Critique 026: Chinese study suggests that alcohol increases angiographically significant coronary disease – 8 December 2010

Zhou X, Li C, Xu W, Hong X, Chen J.   Relation of alcohol consumption to angiographically proved coronary artery disease in Chinese men.  Am J Cardiol 2010;106:1101–1103.

Authors’ Abstract

Light-to-moderate alcohol consumption is believed to be protective against coronary artery disease (CAD) in many studies.  However, the cardioprotective effects of alcohol intake lack epidemiologic evidence in a Chinese population.  The present case–control study was designed to explore the relation between alcohol consumption and angiographically proved CAD in Chinese men.  The study population consisted of 1,476 consecutive men 36 to 84 years of age who underwent coronary arteriography. Participants were categorized as nondrinkers, light drinkers, moderate drinkers, and heavy drinkers.

Adjusted odds ratios for angiographically proved CAD for light, moderate, and heavy drinking were 1.16 (95% confidence interval 0.68 to 1.94), 1.78 (1.35 to 2.27), and 2.18 (1.46 to 3.25).  Adjusted odds ratios were 1.36 (1.08 to 1.83) for drinking alcohol 2 to 3 days/week, 1.58 (1.17 to 2.26) for 4 to 5 days/week, and 2.03 (1.36 to 3.27) for 6 to 7 days/week.  Compared to non-drinking, adjusted odds ratios were 1.03 (0.54 to 1.87) for drinking 0 to 15 years, 1.61 (1.28 to 2.14) for 16 to 30 years, and 1.98 (1.23 to 3.05) for >30 years.  In conclusion, the authors state that moderate-to-heavy alcohol consumption increased the risk of CAD in Chinese men.  CAD risk tended to increase with an increase in frequency and duration of drinking.

Forum Comments

The large majority of studies relating alcohol to CAD have been among Western cultures.  Comparisons between Europeans and Asians may lead to difficulties, as there are large differences in risk factors between Westerners and the Chinese.  Some of these are due to socioeconomic differences, diet, lifestyle, access to care, and diagnostic modalities.  There are also differences between Europeans and Asians in genetic factors, especially related to hepatic metabolism of alcohol. 

Chinese have lower rates of CAD than Americans and Europeans, and only a few studies have evaluated the association between alcohol consumption and CAD in China.  The present study is a case-control comparison of drinking habits of men who were referred for coronary angiography because of chest pain or abnormal EKG changes.  The authors relate alcohol intake prior to the catheterization to their diagnosis of CAD, defining “angiographically proven CAD” as 50% or greater obstruction of one or more arteries.

Background:  Lee et al1 reported from Singapore in 2001 that Chinese and Malay citizens had lower rates of coronary heart disease (CHD) than did Indians in that city.  Overall, however, adjusting for ethnicity, non-drinkers had about 80% higher risk of CAD than did drinkers.  In 2002, Lam et al2 reported on a case control study with 598 CAD hospital cases and 1,100 community controls in Hong Kong.  Stepwise logistic regression models showed adjusted odds ratios of 0.32 (95% CI: 0.22 – 0.45) for alcohol drinking in men, and 0.15 (95% CI: 0.08 – 0.30), respectively, in women.  The protective effect of drinking was observed for different types of drinks and frequency of drinking, although few subjects drank alcohol more than 3 days per week.

A recent large prospective cohort study by Bazzano et al3 reported on the follow up of 64,597 Chinese men aged ≥ 40 years who were free of clinical CAD at baseline examination in 1991.  Follow-up evaluation was conducted in 1999-2000, with a response rate of 94%.  Over 494,084 person-years of follow up, the investigators documented 725 (361 fatal) incident myocardial infarctions (MI) and 976 (588 fatal) incident CAD events.  After stratification by province to account for multi-stage sampling design and adjustment for age, education, physical activity, cigarette smoking, body mass index, systolic blood pressure, urbanization (urban vs. rural), geographic variation (north vs. south) and history of diabetes, relative risk of MI was 0.93 (0.70-1.24) for participants consuming 1 to 6 drinks/week, 0.66 (0.54-0.82) for those consuming 7 to 34 drinks/week, and 0.58 (0.41-0.81) for those consuming ≥ 35 drinks/week (p for linear trend <0.0001) compared to non-drinkers.  The corresponding relative risks for total CAD events were 0.99 (0.77-1.27), 0.67 (0.56-0.81), and 0.58 (0.44-0.78), respectively (p for linear trend <0.0001).  

In the Bazzano et al study,3 there were many more occurrences of cardiovascular disease (CVD) than CAD, as CVD includes stroke as well as heart disease, and stroke is more common in China than CAD.  The lowest risk ratio for CVD was for subjects consuming 1-5 drinks/day (RR 0.79, 95% CI 0.73, 0.86).  This paper adds considerable information on the relation between alcohol consumption and CVD among Asians.  A similar inverse association between alcohol and CVD, as seen in Western populations, seems to be present in China.

Critical comments on present paper:  In the present study, only multivariate-adjusted results are shown (rather than also presenting age-adjusted data and seeing the effects of further adjustment for potential confounders).  This could be an important omission as, for example, in several instances the raw data presented for subjects “with no CAD” (by their definition) indicated that they were more-frequent drinkers (e.g., those reporting 1-6 drinks/week) but the adjusted odds ratio indicated findings in the opposite direction (a tendency for higher risk of CAD for those consuming at this level).  Similarly, 29.4% of subjects with “no CAD” consumed alcohol for up to 15 years, versus 20.9% of those with CAD, but the adjusted OR for CAD was again greater than 1.0; these show a directional change in effect when potential confounding variables were added. 

A Forum member comments on potential confounding by smoking.  “There was a very high prevalence of smokers in this study, with seven out of 10 cases smoking compared to 4 out of 10 controls.  How much is alcohol intake a surrogate for cigarette smoking in this study?  A simple adjustment with smoking status as a categorical variable may have been insufficient to remove such confounding.  Information on smoking amount (in pack years) in cases and controls and by drinking category should have been sought and utilized in this analysis.”  Another Forum member adds: “This population is highly selective, and they have more than one risk factor for CAD (hypertension, smoking, inactivity, obesity, etc.).  Many of the subjects could have the metabolic syndrome.  These are serious confounders, and could have led to inaccurate conclusions.”

The finding in this study that subjects who were shown to have one or more lesions meeting their definition of CAD tended to consume more alcohol “each week,” although the time period of drinking was apparently not specified and potential changes in drinking after the chest pain or EKG changes were noted cannot be determined.  Longer-term alcohol intake was not reported, so ex-drinkers and lifetime abstainers would have been classified as non-drinkers.

In the analyses relating drinking frequency and duration of drinking to coronary artery disease, it appears that no adjustments were made for the amount of drinking of subjects.  Hence, more frequent drinkers may well have been those whose average amount was greater, making it difficult to judge the effects of just the frequency of drinking.

Interpretation of the paper in light of previous research:  This is a very “narrow” study, focusing on subjects presenting with symptoms suggesting CAD, a group not typical of the general Chinese population.  Thus, a key problem in attempting to interpret this paper relates to the characteristics of the subjects studied, all of whom had been referred for angiography because of symptoms or EKG changes.  Angiography detects only fixed obstructive lesions, so subjects with symptoms related to transient intra-vascular clots may not have been diagnosed with CAD at angiography.  As stated, a recent large population-based prospective study of mainland Chinese showed that drinkers had lower risk of CAD and CVD than non-drinkers.3

One Forum member comments:  “This is a symptomatic population of patients with typical and atypical chest pain selected from hospital admissions.  These patients, in general, have an incidence of >90 and > 50%, respectively, of having angiographic CAD.  Most cardiologists consider coronary lesion with < 70% obstruction as non-flow limiting and unlikely to produce symptoms.”  (In the present paper, results using >70% obstruction as a cut point for defining angiographically significant CAD are not given.) 

Renaud4 has stated that moderate alcohol intake does not prevent CAD through an effect on atherosclerosis, but rather primarily through a haemostatic mechanism.  Drinkers have a significantly lower risk of MI, but often do not show a reduced risk of angina, owing to the understanding that MI is a thrombotic event and angina is not.  We do not know how many of this population ‘ruled in’ with positive enzymes as opposed to those without evidence of a MI.


1.  Lee J, Heng D, Chia KS, Chew SK, Tan BY, Hughes K.  Risk factors and incident coronary heart disease in Chinese, Malay and Asian Indian males: the Singapore Cardiovascular Cohort Study.  Int J Epidemiol 2001;30:983-988.

2.  Lam TH, Chung SF, Janus ED, Lau CP, Hedley AJ, Chan HW, Chow L, Keung KK, Li SK.  Smoking, alcohol drinking and non-fatal coronary heart disease in Hong Kong Chinese. Ann Epidemiol 2002;12:560-567.

3.  Bazzano LA, Gu D, Reynolds K, Chen J, Wu X, Chen S, Duan X, Chen J, He J.  Alcohol consumption and risk of coronary heart disease among Chinese men.  Int J Cardiol2009;135:78-85. 

4.  Renaud S, de Lorgeril M.  Wine, alcohol, platelets, and the French paradox for coronary heart disease.  Lancet 1992;339:1523-1526.

Forum Summary:  Among a large number of Chinese men presenting with chest pain or EKG changes, sequential subjects undergoing cardiac angiography were evaluated for obstructive lesions according to their reported recent alcohol intake.  No information was available on drinking patterns or on previous alcohol intake.  The authors concluded that alcohol intake increased the risk of angiographically significant CAD.

However, a recent large population-based study from mainland China showed that consumers of alcohol were less likely to develop coronary disease, results similar to those in most Western populations.  It is not possible from the present study to say that the association of alcohol intake with CAD is different between Chinese and Western populations, as the present study gives results only for a very select group of patients. 

The most important outcome regarding CAD is the occurrence of clinical events (myocardial infarction, cardiac death, etc.).  The detection of such events requires long-term follow-up studies to be able to judge the overall effects of alcohol drinking on CAD.

                                                   *                              *                              *

Comments included in this critique by the International Scientific Forum on Alcohol Research were provided by the following:

Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway

Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA

Tedd Goldfinger, DO, FACC, Desert Cardiology of Tucson Heart Center, Dept. of Cardiology, University of Arizona School of Medicine, Tucson, Arizona, USA

Lynn Gretkowski, MD, Obstetrics/Gynecology, Mountainview, CA, Stanford University, Stanford, CA, USA

Gordon Troup, MSc, DSc, School of Physics, Monash University, Victoria, Australia

David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa

Andrew L. Waterhouse, PhD, Marvin Sands Professor, Department of Viticulture and Enology, University of California, Davis

Ian Puddey, MD, Dean, Faculty of Medicine, Dentistry & Health Sciences, University of Western Australia, Nedlands, Australia

Creina Stockley, clinical pharmacology, Health and Regulatory Information Manager, Australian Wine Research Institute, Glen Osmond, South Australia, Australia

R. Curtis Ellison, MD, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA

Francesco Orlandi, MD, Dept. of Gastroenterology, Università degli Studi di Ancona. Italy