Critique 117: Changes in alcohol intake modify levels of fibrinogen, an important risk factor for coronary heart disease — 18 July 2013
Okwuosa TM, Klein O, Chan C, Schreiner P, Liu K, Green D. Long-term change in alcohol-consumption status and variations in fibrinogen levels: the coronary artery risk development in young adults (CARDIA) study. BMJ Open 2013;3:e002944. doi:10.1136/bmjopen-2013-002944
Authors’ Abstract
Objective: To examine long-term associations between change in alcohol-consumption status and cessation of alcohol use, and fibrinogen levels in a large, young, biracial cohort.
Design: Analysis of covariance models were used to analyse participants within the Coronary Artery Risk Development in Young Adults Study (CARDIA) cohort who had fibrinogen and alcohol use data at year 7 (1992–1993; ages 25–37) and year 20 examinations.
Setting: 4 urban
Patients: 2520 men and women within the CARDIA cohort.
Main outcome measures: 13-year changes in alcohol use related to changes in fibrinogen.
Results: Over 13 years, mean fibrinogen increased by 71 vs 70 mg/dL (p=NS) in black men (BM) versus white men (WM), and 78 vs 68 mg/dL (p<0.05) in black women (BW) versus white women (WW), respectively. Compared with never-drinkers, there were smaller longitudinal increases in fibrinogen for BM, BW and WW (but a larger increase in WM) who became or stayed drinkers, after multivariable adjustment. For BM, WM and WW, fibrinogen increased the most among persons who quit drinking over 13 years (p<0.001 for WM (fibrinogen increase=86.5 (7.1) (mean (SE)), compared with never-drinkers (fibrinogen increase=53.1 (5.4)).
Conclusions: In this young cohort, compared with the participants who never drank, those who became/stayed drinkers had smaller increases, while those who quit drinking had the highest increase in fibrinogen over 13 years of follow-up. The results provide a novel insight into the mechanism for the established protective effect of moderate alcohol intake on cardiovascular disease outcomes.
Forum Comments
This could be an important paper as it reveals the results of changes in alcohol consumption in a well-defined cohort over 13 years. As has been suggested by limited previous studies, people who quit drinking tend to have poorer health outcomes subsequently. This study suggests that an increase in fibrinogen may be one factor leading to increased cardiovascular risk among people who stop alcohol consumption.
However, as Forum reviewer Ellison and others point out, as with other observational research, the reason that some subjects stopped drinking is not known. In the present study, subjects who quit drinking were more likely to have high blood pressure or diabetes at follow up. In an attempt to control for this, baseline and final assessments of these risk factors were adjusted for in the multivariable analyses.
Further, recorded alcohol intake only at baseline and 13 years later were used in the analyses, and no data are included about alcohol intake at other examinations of this cohort between these two points. Forum reviewer Puddey states: “The authors claim that the particularly novel aspect of their study relates to being able to ascertain changes in fibrinogen in relation to many years of alcohol use or abstinence. How can they be certain, however, that any effects of alcohol on fibrinogen seen at follow up related only to alcohol intake in those who have continued to drink or just initiated alcohol intake? The lack of repeatedly assessed data on alcohol consumption throughout the follow-up period makes it difficult to know if the data at both ends of the period reflect usual or chronic alcohol consumption.” In addition, important determinants of cardiovascular risk such as the pattern of drinking (regular moderate, binge) as well as the type of beverage consumed are not presented in the paper.
The authors of the paper conclude: “Overall, persons who continued to use alcohol and those who initiated alcohol consumption during the 13 years of follow-up had smaller changes in fibrinogen levels relative to those who never consumed alcohol.” If other factors are controlled for, this finding could be an important message, and supports
limited data from other cohort studies indicating that moderate drinkers who stop drinking are subsequently at higher risk of cardiovascular disease.
Specific comments by Forum reviewers: Forum reviewer Puddey commented: “The results of this study provide equivocal support for the hypothesis that a long-term effect of alcohol to modulate fibrinogen levels may represent the underlying mechanism for the protective effect of alcohol on cardiovascular disease outcomes. This equivocation is
because of the observation of inconsistent outcomes across race/sex strata within their cohort. In particular, in white men fibrinogen increased more in those who became or stayed drinkers and increased least amongst those who never drank alcohol over a 13 year period of observation.”
Puddey continues: ““An effect of alcohol to reduce fibrinogen is held to reflect an effect of alcohol to reduce platelet aggregation and vascular thrombosis. However fibrinogen levels are also elevated in inflammatory states and as the authors observe, anti-inflammatory effects of alcohol are now well described. Furthermore, interpretation of any influence of alcohol on fibrinogen as overall beneficial without measuring any other markers of coagulation and fibrinolysis is also fraught. In this respect our group (Dimmitt SB, Rakic V, Puddey IB, et al. The effects of alcohol on coagulation and fibrinolytic factors: a controlled trial. Blood Coagul Fibrinolysis 1998;9:39–45) have previously shown that although alcohol directly resulted in a substantial reduction in plasma fibrinogen, it also led to a concomitant increase in factor VII and a relatively greater increase in PAI-1 than tPA – changes which may be interpreted as prothrombotic.”
Forum member Skovenborg stated: “I agree with other reviewers that there is concern about certain aspects of this study — especially the missing information about reasons for quitting, the analysis of alcohol consumption as a dichotomous variable, and the disturbing, unexpected and unexplained difference in the association of alcohol drinking and fibrinogen levels found in white men. Never-the-less the study opens a window of information about long-term alcohol consumption (in the moderate range) on fibrinogen, a co-factor for platelet aggregation and a major determinant of blood viscosity and atherogenesis in a younger healthy population with limited risks of confounding.”
Skovenborg continues: “The differences in fibrinogen levels among the 4 categories (continued non-drinker, became drinker, stayed drinker, quit drinking) are small (in the magnitude of 10 – 30 mg/dl). In a study of risk factors and 10-year risk for cardiovascular disease among 6,371 participants in NHANES III aged 40 to 79 years, the fibrinogen levels associated with low 10-year risk was 299.6 mg/dl, intermediate risk: 306.9 mg/dl and high risk 340.9 mg/dl (Park CS et al. Relation Between C-Reactive Protein, Homocysteine Levels, Fibrinogen, and Lipoprotein Levels and Leukocyte and Platelet Counts, and 10-Year Risk for Cardiovascular Disease Among Healthy Adults in the USA. Am J Cardiol 2010;105:1284 –1288). Hence, at least some effects could be expected with the changes observed in the present study. In a large meta-analysis of plasma fibrinogen level and the risk of cardiovascular disease, the age-and sex-adjusted hazard ratio per 100 mg/dl increase in usual fibrinogen level for CHD was 2.42 (Fibrinogen Studies Collaboration. Plasma Fibrinogen Level and the Risk of Major Cardiovascular Diseases and Nonvascular Mortality.JAMA. 2005;294:1799-1809).” Before concluding that fibrinogen is the key factor, however, potential changes in other clotting factors associated with changes in alcohol consumption should be taken into account.
Forum Summary
A prospective analysis of data from 2,520 subjects in the Coronary Artery Risk Development in Young Adults Study (CARDIA) cohort related reported alcohol intake at two examinations, 13 years apart, to levels of fibrinogen on the two occasions. Fibrinogen is a strong determinant of thrombosis, and is an important risk factor for coronary heart
disease. The authors report that, in comparison with participants who never drank, those who became/stayed drinkers had smaller increases in fibrinogen, while those who quit drinking had the highest increase in fibrinogen over 13 years of follow-up. They conclude that increases in fibrinogen may be a key mechanism for the established protective effect of moderate alcohol intake on cardiovascular disease outcomes.
Forum reviewers considered this to be a potentially important paper, as the finding of higher fibrinogen among never drinkers and those who quit drinking may relate to the higher risk of coronary disease commonly reported from such subjects in prospective studies. However, the reviewers point out that the reasons that subjects quit drinking are not known (although the authors did adjust for conditions related to quitting drinking). Further, blood clotting factors other than fibrinogen may be affected by changes in alcohol consumption, and they were not reported in this paper. More research will be needed to determine if changes in fibrinogen levels are key elements in alcohol’s effects on cardiovascular disease.
Comments on this paper were provided by the following members of the International Scientific Forum on Alcohol Research:
Harvey Finkel, MD, Hematology/Oncology,
Erik Skovenborg, MD, Scandinavian Medical Alcohol Board, Practitioner,
Tedd Goldfinger, DO, FACC, Desert Cardiology of
Ian Puddey, MD, Dean, Faculty of Medicine, Dentistry & Health Sciences, University of Western Australia, Nedlands, Australia
R. Curtis Ellison, MD, Section of Preventive Medicine & Epidemiology,