The prognostic and diagnostic 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

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 10/36/02. The contractual start date was in December 2011. The draft report began editorial review in June 2013 and was accepted for publication in March 2014. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Marie Lordkipanidzé has received speaker honoraria from Eli Lilly, which manufactures the antiplatelet agent prasugrel. David Fitzmaurice has received honoraria from Boehringer Ingelheim, Sanofi-aventis and AstraZeneca, but not in relation to antiplatelet therapy. All other authors declare no competing interests.

Copyright statement

© Queen’s Printer and Controller of HMSO 2015. This work was produced by Dretzke et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Indications for antiplatelet therapy

The use of antiplatelet agents covers a large spectrum of vascular diseases. 4 In primary prevention, antiplatelet agents can be given to patients at high risk of thrombotic events, such as patients with multiple risk factors for CAD or diabetes. 5 In secondary prevention, antiplatelet agents can be given either acutely in patients with acute coronary syndromes (ACSs), following percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG), or chronically in patients with stable CAD, in patients with a history of transient ischaemic attacks (TIAs) or strokes and patients with peripheral arterial disease (PAD). 5 The benefit of aspirin therapy in each of these pathologies is related to the underlying thrombotic risk, and is usually greatest in high-risk individuals and lowest in individuals with no overt atherosclerotic disease (Figure 1).

FIGURE 1.

The benefit of aspirin in terms of risk prevention in different patient groups. From Eikelboom JW, Hirsh J, Spencer FA, Baglin TP, Weitz JI. Antiplatelet drugs: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012;141(2 Suppl.):e89–119. 5 Reproduced with permission from the American College of Chest Physicians. MI, myocardial infarction.

Defining aspirin response

The efficacy of aspirin to prevent thrombotic events in cardiovascular patients is well established, with > 100 randomised trials having been conducted in high-risk patients and demonstrating a reduction in vascular death of approximately 15% and a further reduction in non-fatal vascular events of approximately 30%. 4 Few drugs have demonstrated similar efficacy, with up to 50 major vascular events avoided per 1000 patients treated for 1 year, at a cost of one to two patients experiencing a major GI bleeding event. 5 Both the benefit and the risk associated with aspirin are attributed to its ability to prevent thrombus formation via inhibition of platelet function. 14

The best-characterised mechanism of aspirin is acetylation of a key enzyme in platelet function, the cyclo-oxygenase (COX)-1 enzyme. This enzyme transforms arachidonic acid into thromboxane A₂ (TxA₂), a platelet agonist and vasoconstrictor. 15–17 Aspirin is effective in inhibiting platelet activity at doses as low as 20–40 mg per day,18 and is clinically effective in preventing thrombotic events in daily doses as low as 75 mg with little benefit of higher doses. 5 This is particularly important in view of the fact that though low doses of aspirin appear effective in preventing thrombotic events in patients at risk, the effect on bleeding (especially GI bleeding) has been shown to be aspirin dose dependent. 19 In recent years, it has been shown that even acutely well-managed major bleeding events are associated with worse outcomes in cardiovascular patients, in terms of both major adverse cardiovascular events and mortality. 20,21 It follows that most treatment guidelines advocate the use of the lowest aspirin dose effective in preventing thrombotic complications so as to minimise the risk of major bleeding. 22–25 From this, a need for monitoring of aspirin therapy has emerged and prompted the development and investigation of numerous assays of platelet function.

Platelet function assays

A vast array of platelet function assays is available to test the response of platelets to the inhibitory effect of aspirin (Table 1). Some assays are laboratory based and require extensive expertise to operate, whereas others have been specifically developed to be point of care. Although some assays study global haemostasis, most platelet function assays target a specific phase of platelet function, from platelet adhesion to platelet activation, secretion and aggregation. Important methodological disparities make the assays unique in the way that they assess platelet responses. For example, some of these assays are carried out in whole blood [including whole-blood aggregometry (WBA), platelet counting, platelet function analyser-100 (PFA-100^®; Siemens, Malvern, PA, USA), VerifyNow^® Aspirin (Accumetrics Inc., San Diego, CA, USA), Impact-R^® (DiaMed, Cresier, Switzerland) and flow cytometry], whereas others require sample preparation [such as light transmission aggregometry (LTA), plasma or serum thromboxane B₂ (TxB₂) measurement], and others can be performed on urine (levels of the TxB₂ metabolite 11-dehydro-TxB₂). There is no official guideline recommending one assay over another, and platelet function testing is not recommended for routine clinical testing in patients requiring aspirin therapy. As a result, many of the available platelet function assays have been used in a research capacity, and part of the uncertainty surrounding the definition and clinical relevance of aspirin resistance is due to the non-interchangeable nature of these assays.

From a pharmacological perspective, the monitoring of aspirin efficacy requires assessment of the ability of aspirin to inhibit its pharmacological target (platelet COX-1), and thus inhibit the conversion from arachidonic acid to TxA₂. 29 This is the accepted measurement of the European Agency for the Evaluation of Medicinal Products to assess the efficacy of aspirin. 30 Assays measuring TxA₂ formation in clotting blood or in aggregating platelet-rich plasma thus appear ideal. However, TxA₂ cannot be easily measured in biological samples as it has a very short half-life in plasma (30–60 seconds). 31 As a consequence, assays measuring stable metabolites of TxA₂, most commonly TxB₂ (in serum/plasma) or 11-dehydro-TxB₂ (in urine), are the most widely used.

From a functional perspective, a multitude of platelet function assays are available to assess platelet responsiveness to aspirin. 32 Some assays require extensive technical expertise and are limited to specialised laboratories, whereas others are point of care and are meant as bedside tools. The assays that use arachidonic acid as the agonist require a functioning COX-1 to convert it to the active TxA₂ molecules which then elicit a platelet response; these are referred to as COX-1-specific (Table 2). TxA₂ is a secondary mediator of platelet activation and synergises with other platelet pathways33 to elicit full platelet responses. Therefore, aspirin therapy can also partly inhibit platelet activation induced by other agonists, such as collagen and epinephrine. 34,35 Platelet function assays based on these agonists have been used to quantify the platelet reactivity of platelets in patients taking aspirin, although these do not specifically assess the pharmacological efficacy of aspirin. 36 These are referred to as COX-1-non-specific assays (see Table 2).

Arguably, COX-1-specific assays may capture more faithfully the effect of aspirin on platelets and may therefore be preferable when looking at the pharmacological efficacy of aspirin. Moreover, COX-1-specific assays are directly targeted by aspirin and are not affected by concomitant antiplatelet therapy, whereas COX-1-non-specific and global assays will be influenced by other antiplatelet therapy used (e.g. in cases of dual antiplatelet therapy with aspirin and an ADP receptor blocker, such as clopidogrel).

Prevalence and natural history of ‘aspirin resistance’

When response to aspirin is assessed by COX-1-specific assays, little variability in platelet responses is seen, with almost complete inhibition of TxA₂-dependent platelet aggregation in almost all patients. 36–40 Far greater biological variability in aspirin-induced platelet inhibition has been reported36,37,39,41,42 when COX-1-non-specific assays have been used to assess platelet inhibition by aspirin. The definition of normal response to aspirin has also lacked standardisation, and insufficient platelet response to aspirin, or ‘aspirin resistance’, has been reported in various fashions, including tertiles/quartiles of response as well as dichotomisation based on arbitrary cut-off values. Strikingly, the correlation between the results obtained with the various platelet function assays is disappointingly low,36,37,39,41,42 thus making the studies using different platelet function assays difficult to compare.

Despite the uncertainties surrounding the best way to test for aspirin effects, platelet function assays have provided a number of potential mechanisms to explain some of the variability seen in platelet reactivity in patients taking aspirin. 43,44 As none of these factors fully explain the variability seen in patients, the phenomenon of aspirin resistance is likely to be multifactorial.

In order to assess the efficacy of aspirin, it must be ascertained that the person being assessed has indeed ingested aspirin. However, non-compliance with prescribed aspirin therapy is common and thus compliance needs to be verified. 45,46 Although crucial to the determination of platelet response to aspirin, assessment of compliance is often lacking in studies of aspirin resistance. In a recent report on the use of secondary prevention drugs in patients with established cardiovascular disease, Prospective Urban Rural Epidemiology (PURE) study investigators found that approximately one-quarter of patients with an indication for aspirin therapy were actually taking it,47 making assessment of compliance a necessity prior to platelet function testing. In studies where aspirin administration was actively monitored, the majority of patients who were aspirin resistant on initial testing became responsive to aspirin upon retesting following observed ingestion. 46,48 Thus, in fully compliant patients, aspirin resistance may be a rare but important biological phenomenon. 45,49,50 Another important variable to control for in studies of aspirin resistance is the presence of interacting drugs. A well-described interaction between aspirin and NSAIDs such as ibuprofen and naproxen [but not rofecoxib (Vioxx^®, Merck Sharpe & Dohme), celecoxib (Celebrex^®, Pfizer), meloxicam, acetaminophen or diclofenac] has been shown to have an impact on platelet aggregation responses. 51–54 These drugs prevent aspirin from binding to its target, platelet COX-1. Therefore, current guidelines recommend that concomitant use of NSAIDs with aspirin should be carefully avoided. 23

Other factors have been consistently associated with altered platelet responses to aspirin. Genetic factors are known to be associated with variability in platelet responses to aspirin. 40 In a large study of over 1800 participants treated with aspirin, heritable factors contributed to 27–77% of variability in platelet function assay results, most importantly in COX-1-non-specific assays, whereas COX-1-specific assays were influenced by less than 2% by heritable factors. 40 Among considerable environmental factors, obesity plays an important role. Indeed, increased waist circumference and higher body mass index have been associated with reduced efficacy of aspirin to inhibit platelets. 48 This is especially important when enteric-coated aspirin tablets are used, as these also further reduce aspirin bioavailability. 48,55 In diabetic patients, aspirin resistance is more common, and platelets have an enhanced sensitivity to platelet agonists, which has been associated with metabolic alterations, oxidative stress and endothelial dysfunction. 56–62

Finally, recent evidence suggests that accelerated platelet function recovery may be a potential source of variability in platelet responsiveness to aspirin. The most striking example of platelet turnover involvement in platelet responsiveness to aspirin is in patients suffering from essential thrombocythemia (ET), a natural disease model of enhanced platelet generation. In ET, recovery of platelet function occurs within 24 hours despite daily aspirin therapy and is due to the formation of a large number of new uninhibited platelets from megakaryocytes, resulting in an increased rate of platelet turnover. 63–65 The phenomenon is not, however, limited to ET; both in healthy volunteers and in patients suffering from CAD or diabetes, increased platelet turnover has been associated with insufficient platelet inhibition by aspirin. 66–70 Increasing the frequency of aspirin administration to twice daily has been shown to effectively improve the inhibition of platelet function by aspirin in these settings, although the clinical benefit of this therapy modification remains unknown. 63,64,71–73

Although the characteristics associated with poor response have been explored in detail,74 it is noteworthy that the different studies have used different platelet function methodologies to explore the determinants of platelet responses. In parallel, a number of different studies have shown platelet function assay results to lack correlation and agreement among themselves, thus identifying different patients as poor responders to aspirin and having different determinants of response. 37,75,76 Which platelet function assay, if any, is the most clinically predictive of future major adverse cardiovascular events remains to be established. 77 As a consequence, the natural history of aspirin resistance remains somewhat uncertain. There is a need to address basic questions on the prognostic and diagnostic utility and cost-effectiveness of platelet function testing in the context of aspirin therapy before testing can be recommended in clinical practice. A number of systematic reviews attempting to address this basic question have been published in recent years. In general, these have failed to sufficiently capture the volume of available evidence or consider the heterogeneous nature of the evidence reviewed. These reviews are explored in more detail as part of the results section of this report (see Chapter 5, Systematic reviews). As detailed in Chapter 3, the aims of this report were to address this question of prognostic and diagnostic utility of platelet function testing in the context of aspirin therapy.

Selection criteria

Two broad types of study were considered relevant for this review: those studies that provide information on the prognostic or diagnostic/predictive utility of PFTs and those that report prognostic models, in which a PFT is one of multiple prognostic factors predicting clinical outcomes in a population of interest. The selection criteria for each are outlined below.

Searches

The following bibliographic databases were searched:

• The Cochrane Library (Wiley) (issue 4 of 12) [including the Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects, HTA Database, NHS Economic Evaluation Database (NHS EED) and Cochrane Central Register of Controlled Trials] to April 2012, MEDLINE (Ovid) from 1950 to 2012, MEDLINE In-Process & Other Non-Indexed Citations (Ovid) to 25 April 2012 and EMBASE (Ovid) from 1980 to 2012.

Search strategies combined index and text words encompassing the technologies (platelet function testing) and the patient group (cardiovascular disease, CVD and diabetes), as well as focusing on aspirin resistance. The Zetoc database (The British Library), Conference Proceedings Citation Index and Science Citation Index (Web of Science) were searched for conference proceedings. ClinicalTrials.gov, the UK Clinical Research Network Study Portfolio Database, the World Health Organization International Clinical Trials Registry Platform and the metaRegister of Controlled Trials were also searched for ongoing studies.

Reference lists of relevant articles, particularly systematic reviews, were hand-searched to identify other potentially relevant articles. Furthermore, a subject expert was used to identify any studies which may not be identified using standard methods.

Restrictions on publication language and date were not applied to the searches.

Copies of the search strategies used in electronic databases can be found in Appendix 1.

In addition, abstracts from the following national and international proceedings were hand-searched from 2009 onwards:

• platelet conferences (Platelets International Symposium)

• cardiology conferences (British Cardiovascular Society, American College of Cardiology, European Society of Cardiology, American Heart Association, American College of Chest Physicians)

• stroke conferences (International Stroke Conference, American Stroke Association)

• haematology conferences (British Society for Haematology, International Society on Thrombosis and Haemostasis, International Society for Laboratory Haematology).

Abstracts that were identified were considered for relevance in a similar way to fully published studies/articles.

Search results were entered into reference management software [Reference Manager version 11 (Thomson ResearchSoft, San Francisco, CA, USA)]. Duplicate records were removed by built-in algorithms and subsequent manual checking.

The searches of electronic databases were undertaken in April 2012 and were not updated after this time. A note was made of any additional relevant studies published subsequently that came to the attention of the authors of this report. These studies were not reviewed to avoid bias. A brief comment is made about these studies in Chapter 5, Relevant studies identified after the search cut-off dates.

Study selection

Study selection was undertaken as a two-step process. Titles (and abstracts where available) in records were initially screened by two reviewers, using prespecified screening criteria. These criteria were kept necessarily broad as it was anticipated that not all relevant information would necessarily be presented in an abstract, and thus the use of stricter criteria was likely to lead to the exclusion of relevant articles at this screen stage. These criteria were based on whether or not the records indicated that articles were about, or likely to be about, platelet function testing; reported, or were likely to report, clinical outcomes measured after a PFT; and were about patients who had or were likely to have cardiovascular/cerebrovascular or diabetic disease and were receiving aspirin therapy.

An additional criterion for conference abstracts was that these needed to be published from 2009 onwards to be retained. Letters to journals were not automatically classed as irrelevant, because often new results relevant to this field are made available through this medium.

Full texts of any potentially relevant articles or those where a decision could not be made were sought. In the second part of the two-step selection process, full-text articles were assessed against the full inclusion criteria by two reviewers independently. Any discrepancies between reviewers were resolved by discussion or by referral to a third reviewer. A copy of the selection form used for this process is available on request.

Both stages of the selection process were piloted prior to full implementation.

At title and abstract screening and for full-text screening, appropriate portions of non-English-language articles were translated where necessary to aid the selection process.

A record was kept of all decisions made, the reason for exclusion from the review at the full-text screening stage, articles that were not obtainable even by The British Library and also cases where decisions could not be made owing to missing information in a paper or abstract. In the case of this last scenario, an e-mail was sent to an author requesting further information.

During the selection process, any study identified that was thought to be of relevance to the cost-effectiveness review was cross-checked against the search results for that review to ensure comprehensiveness.

Assessment of risk of bias

Risk of bias was assessed by one reviewer and independently checked by a second reviewer. Disagreements were resolved by discussion.

Data extraction

Data extraction was conducted by one reviewer using a standardised, piloted data extraction form, and independently checked by a second. Disagreements were resolved through discussion or referral to a third reviewer.

The data extraction process was necessarily complex owing to the nature and variability of the included studies. Data extraction was undertaken directly into a specially created sheet in Microsoft Excel 2010 (Microsoft Corporation, Redmond, WA, USA). Extensive data related to the following domains were extracted: study design and characteristics; patient characteristics; antiplatelet regimens; PFT utilised; outcome measures and length of follow-up; data required for analyses; statistical methods employed and their appropriateness. Studies were grouped according to whether patients were on monotherapy (aspirin only) or dual therapy (with a second antiplatelet agent such as clopidogrel added to aspirin), in order to distinguish between patients in a stable (monotherapy) or acute phase (dual therapy) of thrombotic disease. Patients who have experienced ACS, or who have undergone PCI, will generally have a second agent added to their therapy for up to 1 year before reverting back to monotherapy. Note that for reasons outlined in Presentation of results, only results pertaining to monotherapy studies have been presented in this report.

For further details on data extracted, readers can consult a copy of the database via information presented in Appendix 4.

With regard to the data extracted for analysis, details are given in the following section.

Analysis

Amendments to protocol

Initially the protocol did not specify that studies of patients on dual/triple antiplatelet therapy [i.e. aspirin with additional antiplatelet agent(s)] had to employ an aspirin-specific PFT, rather than any PFT. This was changed prior to study selection and the pertinent platelet function assays are reflected in Table 2.

It was originally stated that studies which met all of the inclusion criteria except for reporting clinical outcomes would be noted, as these might provide useful information for cost-effectiveness analysis (e.g. uncertainty around the prevalence of those defined as aspirin resistant from specific assays in specific populations). From very early in the study selection process, the protocol was amended to omit this owing to the very large number of studies being identified and limited benefit of identifying these across all the tests and populations.

These amendments were reported to the NIHR and a revised protocol was submitted.

Presentation of results

Throughout the following sections, our aim has been to highlight the heterogeneity between studies with regard to population, PFT, outcomes and analysis of studies.

Results have therefore been separated according to whether patients were receiving only aspirin as antiplatelet therapy (monotherapy) or aspirin and a second antiplatelet agent (dual therapy) at the time of the PFT. There are a number of reasons for this:

• Populations receiving monotherapy are potentially likely to differ from those receiving dual therapy (e.g. they are less likely to have very recently had an acute cardiovascular event or to be undergoing non-elective PCI).

• The influence of a second antiplatelet agent on an aspirin-specific PFT is unclear.

• The second antiplatelet agent is likely to influence occurrence of clinical outcomes, and occurrence of outcome is fundamental to determination of prognostic utility.

• Resistance to other antiplatelet agents is known, and may affect event rates.

The original intention was to report and analyse studies relating to both patients receiving monotherapy and those receiving dual therapy. It was decided to undertake a stepwise approach to the analysis, starting with monotherapy studies and then moving on to dual-therapy studies; based on the reasons listed above, it is possible that an association between aspirin resistance and clinical outcome may be more apparent within those populations receiving aspirin therapy alone, as it might be more difficult to demonstrate prognostic utility in patients receiving aspirin with additional antiplatelet therapy because of the potential added confounding effect of the other antiplatelet agent.

Furthermore, it is debateable whether or not analysis of studies with dual therapy is warranted in the absence of demonstrated prognostic utility of platelet function testing in patients treated with aspirin as monotherapy. As this criterion was not met (i.e. prognostic utility could not be adequately demonstrated), all results presented in the following sections relate to monotherapy only. However, in the interest of transparency the authors wish for all extracted and analytical data (including those from dual-therapy studies) to be available to readers of this report. The data have been made available through a web portal and further details can be found in Appendix 4, including how to access the data.

Monotherapy studies were further defined as those where all, or the vast majority of, patients were on monotherapy at the time of the PFT, given that treatment strategies may change over time depending on disease progression. Adding a second agent may affect the rate of clinical events, and this may not be independent of the underlying risk, as higher-risk patients are more likely to be receiving or to commence dual therapy. Where studies have clearly specified where a proportion of patients have at some point during the follow-up period switched therapy or received additional therapy, this information has been extracted. It is, however, possible that not all studies have reported this information.

Populations have been broadly classified as having (i) stable CAD, (ii) stable CVD/stroke, (iii) PAD/peripheral vascular disease (PVD) or (iv) unstable angina (UA)/ACS. Where patients are undergoing elective PCI (PCI) or primary PCI (PPCI), this has also been indicated. Where the population comprises several patient groups, this has been classified as miscellaneous. Note that some acute populations have been included where the PFT was undertaken when patients were on monotherapy.

Results have been separated for different PFTs, and where several thresholds or agonists have been used, this has been indicated. Where different PFTs have been used within the same study, results have been presented in Chapter 5, Studies with more than one test.

Outcomes have been classified as (i) death, (ii) MACE, (iii) ischaemic/thrombotic or (iv) haemorrhagic/bleeding. A consistent definition for MACE is not used in the literature;85 for example, it may or may not include stroke. For a composite outcome that includes cerebrovascular complications, the abbreviation MACCE is sometimes used (with the additional ‘C’ indicating the cerebrovascular component), but again, this is not consistent. Rather than devise a definition of what constitutes MACE or MACCE for this report, studies with a composite outcome of adverse cardiovascular events have been grouped together using the abbreviation MACE. Where stroke has been reported as a separate outcome, this has also been highlighted.

Within the categories of MACE/MACCE there are some inconsistencies between studies in how this has been defined; this has been appropriately highlighted where necessary. The category of ‘ischaemic/thrombotic’ events is broad and encompasses a number of different events such as revascularisation, angina, bypass surgery, cardiovascular readmission, graft occlusion, MI, etc.

The different outcome measures used in the studies have been summarised as a first step in deciding whether or not pooling is possible and to give an idea of the range of outcome measures used. They have been grouped according to the following: sensitivity and specificity, unadjusted or adjusted ORs, or unadjusted or adjusted HRs. Where HRs or ORs have not been presented but have been calculated for this report, this has been indicated. Additionally, where groups have been collapsed in order to provide a single threshold, this has also been indicated. Note that where outcomes have been reported for different test characteristics (e.g. different agonist, threshold, etc.), not all results will necessarily have been summarised using the same outcome measures.

Odds ratios and HRs provide information on the usefulness of a PFT as a prognostic risk factor. Adjusted ORs or HRs may take into account differences in clinical characteristics, which are linked to adverse events. At the least informative level, articles have only provided a narrative statement regarding the relationship between PFT results and clinical events.

Quality assessment of studies is also clearly presented to aid interpretation of findings.

Owing to the extensive nature of the data extracted from included studies for this project, it was deemed unfeasible to adequately present all the data in this report (even as appendices). The results section of the prognostic utility review in this report contains, where necessary, details of the studies, including the populations studied, test characteristics and quality-related features, and data for key outcomes are presented in illustrative forest plots.

Quantity of research available

The searches resulted in the identification of 16,583 records (after automatic removal of duplicate records) and one further record from checking reference lists of relevant systematic reviews. Manual removal of duplicate records left 13,795 article records. Screening of titles and abstracts in these records indicated that 12,581 were not relevant. Full-text articles of the remaining 1214 were sought. Twenty of these articles were unobtainable and these are listed in Appendix 5; 65 were reports of ongoing studies and these are commented on later in this chapter (see Ongoing studies); and 1129 full-text articles were obtained for assessment against the inclusion criteria. Nine hundred and thirty-three articles were excluded and these are listed in Appendix 6, Table 85 with reasons for exclusion; 12 of these were excluded because there was insufficient information available to make a decision despite requests by e-mail to the authors for further details (see Appendix 6, Table 86).

One hundred and ninety-six articles met the inclusion criteria. Of these, 62 contained details of PFT results and clinical outcome data but failed to report the outcome data in relation to the test result, and thus provided no relevant information on prognostic utility of the PFT. These studies are listed in Appendix 7.

A further 119 included articles all reported clinical outcome data in relation to the result of one or more PFTs. 46,76,86–202 These articles report the findings of 108 studies that are detailed in the subsequent sections of this report. The remaining 15 articles203–217 reported systematic reviews and these are described below (see Systematic reviews).

A flow diagram presenting the process of selecting studies can be found in Figure 2.

FIGURE 2.

Flow diagram showing study selection.

Study mapping

As outlined in more detail below (see Monotherapy), included studies were separated into categories based on whether enrolled patients were receiving aspirin as their only antiplatelet agent (monotherapy) or aspirin combined with one or more other agents (dual/triple therapy) at the time of the PFT, and by the type of PFT employed in the study. Subcategorisation was undertaken to distinguish studies in which the therapy at the time of platelet function testing remained the same during follow-up from those in which this changed (e.g. patients on monotherapy at the time of testing but subsequently receiving dual therapy). Subjective decision-making was required in some cases where a proportion of patients was receiving a different therapeutic regimen at the time of testing and/or follow-up (e.g. some on monotherapy and some on dual therapy at the time of testing and/or follow-up). If the proportion was considered small (≤ 5%) then these studies were categorised under the therapy of the larger proportion. If large (≥ 11%), then these studies were put into a separate category.

The result of this mapping of studies is shown in Table 3.

Of the 108 included studies with test data linked to clinical outcome data, 57 studies reported on a patient group solely or predominantly receiving aspirin as monotherapy at the time of testing,46,76,86,88,90,92,93,95,99,105,108–110,112,113,115–118,121,123,125,127,128,132,133,135,137,138,142,144–155,159,162–164,166,168,169,171,174,186,187,189,193,195,196,198,201,202 51 studies reported on a group of patients solely or predominantly receiving dual therapy87,89,91,94–98,100–102,104,106,111,114,119,120,122,124,126,129–131,134,136,139–141,143,147,150,156–161,165,167,170–173,175–185,188,190,191,192,194,197,199,200 and one study103,107 contained a mixed population of monotherapy and dual therapy. Five studies95,147,150,159,171 were able to be mapped to both monotherapy and dual therapy categories. Turning to categories of test, LTA and the PFA-100^® were the most frequently used tests in included studies, with VerifyNow^® Aspirin, thromboxane metabolites and WBA also frequently encountered. Thromboelastography (TEG) was less well represented. Several tests that fell outside of these categories were placed in a miscellaneous category and this included small numbers of studies that employed, for example, tests such as flow cytometry methods and various commercial assays not included in other categories. Proportions of tests used within the monotherapy studies were: LTA (25%, 19 studies16,88,90,93,95,112,113,121,125,142,147,149,155,159,163,164,169,174,187,201), VerifyNow^® Aspirin (9%, 7 studies86,92,99,105,133,162,171), PFA-100^® (28%, 21 studies76,99,108,109,112,115,116,118,123,127,132,135,137,138,144,145,150,162,186,187,189,193), thromboxane metabolites measurement (14%, 11 studies46,76,99,108,110,148,151,162,164,195,202), WBA (10%, 8 studies99,117,128,153,162,166,186,196), TEG (4%, 3 studies117,168,174) and miscellaneous tests (9%, 7 studies76,93,125,146,152,154,163,196,198). The corresponding proportions for dual-therapy studies were: LTA (25%, 14 studies95,96,98,100,101,104,111,120,122,124,126,130,136,143,147,157,159,160,172,188), VerifyNow^® Aspirin (21%, 12 studies87,91,98,101,119,157,161,170,171,178,182,190,194,200), PFA-100^® (25%, 14 studies89,94,98,101,102,120,124,129,131,139–141,150,157,173,181,185), thromboxane metabolites measurement (0%), WBA (18%, 10 studies97,100,106,114,158,165,167,175–177,179,180,183,184,191), TEG (5%, 3 studies156,197,199) and miscellaneous tests (5%, 3 studies98,101,134,157,192). Note that several studies utilised a range of tests concurrently in the same study population. These are also identified, along with the tests used, in Table 3.

Prognostic utility of tests

Population characteristics and quality assessment of studies are presented in the following sections. As outlined in more detail in Chapter 4, Presentation of results, the structuring of results has been guided by:

• population receiving monotherapy or dual therapy at the time of the PFT

• therapy received after the PFT

• PFT used

• outcome (death, MACE, ischaemic/thrombotic event, bleeding)

• outcome measures presented or calculable [(un)adjusted OR and HR, sensitivity and specificity]; note that sensitivities and specificities are presented in Appendix 3.

This is followed by a summary for each PFT. Studies where more than one PFT were performed concurrently are reported in Studies with more than one test.

Monotherapy

The tests identified for assessing platelet function in patients on monotherapy (aspirin only) are (i) LTA, (ii) VerifyNow^® Aspirin, (iii) measurement of urinary or serum/plasma 11-dehydro-TxB₂ concentrations, (iv) PFA-100^®, (v) WBA, (vi) TEG and (vii) other miscellaneous tests.

Light transmission aggregometry

Population and test characteristics

Nineteen studies88,90,93,95,112,113,121,125,142,147,149,155,159,162,164,169,174,187,201 were identified in this category, four of which were reported in abstract form only,162,164,169,174 and one as a letter. 88 Populations had CAD (six studies113,142,149,162,164,201), CVD/stroke (six studies88,95,121,125,155,159) or PAD/PVD (four studies90,112,147,169). There were three studies93,174,187 in patients with UA/ACS; in one of these93 patients were all undergoing PPCI. None of the studies reported how long patients had had their primary underlying condition for.

In 12 studies88,90,95,112,121,142,147,155,159,162,164,174 it appeared that patients were exclusively on monotherapy both at the time of the PFT and during follow-up. In two studies,113,125 around 4% and 5% of patients were on dual therapy (aspirin + clopidogrel) at the time of the PFT. Given the small proportion on dual therapy, these studies have been included in the ‘monotherapy’ category.

In a further four studies,149,169,187,201 patients were on monotherapy at the time of the PFT, and around 4%,149 25%169,201 or 45%187 of patients respectively went on to receive an additional antiplatelet agent (clopidogrel) at some point during follow-up. It is possible that not all studies have reported where a proportion of patients commenced additional therapies during follow-up.

In the study where patients underwent PPCI,93 patients were on monotherapy at the time of the PFT and all were on dual therapy (aspirin + clopidogrel) during follow-up. This study has been listed separately, as the addition of clopidogrel therapy in all patients may affect the rate of events, and may also be a reflection of underlying population differences compared with the other studies.

Comedications, where reported, included statins, beta-blockers, angiotensin-converting enzyme (ACE) inhibitors, angiotensin II receptor blockers, calcium channel blockers, nitrate esters, proton pump inhibitors and dalteparin (Fragmin^®, Pfizer). NSAIDs were not permitted (or had to be discontinued within a certain time period) in seven studies;88,90,112,142,149,159,201 one study155 stated that drugs known to affect PFTs were discontinued, and there were no details on NSAIDs in the remaining studies.

The number of participants in the studies ranged from 32 to 583 (see Table 4). Where reported, mean ages of patients ranged from 60 to 75 years, with most means around the mid to late 60s or early 70s. There were more men than women in 14 out of 15 studies that reported on this,88,90,93,95,112,121,125,147,149,155,159,163,174,187,201 with proportions of men ranging from 53% to 81%. Only one study142 included more women (54%). The proportion of patients with diabetes ranged from 11% to 47%, and that of smokers from 5% to 66% (where reported, see Table 4). All studies were conducted in hospital settings.

The dose of aspirin ranged between 75 mg/day and 325 mg/day, with the exception of one study155 where the dose was high, at 1000 mg/day. This study included patients with TIAs or reversible ischaemic neurological deficits. There were no details on dose in one study. 95 Details were variable across studies regarding the length of time patients had been receiving aspirin therapy, with some noting a minimum period and some whether patients were chronic or first-time users, but many giving no details (see Table 4). No study stated whether aspirin was provided in enteric or plain form, though one study93 noted that aspirin was in chewable form.

The main study characteristics are listed in Table 4 below. Note that in some studies baseline characteristics have been reported only according to resistant/sensitive groups or groups with/without adverse clinical events, rather than for the total study population.

TABLE 4 - Population characteristics (LTA, monotherapy)

AA, arachidonic acid; SD, standard deviation.

Study/country	Number of patients	Age (years)	Therapy	Main underlying condition	Selected other population details	Due to undergo vascular intervention?	Aspirin dose/frequency	Duration of aspirin therapy (prior to PFT)	Percentage aspirin resistant	Derivation of threshold/comment
Monotherapy at time of PFT and during follow-up
Abumiya 2011,95 Japan	144	Mean 66.22 (SD 9.73)	Mono	CVD/stroke	Smokers: 16.7% Diabetes: 22.2%	No	No details	No details	No details	Mean % aggregation stated by groups ± adverse event
Cha 2008,121 South Korea	107	Mean 64.3 (SD 12.0)	Mono	CVD/stroke	Smokers: 66% Diabetes: 46%	No	100 mg/day	PFT 5 days after being on aspirin	24 or 75	Depending on how tertiles were aggregated
De Boni 2011,159 Italy	32	Mean/median? 62.8 (range 40–84) for total group (includes patients on monotherapy, dual therapy and clopidogrel only)	Mono	CVD/stroke	No details	No details	100 mg or 200 mg (frequency not stated)	No details	0	Aggregation (AA) ≥ 20%
Feher 2011,88 Hungary (letter)	281	Mean 64.13 (SD 10.92) sensitive, mean 66.74 (SD 10.17) resistant	Mono	CVD/stroke	Smokers: n = 78 (28%) Diabetes: n = 72 (26%)	No	100 mg/day	Mean 8.4 (SD 3.6) months	18.1	No details
Feng 2011,201 China	136	Mean 74.9 (SD 7), range 60–89	Mono (25% dual during follow-up)	CAD	Smokers: n = 41 (30%) Diabetes: n = 37 (27%)	No details	100 mg/day	No details but not previously on aspirin	No details	Quartiles: aggregation results given for patients with events
Gum 2003,149 USA	326	Mean 59 (SD 15) resistant, mean 62 (SD 11) sensitive	Mono (3.7% dual at follow-up)	CAD	Smokers: Aspirin resistant (n = 17): 0% Aspirin sensitive (n = 309): 6% Diabetes: Aspirin resistant (n = 17): 18% Aspirin sensitive (n = 309): 25%	Yes (in ‘some’). Elective cardiac catheterisation	325 mg/day	≥ 7 days before enrolment	5.2	Aggregation ≥ 20% (AA), ≥ 70% (ADP). Unclear if resistance based on AA and/or ADP
Kempfert 2009,113 Germany	59	No details	Mono (5% dual at time of PFT)	CAD	No details	Yes: CABG and combined procedures	500 mg aspirin intravenously postoperatively, then 100 mg/day	No details	28.8	Defined as aspirin resistant if platelet aggregation exceeded the threshold of 30% despite in vitro addition of 25 µM aspirin
Linnemann 2009,112 Germany	57	Median 67.7 (range 44–90)	Mono	PAD/PVD	Smokers: n = 31 (31.6%) Diabetes: n = 45 (45.9%)	No	100 mg/day	At least 14 days	3.5	Resistance defined as the maximum aggregation values within the reference range (≥ 78%) despite aspirin medication
Lordkipanidzé 2011,162 Canada (abstract)	198	No details	Mono	CAD	No details	No	80–325 mg/day	No details	No details	No details (ORs reported)
Miyata 2011,164 Japan (abstract)	583	No details	Mono	CAD	No details	No	More or less than 100 mg/day (proportions not stated)	No details	No details	No details (narrative statement regarding event rate and platelet function)
Modica 2009,187 Sweden	334	Mean 72	Mono (45% dual at discharge)	UA/ACS	Smokers: 21% Diabetes: 20%	No details	75 mg/day	No details	No details	No details (HRs reported)
Ohmori 2006,142 Japan	136	Mean 75.4 (SD 9.4)	Mono	CAD	Smokers: n = 7 (5.1%) Diabetes: n = 15 (11%)	No	81 mg (frequency not stated)	No details	No details	Significant event difference noted for upper quartiles compared with others
Payne 2004,147 UK	54	Mean 69 (SD 8.5)	Mono	PAD/PVD	Smokers: 32% Diabetes: 19%	Yes: carotid endarterectomy	150 mg (frequency not stated)	4 weeks before surgery	0	Aggregation (AA) ≥ 20%
Schwammenthal 2008,125 Israel	105 (79 eligible for analysis)	Mean 63 (SD 12)	Mono (4% dual at time of PFT)	CVD/stroke	Smokers: Good response (N = 40): current smokers, n = 7 (18%); past smokers, n = 13 (33%) Partial response (N = 34): current smokers, n = 5 (15%); past smokers, n = 11 (33%) Complete unresponsiveness (N = 31): current smokers, n = 8 (26%); past smokers, n = 5 (16%) Diabetes: Good response (N = 40): n = 7 (16%) Partial response (N = 34): n = 11 (32%) Complete unresponsiveness (N = 31): n = 10 (32%)	No	100 mg (55%) or 325 mg (45%) (frequency not stated)	40% > 1 week before index event	60.7	Aggregation (AA) ≥ 20% (partial and complete unresponsiveness groups aggregated)
Sørensen 1983,155 Denmark	41	Mean 58 (range 34–73) based on total population, not stated for aspirin group	Mono	CVD/stroke	No details	No	1000 mg/day	No details	41.5	Platelet hyperaggregability defined as ‘secondary aggregation obtained by ADP concentration ≤ 1 µmol/l.’ No further details
Tan 2010,174 China (abstract)	250	Mean 62 (SD 17), range 48–71	Mono	UA/ACS	No details	No details	150–250 mg/day	No details	18.8	Aggregation (AA) ≥ 20%
van der Loo 2011,90 Switzerland	109	Mean 68.1 (SD 11.4) with events, mean 72.3 (SD 9.7) without events	Mono	PAD/PVD	Smokers: Reported by pack-years (1 pack-year of smoking based on someone who had smoked one pack of cigarettes per day, i.e. 20 cigarettes daily for 1 year) With events at follow-up (n = 66): 36.4 (SD 38.8) pack-years Without events at follow-up (n = 43): 27.9 (SD 28.2) pack-years Diabetes: With events at follow-up (N = 66): n = 23 (35%) Without events at follow-up (N = 43): n = 11 (26%)	Yes: percutaneous angioplasty	100 mg/day	No details	No details	Mean levels of platelet aggregation shown for groups with and without events
Zanow 2010,169 Germany (abstract)	75 and 54 (different aspirin regimens)	No details	Mono (25% dual during follow-up)	PAD/PVD	No details	Yes: endovascular revascularisation, peripheral bypass	100 mg/day or one-off dose postoperatively	No details	Daily aspirin group: 31 One-off aspirin group: 32	Aggregation (AA) ≥ 30%
Monotherapy at time of PFT, dual therapy during follow-up
Spectre 2011,93 Israel	54	Mean 59.7	Mono at PFT, all patients on dual post PFT	UA/ACS (undergoing PPCI)	Smokers: 30% Diabetes: 24%	Yes: PPCI	100 mg/day	Previous long-term use of aspirin in 73%	33.3 or 66.7	Depending on whether upper two or lower two tertiles were grouped together to provide a single cut-off

The test performed in 18 out of 19 studies was LTA. Most tests used arachidonic acid as an agonist, with some also using collagen and ADP, and sodium citrate as the anticoagulant (where reported). One study187 used a variant of LTA, an aggregometer that uses laser light scattering (the PA-200).

Most studies reported no details on the timing of the PFT after aspirin ingestion. One study125 noted that there were at least 6 hours between aspirin dose and PFT, and three other studies112,147,149 stated that there were up to 24 hours between aspirin dose and PFT. Table 5 provides details of test characteristics.

TABLE 5 - Test characteristics (LTA, monotherapy)

AA, arachidonic acid.

Study	Details of kit/manufacturer	Anticoagulant (concentration)	Agonist (concentration)	Time since last aspirin dose
Monotherapy at time of PFT and during follow-up
Abumiya 201195	MC Medical, Tokyo, Japan	‘Citrated blood’	Collagen (2 µM) Collagen (5 µM) ADP (1 µM) ADP (10 µM)	No details
Cha 2008121	Model 560 VS (Chrono-log Corporation, Havertown, PA, USA)	Sodium citrate	AA (0.5 mg/ml = 1.6 mM) ADP (10 µM)	No details
De Boni 2011159	Chrono-log 700-4 lumi-aggregation systems (Chrono-log Corporation, Havertown, PA, USA)	Sodium citrate (3.2%)	AA (0.5 mM increasing up to 1 mM)	No details
Feher 201188 (letter)	LTA (no further details)	No details	No details (reported elsewhere)	No details
Feng 2011201	LTA (no further details)	No details	AA	No details
Gum 2003149	PAP4 platelet aggregometer (BioData, Horsham, PA, USA)	Sodium citrate (3.8%)	AA (0.5 mg/ml = 1.6 mM) ADP 10 µM	1–24 hours before blood sampling
Kempfert 2009113	PAP-4 (moeLab, Berlin, Germany)	Citrate	AA (1 mM)	No details
Linnemann 2009112	Behring Coagulation Timer® (BCT®) (Dade Behring, Düdingen, Switzerland)	Sodium citrate (3.2%)	AA (0.5 mg/ml = 1.6 mM)	1–24 hours
Lordkipanidzé 2011162 (abstract)	LTA (no further details)	No details	AA (1.6 mM) ADP (5 µM) ADP (10 µM) ADP (20 µM)	No details
Miyata 2011164 (abstract)	LTA (no further details)	No details	AA Collagen	No details
Modica 2009187	PA-200	Sodium citrate (0.129 M)	Epinephrine (30 µl of a solution containing 0.1 mg epinephrine)	No details
Ohmori 2006142	LTA (no further details) and PA-20 platelet aggregation analyser	Sodium citrate (10%)	Collagen (1 µg/ml)	No details
Payne 2004147	PAP4 platelet aggregometer	Trisodium citrate (3.8% wt/vol)	AA (2.5 mM)	< 24 hours
Schwammenthal 2008125	PACKS-4 (Helena Laboratories, Beaumont, TX, USA)	‘Citrated blood’	AA (1.6 mM)	At least 6 hours before blood sampling
Sørensen 1983155	Turbidimetric aggregation (Born method)	No details	ADP (lowest ADP concentration that could produce secondary aggregation)	No details
Tan 2010174 (abstract)	LTA (no further details)	No details	AA	Unclear: blood samples collected every 2 hours up to 24 hours before and after aspirin administration
van der Loo 201190	APACT 4 aggregometer (Labitec GmbH, Ahrensburg, Germany)	Sodium citrate (3.8%)	Epinephrine (0.1 mM) Collagen (5 µg/ml) ADP (2 mM)	No details
Zanow 2010169 (abstract)	LTA (no further details)	No details	AA ADP	No details
Monotherapy at time of PFT, dual therapy during follow-up
Spectre 201193	PACKS-4	Sodium citrate	AA (1.6 mM)	No details

Study design and quality

Results of the risk-of-bias assessment can be found in Tables 6–9.

TABLE 6 - Risk of bias, patient selection (LTA, monotherapy)

Domain 1: patient selection	Was a consecutive or random sample of patients enrolled?	Was patient selection independent of patient outcomes?	Were reasons for any posteligibility exclusions provided?
Monotherapy at time of PFT and during follow-up
Abumiya 201195	No details	Yes	No details
Cha 2008121	Unclear; patients selected from a larger, consecutively enrolled group	Yes	No details
De Boni 2011159	No details	Yes	Patients who changed therapy, with low/no compliance, intolerance/allergy to aspirin, contraindications to anticoagulants, who did not attend follow-up
Feher 201188 (letter)	Consecutive	Yes	No details
Feng 2011201	Consecutive	Yes	No details
Gum 2003149	Unclear; patients recruited from consecutive patients presenting to the outpatient clinic	Yes	No details
Kempfert 2009113	Consecutive	Yes	No details
Linnemann 2009112	Consecutive	Yes	No details
Lordkipanidzé 2011162 (abstract)	No details	Yes	No details
Miyata 2011164 (abstract)	Consecutive	Yes	No details
Modica 2009187	Consecutive	Yes	No details
Ohmori 2006142	No details	Yes	No details
Payne 2004147	Consecutive	Yes	Unclear; 38/138 patients excluded before randomisation, but unclear if any would have met the inclusion criteria
Schwammenthal 2008125	Consecutive	Yes	No details
Sørensen 1983155	No details	Yes	No details
Tan 2010174 (abstract)	Consecutive	Yes	No details
van der Loo 201190	Unclear (substudy of a trial)	Yes	No details
Zanow 2010169 (abstract)	No details	Yes	No details
Monotherapy at time of PFT, dual therapy during follow-up
Spectre 201193	Consecutive	Yes	No details

TABLE 7 - Risk of bias, PFT (LTA, monotherapy)

AA, arachidonic acid; CT, closure time; ISTH, International Society of Thrombosis and Haemostasis; N/A, not applicable.

Domain 2: PFT	If a threshold was used, was it prespecified?	How was the threshold derived? (e.g. literature cut-off, based on study data)	Is the undertaking and interpretation of the index test blinded to the patient characteristics (including clinical outcomes)?
Monotherapy at time of PFT and during follow-up
Abumiya 201195	No	Quartiles	No details
Cha 2008121	Yes (> 20%)	Reference cited151	No details
De Boni 2011159	Yes (> 20%)	Reference cited149,218	No details
Feher 201188 (letter)	No details	No details	No details
Feng 2011201	No	Quartiles	No details
Gum 2003149	Yes (≥ 20% for AA and ≥ 70% for ADP)	No details	No details
Kempfert 2009113	Yes (aspirin resistant if platelet aggregation exceeded the threshold of 30% despite in vitro addition of 25 µM aspirin)	Unclear; states that platelet aggregation was measured according to the manufacturers’ instructions	No details
Linnemann 2009112	Partially; method yes, actual value no. Based on results from group of 20 healthy volunteers. Resistance defined as the maximum aggregation values within the reference range (≥ 78%) despite aspirin medication	In accordance with recommendations given at the 53rd Annual Scientific and Standardization Committee Meeting of the ISTH in Geneva in 2007, the 5th–95th percentile of maximum aggregation measured in duplicate in a group of healthy volunteers (n = 20) was considered as the reference range (i.e. 78–96%)	No details
Lordkipanidzé 2011162 (abstract)	No details	No details	No details
Miyata 2011164 (abstract)	No details	No details	No details
Modica 2009187	Yes (high residual platelet reactivity was defined as a normal CT value even when the subject was taking aspirin)	Reference cited77	Unclear (‘test results were not accessible by the attending physicians’)
Ohmori 2006142	No details	No details	Yes; laboratory staff were kept unaware of patient information
Payne 2004147	Yes (> 20%)	Reference cited219	Unclear; states that ‘all personnel involved with the trial were blinded to the nature of the patients’ current drug therapy.’ However, this may not apply to PFTs
Schwammenthal 2008125	Yes (good response < 20%, partial response 20–39%, complete unresponsiveness ≥ 40%)	Reference cited149	Unclear; treating physicians and the investigators evaluating the patients were blinded to the results of the platelet function studies
Sørensen 1983155	Yes (platelet hyperaggregability defined as secondary aggregation obtained by ADP concentration ≤ 1 µM)	Reference cited220	No details
Tan 2010174 (abstract)	Yes (> 20%)	No details	No details
van der Loo 201190	No (mean levels of platelet aggregation shown for groups with and without events)	N/A	No details
Zanow 2010169 (abstract)	Yes (≥ 30%)	No details	No details
Monotherapy at time of PFT, dual therapy during follow-up
Spectre 201193	Yes: three groups (good response < 20% aggregation, intermediate 20–40%, poor response > 40%)	No details	No details

TABLE 8 - Risk of bias, outcomes and study attrition (LTA, monotherapy)

GP, general practitioner.

Domains 3 and 4: outcomes and study attrition	Were the outcomes of interest clearly defined in advance?	Were the outcome results interpreted without knowledge of the results of the PFT?	What was the proportion of missing data? (State reasons for loss to follow-up or differences in those who completed or were lost)
Monotherapy at time of PFT and during follow-up
Abumiya 201195	Yes	No details	No details on whether or not there were any missing data
Cha 2008121	Yes	No details	No loss to follow-up
De Boni 2011159	No	No details	Unclear; patients who did not attend follow-up were excluded from study
Feher 201188 (letter)	Unclear	No details	No details
Feng 2011201	Yes	No details	No details
Gum 2003149	Yes	Those performing follow-up interviews were blinded to aspirin sensitivity status	Follow-up data were available on 97% of patients
Kempfert 2009113	No	No details	1/59 patients lost to follow-up. Reason not stated
Linnemann 2009112	Yes	Unclear; reported events were only considered if they were confirmed by medical reports from GPs or admitting hospitals	Data on clinical outcome available only from patients whose platelet function was assessed twice (57/98). Of the 98, four patients died and 16 had their antithrombotic medication changed, mainly because of an acute cardiovascular event. Not clear what the remaining reasons for dropouts were. This might bias the results though authors state that there was no difference observed in aspirin resistance rates between dropouts and those remaining in the study
Lordkipanidzé 2011162 (abstract)	Yes	No details	No details
Miyata 2011164 (abstract)	Yes	No details	No details
Modica 2009187	Yes	Yes (‘test results were not accessible by the attending physicians’)	No loss to follow-up
Ohmori 2006142	Yes	Yes; those performing follow-up were unaware of the aspirin sensitivity status	4/136 (three patients who developed atrial fibrillation and one who did not take aspirin were excluded from analysis)
Payne 2004147	Yes	Unclear; states that ‘all personnel involved with the trial were blinded to the nature of the patients’ current drug therapy.’ However, this may not apply to PFTs and outcomes	No details
Schwammenthal 2008125	Yes	Yes; treating physicians and the investigators evaluating the patients were blinded to the results of the platelet function studies	Follow-up data were available for 81/105 patients (77%)
Sørensen 1983155	No	No details	48/83 patients at last follow-up, but proportion of these in aspirin group (n = 41) unclear
Tan 2010174 (abstract)	Yes (though unclear if composite or individual outcomes)	No details	No details
van der Loo 201190	Yes	No details	Appears to be no loss to follow-up for events (though repeat PFTs in decreasing numbers of patients over time)
Zanow 2010169 (abstract)	No details	No details	No details
Monotherapy at time of PFT, dual therapy during follow-up
Spectre 201193	Yes	No details	7/63 lost to follow-up at 6 months

TABLE 9 - Risk of bias, confounders (LTA, monotherapy)

N/A, not applicable; NIHSS, National Institutes of Health Stroke Scale; TOAST, Trial of ORG 10172 in Acute Stroke Treatment.

Domain 5: confounding	Are confounders accounted for in the design or analysis (e.g. adjustment, stratification)?	If there is an adjusted outcome measure (e.g. OR, HR), what were the factors that were adjusted for?	If a HR was presented, was the proportional hazards assumption met?	Was compliance measured?	How was compliance measured?	Level of compliance
Monotherapy at time of PFT and during follow-up
Abumiya 201195	Design: N/A Analysis: no	N/A	N/A	No details	No details	No details
Cha 2008121	Design: N/A Analysis: yes (for OR)	Age (> 70 years), white blood cell count > 15 points on NIHSS, statin therapy, large artery atherosclerotic infarction based on TOAST classification	N/A	No details	No details	No details
De Boni 2011159	Design: N/A Analysis: no	N/A	N/A	Yes	Exclusion criterion: those with low/no compliance	No details
Feher 201188 (letter)	Design: N/A Analysis: no	N/A	N/A	Yes	‘Study drug compliance was assessed by their general practitioners and by personal interview’	No details
Feng 2011201	Design: N/A Analysis: no	N/A	N/A	No details	No details	No details
Gum 2003149	Design: N/A Analysis: yes (for HR)	Variables in model included age, sex, race, history of tobacco use, diabetes, hypertension, hyperlipidaemia, revascularisation, MI, haemoglobin, platelet count, creatinine, aspirin sensitivity	Yes	Yes	Patient interview both at study enrolment and follow-up	No details
Kempfert 2009113	Design: N/A Analysis: no	N/A	N/A	Yes	If platelets were not sufficiently suppressed (aggregation exceeding the 30% threshold), then the test was repeated after in vitro addition of aspirin (10 µM and 25 µM) to assess whether or not platelets exhibited resistance per se	No details on level. Stated that non-compliance or impaired intestinal uptake excluded by adding in vitro aspirin
Linnemann 2009112	Design: N/A Analysis: no	N/A	N/A	Yes	Interview at study commencement and follow-up	‘All patients confirmed that they had taken aspirin regularly as directed over the last 14 days.’ Assume relates to start of the study
Lordkipanidzé 2011162 (abstract)	Design: N/A Analysis: unclear if adjusted or unadjusted OR	No details	N/A	No details	No details	No details
Miyata 2011164 (abstract)	Design: N/A Analysis: appears that adjustment for possible confounders was undertaken, but no adjusted measures presented	No details	N/A	Yes	Interview and by checking plasma concentration of salicylic acid	No details
Modica 2009187	Design: N/A Analysis: yes (for HR)	Age, sex, diabetes, smoking status, heart failure, atrial fibrillation, baseline glomerular filtration rate, troponin T, platelet aggregation, high residual platelet reactivity, intervention with CABG or PCI	Yes (the assumptions for Cox regression analysis were evaluated by Kaplan–Meier curves for all the variables included)	No details	No details	No details
Ohmori 2006142	Design: N/A Analysis: yes (for HR)	Small aggregates, medium aggregates, large aggregates	No details	Yes	By patient interview at enrolment and follow-up	Unclear; one patient who did not take aspirin was excluded from analysis
Payne 2004147	Design: N/A Analysis: no	N/A	N/A	No details	No details	No details
Schwammenthal 2008125	Design: N/A Analysis: yes (for OR)	Age, NIHSS, diabetes	N/A	No details	No details	No details
Sørensen 1983155	Design: N/A Analysis: no	N/A	N/A	No details	No details	No details
Tan 2010174 (abstract)	Design: N/A Analysis: no	N/A	N/A	No details	No details	No details
van der Loo 201190	Design: N/A Analysis: no	N/A	N/A	No details	No details	No details (N.B. appears to dismiss non-compliance as a reason for variation on PFTs; stated that ‘the variation in platelet response to agonists would imply a high degree of irregularity in taking aspirin, which makes this explanation completely speculative.’)
Zanow 2010169 (abstract)	Design: N/A Analysis: no	N/A	N/A	No details	No details	No details
Monotherapy at time of PFT, dual therapy during follow-up
Spectre 201193	Design: N/A Analysis: yes (for HR)	‘Variables chosen for inclusion into the model were those that tended to be associated with event-free survival on univariate analysis and age’	No details	No details	No details	No details

Patient selection was independent of study outcome in all included studies, with the PFT preceding any outcomes (as specified in the study selection criteria). Ten of 19 studies88,93,112,113,125,147,164,174,187,201 stated that consecutive patients were enrolled into the study. Only one study159 had clear details on posteligibility exclusion of patients; one criterion for exclusion was no or low compliance.

A predetermined threshold percentage (for platelet aggregation) was given in nine studies; in seven of these93,121,125,147,149,159,174 the threshold was 20% (with two studies93,125 defining a further two groups: 20–39% for partial response and ≥ 40% for complete unresponsiveness). In two studies113,169 the threshold was 30%. The remaining studies stated that quartiles were used,95,201 described the method of deriving a threshold but not an actual percentage112,155,187 or gave no details. 88,142,162,164 One study90 stated mean levels of platelet aggregation only (for groups with and without clinical events). Most studies cited a reference for their threshold or method of derivation; there were no details in seven studies. 88,90,95,142,162,164,201 Only one study142 gave clear details on blinding of laboratory staff to patient characteristics.

Outcome measures of interest were clearly predefined in all but five studies. 88,113,155,159,169 Four studies125,142,149,187 had clear details regarding blinding to the PFT results of those assessing outcomes. There appeared to be no loss to follow-up in three studies. 90,121,187 Loss to follow-up was stated in seven studies93,112,113,125,142,149,155 and ranged from 2% to 57% (see Table 8). There were no clear details in nine studies. 88,95,147,159,162,164,169,174,201 The differences in completeness of follow-up may reflect length of follow-up, study design (outcome only followed up in those that had repeat PFTs) or quality of reporting.

Compliance was measured in seven studies. 88,112,113,142,149,159,164 In three studies88,142,149 this was by a general practitioner (GP) assessment and/or patient interview, but no details on the level of compliance were stated; one patient was excluded on the basis of non-compliance in one of these studies. 142 One study112 stated that after interview all patients confirmed that they had taken aspirin as directed over the last 14 days.

A further study164 assessed compliance by interview and checking of plasma concentration of salicylic acid, but there were no details on level of compliance. One study159 did not state the method of assessing compliance, but stated that patients with low/no compliance were excluded. In the study by Kempfert et al. ,113 the PFT was repeated after in vitro addition of aspirin where platelets were not sufficiently suppressed in order to exclude non-compliance.

Thirteen studies did not appear to undertake any adjusted analyses. 88,90,95,112,113,147,155,159,162,164,169,174,201 Six studies93,121,125,142,149,187 attempted to adjust for a number of factors. There was some overlap between factors adjusted for (e.g. age), but no study used all the same factors as another. There may be selective reporting in that only variables that showed significance on univariate analysis may have been included in multivariate analyses. One study142 was unusual in that it adjusted for size of aggregates only.

Overview of outcomes

The main outcome categories reported in the LTA monotherapy studies are shown in Table 10. Note that where a study reports MACEs, individual outcomes (e.g. death, stroke) may additionally have been reported separately. There may also be more than one ischaemic/thrombotic outcome reported in the same study. Follow-up periods ranged from 30 days to 3 years.

TABLE 10 - Outcomes (LTA, monotherapy)

IQR, interquartile range.

Study	Death	MACE	Ischaemic/thrombotic	Bleeding	Length of follow-up
Monotherapy at time of PFT and during follow-up
Abumiya 201195			✓		1 year
Cha 2008121	✓	✓	✓		90 days
De Boni 2011159			✓		3 months
Feher 201188 (letter)	✓				2 years
Feng 2011201				✓	Up to 6 months
Gum 2003149	✓	✓	✓		Mean 679 days
Kempfert 2009113	✓	✓	✓		12 months
Linnemann 2009112		✓	✓		Median 17 months (range 10–37 months)
Lordkipanidzé 2011162 (abstract)		✓			3 years
Miyata 2011164 (abstract)		✓			2 years
Modica 2009187		✓			Median 44 months (IQR 35–55 months)
Ohmori 2006142		✓			Mean 172 days
Payne 2004147	✓		✓		30 days
Schwammenthal 2008125			✓		Median 11.5 months (range 3.9–19.3 months)
Sørensen 1983155			✓		Median 26 months (range 20–36 months)
Tan 2010174 (abstract)	✓	✓	✓		Mean 360 days (range 0–523 days)
van der Loo 201190			✓		Mean 80 months (range 52–94 months)
Zanow 2010169 (abstract)			✓		‘Long-term’
Monotherapy at time of PFT, dual therapy during follow-up
Spectre 201193		✓			Up to 15 months

Death

Only 688,113,121,147,149,174 of the 19 studies reported this outcome, and 5113,121,147,174,188 reported solely the data needed to populate 2 × 2 tables (Table 11).

TABLE 11 - Outcome measures for reporting death (LTA, monotherapy)

Calculated from data given in the publication.

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Monotherapy at time of PFT and during follow-up
Cha 2008121	✓a		✓a		Number of events reported for three groups: low, intermediate, high ADP aggregation	✓a
Feher 201188 (letter)	✓a		✓a			✓a
Gum 2003149			✓
Kempfert 2009113	✓a		✓a			✓a
Payne 2004147					No patients found to be aspirin resistant, therefore not represented in Figures 3 and 4
Tan 2010174 (abstract)					Some percentages presented, but exact numbers not clear

Figures 3 and 4 present the unadjusted ORs and unadjusted HRs reported for death. None of these were directly available from the publications (except for one HR149), but were calculated from other reported data. No study reported adjusted measures. In the study by Payne et al. (2004),147 no patients (out of a total of 54) were found to be aspirin resistant at a threshold of > 20% aggregation, therefore no summary measures could be calculated; one stroke and no deaths occurred in this patient group (follow-up 30 days). Similarly, results could not be presented in forest plots for the study by Tan et al. ;174 here the rate of death was 26% in the resistant and 11% in the sensitive group.

FIGURE 3.

Light transmission aggregometry, monotherapy: death, unadjusted ORs. Cha et al. 121 results obtained by collapsing tertiles presented into two groups defined by the threshold level shown. AA, arachidonic acid.

FIGURE 4.

Light transmission aggregometry, monotherapy: death, unadjusted HRs. Cha et al. 121 results obtained by collapsing tertiles presented into two groups defined by the threshold level shown. AA, arachidonic acid. a, Three patients were on clopidogrel preoperatively, all patients were on aspirin monotherapy post-operatively.

In the studies by Cha et al. 121 and Gum et al. ,149 more deaths occurred in those patients categorised as aspirin resistant; however, none of the unadjusted ORs or HRs were statistically significant. This is also the case for the studies by Feher et al. 88 and Kempfert et al. ,113 where the extremely wide CIs are a reflection of an adjustment factor used in cases where no events occurred in the aspirin-sensitive groups.

In terms of prognostic utility, although there was a trend towards more events in the aspirin-resistant groups, no study was able to show a statistically significant difference (in CAD or CVD/stroke patients).

Major adverse cardiac events

Eleven90,93,112,113,121,142,149,162,164,174,187 of 19 studies reported this outcome (Table 12).

TABLE 12 - Outcome measures for reporting MACEs (LTA, monotherapy)

Calculated from data given in the publication.

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Monotherapy at time of PFT and during follow-up
Cha 2008121	✓a	✓	✓a			✓a
Gum 2003149			✓	✓
Kempfert 2009113	✓a		✓a			✓a
Linnemann 2009112	✓a		✓a			✓a
Lordkipanidzé 2011162 (abstract)	✓
Miyata 2011164 (abstract)					Narrative description
Modica 2009187			✓	✓
Ohmori 2006142				✓
Tan 2010174 (abstract)					Some percentages presented, but exact numbers not clear
van der Loo 201190					Mean platelet aggregation values for groups with and without events
Monotherapy at time of PFT, dual therapy during follow-up
aSpectre 201193	✓a		✓a	✓		✓a

Note that one study112 reporting MACEs included death from cardiovascular causes, MI, ACS and stroke, but also amputation or gangrene.

The study by Miyata et al. 164 reported that no ex vivo measurements for residual platelet function were associated with cardiovascular events (no data presented). In the study by Tan et al. 174 the results for total events (MACEs) were unclear, but based on higher numbers of events for death, recurrent MI or thrombosis, it appeared that a greater number of MACEs occurred in the aspirin-resistant group. van der Loo et al. 90 reported no significant differences in mean platelet aggregation levels in groups with and without MACEs.

The remaining studies are presented in the forest plots in Figures 5–8.

FIGURE 5.

Light transmission aggregometry, monotherapy: MACEs, unadjusted ORs. AA, arachidonic acid.

FIGURE 6.

Light transmission aggregometry, monotherapy: MACEs, adjusted ORs.

FIGURE 7.

Light transmission aggregometry, monotherapy: MACEs, unadjusted HRs. AA, arachidonic acid.

FIGURE 8.

Light transmission aggregometry, monotherapy: MACEs, adjusted HRs. AA, arachidonic acid.

There were 12 unadjusted ORs based on five studies,93,112,113,121,162 (different agonists, thresholds or time points; see Figure 5), 11 of which found no statistically significant differences in event rates; there was also no consistent trend regarding direction of effect. One study121 found a statistically significant result for one of two thresholds, with more events in the aspirin-resistant group in a CVD/stroke population, but this result was based on a non-aspirin-specific agonist (ADP) and the threshold was derived by collapsing tertiles. Most ORs were not directly available from the studies (with the exception of one162).

The one statistically significant OR121 remained so after adjustment, although it moved very close to 1 (see Figure 6). There were no further adjusted ORs.

There were 10 unadjusted HRs (based on six studies,93,112,113,121,149,187 see Figure 7), all calculated from other data presented in the articles. Three showed statistically significant results (more events in the aspirin-resistant group). 121,149,187 All three included different populations (CAD, CVD/stroke and UA/ACS).

Based on the three studies where an unadjusted HR was calculable, one HR149 remained statistically significant after adjustment (see Figure 8); two previously non-significant results93 became statistically significant and one previously significant result became non-significant. 187 Of one further study included,142 one of two results was statistically significant (more events in the aspirin-resistant group).

Based on adjusted measures, there was a consistent trend towards more MACEs in the resistant groups, with some results showing statistical significance. However, this is based on a subsample of studies only (5 of 19 studies93,121,142,149,187), and the choice of adjustment factors and inclusion of certain factors into the models may have affected results. Note that two93,187 of the five studies contributing to these results were in UA/ACS patients, which may differ from the majority of stable populations, and in one study93 all patients were on dual therapy after the PFT.

Ischaemic/thrombotic events

Thirteen studies88,90,95,112,113,121,125,147,149,155,159,169,174 reported additional ischaemic/thrombotic events (Table 13).

TABLE 13 - Outcome measures for reporting ischaemic/thrombotic events (LTA, monotherapy)

CVA, cerebrovascular accident; mRS, modified Rankin scale; PTA, percutaneous transluminal angioplasty; TEA, thoracic epidural analgesia.

Calculated from data given in the publication.

Study	Outcome	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Monotherapy at time of PFT and during follow-up
Abumiya 201195	Recurrent cerebral infarction					Results for quartiles represented graphically. Could not extract data from forest plots
Cha 2008121	Non-fatal stroke	✓a		✓a			✓
	MI	✓a		✓a			✓a
De Boni 2011159	Ischaemic events/clinical relapses					No patient classified as aspirin resistant. No events occurred
Feher 201188 (letter)	Stroke and/or TIA	✓a		✓a			✓
	UA	✓a		✓a			✓a
	MI	✓a		✓a			✓a
	Peripheral arterial occlusion with revascularisation	✓a		✓a			✓a
Gum 2003149	CVA			✓
	MI			✓
Kempfert 2009113	Stroke	✓a					✓a
Linnemann 2009112	Peripheral arterial revascularisation					Numbers in aspirin-resistant and aspirin-sensitive groups not clear
	PTA/stenting
	Bypass surgery or TEA
Payne 2004147	Stroke					No patients classified as aspirin resistant
Schwammenthal 2008125	mRS score ≥ 2 or death during follow-up		✓
	Recurrent ischaemic event
Sørensen 1983155	Stroke or death	✓a		✓a			✓a
Tan 2010174 (abstract)	Recurrent MI or thrombosis					Some percentages presented, but exact numbers not clear
van der Loo 201190	Restenosis or reocclusion					Intrapatient variability of platelet aggregation between groups with and without events
Zanow 2010169 (abstract)	Patency after reconstruction					No numerical data

Seven of 13 studies90,95,112,147,159,169,174 did not provide data which would have allowed their representation in forest plots. There was also heterogeneity across outcome measures (e.g. MI, UA, restenosis, etc.). Many of these measures are also captured in the MACEs described above.

In the study by Abumiya and Houkin95 there appeared to be a trend for more events (recurrent cerebral infarction) to occur in higher quartiles of platelet aggregation, but no statistical significance could be shown. The numbers in Tan et al. 174 were unclear, but it appears that a higher percentage had a recurrent MI or thrombosis in the aspirin-resistant group. van der Loo et al. 90 looked at differences in intrapatient variability of platelet aggregation between groups with and without restenosis or occlusion; no evidence for a difference was found (at adjusted p-value level). Zanow et al. 169 found that long term there was a poorer patency rate for aspirin-resistant patients, but this was not statistically significant.

The study by De Boni et al. 159 provided no useful information, as no patients were classified as aspirin resistant and no events occurred. Similarly, in the study by Payne et al. 147 no patients were classified as aspirin resistant; there was one stroke (in a group of 54 patients). Linnemann et al. assessed a number of ischaemic thrombotic outcomes, but the exact numbers in the aspirin-resistant and aspirin-sensitive groups were unclear; only 2 out of 57 patients were classified as aspirin resistant.

Twelve unadjusted ORs were presented based on four studies88,113,121,155 (different outcomes, thresholds) (Figure 9). All were calculated for this report. CIs were generally very wide and all but one showed no statistical significance, though there was a trend towards more events occurring in the aspirin-resistant groups.

FIGURE 9.

Light transmission aggregometry, monotherapy: ischaemic/thrombotic events, unadjusted ORs. AA, arachidonic acid; PAO, peripheral arterial occlusion.

A different study125 presented an adjusted OR for two thresholds, one of which was statistically significant (Figure 10).

FIGURE 10.

Light transmission aggregometry, monotherapy: ischaemic/thrombotic events, adjusted ORs. AA, arachidonic acid; mRS, modified Rankin scale.

Fourteen unadjusted HRs based on five studies88,113,121,149,155 were presented (Figure 11). Again, all but one showed no statistical significance, though there was a trend towards more events in the aspirin-resistant groups. No adjusted HRs were presented.

FIGURE 11.

Light transmission aggregometry, monotherapy: ischaemic/thrombotic events, unadjusted HRs. AA, arachidonic acid; CVA, cerebrovascular accident; PAO, peripheral arterial occlusion.

Overall, the trend towards more events in the aspirin-resistant group was consistent, but most results were not statistically significant. These results were also consistent with those of the studies not presented in the forest plots. Interestingly, there were three studies where no (or a very small proportion of) aspirin-resistant patients were identified. 112,147,159

Bleeding events

Only one study201 reported bleeding events (GI bleeds) (Table 14).

TABLE 14 - Outcome measures for reporting bleeding (LTA, monotherapy)

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Monotherapy at time of PFT and during follow-up
Feng 2011201					Aggregation values presented for individual patients (according to quartiles for some)

Of the four patients with bleeds, three were in the lowest platelet aggregation quartile (values of 2.7%, 6.75% and 9.12%). The threshold value for the lowest quartile was 9.81%. The remaining patient had a value of 11.2% (not stated which quartile). This is consistent with the assumption that GI bleeds are more likely to occur in aspirin-sensitive patients, but the small number of events precludes any firm conclusions.

Summary: light transmission aggregometry monotherapy

Nineteen studies were identified in this category. 88,90,93,95,112,113,121,125,142,147,149,155,159,162,164,169,174,187,201 There were differences in patient populations, though most appeared to have stable disease; note that although there were only three studies93,174,187 with acute (UA/ACS) populations, two of these93,187 contribute substantially to the MACE results. There was heterogeneity across studies in terms of specific patient characteristics (e.g. smoker, diabetic, comedications, etc.).

There was a lack of detail in reporting of relevant quality criteria, making an overall judgement on risk of bias difficult. Additionally, studies that do report relevant information may be more open to criticism. Lack of detail related in particular to loss-to-follow-up information, blinding and details of compliance. No study provided details on all relevant quality criteria. There were differences in threshold and method of deriving the threshold for defining aspirin resistance, but the most consistent was a threshold of 20% (seven studies93,121,125,147,149,159,174). Only one142 and four125,142,149,187 (of 19) studies respectively gave clear details on blinding to patient characteristics or PFT results. Measurement of compliance was undertaken in seven studies,88,112,113,142,149,159,164 but there was a lack of detail on the results or consequences of this; it appears that in two studies142,159 patients were excluded on the basis of low/no compliance. Some studies provided adjusted analyses; there was overlap but no consistency in factors adjusted for.

Six studies88,113,121,147,149,174 reported on differences in deaths between aspirin-resistant and aspirin-sensitive groups; there was a trend towards more events in the aspirin-resistant group (based on four studies88,113,121,149), but no significant differences were shown in any.

Eleven studies reported on MACEs. 90,93,112,113,121,142,149,162,164,174,187 There was a trend towards more events in the aspirin-resistant groups, but unadjusted measures found mostly statistically non-significant results. Five of seven results (based on five studies93,121,142,149,187) using adjusted measures were statistically significant and the trend was consistent with the unadjusted results; however, this was based on a subset of studies, different adjustment factors and two93,187 (of five93,121,142,149,187) studies in an acute population, which may not be representative of the majority of populations receiving aspirin monotherapy.

Thirteen studies reported additional ischaemic/thrombotic events. 88,90,95,112,113,121,125,147,149,155,159,169,174 Again, there was a trend towards more events in the aspirin-resistant group, but the vast majority of results (mainly unadjusted measures) were not statistically significant. Results of 7 of the 13 studies90,95,112,147,159,169,174 could not be presented in forest plots, but results were consistent (i.e. non-significant).

There was only one study reporting GI bleeds;201 this found a trend for more aspirin-sensitive patients to have more bleeds, but this was based on only four events (in 136 patients).

Note that not all studies reporting the relevant outcomes could be presented in the forest plots; the results of those studies not included in the forest plots were in the main consistent with those included or did not add much useful information. It should also be noted that some studies contributed several results to the forest plots as they presented results or could be analysed for different thresholds. Although no results have been pooled, the visual impact of these forest plots might influence how the overall results are perceived. Given the large amount of heterogeneity between the studies in terms of quality criteria, threshold, population, test characteristics (agonists), aspirin dose, etc., it was not possible to compare results across studies. Despite the heterogeneity and lack of many statistically significant results, the direction of prognostic effect appears to be largely consistent with there being more events in aspirin-resistant patients (ORs and HRs usually > 1). This suggests that LTA is a potential prognostic factor, but this is only a qualitative judgement on the evidence available; meta-analysis was not possible owing to the heterogeneity, and therefore a firm quantitative conclusion regarding whether or not LTA is prognostic is not currently possible.

Summary: light transmission aggregometry

• Nineteen studies were identified with mainly stable populations.

• The most frequently reported threshold was 20% platelet aggregation.

• A lack of detail in reporting of quality criteria, particularly around loss to follow-up, blinding and details (and implications) of compliance, hampered an overall risk-of-bias assessment.

• Heterogeneity in outcomes, patient groups and types of reported statistics meant that meta-analysis was not considered appropriate.

• Adjusted results were rarely presented, and thus the additional prognostic value of the test over other prognostic factors is difficult to ascertain.

• Despite clinical heterogeneity between studies, there was an overall consistent trend for more events to occur in the ‘aspirin-resistant’ group for all relevant outcomes (death, MACEs, ischaemic/thrombotic events); however, most results were not statistically significant.

• There were more statistically significant results (more events in the resistant arm) using adjusted measures for MACEs, but these were based on only five studies.

• One study reporting GI bleeds found a trend for more GI bleeds in ‘aspirin-sensitive’ patients, but this was based on only four events (in 136 patients).

VerifyNow^® Aspirin

Population and test characteristics

Seven studies86,92,99,105,133,162,171 were identified in this category, one162 of which was reported in abstract form only. Populations were mainly classified as having CAD (three studies)99,133,162 or CVD/stroke (two studies). 86,171 One study105 was in patients with UA/ACS, and one92 was in patients with severe CAD undergoing CABG. Six studies did not report for how long patients had had their underlying condition; one study171 stated time from cerebral infarction to randomisation [≤ 90 days: 37 patients (31.1%); 91–364 days: 33 patients (27.7%); ≥ 365 days: 49 patients (41.2%)].

In six studies86,99,105,133,162,171 it appeared that patients were exclusively on monotherapy both at the time of the PFT and at follow-up. In the remaining study,92 patients were on monotherapy at the time of the PFT, and on dual therapy [+ ticlopidine (Ticlid^®, Sanofi Winthrop)] during follow-up.

Comedications across the studies included, where reported, ACE inhibitors, beta-blockers, angiotensin receptor blockers, calcium channel blockers, statins, oral anticoagulants and lipid lowering agents. NSAIDs were not permitted in four studies86,92,133,171 and were taken by 28 out of 314 (9%) patients in another,105 while one study99 stated that ‘concurrent nonsteroidal anti-inflammatory drug use did not correlate with the presence of aspirin non-responsiveness defined by this method at either time point’. There were no details in one study. 162

The number of participants in the studies ranged from 106 to 468 (see Table 15). Mean ages were mainly reported by group (resistant/sensitive) and ranged from 61 years to 70 years. Overall, there were more men than women in the studies, with proportions of men ranging from 50% to 85%. All studies were conducted in hospital settings. The proportion of smokers ranged (where reported) from 11% to 39% and that of diabetics from 21% to 56%.

The dose of aspirin ranged from 75 to 325 mg/day. Four studies86,105,133,171 noted a minimum period (between 1 and 4 weeks) for which patients needed to have been taking aspirin; there were no details in the remaining studies. Two studies99,133 stated that aspirin was provided in enteric form (in 65% of patients in one study133). There were no details in the remaining studies.

The main study characteristics are listed in Table 15. Note that in some studies baseline characteristics have been reported only according to resistant/sensitive groups, rather than for the total study population. All studies used the commercially available VerifyNow^® Aspirin test kit (Table 16), which uses arachidonic acid as an agonist. Four studies noted the timing of the PFT after aspirin ingestion; this was between 1 and 4 hours,86 between 2 and 30 hours,133 up to 24 hours171 or on the same day. 92 There were no details in the remaining studies.

TABLE 15 - Population characteristics (VerifyNow® Aspirin, monotherapy)

ARU, aspirin reaction unit; SD, standard deviation; SVG, saphenous vein graft.

Study/country	Number of patients	Age (years)	Therapy	Main underlying condition	Selected other population details	Due to undergo vascular intervention?	Aspirin dose/frequency	Duration of aspirin therapy (prior to PFT)	Percentage aspirin resistant	Derivation of threshold/comment
Monotherapy at time of PFT and during follow-up
Chen 2007,133 China	468	Mean 66.7 (SD 10.2) resistant, mean 63.2 (SD 11.7) sensitive	Mono	CAD	Smokers: Resistant (N = 128): n = 14 (10.9%) Sensitive (N = 340): n = 44 (14.1%) Diabetes: Resistant (N = 129): n = 50 (39.1%) Sensitive (N = 340): n = 117 (34.4%)	No	80–325 mg/day Mean 114.4 mg (SD 42.2 mg) sensitive, mean 102.0 mg (SD 17.8 mg) resistant	At least 4 weeks	27.4	Resistance defined as ARU ≥ 550. No details for derivation of threshold
Chu 2010,105 New Zealand	314	Mean 67.1 (SD 12.75) resistant, mean 70.1 (SD 11.8) sensitive	Mono (11% dual during follow-up – people who underwent PCI)	UA/ACS	Smokers: n = 7 (23.3%) resistant, n = 38 (13.4%) sensitive Diabetes: n = 10 (33.3%) resistant, n = 60 (21.1%) sensitive	PCI in 11% CABG in 2.9%	75–300 mg/day	At least 4 weeks	9.6	Resistance defined as ARU ≥ 550. No details for derivation of threshold
Gluckman 2011,99 USA	229	For patients with ≥ 1 occluded SVG (n = 70): mean 63 (range 55–72) For patients with patent SVG (n = 159): mean 63 (range 57–71)	Mono	CAD	Smokers: n = 52 (22.7%) Diabetes: n = 84 (36.7%)	Yes: CABG	325 mg/day	No details	No details	Resistance defined as ARU ≥ 550. Derivation according to manufacturer’s instructions
Lee 2010,171 The Republic of Korea	119 (244 enrolled; split into mono and dual therapy) All population characteristics based on 119 (mono group)	Mean 62.8 (SD 10.0)	Mono	CVD/stroke	Smokers: n = 44 (37%) Diabetes: n = 46 (38.7%)	No	100 mg/day	At least 2 weeks	10.9	Resistance defined as ARU ≥ 550
Lordkipanidzé 2011,162 UK (abstract)	198	No details	Mono	CAD	No details	No	80–325 mg/day	No details	No details	No details
Ozben 2011,86 Turkey	106	Mean 64.9 (SD 14.6)	Mono	CVD/stroke	Smokers: n = 41 (38.7%) Diabetes: n = 36 (34%)	No	100 mg/day	At least 1 week	33	Resistance defined as ARU ≥ 550. Derivation according to manufacturer’s clinical studies using optical aggregometry as the comparison standard
Monotherapy at time of PFT, dual therapy during follow-up
Kim 2011,92 The Republic of Korea	220	Mean 65.1 (SD 10.8) resistant, mean 63.9 (SD 8.8) sensitive	Mono (all patients on dual during follow-up)	MI (CABG)	Smokers: Aspirin responders (N = 181): n = 58 (32%) Aspirin non-responders (N = 39): n = 7 (17.9%) Diabetes: Aspirin responders (N = 181): n = 89 (49.2%) Aspirin non-responders (N = 39): n = 22 (56.4%)	Yes: CABG	100 mg/day	No details	17.7	Resistance defined as ARU ≥ 550. Derivation manufacturer’s reference value

TABLE 16 - Test characteristics (VerifyNow® Aspirin, monotherapy)

AA, arachidonic acid.

Study	Details of kit	Anticoagulant (concentration)	Agonist (concentration)	Time since last aspirin dose
Monotherapy at time of PFT and during follow-up
Chen 2007133	VerifyNow® Aspirin	No details	AA	Between 2 and 30 hours
Chu 2010105	VerifyNow® Aspirin	3.2% citrate	AA	No details
Gluckman 201199	VerifyNow® Aspirin	No details	AA	No details
Lee 2010171	VerifyNow® Aspirin	No details	AA	Up to 24 hours
Lordkipanidzé 2011162 (abstract)	VerifyNow® Aspirin	No details	No details	No details
Ozben 201186	VerifyNow® Aspirin	3.2% citrate	AA	Between 1 and 4 hours
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	VerifyNow® Aspirin	No details	AA	Aspirin administered on day of test

Study design and quality

Patient selection was independent of outcome in all studies, as all patients with an available PFT were followed up. Two studies99,171 stated that consecutive patients were enrolled. Two studies92,99 provided details on posteligibility exclusions; one of these99 reported that the study population was deemed to be representative of the eligible population.

As this was a commercial test with a manufacturer-recommended threshold, it was assumed that all studies used the same threshold even where not stated. No study gave clear details on whether or not the undertaking and interpretation of the PFT was blinded to patient characteristics. Outcomes were defined in advance in all studies, and there were details in four studies86,99,105,133 regarding the blinding of outcome assessment (to the results of the PFT). Proportions of missing data were reported in four studies and were less than, or around, 1%105,133 and up to 14%171 and 32%. 99 Longer follow-up times did not correspond to greater loss to follow-up.

It was not stated whether or not the proportional hazards assumption was met in the two studies105,133 that reported HRs. Three studies86,105,133 reported adjusted measures and the factors adjusted for were listed; there was little similarity between the adjustment factors. Four studies86,99,105,171 stated that compliance was assessed (pill counts; ascertained by nurse, verified with patients). Only one study171 gave details on the actual level of compliance: six patients were excluded at 4 weeks and a further six at 6 months owing to poor drug compliance (12/119 in total).

Full details are provided in Tables 17–20.

TABLE 17 - Risk of bias, patient selection (VerifyNow® Aspirin, monotherapy)

SVG, saphenous vein graft.

Domain 1: patient selection	Was a consecutive or random sample of patients enrolled?	Was patient selection independent of patient outcomes?	Were reasons for any posteligibility exclusions provided?
Monotherapy at time of PFT and during follow-up
Chen 2007133	No details	Yes	No details
Chu 2010105	No details	Yes	No details
Gluckman 201199	No details	Yes	Patients where SVG patency not assessed or those not on aspirin monotherapy. Authors stated that the study population was representative of patients undergoing isolated CABG surgery based on comparison with the Society of Thoracic Surgeons National Database
Lee 2010171	Consecutive	Yes	No details
Lordkipanidzé 2011162 (abstract)	No details	Yes	No details
Ozben 201186	Consecutive	Yes	No details
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	No details	Yes	300 patients assessed for eligibility; 75 did not meet inclusion criteria, five declined to participate

TABLE 18 - Risk of bias, PFT (VerifyNow® Aspirin, monotherapy)

ARU, aspirin reaction unit.

Domain 2: PFT	If a threshold was used, was it prespecified?	How was the threshold derived (e.g. literature cut-off, based on study data)?	Is the undertaking and interpretation of the index test blinded to the patient characteristics (including clinical outcomes)?
Monotherapy at time of PFT and during follow-up
Chen 2007133	Yes (ARU ≥ 550)	No details	No details
Chu 2010105	Yes (ARU ≥ 550)	No details	Unclear: the clinical team managing the patients was blinded to aspirin resistance status
Gluckman 201199	Yes (ARU ≥ 550)	Manufacturer’s instructions	No details
Lee 2010171	Yes (ARU ≥ 550)	Manufacturer’s instructions	No details
Lordkipanidzé 2011162 (abstract)	No details	No details	No details
Ozben 201186	Yes (ARU ≥ 550)	Manufacturer’s instructions	No details
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	Yes (ARU ≥ 550)	Manufacturer’s instructions	No details

TABLE 19 - Risk of bias, outcomes and study attrition (VerifyNow® Aspirin, monotherapy)

Domains 3 and 4: outcomes and study attrition	Were the outcomes of interest clearly defined in advance?	Were the outcome results interpreted without knowledge of the results of the PFT?	What was the proportion of missing data? (State reasons for loss to follow-up or differences in those who completed or were lost)
Monotherapy at time of PFT and during follow-up
Chen 2007133	Yes	Yes. Personnel responsible for data collection were not aware of aspirin responsiveness results. Hospital charts were analysed to ascertain whether or not the events qualified for the definition of the end point	4/464 (0.9%) patients lost to follow-up, one in the aspirin-resistant group, three in the aspirin-sensitive group
Chu 2010105	Yes	Yes. The clinical team managing the patients was blinded to aspirin resistance status	2/312 lost to follow-up (death during index hospitalisation)
Gluckman 201199	Yes	Yes. Images were analysed by two blinded reviewers (98% concordance) with a third reviewer adjudicating as necessary	65/229 not included at 6 months
Lee 2010171	Yes	No details	17/119 lost to follow-up [reasons: consent withdrawal (4), poor drug compliance (12), miscellaneous (1)]
Lordkipanidzé 2011162 (abstract)	Yes	No details	No details
Ozben 201186	Yes	Yes. Personnel responsible for data collection were not aware of aspirin responsiveness results	No details
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	Yes	No details	No details

TABLE 20 - Risk of bias, confounders (VerifyNow® Aspirin, monotherapy)

N/A, not applicable; NIHSS, National Institutes of Health Stroke Scale.

Domain 5: confounding	Are confounders accounted for in the design or analysis (e.g. adjustment, stratification)?	If there is an adjusted outcome measure (e.g. OR, HR), what were the factors that were adjusted for?	If a HR was presented, was the proportional hazards assumption met?	Was compliance measured?	How was compliance measured?	Level of compliance
Monotherapy at time of PFT and during follow-up
Chen 2007133	Design: N/A Analysis: yes (for HR)	Diabetes, prior MI, haemoglobin levels	No details	No (‘Compliance to aspirin therapy was not ascertained by pill count or salicylate level monitoring’)	N/A	N/A
Chu 2010105	Design: N/A Analysis: yes (for HR)	Troponin-T (other non-significant factors were removed from model)	No details	Yes (before enrolment)	Aspirin dose and compliance were verified with patients before enrolment	No details
Gluckman 201199	Design: N/A Analysis: no	N/A	N/A	Yes	Pill counts at each postoperative encounter	No details
Lee 2010171	Design: N/A Analysis: no	N/A	N/A	Yes	Counted all returned medications after 4 weeks of treatment and calculated compliance to the trial medications. Patients with poor compliance were defined as those who missed ≥ 25% of 4-week prescribed doses	12/119 noncompliant: six did not undergo follow-up PFT after 4 weeks of treatment owing to poor drug compliance; six excluded after having the follow-up PFT at 4 weeks owing to poor drug compliance
Lordkipanidzé 2011162 (abstract)	Design: N/A Analysis: unclear whether adjusted or unadjusted OR	No details	N/A	No details	N/A	N/A
Ozben 201186	Design: N/A Analysis: yes (for OR)	Age, sex, NIHSS, prior stroke, comorbidities (hypertension, diabetes, hyperlipidaemia, coronary heart disease, renal failure)	N/A	Yes	Ascertained by nurse charts	No details
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	Design: N/A Analysis: no	N/A	N/A	No details	N/A	N/A

Overview of outcomes

The most frequently reported outcome in studies using VerifyNow^® Aspirin was MACEs (four studies92,105,133,162), followed by other ischaemic/thrombotic events (three studies92,99,133), death (two studies86,92) and bleeding events (two studies92,171). Outcomes and follow-up periods are shown in Table 21.

TABLE 21 - Outcomes (VerifyNow® Aspirin, monotherapy)

SD, standard deviation.

Study	Death	MACEs	Ischaemic/thrombotic events	Bleeding	Length of follow-up
Monotherapy at time of PFT and during follow-up
Chen 2007133		✓	✓		Mean 379 (SD 200) days
Chu 2010105		✓			> 30 days and up to 6 months
Gluckman 201199			✓		6 months
Lee 2010171				✓	4 weeks
Lordkipanidzé 2011162 (abstract)		✓			3 years
Ozben 201186	✓				2 years
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	✓	✓	✓	✓	Responders: mean 9.8 (SD 10.5) days; non-responders: mean 10.1 (SD 10.8) days

Death

Two studies86,92 reported this outcome (Table 22).

TABLE 22 - Outcome measures for reporting death (VerifyNow® Aspirin, monotherapy)

Calculated from data given in the publication.

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Monotherapy at time of PFT and during follow-up
Ozben 201186	✓a	✓	✓a			✓a
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	✓a		✓a			✓a

Only two86,92 of seven studies reported deaths and the different outcome statistics are shown in Figures 12–14. One study (Ozben et al. 86 in patients with CVD/stroke) found a statistically significant OR and HR (greater number of deaths in the aspirin-resistant group) for both in-hospital and 2-year mortality. The OR (2-year mortality) remained statistically significant when adjusted for age, sex, National Institutes of Health Stroke Scale (NIHSS), prior stroke and comorbidities (hypertension, diabetes, hyperlipidaemia, coronary heart disease, renal failure). Only one death (in the aspirin-sensitive group) occurred in the study by Kim et al. 92 in patients undergoing CABG, and no significant difference could be shown.

FIGURE 12.

VerifyNow^® Aspirin: death, unadjusted OR.

FIGURE 13.

VerifyNow^® Aspirin, monotherapy: death, adjusted ORs.

FIGURE 14.

VerifyNow^® Aspirin, monotherapy: death, unadjusted HRs.

Major adverse cardiac events

Four studies92,105,133,162 reported this outcome (Table 23).

TABLE 23 - Outcome measures for reporting MACEs (VerifyNow® Aspirin, monotherapy)

Calculated from data given in the publication.

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Monotherapy at time of PFT and during follow-up
Chen 2007133	✓a		✓	✓		✓a
Chu 2010105	✓a		✓/✓a	✓		✓a
Lordkipanidzé 2011162 (abstract)	✓
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	✓a		✓a			✓a

Four92,105,133,162 out of seven studies reported this outcome, and the different outcome statistics are shown in Figures 15–17. The studies by Chen et al. 133 and Chu et al. (for five out of seven subgroups)105 found a statistically significant difference between groups, with more events in the resistant group (unadjusted OR). Two further studies92,162 reported more events in the sensitive group, but there were no statistically significant differences. No study reported adjusted ORs. The pattern was similar for unadjusted HRs, though with statistically significant results for five out of seven subgroups (Chu et al. 105) and a statistically significant result, with more events in the resistant group, also presented for the population undergoing CABG in the study by Chen et al. 133 Note that the unadjusted HR is not statistically significant compared with the unadjusted OR (Chen et al. 133 CAD population); given the relatively long follow-up period (mean 379 days), the HR could be considered the more useful outcome statistic. Adjusted HRs were available for three out of seven subgroups (Chu et al. 105 and Chen et al. 133 CAD population); these were all statistically significant, with more events in the resistant group. Note that the factors adjusted for in the two studies are completely different (troponin-T only in the study by Chu et al. ;105 diabetes, prior MI and haemoglobin levels in the study by Chen et al. 133).

FIGURE 15.

VerifyNow^® Aspirin, monotherapy: MACEs, unadjusted ORs.

FIGURE 16.

VerifyNow^® Aspirin, monotherapy: MACEs, unadjusted HRs.

FIGURE 17.

VerifyNow^® Aspirin, monotherapy: MACEs, adjusted HRs.

Thus, any statistically significant results relate to a greater number of events in the resistant group; however, not all outcome statistics (particularly adjusted HR) have been reported for all four studies/subgroups, so there is some missing information.

Ischaemic/thrombotic events

Three studies92,99,133 reported this outcome (Table 24).

TABLE 24 - Outcome measures for reporting ischaemic/thrombotic events (VerifyNow® Aspirin, monotherapy)

ARU, aspirin reaction unit.

Calculated from data given in the publication.

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Monotherapy at time of PFT and during follow-up
Chen 2007133			✓
Gluckman 201199	✓				Mean ARU presented for groups with and without events
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	✓a		✓a			✓a

Three92,99,133 out of seven studies also reported additional ischaemic/thrombotic outcomes, and the different outcome statistics are reported in Figures 18 and 19. There were no statistically significant differences based on unadjusted ORs (two studies92,99). Two of seven unadjusted HRs were statistically significant (both based on one study133), with more events in the resistant group.

FIGURE 18.

VerifyNow^® Aspirin, monotherapy: ischaemic/thrombotic events, unadjusted ORs. SVG, saphenous vein graft.

FIGURE 19.

VerifyNow^® Aspirin, monotherapy: ischaemic/thrombotic events, unadjusted HRs.

The study by Gluckman et al. 99 also reported mean [standard deviation (SD)] values of aspirin reaction units (ARUs) for groups with one or more occluded saphenous vein grafts (SVGs) versus the group with no occluded SVG, and also for patients undergoing CABG. The mean values in the group with occluded SVGs were slightly higher (indicating greater platelet reactivity), but there were no significant differences and all means were below a threshold of 550 ARUs. No adjusted statistics were reported for any studies.

Thus, there is no evidence of a greater number of events in one or the other group in the two studies with CABG populations,92,99 while the only study113 with the CAD population found significant differences for two outcomes.

Bleeding events

Two studies92,171 reported this outcome (Table 25).

TABLE 25 - Outcome measures for reporting bleeding events (VerifyNow® Aspirin, monotherapy)

Calculated from data given in the publication.

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Monotherapy at time of PFT and during follow-up
Lee 2010171	✓a		✓a			0 events, therefore not calculable
Monotherapy at time of PFT, dual therapy during follow-up
Kim 201192	✓a		✓a			✓

Only two92,171 out of seven studies reported bleeding events; these were postoperative in one study92 and over a 4-week period in ischaemic stroke patients [randomised to aspirin and placebo or aspirin and cilostazol (Pletal^®, Otsuka) in the other study]. 171 The study by Kim et al. 92 also measured postoperative blood loss and transfused units of blood; there were no significant differences between the aspirin-resistant and sensitive groups. Too few events occurred to draw any overall conclusions: none in Lee et al. ,171 and two (re-exploration for bleeding) in Kim et al. 92 (Figures 20 and 21). No studies using VerifyNow^® Aspirin as a PFT were identified that measured long-term adverse bleeding events.

FIGURE 20.

VerifyNow^® Aspirin, monotherapy: bleeding events, unadjusted ORs. ITT, intention to treat; PP, per protocol.

FIGURE 21.

VerifyNow^® Aspirin, monotherapy bleeding events, unadjusted HRs. ITT, intention to treat.

Summary: VerifyNow^® Aspirin

Seven studies86,92,99,105,133,162,171 were identified in this category, most in stable populations, but one in patients with UA/ACS105 and one92 in patients with severe CAD undergoing CABG.

There was a lack of reporting of quality criteria and no study reported all details considered to be important to assess risk of bias. No study reported on blinding to patient characteristics (when undertaking the PFT). Only one study171 gave details on the level of compliance and exclusions on the basis of this. Four99,105,133,171 of the seven studies gave details of missing data and four86,99,105,133 gave details of blinding of outcome assessors.

The risk of death in the resistant and sensitive groups was reported in only two studies. 86,92 In one of these,92 only one death occurred. The other study found statistically significant results based on unadjusted and adjusted ORs, and unadjusted HR (more events in the resistant group); this was based on 43 events in 106 patients.

Major adverse cardiac events were reported in four studies. 92,105,133,162 The direction of effect was consistent across all results (more events in the resistant group). Around half of the unadjusted ORs and unadjusted HRs were statistically significant, but it should be noted that a single study105 contributed to a large proportion of these results as several subgroup results were presented. Adjusted HRs based on two studies105,133 were also statistically significant.

Ischaemic/thrombotic events were reported in three studies;92,99,133 most unadjusted outcome measures were statistically non-significant, though there was a trend towards more events in resistant groups. There were no adjusted outcome measures.

Two studies92,171 measured (short-term) bleeding events (postoperative or re-exploration for bleeding). There were only two events in total and no conclusion can be drawn from the data.

Despite the heterogeneity and lack of many statistically significant results, the direction of prognostic effect appears to be largely consistent with more events in aspirin-resistant patients (ORs and HRs usually > 1). This suggests that VerifyNow^® Aspirin is a potential prognostic factor, but this is only a qualitative judgement on the evidence available; meta-analysis was not possible owing to the heterogeneity, and therefore a firm quantitative conclusion regarding whether or not VerifyNow^® Aspirin is prognostic is not currently possible.

Summary: VerifyNow^® Aspirin

• Seven studies used this commercial PFT.

• A lack of reporting of quality criteria hampered an overall assessment of risk of bias; only one of seven studies gave details on level of compliance.

• Heterogeneity in outcomes, patient groups and types of reported statistics meant that meta-analysis was not considered appropriate.

• Adjusted results were rarely presented, and thus the additional prognostic value of the test over other prognostic factors is difficult to ascertain.

• There was a consistent trend towards a greater number of events in the resistant groups within the studies; some of the results were statistically significant.

• Some studies contributed more results by reporting on several subgroups and not all studies contributed to all outcome measures; therefore, there are potentially some missing data and/or a bias towards certain studies (though results were not pooled).

• No studies were identified that reported on long-term bleeding events.

Thromboxane metabolite measurement

Population and test characteristics

Eleven studies46,76,99,108,110,148,151,162,164,195,202 were identified in this category, three of which were reported in abstract form only. 162,164,202 Populations had CAD (nine studies46,76,99,108,110,162,164,195,202) or CVD disease/stroke (one study148). One study151 included patients with various conditions including CAD, stroke, PVD and diabetes.

Most studies did not report for how long patients had had their primary underlying condition. One study117 reported that patients had their primary underlying condition for a mean period of 41.4 months.

In nine studies46,99,108,110,148,151,162,164,202 it appeared that patients were exclusively on monotherapy both at the time of the PFT and during follow-up. In two studies, patients were on monotherapy at the time of the PFT, and 32%76 and 54.8%195 of patients respectively went on to additionally receive clopidogrel at some point during follow-up. It is possible that not all studies have reported where a proportion of patients commenced additional therapies during follow-up.

Three studies measured thromboxane metabolite levels in serum/plasma46,76,164 and nine studies measured thromboxane metabolite levels in urine. 99,108,110,148,151,162,164,195,202 Data in these groups were analysed separately.

Comedications were reported in five studies108,148,162,164,202 and included ACE inhibitors, angiotensin II antagonists, calcium blockers, statins, beta-blockers, COX-2 antagonists, heparin, warfarin, diuretics, insulin, oral hypoglycaemics, antidepressants, anticoagulants, lipid-lowering agents and vitamin E. NSAIDs were not permitted (or had to be discontinued within a certain time period) in two studies. 108,110 One study99 stated that ‘concurrent nonsteroidal anti-inflammatory drug use did not correlate with the presence of aspirin non-responsiveness defined by this method at either time point’. In two studies, 10%76 and 24%195 of patients respectively were taking NSAIDs. There were no details on NSAIDs in the remaining studies.

The number of participants in the studies ranged from 61 to 3261 (see Table 26). Where reported, average ages of patients ranged from 53 years (mean value) to 69 years (median value), with most average ages around the early 60s. There were more men than women in the eight studies that reported this46,76,99,108,110,148,151,195 (the remaining three studies162,164,202 did not report details), with proportions of men ranging from 59% to 90%. The proportion of patients with diabetes ranged from 19% to 48%, and that of smokers from 16.6% to 71% (where reported, see Table 26). All studies were conducted in hospital settings.

The dose of aspirin ranged between 75 mg/day and 325 mg/day, with the exception of one study,148 where the dose was high at 650 mg/day. This study included patients with a non-cardioembolic, non-incapacitating cerebral infarction. There were no details on dose in one study. 151 Details were variable across studies regarding the length of time patients had been receiving aspirin therapy, with some noting a minimum period and some giving no details (see Table 26). Two studies stated aspirin was provided in enteric or plain form99,110 and no other studies provided this information.

The main study characteristics are listed in Table 26. Note that in some studies baseline characteristics have been reported only according to groups with/without adverse clinical events, or groups with occluded or patent SVG during CABG surgery, rather than for the total study population.

TABLE 26 - Population characteristics (thromboxane metabolite measurement, monotherapy)

SEM, standard error of the mean.

Study/country	Number of patients	Age (years)	Therapy	Main underlying condition	Selected other population details	Due to undergo vascular intervention?	Aspirin dose/frequency	Duration of aspirin therapy (prior to PFT)	Percentage aspirin resistant	Derivation of threshold/comment
Serum/plasma and urinary
Monotherapy at time of PFT and during follow-up
Miyata 2011,164 Japan (abstract)	592 (583 eligible for analyses)	No details	Mono	CAD	No details	No	Some higher doses (> 100 mg/day), some lower doses (≤ 100 mg/day)	No details	No details	No details
Serum/plasma (metabolite used: TxB2)
Monotherapy at time of PFT and during follow-up
Cotter 2004,46 Israel	73 (61 eligible for analyses)	Mean 53 (SD 8)	Mono	CAD	Smokers: 23% Diabetes: 19%	No	100 mg/day	At least 1 month	14.8	5 nmol/1011 platelets (lowest TxB2 production observed in healthy volunteers)
Frelinger 2009,76 USA	700 (555 eligible for analyses)	Mean 60.7 (SEM 0.44)	Mono (32% dual during follow-up)	CAD	Smokers: 22% current smokers (72% prior smokers) Diabetes: 27%	No	81 or 325 mg/day	At least 3 days	8.1	Resistant if ≤ 3.1 ng/ml
Urinary (metabolite used: 11-dehydro-TxB2)
Monotherapy at time of PFT and during follow-up
Addad 2010,108 Tunisia	204	Mean 59 (SD 10)	Mono	CAD	Smokers: n = 145 (71%) Diabetes: n = 79 (38.7%)	No	250 mg/day	At least 2 months on enrolment	No details	No details
Bruno 2004,148 USA	61 (83 in total because 22 patients on ticlopidine; demographics based on 61 patients)	Mean 61 (SD 11)	Mono	CVD/stroke	Smokers: n = 37 (45%) Diabetes: n = 40 (48%)	No	650 mg/day	At least 7 days	No details	No details
Eikelboom 2002,151 Australia	488 (976 enrolled: 488 cases, 488 controls)	Mean 67.3 (SD 7.2)	Mono	Miscellaneous	Smokers: n = 81 (16.6%) Diabetes: n = 159 (32.6%)	No	No details	No details	No details	No details
Eikelboom 2008,195 Australia	3261	With an event (stroke, MI or cardiovascular death; n = 144): median 69 Without an event (n = 3117): median 64	Mono (54.8% dual during follow-up)	CAD	Current smokers: With an event: n = 26 (18.1%) Without an event: n = 632 (20.3%) Former smokers: With an event: n = 76 (52.8%) Without an event: n = 1547 (49.6%) Diabetes: With an event: n = 73 (50.7%) Without an event: n = 1381 (44.3%)	No	81 mg/day (median) Range 75–162 mg/day	At least one month	No details	No details
Eskandarian 2011,202 Iran (abstract)	124	No details	Mono	CAD	No details	No	80 mg/day	At least 7 days before test	65.3	Resistant if > 298 pg/mg Intermediate response 134–298 pg/mg Sensitive if < 134 pg/mg
Gluckman 2011,99 USA	229	For patients with ≥ 1 occluded SVG (n = 70): mean 63 (range 55–72) For patients with patent SVG (n = 159): mean 63 (range 57–71)	Mono	CAD	Smokers: n = 52 (22.7%) Diabetes: n = 84 (36.7%)	Yes: CABG	325 mg/day	No details	No details	Aspirin responsive if TxB2 < 400 pg/mg creatinine
Lordkipanidzé 2011,162 UK (abstract)	198	No details	Mono	CAD	No details	No	80–325 mg/day	No details	No details	No details
Thomson 2009,110 India	63	Mean 57 (SD 10)	Mono	CAD	Smokers: n = 43 (73%) Diabetes: n = 22 (34.9%)	No	75 mg/day	At least 7 days	38.1	Aspirin resistance defined as normalised urinary TxB2 level ≥ 67.9 ng/mmol creatinine

Most studies reported no details on the timing of the PFT after aspirin ingestion. One study148 stated that there were up to 24 hours between aspirin dose and PFT. Table 27 shows details of test characteristics.

TABLE 27 - Test characteristics (thromboxane metabolite measurement, monotherapy)

Study	Details of kit/manufacturer	Anticoagulant (concentration)	Agonist (concentration)	Time since last aspirin dose
Serum/plasma and urinary
Monotherapy at time of PFT and during follow-up
Miyata 2011,164 (abstract)	Serum TxB2 Urinary 11-dehydro-TxB2	No details	Serum TxB2 Urinary 11-dehydro-TxB2	No details
Serum
Monotherapy at time of PFT and during follow-up
Cotter 200446	TxB2 plasma (enzyme immunoassay kit obtained from Amersham, Buckinghamshire, UK)	No details	Collagen (1 µmol)	Aspirin administered on enrolment
Frelinger 200976	Serum TxB2	No details	No details	No details
Urinary
Monotherapy at time of PFT and during follow-up
Addad 2010108	Urinary 11-dehydro-TxB2	Enzyme-linked immunoassay	Enzyme-linked immunoassay	No details
Bruno 2004148	Urinary 11-dehydro-TxB2	No details	No details	Up to 24 hours
Eikelboom 2002151	Urinary 11-dehydro-TxB2	Enzyme immunoassay (Cayman Chemical, Ann Arbor, MI, USA)	Enzyme immunoassay (Cayman Chemical, Ann Arbor, MI, USA)	No details
Eikelboom 2008195	Urinary 11-dehydro-TxB2	No details	No details	No details
Eskandarian 2011202	Urinary 11-dehydro-TxB2	No details	No details	No details
Gluckman 201199	Urinary 11-dehydro-TxB2	No details	No details	No details
Lordkipanidzé 2011162 (abstract)	Urinary 11-dehydro-TxB2	No details	No details	No details
Thomson 2009110	Urinary 11-dehydro-TxB2	No details	No details	No details

Study design and quality

Results of the risk-of-bias assessment can be found in Tables 28–31.

TABLE 28 - Risk of bias, patient selection (thromboxane metabolite measurement, monotherapy)

HOPE, Heart Outcomes Prevention Evaluation.

Domain 1: patient selection	Was a consecutive or random sample of patients enrolled?	Was patient selection independent of patient outcomes?	Were reasons for any posteligibility exclusions provided?
Serum/plasma and urinary
Miyata 2011164 (abstract)	Consecutive	Yes	No details
Serum/plasma
Cotter 200446	Consecutive	Yes	76/82 potentially eligible patients agreed to be interviewed; 73/76 who were on aspirin for at least 1 month were enrolled
Frelinger 200976	Consecutive	Yes	Stated that less than 3% of eligible patients declined participation (reason not given)
Urinary
Addad 2010108	Consecutive	Yes	No details
Bruno 2004148	Unclear: consecutive patients screened for participation in the trial	Yes	98 patients initially gave signed consent; 8/98 withdrew consent and 7/98 did not provide a urine sample
Eikelboom 2002151	Control subjects randomly selected	No	9541 patients in HOPE study; 9282 provided urine samples, samples from 5529 (Canadian centres only) sent to laboratory. Of those, only those who were taking aspirin before and at randomisation, and at each follow-up visit, were eligible for inclusion (number not stated). 488 cases and controls selected from the eligible/included
Eikelboom 2008195	Unclear (patients who complied with a request to provide a sample)	Yes	No details
Eskandarian 2011202 (abstract)	No details	Yes	No details
Gluckman 201199	No details	Yes	Patients for whom SVG patency not assessed or those not on aspirin monotherapy. Authors stated that the study population was representative of patients undergoing isolated CABG surgery based on comparison with the Society of Thoracic Surgeons National Database
Lordkipanidzé 2011162 (abstract)	No details	Yes	No details
Thomson 2012110	Consecutive	Yes	No details

TABLE 29 - Risk of bias, PFT (thromboxane metabolite measurement, monotherapy)

N/A, not applicable; ROC, receiver operating characteristic.

Domain 2: PFT	If a threshold was used, was it prespecified?	How was the threshold derived (e.g. literature cut-off, based on study data)?	Is the undertaking and interpretation of the index test blinded to the patient characteristics (including clinical outcomes)?
Serum/plasma and urinary
Miyata 2011164 (abstract)	No details	No details	No details
Serum/plasma
Cotter 200446	Yes (5 nmol/1011 platelets)	Patients classified as nonresponsive: results in the range observed in volunteers not taking aspirin Patients classified as responsive: results in ranges that are observed in takers Cut-off: lowest TxB2 production value that was observed in non-takers	No details
Frelinger 200976	No (ROC analysis)	ROC analysis of serum TxB2 levels in current study with regard to MACE (resistant if ≤ 3.1 ng/ml)	No details
Urinary
Addad 2010108	No (tertiles)	Tertiles	Yes; stated that all assays were performed in a blinded manner
Bruno 2004148	No (median values for patients with and without events presented)	N/A	Possible; no details, but laboratory off-site
Eikelboom 2002151	No (mean/median values for patients with and without events presented)	N/A	Assays were performed by laboratory staff blinded to patient status (case or control) and also assayed in random order
Eikelboom 2008195	No (quartiles)	Quartiles	Yes; ‘Laboratory staff performing the assays were blinded to treatment allocation and to whether the patients had experienced a primary event’
Eskandarian 2011202 (abstract)	Yes (three groups: resistant > 298 pg/mg, intermediate response 134–298 pg/mg, sensitive < 134 pg/mg)	No details	No details
Gluckman 201199	Yes (aspirin responsive if < 400 pg/mg creatinine; but a threshold of 450 pg/mg creatinine used in model)	‘According to established criteria’ (reference cited221)	No details
Lordkipanidzé 2011162 (abstract)	No details	No details	No details
Thomson 2012110	No (population median value)	Median value of absolute urinary 11-dehydro-TxB2 level of 320 pg/ml used as cut-off in relation to clinical outcomes	No details

TABLE 30 - Risk of bias, outcomes and study attrition (thromboxane metabolite measurement, monotherapy)

Domains 3 and 4: outcomes and study attrition	Were the outcomes of interest clearly defined in advance?	Were the outcome results interpreted without knowledge of the results of the PFT?	What was the proportion of missing data? (State reasons for loss to follow-up or differences in those who completed or were lost)
Serum/plasma and urinary
Miyata 2011164 (abstract)	Yes	No details	No details
Serum/plasma
Cotter 200446	Yes	No details	Appears that there was no loss to follow-up
Frelinger 200976	Yes	Yes; all clinical outcome data obtained by research personnel blinded to results of PFTs	127/682 lost to follow-up (for MACE outcome)
Urinary
Addad 2010108	Yes	Yes; follow-up clinicians were blinded to PFT results	Stated that none of the included patients was lost to follow-up
Bruno 2004148	Yes	Yes; assay results not revealed to investigators until after follow-up examinations and vascular event determinations	Appears that there was no loss to follow-up
Eikelboom 2002151	Yes	Yes (outcome occurred before analysis of sample)	None; retrospective [patients who had a confirmed event were defined as cases and controls were randomly selected from among those with no events (sex and age matched)]
Eikelboom 2008195	Yes	No details	Appears that there was no loss to follow-up
Eskandarian 2011202 (abstract)	Yes	No details	No details
Gluckman 201199	Yes	Yes; stated that images were analysed by two blinded reviewers (98% concordance) with a third reviewer adjudicating as necessary	10/229 not included at 6 months
Lordkipanidzé 2011162 (abstract)	Yes	No details	No details
Thomson 2012110	Yes	No details	11/63 lost to follow-up, unclear if excluded from analysis

TABLE 31 - Risk of bias, confounders (thromboxane metabolite measurement, monotherapy)

BMI, body mass index; N/A, not applicable; TIMI, thrombolysis in myocardial infarction.

Domain 5: confounding	Are confounders accounted for in the design or analysis (e.g. adjustment, stratification)?	If there is an adjusted outcome measure (e.g. OR, HR), what were the factors that were adjusted for?	If a HR was presented, was the proportional hazards assumption met?	Was compliance measured?	How was compliance measured?	Level of compliance
Serum/plasma and urinary
Miyata 2011164 (abstract)	Design: N/A Analysis: appears that adjustment for possible confounders was undertaken, but no adjusted measures presented	No details	N/A	Yes	Interview and by checking plasma concentration of salicylic acid	No details
Serum/plasma
Cotter 200446	Design: N/A Analysis: no	N/A	N/A	Yes	Interview and corroboration with another adult (usually spouse)	12/21 patients (out of a total of 73) whose response was classified as non-responsive stated that they were not taking the medication, the other nine claimed they were
Frelinger 200976	Design: N/A Analysis: yes (OR, HR)	Sex, BMI, TIMI score, aspirin dose, platelet count, use of clopidogrel, statins or oral hypoglycaemic agents	Yes (the assumption of proportionality was tested and found to be valid)	Not specifically	By TxB2 levels	‘Two patients had serum TxB2 levels in the range observed for aspirin-free healthy controls, and their platelet function was therefore consistent with aspirin noncompliance. Because “resistance” cannot be distinguished from noncompliance, these subjects were not excluded from follow-up’
Urinary
Addad 2010108	Design: N/A Analysis: no	N/A	N/A	Yes	Interview at study enrolment and during follow-up period	No details
Bruno 2004148	Design: N/A Analysis: yes (HR)	Age, sex, hypertension, diabetes, smoking, ischaemic heart disease, prior stroke, hypercholesterolaemia, mean 11-dehydro-TxB2 level	No details	No details	N/A	N/A
Eikelboom 2002151	Design: N/A Analysis: yes (OR)	Included conventional vascular risk factors, co-interventions and randomised treatment allocation (full details not given)	N/A	Yes	Assessed compliance with aspirin therapy at each follow-up visit and only considered for inclusion patients who were taking aspirin before randomisation and at 6-month follow-up visits. Patients who discontinued aspirin at any time during the study were not included	No details
Eikelboom 2008195	Design: N/A Analysis: yes (HR)	Subsets of (depending on model): age, sex, BMI, current smoking, hypertension, hypercholesterolaemia, diabetes, past history of MI, stroke, TIA, peripheral artery disease, PCI, CABG, endarterectomy, peripheral angioplasty/bypass, aspirin dose, study clopidogrel, NSAIDs, statins, beta-blockers, diuretics, calcium channel blockers, ACE inhibitors, other blood pressure-lowering agents, oral hypoglycaemic agents, insulin, time of urine collection	No details	Yes	The use of medications, including aspirin dose, was recorded at each follow-up	States in the discussion that non-compliance with aspirin therapy could account for variability in thromboxane concentrations, but that ‘the present data show that factors not related to compliance with aspirin therapy, including age and sex, as well as use of several concomitant therapies, independently determine urinary 11-dehydro thromboxane B2 concentrations’
Eskandarian 2011202 (abstract)	Design: N/A Analysis: no	N/A	N/A	No details	No details	No details
Gluckman 201199	Design: N/A Analysis: no	N/A	N/A	Yes	Pill counts at each postoperative encounter	No details
Lordkipanidzé 2011162 (abstract)	Design: N/A Analysis: unclear whether adjusted or unadjusted OR	No details	N/A	No details	N/A	N/A
Thomson 2012110	Design: N/A Analysis: no	N/A	N/A	No; ‘Aspirin therapy was not supervised and compliance was not verified by assaying’	N/A	N/A

Patient selection was independent of study outcome in 10 of the included studies,46,76,99,108,110,148,162,164,195,202 with the PFT preceding any outcomes (as specified in the study selection criteria). One study151 used a case–control design, so patient selection was not independent of outcome, but the taking of samples for the PFT still preceded the outcomes and so this study was included. Five of 11 studies stated that consecutive patients were enrolled into the study. 46,76,108,110,164 Details on posteligibility exclusion of patients were provided in five studies;46,76,99,148,151 reasons included compliance with aspirin treatment at each follow-up visit,151 patients in whom the outcome was not assessed99 or no provision of urine sample. 148

A predefined threshold was stated in only three studies46,99,202 and this was not consistent across the studies (cut-offs of 298 pg/mg creatinine,202 400 pg/mg creatinine99 and 5 nmol/1011 platelets46). The remaining studies used tertiles,108 quartiles,195 median value,110 derived the value by receiver operating characteristic (ROC) analysis,76 presented mean values for groups with and without events148,151 or gave no details. 162,164 Three studies108,151,195 gave clear details of blinding of laboratory staff to patient characteristics.

Outcome measures of interest were clearly predefined in all studies, and five studies76,99,108,148,151 had details of blinded assessment of outcomes. Five studies appeared to have no loss to follow-up46,108,148,151,195 and there were no details in three studies. 162,164,202 In the remaining three studies the loss to follow-up was 4%,99 17%110 and 19%. 76

Compliance was assessed in six studies. 46,99,108,151,164,195 Methods included interview, plasma concentration of salicylates and pill counts. In one study,46 patients who stated that they were not taking the prescribed aspirin were included as a separate subgroup in the analysis (resistant and non-compliant). Another study76 did not exclude patients as ‘resistance cannot be distinguished from non-compliance’.

Three studies76,148,151 undertook adjusted analyses, with some overlap between the adjustment factors where stated.

Overview of outcomes

Eleven studies were identified; three of these46,76,164 undertook thromboxane measurement in serum/plasma, and nine studies99,108,110,148,151,162,164,195,202 measured thromboxane in urine (one study164 in both categories) (Table 32).

TABLE 32 - Outcomes (thromboxane metabolite measurement, monotherapy)

Study	Death	MACE	Ischaemic/thrombotic	Bleeding	Length of follow-up
Serum/plasma and urinary
Miyata 2011164 (abstract)		✓			2 years
Serum/plasma
Cotter 200446		✓	✓		12 months
Frelinger 200976	✓	✓			Mean 24.8 (SD 0.3) months
Urinary
Addad 2010108		✓			1 year
Bruno 2004148			✓		Mean 2 months (no SD)
Eikelboom 2002151	✓	✓	✓		5 years
Eikelboom 2008195	✓	✓	✓	✓	Median 28 months (no SD)
Eskandarian 2011202 (abstract)		✓			1 year
Gluckman 201199			✓		6 months
Lordkipanidzé 2011162 (abstract)		✓			3 years
Thomson 2012110		✓			Median 36 (range 1–53) months

Death

Death was reported in only 376,151,195 of 11 studies (Table 33). Outcome statistics are shown in Figures 22–25. Unadjusted ORs and HRs were calculable from Frelinger et al. 76 (measurement in serum/plasma), all of which showed a trend towards more events in the resistant arm, though none were statistically significant. The other two studies (measurement in urine) reported adjusted ORs151 and adjusted HRs. 195 Again, all reflected a greater number of events in the resistant arm; two of the adjusted ORs were statistically significant. Note that this is based on comparison of different quartiles rather than using a single cut-off. Overall, the trend was consistent (more events in the resistant arm), but based on few studies.

TABLE 33 - Outcome measures for reporting death (thromboxane metabolite measurement, monotherapy)

Calculated from data given in the publication.

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Serum/plasma
Frelinger 200976	✓a		✓a			✓a
Urinary
Eikelboom 2002151		✓
Eikelboom 2008195				✓

FIGURE 22.

Thromboxane, monotherapy: death, unadjusted ORs.

FIGURE 23.

Thromboxane metabolite measurement, monotherapy: death, adjusted ORs. Q, quartile.

FIGURE 24.

Thromboxane metabolite measurement, monotherapy: death, unadjusted HRs.

FIGURE 25.

Thromboxane metabolite measurement, monotherapy: death, adjusted HRs. Q, quartile.

Major adverse cardiac events

Major adverse cardiac events were reported in nine studies (Table 34). 46,76,108,110,151,162,164,195,202 Outcome statistics are shown in Figures 26–29. For two studies,110,164 results could not be presented in forest plots: one164 stated that ‘no ex vivo measurements for residual platelet functions and COX activities were associated with cardiovascular events. Residual platelet functions correlated poorly with residual COX activities, and were inconsistent with assessments made 6 months later.’ The other110 found that a greater number of MACEs occurred in the upper two quartiles (higher urinary thromboxane levels) than in the lower two and that this difference was statistically significant (p = 0.04); however, the difference was not present when normalised levels of urinary thromboxane were considered.

TABLE 34 - Outcome measures for reporting MACEs (thromboxane, monotherapy)

Calculated from data given in the publication.

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Serum/plasma and urinary
Miyata 2011164 (abstract)					Narrative description of results only
Serum/plasma
Cotter 200446						✓a
Frelinger 200976	✓a	✓	✓a	✓		✓a
Urinary
Addad 2010108	✓a		✓a			✓a
Eikelboom 2002151		✓
Eikelboom 2008195			✓	✓
Eskandarian 2011202 (abstract)	✓a		✓a			✓a
Lordkipanidzé 2011162 (abstract)	✓
Thomson 2012110					Raw data not presented (only a p-value)

FIGURE 26.

Thromboxane metabolite measurement, monotherapy: MACEs, unadjusted ORs. RAVE, recurrent acute vascular event.

FIGURE 27.

Thromboxane metabolite measurement, monotherapy: MACEs, adjusted ORs. Q, quartile.

FIGURE 28.

Thromboxane metabolite measurement, monotherapy: MACEs, unadjusted HRs. RAVE, recurrent acute vascular event.

FIGURE 29.

Thromboxane metabolite measurement, monotherapy: MACEs, adjusted HRs. Q, quartile.

Eight unadjusted ORs were presented, based on four studies, three measuring thromboxane in urine108,162,202 and one in serum/plasma. 76 Six of the ORs reflected more events occurring in the resistant arm, but only two were statistically significant. Four adjusted ORs were presented, based on two studies. 76,151 The direction of effect was consistent and two were statistically significant (more events in the resistant arm).

There were 10 unadjusted HRs based on four studies. 76,108,195,202 Although overall results showed that there were more events in the resistant group, including the three statistically significant results, the direction of effect is not consistent within two studies contributing three195 and four108 unadjusted HRs each; this reflects the effect of using different cut-offs (in this case comparison of different tertiles or quartiles).

There were 12 adjusted HRs based on two studies,76,195 with 11 being based on only one study,195 reflecting the use of different models (adjustment factors) and the comparison of different quartiles rather than one cut-off. One study (serum/plasma)76 showed a statistically significant result (more events in the resistant group), whereas the direction of effect was evenly split in the study generating 11 outcome statistics.

Overall, there is a trend towards more MACEs in the resistant arm, but there is some uncertainty due to the relatively small number of studies contributing MACE results and the fact that there is some inconsistency within studies (depending on thresholds used). Only one study76 measuring thromboxane in serum/plasma was represented in the forest plots. It was therefore not possible to compare results between the two methods, though the direction of effect (more events in the resistant group) was consistent with the majority of results.

Ischaemic/thrombotic events

This outcome measure was reported in five studies (Table 35). 46,99,148,151,195 Outcome statistics are presented in Figures 30–33. Data from one study148 could not be represented in the forest plots. There was no significant difference in thromboxane levels in those with and without a vascular event.

TABLE 35 - Outcome measures for reporting ischaemic/thrombotic events (thromboxane metabolite measurement, monotherapy)

Calculated from data given in the publication.

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Serum/plasma
Cotter 200446	✓a		✓a			✓a
Urinary
Bruno 2004148					Median thromboxane levels compared in those with and without an event
Eikelboom 2002151		✓
Eikelboom 2008195				✓
Gluckman 201199	✓	✓

FIGURE 30.

Thromboxane metabolite measurement, monotherapy: ischaemic/thrombotic events, unadjusted ORs.

FIGURE 31.

Thromboxane metabolite measurement, monotherapy: ischaemic/thrombotic events, adjusted ORs. Q, quartile.

FIGURE 32.

Thromboxane metabolite measurement, monotherapy: ischaemic/thrombotic events, unadjusted HRs.

FIGURE 33.

Thromboxane metabolite measurement, monotherapy: ischaemic/thrombotic events, adjusted HRs. Q, quartile.

Unadjusted and adjusted ORs (based on three studies46,99,151) showed a consistent direction of effect (more events in the resistant group), with the exception of two of six results from one study (Eikelboom,151 adjusted ORs). All four unadjusted ORs and three of the eight adjusted ORs were statistically significant.

Two unadjusted HRs (based on one study46) were statistically significant (more events in the resistant groups), as was one of two adjusted HRs (based on a different study195).

Overall, the direction of effect is consistent (more events in the resistant group), but based on few studies. Note that one study151 contributes to six of eight adjusted ORs, and that the direction of effect is not consistent within this study (reflecting different outcomes and thresholds).

Bleeding events

One study reported this outcome (Table 36). 195

TABLE 36 - Outcome measures for reporting bleeding events (thromboxane metabolite measurement, monotherapy)

Study	Unadjusted OR	Adjusted OR	Unadjusted HR	Adjusted HR	Other measures related to prognosis	Sensitivity/specificity presented or calculable
Urinary
Eikelboom 2008195					Trend across quartiles reported

Only one study195 using a thromboxane test reported bleeding events [Global Utilization of Streptokinase and tPA for Occulded Coronary Arteries (GUSTO) bleeds]. The study found no significant difference when looking at a trend for bleeding rates across quartiles.

Summary: thromboxane metabolite measurement

Eleven studies were identified in this category,46,76,99,108,110,148,151,162,164,195,202 all including stable disease populations, thus making this set of studies more homogenous in terms of population compared with studies reporting other PFTs. There was still heterogeneity, however, for example relating to specific patient characteristics and aspirin dose.

There was a lack of reporting of relevant quality criteria, making overall judgements about risk of bias difficult. No study provided details on all relevant quality criteria. Lack of details related in particular to whether or not assays were performed in a blinded manner and levels of compliance. In one study,46 patients who stated that they were not taking their prescribed aspirin were included as a separate subgroup in the analysis (both resistant and non-compliant, as opposed to resistant and compliant). In the analysis here, these groups have been merged in order to be consistent with the other studies, where it is not possible to make this distinction. There was a lack of consistency in defining thresholds, both in methods and in actual values. Only 346,99,202 of 11 studies gave a predefined threshold. One study151 used a retrospective case–control design, which is more prone to bias than prospective designs; however, as the sampling for the PFT preceded the outcomes, this study was included.

Overall, there was a consistent trend for more deaths reported in the resistant arm, with some statistically significant results, but this was based on only three studies. 76,151,195 There was also a trend for more events in the resistant groups for MACEs and ischaemic/thrombotic events, with some results showing statistical significance, but this is based on relatively few studies (no more than four studies contributed results to any one forest plot).

It is noteworthy that the direction of effect is not consistent within individual studies, reflecting different thresholds used and different outcomes (for ischaemic/thrombotic events). Some studies contributed considerably more to forest plots than others, for example where more outcomes were reported or where results could be presented for different thresholds.

Only one study195 reported bleeding events and this found no significant difference when looking at a trend for bleeding rates across quartiles.

It was not possible to assess any differences between tests measuring thromboxane in urine or serum/plasma, as only one study76 measuring thromboxane in serum/plasma was represented in the forest plots.

Summary: thromboxane

• Eleven studies were identified, all with stable disease populations.

• Methods for deriving thresholds and thresholds themselves were variable.

• A lack of detail on reporting of quality criteria hampered an overall risk-of-bias assessment.

• Heterogeneity in outcomes, test thresholds and types of reported statistics meant that meta-analysis was not considered appropriate.

• Adjusted results were rarely presented, and thus the additional prognostic value of the test over other prognostic factors is difficult to ascertain.

• Despite clinical heterogeneity between studies, there was a general trend for more events to occur in the ‘aspirin-resistant’ group for all relevant outcomes (death, MACEs, ischaemic/thrombotic events); however, this was often based on few studies and there was inconsistency in direction of effect within some studies (based on different thresholds or outcomes).

• Only one study reported bleeding events and this found no significant differences when looking at a trend for bleeding rates.

• Potential differences between measurements of thromboxane in urine or serum/plasma could not be assessed.