Systematically missing confounders in individual participant data meta-analysis of observational cohort studies.

One difficulty in performing meta-analyses of observational cohort studies is that the availability of confounders may vary between cohorts, so that some cohorts provide fully adjusted analyses while others only provide partially adjusted analyses. Commonly, analyses of the association between an exposure and disease either are restricted to cohorts with full confounder information, or use all cohorts but do not fully adjust for confounding. We propose using a bivariate random-effects meta-analysis model to use information from all available cohorts while still adjusting for all the potential confounders. Our method uses both the fully adjusted and the partially adjusted estimated effects in the cohorts with full confounder information, together with an estimate of their within-cohort correlation. The method is applied to estimate the association between fibrinogen level and coronary heart disease incidence using data from 154,012 participants in 31 cohorts

Plasma fibrinogen level and the risk of major cardiovascular diseases and nonvascular mortality: an individual participant meta-analysis.

CONTEXT: Plasma fibrinogen levels may be associated with the risk of coronary heart disease (CHD) and stroke. OBJECTIVE: To assess the relationships of fibrinogen levels with risk of major vascular and with risk of nonvascular outcomes based on individual participant data. DATA SOURCES: Relevant studies were identified by computer-assisted searches, hand searches of reference lists, and personal communication with relevant investigators. STUDY SELECTION: All identified prospective studies were included with information available on baseline fibrinogen levels and details of subsequent major vascular morbidity and/or cause-specific mortality during at least 1 year of follow-up. Studies were excluded if they recruited participants on the basis of having had a previous history of cardiovascular disease; participants with known preexisting CHD or stroke were excluded. DATA EXTRACTION: Individual records were provided on each of 154,211 participants in 31 prospective studies. During 1.38 million person-years of follow-up, there were 6944 first nonfatal myocardial infarctions or stroke events and 13,210 deaths. Cause-specific mortality was generally available. Analyses involved proportional hazards modeling with adjustment for confounding by known cardiovascular risk factors and for regression dilution bias. DATA SYNTHESIS: Within each age group considered (40-59, 60-69, and > or =70 years), there was an approximately log-linear association with usual fibrinogen level for the risk of any CHD, any stroke, other vascular (eg, non-CHD, nonstroke) mortality, and nonvascular mortality. There was no evidence of a threshold within the range of usual fibrinogen level studied at any age. The age- and sex- adjusted hazard ratio per 1-g/L increase in usual fibrinogen level for CHD was 2.42 (95% confidence interval [CI], 2.24-2.60); stroke, 2.06 (95% CI, 1.83-2.33); other vascular mortality, 2.76 (95% CI, 2.28-3.35); and nonvascular mortality, 2.03 (95% CI, 1.90-2.18). The hazard ratios for CHD and stroke were reduced to about 1.8 after further adjustment for measured values of several established vascular risk factors. In a subset of 7011 participants with available C-reactive protein values, the findings for CHD were essentially unchanged following additional adjustment for C-reactive protein. The associations of fibrinogen level with CHD or stroke did not differ substantially according to sex, smoking, blood pressure, blood lipid levels, or several features of study design. CONCLUSIONS: In this large individual participant meta-analysis, moderately strong associations were found between usual plasma fibrinogen level and the risks of CHD, stroke, other vascular mortality, and nonvascular mortality in a wide range of circumstances in healthy middle-aged adults. Assessment of any causal relevance of elevated fibrinogen levels to disease requires additional research.

The long-term within- and between-laboratory variability for assay of antithrombin, and proteins C and S: results derived from the external quality assessment program for thrombophilia screening of the ECAT Foundation.

A stable laboratory performance is important for comparability and transferability of laboratory data both within and between laboratories. The lack of a reference system within hemostasis hampers laboratories in establishing their laboratory performance over a prolonged period of time. Therefore, based on data from an external quality assessment program, we evaluated the between laboratory variation (CVBETWEEN) and the long-term within-laboratory variation (LCVa) for antithrombin, and proteins C and S. We evaluated the CVBETWEEN for the period 1996-2001, including the results of 64-240 laboratories from 23 different surveys (protein S activity 15 surveys). We observed a quite high CVBETWEEN and a broad range for each analyte. The CVBETWEEN was significantly higher for antithrombin and protein S for samples with low levels similar to heterozygous deficiencies. We also evaluated the LCVa, including the results of 136 laboratories. The lowest LCVa[median and 95% content interval (CI)] was observed for antithrombin (7.6%; 3.6-35.5%), intermediate values for protein C activity and antigen (8.6%; 3.5-25.3% and 10.8%; 4.8-33.1%, respectively) and highest values for the protein S variables (13.4%; 6.4-50.6% for total protein S antigen, 14.1%; 6.5-79.1% for free protein S antigen and 17.2%; 7.2-84.3% for protein S activity). We concluded that the main reason for the high CVBETWEEN is the long-term within-laboratory variability. Application of linear regression on data of an external quality assessment program is a useful model to demonstrate per analyte per laboratory the long-term variability (LCVa). It is concluded that improvement of the long-term within-laboratory test performance is the first priority in hemostasis to yield important improvements in the comparability and transferability of laboratory data.

Levels of haemostatic factors, arteriosclerosis and cardiovascular disease.

Plasma levels of haemostasis factors (HFs) such as fibrinogen, tissue-type plasminogen activator (t-PA), plasminogen activator inhibitor-1 (PAI-1) and D-dimer may be markers of arteriosclerosis for the following reasons: There seems to be no difference in levels of HFs between patients with longstanding stable angina and those with an isolated myocardial infarction. HF levels are generally positively associated with subclinical arteriosclerosis as determined by ankle-arm index and carotid ultrasonography in asymptomatic individuals. Levels of most HFs are positively associated with inflammation, which is an essential part of the initiation and progression of the disease. A rough classification is assigned to the associations found in under (2) and (3). Fibrinogen is strongly associated with subclinical arteriosclerosis and with inflammation; Factor VII is not, while an intermediate group is formed by, for instance, von Willebrand Factor (vWF), Factor VIII (F VIII), t-PA, PAI-1, and D-dimer. Also, the associations of HFs with cardiovascular events follow a similar pattern. Fibrinogen is a strong and consistent risk factor in several studies, Factor VII is not, and a similar intermediate group as mentioned under (2) and (3) exists. It suggests that the risk of cardiovascular events in relation to HF levels is explained by their identity as markers of arteriosclerosis. A causal association between HF levels and the disease is not proven. Out of the HF, the markers of coagulation such as thrombin-antithrombin complex and of fibrinolysis such as D-dimer are more likely to act causally. Increased levels indicate that they are markers of arteriosclerosis, but in addition, they may reflect a low-grade, continuous formation and subsequent lysis of fibrin in the disease. As the latter reflects an increased tendency to thrombosis, a causal association of levels of markers of coagulation and fibrinolysis with arteriosclerosis, although not yet proven, seems likely.

Activation products of the haemostatic system in coronary, cerebrovascular and peripheral arterial disease.

To determine the presence of a 'hypercoagulable state' as assessed by indices of thrombin and plasmin generation and of the amount of fibrin that is lysed, in patients with stable coronary, cerebral and peripheral arterial disease a population-based cross-sectional study was performed. From a population-based cohort comprising 7983 men and women aged 55 years and over, we randomly selected 127 subjects with a history of myocardial infarction, 124 with a history of stroke and/or transient ischemic attack, 131 patients with peripheral arterial disease and 263 control subjects in the same age group without arterial disease. Subjects using anticoagulant drugs were not selected. F1+2, TAT, and PAP were not associated with a history of cardiovascular events, nor with peripheral arterial disease. In contrast, positive associations were found for D-Dimer. Mean D-Dimer level was 40 microg/l (95% CI 35, 44) in control subjects; 53 microg/l (47, 61) in those with a history of myocardial infarction and 51 microg/l (45, 58) in those with a history of stroke and or transient ischemic attack. D-Dimer increased gradually with increasing severity of peripheral atherosclerosis; a decrease in ankle/arm systolic blood pressure ratio of 0.1 was associated with an increase in D-Dimer of 3.9 microg/l (p<0.01). This was more pronounced in subjects with higher F1+2, TAT and PAP concentration. In conclusion, the markers of onset of coagulation F1+2, TAT and PAP are not associated with the presence of arterial disease, but increased levels of these markers are necessary for the positive association between D-Dimer and arterial disease.

Fibrinogen Caracas V, an abnormal fibrinogen with an Aalpha 532 Ser-->Cys substitution associated with thrombosis.

A new dysfibrinogenemia associated with thrombophilia has been identified in a Venezuelan kindred. Thrombin and Reptilase times were prolonged and the accelerating capacity of the patient's fibrin on the t-PA-induced plasminogen activation was decreased. In addition the affinity of fibrinogen for plasminogen was diminished. Permeability and electron microscopy studies revealed that the abnormal clot was made up of thin and densely packed fibres giving rise to a reduced fibrin gel porosity. This was confirmed by turbidity studies showing a decreased fibre mass/length ratio. Affected members were heterozygous for an Aalpha 532 Ser-->Cys mutation as demonstrated by genetic analyses. This abnormal fibrinogen has been designated as Fibrinogen Caracas V. The family study showed a convincing association between the mutation and thrombotic manifestations. The thrombotic tendency may be ascribed to lack of accelerating capacity of fibrin to induce fibrinolysis caused by an abnormal clot structure with thin fibres and reduced porosity.

Performance of semiquantitative and quantitative D-dimer assays in the ECAT external quality assessment program.

D-dimer levels are used in clinical practice as markers for the presence or exclusion of venous thrombosis. Therefore, it is important that the performance of the D-dimer tests is well controlled. One of the components of a laboratory quality program is external quality assessment (EQA). The main aim of EQA is to identify the degree of agreement among laboratories. In addition to identifying the individual laboratory performance, it also identifies the characteristics of the different reagents and methods that are used in different laboratories. Recently, the D-dimer test was included in the ECAT External Quality Assessment program. Eighty-one laboratories in four rounds applied semiquantitative and quantitative D-dimer assays to three samples (concentrations normal, slightly elevated, and elevated). Although the reported values varied widely, the classification of the samples was mostly correct for the normal and the elevated samples. The slightly elevated sample was classified as normal by all laboratories using semiquantitative methods and by 60% of the laboratories using quantitative methods, and this varied between 45% and 86% for the different methods. In conclusion, D-dimer methods show a large variation, and the quality of the D-dimer assessment could be improved by standardization and by using cut-off ranges.

Antibodies to Chlamydia pneumoniae and clinical course in patients with unstable angina pectoris.

Inflammation is one of the most important mechanisms that contribute to coronary artery disease (CAD). One of the micro-organisms that is mentioned as a source of the inflammation is Chlamydia pneumoniae. In this study, we investigated the relationship between titres of IgG and IgA antibodies to C. pneumoniae and the clinical course, during hospitalisation and during an 18-month follow-up, in 211 patients admitted to hospital with unstable angina pectoris. Slightly more patients who were refractory during their hospitalisation were positive for C. pneumoniae antibodies than patients who could be stabilised by drug treatment (53 vs. 43%, for IgG and 16 vs. 11% for IgA, respectively)(n.s.). In logistic regression analysis no significant predictive values were observed for the relationship between antibody titres and clinical course. The antibody titres to C. pneumoniae were lower in the unstable angina patients who had plasma levels of interleukin-10 (IL-10) above 5 pg/ml than in the patients with levels below 5 pg/ml, and higher in smokers than in non-smokers. No associations were observed between antibody titres to C. pneumoniae and C-reactive protein (CRP), interleukin-6 (IL-6), age, total cholesterol levels, fibrin degradation products (FDP), plasminogen activator inhibitor-1 (PAI-1) and erythrocyte sedimentation rate (ESR). In conclusion, there was no significant association between antibody titres to C. pneumoniae and risk of events during hospitalisation and the 18-month follow-up period in patients admitted for unstable angina pectoris.

Inactivation of single-chain urokinase-type plasminogen activator by thrombin in human subjects.

Thrombin cleaves single-chain urokinase-type plasminogen activator (scu-PA) into a virtually inactive two-chain form (tcu-PA/T), a process that may protect a blood clot from early fibrinolysis. It is not known under what circumstances tcu-PA/T can be generated in vivo. We have studied the occurrence of tcu-PA/T in human subjects with a varying degree of hypercoagulability. tcu-PA/T was assessed in the plasma of patients with disseminated intravascular coagulation (DIC), endotoxin-treated volunteers, patients with unstable angina pectoris, and patients selected for hip replacement. Relationships between tcu-PA/T and several markers reflecting thrombin generation were examined. tcu-PA/T was observed only in the plasma of patients with DIC and was associated with all thrombin markers and with scu-PA and urokinase antigen. Prothrombin fragment 1 + 2 and urokinase antigen were independent predictors of tcu-PA/T. The fact that tcu-PA/T could not be detected in the other three groups was explained by a lower extent of thrombin generation, a greater inhibition of thrombin by antithrombin, or less available urokinase antigen in these groups. The contribution of scu-PA to total urokinase antigen was decreased in the patients with DIC because of inactivation by thrombin, which may be an additional explanation for the inadequate fibrinolysis observed in these patients. These findings show that scu-PA can be inactivated in the circulation under severe pathophysiologic circumstances and that the process of inactivation depends not only on the generation of thrombin but also on the control of thrombin activity by its inhibitor antithrombin.

Prediction of deep vein thrombosis after elective hip replacement surgery by preoperative clinical and haemostatic variables: the ECAT DVT Study. European Concerted Action on Thrombosis.

The European Concerted Action on Thrombosis (ECAT) DVT Study was a collaborative study of preoperative haemostatic tests in prediction of DVT (diagnosed by routine bilateral venography) after elective hip replacement. 480 patients were recruited in 11 centres across Europe. Clinical risk factors were assessed, and stored citrated plasma aliquots were centrally assayed for 29 haemostatic factors according to the ECAT methodology. 120 (32%) of 375 evaluable patients had DVT, and 41 (11%) had proximal DVT. Among clinical variables, DVT was significantly associated with increased age, obesity, and possibly non-use of stockings. Of the 29 haemostatic factors, mean preoperative levels were significantly higher in patients with subsequent DVT (on univariate analyses) for factor VIII activity, prothrombin fragment F1+2, thrombin-antithrombin complexes, and fibrin D-dimer; and significantly lower for APTT and APC sensitivity ratio. Factor V Leiden was also associated with DVT. Most of these variables were also associated with age, while D-dimer was higher in patients with varicose veins. On multivariate analyses including clinical variables, only a shorter APTT (locally but not centrally performed) and APC resistance showed a statistically significant association with DVT. We conclude that (a) DVT is common after elective hip replacement despite prophylaxis; (b) the study provides some evidence that DVT is associated with a preoperative hypercoaguable state; and (c) preoperative haemostatic tests do not add significantly to prediction of DVT from clinical variables, with the possible exception of APC resistance.

Inflammatory status as a main determinant of outcome in patients with unstable angina, independent of coagulation activation and endothelial cell function.

AIMS: Inflammation, endothelial cell function and the coagulation system have been demonstrated to be involved in the onset and course of unstable angina. Whether a proinflammatory state independently determines outcome is unknown and has not been determined yet in a clinically well defined study population of consecutive patients admitted with unstable angina. METHODS AND RESULTS: Markers of inflammation, coagulation activation and endothelial cell function were determined on admission in blood of 211 consecutive patients with severe unstable angina and were related to the in-hospital course. Refractory unstable angina occurred in 76 patients (36%) during their hospital stay. In a univariate analysis, C-reactive protein (P = 0.03), fibrinogen (P < 0.001) and erythrocyte sedimentation rate (P = 0.001) levels were significantly higher in patients with refractory unstable angina, when compared with patients who had an uneventful clinical course. The odds ratios (95% CI) adjusted for age, sex, body mass index, smoking behaviour and cholesterol levels of the occurrence of refractory unstable angina for patients in the highest quartile compared with patients in the lowest quartile of inflammatory markers were 2.19 (0.94-5.11) for C-reactive protein, 2.83 (1.13-7.10) for fibrinogen and 4.72 (1.70-13.09) for the erythrocyte sedimentation rate. The findings were not affected by the presence or absence of myocardial necrosis or the interval between onset of angina and blood collection. No association was found between markers of coagulation activation or markers of endothelial cell function, and in-hospital outcome. CONCLUSION: We found that in a clinically well-defined study population of patients with severe unstable angina, a proinflammatory state is an important and independent determinant of short-term outcome. The data strengthen the importance of inflammation in this syndrome.

Fibrinolytic activity in peripheral atherosclerosis in the elderly.

Increased concentrations of plasminogen activator inhibitor type 1 (PAI-1) and of D-dimer have jointly been found in subjects with cardiovascular disease. To understand this apparent paradox of increased inhibition of fibrinolysis (high PAI-1) combined with increased fibrinolytic activity (high D-dimer), we examined the relation between D-dimer, PAI-1 and the activator of fibrinolysis, tissue type plasminogen activator (t-PA) in subjects with varying severity of peripheral atherosclerosis. In 325 subjects selected from the Rotterdam Study, a cohort of 7983 men and women aged 55 years and over, the ankle to brachial systolic blood pressure ratio, t-PA antigen and activity, PAI-1 antigen and D-dimer were measured. T-PA antigen and t-PA activity were, independent from each other, increased with degree of atherosclerosis; t-PA antigen increased with 3.5 ng/ml (SE 1.7, p = 0.04) and t-PA activity with 0.46 IU/ml (0.20, p = 0.02) per unit decrease in ankle to brachial pressure ratio (i.e. increase in atherosclerosis). PAI-1 antigen was not related to atherosclerosis. More marked atherosclerosis was associated with increased D-dimer, mainly in subgroups with PAI-1 antigen below 50 ng/ml, t-PA antigen below 10 ng/ml, or t-PA activity above 1.5 IU/ml. In contrast to current beliefs, we found that only a fraction of the variation of t-PA antigen was due to the variation in circulating PAI-1 antigen. A slight positive association was observed between t-PA antigen and D-dimer. PAI-1 and t-PA activity were not associated with D-dimer concentration. In conclusion, in subjects with peripheral atherosclerosis PAI-1 antigen is not increased, but low PAI-1 levels (and possibly also low levels of t-PA antigen and high levels of t-PA activity) appear to be required to increase circulating D-dimer. This suggests that increased D-dimer levels in subjects with atherosclerosis do not reflect increased inhibition, but rather reflect increased fibrinolysis.

Elevated plasma fibrinogen: cause or consequence of cardiovascular disease?

An association between increased plasma fibrinogen and an increased risk for myocardial infarction (MI) is well established, but the nature of this association is subject to debate. Our aim was to shed light on the potentially causal nature of this association. We examined whether increased plasma fibrinogen, due to a condition that is independent of cardiovascular events, also increases the risk for MI. A case-control study was performed in 139 subjects with a history of MI and 287 control subjects selected from the Rotterdam Study, a population-based cohort of 7983 subjects aged 55 years and older. The genotype of the -455G/A polymorphism in the fibrinogen beta-gene was determined by polymerase chain reaction. Functional plasma fibrinogen levels were determined according to von Clauss. The plasma level of fibrinogen was significantly higher in subjects with one or two A alleles compared with subjects with the GG genotype: 3.8 (95% confidence interval [CI], 3.6 to 3.9) g/L and 3.6 (3.5 to 3.7) g/L, respectively. With increasing plasma fibrinogen level, the risk for MI increased gradually; a rise in fibrinogen of 1 g/L was associated with a 45% increased risk (odds ratio adjusted for age, sex, and smoking, 1.45; 95% CI, 1.12 to 1.88). There was no association between the genotype of the -455G/A polymorphism and the risk for MI. The -455G/A polymorphism is therefore associated with increased plasma fibrinogen levels but not with an increased risk for MI. These findings indicate that an increased plasma fibrinogen level due to this genetic factor does not increase the risk for MI.

The Dutch Vascular Factors in Dementia Study: rationale and design.

Dementia is a rapidly increasing health problem in the industrialized countries. With the ageing of the population the number of demented persons increases both in relative and absolute terms. Obviously, there is a need for prevention and intervention strategies. We describe the methods and baseline findings of a large study aimed at identifying potentially modifiable vascular, thrombogenic, and metabolic determinants of dementia. The study population consists of subjects 55 years of age or older. Since the vascular wall of the cerebral vessels is different from that of the coronary or peripheral vessels, we formed three subgroups in which vascular risk factors for dementia are studied. Subjects with stroke were distinguished from subjects with coronary or peripheral artery disease, and from subjects without stroke or coronary or peripheral artery disease. To obtain a large enough number of subjects with stroke, cases and controls from a stroke registry were combined with cases and controls of a population-based study from the same region. For the diagnosis of dementia the DSM-III-R criteria were used. Extensive information on cardiovascular risk factors was collected, including indicators of atherosclerosis. Blood and urine were sampled to study platelet function and thrombogenic and metabolic factors. The study population consists of 7,466 subjects, of whom 300 were recruited from a hospital-based stroke registry. Coronary or peripheral artery disease was present in 956 subjects and stroke in 617. Dementia was present in 434 (5.8%) of all subjects. The prevalence of dementia was 3.0, 24.0, and 4.4% in subjects with a history of coronary or peripheral artery disease, a history of stroke, and subjects without a history of coronary or peripheral artery disease or stroke, respectively. The study will allow us to investigate the role of vascular factors in dementia, irrespective of its cause.

Response to activated protein C in subjects with and without dementia. The Dutch Vascular Factors in Dementia Study.

We performed a cross-sectional case-control study among 295 subjects with dementia and 406 control subjects drawn from participants of the Rotterdam Study, a population-based cohort study among subjects aged 55 years or over, and from participants of the Rotterdam Stroke Databank, a hospital-based stroke registry, to evaluate the association of the factor V Leiden mutation and activated protein C (APC) response with dementia and its subtypes. The risk of dementia was 2.11-fold increased among carriers of factor V Leiden mutation relative to subjects lacking factor V Leiden mutation (95% confidence interval, CI, 0.93-4.77). The increased risks of vascular dementia and of Alzheimer's disease were 4.28 (95% CI 1.26-14.5) and 2.15 (95% CI 0.82-5.63), respectively. No association was found for APC response. We showed a nonsignificant twofold increased risk of dementia among subjects with factor V Leiden. The association appeared to be stronger for vascular dementia.

Coagulation and fibrinolysis markers and risk of dementia. The Dutch Vascular Factors in Dementia Study.

We performed a cross-sectional case-control study among 277 subjects with dementia and 298 control subjects drawn from participants of the Rotterdam Study, a population-based cohort study among subjects aged 55 years or over, and from participants of the Rotterdam Stroke Databank, a hospital-based stroke registry, with the objective to evaluate the association of indicators of coagulability, fibrinogen, prothrombin fragments 1+2, thrombin-antithrombin complex (TAT), and indicators of fibrinolysis, plasmin-inhibitor complex, D-dimer and tissue-type plasminogen activator (t-PA) with dementia. Increased levels of TAT, D-dimer and t-PA activity were associated with an increased risk of dementia. Additional stratified analyses indicated that an increased TAT level was the primary factor related to dementia. The present study provides evidence that predominantly increased thrombin generation is associated with dementia.

Specificity of haemostasis abnormalities for vascular phenotypes.

Atherothrombosis is a systemic disease, hence it is difficult to prove the specificity of haemostasis abnormality for any single vascular phenotype. Associations between haemostatic variables and any given phenotype, e.g. (vascular) dementia, should be interpreted with caution, given the overlaps of vascular disease phenotypes, risk factors, and haemostatic variables.