Effect of Clindamycin and Lincomycin Therapy on Faecal Flora

J Clin Pathol: first published as 10.1136/jcp.31.5.439 on 1 May 1978. Downloaded from

Journal of Clinical Pathology, 1978, 31, 439-443

Effect of clindamycin and lincomycin therapy on faecal flora

D. A. LEIGH AND K. SIMMONS From the Department ofMicrobiology, Wycombe General Hospital, High Wycombe, Buckinghamshire, UK

SUMMARY Bacterial counts were carried out on the faeces of 160 patients receiving clindamycin or lincomycin treatment for bacterial infections. In all the patients the total bacteroides count was significantly reduced while strains ofEnterobacteriaciae, yeasts, and streptococci were corresponding- ly increased. Severe diarrhoea developed in 25 (16%) patients, but this could not be related to a change in faecal flora. Diarrhoea was most common when clindamycin was given prophylactically, women were more affected than men, and the incidence was highest in those aged over 60 years. No cases of pseudomembranous colitis were seen. Although clindamycin is a valuable antibiotic for treating established severe anaerobic bacterial infections it should be used cautiously in elderly patients.

The clinical use of many antibiotics is associated colostomies. Of the 160 patients, 131 were receiving with changes in gastrointestinal function in some clindamycin (66 parenterally, eight orally, and 57 by patients. The incidence varies with the particular both routes) and 29 lincomycin (nine parenterally, antibiotic, and the nature may range from a slight six orally, and 14 by both routes). Other antibiotics, increase in bowel movements or less consistency of most commonly ampicillin or amoxycillin, were the faeces to severe watery diarrhoea with blood being received by 102 patients on clindamycin and http://jcp.bmj.com/ and mucus and possibly pseudomembranous colitis. 22 on lincomycin. Established bacterial infection The cause of the gastrointestinal dysfunction is not related to the gastrointestinal or genital tracts was clear. Many explanations such as a direct toxic present in 143 patients, usually confirmed at opera- effect of the antibiotic or its metabolites; the over- tion, and 17 patients received prophylactic treat- growth of certain pathogenic organisms such as ment: 61 of the infections were considered to be Staphylococcus aureus, Candida albicans, and Pseu- severe, 76 moderate, and six minimal. domonas aeruginosa; viruses; and toxins have been Specimens for bacteriological culture were collec- on September 27, 2021 by guest. Protected copyright. put forward. While quantitative studies have been ted from the site of infection in 106 of the 160 carried out in normal patients there is little in- patients and pathogens were isolated in 82 (77%), formation about changes in the faecal flora of the commonest being Bacterioides fragilis (21 patients with diarrhoea or colitis. patients), various species of streptococci (25 patients), In 1974 a prospective study of all patients in the and Staph. aureus (15 patients). The patients were High Wycombe group of hospitals receiving clinda- asked regularly by a research sister about the mycin or lincomycin was started and quantitative development of gastrointestinal side effects. Anti- bacterial counts were carried out on the faeces of biotic-induced diarrhoea was defined as the passage 160 patients. This paper describes the results and of watery faeces more than five times a day. Many compares the changes with the development of patients had mild changes in bowel function which gastrointestinal side effects. resolved within a few days of starting treatment. These were considered to be a result of surgery and Patients and methods not antibiotic therapy. Specimens of faeces collected from the patients Specimens of faeces were collected from 160 patients during the course of clindamycin or lincomycin -79 men and 81 women-including seven with therapy were delivered to the laboratory immediately and were examined within one hour. The faeces Received for publication 2 November 1977 were selectively cultured to determine the counts of 439 J Clin Pathol: first published as 10.1136/jcp.31.5.439 on 1 May 1978. Downloaded from

440 D. A. Leigh and K. Simmons four bacterial species-bacteroides, representing the of cases this was lower than 5 x 103 organisms per predominant anaerobic organisms, Enterobacteri- gram of faeces. In patients receiving lincomycin aciae, streptococci, and yeasts. streptococci were not increased in numbers as they A sample of faeces weighing about 1 g was were in patients receiving clindamycin. Otherwise disintegrated in 5 ml nutrient broth, serial hundred- both antibiotics produced similar changes, which fold dilutions (0-1 ml in 9-9 ml) were made in were independent of the route of administration of nutrient broth, and surface viable counts using the antibiotics. There was no difference between men 0-1 ml inocula were carried out on blood, and women in the faecal flora changes and the use MacConkey, and a special yeast agar incubated of combined antibiotic therapy did not influence aerobically and blood and neomycin blood agar the findings. (0-1 g/l) incubated anaerobically by the Gaspak Antibiotic-induced diarrhoea was seen in 25 (16%) method. The anaerobic agar was not prereduced. of patients and 87 (57%) patients had mild gastro- The plates were incubated for two days and the intestinal dysfunction related to surgery. The bacterial colonies counted. Bacteroides species were relationship of side effects to changes in flora are identified by their ability to grow on neomycin shown in Table 2. There were no differences be- blood agar and confirmed by either the API tech- tween the groups of patients apart from a higher nique (API Products Limited, Rayleigh, Essex) or incidence of increased yeast count when antibiotic- the Mastring (MID8) (Mast Laboratories Ltd., induced diarrhoea was present. Bootle, Merseyside). Strains of Enterobacteriaciae Although severe diarrhoea was nearly twice as and streptococci were identified by routine lab- common in patients who received prophylactic oratory techniques and yeasts by growth on the clindamycin or lincomycin (Table 3) this was not confirmed (Gram). special yeast agar by staining Table 2 Relation ofgastrointestinal side effects to It was impossible to examine pretreatment faecal changes in faecal flora in 153 patients (excluding specimens because most patients had been admitted patients with colostomies) as surgical emergencies which required immediate treatment. The accuracy of the faecal Normal Mild Antibiotic- antibiotic bowels changes induced collection and culture techniques was tested by diarrhoea examining specimens from a similar group of pre- No. of patients 41 87 25 operative surgical patients not receiving antibiotic Decreased anaerobic count 35 (85%) 86 (98 %) 24 (96%)

therapy. Comparable bacterial counts to those Increased counts http://jcp.bmj.com/ were Enterobacteriaciae 21 (51 %) 52 (59%) 16 (64%) reported in normal volunteers found (Leigh Yeasts 11 (27%) 33 (38 %) 14 (56%) et al., 1976). Changes in the faecal bacterial counts Streptococci 6 (15%) 13 (15%) 7 (28%) were related to these normal findings and were considered significant when there was a decrease of Table 3 Incidence ofgastrointestinal side effects 106 or more organisms in the expected bacteroides related to the severity of the infection. Colostomy count, and an increase in the counts of Entero- patients excluded bacteriaciae (1(06 or more), streptococci (105 or more), Degree of infection Prophylaxis on September 27, 2021 by guest. Protected copyright. and yeast (103 or more). Smaller changes in the bacterial counts were not considered significant in Severe Moderate Minimal view of the techniques used. No side 48 65 4 11 effects or mild side Results effects Antibiotic- 10 (17%) 11 (15%) - 4 (27%) was a in the bacterioides induced There significant decrease diarrhoea count in all faecal specimens (Table 1) and in 95%

Table 1 Changes in faecalflora in patients treated with clindamycin and lincomycin Clindamycin Lincomycin Total Parenteral Oral Parenteral Oral No. of patients 123 8 23 6 160 Decreased anaerobic count* 115 (93%) 8 23 6 152 (95%) Increased counts Enterobacteriaciae 74 (60%.) 6 (75%) 1 1 (48 %) 3 (50 %) 94 (59%) Yeasts 46 (37%) 2 (25%) 9 (39%) 3 (50%) 60 (38%) Streptococci 26 (21%) 2 (25%) - - 28 (18%)

*Count < 5 x 103 bacteria per gram faeces. J Clin Pathol: first published as 10.1136/jcp.31.5.439 on 1 May 1978. Downloaded from

Effect ofclindamycin and lincomycin therapy on faecalflora 441 reflected in changes in the faecal flora. Antibiotic- samples showed similar results and the normal induced diarrhoea was seen more often in women quantitative flora was restored after treatment was (19%) than men (13%) and this was most marked stopped (Leigh, unpublished). in those aged over 60 years of whom 30% of The incidence of diarrhoea (16%) with clinda- women were affected and only 11 % of men. mycin and lincomycin in our patients was similar to that found by Tedesco et al. (1974), but no cases of Discussion pseudomembranous colitis were seen in our 160 patients or in those in a longer follow-up study Gastrointestinal dysfunction after antibiotic treat- (Leigh et al., 1977). Possibly many of the reported ment has been related to many causal factors. cases of pseudomembranous colitis follow the irra- Certain antibiotics such as tetracycline and fusidic tional use of clindamycin and lincomycin for acid are toxic to the gastrointestinal tract either as a respiratory infections. The susceptibility of the direct effect, irrespective of concentration, or owing patient is different when a severe anaerobic infection to poor absorption allowing prolonged contact with is present. Strikingly, Bartlett et al. (1972), Fass the mucosa. The use of broad spectrum antibiotics et al. (1973), Gorbach and Thadephalli (1974), can result in the overgrowth of pathogenic bacteria and Chow et al. (1974) using clindamycin and lin- and yeasts such as Staphylococcus aureus, and Candida comycin in severe anaerobic infections have repor- albicans. In most of these cases quantitative bacteri- ted few cases of diarrhoea and no instances of pseudo- ology has not been investigated and it is not known membranous colitis. whether anaerobic bacteria, particularly bacteroides, Although changes in faecal flora could not be a major bacterial component of the faeces were sig- related to side effects there was a definite correlation nificantly altered. Surprisingly, among the large between clindamycin and lincomycin and gastro- number of reported studies there are few reports intestinal side effects in respect of sex and age in our that have looked prospectively into -the faecal patients, women aged over 60 years being more flora associated with diarrhoea and colitis. likely to be affected. This confirms the findings of We could find no relationship between treatment Swartzberg et al. (1976) and Smart et al. (1976). It is with clindamycin or lincomycin and the develop- still uncertain why most patients receiving clinda- ment of diarrhoea. Marr et al. (1975) reported that mycin develop diarrhoea and why some of them the anaerobic flora was only slightly reduced quanti- progress to pseudomembranous colitis.

tatively in patients with diarrhoea receiving clinda- Larson et al. (1977) have reported an undescribed http://jcp.bmj.com/ mycin but when pseudomembranous colitis was cytotoxin in a patient who developed pseudomem- present the reduction was striking. However, in all branous colitis after penicillin therapy. This con- our patients there was a considerable effect on the dition has also occurred with many other anti- bacteroides count irrespective of the presence of biotics-ampicillin (Keating et al., 1974), amoxy- diarrhoea. Metronidazole, another chemotherapeutic cillin (Simila et al., 1976), chloramphenicol (Reiner agent active against anaerobic bacteria, also causes et al., 1952), tetracycline (Klotz et al., 1953), a significant but lesser reduction in the number of cephalosporins (Tures et al., 1976), and co-tri- bacteroides. moxazole (Cameron and Thomas, 1977). There on September 27, 2021 by guest. Protected copyright. The role of Candida and other yeasts in the causa- have been reports of pseudomembranous colitis tion of diarrhoea is vague since there are increased following lincomycin (Scott et al., 1973; Viteri numbers in the faeces with all antibiotic treatments et al., 1974) and clindamycin (Cohen et al., 1973; and the prescription of antifungal treatment does Tedesco et al., 1974), but although the term clinda- not affect the incidence of diarrhoea. An increase in mycin-associated colitis is in constant use this con- the numbers of strains of Enterobacteriaciae was dition probably cannot be directly related to any common and, as seen with ampicillin (Leigh et al., particular compound. 1976), there was no relation to diarrhoea. The aetiology of pseudomembranous colitis is The problems of relating faecal flora to symptoms not clear and its development seems to be unrelated and side effects are complex. Besides the difficulties to the clinical condition of the patient or any one of collection and laboratory culture, the flora is in a factor. The histology of tissue from patients with dynamic state and is probably constantly changing. pseudomembranous colitis suggests various causes- The relationship of the specimen to the last dose or a local Swartzman phenomenon (Hjort and Rapa- the route of administration may significantly alter port, 1965), ischaemia (McGovern and Goulston, the bacteriology and the condition of the patient 1965), or capillary thrombosis (Whitehead, 1971; may considerably affect the pharmacology of the Bogomoletz. 1976). But as these changes are not antibiotic. However, in a study using amoxycillin, seen in all stages of the condition it seems more ampicillin, and talampicillin consecutive faecal likely that they follow an initiating factor, since J Clin Pathol: first published as 10.1136/jcp.31.5.439 on 1 May 1978. Downloaded from

442 D. A. Leigh and K. Simmoni idiopathic pseudomembranous colitis is rarely anaerobic infections. Annals of Internal Medicine, 78, reported (Jackson and Anders, 1972) and geographical 853-859. differences have been suggested by Geddes (1974). Geddes, A. M. (1974). Lincomycin and clindamycin A viral or bacterial aetiology has been reported colitis. (Letter). British Medical Journal, 4, 591. Gorbach, S. L., and Thadepalli, H. (1974). Clindamycin (Steer, 1975; Marr et al., 1976), but there is little in pure and mixed anaerobic infections. Archives of indication that an altered faecal flora in terms of the Internal Medicine, 134, 87-92. normal bacteria or usual pathogens plays a major Harrod, M. J. E., Brown, M. S., Weinberg, A. G., role. Harkness, W. N., and Goldstein, J. L. (1975). Familial The possibility of genetic influence is interesting pseudomembranous colitis and its relation to lin- and cannot be excluded (Harrod et al., 1975). The comycin therapy. American Journal of Digestive increased incidence of diarrhoea among patients Diseases, 20, 808-812. receiving lincomycin and atropine-diphenoxylate Hjort, P. F., and Rapaport, S. I. (1965). The Schwartz- (Lomotil) or codein phosphate (Novak et al., 1976) man reaction; pathogenetic mechanisms and clinical manifestations. Annual Review of Medicine, 16, suggests that decreased intestinal motility may in- 135-168. crease the contact time between the antibiotic and Jackson, B. T., and Anders, C. J. (1972). Idiopathic its metabolites and the intestinal mucosa, allowing a pseudomembranous colitis successfully treated by toxic reaction to occur. However, decreased in- surgical excision. British Journal of Surgery, 59, testinal motility is common after gastrointestinal 154-156. surgery, and few reports of pseudomembranous Keating, J. P., Frank, A. L., Barton, L. L., and Tedesco, colitis are seen in these patients. Probably the in- F. J. (1974). Pseudomembranous colitis associated with itiating factor is a toxic agent acting directly on the ampicillin therapy. Amer-ican Journal of Diseases of intestinal mucosa (Pittman et al., 1974) leading to a Children, 128, 369-370. Klotz, A. P., Palmer, W. L., and Kirsner, J. B. (1953). two-stage process of a local Schwartzman phenome- Aureomycin proctitis and colitis; a report of five non and capillary microthrombi (Price and Davies, cases. Gastroenterology, 25, 44-47. 1977), but its nature is unknown. Although the Larson, H. E., Parry, J. V., Price, A. B. Davies, D. R., true incidence of pseudomembranous colitis is not Dolby, J., and Tyrrell, D. A. J. (1977). Undescribed well documented, and its frequency is overemphasised toxin in pseudomembranous colitis. British Medical by the large number of single cases reported, a Journal, 1, 1246-1248. common association in its development is antibiotic Leigh, D. A., Reeves, D. S., Simmons, K., Thomas, A. L. new therapy. Since which patients will develop the and Wilkinson, P. J. (1976). Talampicillin; a http://jcp.bmj.com/ complication cannot be predicted chemotherapy derivative of ampicillin. British Medical Journal, 1, must be used especially for prophylaxis. 1378-1380. carefully, Leigh, D. A., Simmons, K., and Williams, S. (1977). The treatment of abdominal and gynaecological Mrs K. Simmons is in receipt of a research grant infections with parenteral clindamycin phosphate. from Beechams Pharmaceuticals. Mastring MID 8 Journal of Atntimicrobial Chemotherapy, 3, 493-500. were kindly supplied by Mr K. Ashley, Mast McGovern, V. J., and Goulston, S. J. M. (1965). Laboratories Ischaemic enterocolitis. Gut, 6, 213-220. Limited, Bootle, Merseyside. on September 27, 2021 by guest. Protected copyright. Marr, J. J., Sans, M. D., and Tedesco, F. J. (1975). References Bacterial studies of clindamycin-associated colitis. Gastroenterology, 69, 352-358. Bartlett, J. G., Sutter, V. L., and Finegold, S. M. (1972). Novak, E., Lee, J. G., Seckman, C. E., Phillips, J. P., Treatment of anaerobic infections with lincomycin and DiSanto, A. R. (1976). Unfavourable effect of and clindamycin. New England Journal of Medicine, atropine-diphenoxylate (Lomotil) therapy in lin- 287, 1006-1010. comycin-caused diarrhea. Journal of the Americaii Bogomoletz, W. V. (1976). Fibrin thrombi, a cause of Medical Association, 235, 1451-1454. clindamycin-associated colitis? Gut, 17, 483-487. Pittman, F. E., Pittman, J. C., and Humphrey, C. D. Cameron, A., and Thomas, M. (1977). Pseudomembran- (1974). Colitis following oral lincomycin therapy. ous colitis and co-trimoxazole. British Medical Journal, Archives of Internal Medicine, 134, 368-372. 1, 1321. Price, A. B., and Davies, D. R. (1977). Pseudomembran- Chow, A. W., Montgomerie, J. Z., and Guze, L. B. ous colitis. Journal of Clinical Pathology, 30, 1-12, (1974). Parenteral clindamycin therapy for severe Reiner, L., Schlesinger, M. J., and Miller, G. M. (1952). anaerobic infections. Archives of Internal Medicine, Pseudomembranous colitis following aureomycin 134, 78-82. and chloramphenicol. Archives of Pathology, 54, Cohen, L. E., McNeill, C. J., and Wells, R. F. (1973). 39-67. Clindamycin-associated colitis. Journal ofthe American Scott, A. J., Nicholson, G. I., and Kerr, A. R. (1973). Medical Association, 223, 1379-1380. Lincomycin as a cause of pseudomembranous colitis. Fass, R. J., Scholand, J. F., Hodges, G. R., and Saslaw, Lancet, 2, 1232-1234. S. (1973). Clindamycin in the treatment of serious Simila, S., Kouvalainen, K., and Makela, P. (1976). J Clin Pathol: first published as 10.1136/jcp.31.5.439 on 1 May 1978. Downloaded from

Effect of clindamycin and lincomycin therapy on faecal flora 443 Pseudomembranous colitis after amoxycillin (Letter). Tedesco, F. J., Barton, R. W., and Alpers, D. H. (1974). Lancet, 2, 317-318. Clindamycin-associated colitis. Annals of Internal Smart, R. F., Ramsden, D. A., Gear, M. W. L., Nicol, Medicine, 81, 429-433. A., and Lennox, W. M. (1976). Severe pseudomem- Tures, J. F., Townsend, W. F., and Rose, H. D. (1976). branous colitis after lincomycin and clindamycin. Cephalosporin-associated pseudomembranous colitis. British Journal of Surgery, 63, 25-29. Journal of the American Medical Association, 236, Steer, H. W., (1975). The pseudomembranous colitis 948-949. associated with clindamycin therapy-a viral colitis. Viteri, A. L., Howard, P. H., and Dyck, W. P. (1974). Gut, 16, 695-706. The spectrum of lincomycin-clindamycin colitis. Swartzberg, J. E., Maresca, R. M., and Remington, J. S. Gastroenterology, 66, 1137-1144. (1976). Gastrointestinal side effects associated with Whitehead, R. (1971). Ischaemic enterocolitis; an clindamycin. Archives of Internal Medicine, 136, expression of the intravascular coagulation syndrome. 876-879. Gut, 12, 912-917. The April 1978 Issue THE APRIL 1978 ISSUE CONTAINS THE FOLLOWING PAPERS Editorial: Authorships and acknowledgements Comparison of susceptibility of Neisseria menin- gitidis to sodium sulphadiazine and sodium fusidate Lazy leucocyte syndrome-disorder of the granulo- in vitro R. S. MILES AND A. MOYES cyte membrane? P. H. PINKERTON, JEAN B. ROBINSON, Free-sporing Cl. welchii in ordinary laboratory media AND J. S. SENN and conditions M. RAHMAN Antibody to hepatitis B surface antigen in haemo- Ethanol formed from arabinose: a rapid method for philiacs on long-term therapy with Scottish factor detecting Escherichia coli P. J. COLOE VIII C. J. BURRELL, S. H. BLACK, AND D. M. RAMSAY Head-space gas liquid chromatography for rapid detection of Escherichia coli and Proteus mirabilis Bone marrow fibrosis in acute lymphoblastic in urine P. J. COLOE leukaemia of childhood I. M. HANN, D. I. K. EVANS, H. B. MARSDEN, P. MORRIS JONES, AND M. K. PALMER Purification of specific precipitinogen and extraction of endotoxin from Haemophilus influenzae http://jcp.bmj.com/ Distribution of transferrin, ferritin, and lactoferrin J. C. VAN DER ZWAN, J. DANKERT, K. DE VRIES, in human tissues D. Y. MASON AND C. R. TAYLOR N. G. M. ORIE, AND H. F. KAUFFMAN Fluorimetric assay of mycobacterial group-specific Role of imprint cytology in intraoperative diagnosis: hydrolase enzymes J. M. GRANGE value and limitations K. C. SUEN, W. S. WOOD, A. A. AND Chemical and kinetic study of the lipoproteins in SYED, N. F. QUENVILLE, P. B. CLEMENT abetalipoproteinaemic plasma JAMES SHEPHERD, on September 27, 2021 by guest. Protected copyright. Human polyomavirus (BK) infection and ureteric MURIEL CASLAKE, ELIZABETH FARISH, AND ADAM stenosis in renal allograft recipients D. V. COLEMAN, FLECK E. F. D. MACKENZIE, S. D. GARDNER, J. M. POULDING, Technical methods B. AMER, AND W. J. I. RUSSELL A rapid paper-strip method for the detection of penicillinase production by penicillin resistant strains Urinary tract infection in young women, with of Staphylococcus aureus D. B. WHELDON AND special reference to Staphylococcus saprophyticus MARY P. E. SLACK W. A. GILLESPIE, MARGARET A. SELLIN, PATRICIA GILL, A simpler and less expensive approach to platelet MARY STEPHENS, L. A. TUCKWELL, AND A. L. HILTON counting with improved precision using the Technicon Autocounter M. M. LAVER AND M. F. Aberrant form of Pseudomonas aeruginosa in HOWELL sputum and cerebrospinal fluid causing infection Letter to the Editor in a compromised patient JOHN MIDDLETON AND HERMAN CHMEL Book reviews Copies are still available and may be obtained from the PUBLISHING MANAGER, BRITISH MEDICAL ASSOCIATION, TAVISTOCK SQUARE, LONDON WC1H 9JR, price £3-00, including postage