The prognostic utility of tests of platelet function for the detection of ‘aspirin resistance’ in patients with established cardiovascular or cerebrovascular disease: a systematic review and economic evaluation

• Main Authors: Janine Dretzke, Richard D Riley, Marie Lordkipanidzé, Susan Jowett, Jennifer O’Donnell, Joie Ensor, Eoin Moloney, Malcolm Price, Smriti Raichand, James Hodgkinson, Susan Bayliss, David Fitzmaurice, David Moore
• Format: Article
• Language: English
• Published: NIHR Journals Library 2015-05-01
• Series: Health Technology Assessment
• Online Access: https://doi.org/10.3310/hta19370
• ISSN: 1366-5278; 2046-4924

Background: The use of aspirin is well established for secondary prevention of cardiovascular disease. However, a proportion of patients suffer repeat cardiovascular events despite being prescribed aspirin treatment. It is uncertain whether or not this is due to an inherent inability of aspirin to sufficiently modify platelet activity. This report aims to investigate whether or not insufficient platelet function inhibition by aspirin (‘aspirin resistance‘), as defined using platelet function tests (PFTs), is linked to the occurrence of adverse clinical outcomes, and further, whether or not patients at risk of future adverse clinical events can be identified through PFTs. Objectives: To review systematically the clinical effectiveness and cost-effectiveness evidence regarding the association between PFT designation of ‘aspirin resistance’ and the risk of adverse clinical outcome(s) in patients prescribed aspirin therapy. To undertake exploratory model-based cost-effectiveness analysis on the use of PFTs. Data sources: Bibliographic databases (e.g. MEDLINE from inception and EMBASE from 1980), conference proceedings and ongoing trial registries up to April 2012. Methods: Standard systematic review methods were used for identifying clinical and cost studies. A risk-of-bias assessment tool was adapted from checklists for prognostic and diagnostic studies. (Un)adjusted odds and hazard ratios for the association between ‘aspirin resistance’, for different PFTs, and clinical outcomes are presented; however, heterogeneity between studies precluded pooling of results. A speculative economic model of a PFT and change of therapy strategy was developed. Results: One hundred and eight relevant studies using a variety of PFTs, 58 in patients on aspirin monotherapy, were analysed in detail. Results indicated that some PFTs may have some prognostic utility, i.e. a trend for more clinical events to be associated with groups classified as ‘aspirin resistant’. Methodological and clinical heterogeneity prevented a quantitative summary of prognostic effect. Study-level effect sizes were generally small and absolute outcome risk was not substantially different between ‘aspirin resistant’ and ‘aspirin sensitive’ designations.: No studies on the cost-effectiveness of PFTs for ‘aspirin resistance’ were identified. Based on assumptions of PFTs being able to accurately identify patients at high risk of clinical events and such patients benefiting from treatment modification, the economic model found that a test–treat strategy was likely to be cost-effective. However, neither assumption is currently evidence based. Limitations: Poor or incomplete reporting of studies suggests a potentially large volume of inaccessible data. Analyses were confined to studies on patients prescribed aspirin as sole antiplatelet therapy at the time of PFT. Clinical and methodological heterogeneity across studies precluded meta-analysis. Given the lack of robust data the economic modelling was speculative. Conclusions: Although evidence indicates that some PFTs may have some prognostic value, methodological and clinical heterogeneity between studies and different approaches to analyses create confusion and inconsistency in prognostic results, and prevented a quantitative summary of their prognostic effect. Protocol-driven and adequately powered primary studies are needed, using standardised methods of measurements to evaluate the prognostic ability of each test in the same population(s), and ideally presenting individual patient data. For any PFT to inform individual risk prediction, it will likely need to be considered in combination with other prognostic factors, within a prognostic model. Study registration: This study is registered as PROSPERO 2012:CRD42012002151. Funding: The National Institute for Health Research Health Technology Assessment programme.