A Multipronged Approach to Decrease the Risk of Clostridium difficile Infection at a Community Hospital and Long-Term Care Facility

Methods

Patients and Data Collection

The study was conducted at a community hospital (59 beds) that has an associated LTC facility (122 beds). We conducted a retrospective analysis of hospital and LTC data from all documented cases of CDI between January 2009 and December 2013. Study subjects included all patients with stools positive for C. difficile antigen and toxin with associated symptoms of infection (n = 123). Institutional review board approval was obtained prior to data collection.

The following information was collected: admission diagnosis, number of days from admission until confirmed CDI, residence prior to admission, duration and type of antibiotics received prior to or during symptoms of CDI, type of GI prophylaxis received within 14 days prior to and during CDI treatment, probiotic received and duration, and the type and duration of antibiotic treatment given for the CDI. The data collected was used to determine the likely origin of each C. difficile  case, dates of recurrences, and the possible effects of the interventions. Antibiotic use was categorized as: (1) recent antibiotic course (antibiotics received within the preceding 4 weeks), (2) antibiotic courses greater than 10 days, and (3) multiple antibiotic courses (more than 1 antibiotic course received sequentially or concurrently).

Positive C. difficile  infections were detected using a 2-step algorithm, starting in 2009. The samples were first screened with a rapid membrane enzyme immunoassay for glutamate dehydrogenase (GDH) antigen and toxin A and B in stool (C. Diff Quik Chek Complete, Techlab, Blacksburg, VA). Discrepant samples (GDH positive and toxin A and B negative) were reflexed to DNA-based PCR testing. The PCR assay was changed to the Verigene C. difficile test (Nanosphere, Northbrook, IL) in 2012. Up to 30 days after discharge from our facility, positive results were considered as acquired from our facility and positive results within 2 days of admission with symptoms of CDI were considered positive on admission and were not attributed to our facility. A primary episode of CDI was defined to be the first identified episode or event in each patient. Recurrent CDI was defined as a repeated case of CDI within 180 days of the original CDI event.

Interventions to Reduce CDI

Reduction of Antibiotic Pressure

Other actions taken to improve antimicrobial prescribing as part of the stewardship program included medication usage evaluations (MUEs) for levofloxacin and carbapenems, implementing an automatic dosing/duration protocol for levofloxacin, and carbapenem restriction to prevent inappropriate use. Nursing and pharmacy staffs were educated on vancomycin appropriateness, benefits of MRSA screening for de-escalation, procalcitonin, and treatment of sepsis. Emergency department staff was educated on (1) empiric antimicrobial treatment recommendations for urinary and skin and soft tissue infections based on outpatient antibiogram data, (2) renal adjustment of antimicrobials, (3) fluoroquinolones: resistance formation, higher CDI risk and higher dosing recommendations, (4) GI prophylaxis recommendations, and (5) probiotics.

Reduction in the Intensity of Acid Suppression for GI Prophylaxis

PPIs were substituted with histamine-2 receptor antagonists (H2RA) whenever acid suppression for GI prophylaxis was warranted. If GI symptoms persisted, sucralfate was added. In May 2010, all eligible LTC patients were converted from PPIs to H2RA.

Expanding the Use of Probiotics

We expanded the use of probiotics as an adjunctive treatment for CDI with metronidazole ± vancomycin oral therapies. Probiotics were included concurrently with any broad-spectrum antibiotic administration, longer antibiotic courses (≥ 7 days), and/or multiple courses of antibiotics. The combination of Saccromyces boulardii plus Lactobacillus acidophilus and L. bulgaricus was given with twice daily dosing until the end of 2011. In January 2012, our facility switched over to daily administration of a probiotic with the active ingredients of Lactobacillus acidophilus and Lactobacillus casei, 50 billion colony-forming units. Probiotics were given during the antibiotic course plus for 1 additional week after course completion. Probiotics were not administered to selected groups of patients: (1) immunocompromised patients, (2) patients who were NPO, or (3) patients excluded by their physicians.

There was no change or enhanced targeting of infection prevention or environmental hygiene strategies during the study period.

Data Analysis and Statistical Methods

All data were collected on data collection sheets and transcribed into Microsoft Office Excel 2007 Service Pack 3. No data were excluded from analysis. Continuous variables, eg, number of cases of CDI, are reported as mean ± standard deviation. Categorical variables, eg, number of recurrent CDI cases, are reported as the count and percentage. Comparison of populations was done with the Wilcoxon rank sum test. Segments of the interrupted time series were assessed using linear regression. Associations were tested using χ2. Statistical tests were deemed significant when the α probability was < 0.05. No adjustments were made for multiplicity. Data descriptive statistics (including frequency histograms for visual examination of distributions) and statistical analyses were performed using Stata 11.1 (StataCorp, College Station, TX).