SEARCH

SEARCH BY CITATION

Keywords:

  • coronary heart disease;
  • factor VII;
  • fibrinogen

Summary.

  1. Top of page
  2. Summary.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Background: The long-term associations of established risk factors for coronary heart disease (CHD), for example cholesterol, are well known, but not for the less familiar hemostatic variables. Objectives: To establish whether associations between hemostatic variables and CHD first identified nearly three decades ago have persisted long-term. Methods: The first Northwick Park Heart Study (NPHS-I) recruited 2167 white men and 941 white women, average age at entry 48 years, on whom measures of factor (F) VII activity (VIIc) and plasma fibrinogen were carried out, both at entry and at follow-up approximately 6 years later. Results: During a median follow-up of 29 years, 231 male and 36 female CHD deaths were recorded from notifications by the Office for National Statistics. VIIc at recruitment was significantly related to CHD mortality, corrected rate ratio, RR, per 1 SD increase 1.56 (95% CI 1.29, 1.88) in men and RR 1.78 (95% CI 1.17, 2.72) in women. Recruitment fibrinogen was also strongly related to CHD mortality in men, RR 1.63 (95% CI 1.33, 1.99) but not in women, RR 0.75 (95% CI 0.40, 1.43). The associations persisted after controlling for confounders and were confirmed using 6-year follow-up measurements and in analyses omitting deaths within 10 years of recruitment. Conclusions: The hemostatic system contributes to CHD mortality, and its effect is stable over time. For VIIc, the effect was similar in men and women, while for fibrinogen it appeared to be present only in men.


Introduction

  1. Top of page
  2. Summary.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The main aim of the first Northwick Park Heart Study (NPHS-I), set up in the early 1970s, was to investigate the contribution of the hemostatic system to the pathogenesis of coronary heart disease (CHD) by epidemiological means. In 1980, preliminary results based on 1510 men aged between 40 and 64 years at recruitment showed increased risks of cardiovascular death [1] (mostly because of CHD) with raised levels of two clotting factors, FVII clotting activity (VIIc) and fibrinogen, in addition to the associations of established risk factors such as cholesterol, blood pressure and smoking. In 1986, the main results confirmed strong and independent associations of the two clotting factors with CHD, the finding on VIIc being due to its association with fatal events [2].

In the PROCAM study, which began in 1979, there was also a significant association with VIIc [3]. However, there was no association for VIIc in the Atherosclerosis Risk In Communities study [4], in the Cardiovascular Health Study in elderly men [5], in the Edinburgh Heart Study [6], or the Caerphilly study [7], all of which started after NPHS-I and PROCAM. The second Northwick Park Heart Study (NPHS-II) [8], which also started later, used the same assay as NPHS-I but found no association in univariable analysis and an inverse association in multivariable analysis. The relation between fibrinogen and CHD has now been confirmed in the Fibrinogen Studies Collaboration [9].

The follow-up of NPHS-I participants has now been going on for some 30 years, and the aim of this paper is to see whether the associations identified nearly three decades ago have persisted in the long-term. Besides the original group reported on, we now have data on a further 608 men who were aged less than 40 years at recruitment and have since moved into the age range in which there is an appreciable risk of CHD. Little is known of the effect of hemostatic variables on women's CHD risk and we also give preliminary results on 941 women.

Methods

  1. Top of page
  2. Summary.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Data

Between 1972 and 1978, NPHS-I recruited men and women of all ages working in three industrial or occupational sites in north west London. The response to invitations to participate was about 80% (there having been uncertainties about staff numbers, i.e. denominator, at one site). Methods for the clinical examination, laboratory measurements and assays and follow-up have been described before [2]. For most participants, repeat blood samples were taken at the follow-up examinations about 6.5 years later (interquartile range 6–7 years). Most were available for the follow-up re-examination (2295/3108) (74%), and comparison of those who were or were not available showed them to be similar in terms of their recruitment demographic, biochemical and clotting factor characteristics, and age-adjusted mortality rates (data not shown). Follow-up for non-fatal events ended in 1986. This paper deals with fatal events up to January 2006, with notifications of deaths and their causes having been provided by the Office for National Statistics since the study began, and which are continuing. Blind assessments of causes of death have been made as before [2], based on information from general practitioners, hospitals and coroners. Analyses are confined to white participants who had not previously had myocardial infarcts, but include men with histories of angina.

NPHS-I was the first prospective study to systematically include hemostatic variables and had no other reports on which to base sample size estimates. Carrying out assays of all the hemostatic variables considered as potentially important was laborious and time-consuming, and to begin with involved designing automated techniques to deal with large numbers of samples to the same standards. In addition, a great deal of time was spent in assaying some clotting factors of theoretical interest, which turned out to have no association with CHD.

A gravimetric clot weight method was assayed for fibrinogen [10], and VIIc was measured by biological assay [11]. Cholesterol, triglycerides, body mass index (BMI, weight(kg)/(height(m2)), and systolic blood pressure (SBP, average of three readings) were measured and smoking history was recorded, all by methods described earlier [2].

NPHS-I started in 1972, which was before the development of ethics committees in the UK. It was therefore not submitted to or approved by a committee. However, those approached about the study were given a full explanation of the reasons for it, its nature and what it would involve and agreed to take part.

Statistical analyses

This paper deals with fatal events only (defined as previously [2]). Follow-up time has been calculated from the date of entry until a fatal CHD event, death from another cause, or 90th birthday. Otherwise, final censoring was on 31 January 2006.

Kaplan–Meier survival curves and Nelson–Aalen cumulative incidence curves [12] were computed (not shown), the latter to assess the assumption, which was confirmed, of constant effects over time, and needed to fit Cox regression proportional hazard models. Cox models were fitted using age as the time-scale and therefore give rate ratio (RR) estimates, which are adjusted for age. The effects of entry and follow-up levels of the hemostatic and biochemical variables were examined, entering each variable on both a continuous scale, in terms of effect per one SD change, and also a categorical scale, in fifths of the overall distribution. In the latter case, the age-adjusted RRs were computed relative to the lowest fifth. The effects on continuous scales require the assumption of linearity, unlike those on categorical scales. Both are presented. Likelihood ratio tests were used throughout to assess linear trends and interactions [12].

To allow for the regression dilution effect [13], we have used the method described by Hughes [14], with reliability coefficients obtained from repeat measures 6 weeks apart in the second Northwick Park Study (NPHS-II) [8], which used the same laboratory methods as NPHS-I. These coefficients were 0.65, 0.62, 0.81 and 0.69, for VIIc, fibrinogen, cholesterol and triglycerides, respectively (we considered the 6-week interval in NPHS-II more appropriate than the 6-year follow-up measures in NPHS-I, although using the latter made little difference to the results).

The age-adjusted RRs associated with each variable of interest were initially computed separately for men and women. In view of the remarkable similarity of the results for the two genders for VIIc, estimates based on combined data for men and women were also computed. Analyses were repeated using the measures made around 6 years after recruitment and also using the average of entry and follow-up values, as this may more accurately reflect the values over the whole of the 6-year period. To further check the stability of the estimated effects, interim censoring times were set for December in 1988, 1993 and 1998 in addition to the final censoring in January 2006.

Results

  1. Top of page
  2. Summary.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Table 1 summarizes the main characteristics of the 2167 white men and 941 white women. Ages at entry ranged from 17 to 88 years, with means and medians of 48 in both genders.

Table 1.   Entry variables of white study participants without history of myocardial infarction by gender
VariablesMen (n = 2167)Women (n = 941)
nMeanSDnMeanSD
  1. n, number of study participants; SD, standard deviation; IQR, interquartile range.

  2. *Median and RR range are shown.

  3. Percentage of standard.

Age (years)*216748.118.194148.314.5
Anthropometric
 Height (m)2165173.17.0941161.06.2
 Weight (kg)216675.110.394163.09.6
 Body mass index (kg m−2)216525.13.094124.33.6
Haemostatic
 VIIc (%)2073107.727.1842114.230.3
 Fibrinogen (g L−1)20722.90.78493.00.6
Biochemical
 Cholesterol (mmol L−1)20475.91.28416.21.3
 Triglycerides (mmol L−1) *20331.41.18371.10.7
Behavioural
 Smoking (at entry)
  No111952% 59063% 
  Yes104848% 35137% 
Clinical
 Systolic blood pressure (mmHg) *2164136.028.0941132.028.0
Previous angina
 No209397% 92498% 
 Possible382% 101% 
 Definite362% 71% 

By the end of January 2006, a total of 809 men and 234 women had died after a median follow-up period of 29 years (interquartile range: 24–31 years), giving all-cause mortality rates, respectively, of 146.1 and 90.0 per 10 000 person-years. Among men, 231 deaths (29%) were due to CHD, while in women the number was 36 (15%), with corresponding rates of 41.7 and 13.9 per 10 000 person-years. Almost without exception, CHD rates rise with increasing levels of the established risk factors (Table 2) in both men and women. In view of the expected findings for these risk factors, results in the rest of the paper are confined to VIIc and fibrinogen.

Table 2.   Gender-specific fatal coronary heart disease rates per 10 000 person-years and rate ratios (RR) by entry variables
Variables*Men (n = 2167; D = 231)Women (n = 941; D = 36)Interaction with gender (P-value)
DRateRR95% CIDRateRR95% CI
  1. *Events listed for each category are less than the total because of missing values.

  2. P-value of the likelihood ratio test for linear trend across the categories of the variable.

  3. P-value of the likelihood ratio test for the heterogeneity across the categories of the variable.

  4. §The RR was corrected for regression dilution effect using reliability estimates obtained in Northwick Park Heart Study II.

Body mass index (kg m−2; in fifths)
 <22.11819.4145.31(0.83)
 22.2–23.,92624.51.09(0.60,2.00)69.91.34(0.38, 4.76)
 24.0–25.44033.21.25(0.71,2.17)817.92.29(0.69, 7.62)
 25.5–27.26250.61.67(0.99,2.83)514.11.62(0.43, 6.05)
 >27.28475.62.45(1.47,4.09)1329.83.41(1.11,10.49)
 Linear trend  <0.001   0.02 
 Effect per 1 SD increase  1.38(1.22,1.56)  1.36(1.04,1.78)
Smoking (at entry)
 No9933.412012.21(0.98)
 Yes13251.31.47(1.13,1.91)1616.71.50(0.78,2.90)
 Heterogeneity  0.004   0.23 
SBP (mmHg; in fifths)
 <1201614.5158.0  (0.71)
 120–1292526.61.67(0.89,3.13)00.01
 130–1393229.21.69(0.93,3.07)510.81.82(0.53,6.29)
 140–1555740.51.99(1.14,3.47)1120.32.86(0.99,8.31)
 >155101103.23.77(2.22,6.40)1534.23.95(1.41,11.02)
 Linear trend  <0.001   0.01 
 Effect per 1 SD increase  1.54(1.38,1.73)  1.51(1.15,1.99)
 Corrected§  1.61(1.43,1.82)  1.57(1.17,2.13)
Previous angina
 No21339.613614.1-
 Possible559.01.03(0.43,2.51)00.0
 Definite13190.23.33(1.90,5.83)00.0
 Linear trend  0.001    
Cholesterol (mmol L−1; in fifths)
 <5.02018.1100.0  (0.60)
 5.0–5.53330.31.18(0.68,2.06)514.31
 5.6–6.14137.11.19(0.70,2.03)49.40.95(0.25,3.57)
 6.2–6.95249.21.50(0.89,2.51)1119.81.57(0.54,4.60)
 >6.97282.32.44(1.48,4.00)1119.81.38(0.47,4.07)
 Linear trend  <0.001   0.44 
 Effect per 1 SD increase  1.38(1.22,1.56)  1.25(0.94,1.67)
 Corrected§  1.48(1.27,1.73)  1.32(0.93,1.88)
Triglycerides (mmol L−1; in fifths)
 <0.81821.4134.31(0.32)
 0.8–1.02326.51.08(0.59,2.01)69.91.94(0.48, 7.75)
 1.1–1.44642.61.46(0.85,2.52)612.82.37(0.59, 9.49)
 1.5–2.14438.31.23(0.71,2.12)618.63.14(0.78,12.60)
 >2.18365.62.02(1.22,3.37)1048.27.45(2.04,27.16)
 Linear trend  0.002   0.001 
 Effect per 1 SD increase  1.24(1.12,1.37)  1.35(1.08,1.69)
 Corrected§  1.36(1.17,1.58)  1.55(1.12,2.13)

Mean values of VIIc were higher in those who died of CHD than in other participants for both men and women, but fibrinogen was noticeably higher only for men (Table 3). Table 4 shows that increasing fifths of VIIc at entry are associated with increasing CHD mortality rates in both men and women. Although the RRs are in some cases significant only in the top fifths for both VIIc and fibrinogen, the linear trends are significant. The regression dilution corrected RRs per 1 SD increase in VIIc are 1.56 (95% CI 1.29, 1.88) in men and 1.78 (95% CI 1.17, 2.72) in women, with no evidence of effect modification by gender (test for interaction with gender P = 0.77). Fibrinogen is significantly and positively associated with CHD mortality in men (corrected RR 1.63, 95% CI 1.33, 1.99). There is no evidence of an association in women.

Table 3.   Entry hemostatic variables of white study participants without history of myocardial infarction by fatal coronary heart disease (CHD) status at the end of follow-up
 CHD deathsOther participants
n*MeanSDn*MeanSD
  1. n, number of study participants; SD, standard deviation; *numbers are less than total because of missing hemostatic values.

  2. Percentage of standard.

Men:
 VIIc (%)222118.029.91851106.426.5
 Fibrinogen (g L−1)2203.20.718522.90.7
Women:
 VIIc (%)33132.231.1809113.530.1
 Fibrinogen (g L−1)343.10.68153.00.7
Table 4.   Gender-specific fatal coronary heart disease death rates per 10 000 person-years and age-adjusted rate ratios (RR) by entry hemostatic variables
VariablesMen (n = 2167; D = 231)Women (n = 941; D = 36)Interaction with gender (P-value)
DRateRR95% CIDRateRR95% CI
  1. n, number of study participants; D, number of fatal CHD events; RR, age-adjusted RR estimated from Cox regression models with age as the time scale.

  2. *Events listed for each fifth are less than the total because of missing hemostatic values.

  3. Test for linear trend in categorical variable.

  4. The RR was corrected for regression dilution effect using reliability estimates obtained in Northwick Park Heart Study II.

VIIc (%, in fifths) *
 <86.32823.4125.01 (0.77)
 86.4–100.03329.51.02(0.62,1.69)24.5 
 100.1–113.04742.71.33(0.84,2.13)614.52.24(0.63,7.96)
 113.1–130.45050.81.43(0.90,2.28)612.21.74(0.49,6.19)
 >130.46472.22.11(1.35,3.28)1729.84.04(1.35,12.09)
 Linear trend  <0.001   0.01 
 Effect per 1 SD increase  1.34(1.18,1.51)  1.46(1.11,1.92)
 Corrected  1.56(1.29,1.88)  1.78(1.17,2.72)
Fibrinogen (g L−1; in fifths) *
 <2.42117.01410.61(0.08)
 2.4–2.73731.31.31(0.77,2.24)920.31.36(0.42,4.44)
 2.8–3.04140.61.30(0.77,2.20)510.00.50(0.13,1.87)
 3.1–3.34444.81.33(0.79,2.24)816.00.70(0.21,2.37)
 >3.37788.12.40(1.48,3.91)815.40.60(0.18,2.02)
 Linear trend  <0.001   0.17 
 Effect per 1 SD increase  1.35(1.19,1.53)  0.84(0.56,1.25)
 Corrected   1.63(1.33,1.99)  0.75(0.40,1.43)

The estimated effects of VIIc are not noticeably confounded by gender, smoking, BMI or SBP. Taking account of triglycerides and cholesterol reduced the effect of VIIc, although the linear trend and the corrected RR remain significant. The effect of fibrinogen in men is clearly reduced by controlling for smoking and lipid measures but the corrected RR remains significant (Table 5).

Table 5.   Multiply adjusted rate ratios (RRs) for fatal coronary heart disease by entry values of hemostatic variables
 Adjusted for:
GenderGender + SmokingGender + Body mass index (BMI)Gender + SBPGender + Smoking + BMI+SBPGender + Smoking + BMI+SBP + Triglycerides + Cholesterol
Rate ratio (RR) 95% CIRR 95% CIRR 95% CIRR 95% CIRR 95% CIRR 95% CI
  1. RR, age-adjusted RR estimated from Cox regression models with age as the time scale and fitted on participants without missing values in the potential confounders.

  2. CI, confidence interval.

  3. *Test for linear trend in categorical variable.

  4. The RR was corrected for regression dilution effect using reliability estimates obtained in the Northwick Park Heart Study.

VIIc (%, in fifths)
 <86.3111111
 86.4–100.00.98(0.60,1.60)0.96(0.59,1.57)0.97(0.60,1.59)0.98(0.60,1.60)0.96(0.59,1.56)0.84(0.51,1.39)
 100.1–113.01.34(0.86,2.11)1.33(0.85,2.09)1.31(0.84,2.06)1.26(0.80,1.97)1.23(0.79,1.93)1.16(0.74,1.83)
 113.1–130.41.41(0.91,2.20)1.41(0.90,2.19)1.31(0.84,2.05)1.28(0.82,2.00)1.18(0.76,1.85)0.98(0.62,1.56)
 >130.42.22(1.46,3.39)2.24(1.47,3.42)1.99(1.30,3.03)2.00(1.31,3.05)1.86(1.22,2.85)1.51(0.97,2.35)
 Linear trend*<0.001 <0.001 <0.001 <0.001 0.001 0.03 
Effect per 1 SD increase
 Uncorrected1.36(1.21,1.52)1.36(1.21,1.51)1.30(1.16,1.46)1.31(1.17,1.47)1.27(1.14,1.43)1.21(1.07,1.37)
 Corrected1.60(1.35,1.90)1.60(1.35,1.89)1.50(1.26,1.79)1.51(1.27,1.80)1.45(1.22,1.73)1.34(1.10,1.62)
Fibrinogen (g L−1; in fifths)*
 <2.4111111
 2.4–2.71.31(0.81,2.15)1.28(0.78,2.08)1.27(0.78,2.07)1.30(0.80,2.12)1.24(0.76,2.02)1.07(0.65,1.79)
 2.8–3.01.16(0.71,1.89)1.12(0.69,1.83)1.05(0.65,1.72)1.05(0.64,1.71)0.93(0.57,1.52)0.87(0.53,1.44)
 3.1–3.31.24(0.77,2.00)1.18(0.73,1.92)1.17(0.72,1.89)1.06(0.66,1.72)0.98(0.60,1.59)0.83(0.50,1.37)
 >3.32.04(1.30,3.20)1.89(1.20,2.98)1.86(1.19,2.92)1.73(1.10,2.72)1.46(0.92,2.31)1.30(0.81,2.08)
 Linear trend*0.001 0.01 0.01 0.02 0.14 0.29 
Effect per 1 SD increase
 Uncorrected1.29(1.15,1.46)1.27(1.12,1.43)1.28(1.13,1.44)1.22(1.07,1.38)1.17(1.03,1.33)1.15(1.00,1.32)
 Corrected1.51(1.25,1.84)1.46(1.20,1.78)1.48(1.21,1.80)1.37(1.12,1.68)1.29(1.05,1.58)1.25(1.01,1.56)

In the analyses based on measurements taken after 6 years, which are restricted to men, because of the small numbers of events in women with follow-up measures, VIIc shows a moderate positive association with CHD (Table 6), with lower RRs than those based on entry values. The association based on the mean of VIIc entry and follow-up measures is stronger and more highly significant, while that for fibrinogen in men remains significant using both the 6-year and mean values (Table 6).

Table 6.   Coronary heart disease rate ratios by (a) follow-up and (b) mean of entry and follow-up hemostatic variables in male participants
 Men (n = 1702; D = 172)
DRate ratio (RR)95% CI
  1. RR, age-adjusted rate ratio estimated from Cox regression models with age as the time scale.

  2. CI, confidence interval.

  3. *Events listed for each category are less than the total because of missing hemostatic values.

  4. Test for linear trend in categorical variable.

  5. The RR was corrected for regression dilution effect using reliability estimates obtained in the Northwick Park Heart Study II.

(a) Follow-up VIIc (%)*
 <86.3111
 86.4–100.0231.39(0.68,2.86)
 100.1–113.0321.53(0.77,3.04)
 113.1–130.4451.85(0.96,3.59)
 >130.4321.90(0.96,3.77)
 Linear trend 0.03 
 Effect per 1 SD increase 1.14(0.97,1.33)
 Corrected 1.22(0.96,1.54)
(b) Mean VIIc (%; in fifths)*
 <92.3141
 92.3–104.1241.55(0.80,2.99)
 104.2–114.8332.09(1.12,3.90)
 114.9–128.3311.87(0.99,3.52)
 >128.3362.56(1.38,4.74)
 Linear trend 0.003 
 Effect per 1 SD increase 1.29(1.10,1.52)
(a) Fibrinogen (g L−1; in fifths)*
 <2.4111
 2.4–2.7170.82(0.38,1.75)
 2.8–3.0240.98(0.48,2.01)
 3.1–3.3381.82(0.93,3.58)
 >3.3541.77(0.92,3.41)
 Linear trend 0.001 
 Effect per 1 SD increase 1.25(1.07,1.46)
 Corrected 1.43(1.12,1.84)
(b)Mean fibrinogen (g L−1; in fifths)*
 <2.56131
 2.56–2.82160.92(0.44,1.91)
 2.83–3.07271.44(0.74,2.80)
 3.08–3.46402.10(1.12,3.97)
 >3.46432.44(1.30,4.58)
 Linear trend <0.001 
 Effect per 1 SD increase 1.34(1.15,1.58)

To examine whether the associations based on entry values vary with increasing length of follow-up, we have censored the follow-up periods at different intervals from recruitment (Table 7). The RRs appear remarkably consistent. As the associations of VIIc and fibrinogen were particularly strong for events within the first 5 or 10 years [2], we have re-estimated them after omitting deaths and person-years of follow-up for all participants within 5 or 10 years of recruitment (Table 8). The results further confirm the presence of long-term associations.

Table 7.   Age-adjusted coronary heart disease (CHD) mortality rate ratios (RR) by entry VIIc and fibrinogen censored at different times
VariableCensored at the end of:
1988199319982006
RR95% CIRR95% CIRR95% CIRR95% CI
  1. n, number of study participants; D, number of fatal CHD events; RR, age–adjusted RR estimated from Cox regression models with age as the time scale; CI, confidence interval.

  2. *Test for linear trend in categorical variable.

  3. The RR was corrected for regression dilution effect using reliability estimates obtained in the Northwick Park Heart Study II.

  4. The number of events was too small to compute reliable estimates.

Men:n = 2167, D = 84n = 2167, D = 133n = 2167, D = 176n = 2167, D = 231
VIIc (%, in fifths)
 <86.31111
 86.4–100.00.63(0.24,1.64)0.87(0.45,1.67)0.94(0.52,1.67)1.02(0.62,1.69)
 100.1–113.01.37(0.62,3.03)1.18(0.65,2.15)1.20(0.70,2.06)1.33(0.84,2.13)
 113.1–130.41.40(0.63,3.09)1.20(0.66,2.18)1.28(0.75,2.17)1.43(0.90,2.28)
 >130.42.40(1.16,4.99)1.76(1.00,3.09)2.15(1.31,3.54)2.11(1.35,3.28)
Linear trend*0.001 0.02 <0.001 <0.001 
Effect per 1 SD increase
 Uncorrected1.46(1.21,1.75)1.30(1.11,1.53)1.39(1.21,1.59)1.34(1.18,1.51)
 Corrected1.78(1.34,2.36)1.50(1.17,1.91)1.66(1.35,2.04)1.56(1.29,1.88)
Fibrinogen (g L−1; in fifths)
 <2.41111
 2.4–2.70.50(0.20,1.27)0.74(0.35,1.56)1.01(0.54,1.90)1.31(0.77,2.24)
 2.8–3.00.82(0.37,1.84)1.02(0.52,2.01)1.10(0.60,2.01)1.30(0.77,2.20)
 3.1–3.30.74(0.34,1.65)0.89(0.45,1.75)1.11(0.61,2.02)1.33(0.79,2.24)
 >3.31.23(0.59,2.53)1.73(0.93,3.19)2.03(1.17,3.53)2.40(1.48,3.91)
 Linear trend*0.14 0.01 0.001 <0.001 
Effect per 1 SD increase
 Uncorrected1.32(1.08,1.61)1.32(1.12,1.55)1.33(1.15,1.52)1.35(1.19,1.53)
 Corrected1.57(1.13,2.16)1.56(1.21,2.02)1.58(1.26,1.97)1.63(1.33,1.99)
Women:n = 941, D = 8n = 941, D = 19n = 941, D = 27n = 941, D = 36
VIIc (%, in fifths)
 <86.3    
 86.4–100.011
 100.1–113.02.24(0.53, 9.47)2.24(0.63,7.96)
 113.1–130.41.07(0.21, 5.32)1.74(0.49,6.19)
 >130.43.67(1.03,13.05)4.04(1.35,12.09)
 Linear trend*    0.06 0.01 
Effect per 1 SD increase
 Uncorrected1.35(0.96,1.90)1.46(1.11,1.92)
 Corrected1.59(0.95,2.68)1.78(1.17,2.72)
Fibrinogen (g L−1; in fifths)
 <2.411
 2.4–2.70.55(0.14,2.21)1.36(0.42,4.44)
 2.8–3.00.33(0.08,1.34)0.50(0.13,1.87)
 3.1–3.30.33(0.08,1.26)0.70(0.21,2.37)
 >3.30.45(0.13,1.54)0.60(0.18,2.02)
 Linear trend*    0.30 0.17 
Effect per 1 SD increase
 Uncorrected0.89(0.56,1.43)0.84(0.56,1.25)
 Corrected0.83(0.39,1.78)0.75(0.40,1.43)
Table 8.   Gender-specific age-adjusted rate ratios (RR) by entry hemostatic variables for participants who survived at least 5 or 10 years
VariablesMenWomen
At least 5 years (n = 2106; D = 208)At least 10 years (n = 2011; D = 170)At least 5 years (n = 929; D = 33)At least 10 years (n = 914; D = 30)
RR95% CIRR95% CIRR95% CIRR95% CI
  1. *P-value of the likelihood ratio test for linear trend across the categories of the variable.

  2. The RR was corrected for regression dilution effect using reliability estimates obtained in the Northwick Park Heart Study II.

VIIc (%, in fifths)
 <86.31111
 86.4–100.01.05(0.63,1.75)1.04(0.59,1.83)0.30(0.03,3.34)0.31(0.03,3.37)
 100.1–113.01.27(0.78,2.05)1.17(0.68,2.01)1.35(0.26,6.96)1.41(0.27,7.30)
 113.1–130.41.24(0.77,2.02)1.27(0.75,2.16)1.23(0.25,6.09)1.08(0.21,5.58)
 >130.41.88(1.19,2.99)1.92(1.15,3.19)2.67(0.61,11.63)2.46(0.56,10.89)
 Linear trend *0.004 0.01 0.01 0.03 
Effect per 1 SD increase
 Uncorrected1.28(1.13,1.47)1.29(1.11,1.49)1.46(1.10,1.95)1.42(1.04,1.93)
 Corrected1.47(1.20,1.80)1.47(1.18,1.84)1.80(1.16,2.79)1.71(1.07,2.75)
Fibrinogen (g L−1; in fifths)
 <2.41111
 2.4–2.71.30(0.75,2.25)1.77(0.95,3.30)1.73(0.47,6.38)1.75(0.47,6.46)
 2.8–3.01.19(0.69,2.06)1.33(0.70,2.53)0.61(0.15,2.57)0.62(0.15,2.63)
 3.1–3.31.27(0.74,2.18)1.51(0.81,2.82)0.86(0.23,3.28)0.56(0.13,2.36)
 >3.32.24(1.36,3.71)2.44(1.35,4.43)0.55(0.13,2.21)0.57(0.14,2.33)
 Linear trend *<0.001 0.01 0.10 0.07 
Effect per 1 SD increase
 Uncorrected1.32(1.15,1.51)1.23(1.05,1.44)0.77(0.50,1.18)0.74(0.47,1.17)
 Corrected1.56(1.26,1.94)1.40(1.09,1.81)0.66(0.33,1.30)0.62(0.30,1.29)

There is a significant trend in all cause mortality with increasing VIIc and fibrinogen in men (Table 9). The results in women are inconclusive.

Table 9.   Gender-specific all-cause mortality rates per 10 000 person-years and age-adjusted rate ratios (RR) by entry hemostatic variables
 Men (n = 2167; D = 809)Women (D = 941; D = 234)Interaction with gender (P-value)
DRateRR95% CIDRateRR95% CI
  1. N, number of study participants; D, number of deaths; RR, age-adjusted rate ratio estimated from Cox regression models with age as the time scale; CI, confidence interval.

  2. *Events listed for each category are less than the total because of missing values.

  3. P-value of the likelihood ratio test for linear trend across the categories of the variable.

  4. The RR was corrected for regression dilution effect using reliability estimates obtained in the Northwick Park Heart Study II.

VIIc (%, in fifths)*
 <86.311999.311742.11(0.93)
 86.4–100.0140125.01.02(0.80,1.30)3681.21.31(0.74,2.34)
 100.1–113.0169153.51.16(0.91,1.46)3277.51.00(0.56,1.81)
 113.1–130.4169171.71.16(0.92,1.47)54109.81.33(0.77,2.30)
 >130.4181204.31.45(1.15,1.82)72126.21.45(0.85,2.47)
 Linear trend  0.001   0.14 
 Effect per 1 SD increase  1.14(1.07,1.23)  1.15(1.00,1.31)
 Corrected  1.23(1.11,1.37)  1.24(1.01,1.52)
Fibrinogen (g L−1; in fifths)
 <2.49980.011745.01(0.20)
 2.4–2.7123104.00.92(0.71,1.20)3170.01.06(0.59,1.92)
 2.8–3.0141139.50.93(0.72,1.20)3876.00.83(0.46,1.47)
 3.1–3.3185188.31.16(0.91,1.49)62123.91.19(0.69,2.05)
 >3.3229261.91.49(1.18,1.90)65125.11.05(0.61,1.81)
 Linear trend  <0.001   0.56 
 Effect per 1 SD increase  1.25(1.16,1.34)  1.08(0.93,1.25)
 Corrected  1.42(1.27,1.59)  1.13(0.89,1.43)
Fibrinolytic activity (100/h; in fifths) *
 <10.3191172.2137107.31(0.69)
 10.3–19.9172167.20.96(0.78,1.18)51112.81.02(0.67,1.56)
 20.0–24.9116147.60.85(0.68,1.07)37102.21.04(0.66,1.64)
 25.0–36.3160127.40.88(0.72,1.09)5073.40.85(0.55,1.29)
 >36.3135121.51.03(0.82,1.28)4679.30.98(0.63,1.51)
 Linear trend  0.73   0.58 
 Effect per 1 SD increase  1.01(0.93,1.09)  0.98(0.85,1.12)

Discussion

  1. Top of page
  2. Summary.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The main advantage of NPHS-I over other studies that have included hemostatic variables is its long follow-up period. It therefore complements long-term studies of established variables (cholesterol, blood pressure and smoking) in now being able to see if VIIc and fibrinogen maintain their associations with fatal CHD beyond 5 or 10 years. The findings in women must be regarded as preliminary in view of the limited number of events so far and also because these women are not yet in the age groups of high CHD mortality. However, the remarkable similarity of the findings on VIIc in men and women is clear, and the data on women can therefore justifiably be shown and also combined with those for men (for example, Table 5). Allowing for regression dilution noticeably increases the strength of the VIIc and fibrinogen associations with CHD.

Factor VII activity

The 1986 observation on VIIc in NPHS-I showed a marked positive association with CHD risk, mainly for fatal events, which was highly significant (P = 0.0005) [2]. The association is less strong in the present data up to 2006 than in 1986, but it has persisted over time and is not simply because of the initially strong association in the immediate 5 or 10 years after study entry.

Two other studies, NPHS-II [8] and the recruitment data in the Thrombosis Prevention Trial [15], both of which started several years after NPHS-I, used the same VIIc assay as NPHS-I. NPHS-II showed no association with CHD death, while the Thrombosis Prevention Trial showed a small but statistically significant association in univariable analysis [16]. Other prospective studies have used different assays that may not have been as sensitive to FVII activity [17]. The NPHS-I results over the time course of the study reported now cannot easily be attributed to chance. It may thus be that the level of VIIc was a clear risk factor for CHD deaths in the 1970s and early 1980s, that its significance then remained, though more weakly (as suggested in the results of the PROCAM study [3], the first to start after NPHS-I), and was then no longer seen (or only weakly, in the Thrombosis Prevention Trial analyses) in the studies that began later [4–8]. A recent study based on Mendelian randomization found no association between a FVII polymorphism and CHD [18]. The shortcomings of this approach are discussed in more detail in connection with fibrinogen (below).

There have been several major changes in aspects of lifestyle in developed countries such as the UK and the USA [19,20] over the last three decades, including lower saturated fat consumption. High fat intakes rapidly and markedly increase cholesterol, triglycerides and VIIc itself [21,22], and this is almost certainly because of an effect of the lipids on VIIc and this was established in our analyses. It is thus necessary to correct for the lipids to determine whether there is still an independent effect of VIIc. Although there was no significant interaction between VIIc and age with CHD, older participants in NPHS-I may have undergone proportionately less change in dietary intake of saturated fat compared with younger participants. The UK National Food Survey (http://statistics.defra.gov.uk/esg/publications/nfs/datasets/74web.xls) supports this possibility and so does a study of dietary trends from the mid-1980s until the early 1990s [23]. Deaths in the older participants in NPHS-II could be one reason for still significant risks conferred by VIIc at the different censoring points (Table 7), while participants in the later studies may have been following the trend of declining fat consumption. In the Thrombosis Prevention Trial, it was in fatal events that the considerable reduction in CHD because of low dose warfarin was seen [15,24] (though warfarin reduces the activity of FII, IX and X as well as of FVII) and case fatality in myocardial infarction has fallen considerably over the last two or three decades [25,26]. The extrinsic coagulation system exerts a predominant effect on the activation of the final common pathway and thrombin production [27]. So high VIIc levels might increase the risk of fatal CHD by raising the probability of life-threatening coronary thrombosis, a mechanism which perhaps operates much less than previously in developed countries. Mortality from CHD and stroke declined by some 60% in the US and Canada between 1970 and 2000, compared with 25 to 40% in Latin America [28]. This observation is also compatible with a role for VIIc, which could still be important in settings in which dietary changes may not have occurred to the same extent as in North America, or where resuscitation or treatment methods are not as well developed as elsewhere. The evidence is certainly circumstantial, but changes in diet over time in different communities could explain the obvious contrast between the NPHS-I findings of the 1970s and early 1980s and the other later studies. This explanation could have potentially important implications for prevention and public health policy in developing countries that may be on the way to developing earlier Western dietary and other habits.

Fibrinogen

In NPHS-1, the plasma fibrinogen level is clearly associated with CHD mortality in the long term in men, based on measurements at both entry to the study and at follow-up 6 years later. This effect is particularly stable over time (Tables 7 and 8). However, there was no clear association in women although mean fibrinogen levels in women may be higher than those in men [29]. The Fibrinogen Studies Collaboration [9], with very much larger numbers, does not show a gender difference, but it is agreed that the results of individual studies contributing to the overview should be reported to allow for possible differences in these studies from the overview findings. Individual studies may truly differ from others in some respects, although the small number of events in women in NPHS-I is acknowledged.

The interpretation of the association of fibrinogen with CHD is controversial. On the one hand, it is suggested that fibrinogen, as an acute phase reactant, is simply a marker of the extent or quality of underlying atheromatous disease and inflammation in the vessel wall. One reason for questioning a causal effect comes from the concept of Mendelian randomization, although there are several limitations to its interpretation in the current state of knowledge [30,31], which also apply to the Mendelian randomization findings on FVII (see above). Some of the reservations over Mendelian randomization [32] are illustrated by a recent overview of a fibrinogen polymorphism, fibrinogen level and CHD risk [33]. For example, studies of fibrinogen gene variants have not so far had adequate power to demonstrate or exclude associations with clinical events. Furthermore, it is not yet clear to what extent the fibrinogen gene polymorphism used is affecting the plasma levels of fibrinogen (a prerequisite in Mendelian randomization) or is acting as a marker for variants in the coding region of the gene that would affect fibrinogen function (which would confound its use in Mendelian randomization). A report from the ISIS study, using polymorphisms that influence plasma levels of lipid traits that are almost certainly causal, unexpectedly found no association with CHD events [34], raising concerns over unresolved problems about Mendelian randomization. Clot structure may be subject to complex fibrinogen-FXIII polymorphism and fibrinogen concentration interactions [35], illustrating the potential limitation of relying on a single polymorphism.

High fibrinogen levels may promote the onset of CHD through involvement in several pathways leading to clinical events [36]. In particular, they increase whole blood and plasma viscosity, which are risk factors for CHD [37]. However, fibrinogen levels within the physiological range also influence platelet activity [38,39], the amount of fibrin deposited when coagulation is initiated [40], and clot deformability [41]. High fibrinogen levels also appear to contribute to the atheromatous process [42], and in trying to understand the role of fibrinogen it is necessary to distinguish between the short-term process of thrombosis and the longer term process of atheroma formation. High fibrinogen levels may be both a marker of the inflammatory consequence of atherogenesis (as well as a possible contributor to this process) and, at the same time, a cause of CHD through its effect on thrombus formation [36].

Randomized controlled trials of selective fibrinogen-lowering agents would be the best way of resolving the role of fibrinogen in CHD. However, there are at present no such agents that can be taken orally. Most of the fibrates lower fibrinogen levels but they also have what would be considered to be beneficial lipid-modifying effects. Fibrates also raise the homocyst(e)ine level [43], a possible risk factor for CHD [44,45]. It is therefore difficult to be certain about the pathways through which fibrates (other than gemfibrozil, which differs chemically from other fibrates) affect CHD.

The association of fibrinogen with all causes of death in men has been reported in the Fibrinogen Studies Collaboration [9] and the explanation for this, along with our result on VIIc, awaits further study. A large contribution to all cause mortality by CHD must partly but not entirely be responsible.

Conclusions

  1. Top of page
  2. Summary.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Our results strengthen the case for a significant contribution of the hemostatic system to CHD, in the long as well as in the short term. VIIc was clearly associated with fatal CHD in men in NPHS-I and equally if not more strongly in women as well, and it seems premature to conclude that VIIc plays no part in CHD in some circumstances. The effects of diet on VIIc, in particular fat intake, may partly explain our results on VIIc and the changes in CHD mortality in developed countries over the last 30 years. This interpretation may be of particular relevance in developing countries, along with the increasing incidence of smoking, in strategies for prevention. Fibrinogen is related to CHD mortality in men, this effect being particularly stable over time. Whatever the explanation for this association, fibrinogen is useful as an additional long-term risk indicator in some people [36].

Acknowledgements

  1. Top of page
  2. Summary.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The study has been funded by the Medical Research Council and the British Heart Foundation. We are most grateful to the late G. Miller and J. Cooper for providing the reliability coefficients from NPHS-II and to G. Miller for commenting on the paper. We thank C. Knottenbelt for obtaining and processing the information from general practitioners, hospitals and coroners needed for the assessment of events, J. Dickinson for independent assessments (initially by G. Rose, until his death in 1991), and Y. Stirling for her comments and managing the coagulation tests and other laboratory work. S. Humphries and P. MacCallum also kindly commented on drafts of the paper. We thank our many other past colleagues in the MRC Epidemiology and Medical Care Unit for their contributions to NPHS-I over the years. Both authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Disclosure of Conflict of Interests

  1. Top of page
  2. Summary.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The authors state that they have no conflict of interest.

References

  1. Top of page
  2. Summary.
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  • 1
    Meade TW, North WRS, Chakrabarti R, Stirling Y, Haines AP, Thompson SG. Haemostatic function and cardiovascular death: early results of a prospective study. Lancet 1980; 1: 10504.
  • 2
    Meade TW, Mellows S, Brozovic M, Miller GJ, Chakrabarti RR, North WRS, Haines AP, Stirling Y, Imeson JD, Thompson SG. Haemostatic function and ischaemic heart disease: principal results of the Northwick Park Heart Study. Lancet 1986; 2: 5337.
  • 3
    Assmann G, Cullen P, Heinrich J, Schulte H. Hemostatic variables in the prediction of coronary risk: results of the 8 year follow-up of healthy men in the Munster Heart Study (PROCAM). Prospective Cardiovascular Munster Study. Isr J Med Sci 1996; 32: 36470.
  • 4
    Folsom AR, Wu KK, Rosamond WD, Sharrett AR, Chambless LE. Prospective study of hemostatic factors and incidence of coronary heart disease: the Atherosclerosis Risk in Communities (ARIC) Study. Circulation 1997; 96: 11028.
  • 5
    Tracy RP, Arnold AM, Ettinger W, Fried L, Meilahn E, Savage P. The relationship of fibrinogen and factors VII and VIII to incident cardiovascular disease and death in the elderly: results from the cardiovascular health study. Arterioscler Thromb Vasc Biol 1999; 19: 177683.
  • 6
    Smith FB, Lee AJ, Fowkes FG, Price JF, Rumley A, Lowe GD. Hemostatic factors as predictors of ischemic heart disease and stroke in the Edinburgh Artery Study. Arterioscler Thromb Vasc Biol 1997; 17: 33215.
  • 7
    Smith A, Patterson C, Yarnell J, Rumley A, Ben Shlomo Y, Lowe G. Which hemostatic markers add to the predictive value of conventional risk factors for coronary heart disease and ischemic stroke? The Caerphilly Study. Circulation 2005; 112: 30807.
  • 8
    Cooper JA, Miller GJ, Bauer KA, Morrissey JH, Meade TW, Howarth DJ, Barzegar S, Mitchell JP, Rosenberg RD. Comparison of novel hemostatic factors and conventional risk factors for prediction of coronary heart disease. Circulation 2000; 102: 281622.
  • 9
    Fibrinogen Studies Collaboration. Plasma fibrinogen level and the risk of major cardiovascular diseases and nonvascular mortality: an individual participant meta-analysis. JAMA 2005; 294: 1799809.
  • 10
    Chakrabarti R, Fearnley GR, Hocking ED, Clarke GM. Fibrinolytic activity related to age in survivors of myocardial infarction. Lancet 1966; 1: 5734.
  • 11
    Brozovic M, Stirling Y, Harricks C, North WR, Meade TW. Factor VII in an industrial population. Br J Haematol 1974; 28: 38191.
  • 12
    Clayton D, Hills M. Statistical Models in Epidemiology. Oxford: Oxford University Press, 1996.
  • 13
    MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, Godwin J, Dyer A, Stamler J. Blood pressure, stroke, and coronary heart disease. Part 1, prolonged differences in blood pressure: prospective observational studies corrected for the regression dilution bias. Lancet 1990; 335: 76574.
  • 14
    Hughes MD. Regression dilution in the proportional hazards model. Biometrics 1993; 49: 105666.
  • 15
    The Medical Research Council's General Practice Research Framework. Thrombosis prevention trial: randomised trial of low-intensity oral anticoagulation with warfarin and low-dose aspirin in the primary prevention of ischaemic heart disease in men at increased risk. Lancet 1998; 351: 23341.
  • 16
    Rudnicka AR, Mt-Isa S, Meade TW. Associations of plasma fibrinogen and factor VII clotting activity with coronary heart disease and stroke: prospective cohort study from the screening phase of the Thrombosis Prevention Trial. J Thromb Haemost 2006; 4: 240510.
  • 17
    Miller GJ, Stirling Y, Esnouf MP, Heinrich J, Van De Loo J, Wu KK, Morrissey JH, Meade TW, Martin JC, Imeson JD, Cooper JA, Finch A. Factor VII-deficient substrate plasmas depleted of protein C raise the sensitivity of the factor VII bio-assay to activated factor VII: an international study. Thromb Haemost 1994; 71: 3848.
  • 18
    Ye Z, Liu EH, Higgins JP, Keavney BD, Lowe GD, Collins R, Danesh J. Seven haemostatic gene polymorphisms in coronary disease: meta-analysis of 66,155 cases and 91,307 controls. Lancet 2006; 367: 6518.
  • 19
    Unal B, Critchley JA, Capewell S. Modelling the decline in coronary heart disease deaths in England and Wales, 1981–2000: comparing contributions from primary prevention and secondary prevention. Br Med J 2005; 331: 614.
  • 20
    Wright JD, Kennedy-Stephenson J, Wang CY, McDowell MA, Johnson CL. Trends in Intake of energy and macronutrients – United States, 1971–2000. JAMA 2004; 291: 11934.
  • 21
    Miller GJ, Martin JC, Webster J, Wilkes H, Miller NE, Wilkinson WH, Meade TW. Association between dietary fat intake and plasma factor VII coagulant activity – a predictor of cardiovascular mortality. Atherosclerosis 1986; 60: 26971.
  • 22
    Miller GJ, Martin JC, Mitropoulos KA, Reeves BEA, Thompson RL, Meade TW, Cooper JA, Cruikshank JK. Plasma factor VII is activated by post-prandial triglyceridaemia irrespective of dietary fat composition. Atherosclerosis 1991; 86: 16371.
  • 23
    Prevost AT, Whichelow MJ, Cox BD. Longitudinal dietary changes between 1984–5 and 1991–2 in British adults: association with socio-demographic, lifestyle and health factors. Br J Nutr 1997; 78: 87388.
  • 24
    Rudnicka AR, Ashby D, Brennan P, Meade T. Thrombosis prevention trial: compliance with warfarin treatment and investigation of a retained effect. Arch Intern Med 2003; 163: 145460.
  • 25
    Ergin A, Muntner P, Sherwin R, He J. Secular trends in cardiovascular disease mortality, incidence, and case fatality rates in adults in the United States. Am J Med 2004; 117: 21927.
  • 26
    Laatikainen T, Critchley J, Vartiainen E, Salomaa V, Ketonen M, Capewell S. Explaining the decline in coronary heart disease mortality in Finland between 1982 and 1997. Am J Epidemiol 2005; 162: 76473.
  • 27
    Rosenberg RD, Aird WC. Vascular-bed-specific hemostasis and hypercoagulable states. N Engl J Med 1999; 340: 155564.
  • 28
    Rodriguez T, Malvezzi M, Chatenoud L, Bosetti C, Levi F, Negri E, La Vecchia C. Trends in mortality from coronary heart disease and cerebrovascular disease in the Americas: 1970–2000. Heart 2006; 92: 45360.
  • 29
    Lowe GD, Rumley A, Woodward M, Morrison CE, Philippou H, Lane DA, Tunstall-Pedoe H. Epidemiology of coagulation factors, inhibitors and activation markers: the Third Glasgow MONICA Survey. I. Illustrative reference ranges by age, sex and hormone use. Br J Haematol 1997; 97: 77584.
  • 30
    Davey SG, Ebrahim S. ‘Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol 2003; 32: 122.
  • 31
    Keavney B. Commentary: Katan's remarkable foresight: genes and causality 18 years on. Int J Epidemiol 2004; 33: 114.
  • 32
    Meade TW, Humphries SE, De Stavola BL. Commentary: fibrinogen and coronary heart disease–test of causality by ‘Mendelian’ randomization by Keavney et al. Int J Epidemiol 2006; 35: 9447.
  • 33
    Keavney B, Danesh J, Parish S, Palmer A, Clark S, Youngman L, Delepine H, Lathrop M, Peto R, Collins R. Fibrinogen and coronary heart disease: test of causality by ‘Mendelian randomization’. Int J Epidemiol 2006; 35: 93543.
  • 34
    Keavney B, Palmer A, Parish S, Clark S, Youngman L, Danesh J, McKenzie C, Delepine H, Lathrop M, Peto R, Collins R. Lipid-related genes and myocardial infarction in 4685 cases and 3460 controls: discrepancies between genotype, blood lipid concentrations, and coronary disease risk. Int J Epidemiol 2004; 33: 100213.
  • 35
    Lim BC, Ariens RA, Carter AM, Weisel JW, Grant PJ. Genetic regulation of fibrin structure and function: complex gene–environment interactions may modulate vascular risk. Lancet 2003; 361: 142431.
  • 36
    Meade TW. Fibrinogen measurement to assess the risk of arterial thrombosis in individual patients: yes. J Thromb Haemost 2005; 3: 6324.
  • 37
    Lowe GDO. Blood viscosity and cardiovascular disease. Thromb Haemost 1992; 67: 4948.
  • 38
    Meade TW, Vickers MV, Thompson SG, Stirling Y, Haines AP, Miller GJ. Epidemiological characteristics of platelet aggregability. Br Med J (Clin Res Ed) 1985; 290: 42832.
  • 39
    Lowe GD, Reavey MM, Johnston RV, Forbes CD, Prentice CR. Increased platelet aggregates in vascular and nonvascular illness: correlation with plasma fibrinogen and effect of ancrod. Thromb Res 1979; 14: 37786.
  • 40
    Gurewich V, Lipinski B, Hyde E. The effect of the fibrinogen concentration and the leukocyte count on intravascular fibrin deposition from soluble fibrin monomer complexes. Thromb Haemost 1976; 36: 60514.
  • 41
    Scrutton MC, Ross-Murphy SB, Bennett GM, Stirling Y, Meade TW. Changes in clot deformability-a possible explanation for the epidemiological association between plasma fibrinogen concentration and myocardial infarction. Blood Coagul Fibrinolysis 1994; 5: 71923.
  • 42
    Smith EB. Haemostatic factors and atherogenesis. Atherosclerosis 1996; 124: 13743.
  • 43
    MacCallum PK, Cooper JA, Rumley A, Lowe GD, Meade TW. Effect of bezafibrate on plasma homocysteine concentration in men with lower extremity arterial disease. J Thromb Haemost 2004; 2: 3645.
  • 44
    Clarke R, Daly L, Robinson K, Naughten E, Cahalane S, Fowler B, Graham I. Hyperhomocysteinemia: an independent risk factor for vascular disease. N Engl J Med 1991; 324: 114955.
  • 45
    Wald DS, Law M, Morris JK. Homocysteine and cardiovascular disease: evidence on causality from a meta-analysis. Br Med J 2002; 325: 1202.