Theme 2: Novel Diagnostics and Therapeutics
2.1 Reducing nosocomial transmission of multi-resistant organisms (MROs)
(AIs: Iredell, Gilroy, Byth; CIs Gilbert)
Background. Delayed treatment of sepsis in the critically ill increases mortality but excessive empirical use of broad-spectrum antibiotics can increase the prevalence of MROs. Antibiotic use may be reduced by earlier diagnosis or targeted therapy. Application of fluorescent in-situ hybridisation (FISH), to blood cultures that have signaled positive in an automated detection system, can reduce time to diagnosis by 1-3 days (Kempf et al, J Clin Microbiol 2000; 38:830)
2.1.1 Rapid identification of MROs using fluorescent in-situ hybridisation (FISH).
Aims 1) To evaluate the use of FISH probes to identify common MROs in blood cultures, on the time to diagnosis and on clinical outcomes in febrile neutropenic patients. 2) To evaluate effects, on MRO transmission, of using PCR and FISH to identify MROs in surveillance swabs. Methods Develop and evaluate multiplex PCR and FISH probes as tools for diagnosis and surveillance, targeting common fixed (eg mecA in MRSA), mobile (eg integrons and common & locally emergent plasmids in gram negative bacteria) and clinically important (eg. confer resistance to cefepime, timentin, carbapenems and gentamicin) resistance genes. Outcomes (morbidity; mortality; cost; time to isolation of patients with MROs) will be compared using PCR/FISH versus standard culture methods and phenotypic resistance testing.
2.1.2 Cycling of empiric antibiotic regimens in febrile neutropenic patients
Antibiotics used for empirical treatment of febrile neutropenia can select for antibiotic resistance in patients’ microflora. Scheduled cycling of antibiotic regimens may be a cost-effective way to reduce MRO carriage and invasive infection (Gruson, D et al, 2003; Crit Care Med 31:1908).
Aim. To evaluate the impact of scheduled changes in antibiotic protocols on 1) clinical outcomes and prevalence of MROs in sterile (eg blood) and surveillance sites (eg perineal) and 2) incidence of resistance markers in endogenous microflora. Methods. We will use PCR/FISH methods, as above, to measure changes in prevalence of resistance markers associated with different empirical antibiotic regimens.
2.2 Laboratory methods to monitor viral infection and disease
(CIs Dwyer, Gottlieb, Bradstock; AIs Gilroy, Shaw; Collaborator Ratanamohan)
Background In HSCT recipients, viral load assays (plasma viral copies using quantitative [Q] PCR) for CMV, and Epstein Barr virus (EBV), can help to distinguish infection from clinical disease. Plasma QPCR has not been evaluated for Human Herpes Virus (HHV) 6/7 or polyomavirus BK in HSCT.
Aims 1) To develop or modify and validate real-time QPCR assays to distinguish asymptomatic reactivation from clinically apparent disease due to CMV, HHV6/7 & EBV (Ratanamohan et al. Transplantation 1998; 66: 877-82). 2) Use QPCR to measure the efficacy of preventive measures including novel cell-based immunoprophylaxis (see 2.4) and valganciclovir chemoprophylaxis for prevention of CMV, EBV and HHV6/7 disease. 4) Measure BK viral load during transplantation, and correlate levels with clinically apparent disease. Methods Blood will be collected weekly from an estimated 120 allogeneic HSCT patients/year, using our routine CMV surveillance protocol. Viral loads will be correlated with clinical, laboratory and imaging parameters of disease.
2.3 Laboratory methods of diagnosis of invasive fungal (IFIs) and other infections
(CIs Sorrell Bradstock, Gottlieb, Kerridge, Gilbert; AIs Gilroy, Shaw; Collaborators Slavin, Byth: Australian Leukemia & Lymphoma Group (ALLG); PhD student of Slavin and Sorrell (Morrissey).
2.3.1 Rapid diagnosis of invasive aspergillosis (IA) and other invasive fungal infections (IFIs)
Background Prophylactic and empirical antifungal drugs reduce IFIs but not overall mortality; adverse reactions and cost are significant. IA incidence is 15% in acute leukemia/HSCT in Australia and mortality is ~60%, mainly due to late diagnosis. The impact of rapid tests (PCR, galactomannan (GM), on clinical outcomes is unknown. We have devised an Aspergillus genus-specific PCR, which, in a pilot study was positive 15d (mean) before other indicators of IA [Shaw (AI-A) et al, Blood 2003;102:972-3a, abst 3621]. We are developing reverse line blot (RLB) assays for key Aspergillus & Candida species and an enzyme immunoassay (EIA) for the virulence factor, phospholipase B (PLB).
Aims 1) To test the hypothesis that more sensitive, rapid diagnostic tests for Aspergillus and Candida species will reduce morbidity and mortality, by allowing earlier initiation of specific therapy. 2) To compare predictive values for fungal disease of the species-specific RLBs and PLB EIA, with those of the Aspergillus genus-specific PCR and GM, in high-risk children and adults.
Methods 1) A 3-year, 7-centre, tri-state, randomised controlled trial of the effect of adding PCR & GM tests to current diagnostic methods, on outcomes and costs of IA in 600 adult high risk patients (allogeneic HSCTs & acute leukemia). Control arm: *high-resolution lung CT scans plus microscopy +/- culture of sputum or bronchial washings - initiated for febrile neutropenia unresponsive after 96 hours of empirical antibacterial treatment. Test arm: * tests initiated if twice weekly screening PCR & GM tests become positive independent of fever. 1o endpoint: use of empirical antifungal therapy (EAFT). 2o endpoints: mortality, drug toxicity, hospital stay, break-through IFIs and costs. Sample size is based on intention-to-treat analysis and 40% current EAFT use, 80% power to detect 12% decrease, <2 interim analyses and 10% predicted dropout rate. 2) Blood will be collected twice weekly and bronchial secretions as indicated, from adults in the IA study and high-risk children. Sensitivity and specificity of RLB and PLB EIA will be compared with PCR & GM and clinical criteria (using consensus definitions of IFI). In 2 years 150 patients (2,000 samples) will accrue (est. from pilot study) 2.3.2 Magnetic resonance spectroscopy (MRS) (CI Sorrell, Collabs Kuchel, Mountford,Somorjai) Background We have obtained high-resolution MR spectra from fasting sera and diagnostic spectra from bacterial and fungal abscess specimens. MRS on urine can diagnose renal graft rejection. Samples require no processing and results are available in 10 min. Aim To test the hypothesis that metabolite profiling of tissues and body fluids by MRS is a rapid, sensitive and inexpensive diagnostic approach to infections and graft versus host disease (GVHD). Methods. Serial sera and urines will be collected from time of diagnosis and stored at –70oC. Test groups will include febrile neutropenics a) with culture-proven infection; b) no microbiological diagnosis, but responsive to antibiotics; c) unresponsive to antibiotics; d) neutropenic, afebrile controls and e) patients with biopsy-proven GVHD. We will pilot the method on 30 patients per group (the minimum required for a valid “training” data-set). MR spectra will be analysed by supervised & non-supervised statistical methods and promising results validated with larger prospective studies.
2.4 Cell based therapies
(CIs Gottlieb, Bradstock, Sorrell, Kerridge; AIs Shaw, Byth, Dodds; Coll Abendroth)
Background Morbidity & mortality from viral and fungal infections remain high, despite improved prophylaxis and treatment. We hypothesize that reconstitution of specific immunological capacity by expanding the donor-derived pool of cytototoxic T-lymphocytes (CTLs) against human cyto-megalovirus (hCMV), other viral pathogens and Aspergillus spp. will reduce infectious morbidity and mortality in allogeneic HSCT recipients. This is based on our unpublished results from HSCT recipients infused with hCMV-reactive CTLs generated by culture of donor lymphocytes with dendritic cells coated with a peptide from the hCMV pp65 antigen. Use of an adenoviral vector containing the entire pp65 gene as immunogen, unlike the hCMV peptide approach, is not HLA-type restricted, and will generate CD4 and CD8 cells that recognise CMV.
Aims: 1) Evaluate the effect of prophylactic infusion of CMV-specific donor CTLs on CMV infection and disease in allogeneic HSCT recipients. 2) Assay proliferative responses of lymphocytes from healthy donors to viral and Aspergillus antigens. Methods: 1) 60 patients (over 3 years) will receive 2 x 107/m2 CMV-specific cells post allogeneic HSCT. Primary endpoint: CMV reactivation (determined by qualitative PCR and pp65 antigen levels) using historical controls. Sample size of 60 is 80% powered to detect a 70% reduction in incidence of reactivation from 26% to 8% (P value 0.05). 2) We will confirm that dendritic cells pulsed with Aspergillus antigens generate proliferative and cytotoxic lymphocyte responses (prelim data) before designing a clinical trial of adoptive immunotherapy. We will study the effect of lysates of VZV, RSV, herpes simplex virus (HSV) on normal donor proliferative and cytotoxic responses. Endpoints: cytotoxicity, cytokine production (flow cytometry, ELISPOT). Dendritic cells not pulsed with antigens will be used as in vitro controls.
2.5. Clinical algorithms and decision support for evidence based practice
2.5.1 Antifungal guidelines (Aim 1)
CIs as for 2.3 & MacIntyre; AIs Sintchenko, Gilroy; Aim 2 AI Selgelid; Kerridge, Ankeny, Sorrell).
Background Guidelines are designed to rationalize therapy based on evidence but factors other than efficacy and toxicity influence their development and use, including competing interests, such as economic factors, regulatory requirements, professional autonomy, clinical practice, patient and consumer demand, academic interest and commercial interests (marketing). Aims 1) To develop and implement best-practice, evidence-based national guidelines for antifungal drug use. 2) To understand factors influencing development and uptake of guidelines and provide a novel structure to explicitly identify competing influences and acknowledge stakeholder interests. Methods 1) We will incorporate data from the tri-state IA study (2.3.1), existing national guidelines for antifungal therapy [CI-A ref #4] and local disease prevalence data (which varies between hospitals), into a management algorithm for use within the NSW BMT network. 2) Methods of philosophical critique and text-based analysis will be used to examine the roles of factors listed above in development and use of national and institutional antifungal guidelines.
2.5.2 Decision support tools for evidence-based management of central venous line (CVL) infections and febrile neutropenia
(CI Gilbert; AIs Sintchenko, Iredell; Byth Collaborator Coriera)
Background Infected CVLs and febrile neutropenia cause significant morbidity, mortality and increased costs in haematology patients. Clinical variables, endogenous microflora and management affect incidence and outcomes. Clinical practice guidelines, especially if combined with decision support systems (DSS), can reduce practice variation and improve quality and cost-effectiveness of clinical decisions and patient care (Shiffman et al. JAMIA 1999; 6: 104). Aims 1) Identify risk factors that affect incidence and outcomes of CVL infection and febrile neutropenia, including prevalence of MROs (based on surveillance – see 2.1.1). 2) Develop DSS for individualised management of CVL infections and febrile neutropenia, incorporating individual clinical and laboratory findings, unit-specific antibiotic resistance, MRO prevalence data and evidence-based antimicrobial guidelines. 3) Prospectively assess efficacy of the DSS. Methods: 1) a) Retrospective/prospective review of CVL infections in the Westmead unit over 3-5 years (total ~300 episodes), to define and grade variables affecting outcomes eg underlying disease status; bacterial cause; antibiotic regimen. b) Case-control study to define and grade variables affecting CVL infection risk eg placement method and management; patients matched for duration of CVL placement & clinical risk group with uninfected controls. c) Selected interventions: RCT of antibiotic line-locks to salvage infected CVLs (see below) and “cycling” of empirical antibiotic regimens for febrile neutropenia (see 2.1.2). Endpoints: recovery, recurrence or death; effect on treatment schedule (eg line removed or salvaged; delays in chemotherapy or BMT); prevalence of MROs and costs. d) Pilot study comparing 2 CVL salvage protocols for treatment of coagulase negative staphylococcal (CNS) infections in neutropenic patients (estimated 30 eligible patients in 1 year). Patients will be alternately assigned to IV vancomycin alone or with a vancomycin line-lock (5 mg/mL), for 10 days and followed for 3 months. 1o endpoint: recurrent bacteraemia with the same CNS strain (phenotype and genotype); 2o endpoints: 2o bacteraemia or fungaemia, duration of line survival and mortality. Promising results will lead to a larger RCT. 2) a) Patients and controls from 1) a)-b) will be grouped according to outcome and, within each group, data will be randomly assigned to training or test sets (equal sizes). b) Data will be analysed by standard statistical methods to determine contribution of each variable. c) Develop classification rules for an electronic DSS to guide CVL infection management. 3) DSS will be implemented using hand-held computers to provide access at point of decision-making and efficacy assessed by comparison of outcomes before and after introduction. Endpoints: CVL infection rates, inpatient mortality from infection, length of stay, antibiotic cost savings, prevalence of MROs, compliance with protocols
