1035

STATE-OF-THE-ART CLINICAL ARTICLE

Peritonitis: Update on Pathophysiology, Clinical Manifestations, and Management

Caroline C. Johnson, James Baldessarre, and Matthew E. Levison
From the Division of Infectious Diseases, Allegheny University of the Health Sciences, and Veterans Affairs Medical Center, Philadelphia, Pennsylvania; and R. W. Johnson Pharmaceutical Research Institute, Raritan, New Jersey

Peritonitis results from any local trigger of inammation. Usually infection is the trigger, although infection may not necessarily be present at the earliest stage. For example, sterile peritonitis may occur in the localized peritoneal space surrounding an infected but resectable intraabdominal organ, such as the appendix or gallbladder. In contrast, there may be contamination of the peritoneum from a defect in the intestinal wall, before establishment of infection or onset of an inammatory response, e.g., immediately following penetrating abdominal trauma. Peritonitis has been categorized as primary, secondary, or (more recently) tertiary. Peritonitis complicating peritoneal dialysis can be considered as an additional category. Each of these categories is reviewed herein, with an emphasis on recent developments in pathogenesis, evaluation, and management. Primary Peritonitis Pathogenesis. Primary peritonitis is an infection of the peritoneal cavity not directly related to other intraabdominal abnormalities. The vast majority of cases are due to bacterial infection; it is also commonly known as spontaneous bacterial peritonitis. Usually it occurs in the presence of ascites from any of a variety of underlying conditions [1]. In the preantibiotic era, primary peritonitis accounted for 10% of all pediatric abdominal emergencies; it now accounts for 1 2% [2]. The decline has been attributed to widespread use of antibiotics for minor upper respiratory tract illness. Although primary peritonitis may occur in children without predisposing disease, it is especially associated with postnecrotic cirrhosis and nephrotic syndrome. In adults, primary peritonitis develops in up to 25% of patients with alcoholic cirrhosis but has also been reported to occur in adults with postnecrotic cirrhosis, chronic active hepatitis, acute viral hepa-

Received 21 January 1997; revised 10 February 1997. Reprints or correspondence: Dr. Matthew E. Levison, Division of Infectious Diseases, Allegheny University MCP, 3300 Henry Avenue, Philadelphia, Pennsylvania 19129.
Clinical Infectious Diseases 1997;24:103547 1997 by The University of Chicago. All rights reserved. 10584838/97/24060001$02.00

titis, congestive heart failure, metastatic malignant disease, systemic lupus erythematosus, and lymphedema and, rarely, in adults with no underlying disease [3]. The presence of ascites appears to be the common link among these various conditions. The route of infection in primary peritonitis is usually not apparent but is thought to be hematogenous, lymphogenous, via transmural migration through an intact gut wall from the intestinal lumen, or, in women, from the vagina via the fallopian tubes. In cirrhotic patients the hematogenous route is most likely. Organisms removed from circulation by the liver may contaminate hepatic lymph and pass through the permeable lymphatic walls into the ascitic uid. In addition, portosystemic shunting greatly diminishes hepatic clearance of bacteremia, which would tend to perpetuate bacteremia and increase the opportunity to cause metastatic infection at susceptible sites such as the ascitic collection. The infrequency of primary peritonitis in forms of ascites other than that due to liver disease emphasizes the importance of intrahepatic shunting in the pathogenesis of this disease. The hepatic reticuloendothelial system is known to be a major site for removal of bacteria from blood, and animal studies have suggested that destruction of blood-borne bacteria by the reticuloendothelial system is impaired in experimental cirrhosis and in alcoholic liver disease. The decrease in phagocytic activity seen in alcoholic cirrhosis is proportional to the severity of the liver disease. Enteric bacteria may also gain access to the peritoneal cavity by directly traversing the intact intestinal wall. In an animal model, Escherichia coli passes from the bowel into the peritoneal cavity after the introduction of hypertonic solutions into the peritoneum. The infrequent occurrence of bacteremia and the multiplicity of species in peritoneal uid when anaerobic bacteria are involved suggest that transmural migration of bacteria is the probable route of infection of ascitic uid in most of these patients. In prepubertal girls, the pathogenesis of primary peritonitis is likely related to an ascending infection of genital origin, as suggested by the simultaneous presence of pneumococci in vaginal secretions and peritoneal uid. Alkaline vaginal secretions that occur in this age group may be less inhibitory to bacterial growth than the acidic secretions of postpubertal females.

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

1036

Johnson, Baldessarre, and Levison

CID 1997;24 (June)

Transfallopian spread is also suggested by the development of primary peritonitis in women with intrauterine devices. The route of spread in women with gonococcal or chlamydial perihepatitis (Fitz-Hugh Curtis syndrome) is presumably via the fallopian tubes and paracolic gutters to the subphrenic space, but it may also be hematogenous. Although tuberculous peritonitis may result from direct entry into the peritoneal cavity of tubercle bacilli (from the lymph nodes, intestine, or genital tract in patients with active disease of these organs), it is more likely to be disseminated hematogenously from remote foci of tuberculosis, most commonly in the lung. Tuberculous peritonitis may become clinically evident after the initial focus has completely healed. Microbiology. Several decades ago, the organisms reported to cause primary peritonitis in children were Streptococcus pneumoniae and group A streptococci. More recently, the number of nephrotic children with streptococcal peritonitis has declined, and the relative frequency of peritonitis due to gramnegative enteric bacilli and staphylococci has increased. In cirrhotic patients, microorganisms of enteric origin account for 70% of the pathogens. E. coli is the most frequently recovered pathogen, followed by Klebsiella pneumoniae, S. pneumoniae, and other streptococcal species, including enterococci. Staphylococcus aureus is an unusual cause of primary peritonitis, accounting for only 2% 4% of cases in most series, but has been noted especially in patients with erosion of an umbilical hernia. Anaerobes and microaerophilic organisms are infrequently reported [4]. Possible explanations include the intrinsic bacteriostatic activity of ascites against Bacteroides species, the relatively high pO2 of ascitic uid, and the lack of optimal anaerobic bacteriologic techniques used to study patients in the past. The presence of anaerobes correlates strongly with polymicrobial infection. Occasionally, primary peritonitis may also be caused by Neisseria gonorrhoeae, Chlamydia trachomatis, Mycobacterium tuberculosis, or Coccidioides immitis. Patients who have positive cultures of ascitic uid with few leukocytes and no clinical signs of peritonitis are considered to have bacterascites. This may represent early colonization before a host response. Mortality among patients with a low leukocyte response is the same as among those with a greater response. Conversely, several series have identied cases of primary peritonitis with negative ascitic uid cultures, termed culture-negative neutrocytic ascites [5]. Blood cultures are positive for one-third of these patients. The frequency of negative results of ascitic uid cultures may be decreased by inoculating blood culture bottles with ascitic uid at the bedside. Clinical presentation and evaluation. In children, primary peritonitis is an acute febrile illness often confused with acute appendicitis. Fever, abdominal pain, nausea, vomiting, and diarrhea usually occur, with diffuse abdominal tenderness, rebound tenderness, and hypoactive or absent bowel sounds. In cirrhotic patients with primary peritonitis, the clinical manifestations may be atypical. Onset may be insidious, with no nd-

ings of peritoneal irritation. Fever (temperature of 100F) is the most common presenting sign, occurring in 50% 80% of cases and even in the absence of abdominal signs or symptoms. Primary peritonitis should always be considered in the differential diagnosis of decompensation of previously stable chronic liver disease, especially hepatic encephalopathy. Primary tuberculous peritonitis usually is gradual in onset, causing fever, weight loss, malaise, night sweats, and abdominal distension. The abdomen may not be rigid and is often characterized as being doughy on palpation. The ndings at surgery or laparoscopy consist of multiple nodules scattered over the peritoneal surface and omentum. Adhesions and a variable amount of peritoneal uid are usually present. Similarly, C. immitis can cause a granulomatous peritonitis with variable clinical manifestations. Although the diagnosis of primary peritonitis can be established with certainty only after thorough laparotomy to exclude a primary intraabdominal site of infection, it can usually be surmised from examination of the peritoneal uid. Fluid obtained at paracentesis should be analyzed for cell count, differential count, and protein concentration, and a gram stain and culture should be performed. Specimens for culture should be sent in an airless, capped syringe or injected directly into aerobic and anaerobic broth culture media; volumes of 10 mL will increase the yield of cultures [6]. A gram stain of the sediment is diagnostic when positive but is more commonly negative. Bacteremia occurs in up to 75% of patients with primary peritonitis due to aerobic bacteria, but it is rare in those with peritonitis due to anaerobes. Usually the same organisms isolated from the peritoneal uid are recovered from the blood. The ascitic uid protein concentration may be low in primary peritonitis (3.5 g/L) because of hypoalbuminemia and dilution of ascitic uid with transudate from the portal system. The leukocyte count in peritoneal uid usually is 300/mm3 (1,000/mm3 in 85%), with granulocytes predominating in 80% of cases. However, the total leukocyte count of some patients with ascites uncomplicated by infection may be similarly elevated. Indeed, an increase in ascitic leukocyte counts has been noted during diuresis in patients with chronic liver disease. Other parameters of ascitic uid that may help in diagnosing primary bacterial peritonitis are pH and lactate concentration [7]. An ascitic uid pH of 7.35 and a lactate concentration of 25 mg/dL are more specic but less sensitive than a leukocyte count of 500/mm3 for diagnostic purposes; using all three parameters together increases the diagnostic accuracy. In patients with negative ascitic uid cultures, endoscopic or laparoscopic peritoneal examination and biopsy may be necessary. Peritonitis secondary to other intraabdominal causes should be excluded, and specimens appropriate for fungal and mycobacteriologic cultures should be obtained. In children, if gram-negative organisms, a mixed ora, or no organisms are obtained from peritoneal uid, full exploratory laparotomy is

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

CID 1997;24 (June)

Peritonitis Update

1037

generally indicated to rule out possible intraabdominal sources of continuing peritoneal contamination. However, in end-stage cirrhotic patients, exploratory laparotomy may be life-threatening, and the likelihood of nding a primary intraabdominal focus may be small. Surgery for these patients can be deferred while the response to antimicrobial therapy is awaited. Patients with primary peritonitis usually respond within 48 hours to appropriate antimicrobial therapy. The observation of an exponential rate of decline in the number of ascitic uid leukocytes after initiation of antimicrobial therapy for primary peritonitis has been found to help differentiate primary from secondary bacterial peritonitis [8]. In patients with a subacute or chronic course of primary peritonitis, other pathogens must be considered, including M. tuberculosis or C. immitis. The diagnosis of tuberculous peritonitis can usually be made at operation or laparoscopy and conrmed by the histologic characteristics of the peritoneal biopsy specimen and bacteriologic examination of the peritoneal biopsy specimen and uid. The diagnosis of C. immitis peritonitis can be made with a wet mount of ascitic uid, by nding the organism in culture, or on histologic examination. Management and prevention. Because the gram stain is frequently negative in primary bacterial peritonitis, the initial choice of antimicrobial drug is often empirical, based on the most likely pathogens. Some of the third-generation cephalosporin antibiotics have been demonstrated to be as efcacious as the combination of ampicillin plus an aminoglycoside in primary bacterial peritonitis. They also eliminate the risk of nephrotoxicity, which is sufciently frequent in this group of patients to warrant avoidance of aminoglycosides if an equally effective alternative antimicrobial regimen can be used. Other antimicrobial agents such as the broad-spectrum penicillins (e.g., mezlocillin, ticarcillin, and piperacillin), carbapenems (e.g., imipenem), and b-lactam antibiotic/b-lactamase-inhibitor combinations (e.g., ticarcillin-clavulanate and ampicillin-sulbactam) are potential alternatives. If peritonitis develops during hospitalization, the therapeutic regimen, such as administration of an aminoglycoside antibiotic and an antipseudomonal penicillin or cephalosporin in combination, should also be active against Pseudomonas aeruginosa. For those situations in which the gram stain is suggestive of a Bacteroides species or polymicrobial peritonitis is evident, antimicrobials with activity against the Bacteroides fragilis group and other anaerobic organisms should be added (e.g., metronidazole, clindamycin). The antimicrobial regimen can be modied once the results of the culture and susceptibility tests are available. In those cases where there is a strong clinical suspicion of primary bacterial peritonitis but all cultures are sterile, antimicrobial therapy should be continued. Clinical improvement, together with a signicant decline in the ascitic uid leukocyte count, should occur after 24 48 hours of antimicrobial therapy if the diagnosis is correct; failure to respond should prompt an

examination for additional pathological conditions. Antimicrobial therapy is usually continued for 10 14 days if improvement is noted, but short-course therapy for 5 days has been shown to be as efcacious as the longer course in some patients. Administration of intraperitoneal antimicrobials is not necessary. Treatment of primary peritonitis is successful in more than one-half of cirrhotic patients, but because of the underlying liver condition, the overall mortality has been reported as high as 95% in some series. Those patients with the poorest prognosis were found to have renal insufciency, hypothermia, hyperbilirubinemia, and hypoalbuminemia [9]. Treatment of peritonitis caused by gram-positive organisms, as well as of early infections, has been more frequently successful than treatment of gram-negative or late infections. In nephrotic patients with gram-positive infections or in patients who do not have a preterminal underlying illness, the survival rate is 90%. Given the common occurrence and high mortality of primary peritonitis in the setting of cirrhosis with ascites, prevention is a desirable strategy. This is particularly true for patients who are awaiting liver transplantation. Selective decontamination of the gut with oral noroxacin (400 mg daily) has been shown to reduce the incidence of spontaneous bacterial peritonitis. Noroxacin has the disadvantage of selecting for gram-positive organisms, including S. aureus, and quinolone-resistant gramnegative organisms. More recently, trimethoprim-sulfamethoxazole (double-strength, given once daily for 5 days each week) has been shown to reduce the incidence of peritonitis and be well tolerated [10]. A survival-rate advantage has not been demonstrated for any of these preventive regimens. Secondary and Tertiary Peritonitis Pathogenesis. Secondary intraabdominal infection usually is due to spillage of gastrointestinal or genitourinary microorganisms into the peritoneal space as a result of loss of integrity of the mucosal barrier. Examples include appendicitis, diverticulitis, cholecystitis, penetrating wound of the bowel, and perforation of a gastric or duodenal ulcer. Secondary infection is relatively common, taking the form of either generalized peritonitis or localized abscesses. Abscesses may be restricted to the immediate peritoneal space around a diseased intraabdominal organ, such as pericholecystic, periappendiceal, or peridiverticular abscesses, or to certain peritoneal recesses, such as interloop, subdiaphragmatic, subhepatic, lesser sac, or pelvic abscesses. Peritoneal signs suggestive of appendicitis in immunocompromised patients, e.g., patients with AIDS, organ transplant recipients, and those receiving chemotherapy or corticosteroids for neoplasms (especially myelosuppressive drugs), may be due to typhlitis, an inammation of the cecum [11]. Cecal ulceration in these patients may progress to perforation and secondary peritonitis with colonic ora. Perforation complicating penetrat-

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

1038

Johnson, Baldessarre, and Levison

CID 1997;24 (June)

ing cytomegalovirus enterocolitis has been described as a common cause of acute abdomen in patients with AIDS [12]. Tertiary peritonitis has been conceived as a later stage in the disease, when clinical peritonitis and systemic signs of sepsis (e.g., fever, tachycardia, tachypnea, hypotension, elevated cardiac index, low systemic vascular resistance, leukopenia or leukocytosis, and multiorgan failure) persist after treatment for secondary peritonitis and either no organisms or lowvirulence pathogens, such as enterococci and fungi, are isolated from the peritoneal exudate. These organisms may gain access to the peritoneal cavity by contamination during operative interventions, by selection from the initial polymicrobial peritoneal inoculum by antibiotic therapy, or by translocation of bowel ora. Translocation may be promoted by intestinal ischemia, endotoxemia, malnutrition, or proliferation of resistant bowel ora by antibiotic pressure. The cytokine response in peritonitis has been the subject of a recent excellent review [13]. Undoubtedly, many of the systemic as well as abdominal manifestations of peritonitis are mediated by cytokines, such as TNF, IL-1, IL-6, IFN-g, and others. Cytokines appear in the systemic circulation of patients with peritonitis and to a much greater extent in the peritoneal exudate [14]. These cytokines are produced by macrophages and other host cells in response to bacteria or bacterial products, such as endotoxin [15], or by tissues traumatized during the operative procedure [16]. Another potential source is direct translocation of cytokines through the intestinal barrier. Cytokine responses have been studied in the peritoneal exudate in experimental animal models of peritonitis [17 20], in patients with spontaneous bacterial peritonitis [21, 22], in patients undergoing continuous ambulatory peritoneal dialysis (CAPD) [23 25], and in patients with severe secondary bacterial peritonitis who were undergoing planned relaparotomy [14]. Antibodies to TNF have been found to fail to protect against death [26] and failed to reduce serum levels of IL-1 and IL-6 in an experimental model of peritonitis [17, 20, 27]. In contrast, antibodies to endotoxin were found to prevent death in this model, as well as to reduce bacterial numbers in the peritoneal exudate [28]. Antibodies to IFN-g also afforded a protective effect both in this model of experimental peritonitis and following intravenous injection of endotoxin [29]. Microbiology. With gastrointestinal perforation as the precipitating event, the number and types of microorganisms isolated from the peritoneal cavity depend on the level of the perforation. The stomach in the fasting state contains a sparse microora of a few relatively more acid-resistant species, e.g., lactobacilli or Candida species. Similarly, the duodenum and proximal small bowel contain a sparse microora in the fasting state, whereas the colon contains a high microbial density, i.e., about 1012/g, of which 99.9% are obligate anaerobes, mainly of the B. fragilis group. E. coli, the predominant colonic facultative anaerobe, is found at counts of 106 8/g, while enterococci are less numerous at counts of 104 6/g. The characteristic microora, however,

can undergo alteration as a result of the primary disease process or previous antimicrobial therapy. For example, diseases of the stomach that result in obstruction, the loss of gastric acidity (e.g., bleeding, gastric ulcer, or carcinoma), or use of acidreducing drugs may cause gastric colonization with oral anaerobes, such as non-fragilis Bacteroides and Fusobacterium species and other oropharyngeal organisms, e.g., viridans streptococci, microaerophilic streptococci, Candida species, and lactobacilli. Gastric perforation is associated with either sterile chemical peritonitis or peritonitis due to the above-mentioned pathogens, depending on the underlying gastric condition. Similarly, the normal sparse ora of the small bowel may be altered by gastric disease or small-bowel ileus. With perforation of the colon, initially enormous numbers (a total of 1012/mL) of the hundreds of different species that constitute the normal colonic microora spill into the peritoneal cavity. A simplication of the numerous species then occurs, so that once peritoneal infection is established, only about ve pathogens on average remain, usually three anaerobic and two aerobic species. In patients with gangrenous or perforated appendicitis, the average number of organisms recovered from clinical specimens has been reported to be as high as 9.8 and 12.7, respectively, with 75% of the ora consisting of anaerobes [30]. The obligate anaerobes isolated have been found to be more oxygen-tolerant and have identiable virulence factors, as compared with the rest of the obligate anaerobic microora. B. fragilis is the most frequently isolated anaerobe following colonic perforation, and E. coli is the most frequently isolated facultative anaerobe. In a now-classic series of experiments in rodents, Weinstein and co-workers claried the sequence of events that occurs following contamination of the peritoneum with fecal ora [31]. In this model, E. coli is responsible for sepsis and mortality with early peritonitis, and B. fragilis, in concert with E. coli and possibly other microorganisms such as enterococci, is responsible for late peritoneal abscess formation. Such synergy between anaerobes and facultative microorganisms has long been recognized in mixed infections [32]. Some of the microorganisms isolated apparently owe their survival largely to the presence of other bacteria in the polymicrobial infection, such as B. fragilis and E. coli, which if eliminated by specic antimicrobial therapy will result in coincident disappearance of these other microorganisms. The lower intestinal ora can be altered in the severely ill, hospitalized patient under the pressure of antibiotic usage that allows proliferation of multiple-drug-resistant microorganisms, such as P. aeruginosa, Enterobacter species, multiple-drug-resistant enterococci, and Candida species. These microorganisms can then contribute to peritoneal infection that may follow colonic perforation. The same microorganisms may also be isolated from patients with so-called tertiary peritonitis, i.e., persistent peritonitis in patients with impaired host defenses and multiple organ dys-

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

CID 1997;24 (June)

Peritonitis Update

1039

function who are unresponsive to initial management [33]. However, monomicrobial infection with microorganisms that have low pathogenicity, such as Candida species, enterococci, and coagulase-negative staphylococci, has been noted even in acute peritonitis in patients with severely impaired defenses [34]. P. aeruginosa usually is thought of as a nosocomial pathogen that emerges under the selective pressure of broadspectrum antimicrobial agents. In several studies, however, this organism has been isolated in up to 24% of patients with acute community-acquired perforating appendicitis [35, 36]. Although enterococci are found in 20% of intraabdominal infections, their exact role in polymicrobial infection and the need for specic antimicrobial therapy remain controversial. Selective therapy against E. coli and B. fragilis, which has minimal in vitro activity against enterococci, has been found to be sufcient to reduce enterococcal counts [37]. Nevertheless, in animal models of experimental polymicrobial intraabdominal infection, enterococci have been found to be a signicant component of the inoculum, which enhances abscess formation, weight loss, bacteremia with B. fragilis and E. coli, and mortality [37, 38]. Similarly, clinical reports have indicated the emergence of enterococcal abscesses and bacteremia after treatment of intraabdominal sepsis with antimicrobial agents that lack signicant in vitro enterococcal activity [39 41]. In addition, a recent multicenter study of intraabdominal infection has found that the presence of enterococci in the initial culture, in addition to APACHE II score, independently predicted treatment failure with broad-spectrum antimicrobial regimens that lacked specic enterococcal activity [42]. APACHE II score, age, length of preinfection hospital stay, and postoperative infections predicted the presence of enterococci. It is unknown if initial inclusion of antienterococcal therapy will improve outcome for these high-risk patients. Clinical presentation and evaluation. CT has become invaluable in evaluating patients suspected of having an intraabdominal infection. CT- or ultrasonography-guided aspiration of suspected intraabdominal abscesses has also become standard in evaluating and managing selected patients. Intraabdominal infection encompasses many different pathological and clinical entities that are associated with widely varying mortality rates that range from zero to 60%. The ability to predict outcome and stratify severity of disease is important for clinical decision-making as well as ensuring comparability in trials that evaluate different management strategies with surgical protocols or antimicrobial agents [43]. Outcome has been found to be related mainly to host factors (e.g., preoperative nutritional status, organ impairment, the severity of the patients systemic response, and the premorbid physiological reserves, as predicted by the APACHE II scoring system) rather than to type and source of the infection [44, 45]. Death from intraabdominal infection, especially when the tertiary phase is reached, is thought to result from an exaggerated uncontrolled cytokine release that is unresponsive to all

therapeutic attempts [46]. This has led to investigation of endotoxin and cytokine levels in circulation to predict outcome. However, the magnitude of levels of endotoxin, TNF, and IL-6, in circulation in patients with peritonitis, has not been invariably related to prognosis [14, 47 51]. It has been suggested that cytokine levels in the peritoneal exudate, rather than the blood, better reect the severity of the compartmentalized peritoneal infection and predict outcome [14]. Management and prevention. A skeptical attitude is necessary when reviewing reports on clinical trials of antimicrobial therapy for intraabdominal sepsis. Surgical therapy alone may be sufcient to cure many otherwise healthy, young patients with intraabdominal infection who have no signs of severe sepsis; even antimicrobial agents to which the pathogens are resistant may be associated with a good outcome for these patients. To establish a specic role for antimicrobial regimens requires clinical trials with sufcient numbers of high-risk patients whose severity of illness has been stratied by APACHE II scoring. Clinical trials are frequently diluted by low-risk patients, since the sickest patients who would test the efcacy of the antimicrobial agent are often excluded by strict enrollment requirements. Medical management of secondary peritonitis includes use of antimicrobial therapy and supportive measures to maintain vital functions, e.g., to improve or maintain circulation, nutrition, and oxygenation to vital organs. Antimicrobial agents must penetrate to the site of infection in concentrations that are sufcient to overcome the effects of high bacterial density, metabolic inactivity and slow growth rate of the bacterial inoculum, low pH, low redox potential, necrotic tissue, and bacterial products that may lower the drugs activity. For example, aminoglycosides and clindamycin are less active at acid pH values, aminoglycosides are less active at low redox potentials, and b-lactam drugs are less active against high bacterial densities. The animal model of intraabdominal sepsis demonstrated the necessity of treating both the facultative enteric gram-negative bacilli, namely E. coli, and the anaerobic gram-negative bacilli, namely B. fragilis. Empirical use of an antimicrobial regimen active against these organisms has been well established. Indeed, the need for intraoperative cultures to document etiological microbes and their antimicrobial susceptibilities has been questioned, because the results are rarely available to inuence clinical decisions [52]. Early clinical trials have established the therapeutic regimen of clindamycin plus aminoglycoside as the gold standard [53, 54]. Aminoglycosides are frequently underdosed because of fear of nephrotoxicity or because of underestimation of the expanded volume of distribution in critically ill patients with intraabdominal sepsis. Once-daily aminoglycoside therapy in which the total daily dose is given to patients with normal renal function as a single dose every 24 hours rather than as multiple smaller, divided doses obviates these dosing prob-

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

1040

Johnson, Baldessarre, and Levison

CID 1997;24 (June)

lems, because both the bactericidal activity and postantibiotic effect of aminoglycosides are concentration-dependent, while their nephrotoxicity is time-dependent. However, too few severely ill patients with intraabdominal sepsis have been studied to allow recommendation of the general use of single daily doses of these drugs. Regimens in which a third-generation cephalosporin is substituted for the aminoglycoside or metronidazole is substituted for clindamycin compare favorably to the gold standard. Resistance, however, emerges readily under selective pressure of antimicrobial therapy with third-generation cephalosporins among certain gram-negative bacilli that produce inducible b-lactamases, such as Enterobacter, Serratia, Citrobacter, Morganella, and Acinetobacter species and P. aeruginosa. These organisms have a high spontaneous mutation rate for constitutive production of large amounts of these b-lactamases that confer resistance to all b-lactam drugs, except the carbapenems (imipenem and meropenem) and cefepime, and are poorly antagonized by sulbactam, clavulanic acid, and tazobactam. Patients likely to be infected with these organisms e.g., those having prolonged hospital stays, prior antibiotic treatment, postoperative peritonitis, or tertiary peritonitis are best treated with a regimen that includes imipenem, meropenem, cefepime, a uoroquinolone, or an aminoglycoside. Resistance among isolates of B. fragilis to metronidazole is rare, while resistance to clindamycin is now unacceptably high in many centers. Furthermore, the relative incidence of Clostridium difcile associated diarrhea and colonization was found to be less following use of metronidazole than that following use of clindamycin in a retrospective study [55]. Microaerophilic gram-positive cocci, which are frequent copathogens in polymicrobial anaerobic infection, are resistant to metronidazole, unlike clindamycin, so if metronidazole is used it should be combined with an agent active against these pathogens, such as a b-lactam antibiotic other than aztreonam. Aztreonam, uoroquinolones, and aminoglycosides also lack activity against microaerophilic gram-positive cocci and theoretically may not be optimal agents to combine with metronidazole, although clinical trials of these combinations have been found to be efcacious in intraabdominal infection [56]. Monotherapy with agents that have activity against both aerobes and anaerobes includes the carbapenems imipenem and meropenem and the b-lactam/b-lactamase combinations, such as ampicillin/sulbactam, ticarcillin/clavulanate, and piperacillin/tazobactam. Unacceptably high B. fragilis resistance to cefoxitin has been found in many centers; cefotetan, another second-generation cephalosporin, is less active than cefoxitin against certain species in the B. fragilis group. The duration of antimicrobial therapy after adequate surgery is usually 5 7 days but depends on severity of infection, clinical response, and normalization of the WBC count [53, 57, 58]. Only a short course of antimicrobial therapy (24 hours) is required for sterile peritonitis that occurs in the peritoneal space around an infected but resectable intraabdominal organ, such

as the appendix or gallbladder, after resection of the organ. Similarly, contamination of the peritoneum from a defect in the intestinal wall, such as immediately following penetrating abdominal trauma, may also require only operative intervention to remove the diseased organ and a brief course of antimicrobial therapy [53]. Persistent signs of sepsis suggest formation of an intraabdominal abscess that requires drainage, continued contamination of the peritoneum from an inadequately controlled source, superimposed nosocomial infection with a resistant pathogen, or tertiary peritonitis. Treatment against Enterococcus or Candida species involved in polymicrobial infections is controversial. Identication of either microorganism in blood cultures, as the sole organism within residual or recurrent intraabdominal infection, or as the predominant pathogen on a gram stain of peritoneal exudate is an indication for specic antimicrobial therapy plus adequate surgical debridement. Enterococci have emerged as leading nosocomial pathogens, undoubtedly as a result of their inherent resistance to many commonly used antimicrobial agents and the recent acquisition of resistance to previous standard therapy, i.e., with ampicillin, aminoglycosides, and vancomycin. Only clinically unproven agents to which the strains are susceptible in vitro, such as doxycycline, novobiocin, or quinupristin/dalfopristin, are available as potential therapeutic agents for infection caused by these multiresistant strains of enterococci. Administration of amphotericin B is standard therapy for invasive candidal infection, although amphotericin lipid complex is an equally efcacious, less nephrotoxic, but more expensive alternative. Azoles such as uconazole and itraconozole are also less toxic, although no comparative trials for candidal peritonitis have been performed. Candida krusei, Candida tropicalis, and Torulopsis glabrata are inherently more resistant to azoles, and resistance to azoles among previously susceptible Candida species is increasing. Although a comprehensive discussion of surgical management of secondary peritonitis is beyond the scope of this review, certain recent trends will be mentioned [59, 60]. Optimal management includes the following: (1) bowel decompression, e.g., by proximal colostomy for perforation, diverticulitis, or colonic carcinoma; (2) closing of traumatic perforations and resection of a diseased, perforated viscus to stop continued peritoneal contamination; and (3) drainage of any purulent collections to reduce the bacterial inoculum and to remove excessive levels of proinammatory cytokines and other adjuvants, such as fecal matter, food, blood, bile, bullets, or barium. In the absence of perforation, when the disease process, e.g., acute appendicitis or necrotic bowel, is anticipated to progress, the involved organ is resected. Laparoscopic appendectomy has been recommended for acute appendicitis, with conversion to open appendectomy for perforated appendix or an inammatory mass, as well as laparoscopic cholecystectomy for acute cholecystitis (with conversion to an open procedure when technical difculties are encountered).

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

CID 1997;24 (June)

Peritonitis Update

1041

Abscesses that present as localized peritonitis may be drained percutaneously with the aid of ultrasonography or CT, followed by denitive surgery, if necessary. Percutaneous drainage of a peridiverticular abscess may permit a subsequent one-stage procedure of primary resection and immediate anastomosis. Peritonitis caused by colonic perforation due to diverticulitis or colon cancer has been treated with one of three procedures, depending on the particular clinical situation: (1) a three-stage procedure, which involves an initial proximal colostomy to decompress the bowel and divert the fecal stream, followed by resection of the diseased bowel, and nally anastomosis to restore bowel continuity several months later; (2) a two-stage procedure, which involves initial resection and temporary proximal colostomy, with either a mucous stula or a blind pouch (Hartmanns procedure), followed by anastomosis; or (3) a one-stage procedure of primary resection and immediate anastomosis. The three-stage procedure has traditionally been reserved for critically ill, high-risk patients; however, the two-stage procedure may be preferred for these patients because it eliminates the source of peritoneal contamination early [61]. The onestage procedure in the patient with colonic perforation has been questioned because the lumen of the bowel is not cleansed preoperatively and because of the assumed risk of breakdown of the anastomosis in the presence of peritonitis. However, the one-stage primary resection and immediate anastomosis eliminate need for further hospitalization and shorten disability due to a colostomy. Indeed, the one-stage procedure has been shown to be efcacious for the moderate-risk patient with colonic perforation (APACHE II score of 15), even in the case of emergency when the bowel has not been cleansed or peritonitis is present [61]. Intraoperative peritoneal lavage with saline, following drainage of purulent peritoneal exudates, fecal matter, food, and other foreign debris, is standard procedure during laparotomy for peritonitis. However, radical peritoneal debridement of all brinous deposits on peritoneal surfaces is no longer thought to be effective [62]. Continuous peritoneal lavage for 48 72 hours postoperatively or until the uid is clear has been studied, but at present the evidence is insufcient to recommend this procedure. Addition of antibiotics or antiseptics to intraperitoneal lavage uid has also been shown to have no added benet [53]. Similarly, there has been interest in the role of planned relaparotomies to treat patients with severe peritonitis. Here, the commitment is made at the rst operation to perform laparotomies at frequent intervals until the abdomen is macroscopically clean, with additional surgical procedures, such as resections, performed as necessary [63]. The abdominal fascia is left open between laparotomies and the defect bridged by saline-soaked gauze or by a temporary abdominal closure device, such as mesh. These demanding and costly procedures have been complicated by multiple stulas, wound contamination, incisional her-

nias, and tertiary peritonitis with organisms such as enterococci or Candida species [63 65]. Indeed, repeated entry into the inamed peritoneum may further escalate the cytokine cascade. A review concluded that in the absence of randomized controlled prospective trials with appropriate stratication of patients by severity of illness, there is insufcient evidence to determine if these procedures improve outcome of severe diffuse peritonitis [66]. Peritonitis Complicating Peritoneal Dialysis Pathogenesis. Since its development in the late 1970s, CAPD has provided a safe, cost-effective treatment for endstage renal disease. Unfortunately, the technique is often abandoned because of its most common complication, peritonitis [67]. The incidence of peritonitis complicating CAPD varies considerably between individual patients and centers. Rates tend to reect the experience of the center, the specic technology used, and the users susceptibility to infection and ability to comply with procedures [68]. Overall, the average incidence of peritonitis is 1.3 1.4 episodes per patient per year [69]. More than half of these episodes are experienced by only 25% of patients. The two most important routes for development of peritonitis in CAPD patients are (1) transluminal, resulting from a break in sterile technique during dialysate exchange, and (2) contiguous spread, in which microorganisms access the peritoneum along the tract of the dialysis catheter. Less common portals of entry are hematogenous spread from a distant site of infection and direct contamination from the gastrointestinal tract. Diverticular disease of the nonsigmoid colon appears to be a risk factor for acquisition of infection of intestinal origin, presumably through occult microperforations; however, isolation of multiple enteric pathogens from peritoneal uid, especially anaerobic bacteria, suggests fecal contamination from gross colonic perforations [68]. Microbial factors that may contribute to pathogenesis of peritonitis include ability to grow in dialysis uids and production of extracellular slime (biolm). Once organisms gain access to the peritoneal cavity, further growth may depend on their survival in the presence of dialysis uid. Fresh dialysate solutions are capable of supporting growth of E. coli but not that of staphylococci. However, after instillation into the peritoneal cavity, dialysis efuent supports growth of both organisms [69]. Other studies have shown that the growth of P. aeruginosa and E. coli is enhanced 1,000-fold in dialysis uids from patients with peritonitis, compared with that in such uids from uninfected controls [70]. Both staphylococci and Candida albicans grow as microcolonies on polymeric surfaces [71]. Surrounding biolm serves to anchor the organisms and protect them from host defenses and antibiotic activity [72]. The clinical observation that some patients remain free of CAPD-related peritonitis for years suggests that host defense

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

1042

Johnson, Baldessarre, and Levison

CID 1997;24 (June)

factors may also be important in the pathogenesis of this infection [73]. Microbial pathogens that reach the dialyzed peritoneal cavity are removed by three major lines of defense. First, despite the dilution of brinogen and coagulation proteins, brin trapping and sequestration of microorganisms operate efciently in the dialyzed peritoneum. Second, removal of the dialysate serves to eliminate or decrease the inoculum of contaminating organisms. Finally, a complex interplay of opsonization, phagocytosis, and intracellular killing by peritoneal macrophages, mesothelial cells, and neutrophils serves to combat bacterial invasion and prevent infection. Unfortunately, the dialyzed peritoneal cavity is not a supportive milieu for operation of these cellular and immunologic defense mechanisms because of its low pH (5.5 6.0), high osmolarity (300 400 mOsmol/kg), and decreased levels of IgG and complement (1% of their normal levels). A correlation between deciencies in these host defenses and patient risk for peritonitis has not been conclusively identied [73]. Microbiology. In patients undergoing CAPD, peritonitis is usually caused by a single pathogen that originates from the normal ora of the skin or upper respiratory tract. Approximately 60% 70% of cases are caused by gram-positive cocci, 20% 30% by gram-negative bacilli, and the remainder by various other bacteria, fungi, and mycobacteria [74]. Coagulasenegative Staphylococcus is the single most common pathogen, followed by S. aureus and species of Streptococcus. Among the gram-negative organisms, most Enterobacteriaceae species have been associated with CAPD peritonitis, but no single species predominates in all reported series. Although a rare pathogen in other types of peritonitis, P. aeruginosa occurs with relative frequency in CAPD peritonitis (5% 10% of cases). It warrants special note because of its association with signicant morbidity and late complications [75, 76]. Fungi have become an important cause of CAPD-related peritonitis in recent years because of their increasing frequency and problematic management. Although many different fungi have been isolated, including Aspergillus, Mucor, Rhizopus, Alternaria, Fusarium, Penicillium, and Drechslera species, C. albicans accounts for 80% 90% of cases [74, 76]. Risk factors for acquisition of fungal peritonitis are bacterial peritonitis within the preceding month, recent hospitalization, presence of extraperitoneal infection, use of immunosuppressive agents, and concomitant HIV infection, but diabetes mellitus does not pose a special risk [77, 78]. Mycobacteria have been described as pathogens in fewer than 3% of cases of CAPD-related peritonitis, but they may also account for a portion of cases labeled as culture-negative [79]. Overall, 86% of mycobacterial isolates reported in the literature have been of group IV (rapid growers), such as M. fortuitum and M. chelonae [80]. One particularly large outbreak involved 17 patients who developed infection with M. chelonae following treatment with contaminated intermittent peritoneal dialysis machines [81]. Other rare causes of peritonitis in patients undergoing CAPD are viruses, algae, and M. tuberculosis.

Clinical presentation and evaluation. Criteria for the diagnosis of CAPD-associated peritonitis are (1) signs and symptoms of peritoneal irritation, (2) cloudy dialysate efuent with a leukocyte count of 100/mm3, and (3) a positive culture of dialysate uid. Any two of these criteria may be adequate to establish the diagnosis [82]. Clinical manifestations of peritonitis vary from mild to severe, depending largely on the virulence of the pathogen and the time course of the infection [67]. Generally, turbid dialysate is the rst and most common symptom to appear, followed shortly thereafter by abdominal pain and tenderness. Laboratory evaluation of dialysate efuent is critical to establishing the diagnosis of CAPD-associated peritonitis. A dialysate leukocyte count of 100/mm3 is the traditional diagnostic value but is not specic. The differential cell count of dialysate may have a better predictive value. In one study, polymorphonuclear leukocytes comprised 50% of the total cell count (mean, 85%) in infected patients, while in uninfected patients the proportion was 40% (mean, 12%) [82]. A preponderance of eosinophils in the uid is seen in the self-limited condition eosinophilic peritonitis, which often follows placement of the Tenckhoff catheter and may represent allergy to the tubing [83, 84]. Peritoneal eosinophilia also occurs with fungal peritonitis or recent intraperitoneal administration of antibiotics. Gram staining of dialysis uid detects only 20% 30% of peritonitis episodes. Gram-positive organisms, especially S. aureus, are more likely to be detected than are gram-negative ones [85]. Cultures offer a greater yield than gram stains but require concentration of the dialysate by centrifugation, ltration, or lysis-centrifugation [86]. Even with these specialized methods, 3% 30% of CAPDrelated peritonitis episodes are culture-negative. Presumably, most are due to fastidious, low-virulence organisms or coagulase-negative staphylococci that survive less well in dialysis uid [70]. Cases of culture-negative peritonitis that do not respond to empirical antibiotic treatment should be further investigated by performance of dialysate uid cultures for mycobacteria and fungi. Blood cultures are usually negative, regardless of the etiologic agent. Management and prevention. The increased use of intraperitoneal antibiotics as therapy for peritonitis has allowed most patients to be treated on an ambulatory basis. Hospitalization is indicated for those who are severely ill or who are unable to manage administration of intraperitoneal antibiotics at home. A variety of antimicrobial agents have been used successfully for treatment, including penicillins, cephalosporins, aztreonam, imipenem, aminoglycosides, uoroquinolones, and macrolide and glycopeptide antibiotics. Initial therapy may be guided by results of the dialysis uid gram staining, but since these are uncommonly positive, empirical treatment with a cephalosporin or vancomycin plus aminoglycoside is typically initiated. In relatively mild episodes of peritonitis, a single agent with antistaphylococcal activity may also be suitable. Antibiotic se-

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

CID 1997;24 (June)

Peritonitis Update

1043

lections are adjusted when the results of cultures and susceptibility tests are known. A single specic agent is adequate treatment in the majority of cases of bacterial peritonitis. P. aeruginosa, however, has been associated with a high therapeutic failure rate and frequent relapses [75, 76]. A synergistic combination of antibiotics, such as an antipseudomonal b-lactam drug plus an aminoglycoside, has been recommended in addition to removal of the dialysis catheter. Intraperitoneal administration of antibiotics is the preferred method for drug delivery in CAPD-associated peritonitis because it achieves high local concentrations and permits selftreatment by the patient [87 89]. Therapy is usually continued for 10 14 days but may need to be extended with unusually severe or slow-to-respond infections. Recommendations for specic drug doses and routes of administration are the subject of several reviews [89, 90]. Treatment of fungal peritonitis is controversial because of the lack of controlled studies and the small numbers of cases seen at any single center. Although there are reports of successful treatment with intraperitoneal and/or systemic antifungal agents alone [91, 92], removal of the dialysis catheter is usually prudent to prevent relapse [68, 93]. A short course of systemic amphotericin B (250500 mg) is often given following catheter removal. Some evidence has suggested that catheter removal alone may be curative in selected patients. Mycobacterial peritonitis also requires removal of the dialysis catheter for cure. Most of these organisms are resistant to conventional antituberculous agents, and susceptibilities vary greatly among the species. Antibiotic selections should be guided by either in vitro susceptibility test results or published recommendations [7981]. Prevention of peritonitis in patients undergoing CAPD requires intensive education regarding aseptic technique and catheter care. Efforts to reduce the incidence of peritonitis by using oral or intraperitoneal antibiotics have largely been unsuccessful [9497]. Although a decrease in the number of random positive dialysate cultures has been observed in clinical studies, occurrence of clinical peritonitis was unaffected by prophylactic antibiotics. In addition, topical mupirocin has been used to eliminate nasal colonization with S. aureus but has yet to be shown to signicantly impact the incidence of CAPD-related peritonitis [98]. The most signicant advances in the prevention of dialysisrelated peritonitis involve instrumentation changes, including (1) devices that facilitate connection of tubing, such as titanium adapters; (2) devices that help maintain eld sterility during exchanges, such as ultraviolet light systems and in-line lters; and (3) devices that protect intraluminal sterility during exchanges, such as connector systems with disinfectant (Y-connector, O-set). Most such devices have been shown to favorably impact the incidence of peritonitis but add appreciably to the overall cost of CAPD.
References 1. Wilcox CM, Dismukes WE. Spontaneous bacterial peritonitis: a review of pathogenesis, diagnosis, and treatment. Medicine (Baltimore) 1987; 66:447 56.

2. Nohr CW, Marshall DG. Primary peritonitis in children. Can J Surg 1984; 27:179 81. 3. Conn HO, Fessel JM. Spontaneous bacterial peritonitis in cirrhosis: variations on a theme. Medicine (Baltimore) 1971; 50:161 97. 4. Sheckman P, Onderdonk AB, Bartlett JG. Anaerobes in spontaneous peritonitis [letter]. Lancet 1977; 2:1223. 5. Runyon BA, Hoefs JC. Culture-negative neutrocytic ascites: a variant of spontaneous bacterial peritonitis. Hepatology 1984; 4:1209 11. 6. Runyon BA, Umland ET, Merlin T. Inoculation of blood culture bottles with ascitic uid: improved detection of spontaneous bacterial peritonitis. Arch Intern Med 1987; 147:73 5. 7. Stassen WN, McCullough AJ, Bacon BR, et al. Immediate diagnostic criteria for bacterial infection of ascitic uid: evaluation of ascitic uid polymorphonuclear leukocyte count, pH, and lactate concentration, alone and in combination. Gastroenterology 1986; 90:1247 54. 8. Runyon BA, Hoefs JC. Ascitic uid analysis in the differentiation of spontaneous bacterial peritonitis from gastrointestinal tract perforation into ascitic uid. Hepatology 1984; 4:447 50. 9. Weinstein MP, Iannini PB, Stratton CW, Eickhoff TC. Spontaneous bacterial peritonitis: a review of 28 cases with emphasis on improved survival and factors inuencing prognosis. Am J Med 1978; 64:592 8. 10. Singh N, Gayowski T, Yu VL, Wagener MM. Trimethoprim-sulfamethoxazole for the prevention of spontaneous bacterial peritonitis in cirrhosis: a randomized trial. Ann Intern Med 1995; 122:595 8. 11. Till M, Lee N, Soper WD, Murphy RL. Typhlitis in patients with HIV-1 infection. Ann Intern Med 1992; 116:998 1000. 12. Wilson SE, Robinson G, Williams RA, et al. Acquired immune deciency syndrome (AIDS): indications for abdominal surgery, pathology, and outcome. Ann Surg 1989; 210:428 34. 13. Schein M, Wittmann DH, Holzheimer R, Condon RE. Hypothesis: compartmentalization of cytokines in intraabdominal infection. Surgery 1996; 119:694 700. 14. Holzheimer RG, Schein M, Wittmann DH. Inammatory response in peritoneal exudate and plasma of patients undergoing planned relaparotomy for severe secondary peritonitis. Arch Surg 1995; 130:1314 20. 15. Giroir BP. Mediators of septic shock: new approaches for interrupting the endogenous inammatory cascade. Crit Care Med 1993; 21:780 9. 16. Baigrie RJ, Lamont PM, Kwiatowski D, Dallman MJ, Morris PJ. Systemic cytokine response after major surgery. Br J Surg 1992; 79:757 60. 17. Bagby GJ, Plessala KJ, Wilson LA, Thompson JJ, Nelson S. Divergent efcacy of antibody to tumor necrosis factor-a in intravascular and peritonitis models of sepsis. J Infect Dis 1991; 163:83 8. 18. Astiz ME, Saha DC, Carpati CM, Rackow EC. Induction of endotoxin tolerance with monophosphoryl lipid A in peritonitis: importance of localized therapy. J Lab Clin Med 1994; 123:89 93. 19. McMasters KM, Cheadle WG. Regulation of macrophage TNF-a, IL-1 b and 1a (1-Aa) mRNA expression during peritonitis is site dependent. J Surg Res 1993; 54:426 30. 20. Zannetti G, Heumann D, Ge rain J, et al. Cytokine production after intravenous or peritoneal gram-negative bacterial challenge in mice: comparative protective efcacy of antibodies to tumor necrosis factor-a and to lipopolysaccharide. J Immunol 1992; 148:1890 7. 21. Propst T, Propst A, Herold M, et al. Spontaneous bacterial peritonitis is associated with high levels of interleukin-6 and its secondary mediators in ascitic uid. Eur J Clin Invest 1993; 23:832 6. 22. Zeni F, Tardy B, Vindimian M, et al. High levels of tumor necrosis factor-a and interleukin-6 in the ascitic uid of cirrhotic patients with spontaneous bacterial peritonitis. Clin Infect Dis 1993; 17:218 23. 23. Nakahama H, Tanaka Y, Shirai D, et al. Plasma interleukin-6 levels in continuous ambulatory peritoneal dialysis and hemodialysis patients. Nephron 1992; 61:132 4. 24. Zemel D, Krediet RT, Koomen GC, Kortekaas WM, Geertzen HG, ten Berge RJ. Interleukin-8 during peritonitis in patients treated with CAPD: an in-vivo model of acute inammation. Nephrol Dial Transplant 1994; 9:169 74.

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

1044

Johnson, Baldessarre, and Levison

CID 1997;24 (June)

25. Zemel D, Koomen GCM, Hart AA, ten Berge IJM, Struijk DG, Krediet RT. Relationship of TNF-a, interleukin-6, and prostaglandins to peritoneal permeability for macromolecules during longitudinal follow-up of peritonitis in continuous ambulatory peritoneal dialysis. J Lab Clin Med 1993; 122:686 96. 26. Stack AM, Saladino RA, Thompson C, et al. Failure of prophylactic and therapeutic use of a murine anti tumor necrosis factor monoclonal antibody in Escherichia coli sepsis in the rabbit. Crit Care Med 1995; 23:1512 8. 27. Echtenacher B, Falk W, Ma nnel DN, Krammer PH. Requirement of endogenous tumor necrosis factor/cachectin for recovery from experimental peritonitis. J Immunol 1990; 145:3762 6. 28. Battafarano RJ, Burd RS, Kurrelmeyer KM, Ratz CA, Dunn DL. Inhibition of splenic macrophage tumor necrosis factor a secretion in vivo by antilipopolysaccharide monoclonal antibodies. Arch Surg 1994; 129: 179 86. 29. Kohler J, Heumann D, Garotta G, et al. IFN-g involvement in the severity of gram-negative infections in mice. J Immunol 1993; 151:916 21. 30. Bennion RS, Thompson JE, Baron EJ, Finegold SM. Gangrenous and perforated appendicitis with peritonitis treatment and bacteriology. Clin Ther 1990; 12:31 44. 31. Weinstein WN, Onderdonk AB, Bartlett JG, Gorbach SL. Experimental intra-abdominal abscesses in rats: development of an experimental model. Infect Immun 1974; 10:1250 5. 32. Rotstein OD, Pruett TL, Simmons RL. Mechanisms of microbial synergy in polymicrobial surgical infections. Rev Infect Dis 1985; 7:151 70. 33. Rotstein OD, Pruett TL, Simmons RL. Microbiologic features and treatment of persistent peritonitis in patients in the intensive care unit. Can J Surg 1986; 29:247 50. 34. Sawyer RG, Rosenlof LK, Adams RB, May AK, Spengler MD, Pruett TL. Peritonitis in the 1990s: changing pathogens and changing strategies in the critically ill. Am Surgeon 1992; 58:82 7. 35. Aronoff SC, Olson MM, Gauderer MWL, Jacobs MR, Blumer JL, Izant RJ Jr. Pseudomonas aeruginosa as a primary pathogen in children with bacterial peritonitis. J Pediatr Surg 1987; 22:861 4. 36. Yellin AE, Heseltine PNR, Berne TV, et al. The role of Pseudomonas species in patients treated with ampicillin and sulbactam for gangrenous and perforated appendicitis. Surg Obstet Gynecol 1985; 161:303 7. 37. Montravers P, Andremont A, Massias L, Carbon C. Investigation of the potential role of Enterococcus faecalis in the pathophysiology of experimental peritonitis. J Infect Dis 1994; 169:821 30. 38. Matlow AG, Bohnen JMA, Nohr C, Christou N, Meakins J. Pathogenicity of enterococci in a rat model of fecal peritonitis. J Infect Dis 1989; 160: 142 5. 39. Dougherty SH, Flohr AB, Simmons RL. Breakthrough enterococcal septicemia in surgical patients: 19 cases and a review of the literature. Arch Surg 1983; 118:232 8. 40. Jones RN. Gram-positive superinfections following beta-lactam chemotherapy: the signicance of the enterococcus. Infection 1985; 13(suppl 1):S81 8. 41. Barrall DT, Kenney PR, Slotman GJ, Burchard KW. Enterococcal bacteremia in surgical patients. Arch Surg 1985; 120:57 63. 42. Burnett RJ, Haverstock DC, Dellinger EP, et al. Denition of the role of Enterococcus in intraabdominal infection: analysis of a prospective randomized trial. Surgery 1995; 118:716 23. 43. Nystro m P-O, Bax R, Dellinger EP, et al. Proposed denitions for diagnosis, severity scoring, stratication, and outcome for trials on intraabdominal infection. World J Surg 1990; 14:148 58. 44. Pacelli F, Doglietto GB, Aleri S, et al. Prognosis in intra-abdominal infections: multivariate analysis on 604 patients. Arch Surg 1996; 131: 641 5. 45. Christou NV, Barie PS, Dellinger EP, Waymack JP, Stone HH. Surgical Infection Society Intra-abdominal Infection Study: prospective evaluation of management techniques and outcome. Arch Surg 1993; 128: 193 9.

46. Goris RJA, te Boekhorst TPA, Nuytinck JKS, Gimbre ` re JSF. Multipleorgan failure: generalized autodestructive inammation? Arch Surg 1985; 120:1109 15. 47. Damas P, Ledoux D, Nys M, et al. Cytokine serum level during severe sepsis in humans: IL-6 as a marker of severity. Ann Surg 1992; 215: 356 62. 48. Fu tzinger P, et al. Perioperative TNF-alpha and gger R, Zadrobilek E, Go IL-6 concentrations correlate with septic state, organ function, and APACHE II scores in intra-abdominal infection. Eur J Surg 1993; 159: 525 9. 49. Patel RT, Deen KI, Youngs D, Warwick J, Keighley MRB. Interleukin 6 is a prognostic indicator of outcome in severe intra-abdominal sepsis. Br J Surg 1994; 81:1306 8. 50. Hamilton G, Hofbauer S, Hamilton B. Endotoxin, TNF-alpha, interleukin6 and parameters of the cellular immune system in patients with intraabdominal sepsis. Scand J Infect Dis 1992; 24:361 8. 51. Barriere SL, Lowry SF. An overview of mortality risk prediction in sepsis. Crit Care Med 1995; 23:376 93. 52. Wittmann DH, Bergstein JM, Frantzides C. Calculated empiric antimicrobial therapy for mixed surgical infections. Infection 1991; 19(suppl 6): S345 50. 53. Bohnen JMA, Solomkin JS, Dellinger EP, Bjornson HS, Page CP. Guidelines for clinical care: anti-infective agents for intra-abdominal infection: a Surgical Infection Society policy statement. Arch Surg 1992; 127:83 9. 54. Shands JW Jr. Empiric antibiotic therapy of abdominal sepsis and serious perioperative infections. Surg Clin North Am 1993; 73:291 306. 55. Gerding DN, Olson MM, Johnson S, Peterson LR, Lee JT Jr. Clostridium difcile diarrhea and colonization after treatment with abdominal infection regimens containing clindamycin or metronidazole. Am J Surg 1990; 159:212 7. 56. Solomkin JS, Reinhart HH, Dellinger EP, et al. Results of a randomized trial comparing sequential intravenous/oral treatment with ciprooxacin plus metronidazole to imipenem/cilastatin for intra-abdominal infections. Ann Surg 1996; 223:303 15. 57. Lennard ES, Derllinger EP, Wertz MJ, Minshew BH. Implications of leukocytosis and fever at conclusion of antibiotic therapy for intraabdominal sepsis. Ann Surg 1982; 195:19 24. 58. Stone HH, Bourneuf AA, Stinson LD. Reliability of criteria for predicting persistent or recurrent sepsis. Arch Surg 1985; 120:17 20. 59. Wittmann DH, Schein M, Condon RE. Management of secondary peritonitis. Ann Surg 1996; 224:10 8. 60. Nathens AB, Rotstein OD. Therapeutic options in peritonitis. Surg Clin North Am 1994; 74:677 92. 61. Nespoli A, Ravizzini C, Trivella M, Segala M. The choice of surgical procedure for peritonitis due to colonic perforation. Arch Surg 1993; 128:814 8. 62. Polk HC Jr, Fry DE. Radical peritoneal debridement for established peritonitis: the result of a prospective randomized clinical trial. Ann Surg 1980; 192:350 5. 63. Wittmann D, Aprahamian C, Bergstein JM. Etappenlavage: advanced diffuse peritonitis managed by planned multiple laparotomies utilizing zippers, slide fastener, and Velcro analogue for temporary abdominal closure. World J Surg 1990; 14:218 26. 64. Teichmann W, Wittmann DH, Andreone PA. Scheduled reoperations (etappenlavage) for diffuse peritonitis. Arch Surg 1986; 121:147 52. 65. Cuesta MA, Doblas M, Castan eda L, Bengoechea E. Sequential abdominal reexploration with the zipper technique. World J Surg 1991; 15:74 80. 66. Schein M, Hirshberg A, Hashmonai M. Current surgical management of severe intraabdominal infection. Surgery 1992; 112:489 96. 67. Korbet SM, Vonesh EF, Firanek CA. A retrospective assessment of risk factors for peritonitis among an urban CAPD population. Perit Dial Int 1993; 13:126 31. 68. Saklayen MG. CAPD peritonitis: incidence, pathogens, diagnosis, and management. Med Clin North Am 1990; 74:997 1010.

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID

CID 1997;24 (June)

Peritonitis Update

1045

69. Linblad AS, Novak JW, Nolph KD, Stablein DM, Cutler SJ. The 1987 USA National CAPD Registry report. Trans Am Soc Artif Intern Organs 1988; 34:150 6. 70. Sheth NK, Bartell CA, Roth DA. In vitro study of bacterial growth in continuous ambulatory peritoneal dialysis uids. J Clin Microbiol 1986; 23:1096 8. 71. Marrie TJ, Noble MA, Costerton JW. Examination of the morphology of bacteria adhering to peritoneal dialysis catheters by scanning and transmission electron microscopy. J Clin Microbiol 1983; 18:1388 98. 72. Holmes CJ, Evands R. Biolm and foreign body infection the signicance to CAPD associated peritonitis. Peritoneal Dialysis Bulletin 1986; 6:168 77. 73. Holmes CJ. Peritoneal host defense mechanisms in peritoneal dialysis. Kidney Int 1994; 46(suppl 48):S58 70. 74. von Graevenitz A, Amsterdam D. Microbiological aspects of peritonitis associated with continuous ambulatory peritoneal dialysis. Clin Microbiol Rev 1992; 5:36 48. 75. Krothapalli R, Duffy WB, Lacke C, et al. Pseudomonas peritonitis and continuous ambulatory peritoneal dialysis. Arch Intern Med 1982; 142: 1862 3. 76. Kaczmarski EB, Tooth JA, Anastassiades E, Manos J, Gokal R. Pseudomonas peritonitis with continuous ambulatory peritoneal dialysis: six year study. Am J Kidney Dis 1988; 14:413 7. 77. Eisenberg ES, Leviton I, Soeiro R. Fungal peritonitis in patients receiving peritoneal dialysis: experience with eleven patients and review of the literature. Rev Infect Dis 1986; 8:309 21. 78. Dressler R, Peters AT, Lynn RI. Pseudomonal and candidal peritonitis as a complication of continuous ambulatory peritoneal dialysis in human immunodeciency virus infected patients. Am J Med 1989; 86: 787 90. 79. Dunmire RB 3d, Breyer JA. Nontuberculous mycobacterial peritonitis during continuous ambulatory peritoneal dialysis: case report and review of diagnostic and therapeutic strategies. Am J Kidney Dis 1991; 18: 126 30. 80. Hakim A, Hisam N, Reuman PD. Environmental mycobacterial peritonitis complicating peritoneal dialysis: three cases and review. Clin Infect Dis 1993; 16:426 31. 81. Band JD, Ward JI, Fraser DW, et al. Peritonitis due to a Mycobacterium chelonei like organism associated with intermittent chronic peritoneal dialysis. J Infect Dis 1982; 145:9 17. 82. Males BM, Walshe JJ, Amsterdam D. Laboratory indices of clinical peritonitis: total leukocyte count, microscopy, and microbiologic culture of peritoneal dialysis efuent. J Clin Microbiol 1987; 25:2367 71. 83. Digenis GE, Khanna K, Panatlony D. Eosinophilia after implantation of the peritoneal catheter. Peritoneal Dialysis Bulletin 1982; 2:98 9. 84. Gokal R, Ramos JM, Ward MK, Kerr DNS. Eosinophilic peritonitis in continuous ambulatory peritoneal dialysis (CAPD). Clin Nephrol 1981; 15:328 30. 85. Ludlam HA, Price TNC, Berry AJ, Phillips I. Laboratory diagnosis of peritonitis in patients on continuous ambulatory peritoneal dialysis. J Clin Microbiol 1988; 26:1757 62.

86. Bailie GR, Eisele G. Continuous ambulatory peritoneal dialysis: a review of its mechanics, advantages, complications, and areas of controversy. Ann Pharmacother 1992; 26:1409 20. 87. Morse GD, Farolino DF, Apicella MA, Walshe JJ. Comparative study of intraperitoneal and intravenous vancomycin pharmacokinetics during continuous ambulatory peritoneal dialysis. Antimicrob Agents Chemother 1987; 31:173 7. 88. Bennet-Jones D, Wass V, Mawson P, et al. A comparison of intraperitoneal and intravenous/oral antibiotics in CAPD peritonitis. Peritoneal Dialysis Bulletin 1987; 7:31 3. 89. Keane WF, Everett ED, Fine RN, et al. CAPD-related peritonitis management and antibiotic therapy recommendations: Travenol Peritonitis Management Advisory Committee. Peritoneal Dialysis Bulletin 1987; 7:55 62. 90. Keller E, Reetze P, Schollmeyer P. Drug therapy in patients undergoing continuous ambulatory peritoneal dialysis: clinical pharmacokinetic considerations. Clin Pharmacokinet 1990; 18:104 17. 91. Rubin J, Kirchner K, Walsh D, Green M, Bower J. Fungal peritonitis during continuous ambulatory peritoneal dialysis: a report of 17 cases. Am J Kidney Dis 1987; 10:361 8. 92. Venning MC, Ford M, Gould GK. Successful treatment of fungal peritonitis in CAPD using oral uconazole [letter]. Nephrol Dial Transplant 1990; 5:555. 93. Nagappan R, Collins JF, Lee WT. Fungal peritonitis in continuous ambulatory peritoneal dialysis the Auckland experience. Am J Kidney Dis 1992; 20:492 6. 94. Sharma BK, Smith EC, Rodriquez H, Pillay VKG, Gandhi VC, Dunea G. Trial of oral neomycin during peritoneal dialysis. Am J Med Sci 1971; 262:175 8. 95. Axelrod J, Meyers BR, Hirschman SZ, Stein R. Prophylaxis with cephalothin in peritoneal dialysis. Arch Intern Med 1973; 132:368 71. 96. Low DE, Vas SI, Oreopoulos DG, et al. Prophylactic cephalexin ineffective in chronic ambulatory peritoneal dialysis [letter]. Lancet 1980; 2: 753 4. 97. Churchill DN, Taylor DW, Vas SI, et al. Peritonitis in continuous ambulatory peritoneal dialysis patients: a randomized clinical trial of co-trimoxazole prophylaxis. Peritoneal Dialysis Bulletin 1988; 8:125 8. 98. Pe rez-Fonta n M, Rosales M, Rodr guez-Carmona A, et al. Treatment of Staphylococcus aureus nasal carriers in CAPD with mupirocin. Adv Peritoneal Dial 1992; 8:242 5.

The Conict-of-Interest Policy of the Ofce of Continuing Medical Education, UCLA School of Medicine, requires that faculty participating in a CME activity disclose to the audience any relationship with a pharmaceutical or equipment company which might pose a potential, apparent, or real conict of interest with regard to their contribution to the program. The author reports no conict of interest.

/ 9c2f$$ju45

05-07-97 17:21:37

cida

UC: CID