Oral antibiotics are often prescribed for children, especially for upper respiratory tract and chest infections, because of the fear of a bacterial cause or bacterial superinfection complicating a viral infection. Adverse effects of antibiotics include various digestive disorders, among which diarrhea occupies a special place: it causes anxiety for parents, justified by the risk of dehydration with which it is associated, especially in children aged less than 2 years (1). Although antibiotic‐associated diarrhea (AAD) has been studied in adult hospitalized patients, there is little available information concerning the epidemiologic characteristics from large, prospective studies of AAD in children. In addition, outpatient studies may reduce the difficulties of interpreting hospital studies undertaken in an environment far removed from the everyday life of children.
The objective of this prospective study was to evaluate the incidence of AAD in children in an outpatient practice and to attempt to determine risk factors for AAD.
PATIENTS AND METHODS
The Research and Study Group in Ambulatory Paediatrics of Northern France performed this epidemiologic study. This group consists of eight pediatricians with partial or total activity in outpatient pediatrics. For 11 consecutive months, patients meeting the following selection criteria were routinely enrolled: children aged 1 month to 15 years, 3 months (French legal definition of the end of childhood), seen as pediatric outpatients, with proven or suspected infection justifying oral antibiotic treatment for at least 5 days, and whose families appeared capable of participating in the study. All the children were seen as outpatients. Patients who had had diarrhea during the week before inclusion or who had taken an antibiotic during the 2 weeks before inclusion, and patients with immune deficiency were not enrolled.
The study protocol involved 2 evaluations: an initial visit when children were enrolled, and a second evaluation (possibly by telephone) 1 week after antibiotics were stopped. Information was recorded using a case report form noting the following variables: gender, age, weight and height, medical and surgical history, diagnosis justifying antibiotic treatment, treatments prescribed before and during antibiotic treatment, and tolerability of treatment. Parents also completed a special daily diary describing intestinal function throughout the antibiotic treatment period and for 1 week after it was stopped. Parents were asked whether they considered each stool produced during this period to be normal, soft, or liquid. AAD was defined by the daily production of at least 3 soft or liquid stools for at least 2 consecutive days.
Analysis of AAD risk factors concerned the age of the child, his or her medical and surgical history, the type of infection leading to antibiotic treatment, the antibiotic prescribed, and the duration of its administration. Antibiotic preparations did not contain sorbitol or fructose in the liquid syrup. Cow's milk protein‐free diet and lactose‐free diet were not associated with the antibiotic treatment. SAS software (SAS Institute, Cary, NC) was used for statistical analysis. Differences in the incidence of risk factors were compared using the χ2 test or Fisher exact probability test for qualitative variables, and a nonparametric Kruskal‐Wallis test or a Student t test was used for quantitative variables. The Bonferroni Holms procedure was used for multiple comparisons to retain a global risk of 5% by adaptation of the decisional threshold to each comparison. Relative risks for risk factors were calculated as the ratio of the incidence of AAD in exposed children divided by the incidence of AAD in children not exposed to the risk factor.
Six hundred fifty‐nine children were enrolled in the study, and 650 with analyzable cases were included during a period of 11 consecutive months. The 345 boys (53%) and 305 girls (47%) were aged 34.1 ± 33.7 months (mean ± SD). Three hundred thirty‐six children (52%) were between 1 and 24 months, and 314 (48%) were between 2 years and 15 years, 3 months. They weighed between 4 kg and 64 kg, with heights between 52 cm and 176 cm. The duration of antibiotic treatment (mean ± SD) was 8.6 ± 1.4 days (range, 5‐18 days) with a total mean monitoring period (from the start of antibiotic treatment until 1 week after it was stopped) of 15.2 ± 2.2 days (range, 5‐20 days). Two hundred twelve children, i.e., one third of those studied, had a history of a medical condition other than the antibiotic‐related infection before enrollment. The most common diagnosis was infectious disease (160 children, i.e., 25%, with upper respiratory tract involvement in 150 of them). Only 11 children (1.7%) had any notable digestive history: gastroesophageal reflux, 8; toddler diarrhea, 2; and cystic fibrosis, 1. No child was lost to follow‐up review.
The reason for prescription of antibiotic treatment was tonsillitis or rhinopharyngitis, 310 (48%); acute otitis media or sinusitis, 195 (30%); and chest infection, 105 (16%). Other reasons for prescription were skin infection, 13 (2%); urinary tract infection, 10 (1.5%); and miscellaneous, 17 (2.5%).
Antibiotics used, prescribed according to diagnosis, are listed in Table 1. Penicillins A and M, cephalosporins, and macrolides were frequently prescribed. One third of antibiotic prescriptions were penicillins A and M (this group not including amoxicillin/clavulanate, which alone accounted for 9% of the antibiotics prescribed).
Five hundred seventy‐eight medications were taken in combination with antibiotics. The medications were theophylline (18%), mucolytics (14.5%), antitussives (14%), antipyretics/analgesics (12.5%), antiinflammatories (10%), nose drops (6%), other systemic treatments (22.5%), and other topical treatments (2.5%). The high rate of prescription of theophylline was the result of its use in patients aged more than 30 months with asthmatic symptoms.
AAD Seen in the Overall Cohort
Among the 650 children, 71 (11%) had an episode of AAD, occurring under the following circumstances: solely during antibiotic treatment, 48 (68%); solely during the week after antibiotics were stopped, 11 (15%); or during antibiotic treatment and the week after it was stopped, 12 (17%). The time between the start of antibiotic treatment and onset of AAD (mean ± SD) was 5.3 ± 3.5 days (range, 1‐15 days). The duration of AAD (mean ± SD) was 4.0 ± 3.0 days (range, 2‐17 days). Cumulative days of diarrhea (mean ± SD) were 21% ± 17% of the observation period (range, 6‐100%). This study was not intended to assess the severity or seriousness of AAD; however, none of the patients included in this study required hospitalization. The only adverse event evaluated in this study was the onset of diarrhea, excluding any other side effect commonly seen during antibiotic treatment (vomiting, abdominal pain, and so on). Initial antibiotic treatment was stopped in 4 of the 650 children (0.6%) for vomiting (n = 2) or diarrhea (n = 2). The antibiotic was changed for only one patient who was seen by another doctor; however, he reported no adverse effects that might be related to the initial antibiotic treatment, especially diarrhea or any other digestive symptom.
The incidence of these AAD episodes was significantly greater in children less than 2 years (61 of 336 = 18%) than in those more than 2 years (10 of 314 = 3%; P < 0.001). The relative risk of onset of an episode of diarrhea in a child less than 2 years was 1.81 (range, 1.50‐2.14). In the group of children more than 2 years, the incidence of AAD was greater in the youngest of them (2‐7 years; 9 of 253 = 4%) than in the older patients (>7 years; 1 of 61 = 2%; Fig. 1), but the difference was not significant. Children with an episode of AAD were younger than those who did not have episodes of diarrhea (13.4 ± 14.7 months vs. 36.6 ± 34.5 months; P < 0.001).
In this study, the rate of onset of AAD differed significantly (P = 0.012) according to the type of antibiotic prescribed: penicillins G and V, 3%; penicillins A and M (except amoxicillin/clavulanate), 11%; amoxicillin/clavulanate, 23%; cephalosporins, 9%; macrolides, 8%; trimethoprim/sulfamethoxazole, 6%; and erythromycin/sulfafurazole, 16% (Fig. 2). There was a statistically significant difference between the rate of onset of AAD associated with amoxicillin/clavulanate compared with all other antibiotics combined (P = 0.003). This difference was statistically significant when diarrhea occurred during the treatment period, but it was not significant if diarrhea occurred after the antibiotics were stopped. The relative risk of onset of an episode of diarrhea in a child receiving amoxicillin/clavulanate was 2.43 (range, 1.4—4.21) and 3.5 (1.89–6.46) when the child also was aged less than 2 years. The incidence of AAD was not influenced by the child's medical or surgical history or by the nature of the infection that led to antibiotic treatment.
This study, in which the diagnostic criterion of diarrhea was particularly strict because it included only episodes of diarrhea lasting at least 2 consecutive days, found an 11% incidence of diarrhea occurring during or after antibiotic treatment. Two risk factors were identified: the young age of the child and the type of antibiotic treatment.
To the best of our knowledge, no epidemiologic study in an outpatient population of children has been published. Studies of AAD are most often focused studies evaluating the responses to only one specific group of antibiotics (2,3). There have been several studies of outpatient children receiving antibiotics for otitis media (4—6) or for upper respiratory infections (7). However, the authors did not provide any epidemiologic data. AAD was found in children given amoxicillin/clavulanate (28.9% (4), 32% (5), 71% (6)) or sultamicillin (6%). The only other studies in children involve pediatric C. difficile disease, which represents a subpopulation of AAD cases, and all of these studies were from hospitalized pediatric patients (8).
In a placebo‐controlled study testing the use of the probiotic organism Lactobacillus GG for the prevention of AAD in children with a median age of 4 years, Vanderhoof et al. (9) reported AAD in 26% of the placebo‐treated patients. AAD was defined in this study as the presence of at least 2 liquid stools per day on at least 2 observation periods during the 10‐day course of the study. This definition of diarrhea is less severe than that given in Food and Drug Administration and World Health Organization guidelines, namely, the production of at least 3 soft or liquid stools for a period of at least 24 hours but less than 48 hours (10,11). In a group of children with a mean age of 4.4 years treated with oral antibiotics for an acute respiratory infection, Arvola et al. (12) reported a 16% incidence of AAD during the first 2 weeks after the beginning of the antimicrobial treatment. These authors used the same definition of diarrhea as used in the present study.
In adults, studies of AAD have been limited to either hospitalized patients (13,14) or C. difficile‐associated diarrhea in an ambulatory population (15,16). McFarland et al. (17) reported an 8% incidence of C. difficile associated diarrhea in adults hospitalized in a single ward. The two studies of adult outpatients reported a low incidence of C. difficile‐associated diarrhea (7.7 to 12 cases/100,000) and did not report the frequency of AAD from other causes (15,16).
Some antibiotics are associated with a particularly high incidence of diarrhea. Unfortunately, the definition of AAD is most often imprecise, or totally absent, in publications (5). A survey reported by Kramer et al. (18) was designed to evaluate the nature and incidence of gastrointestinal adverse effects in a cohort of 2,714 children receiving antibiotic treatment. They reported a 3.6% frequency of diarrhea, but there was no definition of diarrhea given. Kramer et al. also noted the influence of the type of antibiotic taken. The relative risk of diarrhea was between 3 and 5 for penicillins V, amoxicillin, and nystatin; 6.5 for a first‐generation cephalosporin; and 10.2 for cloxacillin. The study protocol involved 2 telephone interviews: one 2 to 4 days after prescription and the other 2 weeks later (or sooner if antibiotic treatment was stopped prematurely). Parents were not asked to complete a diary between these two interviews, which might explain the lower incidence of diarrhea reported in the study compared with our own.
The daily observation diary completed by parents in this study provided the most reliable notification possible of the actual onset of diarrhea. Nevertheless, it is possible that this incidence may have been underestimated because of the fairly short observation period (1 week after antibiotic treatment). It is known that AAD can occur up to 6 weeks after antibiotics are stopped. If this was the case in our study, a number of AAD cases were not taken into account, and the actual incidence of AAD may have been underestimated. The choice of this short observation period was motivated by a desire to limit the number of children lost to follow‐up review. The possibility that diarrhea may not have been the result of the antibiotic treatment but rather of the infectious disease for which antibiotics were prescribed is very unlikely. The incidence of AAD was independent of the site of infection, although it was nevertheless influenced, for a given site of infection, by the type of antibiotic prescribed.
The pathophysiology of AAD remains controversial (13,19). Antibiotic‐induced imbalance of the intestinal flora may predispose patients to the implantation of pathogenic organisms, e.g., C. difficile or bring about untoward changes in intestinal metabolism. Data concerning the potential role of C. difficile in AAD in children are contradictory and even more difficult to interpret given that newborns and young infants are often carriers of C. difficile in the absence of any clinical symptoms (4,8,20). In the present study, penicillins A and M, which have a broad spectrum of action, frequently resulted in AAD. It is not known why clavulanate increases the incidence of diarrhea when it is associated with amoxicillin.
Further studies defining the harmful effects of certain antibiotics and the relative sensitivity of younger children will provide more information about the origins of AAD and ways in which it can be prevented.
1. Kramer MS, Hutchinson TA, Flegel KM, et al. Adverse drug reactions in general pediatric outpatients. J Pediatr
2. Thompson JW, Jacobs RF. Adverse effects of newer cephalosporins. An update. Drug Saf
3. Periti P, Mazzei T, Mini E, et al. Adverse effects of macrolide antibacterials. Drug Saf
4. Mitchell DK, Van R, Mason EH, et al. Prospective study of toxigenic Clostridium difficile
in children given amoxicillin/clavulanate for otitis media. Pediatr Infect Dis J
5. McCarty JM, Phillips A, Wiisanen R. Comparative safety and efficacy of clarithromycin and amoxicillin/clavulanate in the treatment of acute otitis media in children. Pediatr Infect Dis J
1993; 12(suppl 3):S122-7.
6. Feldman W, Sutcliffe T, Dulberg C. Twice-daily antibiotics in the treatment of acute otitis media: trimethoprim-sulfamethoxazole versus amoxicillin-clavulanate. CMAJ
7. Raillard P, Feiner C, Ott V, et al. Worldwide pediatric experience with low-dose sultamicillin oral suspension. Curr Ther Res
8. McFarland LV, Brandmarker SA, Guandalini S. Pediatric C. difficile
: a phantom menace or clinical reality? J Pediatr Gastroenterol Nutr
9. Vanderhoof JA, Whitney DB, Antonson DL, et al. Lactobacillus GG
in the prevention of antibiotic-associated diarrhea in children. J Pediatr
10. Guidelines for the clinical evaluation of antidiarrheal drugs. US Department of Health, Education and Welfare, Public Health Service, Food and Drug Administration. US Government Printing Office; 1977;732-355/237.
11. The treatment of diarrhoea. A manual for physicians and other senior health workers. WHO/CDR/95.3 10/95.
12. Arvola T, Laiho K, Torkkeli S, et al. Prophylactic Lactobacillus GG
reduces antibiotic-associated diarrhea in children with respiratory infections: a randomized study. Pediatrics
13. McFarland LV. Epidemiology, risk factors and treatments for antibiotic-associated diarrhea. Dig Dis Sci
14. McFarland LV, Surawicz CM, Greenberg RN, et al. Prevention of beta-lactam associated diarrhea by Saccharomyces boulardii
compared with placebo. Am J Gastroenterol
15. Hirschhorn L, Trnka Y, Onderdonk A, et al. Epidemiology of community-acquired C.difficile
-associated diarrhea. J Infect Dis
16. Levy DG, Stergachis A, McFarland LV, et al. Antibiotics and Clostridium difficile
diarrhea in the ambulatory care setting. Clin Ther
17. McFarland LV, Surawicz CM, Stamm WE. Risk factors for Clostridium difficile
carriage and C. difficile
-associated diarrhea in a cohort of hospitalized patients. J Infect Dis
18. Kramer MS, Hutchinson TA, Naimark L, et al. Antibiotic-associated gastrointestinal symptoms in general pediatric outpatients. Pediatrics
19. Hogenauer C, Hammer HF, Krejs GJ, et al. Mechanisms and management of antibiotic-associated diarrhea. Clin Infect Dis
20. Dutta P, Niyogi SK, Mitra U, et al. Clostridium difficile
in antibiotic associated pediatric diarrhea. Indian Pediatr
Key Words:: Antibiotics; Diarrhea; Antibiotic‐associated diarrhea