PROPHYLACTIC SYSTEMIC ANTIBIOTICS IN COLORECTAL SURGERY

PROPHYLACTIC SYSTEMIC ANTIBIOTICS IN COLORECTAL SURGERY

4 r CASE INCIDENCE RATES FOR AMOEBIASIS AND HEPATITIS B 1000 POPULATION) BY AGE AND SEX IN (PER SAN FRANCISCO promptly to the health officer, di...

645KB Sizes 0 Downloads 124 Views

4 r

CASE INCIDENCE RATES FOR AMOEBIASIS AND HEPATITIS B

1000

POPULATION) BY AGE AND SEX IN

(PER

SAN FRANCISCO

promptly to the health officer, diagnose and treat their own patients, and help identify and treat source contacts and satellite cases whenever possible. Public-health staff, in addition to a full scale public education programme, are making special efforts to educate patients and their who have food-related occupations, and to alert managers of public baths, massage parlours, and other places where such men congregate. Despite these efforts, the increase in the incidence of apparently sexually transmitted enteric diseases has not been halted. We believe that this problem may also exist in other cities. Requests for reprints should be addressed to S. K. D., Department of Public Health, 101 Grove Street, San Francisco, California 94102, contacts

U.S.A. REFERENCES 1. Ward, R., Borchert, P., Wright, A., Kline, E. Lancet, 1972, ii, 726. 2. Tripatzis, I., Horst, H. G. Nature, 1971,231,266. 3. Heathcote, J., Cameron, C. H., Dane, D. S. Lancet, 1974,i,71.

*1?stimated, July 1975. tNot available for 1969.

PROPHYLACTIC SYSTEMIC ANTIBIOTICS IN COLORECTAL SURGERY

apparent, the incidence of shigella species changed dramatically from Shigella soanei, which was originally predominant, to S. flexneri 2a, then to the 3a serotype, and, late in 1976, back to S. sonnei. The public-health department monitored the changing patterns of antibiotic resistance of the various serotypes, so as to assist physicians in the early choice of effective medication. Most of the S. flexneri strains were resistant to tetracycline, streptomycin, and sulphafurazole diethanolamine (sulfisoxazole diolamine). Late in 1976, some ampicillinresistant strains emerged, but much of the shigellosis could be successfully treated with ampicillin trihydrate and closely related antibiotics, or with co-trimoxazole (trimethoprim and sulphamethoxazole). Symptoms in the early cases of newly endemic amoebiasis were relatively mild compared with some very severe cases of shigellosis. This very mildness may have masked the diagnosis early in the outbreak. Effective treatment of amoebiasis, alone or complicated by giardiasis, has included metronidazole, di-iodohydroxyquinoline, and mepacrine hydrochloride. Some patients with shigellosis or amoebiasis have become chronic carriers. Di-iodohydroxyquinoline and tetracycline have been successfully used to treat carriers of amrebiasis. The development of carrier states does not appear to be related to the initial choice of treatment. Treatment of both types of viral hepatitis remains symptomatic. Saliva,’ urine,2 and semen’ have been implicated as vehicles for transmission of the hepatitis-B virus, and both direct orogenital and oral-anal routes of transmission are possible through minor breaches in skin or mucous membranes. An effective and readily available active vaccine is needed urgently. Very few restaurants are now identified as local sources of multiple-case outbreaks, and no water supplies have been implicated. We believe that these enteric infections have become endemic, sexually transmitted diseases in the City. The lack of active, specific vaccines makes effective prevention virtually impossible in those people in whom such disease is likely to develop. Usual public-health measures have not slowed the increase in cases.

Physicians

now

report

cases

of enteric disease

R. S. FEATHER G. R. SAGOR

A. A. M. LEWIS I. D. AMIRAK PAUL NOONE

Royal Free Hospital, Pond Street, London NW3 2QG The prophylactic value of gentamicin combined with either lincomycin or metronidazole in 52 patients undergoing colorectal surgery was investigated. The results confirmed the value of this practice. In a control group, the sepsis-rate was 48% with 1 death attributable to sepsis, compared with a sepsis-rate of 4% in the treated group. The combination of gentamicin and lincomycin was effective against sepsis but pseudomembranous colitis developed in 2 of the 14 patients treated with this combination of drugs. Lincomycin was discontinued, and when metronidazole was substituted the results were equally good and there were no toxic side-effects.

Summary

Introduction COLORECTAL surgery is associated with a high incidence of postoperative sepsis. Abdominal wound infection is the most common form. Intra-abdominal abscess formation and septicaemia are less common but result in important mortality and morbidity. Preoperative mechanical preparation of the bowel is partly successful in reducing infection-rates but they still range from 40 to

70%.l-s

Anaerobes are responsible for much of the infection that occurs after gastrointestinal surgery.3 Goldring et al. used oral metronidazole and kanamycin for three days preoperatively and reduced the infection-rate from 44% to 8%. The use of prophylactic parenteral antibiotics has increased. Single doses of tobramycin and lincomycin both given with the premedication reduced the infection-rate from 50% to 30%6 and lincomycin given alone preoperatively and continued for five days reduced the infection-rate from 38% to 12%.7 In view of these findings and our own observations that several patients with faecal peritonitis and soiling at operation had made an uneventful recovery when given

5

and lincomycin, a prospective randomised clinical trial of systemic antibiotic prophylaxis was carried out. Combinations of gentamicin with lincomycin and gentamicin with metronidazole were used in the

gentamicin

study. Patients and Methods Patient Selection consecutive patients treated by several surgical studied and consent was obtained before their into the trial. Each patient undergoing large-bowel surentry gery, excluding appendicectomy, was allocated to a control or antibiotic-treated group. Selection was by means of previously mixed, sealed envelopes. Patients with preoperative faecal contamination were all treated with antibiotics as a separate sub-group. Patients were not included in the trial if they had received antibiotics in the previous 10 days or if they were known to be allergic to any of the prophylactic antibiotics. Physical cleaning of the colon was used in all patients. Most received castor oil (30 ml) at 48 and 24 hours preoperatively, were restricted to a fluid diet in the 24 hours before surgery, and given a high colonic wash-out immediately before theatre. Other patients were prepared with an elemental diet for four days and a preoperative colonic wash-out.

Fifty-two

teams were

Antibiotic Regimens Two regimens were used: (i) Gentamicin (1.6 mg/kg)

blood-agar (aerobically) chocolate agar (in carbon dioxide) and blood-agar (anaerobically, Gaspak) for at least two weeks. All aerobes and facultative anaerobes isolated were identified by the methods and tables of Cowan and Steel. The anaerobes were characterised and identified by their growth on BBL brucella agar, Nissui fusobacteria agar and Nissui bacteroides agar9 and by the use of A.P.I. sugars and differential antibiotic sensitivity tests.lO Antibiotic sensitivity tests were performed according to Stokes’ disc diffusion method with Oxford Staphylococcus aureus (NCTC strain 6571) or Bacteroides fragilis (WAL strain 1) as control organisms. Gentamicin assay was performed by the urease method," which has been validated in routine use in this laboratory. 12 Lincomycin was assayed by a standard plate technique on B.H.I. (Oxoid) agar with 0.5% yeast extract and 5% lysed horse blood (Tissue Cultures Ltd.) that was flooded with a 1% dilution of a culture of B. fragilis (NCTC 10581) and grown anaerobically in thiol broth (Difco) overnight at 37°C. Metronidazole assays were performed by standard plate assay on B.H.I. agar incorporating 0-5% yeast extract and 2% haEmin/menaphthone (menadione). The plates were flooded with a 1% dilution (in saline) of a culture of B. fragilis (NCTC 8560) grown at 37°C overnight in thioglycolate broth 135C (Difco) to which had been added 0-5% yeast extract+2% hsmin/menadione. All the plates (NUNC 250 mm x 280 mm bioassay plates) were incubated anaerobically at 37°C over-

night. Results

given 8-hourly intramuscuthen or according to serum conlarly intravenously initially, centrations to achieve a peak of between 5 and 10 flglml (one was

hour after intramuscular injection and 15 min after intrainjection) and a trough of less than 3 flglml (just before

venous

the next injection).

(ii) Gentamicin (1-6 mg/kg) was given as before together with a metronidazole lg suppository 8-hourly, except in cases of low anterior resection where it was given intravenously 500 mg 8-hourly. In patients undergoing abdominoperineal resections the suppository was inserted into the colostomy. The antibiotics were given for 5 days. The first dose was given with the premedication. Possible toxicity of the antibiotics was assessed by preoperative and postoperative hmmoglobin measurements, white-blood-cell count, electrolytes, urea, and liver function tests and by simple clinical assessment of balance and hearing before and after chemotherapy.

Patients in the control and antibiotic-treated groups, well-matched for age, sex, disease, and procedure (table i). There were more patients with malignant disease than in other reported series3-s,7 (84% of the control and 78% of the antibiotic-treated group).

were

Wound Infections

Wound infection developed in twelve of the twentyfive control patients (48%) (table 11). Two of these infections were superficial (8%), the others were deep (40%). Anaerobic organisms were isolated from all but two of the deep wound infections usually in combination with coliforms. Seven patients were given antibiotics in the TABLE I-COMPARISON BETWEEN ANTIBIOTIC-TREATED AND

CONTROL GROUPS

Wound Assessment

Patients were seen on days 5, 7, and 10 by a surgeon unconnected with the operating-team, and the wound was inspected. Wound sepsis was defined as the presence of cellulitis or pus which spontaneously discharged or required releasing. Wounds that contained pus were described as deep or superficial according to whether or not the whole thickness of fat was involved. The development of a localised intra-abdominal abscess or anastomotic breakdown was also noted. Blood-cultures, pus samples, and wound swabs were taken when clinically indicated.

Laboratory Methods Pus samples and wound swabs were cultured on blood-agar and MacConkey agar (Oxoid) aerobically and on Columbia blood-agar (Oxoid) and neomycin blood-agar anaerobically (Gaspak) and incubated for forty-eight hours before examination of plates. Blood cultures were taken into nutrient broth with a Castenada slope (Southern Group) and incubated in 10% carbon dioxide, and into tryptose soy broth (Difco) and thiol broth (Difco) and incubated at 37°C. The contents were subcultured at twenty-four and forty-eight hours and then weekly onto

6

postoperative period in an attempt to counteract suspected sepsis. In six of these patients frank sepsis later developed. One of the twenty-seven patients in the antibiotictreated group (table ii, patient 13) had infection of the abdominal wound after an abdominoperineal resection. One other patient in the gentamicin/metronidazole group had his wound accidentally contaminated by faeces from the colostomy on the 4th day. Skin excoriation occurred by the 6th day and the wound became infected. This infection was almost certainly extrinsic. There was no infection in the fourteen patients in the gentamicin and lincomycin group. None of the four patients with preoperative or peroperative peritoneal soiling had postoperative infection. Each received gentamicin and metronidazole which

were

started

during

operation.

Intra-abdominalSepsis In one patient in the control group (table n, no. 7) a pelvic abscess developed which discharged spontaneously through the anastomosis to the rectum. There was no intra-abdominal sepsis in the antibiotic-treated groups.

,

Death

Sepsis was the cause of death in one patient in the control group (no. 4, table n). A sigmoid colectomy was performed because of extensive recurrent growth from a uterine carcinoma. On the 2nd postoperative day she was clinically diagnosed as having septicaemia and was started on gentamicin; however, septicaemia was not

by blood-culture. On the 7th day thick pus from the wound. This dehisced on the 12th appeared At day. exploration she was found to have an abscess beside a leaking anastomosis overlying the iliac cavity vessels. 8 hours postoperatively she died from heemorrhage from a ruptured mycotic aneurysm. On culture the pus yielded Escherichia coli, sensitive to gentamicin and confirmed

re/ococcMyca/M. In the antibiotic-treated group there was one death. This occurred in a 73-year-old man who had undergone a low anterior resection with a covering transverse colostomy for a rectal carcinoma. He had received a 5-day prophylactic course of gentamicin and lincomycin. His plasma-urea levels rose to a maximum of 120 mg/dl over a 5-day period postoperatively, but then returned to 45 mg/dl. This was thought to be due to poor renal reserve from pre-existing pyelonephritis. The postoperative period was otherwise uncomplicated. 4 weeks later the colostomy was closed and the trial envelope drawn placed him in the group for prophylaxis with gentamicin and lincomycin for the second time. Renal impairment increased postoperatively and by the 9th day he was in acute renal failure. This may have been associated with gentamicin administration. Diarrhcea started on the seventh day and a florid colitis was present by the tenth day. A pseudomembrane was not seen on sigmoidoscopy, but the appearances were otherwise suggestive of pseudomembranous colitis and histology of a biopsy specimen was compatible with this diagnosis. On the 11 th postoperative day he underwent a laparotomy for a suspected subhepatic abscess which was found to be an uninfected hsematoma. He died 24 hours after operation. At necropsy a recent extensive coronary throm-

TABLE II-POSTOPERATIVE INFECTION

* 10 "clean" cases+4 with txcat soiling. None of the 14 patients who received gentamicin and lincomycin became infected. I.R.A.=ileorectal anastomosis. s.c.-sigmoid colectomy. L.H.=left hemicolectomy. c.c.==covering colostomy. A.R.=anterior resection. CL.c.=closure of colostomy.

A.P.R.-abdominoperineal

resection.

R.H.=right hemicolectomy.

7

bosis was found. His death must be attributed to multiple factors. The coronary thrombosis was the main cause but the colitis (presumably lincomycin induced), renal failure, and laparotomy were contributory. There were no deaths in the gentamicin/metronidazole group.

TABLE IV-FACULTATIVE ANAEROBES AND AEROBES ISOLATED FROM SEPTIC WOUNDS, AND THEIR ANTIBIOTIC SENSITIVITY PATTERNS

Toxicity clinical evidence of ototoxicity or nephin rotoxicity any of the patients in this series who received gentamicin. It was later thought that the renal failure of the patient described above was probably not directly related to gentamicin; the serum concentrations were maintained within acceptable limits (highest predose serum concentration was 2.9 µg/ml and peak concentration was 8.9 tJLg/ml). Transient diarrhoea, which settled without treatment, developed in two of the fourteen patients receiving lincomycin. Two had pseudomembranous colitis (one died). In both the colitis was severe but responded to local and systemic steroids; the second patient recovered fully. It may be significant that these two patients were in adjacent beds and acquired the condition within a few days of each other." Serum-lincomycin concentrations in all the patients were within an acceptable range (table ill). As a result of these cases, further prophylaxis with the gentamicin/lincomycin combination was not thought to be advisable. Metronidazole was substituted for lincomycin. No side-effects were noted with its use. There

was no

Organisms from Septic Wounds (tables m and v) The facultative anaerobes most commonly isolated were E. coli (46.6%) and Proteus species (33.3%), while Bacteroides species (47.6%), especially B. fragilis (38%), the most common anaerobes. All the coliforms, staphylococci, and pseudomonads were sensitive to gentamicin. All the anaerobes were sen-

were

Amp.=ampicillin. Tet.=tetracycline. Ceph.=cephalexin. Carb.=carbenicillin. Tri.=trimethoprim. Gent.=gentamicin. sitive

gilis

to

metronidazole, although three strains of B. fraresistant to lincomycin.

were

Fsecal streptococci were isolated from four wounds where they occurred in mixed infections. They were resistant to all three prophylactic drugs.

Discussion

Prophylactic treatment with combinations of gentamicin/lincomycin or gentamicin/metronidazole significantly reduced postoperative infection after large-bowel surgery. B. fragilis, a common infecting organism in deep wound infection, is usually sensitive to lincomycin, clindamycin, chloramphenicol and metronidazole, but when the trial started only lincomycin and chloramphenicol were available commercially in parenteral form. Chloramphenicol was not used because it can cause ’aplastic anmmia. 14 Our experience with lincomycin showed that it was highly effective in preventing anaerobic sepsis; however, in two of fourteen patients treated with lincomycin pseudomembranous colitis de-

TABLE III-ANTIBIOTIC SERUM CONCENTRATIONS

([JLg/mI) ON SECOND POSTOPERATIVE DAY

. *Lincomycin serum concentrations of the two patients in whom pseudomembranous colitis developed were, pre dose 18-0 µg/ml and 8-0 µg/ml; and post dose 27.0 ug/ml and 20.0 µg/ml, respectively. In another patient in whom mild diarrhrea developed serum-lincomycin reached 40-0 1 ,,tg/mi. TABLE V-ANAEROBES AND STREPTOCOCCI ISOLATED FROM SEPTIC WOUNDS AND THEIR ANTIBIOTIC SENSITIVITY PATTERNS

Pen.=penicilhn. Eryth.=erythromycin. Lin.=lincomycin. Clind.=clindamycin. Met.=metronidazole.

8

veloped, and we feel that this makes it unacceptable as a prophylactic agent in colorectal surgery. 15-17 Metronidazole in suppository and intravenous forms became available for assessment later in the trial and was substituted for lincomycin; only one of the fourteen patients treated with gentamicin and metronidazole had a (superficial) wound infection. Significant toxicity or complications were not encountered with this combination. Willis et al. 18 used metronidazole prophylactically in appendicectomies. Anaerobic infection did not develop in any of the forty-nine treated patients but aerobic infections developed in five. Keighley et al. reduced the colorectal infection-rate from 38% to 12% with lincomycin; the infections which did not respond to lincomycin were all caused by aerobic organisms. There would thus seem to be a definite gain in providing prophylaxis against aerobic as well as anaerobic organisms. Gentamicin was chosen for this purpose and provides wide cover against all coliforms, staphylococci, and Pseudomonas œruginosa. To be used satisfactorily the dosage needs to be controlled by serum assay, both to prevent toxic accumulation and to ensure adequate peak concentrations. 19 Gentamicin proved effective in this trial and toxicity was not seen. Nevertheless there are possible disadvantages to the routine use of this agent in prophylaxis: (i) Monitoring is required and this may pose problems in hospitals without sufficient microbiological backup. Blood sampling at fixed times may also make extra demands on junior surgical staff. (ii) Overuse of gentamicin increases the prevalence of resistant organisms.2O Oral aminoglycosides are undoubtedly associated with an even greater risk of promoting the emergence of resistant strains of staphylococci and coliforms.21 This may limit the widespread use of oral kanamycin and metronidazole as advocated by

Goldring et al.4

Our results show that the combination of gentamicin and metronidazole is effective and safe in the prophylaxis of infection in colorectal surgery (P<0.05, onetailed exact test of 2 x 2 contingency table). Investigation is required to see whether co-trimoxazole or a cephalosporin can effectively replace gentamicin as the anti-aerobic agent to be used with metronidazole, since it may be wise to reserve gentamicin for the treatment of life-threatening infection. We thank our surgical and operating-theatre department colleagues, Prof. K. E. Hobbs, for their cooperation, Mr Brian Farrell and the staff of the Microbiology Department for their technical assistance, and Miss Susan Brassett for her patience and secretarial work. Requests for reprints should be addressed to P. N., Microbiology Department, Royal Free Hospital, Pond Street, Hampstead, London

INTRAVENOUS GLUCOSE, AMINOACIDS, AND FAT IN THE POSTOPERATIVE PERIOD A Controlled Evaluation of Each Substrate R. P. CRAIG DAVID TWEEDLE

H. A. DAVIDSON IVAN D. A. JOHNSTON

Department of Surgery, University of Newcastle upon Tyne, Newcastle upon Tyne NE1 4LP The metabolic effects of hypocaloric Summary

(2.51 MJ), equicalorific quantities of inglucose, crystalline aminoacids, and soyabean emulsion were compared with those of fasting in 4 groups of male patients (closely matched for age, weight, and fat-free mass) on the day of vagotomy and pyloroplasty and on the first 3 postoperative days. Patients given glucose alone excreted less nitrogen than the fasting patients but nitrogen sparing was greatest in the group given aminoacids alone. There was no evidence of any nitrogen sparing when intravenous fat was the source of energy. There was no difference in blood glucose, free fatty acids, and insulin concentrations among the groups. Ketone-body concentrations rose in the fasting and fat-fed groups but remained low in the groups who received glucose or aminoacids. This study indicates that, in the short term, the increased preservation of protein achieved by the infusion of aminoacids compared with that produced by the traditional infusion of glucose does not warrant the extra cost involved. travenous

Introduction SIMULTANEOUS infusion of aminoacids with carbo-

hydrate and fat emulsion can meet the metabolic requirements of patients with intestinal failure. However, prolonged administration demands a central venous catheter with its attendant risks. Although the peripheral infusion of isotonic aminoacid solution has been advocated as a method of achieving positive nitrogen balance in surgical patients,2.3 the relative efficacy of peripheral infusion of aminoacids, glucose, and fat as sole substrates in sparing lean body mass is controversial.4-14 This controversy might be explained in part by attempts to compare the effects of substrate infusion in pure starvation with those obtained after operation, injury, or sepsis. We have investigated the effects of these substrates in a controlled trial and have attempted to determine whether the routine postoperative infusion of

especially

NW32QG. REFERENCES 1.

Sellwood, R. A., Barn, J. I., Waterworth, P. M., Welbourn, R. B. Br. J. Surg. 1969,56,610. 2. Rosenburg, I. L., Graham, N. G., de Dombal, F. T., Goligher, J. C. ibid. 1971, 58, 266.

3. Leigh, D. A. ibid. 1975,62,375. 4. Goldring, J., Scott, A., McNaught, W., Gillespie, G. Lancet, 1975, ii, 997. 5. Evans, C., Pollock, A. V., Rosenburg, I. L. Br. J. Surg. 1974,61,133. 6. Griffiths, D. A., Shorey, B. A., Simpson, R. A., Speller, D. C. E., Williams, N. B. Lancet, 1976, ii, 325. 7. Keighley, M. R. B., Crapp, A., R., Burdon, D. W., Cooke, N. T., AlexanderWilliams, J. Br. J. Surg. 1976,63,538. J. Manual for the Identification of Medical Bacteria.

8. Cowan, S. T., Steel, K. 1974, London.

9. Ohtani, F.Jap. J. Bacterial. 1970, 25, 292. 10. Sutter, V. L., Atteberu, H. R., Rosenblatt, J. E., Bricknell, K. S., Finegold, S. M. Anaerobic Bacteriology Manual; p. 89. Los Angeles, 1972. 11. Noone, P., Pattison, J. R., Slack, R. C. B. in Some Methods for Microbiological Assay (edited by R. G. Board and D. W. Lovelock); p. 153. London, 1975. 12. Noone, P., Pattison, J. R., Garfield Davies, D. Postgrad. med. J. 1974, 50, suppl. 7 p. 9. 13. Steer, H. W. Lancet, 1975,i,411. 14. Kucers, A., Bennett, N. McK. The Use of Antibiotics; p. 256. London, 1975. 15. Scott, A. J., Nicholson, G. I., Kerr, A. R. Lancet, 1973, ii, 1232. 16. Viteri, A. L., Howard, P. H., Dyck, W. P. Gastroenterology, 1974, 66, 1137. 17. Crapp, A. R., Powis, J., Clark, C., Keighley, M. R. B., Alexander-Williams, J. Br. med. J. 1975, iii, 227. 18. Willis, A. T. et al. ibid. 1976, i, 318. 19. Noone, P., Parsons, T. M. C., Pattison, J. R., Slack, R. C. B., GarfieldDavies, D., Hughes, K. ibid. 1974,i,477. 20. Speller, D. C. E., Raghunath, D., Stephens, M., Viant, A. C., Reeves, D. S., Wilkinson, P. J., Broughall, J. M., Holt, H. A. Lancet, 1976, i, 464. 21. Lancet, 1965,ii,421.