OBJECTIVES

To evaluate the potential role of PCR-based assays in the over-diagnosis of Clostridium difficile infection (CDI) by using a validated diagnostic algorithm in daily clinical practice.

METHODS

We performed a retrospective cohort study evaluating all C. difficile-positive stool samples identified at our institution during a 12-month period, to compare outcomes and recurrence rates between patients with a positive enzyme immunoassay (EIA) for both glutamate dehydrogenase (GDH) and toxin A/B ('toxin-positive group'), with those with GDH-positive, toxin-negative samples in whom the diagnosis was made by a positive PCR-based assay ('toxin/PCR group'). Medical records were reviewed by two independent investigators blinded to the mode of diagnosis.

RESULTS

We analysed 231 first CDI episodes (106 (45.8 %) in the 'toxin-positive group' and 125 (54.1%) in the 'toxin/PCR group'). Both groups had similar baseline characteristics. Patients in the 'toxin-positive group' presented more frequently with a severe/severe complicated form than those in the 'toxin/PCR group' (36 (33.9%) versus 24 (19.2%); p 0.011) and had more recurrences (27 (25.5%) versus 9 (7.2%); p 0.001). Diagnosis of CDI based on a GDH/toxin-positive EIA independently predicted severe/severe-complicated course (adjusted OR 2.11; 95% CI 1.06-4.22; p 0.033) and recurrence (adjusted OR 3.79; 95% CI 1.65-8.69; p 0.002). There were no differences in all-cause mortality (12.3% versus 12.0%; p 0.95) or CDI-attributable mortality (4.7% versus 4.8%; p 0.93).

CONCLUSIONS

Toxin-positive patients were more likely to have severe-complicated forms of CDI and recurrences. Nevertheless, CDI-related complications may still occasionally occur among toxin-negative patients diagnosed by PCR, which stresses the need for individualized clinical evaluation.