Advertisement

Shift to community-onset Clostridium difficile infection in the national Veterans Health Administration, 2003-2014

Open AccessPublished:November 07, 2017DOI:https://doi.org/10.1016/j.ajic.2017.09.020

      Highlights

      • C. difficile infection (CDI) types were studied in U.S. veterans from 2002 to 2014.
      • Healthcare facility-onset CDI was the predominant CDI type.
      • The proportion of patients with community-associated CDI increased in recent years.
      • Healthcare facility-onset CDI patients had higher rates of severe CDI and mortality.

      Background

      Clostridium difficile infection (CDI) occurs frequently in inpatient settings; however, community-onset cases have been reported more frequently in recent years. This study evaluated hospital-onset and community-onset CDI in the national Veterans Health Administration (VHA) population over a 12-year period.

      Methods

      This was a retrospective cohort study of all adult VHA beneficiaries with CDI between October 1, 2002, and September 30, 2014. Data were obtained from the Veterans Affairs Informatics and Computing Infrastructure. CDI was categorized into community-associated CDI (CA-CDI); community-onset, health care facility-associated CDI; and health care facility-onset CDI (HCFO-CDI). Each type was described longitudinally and was assessed as an independent risk factor for health outcomes using multivariable logistic regression.

      Results

      Overall, 30,326 patients with a first CDI episode were included. HCFO-CDI was the predominant type (60.2%), followed by CO-HCFA-CDI (20.6%) and CA-CDI (19.2%). The proportion of patients with HCFO-CDI decreased from 73.5% during fiscal year 2003 to 53.2% during fiscal year 2014, whereas CA-CDI increased from 8.3% to 26.7%. HCFO-CDI was a positive predictor of severe CDI (odds ratio [OR], 1.71; 95% confidence interval [CI], 1.59-1.84) and 30-day mortality (OR, 1.46; 95% CI, 1.32-1.61), but a negative predictor of 60-day recurrence (OR, 0.41; 95% CI, 0.37-0.46).

      Conclusions

      HCFO-CDI was the predominant CDI type. The proportion of patients with CA-CDI increased and HCFO-CDI decreased in recent years. Patients with HCFO-CDI experienced higher rates of severe CDI and mortality.

      Key Words

      Clostridium difficile infection (CDI) occurs frequently in inpatient settings due to exposure to antibiotic agents, comorbid illness, and environment factors that facilitate organism transmission; however, community-onset cases have been reported more frequently in the literature during recent years. The Centers for Disease Control and Prevention (CDC) estimate that 52% of CDI cases originate in the outpatient setting, predominately in patients with recent contact with a health care system.
      • Centers for Disease Control and Prevention
      Vital signs: preventing Clostridium difficile infections.
      Similarly, a national prevalence study by Lessa et al
      • Lessa F.C.
      • Winston L.G.
      • McDonald L.C.
      Emerging infections program CdST. Burden of Clostridium difficile infection in the United States.
      found that 66% of CDI cases were health care-associated, but only 24% of those patients had onset during hospitalization. Of particular concern in community-onset cases, approximately 10%-20% of patients have no prior exposure to a health care system.
      • Centers for Disease Control and Prevention
      Vital signs: preventing Clostridium difficile infections.
      • Kutty P.K.
      • Woods C.W.
      • Sena A.C.
      • Benoit S.R.
      • Naggie S.
      • Frederick J.
      • et al.
      Risk factors for and estimated incidence of community-associated Clostridium difficile infection, North Carolina, USA.
      • Chitnis A.S.
      • Holzbauer S.M.
      • Belflower R.M.
      • Winston L.G.
      • Bamberg W.M.
      • Lyons C.
      • et al.
      Epidemiology of community-associated Clostridium difficile infection, 2009 through 2011.
      The majority of CDI epidemiologic data are acquired from hospital administrative data; therefore, few studies have described characteristics of patients with community-onset CDI. Even fewer studies have evaluated health outcomes in these distinct patient populations. Therefore, the primary objective of this study was to describe the incidence of hospital-onset and community-onset CDI in the national Veterans Health Administration (VHA) patient population over a 12-year period. This study also compared CDI health outcomes among these subgroups.

      Methods

      This was a retrospective cohort study of patients receiving care at any inpatient or outpatient VHA facility in the United States. Data for this study were obtained from the Veterans Affairs Informatics and Computing Infrastructure. Specifically, Veterans Affairs Informatics and Computing Infrastructure includes administrative, clinical, laboratory, and pharmacy data repositories that are linked using unique patient identifiers. This study was approved by the South Texas Veterans Health Care System Research and Development Committee and the Institutional Review Board at University of Texas Health San Antonio.
      Patients were included in the study if they were aged 18-89 years and had any inpatient or outpatient ICD-9-CM code for CDI (008.45) plus any positive CDI stool test (eg, glutamate dehydrogenase, enzyme immunoassay, and polymerase chain reaction) during or within 7 days of the visit from October 1, 2002, through September 30, 2014. To limit our cohort to first-episode CDI patients only, we excluded patients with an ICD-9-CM code for CDI (008.45) during the year before study inclusion.
      Patient demographic characteristics included age, sex, race, and ethnicity. CDI was categorized as either a principal (ICD-9-CM code in the first position) or secondary diagnosis (ICD-9-CM code in any position except first). CDI type was categorized based on the typical surveillance definitions constructed by the CDC.
      • McDonald L.C.
      • Coignard B.
      • Dubberke E.
      • Song X.
      • Horan T.
      • Kutty P.K.
      Recommendations for surveillance of Clostridium difficile-associated disease.
      Administrative codes do not contain the date of CDI diagnosis; therefore, we used CDI therapy receipt as a surrogate for the first date of CDI diagnosis. Community-onset, health care facility-associated CDI (CO-HCFA-CDI) was defined based on the presence of a prescription for a CDI therapy (eg, metronidazole, oral vancomycin, fidaxomicin, nitazoxanide, rifaximin, or probiotics) as an outpatient or on day 1 or 2 of hospitalization, plus 1 or more prior hospitalizations occurring within 90 days preceding CDI diagnosis. Community-associated CDI (CA-CDI) was defined similarly, except that patients did not have a prior hospitalization. Lastly, health care facility-onset CDI (HCFO-CDI) was defined as CDI therapy on or after day 3 of hospitalization.
      Charlson comorbidities and other relevant diagnoses, as defined by ICD-9-CM codes, were identified for the year before the first CDI episode. We also calculated the Charlson comorbidity score.
      • Deyo R.A.
      • Cherkin D.C.
      • Ciol M.A.
      Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases.
      In addition, we collected other infections that occurred during a CDI episode (between CDI outpatient visit date or hospital admission date and CDI therapy end date), including bacteremia, pneumonia, skin infection, intra-abdominal infection, urinary tract infection, device-related infection, endocarditis, and acute respiratory infection. Severe CDI was defined as the presence of any of the following severity markers at any time during the episode: intensive care unit admission, sepsis/septicemia, shock, acute renal failure, megacolon, prolonged ileus, perforated intestine, colectomy, white blood cell count, C-reactive protein, serum creatinine, and albumin.
      Medication use during the 90 days before the episode (prior use) and during or up to 60 days after the episode (concomitant use) were collected for non-CDI antibiotics (excludes oral vancomycin, metronidazole, fidaxomicin, rifaximin, and nitazoxanide), non-CDI high-risk antibiotics (third- and fourth-generation cephalosporins, fluoroquinolones, and clindamycin), gastric acid-suppressing drugs (antacids, histamine 2 blockers, and proton pump inhibitors), antidiarrhea medications, narcotics, and bowel prep medications.
      Severe CDI was defined as any of the following: intensive care unit admission, sepsis, shock, acute renal failure, megacolon, prolonged ileus, perforated intestine, white blood cell count ≥15,000 cells/µL, C-reactive protein ≥160 mg/L, albumin <2.5 g/dL, serum creatinine >1.5 mg/dL, or colectomy. Mortality was defined as death from any cause within the 30, 60, or 90 days following CDI treatment discontinuation. CDI recurrence was defined as a CDI inpatient or outpatient episode identified with a second CDI ICD-9-CM code within the 30, 60, or 90 days following the initial episode and with a minimum 3-day gap between initial therapy discontinuation and new therapy initiation. Hospital length of stay was defined as date of discharge minus date of admission plus 1 day.
      Data extraction and variable creation were conducted using SAS version 9.2 (SAS Institute Inc, Cary, NC). All other data and statistical analyses were conducted using JMP version 13.0 (SAS Institute Inc).
      All independent and dependent variables were first presented descriptively. For baseline characteristics (eg, sex, race, and ethnicity), we included a missing category. Other variables that were absent from the medical chart (eg, comorbidities) were assumed to have not occurred.
      • Shapiro N.I.
      • Wolfe R.E.
      • Moore R.B.
      • Smith E.
      • Burdick E.
      • Bates D.W.
      Mortality in Emergency Department Sepsis (MEDS) score: a prospectively derived and validated clinical prediction rule.
      • Huang C.C.
      • Kuo S.C.
      • Chien T.W.
      • Lin H.J.
      • Guo H.R.
      • Chen W.L.
      • et al.
      Predicting the hyperglycemic crisis death (PHD) score: a new decision rule for emergency and critical care.
      Baseline characteristics were compared among patients with each surveillance definition using the χ2 test for categorical variables and the Kruskal-Wallis test for continuous variables. Because of the large sample size, a P value < .0001 was considered statistically significant.
      We described the epidemiology of CDI first episodes from fiscal year (FY) 2003 to FY 2014. We first presented CDI type descriptively as the proportion of patients with each surveillance definition. We then described trends longitudinally, with FY as the independent variable, and the surveillance definition as the dependent variable. Finally, we assessed CDI onset type as an independent risk factor for specific outcomes using a series of logistic regression models. CDI outcomes (severe CDI; 30-, 60-, and 90-day mortality; and hospital length of stay ≥14 days) served as the dependent variables and CDI onset-type as the independent variable. For recurrence analyses, we excluded those patients who died before the recurrence. CA-CDI was assigned as the reference category. Control variables included all independent variables listed in Table 1 that were statistically significant between groups at P value < .0001 and occurred in at least 5% of the population. For severe CDI, the model excluded individual severity indicators as covariates. The results were presented as odds ratios (ORs) and 95% confidence intervals (CIs).
      Table 1Baseline characteristics
      CharacteristicCA-CDI

      (n = 5,830)
      CO-HCFA-CDI (n = 6,236)HCFO-CDI (n = 18,260)P value
      Age, y66 (78-59)68 (61-79)68 (60-78)< .0001
      Male sex92.896.496.7< .0001
      Race & ethnicity< .0001
       Non-Hispanic white68.368.165.0
       Non-Hispanic black19.420.721.7
       Hispanic5.45.35.4
       Other3.73.64.8
       Missing3.22.33.1
      Principal CDI diagnosis43.148.216.5< .0001
      Comorbidities
       Hypertension71.482.977.8< .0001
       Dyslipidemia53.560.353.0< .0001
       Obesity16.417.916.0.0022
       Myocardial infarction6.313.712.1< .0001
       Congestive heart failure17.931.327.7< .0001
       Peripheral vascular disease13.422.420.3< .0001
       Cerebrovascular disease14.721.320.4< .0001
       Dementia2.43.94.1< .0001
       COPD31.643.538.0< .0001
       Rheumatologic disease2.43.12.8.0337
       Peptic ulcer disease2.95.15.1< .0001
       Liver disease6.17.97.3.0007
       Diabetes35.843.741.9< .0001
       Hemiplegia or paraplegia2.74.14.6< .0001
       Renal disease21.733.728.6< .0001
       Cancer21.730.830.6< .0001
       HIV/AIDS1.91.91.9.9300
       GERD27.231.325.5< .0001
       Transplant2.02.51.8.0083
       Inflammatory bowel disease4.32.81.7< .0001
       Irritable bowel syndrome1.71.20.9< .0001
      Charlson score2 (1-5)4 (2-6)4 (2-6)< .0001
      Concomitant infections
       Bacteremia4.06.08.3< .0001
       Pneumonia13.617.727.9< .0001
       Skin infection6.89.312.8< .0001
       Intra-abdominal infection7.26.05.6< .0001
       Device-related infection1.82.34.0< .0001
       Acute respiratory infection3.43.23.5.4068
       Endocarditis0.53.21.3< .0001
       Urinary tract infection1.41.32.1< .0001
      CDI severity indicators
       Intensive care unit admission2.22.83.8< .0001
       Sepsis/septicemia11.917.619.6< .0001
       Shock3.45.25.7< .0001
       Acute renal failure24.130.233.1< .0001
       Megacolon0.30.40.3.6089
       Prolonged ileus2.72.94.9< .0001
       Perforated intestine0.20.50.6.0004
       WBC ≥15,000 cells/µL30.137.142.8< .0001
       CRP ≥160 mg/L1.21.12.0< .0001
       Albumin <2.5 g/dL21.927.338.2< .0001
       SCr >1.5 mg/dL20.027.024.8< .0001
       Colectomy0.10.20.1.0129
      Medications
       Prior antibiotics38.386.452.5< .0001
       Prior high-risk antibiotics20.667.134.4< .0001
       Prior GAS drugs39.278.455.4< .0001
       Prior narcotics25.353.138.4< .0001
       Prior antidiarrheals8.312.65.8< .0001
       Prior bowel prep7.423.115.8< .0001
       Concomitant antibiotics42.468.384.8< .0001
       Concomitant high-risk antibiotics35.045.960.3< .0001
       Concomitant GAS drugs65.077.684.4< .0001
       Concomitant narcotics36.745.957.6< .0001
       Concomitant antidiarrheals9.912.711.3< .0001
       Concomitant bowel prep11.412.524.4< .0001
      NOTE. Values are presented as median (interquartile range) or %.
      CA-CDI, community-associated CDI; CDI, Clostridium difficile infection; COPD, chronic obstructive pulmonary disease; CO-HCFA-CDI, community-onset, health care facility-associated CDI; CRP, C-reactive protein; GAS, gastric acid-suppressing; GERD, gastroesophageal reflux disease; HCFO-CDI, health care facility-onset CDI; Scr, serum creatinine; WBC, white blood cells.

      Results

      Baseline characteristics

      Overall 30,326 patients met study inclusion criteria. Table 1 describes the patients' baseline characteristics by CDI type. Patients with HCFO-CDI, CO-HCFA-CDI, and CA-CDI differed with respect to demographic characteristics, comorbidities, and medications received. Overall, patients with HCFO-CDI tended to have more comorbid conditions, more severe CDI indicators, and a higher proportion of prior or concomitant high-risk medication use.

      CDI type and outcomes

      A total of 18,260 (60.2%) patients were classified as HCFO-CDI. Among community-onset patients, 20.6% were classified as CO-HCFA-CDI and 19.2% were classified as CA-CDI.
      There was a shift from HCFO-CDI to CA-CDI over the study period (Figure 1). The proportion of patients with HCFO-CDI decreased from 73.5% during FY 2003 to 53.2% during FY 2014, whereas CA-CDI increased from 8.3% during FY 2003 to 26.7% during FY 2014. The proportion of patients with CO-HCFA-CDI remained relatively stable over the study period.
      Fig 1
      Fig 1Proportion of patients with each Clostridium difficile infection (CDI) type from fiscal year 2003 to fiscal year 2014 (N = 30,326). CA-CDI, community-associated CDI; CO-HCFA-CDI, community-onset, health care facility-associated CDI; HCFO-CDI, health care facility-onset CDI.
      Overall, patients with HCFO-CDI had the highest rates of severe CDI (70.2%); 30-day (24.5%), 60-day (29.4%), and 90-day (32.6%) mortality; and hospital length of stay ≥14 days (62.5%) compared with patients with CA-CDI or CO-HCFA-CDI (Table 2). Alternatively, patients with HCFO-CDI had the lowest rates of 30-day (7.2%), 60-day (9.5%), and 90-day (10.6%) recurrence. Patients with CO-HCFA-CDI had higher rates of severe CDI, mortality, and length of stay ≥14 days, but similar recurrence rates compared with CA-CDI. In multivariable models, HCFO-CDI was a positive independent predictor of severe CDI (OR, 1.71; 95% CI, 1.59-1.84), 30-day mortality (OR, 1.46; 95% CI, 1.32-1.61), 60-day mortality (OR, 1.48; 95% CI, 1.35-1.62), 90-day mortality (OR, 1.54, 95% CI, 1.41-1.68), and hospital length of stay ≥14 days (OR, 3.56; 95% CI 3.26-3.89). HCFO-CDI was a negative predictor of 30-day recurrence (OR, 0.41; 95% CI, 0.37-0.46), 60-day recurrence (OR, 0.40; 95% CI, 0.37-0.45), and 90-day recurrence (OR, 0.41; 95% CI, 0.37-0.46).
      Table 2Clostridium difficile infection (CDI health outcomes by CDI type, N = 30,326)
      OutcomeCA-CDI (n = 5,830)CO-HCFA-CDI (n = 6,236)HCFO-CDI (n = 18,260)
      Severe CDI
      Multivariable logistic regression covariates included age, sex, race, principal CDI diagnosis, hypertension, dyslipidemia, myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, chronic obstructive pulmonary disease, diabetes, renal disease, cancer, GERD, bacteremia, pneumonia, skin infection, prior and concomitant gastric acid-suppressing drugs, prior and concomitant antibiotics, prior and concomitant narcotics, prior and concomitant antidiarrheals, and prior and concomitant bowel prep.
      50.0

      1.00 (reference)
      61.1

      1.34 (1.23-1.46)
      70.2

      1.71 (1.59-1.84)
      30-d mortality
      Multivariable logistic regression covariates included those listed above, plus sepsis, acute renal failure, white blood cell count ≥15,000 cells/µL and albumin <2.5 g/dL.
      12.4

      1.00 (reference)
      18.5

      1.22 (1.09-1.37)
      24.5

      1.46 (1.32-1.61)
      60-d mortality
      Multivariable logistic regression covariates included those listed above, plus sepsis, acute renal failure, white blood cell count ≥15,000 cells/µL and albumin <2.5 g/dL.
      14.8

      1.00 (reference)
      23.7

      1.36 (1.22-1.51)
      29.4

      1.48 (1.35-1.62)
      90-d mortality
      Multivariable logistic regression covariates included those listed above, plus sepsis, acute renal failure, white blood cell count ≥15,000 cells/µL and albumin <2.5 g/dL.
      16.2

      1.00 (reference)
      26.7

      1.40 (1.26-1.56)
      32.6

      1.54 (1.41-1.68)
      30-d recurrence
      Multivariable logistic regression covariates included those listed above, plus sepsis, acute renal failure, white blood cell count ≥15,000 cells/µL and albumin <2.5 g/dL.
      19.0

      1.00 (reference)
      20.6

      1.02 (0.91-1.14)
      7.2

      0.41 (0.37-0.46)
      60-d recurrence
      Multivariable logistic regression covariates included those listed above, plus sepsis, acute renal failure, white blood cell count ≥15,000 cells/µL and albumin <2.5 g/dL.
      23.3

      1.00 (reference)
      26.3

      1.03 (0.93-1.14)
      9.5

      0.40 (0.37-0.45)
      90-d recurrence
      Multivariable logistic regression covariates included those listed above, plus sepsis, acute renal failure, white blood cell count ≥15,000 cells/µL and albumin <2.5 g/dL.
      24.8

      1.00 (reference)
      28.1

      1.02 (0.91-1.15)
      10.6

      0.41 (0.37-0.46)
      Hospital length of stay ≥14 d
      Multivariable logistic regression covariates included those listed above, plus sepsis, acute renal failure, white blood cell count ≥15,000 cells/µL and albumin <2.5 g/dL.
      21.3

      1.00 (reference)
      28.2

      1.43 (1.28-1.59)
      62.5

      3.56 (3.26-3.89)
      NOTE. Boldface type indicates statistical significance compared with CA-CDI. Values are presented as % and odds ratio (95% confidence interval).
      * Multivariable logistic regression covariates included age, sex, race, principal CDI diagnosis, hypertension, dyslipidemia, myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, chronic obstructive pulmonary disease, diabetes, renal disease, cancer, GERD, bacteremia, pneumonia, skin infection, prior and concomitant gastric acid-suppressing drugs, prior and concomitant antibiotics, prior and concomitant narcotics, prior and concomitant antidiarrheals, and prior and concomitant bowel prep.
      Multivariable logistic regression covariates included those listed above, plus sepsis, acute renal failure, white blood cell count ≥15,000 cells/µL and albumin <2.5 g/dL.

      Discussion

      This study documents the burden of CDI by onset type among all adult VHA enrollees over a 12-year period. We found that HCFO-CDI remains the predominant CDI type, but community-onset cases are on the rise. Furthermore, we found that HCFO-CDI patients often have higher rates of severe CDI and mortality and longer hospital stays, but fewer CDI recurrences compared with CA-CDI and CO-HCFA-CDI. Our study is the first to describe health outcomes among patients according to all 3 CDI surveillance definitions. Our study is strengthened by the large national sample and comprehensive data collection.
      Similar to prior studies, we found that the majority of patients developed CDI in the hospital.
      • Centers for Disease Control and Prevention
      Vital signs: preventing Clostridium difficile infections.
      • Kutty P.K.
      • Woods C.W.
      • Sena A.C.
      • Benoit S.R.
      • Naggie S.
      • Frederick J.
      • et al.
      Risk factors for and estimated incidence of community-associated Clostridium difficile infection, North Carolina, USA.
      • Garg S.
      • Mirza Y.R.
      • Girotra M.
      • Kumar V.
      • Yoselevitz S.
      • Segon A.
      • et al.
      Epidemiology of Clostridium difficile-associated disease (CDAD): a shift from hospital-acquired infection to long-term care facility-based infection.
      • Dubberke E.R.
      • Butler A.M.
      • Yokoe D.S.
      • Mayer J.
      • Hota B.
      • Mangino J.E.
      • et al.
      Multicenter study of surveillance for hospital-onset Clostridium difficile infection by the use of ICD-9-CM diagnosis codes.
      • Price M.F.
      • Dao-Tran T.
      • Garey K.W.
      • Graham G.
      • Gentry L.O.
      • Dhungana L.
      • et al.
      Epidemiology and incidence of Clostridium difficile-associated diarrhoea diagnosed upon admission to a university hospital.
      In addition, those patients who developed CDI in the community often had recent health care exposure. We found a shift from HCFO-CDI to CA-CDI over the study period. This could be due to several reasons. First, there has been an increase in the use of high-risk antibiotics, such as fluoroquinolones, in the community.
      • Linder J.A.
      • Huang E.S.
      • Steinman M.A.
      • Gonzales R.
      • Stafford R.S.
      Fluoroquinolone prescribing in the United States: 1995 to 2002.
      This more frequent use could result in a greater number of community-dwelling patients developing CDI. Furthermore, there has also been a significant increase in other high-risk medications, such as proton pump inhibitors, in the community, which could also contribute to higher CDI rates.
      • Rotman S.R.
      • Bishop T.F.
      Proton pump inhibitor use in the U.S. ambulatory setting, 2002-2009.
      Finally, there has been an increase in the number of elderly patients admitted to long-term care facilities. These facilities serve as reservoirs for C difficile and have been previously identified as a risk factor for the development of CDI.
      • McFarland L.V.
      • Mulligan M.E.
      • Kwok R.Y.
      • Stamm W.E.
      Nosocomial acquisition of Clostridium difficile infection.
      In line with our findings, a prior study by Jen et al
      • Jen M.H.
      • Saxena S.
      • Bottle A.
      • Pollok R.
      • Holmes A.
      • Aylin P.
      Assessment of administrative data for evaluating the shifting acquisition of Clostridium difficile infection in England.
      found an increase in the proportion of probable community-acquired CDI of 6% over a 12-year study period (from 7% during 1998 to 13% during 2010) using data from English Hospital Episode Statistics.
      Importantly, patients with HCFO-CDI had significantly worse outcomes compared with community-onset CDI. Garg et al
      • Garg S.
      • Mirza Y.R.
      • Girotra M.
      • Kumar V.
      • Yoselevitz S.
      • Segon A.
      • et al.
      Epidemiology of Clostridium difficile-associated disease (CDAD): a shift from hospital-acquired infection to long-term care facility-based infection.
      found that patients with HCFO-CDI have significantly higher intensive care unit admission (26% vs 3.5%), median hospital length of stay (10 vs 4 days), and median number of comorbidities (2 vs 1). Patients presenting from the community are also traditionally younger, placing them at lower risk for poor health outcomes. Khanna et al
      • Khanna S.
      • Keddis M.T.
      • Noheria A.
      • Baddour L.M.
      • Pardi D.S.
      Acute kidney injury is an independent marker of severity in Clostridium difficile infection: a nationwide survey.
      found that patients with CA-CDI were younger (median, 50 years vs 72 years), had lower comorbidity scores, were less likely to have severe infection (20% vs 31%), and were less likely to have been exposed to antibiotics (78% vs 94%). In addition, patients with CA-CDI often do not present with traditional risk factors for CDI. In a study of 984 patients with CA-CDI, 36% had not received antibiotics during the prior 12 weeks, 18% had no recent health care exposure, and 40% had low-level outpatient exposure.
      • Chitnis A.S.
      • Holzbauer S.M.
      • Belflower R.M.
      • Winston L.G.
      • Bamberg W.M.
      • Lyons C.
      • et al.
      Epidemiology of community-associated Clostridium difficile infection, 2009 through 2011.
      Further research is needed to more accurately predict CDI risk among community-dwelling patients.
      We found that patients with HCFO-CDI have lower rates of recurrence compared with CA-CDI. This reason for this association is unclear. We hypothesize that these patients may receive more aggressive therapy in the inpatient setting, thus decreasing their likelihood for recurrence. Additionally, known risk factors for recurrence may be more easily minimized in controlled hospital environments. However, this association may also represent the use of outdated CDI surveillance definitions. These definitions date back to a CDC working group in 2007, and no longer fully describe the settings in which CDI exposure occurs in the evolving landscape of our health care system.
      • Chitnis A.S.
      • Holzbauer S.M.
      • Belflower R.M.
      • Winston L.G.
      • Bamberg W.M.
      • Lyons C.
      • et al.
      Epidemiology of community-associated Clostridium difficile infection, 2009 through 2011.
      • McDonald L.C.
      • Coignard B.
      • Dubberke E.
      • Song X.
      • Horan T.
      • Kutty P.K.
      Recommendations for surveillance of Clostridium difficile-associated disease.
      • Bloomfield L.E.
      • Riley T.V.
      Epidemiology and risk factors for community-associated Clostridium difficile infection: a narrative review.
      For example, to be considered CO-HCFA-CDI a patient must have a history of hospitalization with a minimum inpatient stay of 48 hours. This narrow definition fails to include a growing population of patients with outpatient or inpatient day procedures occurring at dialysis centers, transplant centers, or chemotherapy clinics that may serve as community reservoirs of spores. These patients are at high risk of recurrence due to repeated exposures or immunosuppression but are currently defined as CA-CDI.
      The need for hospitalization has tremendous impact on health care costs and patient outcomes. Hospitalization inadvertently exposes patients to other risks, avoidable complications, and other nosocomial infections. Therefore, patients with CA-CDI who are older or who have higher comorbidity burden, should be monitored closely and managed more aggressively in the community to prevent poor outcomes.
      There are potential limitations to this study. First, our use of a retrospective cohort study design might be subject to misclassification bias and confounding, particularly due to the use of electronic medical data that is not collected for primary research purposes. Our study attempted to minimize misclassification of CDI by defining CDI cases as ICD-9-CM codes and positive microbiologic data. We limited confounding using multivariable modeling with covariates likely to affect study outcomes; however, we cannot determine whether outcomes were specifically attributable to CDI rather than other unmeasured variables. Next, the large study sample size increases the likelihood of detecting statistically significant differences between groups; therefore, absolute group differences should be considered along with P values. Finally, the veteran population, which is predominately male, might limit the generalizability of our findings to other settings.

      Conclusions

      HCFO-CDI was the predominant CDI onset type among a national cohort of adult veterans. The proportion of patients with CA-CDI and CO-HCFA-CDI increased in recent years, whereas HCFO-CDI decreased. Patients with HCFO-CDI experienced higher rates of severe CDI and mortality and longer hospital stays, but fewer CDI recurrences.

      References

        • Centers for Disease Control and Prevention
        Vital signs: preventing Clostridium difficile infections.
        MMWR Morb Mortal Wkly Rep. 2012; 61: 157-162
        • Lessa F.C.
        • Winston L.G.
        • McDonald L.C.
        Emerging infections program CdST. Burden of Clostridium difficile infection in the United States.
        N Engl J Med. 2015; 372: 2369-2370
        • Kutty P.K.
        • Woods C.W.
        • Sena A.C.
        • Benoit S.R.
        • Naggie S.
        • Frederick J.
        • et al.
        Risk factors for and estimated incidence of community-associated Clostridium difficile infection, North Carolina, USA.
        Emerg Infect Dis. 2010; 16: 197-204
        • Chitnis A.S.
        • Holzbauer S.M.
        • Belflower R.M.
        • Winston L.G.
        • Bamberg W.M.
        • Lyons C.
        • et al.
        Epidemiology of community-associated Clostridium difficile infection, 2009 through 2011.
        JAMA Intern Med. 2013; 173: 1359-1367
        • McDonald L.C.
        • Coignard B.
        • Dubberke E.
        • Song X.
        • Horan T.
        • Kutty P.K.
        Recommendations for surveillance of Clostridium difficile-associated disease.
        Infect Control Hosp Epidemiol. 2007; 28: 140-145
        • Deyo R.A.
        • Cherkin D.C.
        • Ciol M.A.
        Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases.
        J Clin Epidemiol. 1992; 45: 613-619
        • Shapiro N.I.
        • Wolfe R.E.
        • Moore R.B.
        • Smith E.
        • Burdick E.
        • Bates D.W.
        Mortality in Emergency Department Sepsis (MEDS) score: a prospectively derived and validated clinical prediction rule.
        Crit Care Med. 2003; 31: 670-675
        • Huang C.C.
        • Kuo S.C.
        • Chien T.W.
        • Lin H.J.
        • Guo H.R.
        • Chen W.L.
        • et al.
        Predicting the hyperglycemic crisis death (PHD) score: a new decision rule for emergency and critical care.
        Am J Emerg Med. 2013; 31: 830-834
        • Garg S.
        • Mirza Y.R.
        • Girotra M.
        • Kumar V.
        • Yoselevitz S.
        • Segon A.
        • et al.
        Epidemiology of Clostridium difficile-associated disease (CDAD): a shift from hospital-acquired infection to long-term care facility-based infection.
        Dig Dis Sci. 2013; 58: 3407-3412
        • Dubberke E.R.
        • Butler A.M.
        • Yokoe D.S.
        • Mayer J.
        • Hota B.
        • Mangino J.E.
        • et al.
        Multicenter study of surveillance for hospital-onset Clostridium difficile infection by the use of ICD-9-CM diagnosis codes.
        Infect Control Hosp Epidemiol. 2010; 31: 262-268
        • Price M.F.
        • Dao-Tran T.
        • Garey K.W.
        • Graham G.
        • Gentry L.O.
        • Dhungana L.
        • et al.
        Epidemiology and incidence of Clostridium difficile-associated diarrhoea diagnosed upon admission to a university hospital.
        J Hosp Infect. 2007; 65: 42-46
        • Linder J.A.
        • Huang E.S.
        • Steinman M.A.
        • Gonzales R.
        • Stafford R.S.
        Fluoroquinolone prescribing in the United States: 1995 to 2002.
        Am J Med. 2005; 118: 259-268
        • Rotman S.R.
        • Bishop T.F.
        Proton pump inhibitor use in the U.S. ambulatory setting, 2002-2009.
        PLoS ONE. 2013; 8 (e56060)
        • McFarland L.V.
        • Mulligan M.E.
        • Kwok R.Y.
        • Stamm W.E.
        Nosocomial acquisition of Clostridium difficile infection.
        N Engl J Med. 1989; 320: 204-210
        • Jen M.H.
        • Saxena S.
        • Bottle A.
        • Pollok R.
        • Holmes A.
        • Aylin P.
        Assessment of administrative data for evaluating the shifting acquisition of Clostridium difficile infection in England.
        J Hosp Infect. 2012; 80: 229-237
        • Khanna S.
        • Keddis M.T.
        • Noheria A.
        • Baddour L.M.
        • Pardi D.S.
        Acute kidney injury is an independent marker of severity in Clostridium difficile infection: a nationwide survey.
        J Clin Gastroenterol. 2013; 47: 481-484
        • Bloomfield L.E.
        • Riley T.V.
        Epidemiology and risk factors for community-associated Clostridium difficile infection: a narrative review.
        Infect Dis Ther. 2016; 5: 231-251