Published analyses of national hospital databases encompassing the years of 1997 to 2006 document an increasing trend in the number of cases of children hospitalized with Clostridium difficile infection (CDI).1–3 The rate of increase in the incidence of CDI is approximately 9% per year.2,3 This increasing trend is most concerning when viewed in light of the association of CDI with increased morbidity (lengthier hospital stays, higher medical costs and greater risk of colectomy) and mortality when compared with children without the infection.1 Recently, data regarding hospitalized children for the year 2009 became available. Thus, we sought to determine whether the increasing trend in the incidence of CDI and the association of an increased morbidity and mortality with the infection have continued. To this end, we interrogated a United States pediatric inpatient database for the years 2003 to 2009.
MATERIALS AND METHODS
We obtained data for this study from the Healthcare Cost and Utilization Project Kids’ Inpatient Database (HCUP-KID), sponsored by the Agency for Healthcare Research and Quality. Individual hospitalizations of patients who are 20 years of age or less in acute care hospitals within the United States are maintained in the HCUP-KID as unique entries. Each discharge entry includes 1 primary discharge diagnosis and up to 24 secondary diagnoses (based on the International Classification of Diseases, Ninth Revision, and Clinical Modification, diagnosis codes). Individual-level population weights are used to generate national level estimates. All data in this article are presented as national level estimates. For the triennial periods included in this study (2003, 2006 and 2009), the HCUP-KID contains between 7,409,162 and 7,370,203 weighted total pediatric cases.
The predictor variable in this study was the presence of a diagnosis of CDI. We used the International Classification of Diseases, Ninth Revision, and Clinical Modification diagnostic code for CDI (00845) to identify cases. This is the sole code for CDI, and it has be used previously to identify CDI cases in studies of hospitalized adult4 and pediatric patients.1 We extracted all entries with a primary or secondary discharge diagnosis of CDI. The database does not specifically differentiate between community-associated and nosocomial infections. Children less than 1 year of age were excluded from this study due to the high rate of colonization and lack of correlation between disease and detection of C. difficile in the stool in this age group.5 Children older than 1 year of age were categorized into age groups at 5-year intervals coinciding with recognizable descriptive ages (early childhood, late childhood, early adolescence and late adolescence to early adulthood) to allow for categorical analyses. Insurance status was categorized as private, public and none (no pay, self-pay or other). Hospital characteristics were categorized by teaching status (teaching or nonteaching) and setting (urban or rural). Race is incompletely reported within the KID and was not included in this study. To evaluate disease burden, we used the outcome variables of length of hospital stay (LOS), hospital charges (adjusted for inflation using 2009 Consumer Price Index data), performance of a colectomy, in-hospital mortality and All Patient Diagnosis Related Groups (APR-DRGs). APR-DRGs classify patients by severity of illness (minor, moderate, major or extreme) and risk of mortality (minor, moderate, major or extreme).6,7
Evaluation of Associated Comorbid Conditions
Comorbid diagnoses categories were prepared using International Classification of Diseases, Ninth Revision, diagnostic codes based on a modified version of the methods described by Rassekh et al.8 We used specific comorbid diagnosis categories based on their significant association with pediatric CDI as previously demonstrated by Nylund and colleagues.1 The diagnosis categories were sepsis and bacterial infections, fungal infections, cystic fibrosis, neoplastic disease, hematologic disorders, solid organ transplant, hematopoietic stem cell transplantation, malnutrition, fluid and electrolyte imbalances, inflammatory bowel disease, appendicitis, renal disease, liver disease, HIV infection, cardiac disease, pancreatitis, gastrostomy tube status, lupus erythematosus and gastroesophageal reflux disease.
Statistical analyses were performed using SAS version 9.2 (SAS institute, Cary, NC). The Kruskal–Wallis and analysis of variance test were used for comparing differences between non-normal and normal continuous variables, respectively. The χ2 test was used for comparing differences between categorical variables. For trend analysis, we utilized the Cochrane–Armitage test. The threshold for significance for these analyses was P < 0.05.
Univariate regression analyses were performed to describe the association of CDI with the severity outcome variables under study. To account for unobservable confounders, case-control matching using high-dimensional propensity scores was performed. High-dimensional propensity scores were generated by regression analysis of patients with CDI based on their demographics (age, sex, patient insurance status, hospital setting and hospital teaching status) and the previously listed comorbid conditions. Patients with CDI (cases) were matched by high-dimensional propensity score, using a greedy matching algorithm, to patients who did not have CDI (controls) with a 1:5 matching ratio. Odds ratios (ORs) and 95% confidence intervals (CIs) are reported to identify the strength and significance of CDI with severity outcomes.
For the period of our study (2003 to 2009), there were a total of 8,277,876 discharges for children greater than 1 year of age. Of these discharges, 21,973 had a diagnosis of CDI. There were significant differences between the 2 groups in each of the demographic characteristics studied (P < 0.001 for all comparisons of CDI versus non-CDI). Those with CDI were younger (median [interquartile range]; 10  versus 15  years) and more likely to be male (M:F ratio: 1.04 versus 0.64) than those without the infection. Greater percentages of patients discharged with the diagnosis of CDI were from urban (88.0% versus 83.9%) and teaching hospitals (70.8% versus 55.4%) and had private insurance (50.9% versus 41.7%).
From 2003 to 2009, there was a significant trend for an increasing incidence of CDI in hospitalized children overall and in each age group (Table 1). In all pediatric patients, the incidence increased 57% from 20.0 to 31.5 CDI discharges per 10,000 discharges per year (P < 0.001). Increases in incidence of 52% to 67% were observed for each age group (P < 0.001), with the greatest percent increase in the 11- to 15-year group. The 1- to 5-year age group had the highest incidence of CDI in each year (P < 0.001). In this age group, the incidence increased 63% from 2003 to 2009 (P < 0.001).
During 2003 to 2009, a much higher percentage of children with CDI compared with children without the infection had APR-DRGs scores that indicated a major or extreme severity of illness (66.8% versus 10.6%; P < 0.001) and mortality risk (18.5% versus 2.4%; P < 0.001). Also, children with CDI as compared with those without CDI had a longer LOS (median [interquartile range]; 6  versus 2 ; P < 0.001), higher hospital charges (median [interquartile range]; $28,404 [$76,335] versus $9,579 [$12,367]; P < 0.001), greater likelihood of undergoing a colectomy (0.90% versus 0.14%, P < 0.001; OR 6.52, 95% CI: 5.65–7.52) and a higher mortality (1.70% versus 0.30%, P < 0.001; OR 6.04, 95% CI: 5.45–6.69). Using a cohort of patients with and without CDI for all years under study in a 1:5 ratio matched for demographic and comorbid conditions, CDI was independently associated with a greater LOS (adjusted OR [aOR] 2.41, 95% CI: 2.34–2.49; P < 0.001), higher hospital charges (aOR 1.88, 95% CI: 1.82–1.94; P < 0.001) and an increase in the risk for colectomy (aOR 1.36, 95% CI: 1.14–1.61; P < 0.001). However, CDI was not independently associated with increased mortality (aOR 1.12, 95% CI: 0.99–1.26; P < 0.072). Hospital charges demonstrated an increasing linear trend from 2003 to 2009 (41% increase, P < 0.001). No significant trend was observed for the other outcomes including LOS and rates of colectomy and mortality.
In this study, we analyzed the trend in the incidence of pediatric CDI using HCUP-KID data from 2003 to 2009. Our results indicate a continuation of the trend for an increase in the number of cases of CDI in hospitalized children.1–3 Each of the age groups included in our study demonstrated this increasing trend in the incidence of CDI. Despite the increase in incidence, there appears to have been no increase in disease severity as the only outcome measure in which we observed an increase was hospital charges. LOS, colectomy rate and mortality rate were unchanged from 2003 to 2009.
Children with CDI were sicker than children without the disease as demonstrated by differences in APR-DRGs scores, LOS, hospital charges, colectomy rates and mortality rates. Moreover, after controlling for demographic and comorbid confounders, CDI was independently associated with longer LOS, higher hospital charges and a higher colectomy rate. The risk of morality was not independently associated with pediatric CDI. We believe this is because CDI is not frequently a direct cause of death in children. Rather, it is the comorbidities (adjusted for in the analysis for an independent association) that are responsible for the observed mortality in children with CDI. These comorbidities could be conditions that require hospital admission and consequently predispose them to this nosocomial infection or conditions that are a direct consequence of CDI such as fluid and electrolyte imbalances. Similarly our inclusion of sepsis and bacterial infections as a comorbid confounder might have skewed our outcome results. Even though these conditions could be associated with the outcome of CDI, we believe it is essential to include this comorbidity in the case-control matching procedure because sepsis and bacterial infections are likely associated with an increased rate of antibiotic use and, thus, a greater incidence of CDI. The exclusion of this comorbidity would result in a higher percentage of patients with these conditions in the CDI group as compared with the non-CDI group. This would result in a disparity in the severity of illness between the 2 groups and confound the final analysis.
The underlying cause or causes of the increase in CDI in hospitalized children is unknown. We can speculate that increased testing by physicians due to heighted awareness of the disease or the increased use of sensitive methods of C. difficile detection might play a role. However, the HCUP-KID does not provide information regarding the frequency of testing or the laboratory method used to diagnose CDI. An increase in CDI cases in adults has in large part been attributed to the emergence of a hypervirulent strain of bacteria known as North American pulse-field type 1.9 Current evidence suggests that the hypervirulent North American pulse-field type 1 strain is present in less than 20% of C. difficile positive stool samples from children.10,11 Medications associated with the development of CDI in adults (antibiotics and proton-pump inhibitors) are less consistently found to be risk factors for CDI in children.12,13 Therefore, their increased use might not adequately explain the increase in pediatric CDI. In a previous analysis of the HCUP-KID, Nylund et al1 reported a close association between pediatric CDI and medical comorbidities, most notably inflammatory bowel disease, solid organ- and hematopoietic stem cell transplant, HIV and neoplastic disease. Thus it is possible that the increased prevalence of these conditions in children such as inflammatory bowel disease14 may be a causative factor for the rise in CDI cases in hospitalized children.
The recent rise in the incidence of community-associated CDI might have an effect on our incidence and outcome observations.15 However, we did not directly evaluate this condition. Using the data within the HCUP-KID, we were not able to differentiate between community-associated and nosocomial infections. The HCUP-KID does not track individual patients. Thus, even though a discharge entry may carry a primary diagnosis of CDI, it would be impossible to exclude a recent prior hospital admission for that particular patient. Also, the designation of a diagnosis as primary or secondary does not necessarily denote the reason for the admission. Thus, future studies should address the impact of community-associated CDI on the incidence and outcome of CDI in hospitalized children.
Our study demonstrates a continued increase in the incidence of CDI in hospitalized children in the United States. The increase has occurred in all pediatric age groups. Moreover, CDI is associated with a greater morbidity in hospitalized children. These findings suggest that greater effort must be given to the prevention and treatment of this infection in the pediatric population.
1. Nylund CM, Goudie A, Garza JM, et al. Clostridium difficile infection in hospitalized children
in the United States. Arch Pediatr Adolesc Med. 2011;165:451–457
2. Zilberberg MD, Tillotson GS, McDonald C. Clostridium difficile infections among hospitalized children
, United States, 1997-2006. Emerg Infect Dis. 2010;16:604–609
3. Kim J, Smathers SA, Prasad P, et al. Epidemiological features of Clostridium difficile-associated disease among inpatients at children
’s hospitals in the United States, 2001–2006. Pediatrics. 2008;122:1266–1270
4. Pant C, Anderson MP, O’Connor JA, et al. Association of Clostridium difficile infection with outcomes of hospitalized solid organ transplant recipients: results from the 2009 Nationwide Inpatient Sample database. Transpl Infect Dis. 2012;14:540–547
5. Sandora TJ, Fung M, Flaherty K, et al. Epidemiology and risk factors for Clostridium difficile infection in children
. Pediatr Infect Dis J. 2011;30:580–584
6. Heaton PC, Tundia NL, Schmidt N, et al. National burden of pediatric hospitalizations for inflammatory bowel disease: results from the 2006 Kids’ Inpatient Database. J Pediatr Gastroenterol Nutr. 2012;54:477–485
8. Rassekh SR, Lorenzi M, Lee L, et al. Reclassification of ICD-9 codes into meaningful categories for oncology survivorship research. J Cancer Epidemiol. 2010;2010:569517
9. McDonald LC, Killgore GE, Thompson A, et al. An epidemic, toxin gene-variant strain of Clostridium difficile. N Engl J Med. 2005;353:2433–2441
10. Stoesser N, Crook DW, Fung R, et al. Molecular epidemiology of Clostridium difficile strains in children
compared with that of strains circulating in adults with Clostridium difficile-associated infection. J Clin Microbiol. 2011;49:3994–3996
11. Toltzis P, Kim J, Dul M, et al. Presence of the epidemic North American Pulsed Field type 1 Clostridium difficile strain in hospitalized children
. J Pediatr. 2009;154:607–608
12. Benson L, Song X, Campos J, et al. Changing epidemiology of Clostridium difficile-associated disease in children
. Infect Control Hosp Epidemiol. 2007;28:1233–1235
13. Rexach CE, Tang-Feldman YJ, Cantrell MC, et al. Epidemiologic surveillance of Clostridium difficile diarrhea in a freestanding pediatric hospital and a pediatric hospital at a university medical center. Diagn Microbiol Infect Dis. 2006;56:109–114
14. Abramson O, Durant M, Mow W, et al. Incidence
, prevalence, and time trends of pediatric inflammatory bowel disease in Northern California, 1996 to 2006. J Pediatr. 2010;157:233–239.e1
15. Khanna S, Baddour LM, Dibaise JK, et al. Appendectomy is not associated with adverse outcomes in Clostridium difficile infection: a population-based study. Am J Gastroenterol. 2013;108:626–627