- Email: [email protected]
Inverse association between habitual alcohol drinking and d-dimer in patients with type 2 diabetes mellitus
Accepted Manuscript Inverse association between habitual alcohol drinking and d-dimer in patients with type 2 diabetes mellitus Mikio Marumo, Kazumi Ekawa, Shigeyuki Ebara, Ichiro Wakabayashi PII:
S0741-8329(18)30347-1
DOI:
https://doi.org/10.1016/j.alcohol.2019.07.003
Reference:
ALC 6928
To appear in:
Alcohol
Received Date: 15 December 2018 Revised Date:
5 July 2019
Accepted Date: 6 July 2019
Please cite this article as: Marumo M., Ekawa K., Ebara S. & Wakabayashi I., Inverse association between habitual alcohol drinking and d-dimer in patients with type 2 diabetes mellitus, Alcohol (2019), doi: https://doi.org/10.1016/j.alcohol.2019.07.003. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Inverse association between habitual alcohol drinking and d-dimer in
RI PT
patients with type 2 diabetes mellitus
Author:
SC
Mikio Marumo, Kazumi Ekawa, Shigeyuki Ebara, and Ichiro Wakabayashi
Affiliation:
M AN U
Department of Environmental and Preventive Medicine, Hyogo College of Medicine, Nishinomiya, Japan
Correspondence:
TE D
Ichiro Wakabayashi, MD & PhD
Department of Environmental and Preventive Medicine,
EP
Hyogo College of Medicine,
Mukogawa-cho 1-1, Nishinomiya, Hyogo 663-8501, Japan
AC C
Tel: +81-798-45-6561; Fax: +81-798-45-6563 E-mail: [email protected]
1
ACCEPTED MANUSCRIPT
Abstract Alcohol is known to inhibit blood coagulation. Patients with diabetes mellitus are prone
RI PT
to show hypercoagulability. However, it remains to be clarified whether and how habitual alcohol drinking affects coagulability in patients with diabetes. The purpose of this study was to determine the relationship between alcohol intake and d-dimer, a
SC
sensitive marker of blood coagulation, in patients with diabetes. We investigated the relationship between alcohol intake and d-dimer in plasma of 269 patients with type 2
M AN U
diabetes by using analysis of covariance and logistic regression analysis after adjustment for age, gender, body mass index, hemoglobin A1c, and histories of smoking and anti-coagulation therapy. Log-transformed d-dimer and HDL cholesterol were significantly lower and higher, respectively, in regular drinkers than in nondrinkers,
TE D
while there were no significant differences in log-transformed d-dimer and HDL cholesterol in occasional drinkers and nondrinkers. Odds ratios of regular drinkers vs.
EP
nondrinkers for high d-dimer (0.46 [0.21-0.98]) and low HDL cholesterol (0.20 [0.08-0.50]) were significantly lower than the reference level, while the odds ratios of
AC C
occasional drinkers for high d-dimer (1.24 [0.41-3.73] and low HDL cholesterol (0.43 [0.15-1.25]) were not significantly different from the reference level. HDL cholesterol showed a significant inverse correlation with log-transformed d-dimer both in overall subjects and in nondrinkers. Regular drinking, but not occasional drinking, was associated with lower d-dimer levels, suggesting that habitual alcohol drinking suppresses hypercoagulability in patients with diabetes. There is an alcohol intake-independent inverse association between HDL cholesterol and d-dimer.
2
ACCEPTED MANUSCRIPT
Keywords: Alcohol, Blood coagulation, Cardiovascular disease, Diabetes mellitus,
AC C
EP
TE D
M AN U
SC
RI PT
D-dimer, HDL cholesterol.
3
ACCEPTED MANUSCRIPT
Introduction Habitual alcohol drinking has both beneficial and harmful effects on cardiovascular
RI PT
health depending on the amount of alcohol intake: Light-to-moderate drinkers are known to have lower risks of coronary artery disease and ischemic stroke than nondrinkers, while the risk of hemorrhagic stroke is increased in heavy drinkers (Corrao
SC
et al., 2000; Reynolds et al., 2003). The beneficial effect of alcohol is mainly explained by elevation of blood HDL cholesterol levels (Ellison et al., 2004). Alcohol-induced
M AN U
suppression of blood coagulability is also involved in the beneficial effect of alcohol on the cardiovascular risk, and this action of alcohol is thought to be due to inhibition of platelet aggregation (Rubin & Rand, 1994) and reduction in the levels of coagulation
2001).
TE D
factors including fibrinogen (Krobot et al., 1992; Meade et al., 1987; Mukamal et al.,
Diabetes mellitus is a major risk factor for cardiovascular disease, and
EP
thromboatherosclerotic complication is a major determinant for the prognosis of patients with diabetes (Beckman & Creager, 2016; Creager et al., 2003). In patients with
AC C
diabetes, blood coagulation and fibrinolysis are prone to be increased and attenuated, respectively: platelets are easily activated under the condition of hyperglycemia, and hyper-insulinemia facilitates the synthesis of prothrombotic factors, including fibrinogen and PAI-1, in the liver (Alzahrani & Ajjan, 2010). Similar to the findings in a general population, light-to-moderate drinking has been shown to reduce the cardiovascular risk in patients with diabetes (Polsky & Akturk, 2017; van de Wiel, 2004; Wakabayashi et al., 2002). However, it remains unknown whether and how
4
ACCEPTED MANUSCRIPT
habitual alcohol drinking affects blood coagulability in patients with diabetes. The purpose of this concise study was to clarify the relationship between habitual
RI PT
alcohol drinking and coagulation status in patients with type 2 diabetes. D-dimer is a sensitive marker for blood hypercoagulability (Olson, 2015; Tripodi, 2011), and thus we investigated the relationship between alcohol intake and plasma d-dimer levels in
M AN U
SC
diabetes patients.
Methods Subjects
The subjects of this study were 165 male and 104 female outpatients who had been
TE D
diagnosed as having type 2 diabetes mellitus. This study was approved by the ethics committees of Kobe Tokushukai Hospital (number: TGE00313-014) and Hyogo
EP
College of Medicine (number: 1766). Individual histories of medication, cigarette smoking and alcohol drinking were surveyed by questionnaires. The subjects were
AC C
divided by average alcohol consumption into three groups (nondrinkers; occasional drinkers, less than 2 days per week; regular drinkers, 2 days or more per week). The subjects were also divided into four groups by average cigarette consumption (nonsmokers; ex smokers; light smokers, < 20 cigarettes per day; heavy smokers, ≥ 20). The median with 25 and 75 percentile values (in parenthesis) of the average amount of daily ethanol consumption (ml) by regular drinkers was 31.8 (17.5, 45.0). Regarding the types of alcohol beverages in the regular drinker group, the proportions of beer
5
ACCEPTED MANUSCRIPT
drinkers, sake (rice wine) drinkers, shochu (traditional Japanese distilled spirit) drinkers, wine drinkers and whisky drinkers were 57.5% (n = 50), 16.1% (n = 14), 29.9% (n =
RI PT
26), 4.6% (n = 4) and 1.1% (n = 1), respectively. Alcohol is known to show anticoagulant and procoagulant effects depending on the quantity and type of alcoholic beverage (Dimmit et al., 1998). However, we did not perform comparison of each
SC
variable among subgroups according to the type of alcohol beverage because the
M AN U
number of regular drinkers (n = 87) was not large enough for analysis.
Measurements
Height and body weight were measured with each subject wearing light clothes at a health checkup. Body mass index (BMI) was calculated as weight in kilograms divided
TE D
by the square of height in meters.
Fasted blood was collected from each patient in the morning, and serum and plasma
EP
were obtained. Serum HDL cholesterol, LDL cholesterol, total cholesterol and ALT (alanine aminotransferase) concentrations were measured by enzymatic methods using
AC C
commercial kits. Abnormally low HDL cholesterol was defined as < 40 mg/dl for men and < 50 mg/dl for women. Abnormally high ALT was defined as > 45 IU/L. Hemoglobin A1c was measured by using an automatic glycol-hemoglobin analyzer based on high-performance liquid chromatography. Since the standards of hemoglobin A1c used for measurement are different in the NGSP (National Glycohemoglobin Standardization Program) method and the JDS (Japan Diabetes Society) method, hemoglobin A1c values were calibrated by using a formula proposed by the JDS
6
ACCEPTED MANUSCRIPT
(Kashiwagi et al., 2012): hemoglobin A1c (NGSP) (%) = 1.02 x hemoglobin A1c (JDS) (%) + 0.25%. Subjects with diabetes were defined as those receiving drug therapy for
RI PT
diabetes and/or those showing high hemoglobin A1c levels (≥ 6.5%), according to the criteria for diagnosis of diabetes by the American Diabetes Association (Anonymous, 2010). C-reactive protein (CRP) in serum was measured with a latex agglutination assay
SC
using a commercial kit (CRP latex X2, Denka Seiken Co., Ltd., Tokyo, Japan). An abnormally high level of CRP was defined as ≥ 0.3 mg/dl.
M AN U
Plasma fibrinogen concentration was measured by evaluating thrombin clotting time using a commercial kit (Thrombocheck Fib [Sysmex, Kobe, Japan]). D-dimer in plasma was measured with a latex agglutination assay using a commercial kit (Rapid chip d-dimer, Sekisui Medical Co., Ltd, Tokyo, Japan). Abnormally high levels of
EP
Statistical analysis
TE D
fibrinogen and d-dimer were defined as ≥ 400 mg/dl and ≥ 1.0 µg/ml, respectively.
Statistical analyses were performed using a computer software program (SPSS
AC C
version 16.0 J for Windows, Chicago IL, USA). Means of each variable were compared among non-, occasional and regular drinkers by using analysis of variance (ANOVA) followed by Scheffé’s F-test in univariate analysis and analysis of covariance (ANCOVA) followed by Student’s t-test after Bonferroni correction in multivariate analysis. Since d-dimer levels did not show a normal distribution, their medians were compared non-parametrically in univariate analysis by using the Kruskal-Wallis test followed by the Steel-Dwass test. In multivariate analysis (ANCOVA), d-dimer levels
7
ACCEPTED MANUSCRIPT
were compared after log transformation. Categorical variables were compared using the chi-squared test. In logistic regression analysis, odds ratios for low HDL cholesterol,
RI PT
high fibrinogen, or high d-dimer were estimated in the occasional and regular drinker groups vs. the nondrinkers group. In ANCOVA and logistic regression analysis, age, gender, BMI, hemoglobin A1c, history of smoking and history of anti-dyslipidemic
SC
therapy (for HDL cholesterol) or anti-coagulant medication therapy (for fibrinogen and d-dimer) were used as other explanatory variables. Uni- and multivariate linear
M AN U
regression analyses using Pearson’s correlation coefficient and standardized regression coefficient, respectively, were performed to test the relationship between HDL cholesterol and d-dimer. Probability (p) values less than 0.05 were defined as
EP
Results
TE D
significant.
Comparison of variables among non-, occasional and regular drinkers in univariate
AC C
analysis
Table 1 shows the characteristics of each drinker group and overall subjects. The
percentage of men was significantly higher in regular drinkers than in nondrinkers. The percentage of smokers was significantly higher in occasional and regular drinkers than in nondrinkers. The prevalence of arterial disease (at least one of histories of ischemic heart disease, stroke and peripheral arterial disease) was significantly lower in regular drinkers than in nondrinkers. There were no significant differences in ALT levels and
8
ACCEPTED MANUSCRIPT
frequencies of high ALT levels among non-, occasional and regular drinkers. HDL cholesterol and d-dimer levels were significantly higher and lower, respectively, in
RI PT
regular drinkers than in nondrinkers, while there were no significant differences in those levels between occasional drinkers and nondrinkers. The percentages of subjects with low HDL cholesterol and high d-dimer were significantly lower in regular drinkers than
SC
in nondrinkers, while there were no significant differences in those percentages between occasional drinkers and nondrinkers. LDL cholesterol levels were comparable in the
M AN U
non-, occasional and regular drinker groups. Although not statistically significant, total cholesterol level was about 9 ~ 10 mg/dl higher in regular drinkers than in non- and occasional drinkers, which might reflect the differences in HDL cholesterol levels in non- and occasional drinkers and in non- and regular drinkers. Age, BMI, hemoglobin
TE D
A1c level and fibrinogen level were not significantly different among non-, occasional and regular drinkers. The percentage of subjects showing high CRP was also not
EP
significantly different among non-, occasional and regular drinkers.
AC C
Comparison of HDL cholesterol, fibrinogen and d-dimer levels among nondrinkers, occasional drinkers and regular drinkers in ANCOVA Mean levels of HDL cholesterol, fibrinogen and log-transformed d-dimer were
compared among nondrinkers, occasional drinkers and regular drinkers after adjustment for age, gender, BMI, hemoglobin A1c, history of smoking and history of anti-dyslipidemic or anti-coagulant medication therapy (Table 2, upper lines). As was found in univariate analysis, HDL cholesterol and log-transformed d-dimer levels were
9
ACCEPTED MANUSCRIPT
significantly higher and lower, respectively, in regular drinkers than in nondrinkers, while there were no differences in those levels between occasional and nondrinkers.
RI PT
Fibrinogen was not significantly different among non-, occasional and regular drinkers.
Relationships of alcohol intake with low HDL cholesterol, high fibrinogen and high
SC
d-dimer evaluated by using multivariate logistic regression analysis
Odds ratios for low HDL cholesterol, high fibrinogen and high d-dimer of
M AN U
occasional and regular drinkers vs. nondrinkers were estimated (Table 2, lower lines). The odds ratios of regular drinkers vs. nondrinkers for low HDL cholesterol and high d-dimer were significantly lower than the reference level of 1.00, while those odds ratios of occasional drinkers vs. nondrinkers were not significantly different from the
TE D
reference level. Odds ratios for high fibrinogen of occasional and regular drinkers vs.
EP
nondrinkers were not significantly different from the reference level.
Relationship between HDL cholesterol and d-dimer
AC C
Scatter plots for the relationship between HDL cholesterol and d-dimer are shown in Figure 1. Both in overall subjects and nondrinkers, there were significant inverse correlations between HDL cholesterol and log-transformed d-dimer. In multivariate analysis using age, gender, BMI, hemoglobin A1c, and histories of smoking, therapy for dyslipidemia and anticoagulation therapy, there were also significant inverse correlations between HDL cholesterol and d-dimer in overall drinkers and in nondrinkers (standardized regression coefficient: overall subjects, -0.369 [p < 0.01];
10
ACCEPTED MANUSCRIPT
RI PT
nondrinkers, -0.319 [p < 0.01]).
Discussion
The results of ANCOVA and logistic regression analysis indicate that habitual
SC
alcohol drinking is associated with a lower d-dimer level in patients with diabetes. High d-dimer reflects a state of hypercoagulation and consequent activation of fibrinolysis.
M AN U
Therefore, habitual alcohol intake is thought to lower the risk of cardiovascular disease partly through suppressing blood coagulation. This agrees with previously reported results showing lower cardiovascular risk in light-to-moderate drinkers in patients with diabetes (Polsky & Akturk, 2017; van de Wiel, 2004; Wakabayashi et al., 2002) as well
TE D
as in a general population (Corrao et al., 2000; Reynolds et al., 2003). To the best of our knowledge, the present study is the first study demonstrating an inverse association
EP
between alcohol intake and d-dimer in diabetes patients. The results of this study agree with the results of a recent study showing that d-dimer level decreased by 24% after
AC C
consumption of 30 g of alcohol for 8 weeks in healthy postmenopausal women (Stote et al., 2016).
The prevalence of arterial disease was significantly lower in regular drinkers than
in nondrinkers, agreeing with the inverse association between regular drinking and d-dimer in patients with diabetes since high d-dimer levels reflecting hypercoagulation are associated with cardiovascular complications in patients with diabetes (Nwose et al., 2007). We further performed multivariate analyses (ANCOVA and logistic regression
11
ACCEPTED MANUSCRIPT
analysis) with adjustment for history of arterial disease in addition to other variables shown in Table 2: The mean log-transformed d-dimer level was significantly lower in
RI PT
regular drinkers than in nondrinkers, and the odds ratio for high d-dimer level in regular drinkers vs. nondrinkers was significantly lower than the reference level of 1.00 (data not shown). Thus, the inverse association between regular drinking and d-dimer level
SC
was not confounded by a history of arterial disease.
In previous epidemiological studies using databases for general populations, blood
M AN U
coagulation factors, including fibrinogen and factor VII, and a fibrinolysis factor, plasminogen activator, were lower and higher, respectively, in drinkers than in nondrinkers (Krobot et al., 1992; Meade et al., 1987; Mukamal et al., 2001; Ridker et al, 1994; Sumi et al, 1988). Therefore, it is speculated that the inverse association between
TE D
alcohol intake and d-dimer in patients with diabetes is due to alcohol-induced changes in coagulation and fibrinolysis factor levels. However, in the present study, the mean
EP
fibrinogen level was not significantly different in regular drinkers and nondrinkers. One possible reason for this negative result for the relationship between alcohol and
AC C
fibrinogen is that fibrinogen level also reflects acute phase reaction. However, an inverse association between regular drinking and d-dimer (mean log-transformed d-dimer [µg/ml]: nondrinkers, -0.158 ± 0.021 vs. regular drinkers, -0.263 ± 0.028 [p < 0.05]) and no association between alcohol intake and fibrinogen (fibrinogen [mg/dl]: nondrinkers, 315.8 ± 5.8 vs. regular drinkers, 312.9 ± 7.8 [p = 1.00]) were also found when adjusting for CRP, an acute phase reactant, in addition to age, gender, BMI, and histories of smoking and anti-coagulation therapy in ANCOVA. Further studies are
12
ACCEPTED MANUSCRIPT
needed to clarify the reason for a fibrinogen-independent lower d-dimer level, namely lower risk of hypercongulation status, in regular drinkers than in nondrinkers among
RI PT
patients with diabetes. HDL cholesterol is known to be higher in drinkers than in nondrinkers and is used as a marker for alcohol consumption (Berger et al, 2013). In the present study, HDL
SC
cholesterol level was significantly higher in regular drinkers but not in occasional drinkers than in nondrinkers. These results are consistent with the results for d-dimer,
M AN U
which was also significantly higher in regular drinkers but not in occasional drinkers than in nondrinkers. Therefore, beneficial effects of alcohol intake on both blood lipid metabolism and coagulation status from the viewpoint of cardiovascular risk are thought to be expected in habitual drinkers but not in opportunistic drinkers.
TE D
HDL cholesterol has been shown to be inversely associated with d-dimer in patients with diabetes (Ebara et al., 2017). Alcohol showed a positive association with HDL
EP
cholesterol and an inverse association with d-dimer in the present study, suggesting the possibility of alcohol-caused confounding for the relationship between HDL cholesterol
AC C
and d-dimer. However, there was a significant inverse association between HDL cholesterol and d-dimer in nondrinkers as well as in overall subjects (Figure 1). Therefore, the association between HDL cholesterol and d-dimer is independent of alcohol drinking. There are some limitations of this study. Because the number of the subjects in this study was small, we could not perform more detailed analysis using regular drinker categories with different amounts of alcohol intake. Thus, further studies are needed to
13
ACCEPTED MANUSCRIPT
determine whether there is a dose-dependent relationship between alcohol intake and d-dimer levels in patients with diabetes. Although some factors including age, gender,
RI PT
BMI, history of smoking, and history of anticoagulation therapy were adjusted in multivariate analyses, there are other possible confounders for the relationship between alcohol and d-dimer, e.g., diet, nutrition and socioeconomic status, for which
SC
information was not available in this study. There is a possibility that the alcohol-d-dimer relationship is modified by differences in the type of alcohol beverage.
M AN U
However, we could not perform analysis for differences in the type of beverage because the number of subjects was too small. Since this study was a cross-sectional study, further prospective studies are needed to confirm the findings of this study. In conclusion, regular drinking, but not occasional drinking, showed an inverse
TE D
association with d-dimer in patients with type 2 diabetes. This suggests that habitual alcohol drinking suppresses hypercoagulability in diabetes patients and agrees with the
EP
lower risk of atherosclerotic disease in light-to-moderate drinkers. Although a beneficial effect of habitual alcohol drinking on the blood coagulation-fibrinolysis system was
AC C
suggested by this study, it is needless to say that alcohol drinking should never be recommended to patients with diabetes since it may disturb diet therapy for diabetes and there is a danger of alcohol-induced acute hypoglycemia as well as the possibility of alcohol abuse in the future.
References 14
ACCEPTED MANUSCRIPT
Alzahrani, S.H., Ajjan, R.A. (2010). Coagulation and fibrinolysis in diabetes. Diab Vasc Dis Res, 7(4), 260-273.
diabetes mellitus. Diabetes Care, 33(Suppl 1), S62-69.
RI PT
Anonymous: American Diabetes Association (2010) Diagnosis and classification of
Beckman, J.A., Creager, M.A. (2016). Vascular Complications of Diabetes. Circ Res,
SC
118(11), 1771-1785.
Berger, D., Williams, E.C., Bryson, C.L., Rubinsky, A.D., Bradley, K.A. (2013),
M AN U
Alcohol questionnaires and HDL: screening scores as scaled markers of alcohol consumption. Alcohol, 47(6), 439-445.
Corrao, G., Rubbiati, L., Bagnardi, V., Zambon, A., Poikolainen, K. (2000). Alcohol and coronary heart disease: a meta-analysis. Addiction, 95(10), 1505-1523.
TE D
Creager, M.A., Lüscher, T.F., Cosentino, F., Beckman, J.A. (2003). Diabetes and vascular disease: pathophysiology, clinical consequences, and medical therapy:
EP
Part I. Circulation, 108(12), 1527-1532. Dimmitt, S.B., Rakic, V., Puddey, I.B., Baker, R., Oostryck, R., Adams, M.J.,
AC C
Chesterman, C.N., Burke, V., Beilin, L.J. (1998). The effects of alcohol on coagulation and fibrinolytic factors: a controlled trial. Blood Coagul Fibrinolysis, 9(1), 39-45.
Ebara, S., Marumo, M., Yamabata, C., Nishibe, I., Soneda, J.I., Mukai, J., Ohki, M., Uchida, K., Wakabayashi, I. (2017). Inverse associations of HDL cholesterol and oxidized HDL with d-dimer in patients with type 2 diabetes mellitus. Thromb Res, 155, 12-15.
15
ACCEPTED MANUSCRIPT
Ellison, R.C., Zhang, Y., Qureshi, M.M., Knox, S., Arnett, D.K., Province, M.A. (2004). Investigators of the NHLBI Family Heart Study. Lifestyle determinants of
Family Heart Study. Am Heart J, 147(3), 529-535.
RI PT
high-density lipoprotein cholesterol: the National Heart, Lung, and Blood Institute
Kashiwagi, A., Kasuga, M., Araki, E., Oka, Y., Hanafusa, T., Ito, H., Tominaga, M.,
SC
Oikawa, S., Noda, M., Kawamura, T., Sanke, T., Namba, M., Hashiramoto, M., Sasahara, T., Nishio, Y., Kuwa, K., Ueki, K., Takei, I., Umemoto, M., Murakami,
M AN U
M., Yamakado, M., Yatomi, Y., Ohashi, H.; Committee on the Standardization of Diabetes Mellitus-Related Laboratory Testing of Japan Diabetes Society. (2012). International clinical harmonization of glycated hemoglobin in Japan: From Japan Diabetes Society to National Glycohemoglobin Standardization Program values. J
TE D
Diabetes Investig, 3(1), 39-40.
Krobot, K., Hense, H.W., Cremer, P., Eberle, E., Keil, U. (1992). Determinants of
EP
plasma fibrinogen: relation to body weight, waist-to-hip ratio, smoking, alcohol, age, and sex. Results from the second MONICA Augsburg survey 1989-1990.
AC C
Arterioscler Thromb, 12(7), 780-788. Meade, T.W., Imeson, J., Stirling, Y. (1987). Effects of changes in smoking and other characteristics on clotting factors and the risk of ischaemic heart disease. Lancet,
2(8566), 986-988. Mukamal, K.J., Jadhav, P.P., D'Agostino, R.B., Massaro, J.M., Mittleman, M.A., Lipinska, I., Sutherland, P.A., Matheney, T., Levy, D., Wilson, P.W., Ellison, R.C., Silbershatz, H., Muller, J.E., Tofler, G.H. (2001). Alcohol consumption and
16
ACCEPTED MANUSCRIPT
hemostatic factors: analysis of the Framingham Offspring cohort. Circulation, 104(12), 1367-1373.
RI PT
Nwose, E.U., Richards, R.S., Jelinek, H.F., Kerr, P.G. (2007). D-dimer identifies stages in the progression of diabetes mellitus from family history of diabetes to cardiovascular complications. Pathology, 39(2), 252-257.
Clinical Applications. Adv Clin Chem, 69, 1-46.
SC
Olson, J.D. (2015). D-dimer: An Overview of Hemostasis and Fibrinolysis, Assays, and
M AN U
Polsky, S., Akturk, H.K. (2017). Alcohol Consumption, Diabetes Risk, and Cardiovascular Disease Within Diabetes. Curr Diab Rep, 17(12), 136. Reynolds, K., Lewis, B., Nolen, J.D., Kinney, G.L., Sathya, B., He, J. (2003). Alcohol consumption and risk of stroke: a meta-analysis. JAMA, 289(5), 579-588.
TE D
Ridker, P.M., Vaughan, D.E., Stampfer, M.J., Glynn, R.J., Hennekens, C.H. (1994). Association of moderate alcohol consumption and plasma concentration of
EP
endogenous tissue-type plasminogen activator. JAMA, 272(12), 929-933. Rubin, R., Rand, M.L. (1994). Alcohol and platelet function. Alcohol Clin Exp Res,
AC C
18(1), 105-110.
Stote, K.S., Tracy, R.P., Taylor, P.R., Baer, D.J. (2016). The effect of moderate alcohol consumption on biomarkers of inflammation and hemostatic factors in
postmenopausal women. Eur J Clin Nutr, 70(4), 470-474. Sumi, H., Hamada, H., Tsushima, H., Mihara, H. (1988). Urokinase-like plasminogen activator increased in plasma after alcohol drinking. Alcohol Alcohol, 23(1), 33-43.
17
ACCEPTED MANUSCRIPT
Tripodi, A. (2011). D-dimer testing in laboratory practice. Clin Chem, 57(9), 1256-1262.
RI PT
van de Wiel, A. (2004). Diabetes mellitus and alcohol. Diabetes Metab Res Rev, 20(4), 263-267.
Wakabayashi, I., Kobaba-Wakabayashi, R., Masuda, H. (2002). Relation of drinking
SC
alcohol to atherosclerotic risk in type 2 diabetes. Diabetes Care, 25(7),
AC C
EP
TE D
M AN U
1223-1228.
18
ACCEPTED MANUSCRIPT
Figure legend
RI PT
Figure 1. Scatter plots for the relationships between HDL cholesterol and d-dimer in overall subjects (A) and nondrinker subjects (B) of patients with diabetes. D-dimer levels did
SC
not show a normal distribution and were thus used after logarithmic transformation in
AC C
EP
TE D
M AN U
linear regression analysis. Pearson’s correlation coefficients (r) are shown in the figure.
19
ACCEPTED MANUSCRIPT
RI PT
Table 1. Characteristics of subjects in each drinker group and overall subjects of patients with type 2 diabetes.
AC C
EP
TE D
M AN U
SC
Nondrinkers Occasional drinkers Regular drinkers Overall subjects Number 151 31 87 269 Gender (men [%]) 48.3 67.7 81.6** 61.3 Age (years) 70.0 ± 10.9 66.6 ± 11.0 69.3 ± 9.1 69.4 ± 10.4 Current smokers (%) 49.0 54.8* 69.0** 56.1 Anti-dyslipidemia therapy (%) 51.7 58.1 39.1 48.3 Anti-coagulation therapy (%) 37.1 58.1* 36.8 39.4 Arterial disease (%) 27.2 19.4 16.1* 22.7 2 Body mass index (kg/m ) 24.9 ± 4.9 24.8 ± 3.7 24.6 ± 3.8 24.8 ± 4.5 Hemoglobin A1c (%) 7.18 ± 1.27 7.33 ± 1.58 7.07 ± 1.03 7.16 ± 1.23 High C-reactive protein (%) 16.1 21.1 20.0 18.6 ALT (IU/L) 20.00 (14.75, 29.00) 21.00 (19.00, 34.00) 20.00 (16.00, 27.25) 20.00 (15.00, 29.00) High ALT (%) 7.3 9.7 9.2 8.2 HDL cholesterol (mg/dl) 53.1 ± 17.2 54.3 ± 15.9 63.1 ± 21.2** 56.5 ± 19.0 Low HDL cholesterol (%) 33.1 16.1 8.0** 23.0 LDL cholesterol (mg/dl) 110.7 ± 34.5 110.5 ± 33.2 110.6 ± 30.1 110.6 ± 32.9 Total cholesterol (mg/dl) 189.8 ± 38.9 188.7 ± 36.1 198.1 ± 32.2 192.3 ± 36.7 Fibrinogen (mg/dl) 318.1 ± 83.2 317.0 ± 76.7 306.4 ± 81.1 314.2 ± 81.7 High fibrinogen (%) 13.9 16.1 6.9 11.9 D-dimer (µg/ml) 0.68 (0.41, 1.08) 0.58 (0.42, 1.01) 0.51 (0.38, 0.78)* 0.60 (0.40, 0.91) High d-dimer (%) 27.2 25.8 14.9* 23.0 Shown are numbers, means with standard deviations, medians with 25 and 75 percentile values, and frequencies (%) of each variable. Arterial disease, at least one of histories of ischemic heart disease, stroke and peripheral arterial disease. Asterisks denote significant differences from nondrinkers (*, p < 0.05; **, p < 0.01).
ACCEPTED MANUSCRIPT
Table 2. Relationships of frequency of alcohol drinking (non-, occasional and regular drinkers) with HDL cholesterol, fibrinogen, and
RI PT
d-dimer in patients with type 2 diabetes obtained by using ANCOVA and logistic regression analysis. Nondrinkers (n = 151)
Occasional drinkers (n = 31)
Regular drinkers (n = 87)
HDL cholesterol (mg/dl)
52.1 ± 1.4
54.0 ± 3.1
64.9 ± 1.9**
Fibrinogen (mg/dl)
315.9 ± 6.7
320.0 ± 14.7
309.0 ± 9.0
Log(d-dimer)
-0.157 ± 0.023
-0.164 ± 0.050
-0.272 ± 0.030*
M AN U
SC
Means (±SEs)
Low HDL cholesterol
1.00
High fibrinogen
1.00
High d-dimer
1.00
TE D
Odds ratios (CIs)
0.43 (0.15-1.25)
0.20 (0.08-0.50)**
1.24 (0.41-3.73)
0.47 (0.17-1.29)
1.03 (0.40-2.68)
0.46 (0.21-0.98)*
Shown are mean levels with their standard errors (SEs) of HDL cholesterol, fibrinogen and log-transformed d-dimer in each alcohol group (upper lines) and odds ratios with their confidence intervals (CIs) in parentheses of occasional or regular drinkers vs. nondrinkers
EP
for hypo-HDL cholesterolemia, hyper-fibrinogenemia, and high d-dimer (lower lines). Age, gender, BMI, hemoglobin A1c, history of
AC C
smoking and history of medication therapy for dyslipidemia or anti-coagulant therapy were adjusted in ANCOVA and logistic regression analysis. Asterisks denote significant differences (*, p < 0.05; **, p < 0.01) from nondrinkers in ANCOVA or from the reference level of 1.00 in logistic regression analysis.
ACCEPTED MANUSCRIPT
r = -0.340 (p < 0.01)
AC C
EP
TE D
M AN U
r = -0.381 (p < 0.01)
SC
RI PT
Figure 1
Figure. 1. Scatter plots for the relationships between HDL cholesterol and d-dimer in overall subjects (A) and nondrinker subjects (B) of patients with diabetes. D-dimer levels did not show a normal distribution and were thus used after logarithmic transformation in linear regression analysis. Pearson’s correlation coefficients (r) are shown in the figure.
ACCEPTED MANUSCRIPT
Highlights
It remains unknown whether alcohol drinking affects coagulability in patients with
The relationship between alcohol intake and d-dimer was investigated in patients with diabetes.
Regular drinking, but not occasional drinking, was associated with lower d-dimer
M AN U
and higher HDL cholesterol levels.
SC
RI PT
diabetes.
D-dimer was inversely associated with HDL cholesterol independently of alcohol.
Habitual alcohol drinking may suppress hypercoagulability in patients with
AC C
EP
TE D
diabetes.
S0741-8329(18)30347-1
DOI:
https://doi.org/10.1016/j.alcohol.2019.07.003
Reference:
ALC 6928
To appear in:
Alcohol
Received Date: 15 December 2018 Revised Date:
5 July 2019
Accepted Date: 6 July 2019
Please cite this article as: Marumo M., Ekawa K., Ebara S. & Wakabayashi I., Inverse association between habitual alcohol drinking and d-dimer in patients with type 2 diabetes mellitus, Alcohol (2019), doi: https://doi.org/10.1016/j.alcohol.2019.07.003. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Inverse association between habitual alcohol drinking and d-dimer in
RI PT
patients with type 2 diabetes mellitus
Author:
SC
Mikio Marumo, Kazumi Ekawa, Shigeyuki Ebara, and Ichiro Wakabayashi
Affiliation:
M AN U
Department of Environmental and Preventive Medicine, Hyogo College of Medicine, Nishinomiya, Japan
Correspondence:
TE D
Ichiro Wakabayashi, MD & PhD
Department of Environmental and Preventive Medicine,
EP
Hyogo College of Medicine,
Mukogawa-cho 1-1, Nishinomiya, Hyogo 663-8501, Japan
AC C
Tel: +81-798-45-6561; Fax: +81-798-45-6563 E-mail: [email protected]
1
ACCEPTED MANUSCRIPT
Abstract Alcohol is known to inhibit blood coagulation. Patients with diabetes mellitus are prone
RI PT
to show hypercoagulability. However, it remains to be clarified whether and how habitual alcohol drinking affects coagulability in patients with diabetes. The purpose of this study was to determine the relationship between alcohol intake and d-dimer, a
SC
sensitive marker of blood coagulation, in patients with diabetes. We investigated the relationship between alcohol intake and d-dimer in plasma of 269 patients with type 2
M AN U
diabetes by using analysis of covariance and logistic regression analysis after adjustment for age, gender, body mass index, hemoglobin A1c, and histories of smoking and anti-coagulation therapy. Log-transformed d-dimer and HDL cholesterol were significantly lower and higher, respectively, in regular drinkers than in nondrinkers,
TE D
while there were no significant differences in log-transformed d-dimer and HDL cholesterol in occasional drinkers and nondrinkers. Odds ratios of regular drinkers vs.
EP
nondrinkers for high d-dimer (0.46 [0.21-0.98]) and low HDL cholesterol (0.20 [0.08-0.50]) were significantly lower than the reference level, while the odds ratios of
AC C
occasional drinkers for high d-dimer (1.24 [0.41-3.73] and low HDL cholesterol (0.43 [0.15-1.25]) were not significantly different from the reference level. HDL cholesterol showed a significant inverse correlation with log-transformed d-dimer both in overall subjects and in nondrinkers. Regular drinking, but not occasional drinking, was associated with lower d-dimer levels, suggesting that habitual alcohol drinking suppresses hypercoagulability in patients with diabetes. There is an alcohol intake-independent inverse association between HDL cholesterol and d-dimer.
2
ACCEPTED MANUSCRIPT
Keywords: Alcohol, Blood coagulation, Cardiovascular disease, Diabetes mellitus,
AC C
EP
TE D
M AN U
SC
RI PT
D-dimer, HDL cholesterol.
3
ACCEPTED MANUSCRIPT
Introduction Habitual alcohol drinking has both beneficial and harmful effects on cardiovascular
RI PT
health depending on the amount of alcohol intake: Light-to-moderate drinkers are known to have lower risks of coronary artery disease and ischemic stroke than nondrinkers, while the risk of hemorrhagic stroke is increased in heavy drinkers (Corrao
SC
et al., 2000; Reynolds et al., 2003). The beneficial effect of alcohol is mainly explained by elevation of blood HDL cholesterol levels (Ellison et al., 2004). Alcohol-induced
M AN U
suppression of blood coagulability is also involved in the beneficial effect of alcohol on the cardiovascular risk, and this action of alcohol is thought to be due to inhibition of platelet aggregation (Rubin & Rand, 1994) and reduction in the levels of coagulation
2001).
TE D
factors including fibrinogen (Krobot et al., 1992; Meade et al., 1987; Mukamal et al.,
Diabetes mellitus is a major risk factor for cardiovascular disease, and
EP
thromboatherosclerotic complication is a major determinant for the prognosis of patients with diabetes (Beckman & Creager, 2016; Creager et al., 2003). In patients with
AC C
diabetes, blood coagulation and fibrinolysis are prone to be increased and attenuated, respectively: platelets are easily activated under the condition of hyperglycemia, and hyper-insulinemia facilitates the synthesis of prothrombotic factors, including fibrinogen and PAI-1, in the liver (Alzahrani & Ajjan, 2010). Similar to the findings in a general population, light-to-moderate drinking has been shown to reduce the cardiovascular risk in patients with diabetes (Polsky & Akturk, 2017; van de Wiel, 2004; Wakabayashi et al., 2002). However, it remains unknown whether and how
4
ACCEPTED MANUSCRIPT
habitual alcohol drinking affects blood coagulability in patients with diabetes. The purpose of this concise study was to clarify the relationship between habitual
RI PT
alcohol drinking and coagulation status in patients with type 2 diabetes. D-dimer is a sensitive marker for blood hypercoagulability (Olson, 2015; Tripodi, 2011), and thus we investigated the relationship between alcohol intake and plasma d-dimer levels in
M AN U
SC
diabetes patients.
Methods Subjects
The subjects of this study were 165 male and 104 female outpatients who had been
TE D
diagnosed as having type 2 diabetes mellitus. This study was approved by the ethics committees of Kobe Tokushukai Hospital (number: TGE00313-014) and Hyogo
EP
College of Medicine (number: 1766). Individual histories of medication, cigarette smoking and alcohol drinking were surveyed by questionnaires. The subjects were
AC C
divided by average alcohol consumption into three groups (nondrinkers; occasional drinkers, less than 2 days per week; regular drinkers, 2 days or more per week). The subjects were also divided into four groups by average cigarette consumption (nonsmokers; ex smokers; light smokers, < 20 cigarettes per day; heavy smokers, ≥ 20). The median with 25 and 75 percentile values (in parenthesis) of the average amount of daily ethanol consumption (ml) by regular drinkers was 31.8 (17.5, 45.0). Regarding the types of alcohol beverages in the regular drinker group, the proportions of beer
5
ACCEPTED MANUSCRIPT
drinkers, sake (rice wine) drinkers, shochu (traditional Japanese distilled spirit) drinkers, wine drinkers and whisky drinkers were 57.5% (n = 50), 16.1% (n = 14), 29.9% (n =
RI PT
26), 4.6% (n = 4) and 1.1% (n = 1), respectively. Alcohol is known to show anticoagulant and procoagulant effects depending on the quantity and type of alcoholic beverage (Dimmit et al., 1998). However, we did not perform comparison of each
SC
variable among subgroups according to the type of alcohol beverage because the
M AN U
number of regular drinkers (n = 87) was not large enough for analysis.
Measurements
Height and body weight were measured with each subject wearing light clothes at a health checkup. Body mass index (BMI) was calculated as weight in kilograms divided
TE D
by the square of height in meters.
Fasted blood was collected from each patient in the morning, and serum and plasma
EP
were obtained. Serum HDL cholesterol, LDL cholesterol, total cholesterol and ALT (alanine aminotransferase) concentrations were measured by enzymatic methods using
AC C
commercial kits. Abnormally low HDL cholesterol was defined as < 40 mg/dl for men and < 50 mg/dl for women. Abnormally high ALT was defined as > 45 IU/L. Hemoglobin A1c was measured by using an automatic glycol-hemoglobin analyzer based on high-performance liquid chromatography. Since the standards of hemoglobin A1c used for measurement are different in the NGSP (National Glycohemoglobin Standardization Program) method and the JDS (Japan Diabetes Society) method, hemoglobin A1c values were calibrated by using a formula proposed by the JDS
6
ACCEPTED MANUSCRIPT
(Kashiwagi et al., 2012): hemoglobin A1c (NGSP) (%) = 1.02 x hemoglobin A1c (JDS) (%) + 0.25%. Subjects with diabetes were defined as those receiving drug therapy for
RI PT
diabetes and/or those showing high hemoglobin A1c levels (≥ 6.5%), according to the criteria for diagnosis of diabetes by the American Diabetes Association (Anonymous, 2010). C-reactive protein (CRP) in serum was measured with a latex agglutination assay
SC
using a commercial kit (CRP latex X2, Denka Seiken Co., Ltd., Tokyo, Japan). An abnormally high level of CRP was defined as ≥ 0.3 mg/dl.
M AN U
Plasma fibrinogen concentration was measured by evaluating thrombin clotting time using a commercial kit (Thrombocheck Fib [Sysmex, Kobe, Japan]). D-dimer in plasma was measured with a latex agglutination assay using a commercial kit (Rapid chip d-dimer, Sekisui Medical Co., Ltd, Tokyo, Japan). Abnormally high levels of
EP
Statistical analysis
TE D
fibrinogen and d-dimer were defined as ≥ 400 mg/dl and ≥ 1.0 µg/ml, respectively.
Statistical analyses were performed using a computer software program (SPSS
AC C
version 16.0 J for Windows, Chicago IL, USA). Means of each variable were compared among non-, occasional and regular drinkers by using analysis of variance (ANOVA) followed by Scheffé’s F-test in univariate analysis and analysis of covariance (ANCOVA) followed by Student’s t-test after Bonferroni correction in multivariate analysis. Since d-dimer levels did not show a normal distribution, their medians were compared non-parametrically in univariate analysis by using the Kruskal-Wallis test followed by the Steel-Dwass test. In multivariate analysis (ANCOVA), d-dimer levels
7
ACCEPTED MANUSCRIPT
were compared after log transformation. Categorical variables were compared using the chi-squared test. In logistic regression analysis, odds ratios for low HDL cholesterol,
RI PT
high fibrinogen, or high d-dimer were estimated in the occasional and regular drinker groups vs. the nondrinkers group. In ANCOVA and logistic regression analysis, age, gender, BMI, hemoglobin A1c, history of smoking and history of anti-dyslipidemic
SC
therapy (for HDL cholesterol) or anti-coagulant medication therapy (for fibrinogen and d-dimer) were used as other explanatory variables. Uni- and multivariate linear
M AN U
regression analyses using Pearson’s correlation coefficient and standardized regression coefficient, respectively, were performed to test the relationship between HDL cholesterol and d-dimer. Probability (p) values less than 0.05 were defined as
EP
Results
TE D
significant.
Comparison of variables among non-, occasional and regular drinkers in univariate
AC C
analysis
Table 1 shows the characteristics of each drinker group and overall subjects. The
percentage of men was significantly higher in regular drinkers than in nondrinkers. The percentage of smokers was significantly higher in occasional and regular drinkers than in nondrinkers. The prevalence of arterial disease (at least one of histories of ischemic heart disease, stroke and peripheral arterial disease) was significantly lower in regular drinkers than in nondrinkers. There were no significant differences in ALT levels and
8
ACCEPTED MANUSCRIPT
frequencies of high ALT levels among non-, occasional and regular drinkers. HDL cholesterol and d-dimer levels were significantly higher and lower, respectively, in
RI PT
regular drinkers than in nondrinkers, while there were no significant differences in those levels between occasional drinkers and nondrinkers. The percentages of subjects with low HDL cholesterol and high d-dimer were significantly lower in regular drinkers than
SC
in nondrinkers, while there were no significant differences in those percentages between occasional drinkers and nondrinkers. LDL cholesterol levels were comparable in the
M AN U
non-, occasional and regular drinker groups. Although not statistically significant, total cholesterol level was about 9 ~ 10 mg/dl higher in regular drinkers than in non- and occasional drinkers, which might reflect the differences in HDL cholesterol levels in non- and occasional drinkers and in non- and regular drinkers. Age, BMI, hemoglobin
TE D
A1c level and fibrinogen level were not significantly different among non-, occasional and regular drinkers. The percentage of subjects showing high CRP was also not
EP
significantly different among non-, occasional and regular drinkers.
AC C
Comparison of HDL cholesterol, fibrinogen and d-dimer levels among nondrinkers, occasional drinkers and regular drinkers in ANCOVA Mean levels of HDL cholesterol, fibrinogen and log-transformed d-dimer were
compared among nondrinkers, occasional drinkers and regular drinkers after adjustment for age, gender, BMI, hemoglobin A1c, history of smoking and history of anti-dyslipidemic or anti-coagulant medication therapy (Table 2, upper lines). As was found in univariate analysis, HDL cholesterol and log-transformed d-dimer levels were
9
ACCEPTED MANUSCRIPT
significantly higher and lower, respectively, in regular drinkers than in nondrinkers, while there were no differences in those levels between occasional and nondrinkers.
RI PT
Fibrinogen was not significantly different among non-, occasional and regular drinkers.
Relationships of alcohol intake with low HDL cholesterol, high fibrinogen and high
SC
d-dimer evaluated by using multivariate logistic regression analysis
Odds ratios for low HDL cholesterol, high fibrinogen and high d-dimer of
M AN U
occasional and regular drinkers vs. nondrinkers were estimated (Table 2, lower lines). The odds ratios of regular drinkers vs. nondrinkers for low HDL cholesterol and high d-dimer were significantly lower than the reference level of 1.00, while those odds ratios of occasional drinkers vs. nondrinkers were not significantly different from the
TE D
reference level. Odds ratios for high fibrinogen of occasional and regular drinkers vs.
EP
nondrinkers were not significantly different from the reference level.
Relationship between HDL cholesterol and d-dimer
AC C
Scatter plots for the relationship between HDL cholesterol and d-dimer are shown in Figure 1. Both in overall subjects and nondrinkers, there were significant inverse correlations between HDL cholesterol and log-transformed d-dimer. In multivariate analysis using age, gender, BMI, hemoglobin A1c, and histories of smoking, therapy for dyslipidemia and anticoagulation therapy, there were also significant inverse correlations between HDL cholesterol and d-dimer in overall drinkers and in nondrinkers (standardized regression coefficient: overall subjects, -0.369 [p < 0.01];
10
ACCEPTED MANUSCRIPT
RI PT
nondrinkers, -0.319 [p < 0.01]).
Discussion
The results of ANCOVA and logistic regression analysis indicate that habitual
SC
alcohol drinking is associated with a lower d-dimer level in patients with diabetes. High d-dimer reflects a state of hypercoagulation and consequent activation of fibrinolysis.
M AN U
Therefore, habitual alcohol intake is thought to lower the risk of cardiovascular disease partly through suppressing blood coagulation. This agrees with previously reported results showing lower cardiovascular risk in light-to-moderate drinkers in patients with diabetes (Polsky & Akturk, 2017; van de Wiel, 2004; Wakabayashi et al., 2002) as well
TE D
as in a general population (Corrao et al., 2000; Reynolds et al., 2003). To the best of our knowledge, the present study is the first study demonstrating an inverse association
EP
between alcohol intake and d-dimer in diabetes patients. The results of this study agree with the results of a recent study showing that d-dimer level decreased by 24% after
AC C
consumption of 30 g of alcohol for 8 weeks in healthy postmenopausal women (Stote et al., 2016).
The prevalence of arterial disease was significantly lower in regular drinkers than
in nondrinkers, agreeing with the inverse association between regular drinking and d-dimer in patients with diabetes since high d-dimer levels reflecting hypercoagulation are associated with cardiovascular complications in patients with diabetes (Nwose et al., 2007). We further performed multivariate analyses (ANCOVA and logistic regression
11
ACCEPTED MANUSCRIPT
analysis) with adjustment for history of arterial disease in addition to other variables shown in Table 2: The mean log-transformed d-dimer level was significantly lower in
RI PT
regular drinkers than in nondrinkers, and the odds ratio for high d-dimer level in regular drinkers vs. nondrinkers was significantly lower than the reference level of 1.00 (data not shown). Thus, the inverse association between regular drinking and d-dimer level
SC
was not confounded by a history of arterial disease.
In previous epidemiological studies using databases for general populations, blood
M AN U
coagulation factors, including fibrinogen and factor VII, and a fibrinolysis factor, plasminogen activator, were lower and higher, respectively, in drinkers than in nondrinkers (Krobot et al., 1992; Meade et al., 1987; Mukamal et al., 2001; Ridker et al, 1994; Sumi et al, 1988). Therefore, it is speculated that the inverse association between
TE D
alcohol intake and d-dimer in patients with diabetes is due to alcohol-induced changes in coagulation and fibrinolysis factor levels. However, in the present study, the mean
EP
fibrinogen level was not significantly different in regular drinkers and nondrinkers. One possible reason for this negative result for the relationship between alcohol and
AC C
fibrinogen is that fibrinogen level also reflects acute phase reaction. However, an inverse association between regular drinking and d-dimer (mean log-transformed d-dimer [µg/ml]: nondrinkers, -0.158 ± 0.021 vs. regular drinkers, -0.263 ± 0.028 [p < 0.05]) and no association between alcohol intake and fibrinogen (fibrinogen [mg/dl]: nondrinkers, 315.8 ± 5.8 vs. regular drinkers, 312.9 ± 7.8 [p = 1.00]) were also found when adjusting for CRP, an acute phase reactant, in addition to age, gender, BMI, and histories of smoking and anti-coagulation therapy in ANCOVA. Further studies are
12
ACCEPTED MANUSCRIPT
needed to clarify the reason for a fibrinogen-independent lower d-dimer level, namely lower risk of hypercongulation status, in regular drinkers than in nondrinkers among
RI PT
patients with diabetes. HDL cholesterol is known to be higher in drinkers than in nondrinkers and is used as a marker for alcohol consumption (Berger et al, 2013). In the present study, HDL
SC
cholesterol level was significantly higher in regular drinkers but not in occasional drinkers than in nondrinkers. These results are consistent with the results for d-dimer,
M AN U
which was also significantly higher in regular drinkers but not in occasional drinkers than in nondrinkers. Therefore, beneficial effects of alcohol intake on both blood lipid metabolism and coagulation status from the viewpoint of cardiovascular risk are thought to be expected in habitual drinkers but not in opportunistic drinkers.
TE D
HDL cholesterol has been shown to be inversely associated with d-dimer in patients with diabetes (Ebara et al., 2017). Alcohol showed a positive association with HDL
EP
cholesterol and an inverse association with d-dimer in the present study, suggesting the possibility of alcohol-caused confounding for the relationship between HDL cholesterol
AC C
and d-dimer. However, there was a significant inverse association between HDL cholesterol and d-dimer in nondrinkers as well as in overall subjects (Figure 1). Therefore, the association between HDL cholesterol and d-dimer is independent of alcohol drinking. There are some limitations of this study. Because the number of the subjects in this study was small, we could not perform more detailed analysis using regular drinker categories with different amounts of alcohol intake. Thus, further studies are needed to
13
ACCEPTED MANUSCRIPT
determine whether there is a dose-dependent relationship between alcohol intake and d-dimer levels in patients with diabetes. Although some factors including age, gender,
RI PT
BMI, history of smoking, and history of anticoagulation therapy were adjusted in multivariate analyses, there are other possible confounders for the relationship between alcohol and d-dimer, e.g., diet, nutrition and socioeconomic status, for which
SC
information was not available in this study. There is a possibility that the alcohol-d-dimer relationship is modified by differences in the type of alcohol beverage.
M AN U
However, we could not perform analysis for differences in the type of beverage because the number of subjects was too small. Since this study was a cross-sectional study, further prospective studies are needed to confirm the findings of this study. In conclusion, regular drinking, but not occasional drinking, showed an inverse
TE D
association with d-dimer in patients with type 2 diabetes. This suggests that habitual alcohol drinking suppresses hypercoagulability in diabetes patients and agrees with the
EP
lower risk of atherosclerotic disease in light-to-moderate drinkers. Although a beneficial effect of habitual alcohol drinking on the blood coagulation-fibrinolysis system was
AC C
suggested by this study, it is needless to say that alcohol drinking should never be recommended to patients with diabetes since it may disturb diet therapy for diabetes and there is a danger of alcohol-induced acute hypoglycemia as well as the possibility of alcohol abuse in the future.
References 14
ACCEPTED MANUSCRIPT
Alzahrani, S.H., Ajjan, R.A. (2010). Coagulation and fibrinolysis in diabetes. Diab Vasc Dis Res, 7(4), 260-273.
diabetes mellitus. Diabetes Care, 33(Suppl 1), S62-69.
RI PT
Anonymous: American Diabetes Association (2010) Diagnosis and classification of
Beckman, J.A., Creager, M.A. (2016). Vascular Complications of Diabetes. Circ Res,
SC
118(11), 1771-1785.
Berger, D., Williams, E.C., Bryson, C.L., Rubinsky, A.D., Bradley, K.A. (2013),
M AN U
Alcohol questionnaires and HDL: screening scores as scaled markers of alcohol consumption. Alcohol, 47(6), 439-445.
Corrao, G., Rubbiati, L., Bagnardi, V., Zambon, A., Poikolainen, K. (2000). Alcohol and coronary heart disease: a meta-analysis. Addiction, 95(10), 1505-1523.
TE D
Creager, M.A., Lüscher, T.F., Cosentino, F., Beckman, J.A. (2003). Diabetes and vascular disease: pathophysiology, clinical consequences, and medical therapy:
EP
Part I. Circulation, 108(12), 1527-1532. Dimmitt, S.B., Rakic, V., Puddey, I.B., Baker, R., Oostryck, R., Adams, M.J.,
AC C
Chesterman, C.N., Burke, V., Beilin, L.J. (1998). The effects of alcohol on coagulation and fibrinolytic factors: a controlled trial. Blood Coagul Fibrinolysis, 9(1), 39-45.
Ebara, S., Marumo, M., Yamabata, C., Nishibe, I., Soneda, J.I., Mukai, J., Ohki, M., Uchida, K., Wakabayashi, I. (2017). Inverse associations of HDL cholesterol and oxidized HDL with d-dimer in patients with type 2 diabetes mellitus. Thromb Res, 155, 12-15.
15
ACCEPTED MANUSCRIPT
Ellison, R.C., Zhang, Y., Qureshi, M.M., Knox, S., Arnett, D.K., Province, M.A. (2004). Investigators of the NHLBI Family Heart Study. Lifestyle determinants of
Family Heart Study. Am Heart J, 147(3), 529-535.
RI PT
high-density lipoprotein cholesterol: the National Heart, Lung, and Blood Institute
Kashiwagi, A., Kasuga, M., Araki, E., Oka, Y., Hanafusa, T., Ito, H., Tominaga, M.,
SC
Oikawa, S., Noda, M., Kawamura, T., Sanke, T., Namba, M., Hashiramoto, M., Sasahara, T., Nishio, Y., Kuwa, K., Ueki, K., Takei, I., Umemoto, M., Murakami,
M AN U
M., Yamakado, M., Yatomi, Y., Ohashi, H.; Committee on the Standardization of Diabetes Mellitus-Related Laboratory Testing of Japan Diabetes Society. (2012). International clinical harmonization of glycated hemoglobin in Japan: From Japan Diabetes Society to National Glycohemoglobin Standardization Program values. J
TE D
Diabetes Investig, 3(1), 39-40.
Krobot, K., Hense, H.W., Cremer, P., Eberle, E., Keil, U. (1992). Determinants of
EP
plasma fibrinogen: relation to body weight, waist-to-hip ratio, smoking, alcohol, age, and sex. Results from the second MONICA Augsburg survey 1989-1990.
AC C
Arterioscler Thromb, 12(7), 780-788. Meade, T.W., Imeson, J., Stirling, Y. (1987). Effects of changes in smoking and other characteristics on clotting factors and the risk of ischaemic heart disease. Lancet,
2(8566), 986-988. Mukamal, K.J., Jadhav, P.P., D'Agostino, R.B., Massaro, J.M., Mittleman, M.A., Lipinska, I., Sutherland, P.A., Matheney, T., Levy, D., Wilson, P.W., Ellison, R.C., Silbershatz, H., Muller, J.E., Tofler, G.H. (2001). Alcohol consumption and
16
ACCEPTED MANUSCRIPT
hemostatic factors: analysis of the Framingham Offspring cohort. Circulation, 104(12), 1367-1373.
RI PT
Nwose, E.U., Richards, R.S., Jelinek, H.F., Kerr, P.G. (2007). D-dimer identifies stages in the progression of diabetes mellitus from family history of diabetes to cardiovascular complications. Pathology, 39(2), 252-257.
Clinical Applications. Adv Clin Chem, 69, 1-46.
SC
Olson, J.D. (2015). D-dimer: An Overview of Hemostasis and Fibrinolysis, Assays, and
M AN U
Polsky, S., Akturk, H.K. (2017). Alcohol Consumption, Diabetes Risk, and Cardiovascular Disease Within Diabetes. Curr Diab Rep, 17(12), 136. Reynolds, K., Lewis, B., Nolen, J.D., Kinney, G.L., Sathya, B., He, J. (2003). Alcohol consumption and risk of stroke: a meta-analysis. JAMA, 289(5), 579-588.
TE D
Ridker, P.M., Vaughan, D.E., Stampfer, M.J., Glynn, R.J., Hennekens, C.H. (1994). Association of moderate alcohol consumption and plasma concentration of
EP
endogenous tissue-type plasminogen activator. JAMA, 272(12), 929-933. Rubin, R., Rand, M.L. (1994). Alcohol and platelet function. Alcohol Clin Exp Res,
AC C
18(1), 105-110.
Stote, K.S., Tracy, R.P., Taylor, P.R., Baer, D.J. (2016). The effect of moderate alcohol consumption on biomarkers of inflammation and hemostatic factors in
postmenopausal women. Eur J Clin Nutr, 70(4), 470-474. Sumi, H., Hamada, H., Tsushima, H., Mihara, H. (1988). Urokinase-like plasminogen activator increased in plasma after alcohol drinking. Alcohol Alcohol, 23(1), 33-43.
17
ACCEPTED MANUSCRIPT
Tripodi, A. (2011). D-dimer testing in laboratory practice. Clin Chem, 57(9), 1256-1262.
RI PT
van de Wiel, A. (2004). Diabetes mellitus and alcohol. Diabetes Metab Res Rev, 20(4), 263-267.
Wakabayashi, I., Kobaba-Wakabayashi, R., Masuda, H. (2002). Relation of drinking
SC
alcohol to atherosclerotic risk in type 2 diabetes. Diabetes Care, 25(7),
AC C
EP
TE D
M AN U
1223-1228.
18
ACCEPTED MANUSCRIPT
Figure legend
RI PT
Figure 1. Scatter plots for the relationships between HDL cholesterol and d-dimer in overall subjects (A) and nondrinker subjects (B) of patients with diabetes. D-dimer levels did
SC
not show a normal distribution and were thus used after logarithmic transformation in
AC C
EP
TE D
M AN U
linear regression analysis. Pearson’s correlation coefficients (r) are shown in the figure.
19
ACCEPTED MANUSCRIPT
RI PT
Table 1. Characteristics of subjects in each drinker group and overall subjects of patients with type 2 diabetes.
AC C
EP
TE D
M AN U
SC
Nondrinkers Occasional drinkers Regular drinkers Overall subjects Number 151 31 87 269 Gender (men [%]) 48.3 67.7 81.6** 61.3 Age (years) 70.0 ± 10.9 66.6 ± 11.0 69.3 ± 9.1 69.4 ± 10.4 Current smokers (%) 49.0 54.8* 69.0** 56.1 Anti-dyslipidemia therapy (%) 51.7 58.1 39.1 48.3 Anti-coagulation therapy (%) 37.1 58.1* 36.8 39.4 Arterial disease (%) 27.2 19.4 16.1* 22.7 2 Body mass index (kg/m ) 24.9 ± 4.9 24.8 ± 3.7 24.6 ± 3.8 24.8 ± 4.5 Hemoglobin A1c (%) 7.18 ± 1.27 7.33 ± 1.58 7.07 ± 1.03 7.16 ± 1.23 High C-reactive protein (%) 16.1 21.1 20.0 18.6 ALT (IU/L) 20.00 (14.75, 29.00) 21.00 (19.00, 34.00) 20.00 (16.00, 27.25) 20.00 (15.00, 29.00) High ALT (%) 7.3 9.7 9.2 8.2 HDL cholesterol (mg/dl) 53.1 ± 17.2 54.3 ± 15.9 63.1 ± 21.2** 56.5 ± 19.0 Low HDL cholesterol (%) 33.1 16.1 8.0** 23.0 LDL cholesterol (mg/dl) 110.7 ± 34.5 110.5 ± 33.2 110.6 ± 30.1 110.6 ± 32.9 Total cholesterol (mg/dl) 189.8 ± 38.9 188.7 ± 36.1 198.1 ± 32.2 192.3 ± 36.7 Fibrinogen (mg/dl) 318.1 ± 83.2 317.0 ± 76.7 306.4 ± 81.1 314.2 ± 81.7 High fibrinogen (%) 13.9 16.1 6.9 11.9 D-dimer (µg/ml) 0.68 (0.41, 1.08) 0.58 (0.42, 1.01) 0.51 (0.38, 0.78)* 0.60 (0.40, 0.91) High d-dimer (%) 27.2 25.8 14.9* 23.0 Shown are numbers, means with standard deviations, medians with 25 and 75 percentile values, and frequencies (%) of each variable. Arterial disease, at least one of histories of ischemic heart disease, stroke and peripheral arterial disease. Asterisks denote significant differences from nondrinkers (*, p < 0.05; **, p < 0.01).
ACCEPTED MANUSCRIPT
Table 2. Relationships of frequency of alcohol drinking (non-, occasional and regular drinkers) with HDL cholesterol, fibrinogen, and
RI PT
d-dimer in patients with type 2 diabetes obtained by using ANCOVA and logistic regression analysis. Nondrinkers (n = 151)
Occasional drinkers (n = 31)
Regular drinkers (n = 87)
HDL cholesterol (mg/dl)
52.1 ± 1.4
54.0 ± 3.1
64.9 ± 1.9**
Fibrinogen (mg/dl)
315.9 ± 6.7
320.0 ± 14.7
309.0 ± 9.0
Log(d-dimer)
-0.157 ± 0.023
-0.164 ± 0.050
-0.272 ± 0.030*
M AN U
SC
Means (±SEs)
Low HDL cholesterol
1.00
High fibrinogen
1.00
High d-dimer
1.00
TE D
Odds ratios (CIs)
0.43 (0.15-1.25)
0.20 (0.08-0.50)**
1.24 (0.41-3.73)
0.47 (0.17-1.29)
1.03 (0.40-2.68)
0.46 (0.21-0.98)*
Shown are mean levels with their standard errors (SEs) of HDL cholesterol, fibrinogen and log-transformed d-dimer in each alcohol group (upper lines) and odds ratios with their confidence intervals (CIs) in parentheses of occasional or regular drinkers vs. nondrinkers
EP
for hypo-HDL cholesterolemia, hyper-fibrinogenemia, and high d-dimer (lower lines). Age, gender, BMI, hemoglobin A1c, history of
AC C
smoking and history of medication therapy for dyslipidemia or anti-coagulant therapy were adjusted in ANCOVA and logistic regression analysis. Asterisks denote significant differences (*, p < 0.05; **, p < 0.01) from nondrinkers in ANCOVA or from the reference level of 1.00 in logistic regression analysis.
ACCEPTED MANUSCRIPT
r = -0.340 (p < 0.01)
AC C
EP
TE D
M AN U
r = -0.381 (p < 0.01)
SC
RI PT
Figure 1
Figure. 1. Scatter plots for the relationships between HDL cholesterol and d-dimer in overall subjects (A) and nondrinker subjects (B) of patients with diabetes. D-dimer levels did not show a normal distribution and were thus used after logarithmic transformation in linear regression analysis. Pearson’s correlation coefficients (r) are shown in the figure.
ACCEPTED MANUSCRIPT
Highlights
It remains unknown whether alcohol drinking affects coagulability in patients with
The relationship between alcohol intake and d-dimer was investigated in patients with diabetes.
Regular drinking, but not occasional drinking, was associated with lower d-dimer
M AN U
and higher HDL cholesterol levels.
SC
RI PT
diabetes.
D-dimer was inversely associated with HDL cholesterol independently of alcohol.
Habitual alcohol drinking may suppress hypercoagulability in patients with
AC C
EP
TE D
diabetes.