Epidemiologic surveillance of Clostridium difficile diarrhea in a freestanding pediatric hospital and a pediatric hospital at a university medical center

Diagnostic Microbiology and Infectious Disease 56 (2006) 109 – 114 www.elsevier.com/locate/diagmicrobio

Bacteriology

Epidemiologic surveillance of Clostridium difficile diarrhea in a freestanding pediatric hospital and a pediatric hospital at a university medical centerB Carmen E. Rexacha,4, Yajarayma J. Tang-Feldmanb, Mary C. Cantrellb, Stuart H. Cohenb b

a Division of Natural Sciences, University of California, Merced, CA 95344, USA Department of Internal Medicine, Division of Infectious Diseases, University of California, Davis, Davis, CA 95616, USA Received 17 November 2005; accepted 8 March 2006

Abstract To describe the epidemiology of Clostridium difficile in children, we cultured stool specimens from patients at the Children’s Hospital Central California, Madera, CA (CHCC, n = 676) and at the University of California Davis Medical Center Pediatric Hospital, Sacramento, CA (UCDMC-PH, n = 301) for C. difficile, and toxins A and B genes and strain identity of the isolates were determined by polymerase chain reaction assays. A higher percentage of patients from UCDMC-PH were culture positive (148/301, 49%) and colonized with toxigenic strains (45/301, 15%) compared with CHCC (colonized = 178/676, 26%; toxigenic = 96/676, 14%, P V .001). Multiple logistic regression analysis showed decreased colonization with inpatient status (odds ratio [OR] = 0.64; 95% confidence interval [CI] = 0.46, 0.89; P = .007) and use of H-2 antagonists (OR = 0.55; 95% CI = 0.36, 0.84; P = .006), whereas underlying conditions (colonization: OR = 1.42; 95% CI = 1.02, 1.96; P = .04; toxin positive: OR = 1.60; 95% CI = 1.04, 2.44; P = .03) and exposure to z 2 antiinfectives (colonization: OR = 1.56; 95% CI = 1.10, 2.20; P = .01; toxin positive: OR = 1.71; 95% CI = 1.10, 2.66; P = .02) increased colonization. Most isolates appear to be community acquired, although molecular analysis suggests some nosocomial transmission at UCDMC-PH. These data suggest that the epidemiology of colonization with C. difficile in children is different than previously reported. D 2006 Elsevier Inc. All rights reserved. Keywords: Clostridium difficile; Infectious disease; Pediatrics; Epidemiology; Diarrhea

1. Introduction Clostridium difficile is a Gram-positive rod that exists as either a toxigenic organism producing toxins A and B or as a nontoxigenic organism that lacks these toxins. A third toxin, binary toxin, has recently been identified and may be a virulence factor (Gonc¸alves et al., 2004). As an agent of human disease, toxin-producing C. difficile was first described as a major cause of antibiotic-associated diarrhea and pseudomembranous colitis in adult patients (Bartlett, 1994; Lyerly et al., 1988; McFarland, 1993; Simor et al.,

B

This study was supported by the UC Davis Health Systems Grant. 4 Corresponding author. Division of Natural Sciences, University of California, Merced, PO Box 2039, Merced, CA 95344, USA. Tel.: +1-209613-2196; fax: +1-714-530-5922. E-mail address: [email protected] (C.E. Rexach). 0732-8893/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.diagmicrobio.2006.03.002

1993). Risk factors for C. difficile-associated disease (CDAD) in these patients included exposure to antibiotics, such as clindamycin, ampicillin, and third generation cephalosporins, increased age, and long-term hospitalization (Genta et al., 1984). Documented outbreaks of CDAD in geriatric patients also established the importance of C. difficile as a cause of nosocomial diarrhea. Recent studies have demonstrated that CDAD is increasing in incidence and scope in adult patients (Frost et al., 1998). However, its role in disease in children remains controversial. Early studies demonstrated high asymptomatic colonization in neonates, and largely due to these studies, C. difficile was not considered as a possible etiologic agent when children developed diarrhea (Bartlett, 1994). Recently, the importance of diarrhea and acute gastroenteritis in pediatric morbidity and mortality worldwide has been documented and has spurred additional studies into the epidemiology of CDAD in children (Bartlett, 1994). Increasing evidence now

110

C.E. Rexach et al. / Diagnostic Microbiology and Infectious Disease 56 (2006) 109–114

suggests that C. difficile may be a more important pathogen in children than previously appreciated (de Boissieu et al., 1996; Gaglani et al., 1996; Harris and Figueroa-Colon, 1995; Kato et al., 1995; Langley et al., 2002; McFarland et al., 2000). A possible protective role for colonization with both toxigenic and nontoxigenic strains has been suggested and has also increased interest in better understanding the effect of colonization with C. difficile in pediatric populations (Wanahita et al., 2002). We report the results of our epidemiologic surveillance of C. difficile colonization in pediatric patients at the Children’s Hospital Central California (CHCC) and the University of California Davis Medical Center Pediatric Hospital (UCDMC-PH). These results suggest that the epidemiology of C. difficile in children may be different than that reported for adults. 2. Materials and methods 2.1. Patient population 2.1.1. Children’s Hospital Central California Samples obtained from 2 types of facilities, a freestanding pediatric hospital and a hospital serving all ages, were evaluated in this study. The freestanding pediatric hospital was represented by the Valley Children’s Hospital in Fresno, CA, and CHCC in Madera, CA. Valley Children’s Hospital in Fresno was a 186-bed pediatric hospital housing on average 4 patients per room. This facility was closed on September 1, 1998, and all patients remaining in the hospital were transferred to the newly constructed CHCC. Children’s Hospital Central California is a 242-bed pediatric hospital with individual rooms for each patient organized in each ward around a central nurse’s station. Because both hospitals are freestanding pediatric hospitals and 1 hospital replaced the other, the pediatric population served by these 2 facilities is the same, and the subjects are treated as 1 cohort for the purposes of this study and are referred to collectively henceforth as CHCC. The total number of samples from CHCC is 676 and represents all samples that were sent to the Clinical Microbiology Laboratory during the period July 26, 1998, and January 1, 1999, from inpatients and outpatients for evaluation of disease. There were no asymptomatic patients. 2.1.2. University of California Davis Medical Center Pediatric Hospital pediatric hospital The UCDMC-PH is a 100-bed hospital housed in a dedicated portion of the 7th floor of the UCDMC tower within the 525-bed UCDMC hospital serving all ages. The pediatric hospital includes a pediatric intensive care unit and a neonatal intensive care unit (NICU). The NICU is located on the 3rd floor. During the study period, October 1, 2001, to January 16, 2003, all samples submitted to the Virology Department of the Clinical Microbiology Laboratory at UCDMC for evaluation of disease were included.

A total of 301 samples from both inpatients and outpatients were obtained. 2.2. Isolation of C. difficile from patient stools Stool specimens from study subjects were collected from the laboratory and cultured on prereduced cycloserine cefoxitin fructose agar (CCFA) with sodium taurocholate at 37 8C in BBL jars under anaerobic conditions for 36 to 48 h (Gumerlock et al., 1993; Tang et al., 1994). Isolates were confirmed as C. difficile by the Culturette CDT rapid latex agglutination test (Becton-Dickinson, Cockeysville, MD). Latex agglutination test positive isolates were purified by streak isolation, and stock cultures were prepared using meat broth containing thioglycolate. All isolations and laboratory tests were performed in the same laboratory. All patient samples were refrigerated in the respective clinical laboratories until cultured. Samples from CHCC were cultured on site in the clinical laboratories and transported in BBL jars under anaerobic conditions within 2 h of culture to UCDMC research laboratory. Samples from UCDMC were obtained from the clinical laboratories and immediately transported to the research laboratory in an adjacent building. 2.3. Preparation of DNA and PCR of toxins A and B genes To prepare the isolate for polymerase chain reaction (PCR), we streaked a sample from the stock culture on CCFA plates and grown a sample under anaerobic conditions for 48 h. An isolated colony was placed into 100 AL of distilled water and boiled for 10 min. The supernatant containing a crude DNA extract was then used for PCR analysis. The presence of toxins A and B gene sequences was determined by PCR as previously described (Gumerlock et al., 1993; Tang et al., 1994). 2.4. Deoxyribonucleic acid fingerprinting Arbitrarily primed PCR (AP-PCR) was used to fingerprint each of the isolates. Arbitrarily primed PCR was performed as previously described (Silva et al., 1993; Tang et al., 1995) with the arbitrary primer T7. The validity of this method has been described previously (Cohen et al., 2001; Gumerlock et al., 1993; Tang et al., 1994). 2.5. Analysis of amplified products Amplified products from toxins A and B genes were visualized by running 9 AL of PCR product on 2% agarose gels. For AP-PCR, 10.5 AL of PCR product was run in 1.5% agarose gels. All gels were run for 90 min at 110 V after standard protocols. 2.6. Molecular analysis Arbitrarily primed PCR banding patterns produced by gel electrophoresis were visually compared and categorized as indistinguishable or unique. Isolates with indistinguishable patterns were grouped together and given a cluster number. These clusters were evaluated by date of

C.E. Rexach et al. / Diagnostic Microbiology and Infectious Disease 56 (2006) 109–114

sample, inpatient/outpatient status, prior hospitalization, and hospital from which specimen was obtained (CHCC or UCDMC-PH). 2.7. Statistical analysis An unmatched analysis was performed comparing prevalence determinants in study subjects from CHCC (676 total) with those from UCDMC-PH (n = 232) using logistic regression and the EGRET statistical program (Cytel Software, Cambridge, MA). The following exposure variables were considered in the analysis: inpatient status, sex, age, culture positive or negative, toxin positive or negative, oncology status, treatment with antibiotics, H-2 antagonists, antineoplastics, exposure to 2 or more antiinfectives, exposure to 4 or more antiinfectives, prior hospitalization, prior colonization with toxin-positive C. difficile, presence of other potentially pathogenic microorganism in the stool, and underlying condition. Underlying condition was defined as any medical condition requiring frequent and long-term medical intervention, such as cystic fibrosis, cancer, autoimmune disorders, and so on. Current chemotherapeutic use was defined as exposure that began no more than 1 month before date of sample and no less than 24 h before stool sample or ended within 5 days of date the stool specimen was obtained. Inpatient status was defined as patients who were hospitalized for a minimum of 48 h. Information for these variables was abstracted from the patient’s medical records. Two outcomes were considered in the analysis: colonization with C. difficile and colonization with a toxin-positive isolate. Results were reported as prevalence odds ratios (OR) and 95% confidence intervals (CI).

3. Results 3.1. Demographics Pediatric stool specimens from 676 patients from CHCC and 301 from UCDMC-PH (n = 977) were cultured and assayed for the presence of C. difficile and genes for toxins A and B. A higher percentage of the specimens from UCDMC-PH were culture positive (148/301, 49%) compared with 178/676 (26%) at CHCC ( P V .001) (Table 1). Toxigenic strains were identified in 96/676 (14%) specimens from CHCC or 54% (96/178) of the culture-positive isolates, and in 45/301 (15%) specimens or 30% (45/148) of the culture-positive isolates from UCDMC-PH. The average age of children treated at CHCC and UCDMC-PH was 3.5 F 4.7 and 5.3 F 5.6 years, respectively (range, 1 month to 19 years, P V .001). The median age for the patients was 1.3 years for children from CHCC and 2.4 years for UCDMC-PH. Fifty-three percent of study subjects at CHCC were younger than 18 months compared with 41% of patients from UCDMC ( P V .001) (Table 2). The percentage of samples coming from inpatients was similar in each of these populations (56%

111

CHCC and 54% UCDMC-PH). Although the sex distribution in each facility was also similar (56% male and 44% female at CHCC, 57% male and 43% female at UCDMCPH), a greater percentage of the culture-positive patients at UCDMC-PH was male than at CHCC (65% versus 52%, respectively, P V .001). Oncology patients represented a small percentage of the study population from each facility (7% at CHCC and 13% at UCDMC-PH, P V .20). 3.2. Univariate analysis Univariate logistic regression analysis was performed on the total study population (n = 977) for prevalence determinants of 2 outcomes: colonization with C. difficile versus no colonization, and colonization with a toxigenic strain of C. difficile versus all other pediatric patients. For the first outcome, colonization with C. difficile, statistical significance ( P V .05) was observed for 6 of the 14 variables in the analysis. Increased colonization prevalence was observed for patients at UCDMC-PH (OR = 2.71; 95% CI = 2.04, 3.59; P b .001), those treated with 2 or more antiinfectives (OR = 1.44; 95% CI = 1.08, 1.93; P = .01), patients treated with 4 or more antiinfectives (OR = 1.49; 95% CI = 1.01, 2.21; P = .05), those patients who were previously hospitalized (OR = 1.32; 95% CI = 1.01, 1.73; P = .04), and those with underlying medical conditions (OR = 1.40; 95% CI = 1.03, 1.87; P = .03). The prevalence of colonization in pediatric inpatients was less than for outpatients (OR = 0.76; 95% CI = 0.58, 0.99; P = .05). Toxigenic C. difficile isolates were more often found in pediatric patients exposed to 2 or more antiinfectives (OR = 1.68; 95% CI = 1.18, 2.40; P = .004) and those with prior hospitalization (OR = 1.79; 95% CI = 1.24, 2.59; P = .002), prior C. difficile infection (OR = 2.32; 95% CI = 1.26, 4.25; P = .004), or underlying conditions (OR = 2.23; 95% CI = 1.53, 3.23; P b .001). None of the factors were protective for this outcome by univariate analysis. 3.3. Multivariate analysis A multiple logistic regression model was constructed for each of the 2 outcomes (Table 3). For the outcome of colonization with C. difficile, the following variables were entered into the model: site, inpatient status, age, use of H-2 antagonists, prior hospitalization, prior C. difficile infection, exposure to 2 or more antiinfectives, and underlying medical condition. Patients treated at UCDMC-PH were more than 3 times as likely to be colonized with C. difficile as patients treated at CHCC (OR = 3.12; 95% CI = 2.29, 4.24; P b .001). Prior hospitalization (OR = 1.42; 95% CI = 1.02, 1.06; P = .04) and exposure to 2 or more antiinfectives (OR = 1.56; 95% CI = 1.10, 2.20; P = .01) also increased the chances of colonization. Inpatient status (OR = 0.64; 95% CI = 0.46, 0.89; P = .007) and exposure to H-2 antagonists (OR = 0.55; 95% CI = 0.36, 0.84; P = .006) decreased the likelihood of colonization. Underlying condition did not achieve statistical significance in this model (OR = 1.44; 95% CI = 0.98, 2.11; P = .06).

112

C.E. Rexach et al. / Diagnostic Microbiology and Infectious Disease 56 (2006) 109–114

Table 1 Demographic characteristics of pediatric patients, aged 1 month to 19 years old, providing stool specimens: CHCC, Madera, CA, July 1998 to January 1999 and UCDMC-PH, Sacramento, CA, October 2001 to January 2003 CHCC

Total no. of patients Inpatients Outpatients Female Male Oncology patients Antibiotics H-2 antagonists Antineoplastics z 2 antiinfectives z 4 antiinfectives Prior hospitalization Prior C. difficile infection Other infection Underlying condition Toxigenic % of isolates Nontoxigenic % of isolates Average age (years) Median age (years)

UCDMC-PH

P

Culture positive

Culture negative

Total

Culture positive

Culture negative

Total

178 100 78 83 92 14 74 20 7 51 18 90 18 69 67 96/676 96/178 82/676 82/178 3.1

498 315 183 213 285 32 179 88 26 112 36 212 44 212 139

676 415 261 296 377 46 253 108 33 163 54 302 62 281 206

148 80 68 57 96 18 68 39 17 55 31 60 2 34 34 45/301 = 15% 45/148 = 30% 103/301 = 34% 103/148 = 70% 5.1

153 87 66 71 77 20 66 44 16 55 35 43 3 39 32

301 167 134 128 173 38 134 83 33 110 66 103 5 73 66

= = = =

14% 54% 12% 46% 3.7

For the outcome of colonization with a toxigenic strain, the following variables were entered: site, inpatient status, age, exposure to antineoplastics, prior hospitalization, prior C. difficile infection, underlying medical conditions, and exposure to 2 or more antiinfectives. Underlying medical conditions doubled the likelihood of colonization with a toxigenic strain (OR = 1.99; 95% CI = 1.21, 3.28; P = .006). Increased risk of colonization with a toxinpositive strain was also observed for pediatric patients exposed to 2 or more antiinfectives (OR = 1.71; 95% CI = 1.10, 2.66; P = .02) and those who were previously hospitalized (OR = 1.60; 95% CI = 1.04, 2.44; P = .03). Inpatients were less likely to be colonized with toxigenic strains of C. difficile than outpatients (OR = 0.51; 95% CI = 0.33, 0.79; P = .003). 3.4. Molecular analysis Arbitrarily primed PCR fingerprinting was used to determine whether isolates were unique or similar. We analyzed 118/148 isolates from UCDMC-PH; only 9 of these had indistinguishable fingerprints. These were placed into 4 groups; the largest comprised 3 toxigenic isolates. The other 3 groups were nontoxigenic. These samples did not come from patients who were hospitalized at the same time, or for whom there was any evidence of interaction, with the exception of one of the nontoxigenic groups composed of 2 isolates obtained from siblings. The remaining 109 isolates had unique fingerprints (92%). At CHCC, 166/178 isolates were fingerprinted (Table 4). Thirteen groups of toxigenic and 6 groups of nontoxigenic isolates with indistinguishable fingerprints were identified, comprising a total of 52 isolates. The remaining 82 isolates

3.5 F 4.7 1.3

V .001 V .001 V .001 V .001 V .001 V .20 V .001 V .001 V .03 V .01 V .20 V .001 V 1.00 V .001 V .01

V .001 5.3

5.3 F 5.6 2.4

had unique fingerprints. The 2 largest of the toxigenic groups contained 15 and 9 isolates, most of which either came from inpatients or from children who had been previously hospitalized. The largest group of nontoxigenic strains contained 30 isolates and also came from children who either were or had been hospitalized. 3.5. Discussion and conclusions Although the percentage of culture-positive patients at UCDMC-PH was higher than at CHCC, the percentage of patients from both institutions colonized with toxigenic strains was similar (14 –15%). Both study populations were comprised of a similar percentage of inpatients (54 – 56%) and outpatients (44 –46%), the median age of both populations was similar (1.3 versus 2.4 years), and the reported colonization rates for children in this age group was consistent with that reported in previous studies (McFarland et al., 2000). This suggests that the carriage rate for nontoxigenic C. difficile may be related to Table 2 Culture- and toxin-positive isolates from CHCC and UCDMC stratified into 2 age groups, b 18 and z 18 months

CHCC Culture Toxin UCDMC Culture Toxin

b 18 months of age

z 18 months of age

Total

P

Positive

Negative

Positive

Negative

108 57

249 51

70 39

235 31

676 178

V .05 V1

61 20

61 41

86 25

93 61

301 147

V1 V1

C.E. Rexach et al. / Diagnostic Microbiology and Infectious Disease 56 (2006) 109–114 Table 3 Logistic regression analysis of study cohort from CHCC and UCDMC-PH (n = 977) OR

95% CI

Outcome: colonization versus no colonization Site 3.12 2.29, Inpatient 0.64 0.46, Age 0.98 0.96, H-2 antagonists 0.55 0.36, Prior hospitalization 1.42 1.02, Prior C. difficile 1.31 0.72, z 2 antiinfectives 1.56 1.10, Underlying conditions 1.44 0.98,

4.24 0.89 1.01 0.84 1.96 2.39 2.20 2.11

P b 0.001 0.007 0.15 0.006 0.04 0.38 0.01 0.06

Outcome: toxin-positive colonization versus all other study subjects Site 1.20 0.79, 1.84 0.39 Inpatient 0.51 0.33, 0.79 0.003 Age 0.98 0.94, 1.01 0.22 Antineoplastics 1.20 0.58, 2.50 0.62 Prior hospitalization 1.60 1.04, 2.44 0.03 Prior C. difficile 1.66 0.85, 3.28 0.14 Underlying conditions 1.99 1.21, 3.28 0.006 z 2 antiinfectives 1.71 1.10, 2.66 0.02

geographic area or environment. Perhaps the larger population of Sacramento compared with Fresno creates increased opportunities for contact between children and for colonization. Because the nontoxigenic organisms have not been associated with disease, they may be present transiently in the gut as part of the normal microbial flora in healthy children. C. difficile toxin assay, the method used by most clinical laboratories for determining if a patient has CDAD, was only ordered for a small subset of the stool specimens at each hospital. In 38% (69/178) of the samples from CHCC and 23% of those from UCDMC-PH, another possible etiologic agent for disease was identified. These included Salmonella, Shigella, Yersinia, Escherichia coli 0157:H7, and Rotavirus, among others. Reports of coinfection with C. difficile and Salmonella or Shigella were recently reported in elderly patients (Grinblat et al., 2004); however, the remaining samples in this study, which were culture positive for C. difficile, had not other identified pathogen. Therefore, the possibility that these children had CDAD should be considered. A child’s age is considered an important factor when determining if C. difficile is the causative agent of disease. Therefore, data were reevaluated after stratification into 2 groups: b 18 months and z 18 months of age. In both study populations, 16% (57/357 and 20/122) of the children under 18 months of age were colonized with toxigenic strains, as compared with 13% (39/305 at CHCC) and 14% (25/179 at UCDMC-PH) of children over the age of 18 months. These rates are similar to those reported for adults and are not significantly different between the 2 facilities. The high rates of colonization in asymptomatic children b 18 months of age previously reported may reflect transient colonization in otherwise healthy children as the gut flora becomes established.

113

Case reports have described CDAD in children with serious underlying conditions (Gaglani et al., 1996; Harris and Figueroa-Colon, 1995; Kato et al., 1995). This study supports these findings and suggests that underlying conditions may be an important risk factor of colonization with toxigenic C. difficile. These children may be more frequently hospitalized and treated with antiinfectives more often than children without these conditions. Exposure to 2 or more antiinfectives increased the likelihood of both outcomes in this study. The disease model for adults suggests that loss of colonization resistance due to antibiotic exposure is a major factor in allowing toxigenic C. difficile to become established. However, in this study with children, the number of antiinfectives to which a child was exposed at the same time appeared to be more significant than any particular type of antimicrobial. This suggests that exposure to antiinfectives may be a surrogate for underlying condition, and immune status may be more significant in disease development in children than loss of colonization resistance. The use of H-2 antagonists has also been described as a risk factor for disease in adults (Cunningham et al., 2003; Dial et al., 2004). In this study, children treated with H-2 antagonists were less likely to be colonized. How pH effects the composition of the microbial flora of the gut and how this differs in children and in adults are not clear. It has been demonstrated that the composition of the microbial flora of the gut changes throughout life. However, it is suggested that certain groups of bacteria are more commonly found in the gut at specific life stages. The susceptibility of these organisms to pH as they pass through the digestive tract and become established in the large intestines may play a role in the composition of the flora and also in the prevention of the establishment of toxigenic C. difficile. Table 4 Summary of the molecular analysis of DNA fingerprints obtained by AP-PCR for isolates from the CHCC cohort Group no.

No. of isolates

Toxin status

Inpatient status

Prior hospitalization

1 2 3 4 5 6 7 8 9 10 11 12 13 1a 2a 3a 4a 5a 6a

15 9 5 4 2 2 2 3 2 2 2 2 2 30 3 3 2 2 2

Toxigenic Toxigenic Toxigenic Toxigenic Toxigenic Toxigenic Toxigenic Toxigenic Toxigenic Toxigenic Toxigenic Toxigenic Toxigenic Nontoxigenic Nontoxigenic Nontoxigenic Nontoxigenic Nontoxigenic Nontoxigenic

9/15 4/9 0/5 2/4 1/2 2/2 2/2 3/3 1/2 2/2 0/2 1/2 2/2 19/30 2/3 1/3 1/2 1/2 2/2

8/15 (53%) 5/9 (56%) 3/5 (60%) 1/4 (25%) 1/2 (50%) 2/2 (100%) 2/2 (100%) 3/3 (100%) 1/2 (50%) Unknown 2/2 (100%) 1/2 (50%) 1/2 (50%) 15/30 (50%) 2/3 (67%) 1/3 (33%) 0/2 (0%) 2/2 (100%) 1/2 (50%)

Number of unique isolates = 82.

(60%) (44%) (0%) (50%) (50%) (100%) (100%) (100%) (50%) (100%) (0%) (50%) (100%) (63%) (67%) (33%) (50%) (50%) (100%)

114

C.E. Rexach et al. / Diagnostic Microbiology and Infectious Disease 56 (2006) 109–114

Statistical analysis of these data also suggests that children are less likely to be colonized and to carry toxigenic strains if they are inpatients. An additional risk factor in adults is the duration of hospital stay (Bartlett, 1994; Lyerly et al., 1988). Many of the inpatients in this study did not have an extended hospital stay, averaging 3 to 5 days, although some were hospitalized for N 30 days. Prior hospitalization was also associated with the colonization with toxigenic strains in this study, and some of the outpatients had a history of prior hospitalization. The role of C. difficile as an agent of nosocomial diarrhea in adults is well established. Fingerprint analysis of C. difficile isolates obtained from these patients showed evidence of nosocomial transmission. In this study, however, the majority of isolates were unique. Because children are not routinely cultured on admission, it becomes difficult to establish whether a child acquired C. difficile in the hospital or was previously colonized. The large number of unrelated isolates obtained from these children and verified by fingerprint analysis suggests that those that were colonized with toxigenic strains did not acquire the organisms from a common source. Studies of C. difficile colonization in the general population may provide some answers to the prevalence of colonization with toxigenic C. difficile in children. This study suggests that disease resulting from C. difficile may be more common in children than previously described. Underlying medical conditions, exposure to larger numbers of antiinfectives, and H-2 antagonists appear to play a role in colonization with toxin-producing C. difficile. Further studies are required to elucidate how these factors may be associated with CDAD in children. Acknowledgments The authors thank the staff of the Children’s Hospital Central California, Madera, CA and UC Davis Medical Center, Sacramento, CA for their help in this study, especially the personnel from the Clinical Microbiology Laboratory. References Bartlett JG (1994) Clostridium difficile: history of its role as an enteric pathogen and the current state of knowledge about the organism. Clin Infect Dis 18(Suppl 4):S265 – S272. Cohen SH, Tang YJ, Silva Jr J (2001) Molecular typing method for epidemiological identification of Clostridium difficile strains. Expert Rev Mol Diagn 1:61 – 70. Cunningham R, Dale B, Undy B, Gaunt N (2003) Proton pump inhibitors as a risk factor for Clostridium difficile diarrheoa. J Hosp Infect 54:243 – 245.

de Boissieu M, Chaussain M, Badoual J, Raymond J, Dupont C (1996) Small-bowel bacterial overgrowth in children with chronic diarrhea, abdominal pain, or both. J Pediatr 128:203 – 207. Dial S, Alrasadi K, Manoukian C, Huang A, Menzies D (2004) Risk of Clostridium difficile diarrhea among hospitalized inpatients prescribed proton pump inhibitors: cohort and case control studies. CMAJ 171: 33 – 38. Frost F, Craun G, Calderon R (1998) Increasing hospitalization and death due to Clostridium difficile diarrheal disease. Emerg Infect Dis 4:619 – 625. Gaglani MJ, Murray JC, Morad AB, Edwards MS (1996) Chronic osteomyelitis caused by Clostridium difficile in an adolescent with sickle cell disease. Pediatr Infect Dis J 15:1054 – 1056. Genta VM, Gilligan PH, McCarthy LR (1984) Clostridium difficile peritonitis in a neonate: a case report. Arch Pathol Lab Med 108: 82 – 83. Gonc¸ alves C, Decre D, Barbut F, Burghoffer B, Petit JC (2004) Prevalence and characterization of a binary toxin (actin specific ADP ribosyltransferase) from Clostridium difficile. J Clin Microbiol 1933 – 1939. Grinblat J, Weiss A, Grossman B, Dicker D, Beloosesky Y (2004) Diarrhea in elderly patients due to Clostridium difficile associated with Salmonella and Shigella infection. Arch Gerontol Geriatr 39:277 – 282. Gumerlock PH, Tang YJ, Weiss JB, Silva Jr J (1993) Specific detection of toxigenic strains of Clostridium difficile in stool specimens. J Clin Microbiol 31:507 – 511. Harris PR, Figueroa-Colon R (1995) Rectal prolapse in children associated with Clostridium difficile infection. Pediatr Infect Dis J 14:78 – 79. Kato S, Ebina K, Ozawa A, Naganuma H, Nakagawa H (1995) Antibiotic associated hemorrhagic colitis without Clostridium difficile toxin in children. J Pediatr 126:1008 – 1010. Langley JM, LeBlanc JC, Hanakowski M, Goloubeva O (2002) The role of Clostridium difficile and viruses as causes of nosocomial diarrhea in children. Infect Control Hosp Epidemiol 23:660 – 664. Lyerly DM, Krivan HC, Wilkins TD (1988) Clostridium difficile: its disease and toxins. Clin Microbiol Rev 1:1 – 18. McFarland LV (1993) Diarrhea acquired in the hospital. Gastroenterol Clin North Am 22:563 – 577. McFarland LV, Brandmarker S, Guandalini S (2000) Pediatric Clostridium difficile: a phantom menace or clinical reality? J Pediatr Gastroenterol Nutr 31:220 – 231. Silva Jr J, Tang YJ, Gumerlock PH (1993) Genotyping of Clostridium difficile isolates. J Infect Dis 169:661 – 664. Simor AE, Yake SL, Tsimidis K (1993) Infection due to Clostridium difficile among elderly residents of a long-term care facility. Clin Infect Dis 17:672 – 678. Tang YJ, Gumerlock PH, Weiss JB, Silva Jr J (1994) Specific detection of Clostridium difficile toxin A gene sequences in clinical isolates. Mol Cell Probes 8:463 – 467. Tang YJ, Houston ST, Gumerlock PH, Mulligan ME, Gerding DN, Johnson S, Fekety FR, Silva Jr J (1995) Comparison of arbitrarily primed PCR with restriction endonuclease and immunoblot analyses for typing Clostridium difficile isolates. J Clin Microbiol 33:3169 – 3173. Wanahita A, Goldsmith E, Musher D (2002) Conditions associated with leukocytosis in a tertiary care hospital, with particular attention to the role of infection caused by Clostridium difficile. Clin Infect Dis 34:1585 – 1592.