are the main causes of early death and of residual abscesses. Enterococci are recovered in 10% to 30% of patients with community-acquired peritonitis. Although the role for enterococci in postoperative infectious complications is well established, there is no consensus about the need to consider this organism when choosing the initial antibiotic regimen (Grade E).
Recovery of Pseudomonas aeruginosa has been reported in several studies of community-acquired peritonitis. The possible presence of this organism should be considered only in specific situations.
Bacteriological differences across sites, particularly between supracolic and infracolic peritonitis, are not sufficiently marked, even in terms of the size of the inoculum, to influence the choice of antibiotics.
Theoretical advantages associated with use of two or more antibiotics are listed below.
The issue is not whether single-drug or combination-drug therapy is best but rather how well the spectrum of the drug or drugs used covers the organisms involved.
Selection of a synergistic combination may be warranted in patients with severity factors and/or with specific bacteria such as P. aeruginosa.
Combinations of antibiotics can be used to prevent the emergence of resistant strains within the focus of infection; on the other hand, combinations may contribute to increase the selection pressure placed on the commensal flora.
When identifying regimens that meet these criteria, the following should be considered: distribution in France of resistance among bacterial species usually found in patients with community-acquired peritonitis; risk of emergence of strains with resistance to fluoroquinolones used as single-drug therapy; absence of direct information on the susceptibility of bacteria recovered from peritonitis fluid; and fragmentary nature of available studies. Regimens that meet the above-mentioned criteria include:
The following should be considered when designing treatment regimens:
- The panel recommends separate selection of antibiotics used for prophylaxis and of antimicrobials used as curative treatment (Grade E).- Addition to some regimens of an aminoglycoside with the goal of optimising the treatment by obtaining a synergistic effect can be considered in patients with severity criteria (Grade D).
- In patients with valvular disease, prophylaxis of infectious endocarditis caused by enterococci and/or streptococci is imperative.
- Administration modalities (dosages, dosing times) must be selected according to pharmacodynamic constraints (Grade A).
In situations that carry a high risk of death, antibiotherapy must meet two criteria: the regimen must provide optimal bactericidal efficacy, and the spectrum of the regimen must cover all the organisms that might be present. There are no clinical data confirming that these criteria are useful in patients with community-acquired peritonitis. Published studies have not provided compelling data in favour of intensified probabilistic antibiotherapy; however, this strategy seems reasonable in situations associated with a high risk of death. In patients with community-acquired peritonitis, the regimens "amoxicillin/ clavulanic acid+ gentamicin", "ticarcillin/ clavulanic acid + gentamicin", cefoxitin", "cefotaxime or ceftriaxone + imidazole derivative", and "gentamicin + imidazole derivative" (only in patients allergic to beta-lactams) should be considered inadequate only in clinical situations that carry a high risk of death (Grade D). In these situations, disease related to P. aeruginosa, Enterobacter, other resistant Enterobacteriaceae, or nonfermenting Gram-negative rods, although rare, should be considered, most notably in patients at risk for changes in gastrointestinal tract bacterial ecology (institutionalisation, previous anti-microbial therapy) (Grade D). For all these regimens, the need for continued treatment should be reappraised based on the clinical course and on the results of bacterial cultures and identification tests (Grade E).
The panel recommends that each health-care institution establish a set of written procedures that are developed by consensus, described clearly, and made available to all health care workers.
Few clinical studies with a good level of proof provide information of assistance for answering this question. Clinical experience and "medical cornmon sense" are used to evaluate how well the patient tolerates the treatment, to look for persistent infection, and to confirm that full recovery has occurred.
In the surgical ward, monitoring requires only an evaluation of clinical parameters including the general appearance of the patient, vital functions, and return to normal of body temperature and gastrointestinal tract function.
The same clinical parameters are monitored in the intensive care unit. Weaning from artificial ventilation and discontinuation of vasopressors are indicators of a favourable course. Simple laboratory tests should be done to monitor any changes in organ failures during the first few days. Isolated leucocytosis is not specific for persistent sepsis. Drainage fluid cultures, tests for inflammation, and abdominal imaging studies are not necessary.
The duration of antibiotherapy depends on the time from peritoneal contamination to surgery, on the severity of organ lesions, on the nature of the peritoneal effusion, on the initial clinical severity, and on whether immunodepression is present. The durations below are suggested as examples, and decisions should be made on a case-by-case basis.
- penetrating abdominal wounds with a breach in the gastrointestinal tract treated surgically within 12 hours of the injury: 24 hours (Grade A).- localized peritonitis, perforated gastric or duodenal ulcer: 48 hours (Grade A).
- generalized peritonitis with prompt surgical treatment: five days (Grade D).
- generalized stercoral peritonitis or generalized peritonitis seen late regardless of the location: seven to ten days (Grade D).
Abdominal or extra-abdominal manifestations (respiratory, haemodynamic, metabolic, psychiatric ...), fever, and/or leucocytosis that are not explained and persist beyond the scheduled duration of antibiotherapy suggest intra-abdominal sepsis and invite an appraisal of the need for an additional therapeutic procedure.
Medical imaging studies are useful for determining whether an additional procedure is needed. They rarely contribute to the diagnosis when performed before the fourth postoperative day.
Advantages of ultrasonography include wide availability, absence of adverse effects, and the possibility of repeating the investigation if needed. The results are inconsistent, and there are many limitations to this technique. Nevertheless, ultrasonography has good sensitivity for detecting subphrenic abscesses and biliary complications. A normal result does not rule out an abdominal complication (Grade B).
Computed tomography (CT) has several disadvantages: the patient is exposed to radiation and must be taken to a radiology room, and the investigation is sensitive only if performed after injection of an iodinated contrast agent. CT contributes to the detection of extra-abdominal foci of infection (Grade A).
Ultrasonography or CT can be used to guide percutaneous drainage of fluid collection. CT is more accurate than ultrasonography and, consequently, is more helpful for dilfficult locations (Grade D).
Percutaneous drainage should induce a rapid and marked clinical improvement, and presence of an abscess does not rule out a progressive lesion. Progressive lesions usually require surgery.
When there is no evidence of extra-abdominal infection, persistent organ failure, or sepsis, a repeat surgical procedure should be considered even in patients with normal CT findings (Grade D).
A midline and very long incision should be performed to allow exploration of the entire abdominal cavity. No digestive tract sutures or anastomoses should be performed at the infracolic level, where ostomies should be preferred after removal of the causative lesion. At the supracolic level, contact drainage is used to avoid development of a fistula; extensive drainage of dependent areas of the abdominal cavity and of both lateral gutters is performed routinely. Prolonged antibiotherapy should be given, with no antibiotic-free interval. The antibiotics should be selected based on the microbiological data.
Back to the Homepage
of the SFAR