Br. J. Surg. Vol. 62 (1975) 375-378
Wound infections due to Bacteroides fragi/is foI lowing intestinal surgery D. A. LEIGH* SUMMARY
Ninety-six wound infections due to Bacteroides fragilis occurring after intestinal surgery are described. Most infestions followed appendicectomy or colectomy f o r diverticulitis. B. fragilis was isolated in pure growth from 47 per cent of infections, and in mixed culture the other common bacteria were Klebsiella spp., Escherichia coli and Enterobacter spp. Fifty-three per cent of the infections responded rapidly to either non-specific measures or antibiotic therapy, but 47 per cent of patients developed complications, usually abscess ,formation. The successfirl isolation of bacteroides f i o m clinical material requires special precautions. The specimen for basteriological examination must be sent to the laboratory in a transport medium which maintains an anaerobic environment and the culture techniques must include a selective agar f o r the isolation of anaerobic bacteria. The role of antibiotic therapy in preventing postoperative wound injection after intestinal surgery is discussed. Bacteroides has an unusual pattern of antibiotic susceptibility and it is essential that an effectiue anfibiotic such as lincomycin is included in the treatment o f infection originating ,from the intestinal tract. THERE has recently been revived interest in the importance of anaerobic bacteria and their role in postoperative infection, particularly following gastro-intestinal surgery. Anaerobic bacteria are the predominant group of organisms in the faeces and account for over 90 per cent of the total bacterial population. These bacteria are highly sensitive to aerobic conditions and changes in temperature and are difficult to isolate routinely in the laboratory unless special culture techniques are used. Bacteroidesfragilis is the species most frequently isolated but its incidence in wound infections following intestinal surgery is not known. After appendicectomy 93 per cent of wound infections were due to bacteroides (Leigh, 1974a), and in patients with anaerobicbacteriaemia73 per cent had underlying gastrointestinal disease (Wilson et al., 1972). Gillespie and Guy (1956) and Gunn (I 956) reported that the majority of bacteroides infections followed intestinal surgery. The successful isolation of anaerobic bacteria from surgical infections requires careful collection of the specimen and protection in a transport medium which maintains an anaerobic environment during transit to the laboratory. This paper presents the details of a group of 96 patients who developed wound infection due to B. fiagilis following intestinal surgery, describes the 27*
measures necessary for successful isolation and discusses the role of antibiotic therapy. Patients and collection of specimens All the patients included in the study were admitted for emergency or elective surgery to the intestinal tract and had developed severe wound infections which were associated with fever, pain and a purulent discharge. Swabs were taken from the wound infection and were sent to the laboratory in either Stuart’s transport medium or Robertson’s meat broth. Specimens of pus or purulent fluid were also collected at the time of the operation from patients with obvious intra-abdominal infection. Patients with formed abscesses at the time of surgery were excluded. Details of antibiotic therapy and the postoperative progress of the patients were obtained either from the Control of Infection records or the inpatients’ notes. All wound infections were inspected by the Control of Infection sister. The outcome of the infection was considered successful when there was a rapid response to either antibiotic therapy or non-specific therapeutic measures. Complications were present when an abscess developed from the initial infection or the infection persisted despite appropriate therapy. Laboratory techniques The specimens were examined as soon as possible after arrival in the laboratory. All the swabs were cultured on to simple blood agar, MacConkey’s agar, a selective blood agar plate containing 0.01 per cent neomycin, haemin 5 pg/ml and vitamin K 0.2 pg/ml and into Robertson’s meat broth. The plates were incubated aerobically and anaerobically using the Gaspak (BBL, Maryland) method and examined at 24 and 48 hours. Subculture from the Robertson’s meat broths was carried out on to a similar set of agar plates. The strains of B. fragilis were identified by routine bacteriological methods (Cowan and Steel, 1965), and antibiotic sensitivity tests were carried out by the impregnated disc technique. Results The primary site of surgery and probable source of the 96 B. fragilis infections is shown in Table I . Sixtyone of the wound infections were associated with appendicectomy, and in 37 perforation of the appendix had occurred. Four infections followed an elective appendicectomy carried out in association with other surgical procedures. There were 4 infections following
* Wycombe and Amersham Wycombe, Buckinghamshire.
General
Hospitals,
High
375
D. A. Leigh Table I: SOURCE OF INFECTION IN 96 PATIENTS Site of surgery Disease No. of cases Appendix
Ileum Colon and rectum
Acute appendicitis Perforated appendicitis Routine appendicectomy Ischaemia Crohn's disease Ischaemia Diverticulitis Perforated diverticulitis Carcinoma Caecum Colon Rectum
61
4
:>
Table 11: ASSOCIATED BACTERIA IN 52 WOUND INFECTIONS WITH MIXED CULTURES Bacteria No. Y Gram negstive: Klebsiella spp. 21 43 Escherichia coli 14 27 Enterobocter spp. 13 25 4 8 Proteics spp. Pseudomonas aeruginosa 3 6 Gram positive: Anaerobic streptococci 4 8 Streptococcus faecalis 3 6 P-Haemolytic streptococci 1 Ciostridiuni welchii 1 StaDh vlococcus albus 1 Table 111: INCIDENCE OF COMPLICATIONS Incidence of Nature of Disease complications (%) complication Acute appendicitis
45
Perforated appendicitis
46
Routine appendicectomy
50
Ileum
33
Diverticulitis
70
Perforated diverticulitis
50
Carcinoma
50
Abscess: wound 7, pelvic 1, appendix 1 Abscess: wound 13, subphrenic 2, pelvic I , abdominal wall 1 Abscess: wound 1, abdominal wall 1 Abscess : generalized abdominal 1 Abscess: wound 3, pelvic 1, persisting sinus 3 Abscess: pelvic 2 (death 1) Abscess: wound 5, generalized abdominal 1
Table 1V: ISOLATION RATE OF BACTEROIDES IN MAJOR INTESTINAL SURGERY Wound Bacteroides infections isolated Total no. Year of ooerations No. Y No. "A 1970 1200* 123 10 16 17 1971 1272 145 11 20 ;4 1972 1380 155 11 35 22 1973 1328 200 15 120 60
*
Approximate total
surgery to the ileum, 3 of which were in patients with Crohn's disease. Thirty-one infections followed operations on the large intestine; 16 in patients with diverticulitis including 6 whereperforation of a diverticulum had occurred, and 12 where carcinoma was found.
376
In 26 (27 per cent) of the 96 patients who developed wound infection, swabs taken at the time of surgery grew the bacteria later responsible for the wound infection. B. fkgilis was isolated in pure growth from 44 (47 per cent) of the infections. Where a mixed growth was present (Table ZI) the commonest associated organism was Klebsiella spp. (40 per cent). A pure growth was found more frequently where carcinoma was present (83 per cent), and less commonly where perforation of the appendix (29 per cent) or large bowel (40 per cent) had occurred. Fifty-one (53 per cent) of the wound infections responded rapidly either without specific therapy or with antibiotic therapy. Forty-five (47 per cent) patients developed complications, usually wound abscess or metastatic abscesses such as subphrenic or pelvic, and 1 patient died of the infection. The incidence of complications was highest (70 per cent) in diverticulitis (Table H I ) . Where complications developed a B. fragilis infection had been diagnosed at the time of surgery in 33 per cent of patients. Forty-six of the patients received chemotherapy but in only 21 was the antibiotic given effective against bacteroides. Ten patients received either clindamycin or lincomycin, 5 tetracycline and the remainder either co-trimoxazole or metronidazole. Where ineffective treatment was prescribed 20 patients received ampicillin, 7 kanamycin, 4 cephaloridine and 3 cotrimoxazole. In several cases combined therapy was given. Antibiotics were more commonly prescribed at the time of the operation when perforation of the intestine had occurred, but in many cases treatment was only given after the wound infection had developed. In this retrospective analysis antibiotic therapy did not appear to improve the rate of recovery or reduce the incidence of complications. The increase in the incidence of isolation of B. fragilis from postoperative wound infections following major intestinal surgery is shown in Table ZV.Although the incidence of all wound infections rose only slightly between 1970 and 1973, the isolation rate of B.fvagilis increased from 13 to 60 per cent. The increase in the isolation rate was due mainly to the use of special transport media which protected the bacteria present on the swab during delivery to the laboratory. In addition the swab was inoculated on to a special selective culture medium which contains neomycin to inhibit other bacteria.
Discussion Infections due to B. fragilis occur most frequently after intestinal surgery (Gillespie and Guy, 1956), and in the Wycombe and Amersham Hospitals 67 per cent followed operations on the small and large bowel. In addition, where bacteroides organisms were isolated from abscesses 78 per cent of these had originated from the intestine. In this study the majority of wound infections followed appendicectomy, but this is probably only a reflection of the frequency of this operation.
Wound infections following intestinal surgery All the wound infections were associated with severe inflammation, a purulent discharge and fever. There was frequent necrosis of the skin edges of the operative wound and in a few cases the suture line burst. Bacteria were isolated in large numbers and B. frugilis was always present in the highest count and was usually seen as the predominant organism in the Gram film. There was no clinical doubt that the wound was infected and not simply colonized in the superficial layers by bacteria. B. frugilis was isolated in pure culture in nearly 50 per cent of the wound infections. In the past the incidence of sterile swabs from clinically infected wounds has been a common finding and there is little doubt that in many of these patients anaerobic bacteria which failed to grow on culture were responsible for the infection. The high incidence of Klebsiella spp. and Enterobacter spp. in mixed cultures (Table ZZ) may be a reflection of previous antibiotic therapy as these organisms are resistant to ampicillin, an antibiotic commonly prescribed. Wound infection following surgery for malignant conditions of the intestine was usually caused by a single organism, and 4 mixed culture was found more frequently following perforation of the intestine. The outcome of bacteroides infections is relativeIy good. Only I patient in this series died but 47 per cent of patients developed complications which prolonged their period of hospitalization. The commonest complication was metastatic abscess formation which frequently required further surgical intervention. The development of bacteriaemia is a serious feature of bacteroides infections which can increase the mortality to over 30 per cent (Bodner et al., 1970; Wilson et al., 1972; Leigh, 1974b), but this can be controlled by the early administration of effective antibiotic therapy (Gelb and Seligman, 1970; MacKenzie and Litton, 1974). The role of antibiotic therapy in the prevention and treatment of wound infections after surgery is not generally known. Many antibiotic and antiseptic preparations have been used locally in the wound area to prevent infections (Crosfil et al., 1969; Longland et al., 1971; Gilmore et al., 1973; Evans et al., 1974). Parenteral antibiotic therapy has been shown to reduce the development of complications but not to influence wound healing (Magarey et al., 1971). However, in patients with perforated appendicitis the incidence of wound infection was reduced from 50 to 15 per cent when lincomycin was given early in the postoperative period (Leigh, 1974a). The success of antibiotic therapy depends on the activity of the chemotherapeutic agent against the infecting organisms and the dosage in providing adequate serum and tissue concentrations. Many studies have not prescribed antibiotics with significant bactericidal activity against anaerobic bacteria, and in the group of patients reported here less than 50 per cent received an antibiotic effective against bacteroides. Bacteroides has an unusual spectrum of antibiotic susceptibility and it is rarely sensitive to the commonly prescribed chemotherapeutic agents. It is totally
resistant to the aminoglycoside group of antibiotics such as gentamicin and kanamycin, and of the penicillins, only carbenicillin has consistent bactericidal activity because of the high blood levels that are achieved on standard parenteral therapy. Tetracycline has been used effectively in the past, but recent workers (Bodner et al., 1972; Kislak, 1972; Okubadejo et al., 1973) have found nearly 40 per cent of strains resistant to easily attainable serum concentrations. Clindamycin is the most active antibiotic against bacteroides and has been used with success in the treatment of infections (Bartlett et al., 1972), but it can only be given orally, and in severe infections parenteral lincomycin is the antibiotic of choice. There are three main difficulties associated with the diagnosis of anaerobic infections following intestinal surgery. First, the failure of the surgeon to take specimens for bacteriological examination from areas of potential infection seen at operation. In appendicectomy it has been shown (Leigh, 1974a) that the isolation of B. frngilis from the appendix fossa is associated with a high incidence of wound infection, and that the administration of effective antibiotic therapy early in the postoperative period can reduce this incidence. Secondly, it is necessary to protect anaerobic bacteria which are highly oxygen-sensitive, and the specimen must be collected from deep in the wound and placed into a transport medium, either Stuart’s or Robertson’s meat broth, which provides an anaerobic environment. Many operations are performed outside the normal laboratory working hours and immediate examination and culture of the specimen are not possible. In the laboratory it is always necessary to use selective media to isolate anaerobic bacteria, as it is time consuming and difficult to identify these bacteria amongst large numbers of aerobic organisms which may grow well even under anaerobic conditions. The increase in the isolation rate of B. fragilis when these precautions are taken is startlingly high (Table ZV) and has confirmed the report of other workers (Hoffmann and Gierhake, 1969). The incidence of wound infection following intestinal surgery is high; in the past it has been accepted as a natural sequel to the opening of the bowel and subsequent contamination of the peritoneal cavity and wound area. In the High Wycombe area in 1973, 15 per cent of patients undergoing major intestinal surgery developed a wound infection, and in 60 per cent B. fragilis was the causative organism either alone or in association with other bacteria. It is possible by bacteriological examination of the peritoneal cavity at the time of operation, especially in patients with perforation of the intestine where the incidence of wound infection is high, to diagnose the likely bacterial pathogens and to prescribe antibiotic treatment early in the postoperative period and prevent the subsequent wound infection. This procedure requires close cooperation between the surgeons and the laboratory, but it would be of great benefit to the patients in whom a wound infection may greatly delay their discharge from hospital.
377
D. A. Leigh
In view of the high frequency of isolation of bacteroides from wound infections following intestinal surgery blind chemotherapy must include an antibiotic such as linconiycin which is effective against these organisms. Acknowledgements I would like to thank Mrs E. Norman, Control of Infection sister, who was responsible for the wound infection survey; the consultant surgeons, Mr P. H. Lord and Mr B. Higgs, for allowing the study of their patients; Mrs K. Simmons for the details of antibiotic susceptibility of the strains of bacteroides; and the technicians of the Microbiology Laboratory. References
and FlNEGOLD S. M. (1972) Treatment of anaerobic infections with lincomycin and clindamycin. New. Engl. J . Med. 287, 1006-1010. BODNER S. J., KOENIG M. G. and GOODMAN J. S. (1970) Bacteremia bacteroides infections. Ann. Intern. Med. 73,537-544. BODNER s. J., KOENIG M. G., TREANOR L. L. and GOODMAN J. s. (1972) Antibiotic susceptibility testing of Bacteroides. Antimicrob. Agents CAemother. 2, 57-60. COWAN s. T. and STEEL K. J . (1965) Manual for the Identification o f Medical Bacteria. London, Cambridge University Press. CROSFILL M., HALL R. and LONDON D. (1969) The use of chlorhexidine antisepsis in contaminated surgical wounds. Br. J. Surg. 56, 906-908. EVANS c., POLLOCK A. v. and ROSENBERG I. L. (1974) The reduction of surgical wound infections by topical cephaloridine: a controlled clinical trial. Br. J. Surg. 61, 133-135. GELB A. F. and SELIGMAN s. J. (1970) Bacteroidaceae bacteremia, effect of age and focus of infection on clinical course. J A M A 212, 1038-1041. BARTLETT J. G., SUTTER V. L.
378
w. A. and GUY J. (1956) Bacteroides in intra-abdominal sepsis. Lancet 1, 1039-1042. GILMORE 0. J. A., MARTIN T. D. M. and FLETCHER B. N. (1973) Prevention of wound infection after appendicectomy. Lancet 1, 220-222. G U N N A. A. (1956) Bacteroides septicaemia. J. R . toll. Surg. Edinb. 2, 41-50. HOFFMANN K. and GIERHAKE F. w. (1969) Postoperative infection of wounds by anaerobes. Ger. Med. Mon. 14, 31-33. KlSLAK J. w. (1972) The susceptibility of Bacteroides fragilis to 24 antibiotics. J . Infect. Dis. 125, 295-299. LEIGH D. A. (1974a) Bacterial flora of the appendix fossa and the incidence of wound infection. J. Clin. Pathol. 27, 997-1000. LEIGH D. A. (1974b) The clinical importance of infections due to Bacfevoides fragilis and the role of antibiotic therapy. Br. Med. J. 3, 225-228. LONGLAND c. J., GRAY J. G., LEES w. and GARRETT J. A. M. (1971) The prevention of infection in appendicectomy wounds. Br. J. Surg. 58, 117-1 19. MACKENZIE I. and LITTON A. (1974) Bacteroides bacteriaernia in surgical patients. Br. J . Surg. 61, 288-290. GILLESPIE
MAGAREY C. J., CHANT A. D. B., RICKFORD C. R. K. and MAGAREY J. R. (1971) Peritoneal drainage
and systemic antibiotics after appendicectorny. Lancet 2, 179-1 82. OKUBADEJO 0. A , , GREEN P. J. and PAYNE D. J. H. (1973) Bacteroides infection among hospital patients. Br. Med. J. 2, 212-214. WILSON w. R., MARTIN w. J . , WILKOWSKE c. J. and WASHINGTON J. A. (1972) Anaerobic bacteraemia. Mayo Clin. Prvc. 47, 649-646.
Wound infections due to Bacteroides fragi/is foI lowing intestinal surgery D. A. LEIGH* SUMMARY
Ninety-six wound infections due to Bacteroides fragilis occurring after intestinal surgery are described. Most infestions followed appendicectomy or colectomy f o r diverticulitis. B. fragilis was isolated in pure growth from 47 per cent of infections, and in mixed culture the other common bacteria were Klebsiella spp., Escherichia coli and Enterobacter spp. Fifty-three per cent of the infections responded rapidly to either non-specific measures or antibiotic therapy, but 47 per cent of patients developed complications, usually abscess ,formation. The successfirl isolation of bacteroides f i o m clinical material requires special precautions. The specimen for basteriological examination must be sent to the laboratory in a transport medium which maintains an anaerobic environment and the culture techniques must include a selective agar f o r the isolation of anaerobic bacteria. The role of antibiotic therapy in preventing postoperative wound injection after intestinal surgery is discussed. Bacteroides has an unusual pattern of antibiotic susceptibility and it is essential that an effectiue anfibiotic such as lincomycin is included in the treatment o f infection originating ,from the intestinal tract. THERE has recently been revived interest in the importance of anaerobic bacteria and their role in postoperative infection, particularly following gastro-intestinal surgery. Anaerobic bacteria are the predominant group of organisms in the faeces and account for over 90 per cent of the total bacterial population. These bacteria are highly sensitive to aerobic conditions and changes in temperature and are difficult to isolate routinely in the laboratory unless special culture techniques are used. Bacteroidesfragilis is the species most frequently isolated but its incidence in wound infections following intestinal surgery is not known. After appendicectomy 93 per cent of wound infections were due to bacteroides (Leigh, 1974a), and in patients with anaerobicbacteriaemia73 per cent had underlying gastrointestinal disease (Wilson et al., 1972). Gillespie and Guy (1956) and Gunn (I 956) reported that the majority of bacteroides infections followed intestinal surgery. The successful isolation of anaerobic bacteria from surgical infections requires careful collection of the specimen and protection in a transport medium which maintains an anaerobic environment during transit to the laboratory. This paper presents the details of a group of 96 patients who developed wound infection due to B. fiagilis following intestinal surgery, describes the 27*
measures necessary for successful isolation and discusses the role of antibiotic therapy. Patients and collection of specimens All the patients included in the study were admitted for emergency or elective surgery to the intestinal tract and had developed severe wound infections which were associated with fever, pain and a purulent discharge. Swabs were taken from the wound infection and were sent to the laboratory in either Stuart’s transport medium or Robertson’s meat broth. Specimens of pus or purulent fluid were also collected at the time of the operation from patients with obvious intra-abdominal infection. Patients with formed abscesses at the time of surgery were excluded. Details of antibiotic therapy and the postoperative progress of the patients were obtained either from the Control of Infection records or the inpatients’ notes. All wound infections were inspected by the Control of Infection sister. The outcome of the infection was considered successful when there was a rapid response to either antibiotic therapy or non-specific therapeutic measures. Complications were present when an abscess developed from the initial infection or the infection persisted despite appropriate therapy. Laboratory techniques The specimens were examined as soon as possible after arrival in the laboratory. All the swabs were cultured on to simple blood agar, MacConkey’s agar, a selective blood agar plate containing 0.01 per cent neomycin, haemin 5 pg/ml and vitamin K 0.2 pg/ml and into Robertson’s meat broth. The plates were incubated aerobically and anaerobically using the Gaspak (BBL, Maryland) method and examined at 24 and 48 hours. Subculture from the Robertson’s meat broths was carried out on to a similar set of agar plates. The strains of B. fragilis were identified by routine bacteriological methods (Cowan and Steel, 1965), and antibiotic sensitivity tests were carried out by the impregnated disc technique. Results The primary site of surgery and probable source of the 96 B. fragilis infections is shown in Table I . Sixtyone of the wound infections were associated with appendicectomy, and in 37 perforation of the appendix had occurred. Four infections followed an elective appendicectomy carried out in association with other surgical procedures. There were 4 infections following
* Wycombe and Amersham Wycombe, Buckinghamshire.
General
Hospitals,
High
375
D. A. Leigh Table I: SOURCE OF INFECTION IN 96 PATIENTS Site of surgery Disease No. of cases Appendix
Ileum Colon and rectum
Acute appendicitis Perforated appendicitis Routine appendicectomy Ischaemia Crohn's disease Ischaemia Diverticulitis Perforated diverticulitis Carcinoma Caecum Colon Rectum
61
4
:>
Table 11: ASSOCIATED BACTERIA IN 52 WOUND INFECTIONS WITH MIXED CULTURES Bacteria No. Y Gram negstive: Klebsiella spp. 21 43 Escherichia coli 14 27 Enterobocter spp. 13 25 4 8 Proteics spp. Pseudomonas aeruginosa 3 6 Gram positive: Anaerobic streptococci 4 8 Streptococcus faecalis 3 6 P-Haemolytic streptococci 1 Ciostridiuni welchii 1 StaDh vlococcus albus 1 Table 111: INCIDENCE OF COMPLICATIONS Incidence of Nature of Disease complications (%) complication Acute appendicitis
45
Perforated appendicitis
46
Routine appendicectomy
50
Ileum
33
Diverticulitis
70
Perforated diverticulitis
50
Carcinoma
50
Abscess: wound 7, pelvic 1, appendix 1 Abscess: wound 13, subphrenic 2, pelvic I , abdominal wall 1 Abscess: wound 1, abdominal wall 1 Abscess : generalized abdominal 1 Abscess: wound 3, pelvic 1, persisting sinus 3 Abscess: pelvic 2 (death 1) Abscess: wound 5, generalized abdominal 1
Table 1V: ISOLATION RATE OF BACTEROIDES IN MAJOR INTESTINAL SURGERY Wound Bacteroides infections isolated Total no. Year of ooerations No. Y No. "A 1970 1200* 123 10 16 17 1971 1272 145 11 20 ;4 1972 1380 155 11 35 22 1973 1328 200 15 120 60
*
Approximate total
surgery to the ileum, 3 of which were in patients with Crohn's disease. Thirty-one infections followed operations on the large intestine; 16 in patients with diverticulitis including 6 whereperforation of a diverticulum had occurred, and 12 where carcinoma was found.
376
In 26 (27 per cent) of the 96 patients who developed wound infection, swabs taken at the time of surgery grew the bacteria later responsible for the wound infection. B. fkgilis was isolated in pure growth from 44 (47 per cent) of the infections. Where a mixed growth was present (Table ZI) the commonest associated organism was Klebsiella spp. (40 per cent). A pure growth was found more frequently where carcinoma was present (83 per cent), and less commonly where perforation of the appendix (29 per cent) or large bowel (40 per cent) had occurred. Fifty-one (53 per cent) of the wound infections responded rapidly either without specific therapy or with antibiotic therapy. Forty-five (47 per cent) patients developed complications, usually wound abscess or metastatic abscesses such as subphrenic or pelvic, and 1 patient died of the infection. The incidence of complications was highest (70 per cent) in diverticulitis (Table H I ) . Where complications developed a B. fragilis infection had been diagnosed at the time of surgery in 33 per cent of patients. Forty-six of the patients received chemotherapy but in only 21 was the antibiotic given effective against bacteroides. Ten patients received either clindamycin or lincomycin, 5 tetracycline and the remainder either co-trimoxazole or metronidazole. Where ineffective treatment was prescribed 20 patients received ampicillin, 7 kanamycin, 4 cephaloridine and 3 cotrimoxazole. In several cases combined therapy was given. Antibiotics were more commonly prescribed at the time of the operation when perforation of the intestine had occurred, but in many cases treatment was only given after the wound infection had developed. In this retrospective analysis antibiotic therapy did not appear to improve the rate of recovery or reduce the incidence of complications. The increase in the incidence of isolation of B. fragilis from postoperative wound infections following major intestinal surgery is shown in Table ZV.Although the incidence of all wound infections rose only slightly between 1970 and 1973, the isolation rate of B.fvagilis increased from 13 to 60 per cent. The increase in the isolation rate was due mainly to the use of special transport media which protected the bacteria present on the swab during delivery to the laboratory. In addition the swab was inoculated on to a special selective culture medium which contains neomycin to inhibit other bacteria.
Discussion Infections due to B. fragilis occur most frequently after intestinal surgery (Gillespie and Guy, 1956), and in the Wycombe and Amersham Hospitals 67 per cent followed operations on the small and large bowel. In addition, where bacteroides organisms were isolated from abscesses 78 per cent of these had originated from the intestine. In this study the majority of wound infections followed appendicectomy, but this is probably only a reflection of the frequency of this operation.
Wound infections following intestinal surgery All the wound infections were associated with severe inflammation, a purulent discharge and fever. There was frequent necrosis of the skin edges of the operative wound and in a few cases the suture line burst. Bacteria were isolated in large numbers and B. frugilis was always present in the highest count and was usually seen as the predominant organism in the Gram film. There was no clinical doubt that the wound was infected and not simply colonized in the superficial layers by bacteria. B. frugilis was isolated in pure culture in nearly 50 per cent of the wound infections. In the past the incidence of sterile swabs from clinically infected wounds has been a common finding and there is little doubt that in many of these patients anaerobic bacteria which failed to grow on culture were responsible for the infection. The high incidence of Klebsiella spp. and Enterobacter spp. in mixed cultures (Table ZZ) may be a reflection of previous antibiotic therapy as these organisms are resistant to ampicillin, an antibiotic commonly prescribed. Wound infection following surgery for malignant conditions of the intestine was usually caused by a single organism, and 4 mixed culture was found more frequently following perforation of the intestine. The outcome of bacteroides infections is relativeIy good. Only I patient in this series died but 47 per cent of patients developed complications which prolonged their period of hospitalization. The commonest complication was metastatic abscess formation which frequently required further surgical intervention. The development of bacteriaemia is a serious feature of bacteroides infections which can increase the mortality to over 30 per cent (Bodner et al., 1970; Wilson et al., 1972; Leigh, 1974b), but this can be controlled by the early administration of effective antibiotic therapy (Gelb and Seligman, 1970; MacKenzie and Litton, 1974). The role of antibiotic therapy in the prevention and treatment of wound infections after surgery is not generally known. Many antibiotic and antiseptic preparations have been used locally in the wound area to prevent infections (Crosfil et al., 1969; Longland et al., 1971; Gilmore et al., 1973; Evans et al., 1974). Parenteral antibiotic therapy has been shown to reduce the development of complications but not to influence wound healing (Magarey et al., 1971). However, in patients with perforated appendicitis the incidence of wound infection was reduced from 50 to 15 per cent when lincomycin was given early in the postoperative period (Leigh, 1974a). The success of antibiotic therapy depends on the activity of the chemotherapeutic agent against the infecting organisms and the dosage in providing adequate serum and tissue concentrations. Many studies have not prescribed antibiotics with significant bactericidal activity against anaerobic bacteria, and in the group of patients reported here less than 50 per cent received an antibiotic effective against bacteroides. Bacteroides has an unusual spectrum of antibiotic susceptibility and it is rarely sensitive to the commonly prescribed chemotherapeutic agents. It is totally
resistant to the aminoglycoside group of antibiotics such as gentamicin and kanamycin, and of the penicillins, only carbenicillin has consistent bactericidal activity because of the high blood levels that are achieved on standard parenteral therapy. Tetracycline has been used effectively in the past, but recent workers (Bodner et al., 1972; Kislak, 1972; Okubadejo et al., 1973) have found nearly 40 per cent of strains resistant to easily attainable serum concentrations. Clindamycin is the most active antibiotic against bacteroides and has been used with success in the treatment of infections (Bartlett et al., 1972), but it can only be given orally, and in severe infections parenteral lincomycin is the antibiotic of choice. There are three main difficulties associated with the diagnosis of anaerobic infections following intestinal surgery. First, the failure of the surgeon to take specimens for bacteriological examination from areas of potential infection seen at operation. In appendicectomy it has been shown (Leigh, 1974a) that the isolation of B. frngilis from the appendix fossa is associated with a high incidence of wound infection, and that the administration of effective antibiotic therapy early in the postoperative period can reduce this incidence. Secondly, it is necessary to protect anaerobic bacteria which are highly oxygen-sensitive, and the specimen must be collected from deep in the wound and placed into a transport medium, either Stuart’s or Robertson’s meat broth, which provides an anaerobic environment. Many operations are performed outside the normal laboratory working hours and immediate examination and culture of the specimen are not possible. In the laboratory it is always necessary to use selective media to isolate anaerobic bacteria, as it is time consuming and difficult to identify these bacteria amongst large numbers of aerobic organisms which may grow well even under anaerobic conditions. The increase in the isolation rate of B. fragilis when these precautions are taken is startlingly high (Table ZV) and has confirmed the report of other workers (Hoffmann and Gierhake, 1969). The incidence of wound infection following intestinal surgery is high; in the past it has been accepted as a natural sequel to the opening of the bowel and subsequent contamination of the peritoneal cavity and wound area. In the High Wycombe area in 1973, 15 per cent of patients undergoing major intestinal surgery developed a wound infection, and in 60 per cent B. fragilis was the causative organism either alone or in association with other bacteria. It is possible by bacteriological examination of the peritoneal cavity at the time of operation, especially in patients with perforation of the intestine where the incidence of wound infection is high, to diagnose the likely bacterial pathogens and to prescribe antibiotic treatment early in the postoperative period and prevent the subsequent wound infection. This procedure requires close cooperation between the surgeons and the laboratory, but it would be of great benefit to the patients in whom a wound infection may greatly delay their discharge from hospital.
377
D. A. Leigh
In view of the high frequency of isolation of bacteroides from wound infections following intestinal surgery blind chemotherapy must include an antibiotic such as linconiycin which is effective against these organisms. Acknowledgements I would like to thank Mrs E. Norman, Control of Infection sister, who was responsible for the wound infection survey; the consultant surgeons, Mr P. H. Lord and Mr B. Higgs, for allowing the study of their patients; Mrs K. Simmons for the details of antibiotic susceptibility of the strains of bacteroides; and the technicians of the Microbiology Laboratory. References
and FlNEGOLD S. M. (1972) Treatment of anaerobic infections with lincomycin and clindamycin. New. Engl. J . Med. 287, 1006-1010. BODNER S. J., KOENIG M. G. and GOODMAN J. S. (1970) Bacteremia bacteroides infections. Ann. Intern. Med. 73,537-544. BODNER s. J., KOENIG M. G., TREANOR L. L. and GOODMAN J. s. (1972) Antibiotic susceptibility testing of Bacteroides. Antimicrob. Agents CAemother. 2, 57-60. COWAN s. T. and STEEL K. J . (1965) Manual for the Identification o f Medical Bacteria. London, Cambridge University Press. CROSFILL M., HALL R. and LONDON D. (1969) The use of chlorhexidine antisepsis in contaminated surgical wounds. Br. J. Surg. 56, 906-908. EVANS c., POLLOCK A. v. and ROSENBERG I. L. (1974) The reduction of surgical wound infections by topical cephaloridine: a controlled clinical trial. Br. J. Surg. 61, 133-135. GELB A. F. and SELIGMAN s. J. (1970) Bacteroidaceae bacteremia, effect of age and focus of infection on clinical course. J A M A 212, 1038-1041. BARTLETT J. G., SUTTER V. L.
378
w. A. and GUY J. (1956) Bacteroides in intra-abdominal sepsis. Lancet 1, 1039-1042. GILMORE 0. J. A., MARTIN T. D. M. and FLETCHER B. N. (1973) Prevention of wound infection after appendicectomy. Lancet 1, 220-222. G U N N A. A. (1956) Bacteroides septicaemia. J. R . toll. Surg. Edinb. 2, 41-50. HOFFMANN K. and GIERHAKE F. w. (1969) Postoperative infection of wounds by anaerobes. Ger. Med. Mon. 14, 31-33. KlSLAK J. w. (1972) The susceptibility of Bacteroides fragilis to 24 antibiotics. J . Infect. Dis. 125, 295-299. LEIGH D. A. (1974a) Bacterial flora of the appendix fossa and the incidence of wound infection. J. Clin. Pathol. 27, 997-1000. LEIGH D. A. (1974b) The clinical importance of infections due to Bacfevoides fragilis and the role of antibiotic therapy. Br. Med. J. 3, 225-228. LONGLAND c. J., GRAY J. G., LEES w. and GARRETT J. A. M. (1971) The prevention of infection in appendicectomy wounds. Br. J. Surg. 58, 117-1 19. MACKENZIE I. and LITTON A. (1974) Bacteroides bacteriaernia in surgical patients. Br. J . Surg. 61, 288-290. GILLESPIE
MAGAREY C. J., CHANT A. D. B., RICKFORD C. R. K. and MAGAREY J. R. (1971) Peritoneal drainage
and systemic antibiotics after appendicectorny. Lancet 2, 179-1 82. OKUBADEJO 0. A , , GREEN P. J. and PAYNE D. J. H. (1973) Bacteroides infection among hospital patients. Br. Med. J. 2, 212-214. WILSON w. R., MARTIN w. J . , WILKOWSKE c. J. and WASHINGTON J. A. (1972) Anaerobic bacteraemia. Mayo Clin. Prvc. 47, 649-646.
