The Interaction of Age and Levofloxacin Exposure on the Incidence of Clostridium difficile

Background: Clostridium difficile (cdiff) infection is associated with significant morbidity and mortality. Advanced age and antibiotic exposure are established risk factors. Among the antibiotics, cephalosporins have been associated with cdiff for decades while exposure to fluoroquinolones has more recently emerged as an important risk.

Objective: This cohort study was designed to explore the interaction between age and exposure to levofloxacin (levo) on the risk of hospital-acquired cdiff.

Methods: Fletcher Allen Health Care is a 460 bed academic medical center. A retrospective cohort study was performed including all adults aged 20-99 hospitalized between 1/1/07 and 12/31/08. Psychiatry admissions were excluded. Patients were stratified by age into 10 year intervals. A pharmacy database was used to identify which patients received at least one dose of levo or ceftriaxone, and an infection control database was used to identify which patients were diagnosed with nosocomial cdiff based on a positive stool toxin enzyme immunoassay. The outcome of interest was a diagnosis of hospital-acquired cdiff within 30 days following exposure to levo. Comparison groups included those not exposed to levo as well as those treated with ceftriaxone.

Results: There were 202 cases of cdiff during or within 30 days of discharge for 38,914 adult admissions (0.5%, 0.8/1,000 patient-days). Ten of 202 (5%) died or had a colectomy judged to be attributable to cdiff. Exposure to levo occurred during 7% of all admissions, and 2.5% of patients who received levo developed cdiff compared to 0.4% who did not (p < .01) and 1.7% who got ceftriaxone (p = .14). 33% of all cdiff cases had exposure to levo vs. 11% for ceftriaxone (p < .01). The relative risk of cdiff for those with levo exposure vs. those without was 6.68 (4.99-8.94, p < .01), nearly twice the relative risk for ceftriaxone exposure = 3.63 (2.34-5.63, p < .01). For all admissions, the rate of cdiff increased in a precisely linear fashion with increasing age (see graph, slope = .12% increase per 10 years in age, p <.01, r = .96). When exposed to levo, the linear increase in cdiff with age became more dramatic (see graph, slope = .35% increase per 10 years, p = .01, r = .81). This relationship was not statistically significant for ceftriaxone (slope = .27%, p = .32, r = .41).

Conclusions: The rate of cdiff increases linearly with increasing age. Exposure to levo is common during adult hospitalizations, and occurs during one-third of admissions complicated by cdiff. The increase in cdiff with age is accentuated following levo exposure with a steeper slope of increase. This increase in risk was greater than that following exposure to ceftriaxone. Elderly hospitalized patients treated with levo should be targeted for cdiff prevention efforts.