Abstract

The frequency of antibiotic-associated diarrhoea (AAD) and Clostridium difficile-associated diarrhoea (CdAD) was prospectively determined in a population of 2462 patients recruited from five Swedish hospitals, including divisions for infectious diseases, orthopaedics, surgery, geriatrics, nephrology and internal medicine. AAD developed in 4.9% of the treated patients. Faecal samples were obtained from 69% of patients with AAD and 55.4% were positive for C. difficile cytotoxin B. The frequency of AAD varied from 1.8 to 6.9% at the participating centres (P < 0.001). The frequency of AAD also varied considerably between medical disciplines and wards within different hospitals and was highest in the nephrology and geriatric units (6.7 and 7.1%, respectively). There was no difference in frequency of AAD when analysed with respect to gender or age. Medical interventions (laxative treatment, endoscopy and abdominal surgery) or presence of one concomitant disease (diabetes, malignancy, chronic renal disease and inflammatory bowel disease) did not significantly affect the frequency of AAD, whereas patients suffering from two or more of these illnesses had significantly (P = 0.001) higher frequencies of AAD. Patients treated with antibiotics for 3 days had a significantly (P = 0.009) lower frequency of AAD than those treated for longer periods. Treatment with cephalosporins, clindamycin or broad-spectrum penicillins was associated with an increased risk of AAD. With specimens from one centre, 62.5% of tested patients with AAD and 33.8% of asymptomatic patients were positive for cytotoxin B. Although C. difficile cytotoxin B in stool samples was significantly associated with AAD (P = 0.003), the causal relationship with diarrhoea is not always evident.

Introduction

Diarrhoea is one of the most frequent side effects of antibiotic treatment. The symptoms may vary from slight abdominal discomfort to severe diarrhoea and colitis. The aetiology of antibiotic-associated diarrhoea (AAD) varies. The disruption of the normal enteric flora caused by antibiotics may lead to overgrowth of pathogens and functional disturbances of the intestinal carbohydrate and bile acid metabolism, resulting in osmotic diarrhoea.1 Allergic, toxic and pharmacological effects of antibiotics may also affect the intestinal mucosa and motility.1 Cytotoxin-producing Clostridium difficile is held to be the causative agent of approximately 20% of AAD and of nearly all cases of pseudomembranous colitis, the most severe manifestation of AAD.2–4

In hospitals, C. difficile is an increasing problem, especially among elderly patients with serious underlying diseases.2,4 Other infectious agents with less convincing correlations with AAD include Clostridium perfringens, Staphylococcus aureus, Salmonella spp. and Candida spp.1,3 In hospitalized patients, AAD has been associated with increases in mortality, length of stay and cost of medical care.4–7

Virtually all antibiotics have been implicated in AAD.3,8–10 The antimicrobial spectrum of an antibiotic (particularly the activity against anaerobic bacteria) and the faecal concentration of antibiotic are probably important factors in the development of diarrhoea. Cephalosporins, clindamycin and ampicillin have been reported to present the highest risk, especially for C. difficile-associated diarrhoea (CdAD).3,8–12 A single dose of an antibiotic may be sufficient to trigger symptomatic disease, e.g. in surgical prophylaxis.8,13

The magnitude of AAD and CdAD in different patient populations is not well known. Frequencies of AAD varying between 10 and 30% have been reported among hospitalized patients treated with antibiotics.3,14–18 Many reports have, however, been retrospective studies of relatively small samples often restricted to single hospitals, wards or treatment regimens, while only a limited number of large prospective studies involving different hospitals and medical disciplines has been performed.

The main purpose of the present study was to determine prospectively the frequency of AAD in a large sample of patients, hospitalized at five Swedish hospitals, and to include patients from different medical disciplines. Possible risk factors for acquisition of AAD, including antibiotics, intestinal interventions and concomitant diseases, were also studied.

Materials and methods

Study design

This prospective, epidemiological study was conducted between November 1995 and July 1996 at the hospitals of Danderyd, Huddinge, Lund, Norrköping and Umeå, Sweden and was approved by regional research ethics committees. All patients, and for those aged 12–17 years, their parents, provided written consent. All who received systemic antibiotic therapy were invited to participate and were enrolled in the study if inclusion criteria were fulfilled. The aim was to identify ≥50 patients with CdAD, 10 patients each from divisions of orthopaedic surgery, general surgery and internal medicine/nephrology/geriatrics and 20 patients from divisions of infectious diseases.

Patients and treatment

Patients aged ≥12 years who had started antibiotic treatment within 24 h before enrolment and who gave their informed consent were eligible for the study. Patients with antibiotic treatment who were screened and excluded from the study were registered in a reject log. Exclusion criteria were the presence of any kind of diarrhoeal illness on admission to hospital, inability to co-operate (because of advanced age or mental confusion), and infection with human immunodeficiency virus (HIV).

Patients included in the study were followed for 6 weeks. All antibiotics prescribed in the 4 weeks before inclusion and during the study period were recorded on a case report form, as were gender, age, date of admission and discharge from hospital. Certain presumed risk factors for AAD such as medical interventions (use of laxatives, rectal and/or colon endoscopy, transplantation and abdominal surgery) and concomitant diseases (chronic renal disease, diabetes mellitus, malignancy and inflammatory bowel disease) were also recorded. On discharge from hospital the patients were given diary cards in which they were asked to register antibiotic treatment and, in the case of diarrhoeal illness, the number of loose stools per day. A faecal sample for C. difficile cytotoxin B test was requested from patients who developed diarrhoea during the study period. Patients who developed diarrhoea after discharge from hospital were asked to mail their faecal samples to one central laboratory using vials that they had received on discharge from hospital. Day 1 was defined as the day when antibiotic treatment was started. A randomly selected group of 200 asymptomatic patients from one of the study centres was also asked to send faecal samples for detection of C. difficile cytotoxin B, 1 week after completing their antibiotic treatment.

Evaluation of clinical symptoms

All decisions concerning evaluation of clinical and bacteriological outcome at follow up were made before the study was initiated. AAD was defined as three or more loose stools per day for two or more consecutive days;19 mild illness was defined as development of any diarrhoeal symptoms not fulfilling the definition of AAD. Either AAD or mild illness could be associated with a positive or negative C. difficile cytotoxin B test. CdAD was defined as AAD with a positive cytotoxin B test in a stool specimen.

Microbiological procedures

All faecal specimens were tested for C. difficile cytotoxin B according to Orrhage et al.20

Statistical methods

χ2 Tests were used for dichotomous data. Confidence intervals for risk ratios were calculated using the Mantel–Haenszel method, while those for proportions were calculated using exact binomial probabilities. Wilcoxon two-sample tests were used for ordinal data. All tests were two-sided and P values <0.05 were considered significant.

Results

Patient characteristics and clinical outcome

Of 6369 patients screened, 2799 (44%) satisfied the inclusion criteria and were included in the study. Reasons for exclusion of 3570 patients were: patient refusal (18%), inability to co-operate because of advanced age or mental confusion (54%), ongoing diarrhoeal illness at start of antibiotic treatment (17%), previous inclusion (6%), HIV infection (3%) and others (2%). Of the 2799 patients included, 2462 were evaluable for clinical outcome (Table I). Of the remaining 337 non-evaluable patients (12% of all included), 93 refused to participate further in the study, 139 were lost to follow up, 75 died during the study period and 30 were excluded from evaluation because they did not fulfil inclusion criteria.

The distribution of patients in different centres and medical disciplines is shown in Table II. Of them 2111 (86%) were admitted from home while 351 (14%) were referred from another hospital department. Of the evaluable patients, 523 (21%) had previously (within 3 months) been admitted to hospital and 2388 (97%) were discharged from hospital during the 6 week follow-up period.

Of the evaluable patients, 12% experienced diarrhoea during antibiotic treatment or within 45 days after the start of antibiotic treatment (Table I). The median time to occurrence of symptoms was 9 days (range 0–45 days) after start of antibiotic treatment and 13 days after admittance to hospital for patients with AAD and those with mild illness.

Symptoms of AAD or mild illness did not significantly affect the length of hospital stay. However, 12 of the patients with AAD or mild illness (4.1%) were readmitted because of relapse of symptoms and were again hospitalized for a median of 5 days.

There was no statistically significant difference in the frequency of AAD when analysed with respect to gender (4.6% in males and 5.1% in females) or age (5.2% in 1165 patients aged ≥65 years and 4.6% in 1297 patients aged <65 years).

Clinical outcome according to centre and medical disciplines

The highest frequency of AAD, 6.9%, was observed at centre V and the lowest, 1.8%, at centre IV (P < 0.001) (Table II). At centre IV, a relatively high number of orthopaedic patients (48%) was included. These patients had a low total frequency of AAD (0.4%) and mild illness (1.1%). Among the non-orthopaedic patients at this centre the frequencies of AAD (4.5%) and mild illness (7.6%) were similar to those observed at other centres.

The frequency of AAD and mild illness varied considerably between medical disciplines and wards within different hospitals. The overall frequency of AAD among patients treated in geriatric units was 7.1% and varied from 2.6 to 6.7% among those treated in other medical disciplines (Table II).

Clinical outcome according to medical interventions or concomitant diseases

The medical interventions recorded during the study period seemed not to affect the risk of developing AAD. The highest frequency (9.3%) was found among patients investigated with rectoscopy or colonoscopy [confidence interval (CI) 5.2–16.4%]. Corresponding frequencies were 5.4% for patients treated with laxatives, 2.1% for transplant patients and 6.7% for subjects who had abdominal surgery.

Of the 540 subjects with concomitant illness, 6.3% developed AAD (Table III). Patients suffering from two or more of these illnesses had a significantly increased risk of AAD (risk ratio 3.01; CI 1.55–5.86; P = 0.001) compared with patients with at most one concomitant disease (Table III). No such differences where found among patients with diabetes, chronic renal disease or malignancy only.

Clinical outcome according to antibiotic treatment

A significant increase in risk for AAD was observed for patients treated with antibiotics for >3 days (risk ratio 2.28; CI 1.23–4.21; P = 0.009) (Table IV). No significant differences were found among groups of patients treated for >3 days (Table IV).

Among groups of patients treated with only one antibiotic, the highest frequencies of AAD (6.7%) were found in patients treated with broad-spectrum penicillins (ampicillin derivatives, pivmecillinam and piperacillin alone or in combination with tazobactam) and those treated with cephalosporins (6.1%) (Table V). None of those treated only with quinolones or co-trimoxazole developed AAD.

The frequencies of AAD and total illness were also analysed in groups of patients treated with more than one antibiotic either concomitantly or sequentially (Table V). The highest frequency of AAD (11%) was observed among those treated with clindamycin in combination with other antibiotics. Among 220 patients treated with metronidazole and 40 treated with glycopeptides in combination with other antibiotics, 7.7 and 5.0%, respectively, developed AAD.

Only one of 80 patients treated solely with cefuroxime experienced symptoms of AAD. When cefuroxime was given in combination with other antibiotics, 4.9% of the patients developed AAD compared with 7.5% for those who received cefotaxime in combinations and 11.4% for patients in whom ceftazidime was the cephalosporin in combinations (Table V). The frequency of AAD varied substantially among groups of patients treated with an oral cephalosporin in combination with other antibiotics. The highest frequency of AAD (28.6%, CI 15.9–48.7%) was found in patients treated with cefpodoxime proxetil, as compared with 1.6% in patients given loracarbef (Table V).

Bacteriological results

Faecal samples for C. difficile cytotoxin B test were obtained from 83 (69%) of 120 patients with AAD and 78 (45%) of 174 patients with mild illness. Cytotoxin B was detected in 55.4 and 36% of the samples from patients with AAD and mild illness, respectively. The frequency of CdAD among tested patients with AAD varied between 41.7 and 62.5% in the five participating hospitals (Table II). CdAD was most common among patients treated in nephrology wards (66.7%); lower numbers were observed in orthopaedic (22.2%) and surgical wards (28.6%) (Table II).

Among the 522 patients evaluated at centre I, cytotoxin B was detected in 63% of 32 samples obtained from 34 patients with AAD and in 29% of 41 samples from 66 patients with mild illness (Table VI). Faecal samples were also obtained from 139 asymptomatic patients (63 males and 76 females; mean age 55 years) at this centre, and 34% had a positive C. difficile cytotoxin B test (Table VI). Cytotoxin B was detected in 40% of asymptomatic patients treated in the infectious diseases ward and 41.9% in the nephrology ward compared with 21.7 and 22.2% of patients treated in the orthopaedic or geriatric wards, respectively (Table VI).

Discussion

This is to our knowledge the largest prospective study evaluating the frequency of AAD and CdAD among hospitalized antibiotic-treated patients. Of the 2462 patients, recruited from five geographically unrelated hospitals and from a variety of medical disciplines, 4.9% developed AAD. An additional 7.1% experienced mild illness not fulfilling the proposed criteria for AAD. Thus, 12% of the antibiotic-treated patients developed diarrhoeal symptoms.

AAD was not a major clinical problem in our patients. The frequency of AAD was somewhat lower than expected from previous studies, which found AAD in 10–30% of patients.3,14–18 Comparison with other reports is, however, difficult since there are differences in study populations and definitions of AAD. Many studies included fewer patients and were often restricted to specific hospitals and wards.14–17 The frequency must also depend on the antibiotic agents used as well as on host factors. In recent reports, with the same definition of AAD as in the present study, the frequencies still varied but were nevertheless higher, at 22,14 1515 and 10%.17

Although <70% of patients with AAD and 45% with mild illness were tested for C. difficile cytotoxin B, the results indicate that over half of the patients with AAD and one-third of those with mild illness may be cytotoxin B positive. The assumed risk of developing diarrhoea associated with C. difficile among patients with AAD in the present study was higher than found in many previous reports, where figures ranging from 10 to 40% are commonly given.1–3,18 However, since the sampling rate was rather low, we cannot exclude the possibility that patients with severe symptoms were more willing to be tested.

One-third of asymptomatic patients, recruited from one of the five study hospitals, were positive for C. difficile cytotoxin B. As in many other enteric infections, asymptomatic carriage of the pathogen is probably much more common than the disease itself. The reported carriage rates of C. difficile in healthy adults vary from 1 to 3% and those in healthy neonates from 35 to 65%, as estimated by established methods.18,21,22 Epidemiological studies have shown that C. difficile may be acquired by 3–30% of in-patients in different hospitals.2,19,23–25 Only a minority of these patients will, however, develop symptoms. The true carriage rate is probably much higher. Studies on healthy volunteers have shown colonization rates of ≤60% after they were given antibiotics,26 and virtually all individuals may carry the microorganism permanently or intermittently. Owing to high carriage rates in selected populations, the causal relationship between diarrhoea and the finding of C. difficile in faeces is not always evident. This emphasizes the need for simple and more reliable methods to distinguish between colonization and active disease.

Elderly patients have traditionally been considered to have an increased risk of developing AAD, especially caused by C. difficile.4,9,18,24,27,28 We found only minor differences in the frequency of AAD between patients below and above 65 years of age. It should be stressed, however, that a large proportion of the elderly patients screened was excluded because of inability to co-operate. On the other hand, the risk of AAD among elderly patients might previously have been overestimated since many studies have been performed in units mainly treating elderly patients .7,16,23,24 We were not able to adjust the frequency of AAD in different age groups for other confounding factors such as concomitant disease, length of antibiotic treatment or length of hospital stay. It is most probable, however, that such risk factors were at least as common in patients above 65 years as in those below 65 years of age. This implies that the risk of developing AAD is not related to age per se but rather to other host factors or medical interventions.

The frequency of AAD and CdAD was similar at the five hospitals but varied between different medical disciplines. The extremely low rate of AAD at one hospital was to a large extent explained by the fact that nearly half of the patients evaluated had been treated in orthopaedic wards with a low frequency of AAD. The great variation in frequency of AAD and CdAD between different medical disciplines was expected and might be explained by factors such as antibiotics used, duration of treatment, concomitant diseases, measures of hospital hygiene, length of hospital stay and medical interventions. In general, the lowest frequencies of AAD and CdAD were found in orthopaedic disciplines, where antibiotics to a large extent were given as short perioperative prophylaxis. Moreover, most patients in the orthopaedic wards were admitted from home and had a rather short duration of medical care in hospital, which reduces the risk of CdAD.9,25

Patients treated in units for chronic renal diseases and long-term care facilities are at increased risk of developing AAD.4,24,27 This was also evident in the present study, where the highest frequencies were found in certain nephrology and geriatric units, mainly treating patients with concomitant chronic diseases. However, concomitant diseases, such as diabetes mellitus, malignancy, chronic renal disease and organ transplantation, did not seem to involve an increased risk of AAD unless two or more such underlying diseases were present. The frequency of AAD was not significantly higher among patient groups exposed to medical interventions such as treatment with laxatives, rectal endoscopy and abdominal surgery, factors that have previously been shown to be associated with increased risk of CdAD.9,28

Prolonged courses of antibiotic treatment have been related to an increased risk of AAD.8,9 This was also found in the present study, where a short duration of treatment, <3 days, involved a much lower risk than longer treatment periods. The median time for occurrence of symptoms was 9 days after the start of treatment, which is in accordance with other studies.17 The frequency of AAD seemed to be about the same if the duration of treatment was 1 week or >3 weeks. This suggests that disturbance of the normal colonic flora, eventually resulting in diarrhoea, takes place within about 1 week of antibiotic treatment and that the risk for symptoms thereafter is independent of duration of therapy. Prolonged duration of hospital stay has been reported in cases of AAD and CdAD.9,25 In contrast, the length of hospitalization in the present study was the same irrespective of diarrhoea.

As in previous studies, AAD was most frequently associated with cephalosporins, clindamycin and broad-spectrum penicillins, like amoxycillin and piperacillin,3,8–12 whereas quinolones, co-trimoxazole, penicillins and tetracyclines constituted a lower risk. This was evident also when these preparations were given as single treatment. Among oral cephalosporins, cefpodoxime-treated patients were commonly implicated. This might be expected owing to its great influence on the colonic flora.29 Metronidazole and vancomycin, the most common treatment alternatives for CdAD, were only given in combination with other antibiotics. The frequency of AAD was 7.7 and 5.0%, respectively, when these two preparations were included in the therapy. This suggests that metronidazole and vancomycin are not effective as prophylaxis for AAD.

In conclusion, AAD occurred in approximately 5% of the patients in this comprehensive, prospective study, which largely confirms the risk factors reported by others. In specimens from one centre, C. difficile cytotoxin B was detected in 62.5% of cases of AAD tested but also in 33.8% of asymptomatic patients treated with antibiotics. While detection of cytotoxin B is significantly associated with AAD, the results suggest a limited diagnostic specificity of such tests and emphasize the need for simple and reliable diagnostic tools to distinguish active disease.

Table I.

Patient characteristics for all 2462 evaluable subjects including patients with antibiotic-associated diarrhoea (AAD) or mild illness (MI)

   Age (years) 
Patient category Number of patients (%) Gender male/female mean median range 
No diarrhoea 2168 (88.0) 1019/1149 60.2 63.5 18–103 
AAD 120 (4.9) 53/67 62.5 65.0 22–91 
MI 174 (7.1) 78/96 61.5 64.0 20–94 
Total with diarrhoea 294 (12.0) 131/163 62.1 64.6 18–103 
   Age (years) 
Patient category Number of patients (%) Gender male/female mean median range 
No diarrhoea 2168 (88.0) 1019/1149 60.2 63.5 18–103 
AAD 120 (4.9) 53/67 62.5 65.0 22–91 
MI 174 (7.1) 78/96 61.5 64.0 20–94 
Total with diarrhoea 294 (12.0) 131/163 62.1 64.6 18–103 
Table II.

Frequency of antibiotic-associated diarrhoea (AAD) and mild illness (MI), and of positive C. difficile cytotoxin B test among patients with AAD according to centre and medical discipline

  Number (%) of patients with  
Centre/discipline (number of patients) Mean age (years) AAD MI AAD + MI Number of AAD patients tested/ number positive for cytotoxin B (%) 
aSignificantly different from other centres (P < 0.001). 
Centre I (522) 60.3 34 (6.5) 66 (12.6) 100 (19.1) 32/20 (62.5) 
Centre II (326) 59.4 12 (3.7) 27 (8.3) 39 (12.0) 8/4 (50.0) 
Centre III (526) 58.1 27 (5.1) 44 (8.4) 71 (13.5) 12/5 (41.7) 
Centre IV (554) 58.4 10 (1.8)a 16 (2.9) 26 (4.7)a 8/5 (62.5) 
Centre V (534) 65.4 37 (6.9) 21 (3.9) 58 (10.8) 23/12 (52.2) 
Infectious diseases (993) 58.1 56 (5.6) 100 (10.1) 156 (15.7) 42/27 (64.3) 
Orthopaedics (695) 57.2 18 (2.6) 36 (5.2) 54 (7.8) 9/2 (22.2) 
Surgery (288) 60.2 13 (4.5) 7 (2.4) 20 (6.9) 7/2 (28.6) 
Geriatrics (170) 77.5 12 (7.1) 17 (10.0) 29 (17.1) 10/6 (60.0) 
Nephrology (164) 61.2 11 (6.7) 12 (7.3) 23 (14.0) 9/6 (66.7) 
Internal medicine (152) 70.1 10 (6.6) 2 (1.3) 12 (7.9) 6/3 (50.0) 
Total (2462) 60.4 120 (4.9) 174 (7.1) 294 (12.0) 83/46 (55.4) 
  Number (%) of patients with  
Centre/discipline (number of patients) Mean age (years) AAD MI AAD + MI Number of AAD patients tested/ number positive for cytotoxin B (%) 
aSignificantly different from other centres (P < 0.001). 
Centre I (522) 60.3 34 (6.5) 66 (12.6) 100 (19.1) 32/20 (62.5) 
Centre II (326) 59.4 12 (3.7) 27 (8.3) 39 (12.0) 8/4 (50.0) 
Centre III (526) 58.1 27 (5.1) 44 (8.4) 71 (13.5) 12/5 (41.7) 
Centre IV (554) 58.4 10 (1.8)a 16 (2.9) 26 (4.7)a 8/5 (62.5) 
Centre V (534) 65.4 37 (6.9) 21 (3.9) 58 (10.8) 23/12 (52.2) 
Infectious diseases (993) 58.1 56 (5.6) 100 (10.1) 156 (15.7) 42/27 (64.3) 
Orthopaedics (695) 57.2 18 (2.6) 36 (5.2) 54 (7.8) 9/2 (22.2) 
Surgery (288) 60.2 13 (4.5) 7 (2.4) 20 (6.9) 7/2 (28.6) 
Geriatrics (170) 77.5 12 (7.1) 17 (10.0) 29 (17.1) 10/6 (60.0) 
Nephrology (164) 61.2 11 (6.7) 12 (7.3) 23 (14.0) 9/6 (66.7) 
Internal medicine (152) 70.1 10 (6.6) 2 (1.3) 12 (7.9) 6/3 (50.0) 
Total (2462) 60.4 120 (4.9) 174 (7.1) 294 (12.0) 83/46 (55.4) 
Table III.

Frequency of antibiotic-associated diarrhoea (AAD) or mild illness (MI) according to presence of one or several concomitant diseases (the same patient may appear several times)

Concomitant disease Number (%) treated Number (%) with AAD Number (%) with MI Total number (%) 
aRisk ratio 3.01; CI 1.55–5.86; significantly different from patients with at most one concomitant disease (P = 0.001). 
No concomitant disease 1922 86 (4.5) 129 (6.7) 215 (11.2) 
Any concomitant disease 540 34 (6.3) 45 (8.3) 79 (14.6) 
One concomitant disease only 483 26 (5.4) 39 (8.1) 65 (13.5) 
Two or more concomitant diseases 57 8 (14.0)a 6 (10.5) 14 (24.6)b 
Diabetes only 234 12 (5.1) 24 (10.3) 36 (15.4) 
Diabetes + one or more other diseases 51 7 (13.7) 5 (9.8) 12 (23.5) 
Malignancy only 152 8 (5.3) 7 (4.6) 15 (9.9) 
Malignancy + one or more other diseases 23 4 (17.4) 4 (17.4) 8 (34.8) 
Chronic renal disease only 79 5 (6.3) 7 (8.9) 12 (15.2) 
Chronic renal disease + one or more other diseases 41 6 (14.6) 3 (7.3) 9 (21.9) 
Inflammatory bowel disease 18 1 (5.6) 1 (5.6) 2 (11.1) 
All evaluable 2462 120 (4.9) 174 (7.1) 294 (12.0) 
Concomitant disease Number (%) treated Number (%) with AAD Number (%) with MI Total number (%) 
aRisk ratio 3.01; CI 1.55–5.86; significantly different from patients with at most one concomitant disease (P = 0.001). 
No concomitant disease 1922 86 (4.5) 129 (6.7) 215 (11.2) 
Any concomitant disease 540 34 (6.3) 45 (8.3) 79 (14.6) 
One concomitant disease only 483 26 (5.4) 39 (8.1) 65 (13.5) 
Two or more concomitant diseases 57 8 (14.0)a 6 (10.5) 14 (24.6)b 
Diabetes only 234 12 (5.1) 24 (10.3) 36 (15.4) 
Diabetes + one or more other diseases 51 7 (13.7) 5 (9.8) 12 (23.5) 
Malignancy only 152 8 (5.3) 7 (4.6) 15 (9.9) 
Malignancy + one or more other diseases 23 4 (17.4) 4 (17.4) 8 (34.8) 
Chronic renal disease only 79 5 (6.3) 7 (8.9) 12 (15.2) 
Chronic renal disease + one or more other diseases 41 6 (14.6) 3 (7.3) 9 (21.9) 
Inflammatory bowel disease 18 1 (5.6) 1 (5.6) 2 (11.1) 
All evaluable 2462 120 (4.9) 174 (7.1) 294 (12.0) 
Table IV.

Frequency of antibiotic-associated diarrhoea (AAD) or mild illness (MI) according to duration of antibiotic treatment

Duration of treatment (days) Number treated Number (%) with AAD Number (%) with MI Total number (%) 
aRisk ratio 2.28; CI 1.23–4.21; significantly different from patients with longer duration of treatment (P = 0.009). 
bRisk ratio 1.89; CI 1.33–2.69; significantly different from patients with longer duration of treatment (P = 0.001). 
≤3 422 10 (2.4)a 19 (4.5) 29 (6.9)b 
4–7 286 15 (5.2) 15 (5.2) 30 (10.5) 
8–10 269 12 (4.5) 23 (8.6) 35 (13.0) 
10–21 932 46 (4.9) 74 (7.9) 120 (12.9) 
>21 553 37 (6.7) 43 (7.8) 80 (14.5) 
Total 2462 120 (4.9) 174 (7.1) 294 (12.0) 
Duration of treatment (days) Number treated Number (%) with AAD Number (%) with MI Total number (%) 
aRisk ratio 2.28; CI 1.23–4.21; significantly different from patients with longer duration of treatment (P = 0.009). 
bRisk ratio 1.89; CI 1.33–2.69; significantly different from patients with longer duration of treatment (P = 0.001). 
≤3 422 10 (2.4)a 19 (4.5) 29 (6.9)b 
4–7 286 15 (5.2) 15 (5.2) 30 (10.5) 
8–10 269 12 (4.5) 23 (8.6) 35 (13.0) 
10–21 932 46 (4.9) 74 (7.9) 120 (12.9) 
>21 553 37 (6.7) 43 (7.8) 80 (14.5) 
Total 2462 120 (4.9) 174 (7.1) 294 (12.0) 
Table V.

Frequency of antibiotic-associated diarrhoea (AAD) or mild illness (MI) when treated with either one group of antibiotics (‘only’) or in combination/sequentially with other antibiotics (‘other’)a

Antibiotic Number treated Number (%) with AAD Number (%) with MI Total number (%) 
aThe same patient may appear several times in the groups treated with several antibiotics. 
bIncludes ampicillin derivatives, pivmecillinam and piperacillin alone or in combination with tazobactam. 
Co-trimoxazole only 25 1 (4.0) 1 (4.0) 
Co-trimoxazole + other 89 5 (5.6) 5 (5.6) 10 (11.2) 
Quinolones only 117 12 (10.3) 12 (10.3) 
Quinolones + other 391 32 (8.2) 35 (9.0) 67 (17.1) 
Tetracycline only 50 1 (2.0) 1 (2.0) 
Tetracycline + other 171 3 (1.8) 10 (5.8) 13 (7.6) 
β-Lactamase-resistant organism     
penicillin only 392 13 (3.3) 20 (5.1) 33 (8.4) 
penicillin + other 302 16 (5.3) 23 (7.6) 39 (12.9) 
β-Lactamase-sensitive organism     
penicillin only 181 6 (3.3) 17 (9.4) 23 (12.7) 
penicillin + other 324 13 (4.0) 34 (10.5) 37 (14.4) 
Clindamycin only 40 2 (5.0) 2 (5.0) 4 (10.0) 
Clindamycin + other 119 13 (10.9) 17 (14.3) 30 (25.2) 
Cephalosporins only 330 20 (6.1) 19 (5.8) 39 (11.8) 
Cephalosporins + other 771 43 (5.6) 58 (7.5) 101 (13.1) 
Loracarbef + other 64 1 (1.6) 7 (10.9) 8 (12.5) 
Cefadroxil + other 317 16 (5.0) 16 (5.0) 32 (10.1) 
Cefpodoxime proxetil + other 28 8 (28.6) 2 (7.1) 10 (35.7) 
Ceftibuten + other 10 2 (20.0) 2 (20.0) 
Cefuroxime + other 223 11 (4.9) 21 (9.4) 32 (14.3) 
Cefotaxime + other 146 11 (7.5) 14 (9.6) 25 (17.1) 
Ceftazidime + other 35 4 (11.4) 5 (14.3) 9 (25.7) 
Broad-spectrum penicillina only 60 4 (6.7) 1 (1.7) 5 (8.3) 
Broad-spectrum penicillin + other 223 11 (4.9) 21 (9.4) 32 (14.3) 
Glycopeptides + other 40 2 (5.0) 4 (10.0) 6 (15.0) 
Imidazole derivatives + other 220 17 (7.7) 9 (4.1) 26 (11.8) 
Antibiotic Number treated Number (%) with AAD Number (%) with MI Total number (%) 
aThe same patient may appear several times in the groups treated with several antibiotics. 
bIncludes ampicillin derivatives, pivmecillinam and piperacillin alone or in combination with tazobactam. 
Co-trimoxazole only 25 1 (4.0) 1 (4.0) 
Co-trimoxazole + other 89 5 (5.6) 5 (5.6) 10 (11.2) 
Quinolones only 117 12 (10.3) 12 (10.3) 
Quinolones + other 391 32 (8.2) 35 (9.0) 67 (17.1) 
Tetracycline only 50 1 (2.0) 1 (2.0) 
Tetracycline + other 171 3 (1.8) 10 (5.8) 13 (7.6) 
β-Lactamase-resistant organism     
penicillin only 392 13 (3.3) 20 (5.1) 33 (8.4) 
penicillin + other 302 16 (5.3) 23 (7.6) 39 (12.9) 
β-Lactamase-sensitive organism     
penicillin only 181 6 (3.3) 17 (9.4) 23 (12.7) 
penicillin + other 324 13 (4.0) 34 (10.5) 37 (14.4) 
Clindamycin only 40 2 (5.0) 2 (5.0) 4 (10.0) 
Clindamycin + other 119 13 (10.9) 17 (14.3) 30 (25.2) 
Cephalosporins only 330 20 (6.1) 19 (5.8) 39 (11.8) 
Cephalosporins + other 771 43 (5.6) 58 (7.5) 101 (13.1) 
Loracarbef + other 64 1 (1.6) 7 (10.9) 8 (12.5) 
Cefadroxil + other 317 16 (5.0) 16 (5.0) 32 (10.1) 
Cefpodoxime proxetil + other 28 8 (28.6) 2 (7.1) 10 (35.7) 
Ceftibuten + other 10 2 (20.0) 2 (20.0) 
Cefuroxime + other 223 11 (4.9) 21 (9.4) 32 (14.3) 
Cefotaxime + other 146 11 (7.5) 14 (9.6) 25 (17.1) 
Ceftazidime + other 35 4 (11.4) 5 (14.3) 9 (25.7) 
Broad-spectrum penicillina only 60 4 (6.7) 1 (1.7) 5 (8.3) 
Broad-spectrum penicillin + other 223 11 (4.9) 21 (9.4) 32 (14.3) 
Glycopeptides + other 40 2 (5.0) 4 (10.0) 6 (15.0) 
Imidazole derivatives + other 220 17 (7.7) 9 (4.1) 26 (11.8) 
Table VI.

Frequency of positive C. difficile cytotoxin B test among patients with antibiotic-associated diarrhoea (AAD), mild illness (MI) and asymptomatic patients treated on different wards in centre I

 Number of patients tested/number positive for cytotoxin B (% positive) 
Type of ward (number treated) with AAD with MI asymptomatic 
aRisk ratio 1.69; CI 1.04–2.73; percentage of positive cytotoxin B tests significantly different from that in asymptomatic patients (P = 0.034). 
bRisk ratio 1.85; CI 1.24–2.77; percentage of positive cytotoxin B tests significantly different from that in asymptomatic patients (P = 0.003). 
Infectious diseases (305) 17/12 (70.6)a 31/11 (35.5) 74/31 (41.9) 
Nephrology (63) 6/3 (50.0) 4/1 (25) 10/4 (40.0) 
Geriatrics (40) 7/5 (71.4) 4/0 (0) 9/2 (22.2) 
Orthopaedics (114) 2/0 (0) 2/0 (0) 46/10 (21.7) 
Total (522) 32/20 (62.5)b 41/12 (29.3) 139/47 (33.8) 
 Number of patients tested/number positive for cytotoxin B (% positive) 
Type of ward (number treated) with AAD with MI asymptomatic 
aRisk ratio 1.69; CI 1.04–2.73; percentage of positive cytotoxin B tests significantly different from that in asymptomatic patients (P = 0.034). 
bRisk ratio 1.85; CI 1.24–2.77; percentage of positive cytotoxin B tests significantly different from that in asymptomatic patients (P = 0.003). 
Infectious diseases (305) 17/12 (70.6)a 31/11 (35.5) 74/31 (41.9) 
Nephrology (63) 6/3 (50.0) 4/1 (25) 10/4 (40.0) 
Geriatrics (40) 7/5 (71.4) 4/0 (0) 9/2 (22.2) 
Orthopaedics (114) 2/0 (0) 2/0 (0) 46/10 (21.7) 
Total (522) 32/20 (62.5)b 41/12 (29.3) 139/47 (33.8) 
*
Corresponding author. Tel: +46-90-7852305; Fax: +46-90-133006; E-mail: johan.wistrom.us@vll.se

We wish to acknowledge the assistance of the staff at the participating centres, especially that of Ami Hommel, Eva Hynning-Lechter, Ann Johansson, Kurt Kero, Sonia Lundberg, Annika Sundholm, Kristina Svensson, Ulla Britt Thollström and Solveig Wennerholm. We also wish to acknowledge the skilful technical assistance of AnnChatrin Palmgren and Olle Stuermer. This study was supported by a grant from Astra Läkemedel AB, Södertälje, Sweden.

References

1
Högenauer, C., Hammer, H. F., Krejs, G. J. & Reisinger, E. C. (
1998
). Mechanisms and management of antibiotic-associated diarrhea.
Clinical Infectious Diseases
 
27
,
702
–10.
2
Tabaqchali, S. & Wilks, M. (
1992
). Epidemiological aspects of infections caused by Bacteroides fragilis and Clostridium difficile.
European Journal of Clinical Microbiology and Infectious Diseases
 
11
,
1049
–57.
3
Bartlett, J. G. (
1996
). Management of Clostridium difficile infection and other antibiotic-associated diarrhoeas.
European Journal of Gastroenterology and Hepatology
 
8
,
1054
–61.
4
Spencer, R. C. (
1998
). Clinical impact and associated costs of Clostridium difficile-associated disease.
Journal of Antimicrobial Chemotherapy
 
41
, Suppl. C
5
–12.
5
Eriksson, S. & Aronsson, B. (
1989
). Medical implications of nosocomial infection with Clostridium difficile.
Scandinavian Journal of Infectious Diseases
 
21
,
733
–4.
6
Wilcox, M. H., Cunniffe, J. G., Trundle, C. & Redpath, C. (
1996
). Financial burden of hospital-acquired Clostridium difficile infection.
Journal of Hospital Infection
 
34
,
23
–30.
7
MacGowan, A. P., Feeney, R., Brown, I., McCulloch, S. Y., Reeves, D. S. & Lovering, A. M. (
1997
). Health care resource utilization and antimicrobial use in elderly patients with community-acquired lower respiratory tract infection who develop Clostridium difficile-associated diarrhoea.
Journal of Antimicrobial Chemotherapy
 
39
,
537
–41.
8
Spencer, R. C. (
1998
). The role of antimicrobial agents in the aetiology of Clostridium difficile-associated disease.
Journal of Antimicrobial Chemotherapy
 
41
, Suppl. C
21
–7.
9
Bignardi, G. E. (
1998
). Risk factors for Clostridium difficile infection.
Journal of Hospital Infection
 
40
,
1
–15.
10
Aronsson, B., Möllby, R. & Nord, C. E. (
1982
). Clostridium difficile and antibiotic associated diarrhoea in Sweden.
Scandinavian Journal of Infectious Diseases Supplementum
 
35
,
53
–8.
11
Impallomeni, M., Galletly, N. P., Wort, S. J., Starr, J. M. & Rogers, T. R. (
1995
). Increased risk of diarrhoea caused by Clostridium difficile in elderly patients receiving cefotaxime.
British Medical Journal
 
311
,
1345
–6.
12
Pear, S. M., Williamson, T. H., Bettin, K. M., Gerding, D. N. & Galgiani, J. N. (
1994
). Decrease in nosocomial Clostridium difficile-associated diarrhea by restricting clindamycin use.
Annals of Internal Medicine
 
120
,
272
–7.
13
Westh, H., Iversen, J. T. & Gyrtrup, H. J. (
1991
). Clostridium difficile in faecal flora after perioperative prophylaxis with ampicillin or ceftriaxone.
Journal of Infection
 
23
,
347
–50.
14
Surawicz, C. M., Elmer, G. W., Speelman, P., McFarland, L. V., Chinn, J. & van Belle, G. (
1989
). Prevention of antibiotic-associated diarrhea by Saccharomyces boulardii: a prospective study.
Gastroenterology
 
96
,
981
–8.
15
McFarland, L. V., Surawicz, C. M., Greenberg, R. N., Elmer, G. W., Moyer, K. A., Melcher, S. A. et al. (
1995
). Prevention of β-lactam-associated diarrhea by Saccharomyces boulardii compared with placebo.
American Journal of Gastroenterology
 
90
,
439
–48.
16
Lewis, S. J., Potts, L. F. & Barry, R. E. (
1998
). The lack of therapeutic effect of Saccharomyces boulardii in the prevention of antibiotic-related diarrhoea in elderly patients.
Journal of Infection
 
36
,
171
–4.
17
Thamlikitkul, V., Danpakdi, K. & Chokloikaew, S. (
1996
). Incidence of diarrhea and Clostridium difficile toxin in stools from hospitalized patients receiving clindamycin, β-lactams, or nonantibiotic medications.
Journal of Clinical Gastroenterology
 
22
,
161
–3.
18
Aronsson, B., Möllby, R. & Nord, C.E. (
1985
). Antimicrobial agents and Clostridium difficile in acute enteric disease: epidemiological data from Sweden, 1980–1982.
Journal of Infectious Diseases
 
151
,
476
–81.
19
McFarland, L. V., Mulligan, M. E., Kwok, R. Y. & Stamm, W. E. (
1989
). Nosocomial acquisition of Clostridium difficile infection.
New England Journal of Medicine
 
320
,
204
–10.
20
Orrhage, K., Brismar, B. & Nord, C. E. (
1994
). Effects of supplements with Bifidobacterium longum and Lactobacillus acidophilus on the intestinal microbiota during administration of clindamycin.
Microbial Ecology in Health and Disease
 
7
,
17
–25.
21
Brazier, J. S. (
1998
). The epidemiology and typing of Clostridium difficile.
Journal of Antimicrobial Chemotherapy
 
41
, Suppl. C
47
–57.
22
Tullus, K., Aronsson, B., Marcus, S. & Möllby, R. (
1989
). Intestinal colonization with Clostridium difficile in infants up to 18 months of age.
European Journal of Clinical Microbiology and Infectious Diseases
 
8
,
390
–3.
23
Simor, A. E., Yake, S. L. & Tsimidis, K. (
1993
). Infection due to Clostridium difficile among elderly residents of a long-term-care facility.
Clinical Infectious Diseases
 
17
,
672
–8.
24
Samore, M. H., DeGirolami, P. C., Tlucko, A., Lichtenberg, D. A., Melvin, Z. A. & Karchmer, A. W. (
1994
). Clostridium difficile colonization and diarrhea at a tertiary care hospital.
Clinical Infectious Diseases
 
18
,
181
–7.
25
Clabots, C. R., Johnson, S., Olson, M. M., Peterson, L. R. & Gerding, D. N. (
1992
). Acquisition of Clostridium difficile by hospitalized patients: evidence for colonized new admissions as a source of infection.
Journal of Infectious Diseases
 
166
,
561
–7.
26
Chachaty, E., Depitre, C., Mario, N., Bourneix, C., Saulnier, P., Corthier, G. et al. (
1992
). Presence of Clostridium difficile and antibiotic and β-lactamase activities in feces of volunteers treated with oral cefixime, oral cefpodoxime proxetil, or placebo.
Antimicrobial Agents and Chemotherapy
 
36
,
2009
–13.
27
Karlström, O., Fryklund, B., Tullus, K. & Burman, L. G. (
1998
). A prospective nationwide study of Clostridium difficile-associated diarrhea in Sweden. The Swedish C. difficile Study Group.
Clinical Infectious Diseases
 
26
,
141
–5.
28
McFarland, L. V., Surawicz, C. M. & Stamm, W. E. (
1990
). Risk factors for Clostridium difficile carriage and C. difficile-associated diarrhea in a cohort of hospitalized patients.
Journal of Infectious Diseases
 
162
,
678
–84.
29
Edlund, C., Stark, C. & Nord, C. E. (
1994
). The relationship between an increase in β-lactamase activity after oral administration of three new cephalosporins and protection against intestinal ecological disturbances.
Journal of Antimicrobial Chemotherapy
 
34
,
127
–38.