What is the Candida Score?
The Candida score is a clinical prediction tool developed to identify critically ill patients at high risk for invasive candidiasis by combining clinical risk factors with colonization status, using a point-based system where a cutoff value ≥3 indicates patients who may benefit from empirical antifungal therapy. 1
Score Components and Calculation
The Candida score assigns points based on the following risk factors 1:
- Surgery (recent major surgery, particularly abdominal): 1 point
- Multifocal Candida colonization: 1 point
- Total parenteral nutrition (TPN): 1 point
- Severe sepsis/clinical sepsis: 2 points
The total score ranges from 0 to 5 points, with a cutoff value of ≥2.5 (typically rounded to ≥3 in clinical practice) demonstrating sensitivity of 81% and specificity of 74% for predicting invasive candidiasis. 1
Clinical Performance and Validation
The score was originally developed in 2006 by a Spanish group using the Estudio de Prevalencia de CANdidiasis database, with subsequent validation in 2009 demonstrating a significant linear association between increasing score values and rates of invasive Candida infections 1.
Prospective validation studies have confirmed that patients with scores ≤3 have extremely low rates of invasive candidiasis (0%), while those with scores of 4 or 5 have substantially higher rates (17.6% and 50%, respectively). 2
Key Performance Characteristics:
- Negative predictive value: 84-96% across validation studies 3
- Positive predictive value: 25-47% (relatively poor) 3
- Area under ROC curve: 0.66-0.72 in external validation studies 3, 4
Clinical Application and Limitations
The Candida score is most useful for ruling out invasive candidiasis rather than confirming it, given its high negative predictive value but poor positive predictive value. 3 This makes it particularly valuable for identifying patients who do not require empirical antifungal therapy, thereby avoiding unnecessary antifungal use and associated costs 1.
Important Caveats:
- The score has high specificity but low sensitivity, meaning it may miss some patients with invasive candidiasis 1
- Performance may be limited in certain populations, including COVID-19 patients where one study found no difference in Candida scores between those with and without candidemia 1
- The score performs better when combined with biomarkers such as β-D-glucan (BDG), with one study showing that an integrated strategy using both Candida score >3 and BDG helped optimize antifungal therapy decisions without increasing mortality 1
Target Population
The Candida score is specifically designed for non-neutropenic critically ill patients in the ICU setting, particularly those with hospital-acquired severe sepsis or septic shock. 2 It is not validated for neutropenic patients or those outside the ICU environment 5.
Risk Stratification:
- Score 0-2: Low risk - invasive candidiasis highly improbable 2
- Score 3-4: Moderate risk - consider additional diagnostic testing 4
- Score ≥5: High risk - strong consideration for empirical antifungal therapy 2
Integration with Other Diagnostic Approaches
While the Candida score provides clinical guidance, it should ideally be combined with non-culture diagnostic methods such as β-D-glucan testing (sensitivity 56-93%, specificity 71-100%) or mannan/anti-mannan antibody assays (combined sensitivity 83%, specificity 86%) to improve diagnostic accuracy 1. Blood cultures remain the gold standard but are negative in up to 50% of invasive candidiasis cases, highlighting the need for complementary diagnostic strategies 1.
The score has been superseded by newer prediction models in some contexts, with one recent study developing a risk score incorporating CD8+ T-cell counts that demonstrated superior performance (AUROC 0.820 versus 0.711 for Candida score alone) 4.