Bayesian estimation of post-test probability of Candida glabrata fungemia by means of serum creatinine

New Microbiol. 2022 Dec;45(4):324-330.

Abstract

Fungemia is a life-threatening condition associated with high mortality; the most frequently isolated genus is Candida. Candida glabrata is of particular concern because of its increasing resistance to azoles. We evaluated common lab tests accessible by almost all healthcare professionals to estimate the post-test probability of recovery of C. glabrata from a blood culture collected by venipuncture, positive for fungi identified by microscopic examination. Patients with blood cultures positive for C. glabrata had significantly higher median values of serum creatinine (P=0.006), and a value of ≥1.45 mg/dL was the best cut-off in discriminating C. glabrata from other Candida spp., with 0.67 [95% Confidence Interval (CI): 0.49-0.85] sensitivity and 0.75 (95% CI: 0.66-0.84) specificity; Youden's J statistic: 0.42. The receiver operator characteristic curve analysis showed an area under the curve of 0.718 (95% CI: 0.603-0.833); P=0.001. Therefore, given a pre-test probability of 24% and applying the Bayes' theorem, the post-test probability of C. glabrata fungemia with creatinine values ≥1.45 mg/dL increased to 45.8%. In conclusion, we showed how the probability of recovery of C. glabrata from blood cultures collected by venipuncture and positive for fungi can be better estimated using concurrent creatinine values.

Keywords: Bayesian Analysis; Blood Cultures; Candida glabrata; Fungemia.

MeSH terms

  • Antifungal Agents / pharmacology
  • Antifungal Agents / therapeutic use
  • Bayes Theorem
  • Candida
  • Candida glabrata
  • Candidiasis* / diagnosis
  • Creatinine
  • Fungemia* / etiology
  • Fungemia* / microbiology
  • Humans
  • Microbial Sensitivity Tests

Substances

  • Creatinine
  • Antifungal Agents