help button home button ClinMed NetPrints
Warning: This article has not yet been accepted for publication by a peer reviewed journal. It is presented here mainly for the benefit of fellow researchers. Casual readers should not act on its findings, and journalists should be wary of reporting them.

This Article
Right arrow Abstract Freely available
Right arrow Similar articles in this netprints
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Elwood, P. C.
Right arrow Articles by Flower, R.
Right arrow Search for Related Content
Right arrow Articles by Elwood, P. C.
Right arrow Articles by Flower, R.

clinmed/2000080002v1 (August 29, 2000)
Contact author(s) for copyright information

Platelets.BMJ 25th July 2000







PC Elwood MD, FRCP1, A Beswick BSc1, Janet Pickering BSc2, P McCarron MD2,

JR O’Brien DM,FRCP3, SR Renaud PhD4, RJ Flower PhD,DSc5

1. University of Wales College of Medicine, Cardiff

2. Dept of Social Medicine, University of Bristol, Bristol

3. Department of Haematology, St Mary’s Hospital, Portsmouth

4. INSERM, Cedex, France

5. The William Harvey Research Institute, London

Author’s address etc:

Professor PC Elwood,

MRC Unit, Llandough Hospital, Penarth, S. Glam CF64 2XW

029 20715505 Fax 0029 20716339 e-mail



KEYWORDS: Prediction; stroke; myocardial infarction; platelet aggregation

Number of words: Abstract 250; condensed abstract 117; text 2,620 (excluding abstract)


Objectives: A platelet test which is predictive of myocardial infarction (MI) and/or stroke would enable the targeting of antiplatelet drugs on patients at highest risk. The predictive power of several platelet tests for myocardial infarction (MI) and for stroke was examined. The tests were: aggregation to adenosine di-phosphate (ADP) in platelet rich plasma (PRP); aggregation to ADP in whole blood measured by an impedance method, and a test of platelet aggregation induced in whole blood by high-shear flow.

Design: Prospective: base-line platelet tests were related to incident myocardial infarction (MI) and stroke events.

Setting: The Caerphilly Cohort Study of Heart Disease, Stroke and Cognitive Decline.

Subjects: Around 2,000 men aged 50 to 69 years, representing almost 90% of men in the cohort.

Outcome measures: Around 200 MIs and 100 ischaemic strokes occurred during the following ten years.

Results: Neither primary nor secondary aggregation in PRP was predictive of MI. The fifth of men in whom the primary response to ADP was least, showed the highest risk of a subsequent stroke (RO 1.64; 95% CI 1.12 to 2.43). Aggregation in whole blood was not predictive of MI but again the fifth of men with the least platelet response showed the highest stroke incidence (RO 1.79; 95% CI 1.06 to 3.00). Retention of platelets in the high-shear test was not predictive of either event.

Conclusions: None of three platelet tests was found to be predictive of MI. Two tests of aggregation to ADP predicted stroke, but in a way opposite to that expected.




Paragraph for ‘This Week’:

Unlike anti-hypertensives and cholesterol lowering drugs, aspirin and other anti-platelet drugs are given to all patients at risk of a thrombotic event. A test of platelet function that is predictive of a vascular disease event would enable antiplatelet prophylaxis to be targeted on patients at greatest risk. The predictive power of two tests of aggregation to ADP and a test of aggregation by high-shear flow was examined in the Caerphilly Cohort of around 2,000 men. Around 200 MIs and 100 strokes occurred during ten years of follow-up. No test was predictive of MI. The two aggregation tests significantly predicted stroke, but it was the men with the lowest response to ADP who showed the greatest stroke incidence.

Key message box:

  1. A platelet test that is predictive would enable aspirin and other anti-platelet drugs to be focussed on patients at greatest risk of a vascular event.
  2. Platelet tests in current use do not predict MI.
  3. The tests appear to be predictive of stroke, but it is men with the least platelet activity that appear to be at greatest risk.





INTRODUCTION: Platelets play a key role in thrombotic vascular disease.1 Relevant evidence includes the histology of intra-coronary thrombi2, platelet emboli in the coronary micro-circulation of subjects who have died suddenly3,4 and an enhancement of platelet activity after myocardial infarction.5 The best evidence however is the reduction in myocardial and cerebral thrombosis with aspirin6 and with other anti-platelet agents.

There have been few prospective studies of the prediction of thrombotic events by platelet tests, and no platelet test seems to have been evaluated against cerebrovascular disease. Two studies have shown prediction of vascular events by a test of platelet aggregation, but both these were based on very small numbers7,8 but a larger study of a cohort of 740 followed for16 years, gave no evidence of prediction by aggregation in platelet rich plasma (PRP) induced by adenosine diphosphate (ADP) or adrenaline.9

In the first re-examination of the men in the Caerphilly Cohort Study of Heart Disease, Stroke and Cognitive Decline study (Phase II) platelet aggregation to ADP was measured in PRP.10 There was no evidence of prediction of incident ischaemic heart disease (IHD) events during the following 8-10 years11,12 We now present results on an extended follow-up of this cohort together with evidence on the prediction of ischaemic stroke by two further tests of platelet activity: aggregation to ADP in whole blood measured by an impedance method,13,14 and a shear induced platelet retention filter test.15,16 This report also includes an examination of prediction of ischaemic stroke by these tests.

Prediction of vascular events by a platelet test would enable the focussing of aspirin and other anti-platelet drugs on patients at greatest risk of a vascular thrombotic event.


The Caerphilly Study, set up in 1979-83, comprises a population sample of 2502 men aged 45 to 59 years17 This report is based on the results of a platelet test conducted in the first re-examination of the men (Phase II), when they were aged 50 to 64 years, and two platelet tests conducted in Phase III, five years later.

In each Phase of the study, special clinics were held at which questions were asked about a wide range of items including social and life-style factors, diet, medical history, medication. Blood pressure was measured and an ECG recorded. A few days later, blood was taken after an overnight fast. In both Phase II and Phase III of the study, blood was drawn into a one-tenth volume of 0.13M sodium citrate for the platelet tests and these were performed without delay. Two specially trained technicians performed the platelet aggregation tests11,14,18 Other haemostatic and rheological tests were performed by colleagues in collaborating laboratories.19,20,21

The platelet tests: Aggregation in platelet rich plasma (PRP) was measured by the method of O’Brien22 and Born23 as modified by Renaud et al.10 The methodology and preliminary results for this test as performed in this study is described elsewhere.11,12 In brief: ADP is added to a preparation of PRP and the changes in light transmission through the PRP traced by a pen-recorder. The response is divided into a primary wave of reversible aggregation of platelets, and a secondary wave which represents the degree of irreversible aggregation. The primary wave is measured for each subject as the percentage of the maximum change in light transmission between platelet rich and platelet poor plasma, but as we have described elsewhere, the secondary wave was found to be bimodal and has therefore been divided into ‘low’ and ‘high’ secondary aggregation, indicative, respectively, of a small and a large degree of irreversible aggregation.11,12

In the impedance method, platelet aggregation was measured in citrated whole blood. The lowest concentration of ADP required to provoke a defined degree of platelet response was determined with a Chronolog 560 Aggregometer, using the impedance method described by Cardinal and Flower.13 The ‘threshold’ concentration was judged from serial testing of separate aliquots of blood, ranging from 0.1-21.5 m mol/l blood-saline, in a 19 step logarithmic series. The defined degree of platelet aggregation was an increase in impedance of 1.5 W or more within 2.5 minutes of adding ADP.18 A high figure therefore represents a low sensitivity to ADP.

The filter test15,16 was performed immediately after venesection. Lithium heparinised blood was forced at a pressure of 40 mmHg through a 10mm diameter filter (10 m m or finer pore). The patelets are activated, they aggregate, are retained in the filter and block it. The time to blocking can be measured, but as this correlates highly with the percentage retention of platelets the latter measure is reported here. Platelet counts were made using a Coulter S-plus.

Evidence on the reproducibility of these last two, and other haemostatic tests, has been reported.24 In summary, the coefficient of variation based on split samples was 11.5% for aggregation in whole blood and 8.2% for the platelet retention test. The coefficient of variation within subjects, based on the results from 41 men tested twice, two weeks apart, was 7.6% for platelet aggregation in whole blood and 8.7% for the platelet retention filter test.

Vascular disease: Standard diagnostic criteria were used to define ischaemic heart disease and ischaemic stroke. The men were asked about admission to hospital for possible vascular disease and the London School of Hygiene and Tropical Medicine questionnaire on chest pain and intermittent claudication25 was administered, together with questions about stroke, based on the Oxford study. Data from all these, together with admission lists from local hospitals were the basis of a search of hospital and GP notes for possible vascular disease events.

Strokes were classified as either ischaemic or haemorrhagic by a procedure using two independent expert assessors and an epidemiologist, using all available clinical, radiological and pathological information.26 Haemorraghic and other non-ischaemic cerebral events are omitted in all that follows.

Statistical methods: Standard statistical techniques were used to test the significance of the differences in mean values between those who experienced a vascular event and those who did not. Logistic regression was used to calculate the relative odds (RO) for an MI or ischaemic stroke occurring at different levels of an independent variable, while allowing for the effects of confounding factors.


Results are available for each of the tests for around 2,000 men, representing almost 90% of the survivors of the original cohort at each re-examination (table 1). Results for men from whom blood had been obtained with difficulty were omitted because of a possible activation of platelets.

Table 2 summarises the results of all the tests. For the Phase II PRP aggregation test the mean extent of the primary wave in men who went on to experience an MI (25.2%) or an ischaemic stroke (24.1%) was closely similar to the extent in men who experienced no vascular event (25.1%). On the basis of the distribution of secondary waves, which represent irreversible aggregation, the men were divided into ‘low’ and ‘high’. Within the subgroup of men with a ‘high’ response the proportion who subsequently had an MI (13.9%) was closely similar to that in the men who experienced no incident vascular event (12.9%). The proportion of men who had shown a ‘high’ secondary response and later experienced an ischaemic stroke (16.3%) is however significantly greater (P < 0.05) than the proportion in men who experienced no vascular event (12.9%).

In table 3 these Phase II PRP aggregation tests are examined further. Dividing the men into fifths by their primary responses to ADP enables ROs for incident disease events to be calculated. The incidence of MI shows no convincing trend with primary aggregation, and neither this, nor the difference between the low and high responses in secondary aggregation, is significant for MI. Prediction of ischaemic stroke by primary aggregation shows no significant overall trend, yet the men with the lowest levels of aggregation appear to have had the greatest risk of a subsequent stroke. In fact, the odds for stroke in the fifth of men with the least responsive platelets, relative to the other four fifths of men are 1.64 (95% CI 1.12-2.43). This is unexpected. It is also inconsistent with the data for stroke prediction by secondary aggregation, which, although not significant (at P<0.05) suggests an excess incidence of stroke in the men with the greatest degree of irreversible aggregation, as would be expected.

The Phase III whole blood tests are summarised in table 2. The mean dose of ADP required to achieve a standard change in impedance was 4.69 m mol l-1 in the men who had no incident vascular event during follow-up and this was almost identical to the dose required in the men who went on to experience an MI (4.69) or a stroke (4.81). In the same way, the proportion of platelets retained in the filter under conditions of high shear-stress was almost identical in the three groups (65.1%, 64.1% and 64.9%).

Table 4 shows the ROs for an ischaemic event in the men subdivided by their responses in the whole blood impedance test. Again there is no evidence of any significant trend in the incidence of MI nor do the odds for an MI in the group with the highest primary aggregation differ significantly from the base group. The ROs for stroke do however show a suggestive trend and though this is not significant at P < 0.05, the evidence suggests that it is the men with the least responsive platelets that had the highest odds of a subsequent stroke event. Again, as with the PRP test, the RO for stroke in the fifth of men with the least responsive platelets, relative to the other four fifths of men are significantly raised (RO = 1.79; 95% CI 1.06-3.00). This again is unexpected, and as table 4 shows, the odds for an incident stroke in the men with the highest fifth of platelet responses is highly and significantly reduced in relation to the men with the least responsive platelets (RO = 0.25; 95%CI 0.11–0.56).

Table 5 presents the data for the filter test with the percentage platelet retention used to divide the men into fifths. Within these fifths the incidence of MI and stroke is shown. There is no suggestion of any trend, nor does the group of men with the highest retentions show any excess in MI or ischaemic stroke incidence.


A simple test of platelets could be of considerable value in clinical practice if subjects at increased risk of a vascular event could be identified. This would enable platelet active medication, such as aspirin, to be more specifically targeted than at present.

Evidence that platelets play a key role in ischaemic heart disease is virtually conclusive. In particular, the survival benefits of aspirin prophylaxis following myocardial infarction6 suggest a major causal role for platelet activation mediated through cyclo-oxygenase dependent formation of the pro-aggregatory thromboxane A2. It is this mechanism which accounts for both the immediate platelet aggregation induced by ADP in the PRP test, and the irreversible ADP induced platelet aggregation, the primary and the secondary responses respectively.

A test of platelet aggregation based upon whole blood has clear advantages compared to tests performed on blood which has been centrifuged and manipulated in various ways to yield PRP. We therefore included the test devised by Cardinal and Flower.13 This test also takes account of the fact that the removal of blood-borne cells may have important effects upon platelets and may alter their state of aggregability. There are several reports which lend support to this. For example, several authors have described an effect of haematocrit on platelet aggregation: Balduini et al27 noted that the presence of erythrocytes between 40-60% haematocrit inhibits platelet aggregation induced by ADP and adrenaline.

The results of the platelet tests we describe here have already been reported to show meaningful results in relation to prevalent vascular disease, smoking, alcohol and other haematological factors.10,11,14,28,29 In particular we have already reported a positive and significant association between the whole-blood impedance aggregation test results and prevalent IHD18 and significant inter-relationships between it and other haemostatic tests.14,24

The shear induced platelet retention test was also of special interest because of evidence that platelet activities, other than the aggregation consequent upon cyclo-oxygenase production, are of likely relevance to vascular thrombotic events. Thus, platelets react to shear-stress, even when their aggregation response has been blocked by aspirin,30 and platelet adhesion, rather than aggregation, appears to be related to heart rhythm disturbances.31 This test therefore supplemented the other tests we were using, both of which are dependent upon GpIIb/IIIa related actions.

In fact, very few studies of prediction by platelet aggregation have been reported. Trip et al.8 recorded spontaneous aggregation in PRP, without the addition of an agonist, in 149 post MI patients. They reported a significant relationship; the 26 patients who had shown spontaneous aggregation had a relative risk of re-infarction of 5.4, compared with a RR of 3.1 in the 94 who had shown no aggregation. Thaulow et al.7 performed ADP induced aggregation on 487 healthy men. These were followed and the 75 men with the ‘fastest’ ADP induced aggregation had a significantly greater mortality than the 75 with the ‘slowest’ aggregation. Another study of primary ADP-induced aggregation was performed in a sample of 740 men within a larger cohort.9 66 IHD events occurred during the subsequent ten years, but no measure of ADP aggregation at baseline showed an association with IHD incidence. Nor did similar measurements of adrenaline-induced aggregation in 460 men show significant prediction.

We have already reported an absence of prediction of new IHD events by aggregation to ADP12 and the extended data we now present confirm this. There is not even a suggestion of prediction of MI either by aggregation in PRP or in whole blood, or by the platelet blocking filter test.

The aggregation tests appear to predict stroke though the results are not entirely consistent. Secondary, or irreversible aggregation of platelets in PRP following exposure to ADP appears to indicate an increased risk (table 2). Thus, overall, 16% of men with a high level of irreversible aggregation experienced a stroke, compared with 13% in men who remained stroke and MI free; and the RO in these men was 1.26 (n.s.) when compared with men with low levels of irreversible aggregation. These results are what one would reasonably expect.

Prediction of stroke by both primary platelet aggregation in PRP, and on exposure to ADP in whole blood, is however the reverse of what would reasonably be expected. Both techniques suggest that it is the men with the least active platelets who are at highest risk of an ischaemic stroke. Thus the RO for stroke in the fifths of men with the lowest responses to ADP, compared with the men with more active responses to ADP is 1.64 (P<0.05) for the test in PRP and 1.79 (P<0.05) for the whole blood test. Tables 3 and 4 show the ROs for the individual fifths of men and confirm these differences.

We are unable to explain why the platelet responses to ADP, as we have measured them, relate to ischaemic stroke, and not to MI. Nor can we explain the possible inconsistency between primary and secondary aggregation in PRP and the risk of stroke. The most unexpected result is however the fact that the lowest incidence of stroke was in the groups of men judged to have the highest platelet activities, as judged by two tests of the response to ADP, one in PRP, the other in whole blood. It would seem unlikely that these results could be pure chance because of the size of the association and the fact that similar results were shown by two rather different techniques on blood samples taken five years apart. Nevertheless, the present study should be repeated and if the findings are confirmed, further basic work on the tests should be undertaken.



These data were collected by the MRC Epidemiology Unit. Janet Pickering was supported by a British Heart Foundation grant. We are grateful to Mike Etherington of St. Mary’s Hospital , Portsmouth for much of the technical work. We are also grateful to Dr John Giddings and Professor Stan Heptinstall for helpful comments.


1. Flores NA, Sheridan DJ. The pathophysiological role of platelets during myocardial ischaemia. Cardiovasc Res 1994;28:295-302

2. Davies MJ, Thomas A. Thrombosis and acute coronary-artery lesions in sudden cardiac ischemic death. N Engl J Med 1984;310:1137-40.

3. Haerem JW. Platelet aggregates in intramyocardial vessels of patients dying suddenly and unexpectedly of coronary artery disease. Atherosclerosis 1972;15:199-213

4. Haerem JW. Platelet aggregates and mural microthrombi in the early stages of acute, fatal coronary disease. Thromb Res 1974;5:243-9

5. Rubenfire M, Blevins RD, Barnhart M, Housholder S, Selik N, Mammen EF. Platelet hyperaggregability in patients with chest pain and angiographically normal coronary arteries. Am J Cardiol 1986;57:657-60.

6. ISIS-2 (Second Internation Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17,187 cases of suspected acute myocardial infarction: ISIS-2. Lancet 1988;ii:349-60.

7. Thaulow E, Erikssen J, Sandvik L, Stormorken H, Cohn PF. Blood platelet count and function are related to total and cardiovascular death in apparently healthy men. Circulation 1991;84:613-7.

8. Trip MD, Cats VM, Von Capelle FJ, Vreeken J. Platelet hyperreactivity and prognosis in survivors of myocardial infarction. New Engl J Med 1990;322:1549-54.

9. Meade TW, Cooper JA, Miller GJ. Platelet counts and aggregation measures in the incidence of ischaemic heart disease (IHD). Thromb Haemostas 1997;78:926-9.

10. Renaud SC, Beswick AD, Fehily AM, Sharp DS, Elwood PC. Alcohol and platelet aggregation: the Caerphilly Prospective Heart Disease Study. Am J Clin Nutr. 1992;55:1012-17.

11. Elwood PC, Renaud S, Sharp DS, Beswick AD, O'Brien JR, Yarnell JWG. Ischemic heart disease and platelet aggregation. The Caerphilly Collaborative Heart Disease Study. Circulation l991;83:38-44.

12. Elwood PC, Renaud S, Beswick AD, O’Brien JR, Sweetnam PM. Platelet aggregation and incident ischaemic heart disease in the Caerphilly cohort. Heart 1998;80:578-82

13. Cardinal DC, Flower RJ. The electronic aggregometer: a novel device for assessing platelet behavior in blood. J Pharmacol Methods 1980;3:135-58

14. Beswick AD, O'Brien JR, Limb ES, Yarnell JWG, Elwood PC. Shear induced filter blockage. A population based appraisal of a method for the assessment of platelet, white cell and von Willebrand factor interactions. Platelets 1994;5:186-92.

15. O’Brien JR, Salmon GP. Shear stress activation of platelet glycoprotein IIb/IIIa plus von Willebrand factor causes aggregation: filter blockage and the long bleeding time in von Willebrand’s disease. Blood 1987;70:1354-61

16. O’Brien JR. Shear-induced platelet aggregation. Lancet 1990;335:711-3.

17. Caerphilly and Speedwell Collaborative Group. The Caerphilly and Speedwell collaborative heart disease studies. J Epidemiol Community Health 1984;38:259-262.

18. Elwood PC, Beswick AD, Sharp DS, Yarnell JWG, Rogers S, Renaud S. Whole blood impedance platelet aggregometry and ischemic heart disease. The Caerphilly Collaborative Heart Disease Study. Arteriosclerosis 1990;10:1032-36.

19. Sweetnam PM, Thomas HF, Yarnell JW Beswick AD, Baker IA, Elwood PC. Fibrinogen, viscosity and the 10-year incidence of ischaemic heart disease. The Caerphilly and Speedwell Studies. Eur Heart J 1996;17: 1814-20.

20. Sweetnam PM, Thomas HF, Yarnell JW, Baker IA, Elwood PC. Total and differential leucocyte counts as predictors of ischemic heart disease: the Caerphilly and Speedwell Studies. Am J Epidemiol 1997;145:416-21.

21. Lowe GD, Yarnell JW, Sweetnam PM, Rumley A, Thomas HF, Elwood PC. Fibrin D-dimer, tissue plasminogen activator, plasminogen activator inhibitor and the risk of major ischaemic heart disease in the Caerphilly study. Thromb Haemost 1998;79:129-33.

22. O’Brien JR. Platelet aggregation Part II. Some results from a new method of study. J.Clin.Path. 1962;15:452-5.

23. Born GVR Aggregation of blood platelets by adenosine di-phosphate and its reversal. Nature (Lond) 1962;194:927-9.

24. Elwood PC, Beswick AD, O'Brien JR, Yarnell JWG, Layzell JC, Limb ES. Inter-relationships between haemostatic tests and the effects of some dietary determinants in the Caerphilly cohort older men. Blood Coagul Fibrinolysis 1993;4:529-36.

25. Rose GA. The diagnosis of ischaemic heart pain and intermittent claudication in field surveys.Bull WHO 1962;27:645-58

26. Greenwood R, McCarron P, Elwood PC, et al. The incidence and aetiology of stroke in the Caerphilly and Speedwell collaborative studies -–I: Methods and incidence of events. Public Health 2000 (in press).

27. Baldini CL et al. The effect of red cells on platelet aggregation: a study with the electronic whole blood aggregometer. Scand J Clin Lab Invest. 1988;48:337-40.

28. Sharp DS, Bath PMW, Martin JF, Beswick AD, Sweetnam PM. Platelet and erythrocyte volume and count: epidemiological predictors of impedance measured ADP-induced platelet aggregation in whole blood. Platelets 1994;5:252-257.

29. Sharp DS, Beswick AD, O’Brien JR, Renaud S, Yarnell JW, Elwood PC. The association of platelet and red cell count with platelet impedance changes in whole blood and light-scattering changes in platelet rich plasma: evidence from the Caerphilly Collaborative Heart Disease Study. Thromb Haemost 1990;64:211-215

30. Ratnatunga CP, Edmondson SF, Rees GM, Kovacs IB. High-dose aspirin inhibits shear-induced platelet reaction involving thrombin generation. Circulation 1992;85:1077-82.

31. Gebalska J, Herbaczynska-Cedro K, Ceremuzynski L. Platelet adhesion is related to heart rhythym disturbances in the acute phase of myocardial infarction. Int J Cardiol 1993;38:19-24.


Table 1

Numbers of men in the cohort, numbers for whom acceptable platelet tests

are available and the numbers of incident vascular events.




Number of men Incident vascular events:

MI Stroke

Phase II platelet rich plasma tests:

Number of men in cohort 2398 345 156

platelet tests omitted etc. 232 28 9

number with platelet data 2175 317 147

Phase III whole blood tests:

Number of men in cohort 2236 218 115

platelet tests omitted etc. 350 37 19

number with filter test 1749 161 88

number with impedance test 1886 181 96



* NOTE: Men who experienced cerebral lesions other than ischaemic stroke have been omitted, and events judged to have been TIAs have been ignored throughout.





Table 2

Results of the platelet tests in men with and without an incident ischaemic event.



No incident Incident Incident

vascular event MI isch. stroke

Phase II tests:

10 aggregation in PRP

numbers of men 17271 317 147

response % of maximum 25.1 25.2 24.1

SD 7.7 SD 7.6 SD 7.9

Difference from men with no event n.s. n.s.


20 aggregation in PRP

men with a ‘low’ response* 1505 273 123

men with a ‘high’ response* 222 (12.9%) 44 (13.9%) 24 (16.3%)

Differences from men with no event n.s. P < 0.05


Phase III tests:

Whole blood platelet aggregation

numbers of men 16092 181 96

mean dose of ADP (loge) 4.69 4.69 4.81

SD 0.84 SD 0.95 SD 0.74

Difference from men with no event n.s. n.s.


Platelet retention in filter

numbers of men 15033 161 88

% platelets retained 65.0% 64.1% 64.6%

SD 15.7 SD 16.2 SD 14.9

Difference from men with no event n.s. n.s.


* NOTE: As explained in the text the secondary response wave was found to be bimodal and has therefore been divided into ‘low’ and ‘high’ irreversible aggregation following recovery from the peak of the primary wave (Elwood et al 1991; 1998)

The small discrepancies in the numbers of men shown in this and in later tables are because some men had both an MI and an ischaemic stroke. Omissions because of other cerebral events or missing stroke data are as follows:

1: sixteen omissions, 30 men had an ischaemic stroke and an MI

2. eleven omissions, 11 men had an ischaemic stroke and an MI

3. seven omissions, 11 men had an ischaemic stroke and an MI





Table 3

Incident ischaemic stroke and MI in men subdivided by

degree of primary and secondary platelet aggregation to adenosine phosphate (ADP)

measured in platelet rich plasma (PRP)


Total Number of men

Incident MI

No. RO1 RO2

Incident ischaemic stroke

No. RO1 RO2

10 aggregation

lowest fifth


highest fifth

Probability of trend2






















1.08 (0.73-1.06)

1.23 (0.84-1.82)

0.87 (0.58-1.31)

1.21 (0.82-1.79)


















0.51 (0.30-0.89)

0.79 (0.48-1.29)

0.49 (0.28-0.84)

0.64 (0.38-1.07)


2nd aggregation4

  • ‘low’
  • ‘high’















1.04 (0.73-1.48)












1.26 (0.69-2.02)



    1. ROs shown with no adjustment for confounding
    2. ROs adjusted for age, social class, smoking, previous vascular disease and medication relevant to platelet function
    3. Probability for primary aggregation estimated using the test as a continuous variables
    4. A ‘high’ secondary aggregation wave indicates a high degree of irreversible aggregation, while a ‘low’ secondary wave indicates a low amount of irreversible aggregation (see Elwood et al 1991; 1998).
    5. 16 cases with other cerebral events or missing stroke data excluded from these analyses.






Table 4

Prediction by platelet aggregation in whole blood.




Total Number of men

Incident MI

No. RO1 RO2

Incident ischaemic stroke

No. RO1 RO2

WB aggregation3

least: (2.20 ADP +)

(1.20 – 1.80 ADP)

(0.82 – 1.00 ADP)

(0.68 + ADP)

most: (0.56 or less ADP)

Probability of trend 4



























0.87 (0.52 – 1.45)

0.82 (0.50 – 1.33)

0.65 (0.36 – 1.19)

1.00 (0.61 – 1.66)


















0.72 (0.38 – 1.36)

0.65 (0.35 – 1.19)

0.58 (0.28 – 1.23)

0.255 (0.11 – 0.56)


NOTES: Men on aspirin have been omitted

    1. Relative odds (Ros) shown with no adjustment for confounding
    2. ROs adjusted for age, social class, smoking, previous vascular disease and disease and medication relevant to platelet function
    3. The figures shown in brackets are the lowest concentrations of ADP required to produce 1.5 ohms change in impedance after 2.5 minutes
    4. Probabilities were estimated using the test results as a continuous variable
    5. The ROs (0.34 and 0.25) differs from that of the base group (0.34 at P< 0.05 and 0.25 at < 0.005 respectively) although the trend in the whole series is not significant at P < 0.05

Note: In the whole blood aggregation test the index used is the concentration of ADP required to produce 1.5 ohms change in impedance after 2.5 minutes. The results, in terms of concentrations of ADP, were grouped as shown in the table, and the extent of aggregation shown as ‘highest’ and ‘lowest’.

Table 5:

Prediction by the filter retention test




Total Number of men

Incident MI

No. RO1 RO2

Incident ischaemic stroke

No. RO1 RO2

Platelet retention

  • lowest1/5
  • middle 1/5
  • highest 1/5
  • Probability of trend2

























0.86 (0.51 – 1.45)

0.97 (0.58 – 1.62)

1.03 (0.62 – 1.73)

0.80 (0.47 – 1.37)

















1.27 (0.66 – 2.47)

0.96 (0.47 – 1.94)

0.57 (0.25 – 1.32)

1.05 (0.53 – 2.08)



    1. No adjustment for confounding
    2. Adjusted for age, social class, smoking, previous vascular disease and disease and medication relevant to platelet function
    3. Probabilities were estimated using the test results as a continuous variable

This Article
Right arrow Abstract Freely available
Right arrow Similar articles in this netprints
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Elwood, P. C.
Right arrow Articles by Flower, R.
Right arrow Search for Related Content
Right arrow Articles by Elwood, P. C.
Right arrow Articles by Flower, R.