Pediatrics International (2003) 45, 123–129

Invited Paper

Microbiology and management of intra-abdominal infections in children ITZHAK BROOK Department of Pediatrics, Georgetown University School of Medicine, Washington DC, United States of America Abstract

The present review describes the microbiology, diagnosis, and management of intra-abdominal infections in children. Infection generally occurs due to the entry of enteric micro-organisms into the peritoneal cavity through a defect in the wall of the intestine or other viscus as a result of obstruction, infarction, or direct trauma. Mixed aerobic and anaerobic flora can be recovered from the peritoneal cavity of these patients. The predominant aerobic isolates are Escherichia coli, and enterococci; the main anaerobic bacteria are Bacteroides fragilis group, Peptostreptococcus spp. and Clostridium spp. The treatment of abdominal infection includes surgical correction and drainage, and administration of antimicrobials that are effective against both aerobic and anaerobic micro-organisms.

Key words

anaerobic, bacteria, Bacteroides fragilis, Escherichia coli, peritonitis, therapy.

Introduction

Microbiology

Intra-abdominal infections that are the result of secondary peritonitis generally occur due to the entry of enteric microorganisms into the peritoneal cavity through a defect in the intestinal wall or other viscus as a result of infarction, obstruction, or direct trauma. In children, peritonitis is mainly associated with appendicitis, but may also occur with intussusception, incarcerated hernia, volvulus, or rupture of a Meckel’s diverticulum. A rare form of peritonitis in children can occur as a complication of intestinal mucosal disease, including peptic ulcers, ulcerative colitis, and pseudomembranous enterocolitis. Intra-abdominal infection in the neonatal period is often a complication of necrotizing enterocolitis but may also be associated with meconium ileus or spontaneous rupture of the stomach or intestines. The peritonitis that follows introduction of the enteric flora to the peritoneal cavity is usually a synergistic polymicrobial. Generally, the more types of bacteria that participate in the infection, the graver the morbidity. The specific micro-organisms involved in peritonitis generally are those of the normal gastrointestinal tract (GIT) flora at the site of entry to the peritoneal cavity.

Anaerobic bacteria are the main component of the gastrointestinal tract flora,1 where they outnumber aerobic and facultative bacteria in the ratio 1:1000–1:10 000.1 This explains their predominance in infections that are associated with bowel perforation such as perforated appendicitis, inflammatory bowel disease with perforation, and perforation following GIT surgery. Mixed aerobic and anaerobic flora were recovered in the peritoneal cavity of adults with ruptured appendix or intestinal viscus,2 and were also occasionally recovered from post-operative wounds.3 A few retrospective studies evaluated the microbiology of the peritoneal cavity and post-operative wounds following perforated appendix in children.4,5 Anaerobes were recovered from most of the peritoneal cavity specimens and from twothirds of the complicating wounds in children who underwent surgery for perforation of the appendix or other viscus.4 Clostridium spp. were isolated from 43%, and Bacteroides fragilis from 93% of the peritoneal fluids, along with aerobic Gram-negative bacteria and enterococci. Similar isolates were also isolated from liver, pelvic and subphrenic abscesses, surgical wounds, and blood cultures of patients.5 Brook, in a prospective study,6 evaluated peritoneal cavity specimens from 100 children who had a ruptured appendix (Table 1). Specimens were also collected from 11 of these patients who developed post-operative surgical wound infection. Polymicrobial aerobic and anaerobic flora were

Correspondence: Itzhak Brook, MD, MSc, 4431 Albemarle St. NW, Washington DC, 20016, USA. Email: [email protected] Received 18 September 2002.

124

I Brook

Table 1 Predominant organisms isolated from peritoneal fluid from 100 patients with perforated appendix and 11 patients with postoperative wound infection6 † Aerobic and facultative isolates

Gram-positive cocci (total) Group D enterococcus Gram-positive bacilli Gram-negative bacilli (total) Pseudomonas aeruginosa Escherichia coli Klebsiella pneumoniae Total number of aerobes and facultatives †

No. isolates (No from wound infection)

Anaerobic isolates

No. isolates (No from wound infection)

53 (6) 12 (1) 4 (1) 87 (10) 9 (3) 57 (6) 7

Gram-positive cocci (total) Gram-positive bacilli (total) Clostridium sp. Gram-negative bacilli Fusobacterium spp. Bacteroides spp. Pigmented Prevotella and Porphyromonas spp. B. fragilis group Total number of anaerobes

62 (9) 52 (7) 16 (2)

144 (17)

27 (3) 32 (4) 26 (2) 102 (8) 301 (33)

Only the important pathogens are listed in detail. The total number of the groups of organisms is represented.

isolated from both peritoneal cavity and post-operative wound flora of these children. Anaerobic bacteria alone were found in 14 specimens, aerobes alone in 12, and mixed aerobic and anaerobic flora in 74. There were 144 aerobic isolates (1.4 per specimen) and 301 anaerobic isolates (3 per specimen). The predominant aerobes were Escherichia coli, alpha-hemolytic streptococci, gamma-hemolytic streptococci, group D enterococcus, and Pseudomonas aeruginosa. The predominant anaerobes were Gram-negative bacilli. (Bacteroides fragilis group and pigmented Prevotella and Porphyromonas), Gram-positive anaerobic cocci, Fusobacterium spp., and Clostridium spp. The combination of B. fragilis and E. coli occurred in 43 instances, and B. fragilis and Peptostreptococcus spp. in 23. Beta-lactamase production was detectable in 108 isolates recoverted from 78 patients. These included all isolates of B. fragilis group and six of the 37 other Bacteroides spp. Forty-nine isolates (16 aerobic and 33 anaerobic) were recovered from the 11 post-surgical wounds and were predominantly B. fragilis group, E. coli, Peptostreptococcus spp., and P. aeruginosa. Similar isolates were also recovered from the majority of the peritoneal cavity of the children. The microbiology of peritonitis is different in newborns than in older children. Mollitt et al.7 recovered fewer anaerobes in peritonitis associated with necrotizing enterocolitis (NEC) as compared with their isolation from perforated appendix in older children. The main aerobes were Klebsiella, Enterobacter, and Streptococcus spp. Clostridium difficile was isolated from peritoneal fluid in newborns who had peritonitis associated with NEC8 or obstruction.9 The recovery of C. difficile in this age group probably related to its presence as part of the normal GIT flora in newborns.8 Peritoneal dialysis associated peritonitis has higher incidence in children than adults.10,11 The isolates in this type of peritonitis are Gram-positive and Gram-negative aerobic bacteria as well as fungi. These include Staphylococcus aureus, Enterobacteriaceae, Streptococcus spp., Pseudomonas

spp., Enterococci and Candida. Peritoneal dialysis-associated peritonitis tends to be less severe than post perforation infection and is more amenable to treatment. However, fungal infections, which account for 2–6% of episodes,10,11 are more severe and often cause extensive adhesions.

Pathogenesis The dynamics and changes in the microbial flora of the GIT influence the nature and severity of infections that follow perforation. The number of micro-organisms increases at the distal portions of the GIT. At the stomach and upper bowel there are 104 organisms/g or fewer, at the lower ileum the number increases to up to 108 organisms/g, and at the colon it reaches up to 1011 organisms/g,1 most of which are anaerobes. These slow changes are believed to be caused by a variety of factors, including the detrimental effect of the low pH of the stomach, which decreases the number of bacteria ingested from the oral cavity. These conditions are slowly balanced by the alkaline environment of the lower intestine, the effect of bile, and the decrease in oxygen tension in the lower intestine. The changes support and favor the growth of anaerobic bacteria. A greater number of organisms per gram in the upper intestine can be present in patients with decreased stomach acidity or in those with a shorter GIT or anastomosis. The changes in the number of intestinal bacteria account for some of the differences observed in cultures of the peritoneal cavity after perforations. An average of three isolates per specimen and about 107 organisms/g were recovered in perforation of the small intestine, and 26 bacterial isolates and 1012 organisms/g were found in specimens of colonic perforation. The presence of the greater number of organisms in the distal portion of the colon can explain why infection developed in 45% of patients with descending-colon injuries, as compared with only 13% in other colon sites.12

Intra-abdominal infections Peritonitis is a synergistic infection caused by various aerobic and anaerobic bacteria. The two types of organisms have opposite and therefore complementary oxygen requirements, and the changes each causes in its environment allows for the rapid proliferation of their other partners.13,14 Studies that utilized appropriate culture techniques documented that the great majority of intra-abdominal infections are symbiotic.13,15 Post-surgical morbidity increases as more types of bacteria are present.13,14 The synergistic relationship between the aerobic and anaerobic bacteria in intra-abdominal infections has been demonstrated.13,14 Altemeier13 illustrated the pathogenicity of bacterial isolates recovered from peritoneal cavity after appendiceal rupture. Individual isolates were relatively innocuous when implanted subcutaneously in animals, but combinations of facultative and anaerobic strains were more virulent. Similar observations were reported by Meleney et al.,14 Hite et al.,16 and Brook et al.17,18 The number of bacterial species present in infections associated with colonic perforation is five to 10, even though more than 400 bacterial species reside in the colon.19 From the multiple anaerobic bacteria present in the normal flora, only a few are common in septic processes. It is likely that virulence is a critical factor in determining their selection. The dominant anaerobic bacteria in intra abdominal infection include B. fragilis group, pigmented Prevotella and Porphyromonas, Fusobacterium nucleatum, Clostridium perfringens, Peptostreptococcus anaerobius, and Peptostreptococcus asaccharolyticus. These species account for the majority of anaerobic isolates in clinical laboratories.20 A newly recognized species in intra-abdominal infection is Bilophila wadsworthia. Because of its fastidious growth it is often missed in bacterial cultures.21 Bacteroides fragilis is the most common anaerobe in intra-abdominal sepsis or bacteremia. The B. fragilis group possesses several virulence factors. These include resistance to beta-lactam antibiotics through the production of the enzyme beta-lactamase,22 expression of a capsule that inhibits phagocytosis,23 and elaboration of enzymes and metabolic by-products that increases their virulence. An example for an important metabolic by-product is succinic acid that can reduce polymorphonuclear migration.24 Organisms of the B. fragilis group share many phenotypic characteristics, including resistance to penicillins. These organisms were subdivided until 1976 into at least six subspecies: fragilis, distasonis, vulgatus, thetaiotaomicron, ovatus and uniformis. They were promoted to a species level in 1976.25 Their separation into species is based on minor variations in their biochemical reactions and differences in DNA. The distribution of the members of B. fragilis group is significantly different in normal flora and the site of infection. In the colon, the predominant strains of B. fragilis

125

group a septic processes are B. distasonis, B. vulgatus, and B. thetaiotaomicron; while B. fragilis accounts for only about 0.5% of the colonic microflora.16 In clinical specimens, however, B. fragilis is the most often encountered species. The predominance of B. fragilis in infection sites suggests that it has unique virulence properties. The presence of a capsule is one of the important virulence factors of B. fragilis group. The addition of an aerobic organism was required to form an abscess in mice, when unencapsulated B. fragilis were injected intraabdominally. In contrast, encapsulated B. fragilis produced an infection in animals even when injected alone.26 Heatkilled encapsulated or purified capsular material of B. fragilis can also produce abscesses indistinguishable from those resulting from infection with viable organisms. The capsule enhances the virulence B. fragilis through the inhibition of phagocytosis.26 It also enables the organism to induce abscesses even when injected alone,27 making it a greater contributor to the infectious process than its nonencapsulated counterparts in mixed infection.28 Most Bacteroides spp. and Peptostreptococcus spp. do not produce a capsule when they colonize mucous surfaces.29 However, up to 75% of the abscesses harbor encapsulated strains of these organisms. The ability to possess a capsule seems to be expressed only in an inflammatory process. The process of emergence of a capsule has been shown to occur in animal models.30 The chain of events that follows gut perforation can be divided into two stages. The initial stage evolves generalized peritonitis that is often associated with bacteremia. The bacteria that induce the peritonitis originate from the bacterial flora at the site of perforation, and include multiple aerobic and anaerobic bacteria that spill into the peritoneal cavity.6 The early stage may last up to 1 week, and is followed by localization of the infection and formation of abscesses which can be retroperitoneal, peritoneal, or within viscera. The Weinstein and Onderdonk group were the first to study the pathogenicity and principles of management of infection after perforated viscus.31–34 Peritonitis was induced by introducing capsules of cecal material into the abdominal cavity of rats. A biphasic disease emerged, and about 40% of the rodents died within the first week from peritonitis and sepsis, and 100% of the survivors developed intra-abdominal abscesses. Escherichia coli was isolated during the peritonitis stage, and B. fragilis was recovered in the abscesses. When gentamicin that is effective against only E. coli was given, mortality was reduced to 4%, but abscess formation was unchanged. When clindamycin that is active only against B. fragilis was delivered, mortality stayed the same, but abscess was prevented. Only the combination of gentamicin and clindamycin reduced both mortality and morbidity. Other single drugs or drug combinations were effective only when they were active against both Enterobacteriaceae and

126

I Brook

B. fragilis group.33 These included agents such as cefoxitin, carbapenems, combination of penicillin and beta-lactamase inhibitors, and the newer quinolones (i.e. trovafloxacin).32 These findings demonstrate that anaerobes may account for the complications that follow abdominal perforation (i.e., intra-abdominal abscess), and that optimal management of intestinal perforation requires a drug(s) effective against both aerobic and anaerobic bacteria. Studies that utilized subcutaneous abscesses induced by B. fragilis alone highlighted the differences between the antimicrobial’s killing ability and the effect of delay of beginning of therapy. Metronidazole reduced the bacterial counts of B. fragilis in 6.7 log, clindamycin in 5 log, moxalactam in 3.8 log, and cefoxitin in 3.5 log.34,35 Chloramphenicol, carbenicillin, and cephalothin were less cidal. Delays in initiation of therapy minimized the efficacy of clindamycin and cefoxitin but were less significant with metronidazole. However, when fewer organisms were used to induce abscesses, clindamycin or cefoxitin were still effective even after a delay of 24 h. These data illustrate the in vivo differences of various antimicrobials, the importance of starting therapy as soon as possible, and the efficacy of surgical drainage in facilitating the effect of antimicrobials. Administration of single agents that are effective against both aerobic and anaerobic bacteria (i.e., carbapenems), or the newer quinolones, were also effective against both aerobic and anaerobic bacteria.36,37

Management Surgical correction and drainage are of utmost importance in the treatment of abdominal infection. The surgical intervention should be performed as soon as possible, preferably after the patient is stabilized. Stabilization of the patient is achieved by correcting any fluid and electrolyte deficiencies with parenteral fluids, and by alleviation of intestinal obstruction, and controlling the peritoneal infection with antibiotics. The proper management of mixed aerobic and anaerobic infections requires the administration of antimicrobials that are effective against both aerobic and anaerobic components of the infection.31 If such therapy is not delivered, the infection may progress, and more serious complications may develop.15 Failure of the antimicrobial therapy may occur because of the development of bacterial resistance, achievement of insufficient tissue levels, incompatible drug interaction, or the development of an abscess. The abscess capsule can interfere with the penetration of some antimicrobial agents, and the low pH and the presence of binding proteins or antimicrobial inactivating enzymes (beta-lactamases) can impair the activity of many antimicrobials. The low pH and the anaerobic environment within the abscess are especially detrimental toward the

aminoglycosides and quinolones.38 However, an acidic environment, presence of an anaerobic environment, and high osmolarity can also develop in an infectious site without the presence of an abscess. These conditions are especially deleterious for aminoglycosides that depend on oxygen to penetrate bacteria. In contrast, beta-lactam antibiotics are unimpaired in this environment. Antibiotics effective in vivo against B. fragilis group include clindamycin, cefoxitin, metronidazole,30,33,39,40 the combination of a penicillin and a beta lactamase inhibitor (e.g. ticarcillin and clavulanic acid), and carbapenems (e.g. imipenem, meropenem) and some of the newer quinolones (i.e. trovafloxacin). Because some anaerobic bacteria, especially strains of B. fragilis group, might acquire resistance to one of these antibiotics, susceptibility-testing of these organisms should be done in serious infections. Using antimicrobial coverage effective against both the aerobic and anaerobic bacteria involved in intra-abdominal infections has become the cornerstone of practice, and has been confirmed by numerous studies.40,41 These studies employed combination therapies of metronidazole, clindamycin, or cefoxitin directed against anaerobes, and aminoglycosides, fourth-generation cephalosparin; (e.g. cefepime, ceftazidine) and quinolones42 directed at the Enterobacteriaceae. Equal efficacy with most therapies was evident whenever therapies were adequately effective against both B. fragilis group and Enterobacteriaceae. Triple-agent therapy that includes also ampicillin to cover Enterococcus spp. is advocated by some.40,41 The efficacy of single-agent therapy was demonstrated in the management of intra-abdominal infection following trauma. Single-agent therapy with either cefoxitin or moxalactam was found to be equally effective as clindamycin plus an aminoglycoside,43–48 and superior to cephalosporins less active against B. fragilis group.44–46 Other studies illustrated that single-agent therapy with carbapenems (i.e. imipenem, meropenem) or a penicillin plus a beta lactamase inhibitor (i.e., ticarcillin-clavulate, piperacillin-tazobactam) were at least as effective as combination therapies.42,49–54 The advantage of single-agent therapy is avoiding the ototoxicity and nephrotoxicity of aminoglycosides, and may also be less expensive. Single-agent may, however, not always be effective against hospital-acquired, resistant bacteria. Guidelines for the selection of antibiotic therapy for intraabdominal infections have been developed by the Surgical Infection Society Executive Council.53 These recommendations are founded on in vitro activity against enteric organisms, experience in animal models, and documented efficacy in clinical trials. The pharmacokinetics, mechanisms of action, microbial resistance, and safety issues were also used in the formation of these guidelines. Single-agent therapy with cefoxitin, cefotetan, or cefmetazole or ticarcillinclavulanic acid is recommended for community acquired

Intra-abdominal infections infections of mild to moderate severity. Single-agent therapy with carbapenems (i.e., imipenem), or combination therapy with either a third-generation cephalosporin, a monobactam (aztreonam), or an aminoglycoside plus clindamycin or metronidazole is recommended for more severe infections. Antimicrobial regimens with minimal or no activity against facultative Gram-negative bacilli or anaerobic Gram-negative rods are not considered acceptable. Recent work also demonstrated the ability to combine antianaerobic coverage with a quinolone42,51 or a fourth-generation cephalosporin.53 Prophylactic antimicrobial therapy prior to surgery can ensure adequate tissue levels of effective agents at the time of the procedure. The choice of antimicrobial therapy depends on the site of intestinal perforation. In perforation of the upper part of the GIT (that generally does not involve Enterobacteriaceae and B. fragilis group), a first-generation cephalosporin such as cefazolin is adequate. However, in perforation of the lower intestinal lumen or in those with perforation of the upper GIT who are at risk of harboring these pathogens, agents effective against Enterobacteriaceae and B. fragilis group should be administered as prophylaxis, as well as for therapy. The recommended duration of antimicrobials therapy following intra-abdominal perforation varies.54 It should be adjusted according to the patient’s response to therapy, and the patient’s age, immune status, general health, type of injury, degree of contamination, and the effect of any delay in initiation of surgical and medical therapy. Parameters such as serial erythrocyte sedimentation rate or CT scans can influence the duration of therapy. When perforation has not been demonstrated, the duration of antimicrobial therapy can be short (5–7 days). However, if perforation of GIT has occurred, and especially if intraabdominal abscesses are present, therapy should be prolonged to 2–6 weeks. Parenteral therapy is administered in the earlier stages of therapy, and can be switched to oral medication if an extension of therapy is needed. This can be accomplished after oral feeding has been resumed.55 Complications

The complications that can develop following perforation of viscus include septic shock, respiratory failure, retroperitoneal or intra-abdominal abscesses, small bowel obstruction from adhesions, fistula formation, and infection of the postsurgical wounds. Anaerobic bacteria are major pathogens in these infections, and were recovered from all of these infection sites.

Conclusions Intra-abdominal infections generally emerge following the entry of enteric micro-organisms into the peritoneal cavity

127

through a defect in the intestinal wall or other viscus as a result of obstruction, infarction, or direct trauma. The specific micro-organisms involved in peritonitis generally are those of the normal flora of the gastrointestinal tract at the site of entery to the peritoneal cavity. Because anaerobic bacteria outnumber aerobes in the ratio 1:1000–1:10 000, they predominate in these infections. Polymicrobial infection due to aerobic and anaerobic flora can be recovered from the peritoneal cavity of these patients. The predominant aerobic isolates are E. coli, and enterococci, and anaerobic are B. fragilis group Peptostreptococcus spp. and Clostridium spp. The treatment of intra-abdominal infection always includes surgical correction and drainage, and administration of antimicrobials effective against both aerobic and anaerobic micro-organisms. Proper and judicious use of surgical and medical therapy can reduce the rate of complications.

References 1 Gorbach SL. Intestinal microflora. Gastroenterology 1971; 60: 1110–29. 2 Stone HH, Kolb LD, Geheber CE. Incidence and significance of interperitoneal anaerobic bacteria. Ann. Surg. 1975; 181: 705–15. 3 Sanderson PJ, Wren MWP, Baldwin AWF. Anaerobic organisms in postoperative wounds. J. Clin. Pathol. 1979; 32: 143–7. 4 Marchildon MB, Dudgeon DL. Perforating appendicitis: a current experience in children’s hospital. Ann. Surg. 1977; 185: 84–7. 5 Stone HH. Bacterial flora of appendicitis in children. J. Pediatr. Surg. 1976; 11: 37–42. 6 Brook I. Bacterial studies of peritoneal cavity and postoperative wound infection following perforated appendix in children. Ann. Surg. 1980; 192: 208–12. 7 Mollitt DL, Tepas JJ, Talbert JL. The microbiology of neonatal peritonitis. Arch. Surg. 1988; 123: 176–9. 8 Brook I, Avery G, Glasgow A. Clostridium difficile in pediatric patients. J. Infect. 1982; 4: 253–7. 9 Genta VM, Gilligan PH, McCarthy LR. Clostridium difficile peritonitis in a neonate. Arch. Pathol. Laboratory Med. 1984; 108: 82–3. 10 Mocan. H, Murphy AV, Beattie TJ, McAllister TA. Peritonitis in children on continuous ambulatory peritoneal dialysis. J. Infection 1988; 16: 243–51. 11 Spencer RC. Infections in continuous ambulatory peritoneal dialysis. J. Med. Microbiol. 1988; 27: 1–9. 12 Mandal AK, Thadepalli H, Matory E, Lou MA, O’Donnell VA Jr. Evaluation of antibiotic therapy and surgical techniques in areas of homicidal wounds of the colon. Am. Surg. 1984; 50: 254–7. 13 Altemeier WA. The bacterial flora of acute perforated appendicitis with peritonitis. Ann. Surg. 1938; 107: 517–28. 14 Meleney FL, Olpp J, Harvey HD, Jern HZ, Peritonitis. II. Synergism of bacteria commonly found in peritoneal exudates. Arch. Surg. 1932; 25: 709–21. 15 Thadepalli H, Gorbach SL, Broido PW, Norsen J, Nyhus L. Abdominal trauma, anaerobes and antibiotics. Surg. Gynecol. Obstet. 1973; 137: 270–6.

128

I Brook

16 Hite KE, Locke M, Hesseltine HC. Synergism in experimental infections with nonsporulating anaerobic bacteria. J. Infect. Dis. 1949; 84: 1–9. 17 Brook I, Hunter V, Walker RI. Synergistic effects of anaerobic cocci, Bacteroides, Clostridium, Fusobacterium, and aerobic bacteria on mouse mortality and induction of subcutaneous abscess. J. Infect. Dis. 1984; 149: 924–8. 18 Brook I. Enhancement of growth of aerobic and facultative bacteria in mixed infections with Bacteroides species. Infect. Immun. 1985; 50: 929–31. 19 Finegold SM. Anaerobic bacteria in human disease. Academic Press, New York, 1977. 20 Finegold SM, Shepard WE, Spaulding EH. Practical anaerobic bacteriology. In: CUMITECH, no. 5. American Society for Microbiology, Washington, D.C., 1977. 21 Baron EJ, Curren M, Henderson G et al. Bilophila wadsworthia isolates from clinical specimens. J. Clin. Microbiol. 1992; 30: 1882–4. 22 Brook I, Calhoun L, Yocum P. Beta-lactamase-producing isolates of Bacteroides species from children. Antimicrob. Agents Chemother. 1980; 18: 164–6. 23 Tofte RW, Peterson PK, Schmeling D, Bracke J, Kim Y, Quie PG. Opsonization of four Bacteroides species. Role of the classical complement pathway and immunoglobulin. Infect. Immun. 1980; 27: 784–92. 24 Rotstein D, Pruett TL, Fiegel VD, Nelson RD, Simmons RL. Succinic acid, a metabolic by-product of Bacteroides species, inhibits polymorphonuclear leukocytes function. Infect. Immun. 1985; 48: 402–8. 25 Cato EP, Johnson JL. Reinstatement of species rank for Bacteroides fragilis. B. ovatus, B. distasonis, B. thetaiotoamicron, and B. vulgatus. Intern. J. System. Bacteriol. 1976; 26: 230. 26 Onderdonk AB, Kasper DL, Cisneros RL, Bartlett JG. The capsular polysaccharide of B. fragilis as a virulence factor: comparison of the pathogenetic potential of encapsulated and unencapsulated strains. J. Infect. Dis. 1977; 136: 82–9. 27 Brook I, Walker RI. Infectivity of organisms recovered from polymicrobial abscesses. Infect. Immun. 1983; 41: 986–9. 28 Brook I, Walker RI. Significance of encapsulated Bacteroides melaninogenicus and Bacteroides fragilis groups in mixed infections. Infect. Immun. 1984; 44: 12–15. 29 Brook I. Isolation of casulate anaerobic bacteria from orofacial abscesses. J. Med. Microbiol. 1986; 22: 171–4. 30 Brook I, Gillmore JD, Coolbaugh JC, Walker RI. Pathogenicity of encapsulated Bacteroides melaninogenicus group, Bacteroides oralis and Bacteroides ruminicola in abscesses in mice. J. Infect. 1983; 7: 218–26. 31 Weinstein WM, Onderdonk AB, Bartlett JG, Gorbach SL. Experimental intraabdominal abscesses in rats. I. Development of an experimental mode. Infect. Immun. 1974; 10: 1250–5. 32 Onderdonk AB, Weinstein WM, Sullivan NM, Bartlett JG, Gorbach SL. Experimental intra abdominal abscesses in rats. II. Quantitative bacteriology of infected animals. Infect. Immun. 1974; 10: 1256–9. 33 Weinstein AB, Onderdonk AB, Bartlett JG, Louie TJ, Gorbach SL. Antimicrobial therapy of experimental intra abdominal sepsis. J. Inf. Dis. 1975; 132: 282–6. 34 Bartlett JG, Onderdonk AB, Louie T, Kasper DL, Gorbach SL. Lessons from an animal model of intra abdominal sepsis. Arch. Surg. 1978; 113: 853–7. 35 Brook I. In vivo efficacies of quinolones and clindamycin for treatment of infections with Bacteroides fragilis and/or

36

37

38 39

40

41

42 43

44

45

46

47

48

49

50

51

52

53

Escherichia coli in mice: correlation with in vitro susceptibilities. Antimicrob Agents Chemother 1993; 37: 997–1000. Joiner KA, Lowe B, Dzink J, Bartlett JG. Comparative efficacy of 10 antimicrobial agents in experimental infections with Bacteroides fragilis. J. Infect. Dis. 1982; 165: 561–8. Brook I, Ledney GD. The treatment of irradiated mice with polymicrobial infection with Bacteroides fragilis and Escherichia coli. J. Antimicrob. Chemother. 1994; 33: 243–52. Verklin RM, Mandell GL. Alteration of antibiotics by anaerobiosis. J. Laboratory Clin. Med. 1977; 89: 65–71. Sutter VL, Finegold SM. Susceptibility of anaerobic bacteria to 23 antimicrobial agents. Antimicrob. Agent Chemother. 1976; 10: 736–52. Bartlett JG, Louie TJ, Gorbach SL, Onderdonk AB. Therapeutic efficacy of 29 antimicrobial regimens in experimental intraabdominal sepsis. Rev. Infect. Dis. 1981; 3: 535–42. Solomkin JS, Meakins JL Jr, Allo MD, Dellinger EP, Simmons RL. Antibiotic trials in intra-abdominal infections: a critical evaluation of study design and outcome reporting. Ann. Surg. 1984; 200: 29–39. Bartlett JG. Recent developments in the management of anaerobic infection. Rev. Inf. Dis. 1983; 5: 235–45. Hofstetter SR, Pachter HL, Bailey AA, Coppa GF. A prospective comparison of two regimens of prophylactic antibiotics in abdominal trauma: cefoxitin versus triple drug. J. Trauma 1984; 24: 307–10. Gentry LO, Feliciano DV, Lea AS, Short HD, Mattox KL, Jordan GL Jr. Perioperative antibiotic therapy for penetrating injuries of the abdomen. Ann. Surg. 1984; 200: 561–6. Nichols RL, Smith JW, Klein DB, Trunkey DD, Cooper RH, Adinolfi MF, Mills J. Risk of infection after penetrating abdominal trauma. N. Eng. J. Med. 1984; 311: 1065–70. Jones RC, Thal ER, Johnson NA, Gollihar LN. Evaluation of antibiotic therapy following penetrating abdominal trauma. Ann. Surg. 1985; 201: 576–85. Malangoni MA, Condon RE, Spiegel CA. Treatment of intraabdominal infections as appropriate with single-agent or combination antibiotic therapy. Surg. 1985; 98: 648–55. Tally FP, McGowan K, Kellum JM, Gorbach SL, O’Donnell TF. A randomized comparison of cefoxitin with or without amikacin and clindamycin plus amikacin in surgical sepsis. Ann. Surg. 1981; 193: 318–23. Solomkin JS, Dellinger EP, Christou NV et al. Results of a multicenter trial comparing imipenem/cilastatin to tobramycin/ clindamycin for intra-abdominal infections. Ann. Surg. 1990; 212: 581–91. Dougherty SH, Sirinek KR, Schauer PR et al. Ticarcillin/ clavulanate compared with clindamycin/gentamicin (with or without ampicillin) for the treatment of intra-abdominal infections in pediatric and adult patients. Am. Surg. 1995; 61: 297–303. Solomkin JS, Reinhart HH, Dellinger EP et al. Results of a randomized trial comparing sequential intravenous/oral treatment with ciprofloxacin plus metronidazole to imipenem/ cilastatin for intra-abdominal infections. The Intra-Abdominal Infection Study Group. Ann. Surg. 1996; 223: 303–15. Nord CE. The treatment of severe intra-abdominal infections. the role of piperacillin/tazobactam. Intensive Care Med. 1994; 20 (Suppl. 3): S35–8. Bohnen JM, Solomkin JS, Dellinger EP, Bjornson HS, Page CP. Guidelines for clinical care: anti-infective agents for intraabdominal infection. A Surgical Infection Society policy statement. Arch. Surg. 1992; 127: 83–9.

Intra-abdominal infections 54 Wilson SE. Results of a randomized, multicenter trial of meropenem versus clindamycin/tobramycin for the treatment of intra-abdominal infections. Clin. Infect. Dis. 1997; 24 (Suppl. 2): S197–206.

129

55 Solomkin JS, Dellinger EP, Bohnen JM, Rostein OD. The role of oral antimicrobials for the management of intra-abdominal infections. New Horiz. 1998; 6 (Suppl. 2): S46–52.