Prediction of invasive candidal infection in critically ill patients with severe acute pancreatitis
1 Intensive Care Unit, Royal Liverpool University Hospital, Prescot Street, Liverpool, L7 8XP, UK
2 Acute Medical Unit, Royal Liverpool University Hospital, Prescot Street, Liverpool, L7 8XP, UK
3 Medical Microbiology Department, Royal Liverpool University Hospital, Prescot Street, Liverpool, L7 8XP, UK
4 NIHR Pancreas Biomedical Research Unit, Department of Molecular and Clinical Cancer Medicine, Institute of Translational Medicine, University of Liverpool, Daulby Street, L69 3GA, UK
5 Department of Biomedical Statistics, University of Liverpool, Daulby Street, L69 3GA, UK
Critical Care 2013, 17:R49 doi:10.1186/cc12569
See related commentary by Montravers et al.,http://ccforum.com/content/17/3/137Published: 18 March 2013
Patients with severe acute pancreatitis are at risk of candidal infections carrying the potential risk of an increase in mortality. Since early diagnosis is problematic, several clinical risk scores have been developed to identify patients at risk. Such patients may benefit from prophylactic antifungal therapy while those patients who have a low risk of infection may not benefit and may be harmed. The aim of this study was to assess the validity and discrimination of existing risk scores for invasive candidal infections in patients with severe acute pancreatitis.
Patients admitted with severe acute pancreatitis to the intensive care unit were analysed. Outcomes and risk factors of admissions with and without candidal infection were compared. Accuracy and discrimination of three existing risk scores for the development of invasive candidal infection (Candida score, Candida Colonisation Index Score and the Invasive Candidiasis Score) were assessed.
A total of 101 patients were identified from 2003 to 2011 and 18 (17.8%) of these developed candidal infection. Thirty patients died, giving an overall hospital mortality of 29.7%. Hospital mortality was significantly higher in patients with candidal infection (55.6% compared to 24.1%, P = 0.02). Candida colonisation was associated with subsequent candidal infection on multivariate analysis. The Candida Colonisation Index Score was the most accurate test, with specificity of 0.79 (95% confidence interval [CI] 0.68 to 0.88), sensitivity of 0.67 (95% CI 0.41 to 0.87), negative predictive value of 0.91 (95% CI 0.82 to 0.97) and a positive likelihood ratio of 3.2 (95% CI 1.9 to 5.5). The Candida Colonisation Index Score showed the best discrimination with area under the receiver operating characteristic curve of 0.79 (95% CI 0.69 to 0.87).
In this study the Candida Colonisation Index Score was the most accurate and discriminative test at identifying which patients with severe acute pancreatitis are at risk of developing candidal infection. However its low sensitivity may limit its clinical usefulness.