Despite the increasing incidence of Clostridium difficile infection (CDI) in hospitalized children,1 little is known regarding risk factors associated with the recurrence of CDI in the pediatric population. Recurrence of disease is well-recognized in adults and estimated to affect between 7% and 50% of patients.2 Increasing recurrence rates in adults have been reported in the last decade and associated with the emergence of novel, more virulent strains of C. difficile. 3 Risk factors for disease recurrence in adults include increasing age, initial disease severity, use of C. difficile-provocative antibiotics and antacids.4 Among children, prior studies have reported recurrence rates of 20% and 24%.5,6 Analyses regarding predictors in children are lacking and hampered by small sample sizes. Identifying risk factors for recurrent CDI is crucial to guide clinicians regarding effective treatment strategies, follow-up care and parent as well as patient education. Our objectives were to identify the recurrence rate of CDI and predictors of first recurrence in hospitalized children with CDI.
MATERIALS AND METHODS
We conducted a retrospective cohort study at The Johns Hopkins Children’s Hospital from July 2003 to July 2012 to identify all hospitalized patients treated for CDI to assess the recurrence rate and predictors for the first episode of recurrent CDI in children. The primary outcome of interest was recurrent CDI, defined as an episode of CDI occuring ≤ 8 weeks after the onset of a previous episode, provided that CDI symptoms from the earlier episode completely resolved, with or without therapy in accordance with the Society for Healthcare Epidemiology of America (SHEA)/Infectious Diseases Society of America (IDSA) guidelines.7 A child met criteria for CDI if appropriate clinical symptoms were identified (ie diarrhea defined as ≥3 loose stools/d alone or in combination with other gastrointestinal symptoms as abdominal pain or emesis), the exclusion of other identified causes of diarrhea as well as clinical confirmation of the diagnosis by the treating physician and respective initiation of treatment in the presence of confirmatory laboratory testing. Patients with missing follow-up information after 8 weeks of the initial CDI diagnosis were excluded. Furthermore, we excluded children whose presenting CDI episode was recurrent disease. Relevant demographic, clinical course and treatment data were extracted from medical records. We were able to capture all inpatient and outpatient medical visits and laboratory work performed within the Johns Hopkins Health System. Standard therapies for CDI in children hospitalized at Johns Hopkins Hospital are described elsewhere.8
We classified episodes of CDI with respect to healthcare exposure and onset of clinical symptoms. Healthcare Facility-onset Healthcare Facility-associated CDI (HO-HCFA CDI), Community-onset Healthcare Facility-associated CDI (CO-HCFA CDI), Community-associated CDI and “indeterminate” CDI were defined according to standard definitions.7 The IDSA and SHEA define CDI in adults as severe if peripheral leukocyte count is ≥ 15,000/mm3 or if serum creatinine increases to ≥ 50% from baseline.7
Before June 2009, a 2-step algorithm was applied to identify C. difficile,7 consisting of screening of stool specimens for C. difficile glutamate dehydrogenase antigen and consecutive cell culture cytotoxicity assay for screen-positive specimens. In June 2009, polymerase chain reaction (BD GeneOhm Cdiff, Sparks MD assay) was introduced.
χ2 and Fisher’s exact test (where appropriate) were used for comparisons of proportions and the Mann-Whitney U test was applied to analyze non-normally distributed continuous variables. All variables found to be significant in univariable analyses were included in the multivariable regression model. The Hosmer-Lemeshow goodness-of-fit test was applied to assess the final model. Two-sided, P-values ≤ 0.05 were considered significant. Analyses were performed using STATA Statistical Software Version 12.0 (Stata Corp., College Station, TX). The Johns Hopkins University School of Medicine Institutional Review Board approved this study with a waiver for patients’ informed consent.
From July 2003 to July 2012, 202 pediatric inpatients with the clinical diagnosis of CDI were identified and 175 patients met eligibility criteria. Nine cases were excluded as their presenting episode was recurrent disease and 18 were excluded as no follow-up information was available after 8 weeks from discharge. Twenty-one episodes (12%) were first recurrences of children with prior CDI, with mean time to recurrence of 34 days (range 11–58 days).
After introduction of polymerase chain reaction for diagnosis of CDI (2009–2012), the rates ranged from 2.7/1000 admissions to 6.6/1000 admissions, when compared with 1.8/1000 admissions to 3.8/1000 admissions from 2003 to 2007.
Patients with recurrence tended to be younger (median 3.8 vs. 9.9 years) and their initial CDI episode was more commonly classified as community-associated (38.1% vs. 15.6%), rather than HO-HCFA CDI (28.6% vs. 53.3%) when compared with patients without recurrence (see Table, Supplemental Digital Content 1, http://links.lww.com/INF/B719). Neither disease severity nor therapy received differed between the 2 groups. Patients with recurrence tended to have received additional antibiotics prescribed for non-CDI indications during their course of treatment for their initial episode of CDI (71.4% vs. 49.4%, P = 0.06). The most common antibiotics continued during CDI treatment were trimethoprim/sulfamethoxazole (42.9%), penicillins (27.5%), cephalosporins (26.4%) aminoglycosides (22.0%), macrolides (16.5%), carbapenems (11%), quinolones (7.7%) and clindamycin (2.2%). There was no association with recurrence for any single class of antibiotics, antacids or antiretrovirals administered concomitantly (data not shown).
Both community-associated CDI and receipt of additional antibiotics were included in a multivariable logistic regression model and were independently associated with recurrent disease (OR: 5.07, 95% confidence interval: 1.71–15.07, P = 0.003 and OR: 3.85, 95% confidence interval: 1.28–11.58, P = 0.02).
In our cohort, the first symptomatic recurrence of CDI occurred in 12% of pediatric inpatients at a mean of 34 days after the onset of initial disease. Community-associated first CDI episode and receipt of concomitant antibiotics during CDI treatment for the first episode were independent predictors for recurrent disease.
The proportion of recurrence in our cohort is similar to what has been previously identified in adults, ranging from 7 to 50%,2 but lower than previously reported in 92 children with CDI from a population-based cohort (recurrence rate 20%)5 and 82 hospitalized pediatric patients with CDI (recurrence rate 24.4%).6 The latter study, however, relied on a broader definition of recurrence than we applied, by defining the receipt of an additional course of antibiotics at least 2 weeks after completion of the initial course as recurrence, possibly explaining the higher rate identified in their study. Similar to our results, severe CDI as defined by the SHEA/IDSA guidelines7 was not associated with recurrence in both studies.5,6
Our finding of an independent association of community-associated CDI with recurrence could be explained by the circulation of more virulent strains (ie, the BI/NAP1/027 strain) in the community.9 In adults, the association between higher CDI recurrence rates caused by the BI/NAP1/027 strain has been recently confirmed by phase 3 clinical trials of fidaxomicin3 and by a study using multilocus variable-number tandem-repeat analysis to distinguish relapse from reinfection.10 In children, the BI/NAP1/027 strain was, however, not identified as a predictor for recurrent disease in a previous study, albeit underpowered to analyze this correlation.6 In contrast to our finding, a higher rate of recurrence was identified in hospital-acquired CDI, rather than community-acquired CDI in a previous study (26% vs. 17%). This population however consisted of only 23 cases of hospital-acquired CDI, relied on ICD-9 codes for case definition5 and derived from a population-based, rather than a hospitalized cohort. Our finding may, therefore, result from a referral bias, as children with community-associated disease not resulting in hospitalization may have lower recurrence rates. Nevertheless, our finding is important for clinicians to consider while caring for pediatric inpatients.
Concomitant antibiotics were associated with a lower cure rate and an extended time to resolution of diarrhea in an adult population.11 The independent association between recurrence and coadministration of other antibiotics for treatment of additional infections detected in our study is plausible as sustained alteration of the bacterial gut flora favors the proliferation of C. difficile.2 This result has significant clinical implications for the treating physician by underscoring the importance of cessation of antibiotics intended for treatment of concurrent infections whenever possible in the setting of CDI.
Neither duration nor type of treatment and route of administration for patients treated with metronidazole differed significantly between children with and without recurrence in our cohort. Treatment with vancomycin was without failure in 7 pediatric patients when compared with treatment with metronidazole (n = 75), which was associated with recurrence in 21.3%.5 This study, as ours, was however limited to detect associations between treatment and recurrence by the small sample size.
Important limitations of this study include its retrospective, single-center design. Point estimates must be interpreted with caution due to the small sample size, reflected by wide confidence intervals. In addition, the identification of causative strains was missing, to further elucidate the association of community-associated CDI with recurrence in our population. Disease severity was classified based on the recommendations issued by the IDSA and SHEA7 and its predictive value has not been assessed in children. Furthermore, we were not able to assess the association of administered dose of metronidazole and vancomycin with recurrence, which may have be important especially in younger children. Finally, we may have underestimated cases of recurrence if recurrent CDI cases were managed in other institutions.
In conclusion, recurrence rate was 12% during the 9-year study period in pediatric inpatients. Community-associated disease and concomitant use of systemic antibiotics were independent predictors for recurrence. Antibiotics administered for reasons other than treatment of CDI should be discontinued whenever possible.
1. Schutze GE, Willoughby RE. Committee on Infectious Diseases; American Academy of Pediatrics
. Clostridium difficile
infection in infants and children. Pediatrics
2. Aslam S, Hamill RJ, Musher DM. Treatment of Clostridium difficile
-associated disease: old therapies and new strategies. Lancet Infect Dis. 2005;5:549–557
3. Petrella LA, Sambol SP, Cheknis A, et al. Decreased cure and increased recurrence rates for Clostridium difficile
infection caused by the epidemic C. difficile BI strain. Clin Infect Dis. 2012;55:351–357
4. Garey KW, Sethi S, Yadav Y, et al. Meta-analysis to assess risk factors for recurrent Clostridium difficile
infection. J Hosp Infect. 2008;70:298–304
5. Khanna S, Baddour LM, Huskins WC, et al. The epidemiology of Clostridium difficile
infection in children: a population-based study. Clin Infect Dis. 2013
6. Kim J, Shaklee JF, Smathers S, et al. Risk factors and outcomes associated with severe clostridium difficile infection in children. Pediatr Infect Dis J. 2012;31:134–138
7. Cohen SH, Gerding DN, Johnson S, et al.Society for Healthcare Epidemiology of America; Infectious Diseases Society of America. Clinical practice guidelines for Clostridium difficile
infection in adults: 2010 update by the society for healthcare epidemiology of America (SHEA) and the infectious diseases society of America (IDSA). Infect Control Hosp Epidemiol. 2010;31:431–455
8. Tamma PD, Sandora TJ. Clostridium difficile
infection in children: current state and unanswered questions. J Pediatric Infect Dis Soc. 2012;1:230–243
9. . Centers for Disease Control and Prevention (CDC). Severe Clostridium difficile
-associated disease in populations previously at low risk--four states, 2005. MMWR Morb Mortal Wkly Rep. 2005;54:1201–1205
10. Marsh JW, Arora R, Schlackman JL, et al. Association of relapse of Clostridium difficile
disease with BI/NAP1/027. J Clin Microbiol. 2012;50:4078–4082
11. Mullane KM, Miller MA, Weiss K, et al. Efficacy of fidaxomicin versus vancomycin as therapy for Clostridium difficile
infection in individuals taking concomitant antibiotics for other concurrent infections
. Clin Infect Dis. 2011;53:440–447