JOINT USE OF CLINICAL PARAMETERS, BIOLOGICAL MARKERS AND CAGE QUESTIONNAIRE FOR THE IDENTIFICATION OF HEAVY DRINKERS IN A LARGE POPULATION-BASED SAMPLE

Vincent Bataille1, Jean-Bernard Ruidavets1, Dominique Arveiler2, Philippe Amouyel3, Pierre Ducimetière4, Bertrand Perret5 and Jean Ferrières1,*

1 INSERM U 558, Département d’Epidémiologie, Faculté de Médecine, Toulouse,
2 Laboratoire d’Epidémiologie et de Santé Publique, Faculté de Médecine, Strasbourg,
3 INSERM U 508, Institut Pasteur, Lille,
4 INSERM U 258, Hôpital Broussais, Villejuif and
5 INSERM U 326, Département de Biochimie, Hôpital La Grave, Toulouse, France

Received 16 January 2002; in revised form 6 May 2002; accepted 12 September 2002


    ABSTRACT
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Aims: Alcohol consumption in France is one of the highest in the world. Factors associated with excessive alcohol drinking are numerous. However, taken separately, none of the existing clinical or biological markers of excessive alcohol intake enables an adequate identification of heavy drinkers. The aim of this cross-sectional survey was to identify socio-demographic, clinical and biological factors associated with excessive alcohol drinking, to develop a model and to assess its reliability, thus enabling the detection of heavy drinkers. Methods: Subjects were 1619 men and 1559 women, aged 35–64 years, living in three French areas (Lille, Strasbourg and Toulouse) and randomly selected from polling lists. Socio-demographic status, lifestyle, reported alcohol intake and answers to the CAGE questionnaire (alcohol dependence) were obtained by questionnaire. A blood sample was taken for quantification of biological parameters. Men who drank 60 g of ethanol a day (g/day) or above and women who drank 30 g/day or above were classified as heavy drinkers. The reference class (RC) gathered non-drinkers and moderate drinkers together. The sample was divided into two sub-samples: the first was used to estimate the parameters of a logistic regression model (heavy drinkers vs others), and the second to assess the accuracy of this model for the identification of heavy drinkers, using receiver operating characteristic (ROC) curves. A specific analysis was performed for each gender. Results: Fourteen per cent of men and 40.8% of women were non-drinkers. Nine per cent of women and 14.4% of men were heavy drinkers. Wine was the most consumed alcoholic beverage. In the univariate analyses, differences were observed between the two groups of alcohol consumers for most of the socio-demographic, clinical and biological variables considered. In the multivariate analyses, low educational level, smoking, apoprotein B, high density lipoprotein cholesterol, mean corpuscular volume (MCV), {gamma}-glutamyl-transferase (GGT) and the CAGE score for men, and living area, age, MCV, GGT and the CAGE score for women remained independently and significantly associated with heavy drinking. In the validation sub-sample, these models combining different types of markers enabled a good discrimination between heavy drinkers and the RC, with an area under the ROC curve of 82% for men and of 79% for women. Conclusions: In this study, socio-demographic, clinical and biological factors and the CAGE score were independently related to excessive alcohol drinking and their joint utilization in a screening model enabled a good recognition of heavy drinkers.


    INTRODUCTION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The relationship between alcohol and mortality has been widely investigated and is often summarized by the notion of U-shaped or J-shaped curves. Thus, in most prospective studies (Marmot et al., 1981Go; Klatsky et al., 1992Go; Fuchs et al., 1995Go; Renaud et al., 1999Go), all-cause mortality rates are lower for moderate drinkers [1–30 g (women) or 50 g (men) of pure alcohol a day, depending on the studies] than for non-drinkers, because of a lower mortality by cardiovascular diseases, particularly coronary heart diseases. This beneficial effect of moderate alcohol consumption might be explained by a rise of high density lipoprotein cholesterol (HDL-c) induced by alcohol consumption (Rimm et al., 1999Go), but also by other mechanisms such as alcohol anti-aggregation properties (Meade et al., 1985Go).

In contrast, excessive alcohol drinking (above 30 or 50 g of pure alcohol a day) is associated with a significant increase of all-cause and non-cardiovascular mortality rates (Marmot et al., 1981Go; Klatsky et al., 1992Go; Fuchs et al., 1995Go; Renaud et al., 1999Go), especially by cirrhosis, cancer and violent deaths. According to the official French national mortality statistics (INSERM, 1997Go), even when taking into account only alcoholic psychosis and cirrhosis, excessive alcohol consumption was involved in 11 000 deaths in France in 1997 (2.1% of total mortality); and up to 23 000 deaths (4.6% of total mortality) when including oral cancers (buccal cavity, pharynx, larynx and oesophagus), which are known to be strongly associated with alcohol misuse.

In the same way, excessive alcohol drinking has been associated with an increase in blood pressure (Milon et al., 1982Go; Saunders, 1987Go; Marmot et al., 1994Go) and often coexists with heavy smoking (Johnson and Jennison, 1992Go). Heavy drinking also induces changes in some biological parameters, such as {gamma}-glutamyl-transferase (GGT) (Whitehead et al., 1978Go; Yersin et al., 1995Go; Hoffmeister et al., 1999Go) or mean corpuscular volume (MCV) (Whitehead et al., 1978Go; Yersin et al., 1995Go), which are the most widely used among markers of excessive alcohol drinking, though their diagnostic accuracy remains controversial (Hoeksema and de Bock, 1993Go). Other tools, such as dependence questionnaires, have been tested and among these, the CAGE questionnaire (Ewing, 1984Go) has proved its effectiveness in the detection of dependent drinkers in hospital settings (Bush et al., 1987Go; Beresford et al., 1990Go).

Accordingly, a large number of scientific papers have been published concerning excessive alcohol drinking markers, but many of these have studied one or two markers — mainly biological ones — thus not achieving a very accurate discrimination between heavy drinkers and other categories of alcohol consumers. The use of combined markers (clinical markers, biological markers, dependence questionnaire) could help to identify heavy drinkers better than the use of biological markers only. Moreover, French drinkers have specific alcohol consumption habits, and studies concerning alcohol intake assessment in wine-drinking countries of Southern Europe are not particularly widespread (Feunekes et al., 1999Go). Indeed, studies on markers of alcohol misuse have mainly been conducted in North America or in countries where alcohol consumption patterns differ substantially from those found in France, either as regards quantities of alcohol ingested or types of beverages consumed.

The aim of this cross-sectional study was to investigate socio-demographic, clinical and biological factors associated with excessive alcohol drinking, and to develop a model and evaluate its reliability, thus enabling the detection of heavy drinkers in a population-based sample.


    SUBJECTS AND METHODS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Study population
Our sample population was recruited among the third MONICA (MONItoring of trends and determinants of CArdiovascular disease) French population survey. The MONICA Project was a 10-year international programme on cardiovascular risk factors coordinated by the World Health Organisation. It was set up in 27 countries. In France, the three MONICA centres — Lille (North), Strasbourg (East) and Toulouse (Southwest) — conducted three population surveys in the framework of this programme. The third one was carried out in a large sample of 1778 men and 1730 women, aged 35–64 years, randomly selected from the polling lists in the three different centres and interviewed between December 1994 and April 1997.

Questionnaire and medical examination
Interviews and medical examinations were performed by a trained nurse and checked by an epidemiologist. Each participant signed an informed consent and completed a standardized questionnaire collecting data about socio-economic status, lifestyle and self-reported alcohol consumption during a typical week (the nurse asked the subject to report his mean alcohol consumption in units of wine, beer, cider and spirits for each day of the week), and answers to the four questions of the CAGE questionnaire: Have you ever felt you ought to Cut down on your drinking? Have people Annoyed you by criticising your drinking? Have you ever felt bad or Guilty about your drinking? Have you ever had a drink first thing in the morning to steady your nerves? (Eye-opener).

Intake of alcohol (expressed in ml of pure alcohol per day) was estimated from the average number of ml of ethanol in a measure of each type of alcoholic beverage: wine, 12 cl serving, 10 or 12% alcohol; beer, 12 cl serving, 5% alcohol, 25 or 33 cl serving, 6 or 8% alcohol; cider, 12 cl serving, 5% alcohol; spirits, 2 or 6 cl serving, 20 or 40% alcohol. Total alcohol consumption expressed in ml per day was translated into g of pure alcohol per day (g/day) by multiplying the alcohol volume by alcohol density (0.8). Heavy drinking was defined as a consumption of 60 g/day or above for men (about six glasses a day of any beverage), and of 30 g/day or above for women (about three glasses a day). These relatively high thresholds were chosen in accordance with French alcohol consumption habits. Indeed, alcohol consumption (and particularly wine consumption) in France is known to be considerably higher than in other countries (Pyörälä, 1990Go). Heavy drinkers were compared with a reference class (RC) composed of both teetotallers and moderate drinkers.

Medical examination included anthropometric measurements (height, weight, waist girth, hip girth) and two blood pressure measurements performed according to standardized conditions (sitting position, after a 5 min rest, mercury sphygmomanometer). The average of the two blood pressure measurements was used in the analysis. Body mass index (BMI) was calculated as weight (kg) divided by squared height (m2), and waist-to-hip ratio (WHR) as waist girth (cm) divided by hip girth (cm). BMI and WHR were used in their continuous form for the analysis.

Blood analyses
Subjects were asked to comply with a minimum fasting period of 10 h before blood collection (20 ml). The blood samples were collected in tubes containing Na2EDTA and were centrifuged within 4 h. Total cholesterol and triglycerides were measured by enzymatic methods (Roche Diagnostics, Mannheim, Germany). HDL-c measurement was performed after sodium phosphotungstate-magnesium chloride precipitation of apolipoprotein B-containing lipoproteins (Roche Diagnostics). Apolipoprotein B was determined after first-order immunoprecipitation in an automated analyser. Fasting plasma glucose (FPG) was measured with a hexokinase-glucose-6-phosphate dehydrogenase method (Roche Diagnostics). MCV was calculated. These variables were used in their continuous form, after logarithmic transformation when their distribution was skewed (GGT, FPG and triglycerides).

Statistical analysis
Univariate and multivariate analyses were performed using the SAS® Statistical software version 6.12 (Sas Institute Inc., Cary, NC, USA). Incomplete data were excluded and a specific analysis was conducted for each gender.

For each gender, the sample was randomly divided into two sub-samples. The first one, the so-called ‘estimation sample’ (ES), was used to study differences between heavy drinkers and the RC, and to assess the parameters of a logistic regression model. The second one, ‘validation sample’ (VS), was used to achieve the validation of the model.

In the ES, univariate analyses were performed using the {chi}2 test for categorical data and analysis of variance for continuous data. Variables corresponding to a P < 0.25 in univariate analyses were kept for multivariate analyses. Stepwise descending logistic regression was then performed using a significance limit of 5%. Four groups of independent variables were studied: A, socio-demographic characteristics (age group, education); B, lifestyle (physical activity, smoking); C, biological variables (lipids, FPG, GGT, MCV); D, alcohol dependence (CAGE questionnaire). Interaction terms were introduced for variables remaining in the final model and first-degree interactions were tested.

The accuracy of the logistic function for the identification of heavy drinkers was assessed in the VS by comparing the model predictions and the actual existence of self-reported excessive alcohol consumption, using receiver operating characteristic (ROC) curves. The incremental value of the various groups of independent variables was assessed by comparing the ROC curves for models including different types of independent variables: A vs AB, AB vs ABC and ABC vs ABCD (full model).

Stata Statistical Software (Stata Statistical Software: Release 6.0, StataCorp. 1999, Stata Corporation, College Station, TX, USA) was used to assess ROC curves and to compare areas under the ROC curves, using the calculations from DeLong et al. (1988)Go, taking into account the correlated nature of the data (several tests performed in the same subjects). The ROCFIT software (University of Chicago, Department of Radiology, Chicago, IL, USA) was used to plot graphical representations of the ROC curves.


    RESULTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The participation rate was 63.5% for men and 60.2% for women. From the initial sample composed of 3508 subjects, 330 subjects were excluded because of incomplete data, and thus the analysis was performed on a sample of 1619 men and 1559 women, aged 50.3 years (SD ±8.6).

The random division of the sample resulted in an ES composed of 810 men and 780 women, and a VS composed of 809 men and 779 women. No differences were observed between the two sub-samples for alcohol intake, sex, age, educational level and smoking, but the mean BMI was a little higher in the ES (26.6 ± 0.1) than in the VS (26.1 ± 0.1) (P < 0.01).

Results from the estimation sample
Alcohol intake and the number of heavy drinkers were both higher in men than in women. Indeed, in the ES, 40.8% of women were non-drinkers, whereas, in men, these were only 14.1%. In the same way, the mean alcohol consumption among drinkers was 35 ± 29 g/day in men and 16 ± 17 g/day in women. Finally, in the ES, 117 men (14.4% of all men) and 68 women (8.7% of all women) were classified as heavy drinkers. Wine was the most widely consumed alcoholic beverage, representing 66% of total alcohol consumption. However, heavy drinkers had lower wine and higher beer intakes when compared with moderate drinkers (wine represented 69% of total alcohol intake in moderate drinkers, but only 62% in heavy drinkers; and beer represented 19% of total alcohol intake in moderate drinkers vs 28% in heavy drinkers).

Lower educational level and heavy smoking were more frequent in male heavy drinkers, than in male subjects included in the RC (Table 1Go). In women, differences were found between heavy drinkers and subjects in the RC for age group, living area and physical activity. For both sexes, BMI was identical for heavy drinkers and for subjects in the RC, but the WHR was a little higher for heavy drinking women. Systolic blood pressure was significantly higher for heavy drinking men, than for men from the RC. A similar difference was observed in women, but statistical significance was not reached. In the same way, diastolic blood pressure seemed to be somewhat higher for heavy drinkers in both sexes, but the difference was not significant. Almost all the biological parameters studied (Table 1Go) were found to be related to heavy drinking, with elevated GGT, MCV, triglycerides and apolipoprotein B for heavy drinkers in both genders. FPG was found to be significantly elevated only for heavy drinking women, and HDL-c only for heavy drinking men, when compared with subjects in the RC (Table 1Go).


View this table:
[in this window]
[in a new window]
 
Table 1. Differences between heavy drinkers (HD) and subjects from the reference class (RC), by sex
 
Independent variables from the four groups (socio-economic, clinical, biological and dependence) remained significantly and independently related to heavy drinking in the multivariate analyses for both genders (Table 2Go). For men, almost all the independent variables studied remained in the final model, except blood pressure and triglycerides. For women, only age class, living area, MCV, GGT and the CAGE score were kept in the final model. No significant interactions were found.


View this table:
[in this window]
[in a new window]
 
Table 2. Factors associated with heavy drinking by sex — results from the multivariate analyses by logistic regression
 
Results from the validation sample
Among the single markers taken into account, the CAGE questionnaire was found to be the most effective in identifying heavy drinking in both genders, with areas under the ROC curve (AUC) of 73% (68–77) and 71% (64–77), respectively, for men and for women. The best biochemical marker was GGT, with AUC of 70% (65–75) for men and 68% (59–76) for women.

Applied to subjects of the VS, the full models including several types of independent variables (socio-economic, clinical, biological, dependence) enabled adequate identification of heavy drinkers with AUC of 82% for men and 79% for women (Fig. 1Go). AUC for these full models were significantly higher than AUC for any other parameter taken individually. With GGT and MCV, the CAGE score was an independent and effective identifier of heavy drinking for men, and its inclusion in the model resulted in a significant increase of the AUC. On the other hand, inclusion of the CAGE score in the model for women did not result in enhanced performance of the model in identification of heavy drinkers, in comparison with the model including other types of markers (age class, living area, MCV, GGT).



View larger version (26K):
[in this window]
[in a new window]
 
Fig. 1. Results from the validation sample: model validation for men and women. Receiver operating characteristic curves and areas under the curves (AUC) according to the different types of markers included in the model. *P < 0.05, ***P < 0.001, significant degree in comparison with the preceding curve.

 

    DISCUSSION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
In this work, some socio-demographic, clinical and biological factors were independently related to excessive alcohol drinking, and their joint use in a screening model enabled a better recognition of heavy drinkers, than the mere use of biological markers or of the CAGE dependence questionnaire.

Nevertheless, the data used in this analysis were cross-sectional, without a time dimension, and the models presented should not be considered as predictive. So, independent variables used in this work may be either determinants of heavy drinking or consequences of alcohol misuse on health. These two types of variables were examined in the same way, insofar as the objective of this study was to describe the characteristics of French heavy drinkers and to provide a set of markers reliable for the screening of alcohol misusers.

Subjects were recruited in the three French MONICA centres and not in the general French population, and alcohol consumption habits can vary widely between geographical areas in France. But, even if it is not really representative of the general French population, the sample gives a good indication of alcohol consumption behaviours in France.

Moreover, subjects were randomly selected from polling lists, and some refused to participate. Some non-participants agreed to answer a few questions by telephone. Non-participants’ answers were compared with participants data. The results showed no difference for sex, age and prevalence of arterial hypertension, but differences were found for educational level and smoking habits, with lower educational levels and more frequent smoking in non-participants than in participants (data not shown). The fact that alcohol misuse and heavy smoking are strongly associated may account for the refusal of some heavy drinkers to participate in this study. Subjects excluded because of incomplete data were not different from subjects kept in the final sample.

In this work, alcohol consumption was self-reported and excessive alcohol drinking in France remains a strong taboo. Consequently, heavy drinkers may have been tempted to give ‘socially desirable’ answers, underestimating the quantities of alcohol actually ingested. Indeed, the reliability of self-reported alcohol consumption is a common drawback of studies on alcohol intake. There is a lack of reference methods of alcohol intake assessment, but questionnaires detailing alcohol intake for each day of the week (Romelsjö et al., 1995Go) and for each type of alcoholic beverage (Romelsjö et al., 1995Go; Feunekes et al., 1999Go), such as ours, do limit the risk of underestimation.

Alcohol misuse is known to be related to socio-demographic factors [educational level (Crum et al., 1992Go; Marques-Vidal et al., 2000Go)], clinical [smoking (Friedman et al., 1991Go)], blood pressure [(Milon et al., 1982Go; Saunders, 1987Go; Marmot et al., 1994Go; Marques-Vidal et al., 2001Go)], central obesity (Dallongeville et al., 1998Go) and biological data [triglycerides (Whitehead et al., 1978Go; Hoffmeister et al., 1999Go; Rimm et al., 1999Go; Marques-Vidal et al., 2001Go; Ruidavets et al., 2002Go), HDL-c (Miller et al., 1988Go; Patsch et al., 1992Go; Marques-Vidal et al., 1995Go; Hoffmeister et al., 1999Go; Rimm et al., 1999Go; Marques-Vidal et al., 2001Go; Ruidavets et al., 2002Go), MCV (Whitehead et al., 1978Go; Yersin et al., 1995Go), GGT (Whitehead et al., 1978Go; Hoffmeister et al., 1999Go)], and our results are in agreement with these previous findings.

But the aim of this work was not aetiological. The objective was not so much to identify determinants of heavy drinking, or the influence of alcohol on the level of clinical or biological markers, but rather to specify which ones, among the latter, were the most related to heavy drinking, in order to set-up an alcohol misuser’s screening model.

MCV and GGT are the most studied and the most widely used among markers of heavy drinking, yet, their reliability for the detection of heavy drinkers is limited (Hoeksema and de Bock, 1993Go; Yersin et al., 1995Go; Reynaud et al., 2000Go). Carbohydrate-deficient transferrin (CDT), another biological marker of alcohol misuse, has been found to have a better diagnostic accuracy than MCV and GGT, even if its performance is still considered as too low to be useful for screening procedures in a general population (Meerkerk et al., 1999Go). Unfortunately, these data were not available in the present study and it was not possible to study CDT and include it in our models.

Studies providing scores including several biochemical markers of high alcohol intake showed that a combination of markers enabled an improved detection of heavy drinkers. However, these combined tests were often insufficient for the detection of heavy drinkers. Thus, in 1981, Whitfield et alGo. proposed an index including three biological parameters they found to be the most closely related to heavy drinking: GGT, MCV and uric acid. The combination of these three parameters enabled a better recognition of heavy drinkers than using them individually. Shaper et al. (1985)Go studied a score combining five biological parameters (GGT, HDL-c, urate, mean corpuscular haemoglobin and lead), but only 50% of heavy drinkers could be identified.

Many papers about the identification of heavy drinkers have been devoted to dependence questionnaires, such as the CAGE. This short questionnaire was validated in middle-aged people in hospital settings (Bush et al., 1987Go; Beresford et al., 1990Go; Girela et al., 1994Go) and, in these studies, it was found to have a better diagnostic accuracy than biochemical markers. However, up to now, the reliability of the CAGE questionnaire in primary care (Aertgeerts et al., 2001Go) or in the general population (Bisson et al., 1999Go) remains controversial. In our work, the CAGE questionnaire was found to be a better device for the detection of heavy drinkers than any of the biochemical markers we studied. Actually, biochemical markers and dependence questionnaire may be considered as two different ways of approaching the problem of alcohol. Biochemical changes may illustrate the consequences of drinking on purely biological functions, whereas dependence questionnaires may be more likely to approach the patient’s ambiguous feelings about their problem towards alcohol. They may be two different types of information, and therefore, it is no surprise that biochemical markers and the CAGE questionnaire have remained independently associated with heavy drinking in this work, and that their combination in the model was found to be more effective for the screening of heavy drinkers than the use of either alone. Thus, in conclusion, in the absence of a sensitive and specific marker, the way to improve the detection of heavy drinkers may be based upon the combined use of several types of markers: clinical and biological markers, and dependence questionnaires.


    FOOTNOTES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
* Author to whom correspondence should be addressed at: Department of Epidemiology, INSERM U 558, Faculté de Médecine, 37 allées Jules Guesde, 31073 Toulouse Cedex, France. Back


    REFERENCES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Aertgeerts, B., Buntinx, F., Ansoms, S. and Fevery, J. (2001) Screening properties of questionnaires and laboratory tests for the detection of alcohol abuse or dependence in a general practice population. British Journal of General Practice 51, 206–217.[ISI][Medline]

Beresford, T. P., Blow, F. C., Hill, E., Singer, K. and Lucey, M. R. (1990) Comparison of CAGE questionnaire and computer-assisted laboratory profiles in screening for covert alcoholism. Lancet 336, 482–485.[ISI][Medline]

Bisson, J., Nadeau, L. and Demers, A. (1999) The validity of the CAGE scale to screen for heavy drinking problems in a general population survey. Addiction 94, 715–722.[CrossRef][ISI][Medline]

Bush, B., Shaw, S., Cleary, P., Delbanco, T. L. and Aronson, M. D. (1987) Screening for alcohol abuse using the CAGE questionnaire. American Journal of Medicine 82, 231–235.[ISI][Medline]

Crum, R. M., Bucholz, K. K., Helzer, J. E. and Anthony, J. C. (1992) The risk of alcohol abuse and dependence in adulthood: the association with educationnal level. American Journal of Epidemiology 135, 989–999.[Abstract]

Dallongeville, J., Marécaux, N., Ducimetière, P., Ferrières, J., Arveiler, D., Bingham, A., Ruidavets, J. B., Simon, C. and Amouyel, P. (1998) Influence of alcohol consumption and various beverages on waist girth and waist-to-hip ratio in a sample of French men and women. International Journal of Obesity 22, 1178–1183.[CrossRef][ISI]

DeLong, E. R., DeLong, D. M. and Clarke-Pearson, D. L. (1988) Comparing the areas under two or more correlated receiver operating characteristics curves: a non parametric approach. Biometrics 44, 837–845.[ISI][Medline]

Ewing, J. A. (1984) Detecting alcoholism: the CAGE questionnaire. Journal of the American Medical Association 252, 1905–1907.[Abstract]

Feunekes, G. I. J., Van’t Veer, P., Van Staveren, W. A. and Kok, F. J. (1999) Alcohol intake assessment: the sober facts. American Journal of Epidemiology 150, 105–112.[Abstract]

Friedman, G. D., Tekawa, I., Klatsky, A. L., Sidney, S. and Armstrong, M. A. (1991) Alcohol drinking and cigarette smoking: an exploration of the association in middle-aged men and women. Drug and Alcohol Dependence 27, 283–290.[CrossRef][ISI][Medline]

Fuchs, C. S., Stampfer, M. J., Colditz, G. A., Giovannuchi, E. L., Manson, J. E., Kawachi, I., Hunter, D. J., Hankinson, S. E., Hennekens, C. H., Rosner, B., Speizer, F. E. and Willett, W. C. (1995) Alcohol consumption and mortality among women. New England Journal of Medicine 332, 1245–1250.[Abstract/Free Full Text]

Girela, E., Villanueva, E., Hernandez-Cueto, C. and Luna, J. D. (1994) Comparison of the CAGE questionnaire versus some biochemical markers in the diagnosis of alcoholism. Alcohol and Alcoholism 29, 337–343.[Abstract]

Hoeksema, H. L. and de Bock, G. H. (1993) The value of laboratory tests for the screening and recognition of alcohol abuse in primary care patients. Journal of Family Practice 37, 268–276.[ISI][Medline]

Hoffmeister, H., Schelp, F. P., Mensink, G. B. M., Dietz, E. and Böhning, D. (1999) The relationship between alcohol consumption, health indicators and mortality in the German population. International Journal of Epidemiology 28, 1066–1072.[Abstract]

Institut National de la Santé et de la Recherche Médicale (INSERM) (1997) Causes médicales de décès. Résultats définitifs. INSERM, Paris, France.

Johnson, K. A. and Jennison, K. M. (1992) The drinking–smoking syndrome and social context. International Journal of the Addictions 27, 749–792.[ISI][Medline]

Klatsky, A. L., Armstrong, M. A. and Friedman, G. D. (1992) Alcohol and mortality. Annals of Internal Medicine 117, 646–654.[ISI][Medline]

Marmot, M. G., Shipley, M. J., Rose, G. and Thomas, B. J. (1981) Alcohol and mortality: a U-shaped curve. Lancet i, 580–583.

Marmot, M. G., Elliot, P., Shipley, M. J., Dyer, A. R., Ueshima, H., Beevers, D. G., Stamler, R., Kesteloot, H., Rose, G. and Stamler, J. (1994) Alcohol and blood pressure: the INTERSALT study. British Medical Journal 308, 1263–1267.[Abstract/Free Full Text]

Marques-Vidal, P., Cambou, J. P., Nicaud, V., Luc, G., Evans, A., Arveiler, D., Bingham, A. and Cambien, F. (1995) Cardiovascular risk factors and alcohol consumption in France and Northern Ireland. Atherosclerosis 115, 225–232.[CrossRef][ISI][Medline]

Marques-Vidal, P., Arveiler, D., Evans, A., Montaye, M., Bingham, A., Ruidavets, J. B., McMaster, D., Haas, B., Amouyel, P. and Ducimetière, P. (2000) Patterns of alcohol consumption in middle-aged men from France and Northern Ireland. The PRIME study. European Journal of Clinical Nutrition 54, 321–328.[CrossRef][ISI][Medline]

Marques-Vidal, P., Montaye, M., Haas, B., Bingham, A., Evans, A., Juhan-Vague, I., Ferrières, J., Luc, G., Amouyel, P., Arveiler, D., Yarnell, J. W. G., Ruidavets, J. B., Scarabin, P. and Ducimetiere, P. (2001) Relationships between alcoholic beverages and cardiovascular risk factor levels in middle-aged men, the PRIME study. Atherosclerosis 157, 431–440.[CrossRef][ISI][Medline]

Meade, T. W., Vickers, M. V., Thompson, S. G., Stirling, Y., Haines, A. P. and Miller, G. J. (1985) Epidemiological characteristics of platelet aggregability. British Medical Journal 290, 428–432.[ISI][Medline]

Meerkerk, G. J., Njoo, K. H., Bongers, I. M., Trienekens, P. and van Oers, J. A. (1999) Comparing the diagnostic accuracy of carbohydrate-deficient transferrin, gamma-glutamyltransferase, and mean cell volume in a general practice population. Alcoholism: Clinical and Experimental Research 23, 1052–1059.[ISI][Medline]

Miller, N. E., Bolton, C. H., Hayes, T. M., Bainton, D., Yarnell, J. W. G., Baker, I. A. and Sweetnam, P. M. (1988) Associations of alcohol consumption with plasma high density lipoprotein cholesterol and its major subfractions: the Caerphilly and Speedwell collaborative heart disease studies. Journal of Epidemiology and Community Health 42, 220–225.[Abstract]

Milon, H., Froment, A., Gaspard, P., Guidollet, J. and Ripoll, J. P. (1982) Alcohol consumption and blood pressure in a French epidemiological study. European Heart Journal 3, 59–64.

Patsch, W., Sharrett, A. R., Sorlie, P. D., Davis, C. E. and Brown, S. A. (1992) The relation of high density lipoprotein cholesterol and its subfractions to apolipoprotein A-I and fasting triglycerides: the rôle of environmental factors. American Journal of Epidemiology 136, 546–557.[Abstract]

Pyörälä, E. (1990) Trends in alcohol consumption in Spain, Portugal, France and Italy from the 1950s until the 1980s. British Journal of Addiction 85, 469–477.[ISI][Medline]

Renaud, S. C., Guéguen, R., Siest, G. and Salamon, R. (1999) Wine, beer, and mortality in middle-aged men from eastern France. Archives of Internal Medicine 159, 1865–1870.[Abstract/Free Full Text]

Reynaud, M., Schellenberg, F., Loisequx-Meunier, M. N., Schwan, R., Maradeix, B., Planche, F. and Gillet, C. (2000) Objective diagnosis of alcohol abuse: compared values of carbohydrate-deficient transferrin, gamma-glutamyltransferase, and mean corpuscular volume. Alcoholism: Clinical and Experimental Research 24, 1414–1419.[ISI][Medline]

Rimm, E. B., Williams, P., Fosher, K., Criqui, M. and Stampfer, M. J. (1999) Moderate alcohol intake and lower risk of coronary heart disease: meta-analysis of effects on lipids and haemostatic factors. British Medical Journal 319, 1523–1528.[Abstract/Free Full Text]

Romelsjö, A., Leifman, H. and Nyström, S. (1995) A comparative study of two methods for the measurement of alcohol consumption in the general population. International Journal of Epidemiology 24, 929–936.[Abstract]

Ruidavets, J. B., Ducimetière, P., Arveiler, D., Amouyel, P., Bingham, A., Wagner, A., Cottel, D., Perret, B. and Ferrières, J. (2002) Types of alcoholic beverages and blood lipids in a French population. Journal of Epidemiology and Community Health 56, 24–28.[Abstract/Free Full Text]

Saunders, J. B. (1987) Alcohol: an important cause of hypertension. British Medical Journal 294, 1045–1046.[ISI][Medline]

Shaper, A. G., Pocock, S. J., Ashby, D., Walker, M. and Whitehead, T. P. (1985) Biochemical and haematological response to alcohol intake. Annals of Clinical Biochemistry 22, 50–61.[ISI][Medline]

Whitehead, T. P., Clarke, C. A. and Whitfield, A. G. W. (1978) Biochemical and haematological markers of alcohol intake. Lancet i, 978–981.

Whitfield, J. B., Allen, J. K., Adena, M., Gallagher, H. G. and Hensley, W. J. (1981) A multivariate assessment of alcohol consumption. International Journal of Epidemiology 10, 281–288.[Abstract]

Yersin, B., Nicolet, J. F., Decrey, H., Burnier, M., Van Melle, G. and Pécoud, A. (1995) Screening for excessive alcohol drinking. Archives of Internal Medicine 155, 1907–1911.[Abstract]