The use of a risk group stratification in the management of invasive fungal infection: a prospective validation.

Abstract

Invasive fungal infection (IFI) is the major cause of infection-related morbidity and mortality in immunocompromised haematology patients receiving cytotoxic chemotherapy and stem cell transplantation (SCT). Following SCT, 10% of patients will develop invasive aspergillosis (IA) by 1 year (Marr et al, 2002) with an associated mortality of up to 80–90% (Lin et al, 2001; Marr et al, 2002). Several management strategies for IA have been developed, including prophylaxis, empirical therapy and pre-emptive therapy. Common to each of these approaches is the concept of risk. While empirical therapy represents a risk-based intervention for a narrowly defined group of patients with antibiotic resistant neutropenic fever, pre-emptive therapy, also known as targeted prophylaxis, recognizes an even more narrowly defined subset of patients at highest risk for IFI. Newer approaches to the management of IFI such as pre-emptive therapy require the availability of rapid, sensitive, non-culture diagnostic methods such as polymerase chain reaction (PCR) and galactomannan enzyme-linked immunosorbent assay. The performance of such diagnostic tests is likely to be optimal in patients at highest risk of IFI. Therefore, it is clear that if new management strategies are going to impact on mortality then the definition of risk of IFI is crucial. In response to this need, Prentice et al (2000) published a risk group stratification assigning patients to low, intermediate low, intermediate high and high risk of developing IFI. This was based on a detailed review of the literature and analysis of the relationship of IFI with clinical parameters including underlying disease, type of SCT, duration of neutropenia, graft versus host disease (GVHD), steroid use and fungal colonization. We now report a prospective validation of this risk stratification. We have performed a blinded prospective study of the use of a real-time, pan-fungal PCR assay for the early diagnosis of IFI in haematology patients undergoing intensive chemotherapy or SCT. Local ethical approval was granted and written informed consent obtained from each patient. During each treatment episode, blood was collected twice weekly for PCR and clinical, radiological and microbiological data were recorded. This was continued until neutrophil recovery or, for allogeneic SCT recipients, until day 100 (or beyond if GVHD developed). For each treatment episode, patient data were used to determine the highest risk group achieved according to the Prentice risk stratification (Prentice et al, 2000). Invasive fungal infection was defined according to European Organization for Research and Treatment of Cancer/Mycoses Study Group criteria (Ascioglu et al, 2002). A genuine positive PCR result was defined as at least two consecutive positive results. We determined the relationship between risk group, evidence of IFI, use of empirical therapy and sequential positive PCR (Fig 1). A total of 133 patients with 204 treatment episodes were included in the analysis. Ninty-five per cent of patients received antifungal prophylaxis. Evidence of IFI was documented in 22 (10·8%) episodes: three proven, nine probable and 10 possible. The development of IFI and sequential positive PCR results correlated directly with risk stratification, with the highest proportion of each occurring in the high-risk group and the proportion falling as the risk declined. The use of empirical therapy was almost equal in the high- and intermediate high-risk groups but was twice that seen in the intermediate low-risk group. Evidence of invasive fungal infection (IFI), use of empirical antifungal therapy and sequentially positive PCR results for each risk group* (n = number of treatment episodes per risk group). *No patients were in the low risk group due to the widespread use of central venous catheters. The data demonstrate that the risk stratification described by Prentice et al (2000) is a valid method for identifying those patients at highest risk for the development of IFI. The incorporation of this protocol into prophylactic and pre-emptive treatment strategies may improve the outcome for immunocompromised haematology patients at highest risk of developing IFI. We are grateful to Prof. I. Franklin, Dr G. McQuaker, Dr N. P. Lucie, Dr E. J. Fitzsimons, Dr R. Soutar, Dr M. Leach, Dr P. R. E. Johnson and Dr J. Davies for their encouraging recruitment of patients to the study. We also thank Dr L. Mitchell and Ms C. Nugent for obtaining the consent of patients and collecting samples and Dr A. Prentice for initial appraisal of these data. This work was funded by the Chief Scientist Office for Scotland, Gilead Sciences Ltd and Wyeth Health Care. Dr L. A. McLintock is funded by an educational grant from Gilead Sciences Ltd.