The authors of this research note that most cases of Clostridium difficile infection (CDI) are acquired in hospitals or long-term care facilities, and that patients with community-associated CDI (CA-CDI) may lack traditional risk factors. Although use of antibiotics in the outpatient setting is an established risk factor for CA-CDI, research suggests that some cases may occur in the absence of prior antibiotic exposure.
The purpose of the current meta-analysis was to evaluate the association between antibiotic exposure and CA-CDI, and to determine the antibiotic classes associated with the highest risk. A systematic review (last search December 2012) was conducted and eight retrospective case-control studies (varied n=152-13,563) meeting the inclusion criteria were included in the analysis (3 UK [including 2 using the General Practice Research Database], 3 USA and 1 Canadian). The definition of community acquisition varied (no history of hospital admission within the previous 8-52 weeks; one study included all stool samples submitted via GPs). All studies addressed antibiotic exposure but only five addressed individual antibiotic class. The antibiotic exposure assessment ranged from 30-180 days.
The main results reported were as follows:
• Recent antibiotic usage was associated with a higher risk of developing CA-CDI than no such exposure (8 studies; odds ratio [OR] 6.91; 95% CI 4.17-11.44; p<0.00001). This association remained significant when limiting the analysis to studies that provided adjusted data (7 studies; OR 7.33; 95% CI 4.26-12.59; p<0.0001).
• There was high heterogeneity across included studies (I2=95-6%) and the summary OR should therefore be interpreted with caution.
• Meta-analysis of antibiotic class data showed that the risk of CA-CDI was highest with clindamycin (2 studies, OR 20.43, 95% CI 8.50–49.09), followed by fluoroquinolones (3 studies, OR 5.65, 95% CI 4.38–7.28), cephalosporins (3 studies, OR 4.47, 95% CI 1.60–12.50), penicillins (4 studies, OR 3.25, 95% CI 1.89–5.57), macrolides (3 studies, OR 2.55, 95% CI 1.91–3.39) and sulphonamides/trimethoprim (3 studies, OR 1.84, 95% CI 1.48–2.29).
• Tetracyclines were not associated with an increased CDI risk (3 studies, OR 0.91, 95% CI 0.57–1.45).
A sensitivity analysis was conducted to explore the causes of persistent heterogeneity. High heterogeneity remained among studies grouped according to time of antibiotic exposure and sample size, and it was moderate when they were grouped according to the way that cases were selected (CD toxin assay; 4 studies; n=597, OR 7.67; 95% CI 4.19-14.06; I2=39%). Possible reasons for the significant heterogeneity postulated by the authors include variations in CD strain, dose and duration of antimicrobial therapy, concomitant illnesses and other medications (there was insufficient data on these to permit subgroup analysis). The authors comment that whilst there was significant heterogeneity in the estimates, the majority of studies had a similar direction of effect in demonstrating a significant association.
As the included studies were all observational, they are subject to residual confounding, and causality cannot be established. Other limitations include the possibility of confounding by indication (it could be that patients receiving antibiotics and those developing CDI are both sicker at baseline), and use of the random effects model for meta-analysis (greater reliance could have been placed on smaller studies with potentially inferior data).
The authors say that clinicians and patients should be aware that use of antibiotics in an outpatient setting can predispose to CA-CDI, including in patients with no other risk factors. Risk appears to differ between antibiotic classes and a lower risk agent should be selected where possible. Further research is required to determine whether dose or duration of antibiotic therapy alters the risk of CDI in the outpatient setting.