Long-term surveillance of cefotaxime and piperacillin-tazobactam prescribing and incidence of Clostridium difficile diarrhoea

Objectives: We followed the effects of changes to a new antibiotic policy favouring a ureidopenicillin as opposed to a third-generation cephalosporin on the long-term incidence of Clostridium difficile diarrhoea (CDD) and antibiotic utilization in a large Elderly Medicine Unit. Patients and methods: In 1999, piperacillin-tazobactam was added to the formulary in Elderly Medicine and its use promoted in preference to cefotaxime. Following review and feedback to clinicians of surveillance data, cefotaxime prescribing was actively restricted during 2000-2001. An audit of prescriber adherence to antibiotic policy was carried out by reviewing the records of 159 patients during February-April 2001. In December 2001, due to manufacturer production problems, supply of piperacillin-tazobactam was stopped. We performed standardized period prevalence surveillance (February-April) allowing comparisons of antibiotic utilization and CDD incidence during the 5 year study period (1998-2002). Results: CDD incidence did not change significantly (P>0.1) during 1998-1999 despite a marked increase in piperacillin-tazobactam prescribing. However, when cefotaxime prescribing was curtailed in 2001, CDD rates decreased (in four of five wards) and overall by 52% (P = 0.008). When piperacillin-tazobactam became unavailable in 2002, despite advice to the contrary cefotaxime prescribing rose five-fold, and CDD rates increased in four of five wards and by 232% (P<0.01) overall. Adherence to antibiotic policy introduced in 2000 was good (81% accordance); 94%, 88% and 73% of patients with cellulitis, urinary tract and respiratory tract infection, respectively, received appropriate antibiotics. Conclusions: Long-term prescribing of piperacillin-tazobactam in Elderly Medicine in preference to cefotaxime is associated with reduced rates of CDD. However, unless cephalosporin prescribing is curtailed, the beneficial effects on CDD rates may be missed. This is one of few studies to document adverse effects due to loss of antibiotic supply.