© 1995 American Heart Association, Inc.
Articles |
Increased Fibrinogen Levels in the Offspring of Hypertensive Men
Relation With Hyperinsulinemia and the Metabolic Syndrome
í Válek
e VálkováFrom the Institute for Clinical and Experimental Medicine (J.V., Z.V.) and the Postgraduate Medical School (L.V.), Abakus-Distribution (V.T.), Prague, Czech Republic.
Correspondence to Assistant Professor Jií Válek, MD, PhD, Department of Diabetes and Experimental Therapy, Institute for Clinical and Experimental Medicine, Víde
ská 800, 14000 Prague 4, Czech Republic.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Abstract Epidemiological studies have revealed that elevated fibrinogen concentrations are associated with an increased risk of myocardial infarction, stroke, intermittent claudication, and cardiovascular mortality. The manner in which fibrinogen operates in atherogenesis has not yet been elucidated, but genetic control of fibrinogen levels is partially responsible. Fibrinogen frequently acts in concert with hyperlipidemia, diabetes, hypertension, physical inactivity, and age, variables that are influenced by insulin action. Because the offspring of hypertensive men tend to be hyperinsulinemic and insulin resistant from a young age, we hypothesized that their increased fibrinogen levels might reflect decreased insulin action and thus play a role in the metabolic syndrome. We chose 48 adult offspring (mean age, 38.4 years) of 30 fathers who had been treated for hypertension, and the former were matched by age, body mass index, sex, and smoking habits with 37 control subjects. Elevations in fibrinogen concentration (3.63±0.93 versus 2.87±0.54 g/L, P<.001) paralleled increases in blood glucose and insulin levels, estimates of insulin resistance, and blood pressure. In the offspring, in contrast to the control group, correlations between fibrinogen and metabolic-syndrome variables (ie, insulin, glucose, and waist and hip circumferences) were found. In stepwise multiple regression analyses, age and smoking habits were entered as variables in both study groups, but postload insulin and high-density lipoprotein cholesterol were entered as variables in the offspring group only. We propose that familial predisposition influences the relationship between insulin concentration and fibrinogen, an effect that may contribute to the clinical importance of the metabolic syndrome.
Key Words: fibrinogen • hypertension metabolism • insulin resistance • risk factors
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Prospective studies have shown that fibrinogen levels are strong predictors for detecting the risk of cardiovascular mortality, myocardial infarction, stroke, and intermittent claudication.1 2 3 4 5 6 7 However, our understanding of the mechanisms by which fibrinogen operates in atherogenic processes is still incomplete.3 Although elevated fibrinogen levels may be a sign of advanced atherosclerosis,4 they may also indicate conditions that predispose to atherosclerosis.8 9 Elevated fibrinogen levels have also been linked to the metabolic syndrome and insulin resistance.10 The latter, which leads to hyperinsulinemia, has been related to hypertension, dyslipidemia, and CHD in some studies.11 12 13
Fibrinogen had been known to be under genetic control14 before its association with hyperlipidemia, diabetes, hypertension, and a family history of CHD was demonstrated in large cohorts.15 16 17 It follows that a familial predisposition toward high fibrinogen levels, when combined with some biochemical and/or hemostatic changes,14 15 may enhance the risk for CHD. We hypothesized that elevated plasma fibrinogen levels in such predisposed persons might reflect decreased insulin action. To test this hypothesis, we examined the offspring of hypertensive fathers (the former tend to be hyperinsulinemic and insulin resistant18 19 20 21 22 ) in an effort to analyze the relationship between fibrinogen levels and insulin concentrations, EIRs, and several components of the metabolic syndrome. Our aim was to show whether differences exist in the variables in question (in addition to age and other environmental factors) in individuals with a familial predisposition to hypertension compared with healthy control subjects.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Subjects
We enrolled a total of 60 offspring from the families of 30 probands with hypertension, a subgroup of former participants in the Heart Attack Primary Prevention in Hypertension study.23 During the period 1991 through 1993, 48 men and women responded to our invitation and were examined. They were generally fairly young and mainly of pyknic habitus. Their blood pressure readings did not exceed normal levels, and laboratory measurements showed no signs of major deviations from the norm. Dietary patterns were consistent with those of an age-matched group of the population.
For comparison, we recruited and performed physical examinations on a group of healthy volunteers who were offered financial compensation. Those eligible for the study had to be in good health, as confirmed by clinical examination, and had to have a negative family history for hypertension, diabetes mellitus, and vascular disease, specifically, myocardial infarction, in parents, siblings, and grandparents less than 65 years of age. Examinations were performed on an outpatient basis between 6:30 and 8:30 AM after a 12-hour fast. The aim of the study was explained individually, and each patient's written informed consent was obtained. Study protocols were approved by the local ethics committee.
Using a freestanding metal rule, we measured each subject's standing height (without shoes) to the nearest 10 mm. Weight (without outer garments and shoes) was measured to the nearest 500 g on a digital scale. Waist circumference was measured midway between the lowest rib and the ventral iliac spine and hip circumference as the maximum value at the level of the greater trochanter, with the subject standing erect. Weight and height were used to calculate BMI (weight in kilograms divided by the square of height in meters), whereas waist and hip measures were used for the WHR.
Venous blood samples were obtained to determine TC, TGs, HDL-C, uric acid, fibrinogen, blood glucose, and immunoreactive insulin. With use of the Friedewald equation, LDL-C was calculated as TC-(TGs/2.18+HDL-C).
An oral glucose tolerance test was performed after a 75-g glucose load had been given to each subject. Blood glucose (from capillary blood) and insulin levels were determined in fasting subjects (at 0 minutes) and then after 30, 60, and 120 minutes (postload).
Information about each subject's dietary regimen was obtained by individual interview with a qualified dietary nurse and was based on recall of all food and drink consumed during the previous 3 days. Conversion of dietary information to energy and nutrient contents was performed by consulting appropriate tables.24 Data on smoking habits were obtained by asking the patient if he or she was a nonsmoker (1), ex-smoker (2), or someone who smoked 1 to 4 cigarettes per day (3), 5 to 14 cigarettes per day (4), or more than 15 cigarettes per day (5).
Laboratory Methods
TC and HDL-C were determined by enzymatic methods using the
cholesterol CHOD-PAP monotest and TGs by the
Peridochrom-Triglyceride CHOD-PAP monotest
(Boehringer). HDL-C was assessed in plasma supernatant after
precipitation with dextran sulfate and MgCl2. Blood glucose
was measured by the glucose oxidase method on a Beckman Glucose
Analyzer 2 (Beckman Instruments Inc). Insulin was determined
with OPIDI RIA kits (Institute of Atomic Energy, Swierk, Poland). For
fibrinogen thrombin time determinations, the method of
Claus25 with a coagulometer (series 3.1.F22854
Fibrintimer, Behringwerke) and the Multifibrintest from the same
manufacturer were employed.
For insulin analysis, the blood samples were kept frozen at -80°C; all other analyses were completed on the day of blood withdrawal. Insulin resistance was estimated from fasting insulin concentrations and blood glucose levels by using a published formula.26
All examination data for both the offspring and control groups were processed on a Logostar 386 PC with CSS Statistica software.27
For all continuous variables, the significance of differences between offspring and control subjects was estimated by Student's two-tailed t tests. Pearson rank correlations between each factor and fibrinogen were calculated. The Z test, based on Fisher's z transformation of the correlation coefficient, was used to determine the significance of differences between respective values for each variable in the matched groups. With fibrinogen as the dependent variable, in a stepwise fashion we selected variables that were either bivariately correlated or otherwise related to the differences between offspring and control subjects (P<.001 to .1). Statistical evaluation was considered significant at P<.05.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The offspring (n=48) of hypertensive parents were very similar in age and BMI to the group of 37 control subjects, and the male-female ratio was also similar. There were no differences in the number of cigarettes smoked daily (smoking score; Table 1
). Clear differences were found in the levels of
fasting and 120-minute postload blood glucose, fasting and 120-minute
postload insulin, and EIRs. LDL-C as well as blood pressure means were
found to be higher in the offspring group (LDL-C, 141.7 versus 119.35
mg/dL). Highly significant differences were also seen in fibrinogen
concentrations.
|
Fibrinogen was found to be very significantly correlated with age in
both groups; other correlations with fibrinogen were evident only in
the offspring group. These included correlations between fibrinogen and
BMI, glycemia (fasting and 60-minute postload), insulin (fasting and
60-minute postload, EIR, and waist and hip circumferences. Differences
between correlation coefficients for offspring (significant or nearly
so) and control subjects (nonsignificant, or opposite in direction)
were present in insulin concentrations, EIRs, TG levels, hip
circumferences, and blood pressures (Table 2).
|
Stepwise regression analysis was used to select four
covariables with fibrinogen as the dependent variable, with
a highly significant determination coefficient of 52% for offspring. A
similar procedure was used to perform a stepwise regression
analysis in control subjects, for whom the coefficient was
lower (-.31) and only two variables were entered (age and
smoking habits; Table 3). Partial correlations for age,
insulin concentration, smoking score, and HDL-C are shown in the
Figure. The completely different positions of two
independent variables (ie, 60-minute-postload insulin and
HDL-C) in offspring and control subjects are shown for the two compared
groups.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Increases in plasma fibrinogen level were found in the generally young offspring of hypertensive fathers who had been treated for this condition since their 50s, and these increases were not the only differences between these offspring and age-, BMI-, sex-, and smoking status–matched healthy volunteer control subjects. An association between increased fibrinogen and decreased insulin action was described in a clamp study of a group of untreated hypertensives.10
Insulin resistance has also been associated with increases in the activity of PAI-1, responsible for disorders of fibrinolysis.28 29 30 The implications of elevated PAI-1 levels have been confirmed in subsequent trials that have explored the association between PAI-1 and angina pectoris or myocardial infarction.28 An increased PAI-1 level has become the key component in hemostatic disorders and is believed to be an integral part of the metabolic syndrome.28 29 30 The association between the metabolic syndrome, PAI-1, and other disorders that involve clotting factors (ie, factor VII)30 has effected a substantial shift in our understanding of the relevance of combining both metabolic and hemostatic components in models of atherogenesis and infarctogenesis. Fibrinogen levels tend to fluctuate, depending on infection, tissue injury, and other factors.3 The extent to which genetic factors may determine plasma fibrinogen levels is unknown,30 although path analysis has suggested a high estimate (0.3 to 0.5) of heritability.14 However, fibrinogen level fluctuations may lead to an underestimation of its importance.30
Familial predisposition to increased fibrinogen levels in the offspring of probands who had had a myocardial infarction31 became an attractive topic for investigation in the extensive European Atherosclerosis Research Study.32 In addition to the universal predisposition of the G/A genotype to increased fibrinogen levels, it appears that a number of environmental factors interact with each individual's genetic make-up. Transcription of the ß-fibrinogen gene is affected by cytokines that originate from pulmonary macrophages, the stimulus of which results in repeated bronchial infections, especially in smokers.32 Hence, fibrinogen could also mediate the effects of smoking in the arterial wall.30 32
At least half of the offspring of hypertensive parents have demonstrable metabolic changes21 that could be explained by increased insulin levels. Because of the high correlation coefficients between insulin and its actions (as determined during clamp studies), the term "insulin resistance" is often used despite the fact that it may not meet the strictly defined criteria for true insulin resistance.12 33 Our study is based on EIRs calculated using equations derived from results of investigations of nondiabetic persons by the euglycemic clamp method.26 However, the current definition of the metabolic syndrome does not require the simultaneous presence of its four main components (dyslipidemia, hypertension, androgenous type of obesity, and insulin resistance).33 Familial genetic traits and ethnic specificities may be reasons for a single, common defect that has different manifestations.34 At present, hemostatic disorders are also considered to be part of the metabolic syndrome, made all the more serious by the contribution of hemostasis to thrombogenicity, atherosclerotic plaque formation, and an increased risk of thrombus formation in cases of plaque rupture.30
Although our group of offspring of hypertensive probands did not have a fully developed metabolic syndrome (signs of dyslipidemia and obesity were absent), significant differences were seen in insulin levels, EIRs, glycemia, and blood pressures. In these offspring, fibrinogen was associated exclusively with those components of the metabolic syndrome that reflect an abnormal body make-up and insulin resistance. Differences in correlation coefficients (close to statistical significance in offspring) suggested different relations between fibrinogen and two other cardinal symptoms of the metabolic syndrome, ie, blood pressure and TGs. Almost all metabolic-syndrome components showed different associations with fibrinogen levels in offspring compared with control subjects. The biggest differences in correlation coefficients were seen between fibrinogen and insulin concentration, EIR, and blood pressure. From the point of view of the questions to be answered, these are variables of utmost importance.
In multiple regression analysis with fibrinogen as the dependent variable, age was entered as a variable in both the offspring and control subjects. This may also serve to validate our results, which are consistent with data from epidemiological studies.3 8 9 As much as 52% of the variation in fibrinogen level can be explained by four independent variables. Insulin concentration is one such variable that was entered into the regression, but only in offspring. Also for the offspring group only, cigarette smoking was another important factor, while decreased HDL-C was less important. The differences between offspring and control subjects can be readily derived from fibrinogen partial correlations. The insulin correlation coefficients for both groups are generally opposite in sign, as highly significant positive correlations for the offspring contrast with negative coefficients for the control subjects.
Although the number of offspring in this study was not very high, the group was nevertheless a sample from an ethnically homogeneous population. Fibrinogen is closely related to plasma insulin in a multivariate analysis and bivariately related to most other components of the metabolic syndrome. This fact differentiates predisposed persons from control subjects with lower insulin and fibrinogen levels, which were not significantly correlated with any variable except age. The different levels and associations of fibrinogen with the metabolic syndrome in the offspring of hypertensive fathers may be due to additive effects of their familial predisposition with some environmental factors. Although relations between nutrient composition and energy versus fibrinogen were not significant, smoking did have an effect on fibrinogen levels.
We are well aware of the shortcomings of this clinical study, which was focused on fibrinogen in the offspring of persons with the metabolic syndrome10 while ignoring those hemostatic variables whose associations have repeatedly been demonstrated.29 30 31 To provide an unequivocal answer to the question of whether or not fibrinogen reflects a decrease in insulin action and is therefore a component of the metabolic syndrome, some limitations will need to be overcome. Either a clamp study should be performed in all these subjects (which was not possible, due to disagreement of some offspring) or some approximations, eg, EIRs, insulin concentrations, and components of the metabolic syndrome, should be used. The association between fibrinogen and metabolic syndrome variables that were found in this study could be supported by statistically documented findings in an unselected group of 80% of living offspring of probands with hypertension. If hyperinsulinemia and insulin resistance are indeed the common denominators for metabolic and thrombogenic factors, then these data underscore the importance of implementing lifestyle changes to prevent the adverse consequences of the metabolic syndrome in those with a family history of hypertension. Although the pathogenesis of the metabolic syndrome has not yet been fully elucidated, we should not shun efforts to gain a better understanding of this condition.34 35
|
Selected Abbreviations and Acronyms |
|---|
|
|
Acknowledgments |
|---|
This work was supported by a research grant (No. IGA 2022-3) from the Ministry of Health of the Czech Republic.
Received December 1, 1994; accepted September 19, 1995.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1. Wilhelmsen L, Svärdsudd K, Korsan-Bengtsen K, Larsson B, Welin L, Tibblin G. Fibrinogen as a risk factor for stroke and myocardial infarction. N Engl J Med. 1984;311:501-505. [Abstract]
2. Kannel WB, D'Agostino RB, Belanger AJ. Update on fibrinogen as a cardiovascular risk factor. Ann Epidemiol. 1994;2:457-466.
3. Ernst E. The role of fibrinogen as a cardiovascular risk factor. Atherosclerosis. 1993;100:1-12. [Medline] [Order article via Infotrieve]
4. Bolibar I, Kienast J, Thompson SG, Mathias R, Niessner H, Fechtrup C. Relation of fibrinogen to presence and severity of coronary artery disease is independent of other coexisting heart disease: the ECAT Angina Pectoris Study Group. Am Heart J. 1993;125:1601-1605. [Medline] [Order article via Infotrieve]
5.
Elwood PC, Yarnell JW, Pickering J, Fehilly AM,
O'Brien JR. Exercise, fibrinogen and other risk factors for
Caerphilly Prospective Heart Disease Study. Br Heart
J. 1993;69:183-187.
6.
Heinrich J, Balleisen L, Schulte H, Assmann G, van de
Loo J. Fibrinogen and factor VII in the prediction of
coronary risk: results from the PROCAM study in healthy
men. Arterioscler Thromb. 1994;14:54-59.
7. Banerjee AK, Pearson J, Gilliland EL, Goss D, Lewis JD, Sterling Y, Meade TW. A six year prospective study of fibrinogen and other risk factors associated with mortality in stable claudicants. Thromb Haemost. 1992;68:261-263. [Medline] [Order article via Infotrieve]
8. Folsom AR, Wu KK, Davis CE, Conlan MG, Sorlie PD, Szklo M. Population correlates of plasma fibrinogen and factor VII, putative cardiovascular risk factors. Atherosclerosis. 1991;91:191-205. [Medline] [Order article via Infotrieve]
9.
Krobot K, Hense WH, Cremer P, Eberle E, Keil M.
Determinations of fibrinogen: relation to body weight, waist to hip
ratio, smoking, alcohol and age. Arterioscler
Thromb. 1992;12:780-788.
10. Landin K, Tengborn L, Smith U. Elevated fibrinogen and plasminogen-activator inhibitor (PAI-1) in hypertension are related to metabolic risk factors for cardiovascular disease. J Intern Med. 1990;227:273-278. [Medline] [Order article via Infotrieve]
11. DeFronzo RA, Ferrannini E. Insulin resistance: a multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidemia and atherosclerotic cardiovascular disease. Diabetes Care. 1991;14:173-194. [Abstract]
12. Reaven GM. Relationship between insulin resistance and hypertension. Diabetes Care. 1991;14(suppl 4):33-38.
13.
Haffner SM, Ferrannini E, Hazuda HP, Stern MP.
Clustering of cardiovascular risk factors in confirmed
prehypertensive individuals. Hypertension. 1992;20:38-45.
14. Hamsten A, Iselius L, DeFaire U. Genetic and cultural inheritance of plasma fibrinogen concentration. Lancet. 1987;2:988-991. [Medline] [Order article via Infotrieve]
15. Brunner D, Graff E, Hotting I, Blettner M, Wahrendorf J, Schettler G. Intrafamilial correlations of plasma fibrinogen levels and their relations to lipid and lipoprotein values: Tel-Aviv Heidelberg three generation offspring study. Atherosclerosis. 1994;109:223. Abstract.
16. Ganda OP, Arkin CF. Hyperfibrinogenemia: an important risk factor for vascular complications in diabetes. Diabetes Care. 1992;15:1245-1250. [Abstract]
17.
Lee AJ, Lowe GD, Woodward M, Tunstall-Pedoe H.
Fibrinogen in relation to personal history of prevalent hypertension,
diabetes, stroke, intermittent claudication, coronary heart
disease and family history: the Scottish Heart Health Study.
Br Heart J. 1993;69:338-342.
18. Válková L, Válek J, Karasová L. Aimed prevention of hypertension in families. Prakt Lék. 1993;73:394-396. [in Czech]
19. Beatty OL, Harper R, Sheridan B. Insulin resistance in offspring of hypertensive parents. BMJ. 1993;307:92-96.
20. Facchini F, Chen YD, Clinkingbeard C. Insulin resistance, hyperinsulinemia, and dyslipidemia in nonobese individuals with a family history of hypertension. Am J Hypertens. 1992;5:694-699. [Medline] [Order article via Infotrieve]
21. Neutel JM, Smith HG, Graetinger WF, Winere RL, Weber MA. Metabolic characteristics of hypertensives: importance of positive family history. Am Heart J. 1993;126:924-929. [Medline] [Order article via Infotrieve]
22. Endre T, Mattiason I, Hulthén LV, Lingarde F, Berglund G. Insulin resistance is coupled to physical fitness in normotensive men with a family history of hypertension. J Hypertens.. 1994;12:81-88. [Medline] [Order article via Infotrieve]
23. Wilhemsen L, Berglund G, Elmfeldt D. Beta-blockers versus diuretics in hypertensive men: main results from the HAPPHY trial. J Hypertens. 1987;5:561-572. [Medline] [Order article via Infotrieve]
24.
Krondlová-
kopková M, mrha
O. Tables of Food Nutritive Values. Prague,
Czechoslovakia: Avicenum; 1965:1-575. [in Czech]
25. Claus A. Gerinnungsphysiologische Schnellmethode zur Bestimmung des Fibrinogens. Acta Haematol (Basel). 1957;7:243-246.
26. Mathews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta cell function from fasting plasma glucose and insulin concentrations in men. Diabetologia. 1985;28:412-417. [Medline] [Order article via Infotrieve]
27. CSS Statistica. Tulsa, Okla: StatSoft Inc; 1991:1-968.
28.
Juhan-Vague I, Thompson SG, Jespersen J.
Involvement of the hemostatic system in the insulin resistance
syndrome: a study of 1500 patients with angina pectoris: the ECAT
Angina Pectoris Study Group. Arterioscler
Thromb. 1993;13:1865-1873.
29. Mykkänen L, Rönnemaa T, Marniemi J, Haffner SM, Bergnan R, Laakso M. Insulin sensitivity is not an independent determinant of plasminogen activator inhibitor-1 activity. Arterioscler Thromb. 1994;14,1264-1271.
30. Hamsten A, Erikson P, Karpe T, Silveira A. Insulin, lipoproteins and hemostatic functions. In: Woodford FP, Davignon J, Sniderman A, eds. Atherosclerosis X: Proceedings of the International Symposium in Montreal. Amsterdam, Netherlands: Elsevier Publishers; 1995:628-631.
31.
Hoffman CJ, Burns P, Lawson WE, Katz JP, Miller RH,
Hultin MB. Plasma fibrinogen levels in young adults from
families with premature ischemic disease.
Arterioscler Thromb. 1993;13:800-803.
32.
Humphries SE, Ye S, Talmud P, Bara L, Wilhelmsen L,
Tiret L. European Atherosclerosis Research
Study: genotype at the fibrinogen locus (G-455-A beta-gene)
is associated with differences in plasma fibrinogen levels in young men
and women from different regions in Europe. Arterioscler
Thromb Vasc Biol. 1995;15:96-104.
33. Cigolini M, Seidell JC, Targher G, Deslypere JP, Elsinger BM, Charzewska J, Cruz A, Bjorntorp P. Fasting serum insulin in relation to components of metabolic syndrome in European healthy men: the European Fat Distribution Study. Metabolism. 1995;44:35-40. [Medline] [Order article via Infotrieve]
34. Crepaldi G. The plurimetabolic syndrome or syndrome X. In Woodford FP, Davignon J, Sniderman A, eds. Atherosclerosis X. Proceedings of the International Symposium in Montreal. Amsterdam, Netherlands: Elsevier Publishers; 1995:511-515.
35. Ginsberg HN. Syndrome X: what's old, what's new, what's etiologic. J Clin Invest. 1993;92:1. Editorial.
This article has been cited by other articles:
 |
|
 |
|
  A. Shankar, J. J. Wang, E. Rochtchina, and P. Mitchell Positive Association Between Plasma Fibrinogen Level and Incident Hypertension Among Men: Population-Based Cohort Study Hypertension, December 1, 2006; 48(6): 1043 - 1049. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Arikan and S. Sen Endothelial Damage and Hemostatic Markers in Patients with Uncomplicated Mild-to-Moderate Hypertension and Relationship with Risk Factors Clinical and Applied Thrombosis/Hemostasis, April 1, 2005; 11(2): 147 - 159. [Abstract] [PDF]
|
|
|
|
 |
|
 P. A. Sakkinen, P. Wahl, M. Cushman, M. R. Lewis, and R. P. Tracy Clustering of Procoagulation, Inflammation, and Fibrinolysis Variables with Metabolic Factors in Insulin Resistance Syndrome Am. J. Epidemiol., November 15, 2000; 152(10): 897 - 907. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||