Evaluation of the ATLAS scoring system in the prediction of clinical cure and recurrence of Clostridium difficile infection in a large academic hospital

Purpose: CDI treatment has become challenging with the advent of hypervirulent strains and increased presence of potential CDI risk factors. Several risk factors have been associated with the occurrence of severe CDI or CDI recurrence. Risk and severity assessment tools have been developed to aid treatment decisions. Recently a scoring system known as ATLAS based on 5 characteristics (age, temperature, leukocytosis, albumin, and concomitant antibiotics) was tested retrospectively in two studies. ATLAS scores correlated with mortality and cure rates (inversely). Thirty day recurrence rates did not correlate statistically with ATLAS scores. Neither study assessed patients treated with metronidazole. The purpose of this study is to determine if ATLAS predicts cure and recurrence (< 90 days) in patients receiving metronidazole and/or vancomycin.

Methods: This was an observational study conducted with adult patients with a positive C. difficile toxin-PCR 2-step test treated with metronidazole and/or vancomycin. Potential risk factors for severe CDI, decreased cure rates, or recurrence were recorded as were variables to calculate ATLAS scores. Cases were considered complicated if ileus, perforation, megacolon, colitis, colectomy, or hypotension was present.

Results: A total of 245 patients met inclusion criteria for the study. Median ATLAS score was 5.  ATLAS scores showed a significant inverse association with cure (P=0.009), but not with recurrence (P=0.901). The only ATLAS component to be independently associated with cure was concomitant use of antibiotics (P=0.022). Longer courses of vancomycin were associated with more cures (P=0.0009), but no difference in recurrence was found (P=0.170). Complicated cases were less likely to be cured (P=0.027). The only factors associated with recurrence were complicated CDI (P=0.002) and antibiotics continued after CDI treatment (P=0.055).

Conclusion: The ATLAS score appears to correlate well with cure rates in patients receiving vancomycin and/or metronidazole. The ATLAS score does not appear to be useful for predicting recurrence.