ALCOHOL-ATTRIBUTABLE MORTALITY IN A HIGH PER CAPITA CONSUMPTION COUNTRY — GERMANY

Ulrich John* and Monika Hanke

University of Greifswald, Medical Faculty, Institute of Epidemiology and Social Medicine, Ernst-Moritz-Arndt University, Walther-Rathenau-Strasse 48, D-17487 Greifswald, Germany

Received 23 January 2002; in revised form 19 April 2002; accepted 3 May 2002


    ABSTRACT
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Aims: Little is known about age and gender distributions in alcohol-attributable mortality (AAM). The aim of this paper is to describe age and gender in AAM in a high per capita consumption country. Methods: The AAM was determined using national mortality statistics and data on the prevalence of alcohol risk drinking from Germany. According to this approach, alcohol-attributable fractions (AF) of causes of death are estimated. Results: In males, as well as females, more than 70% of the AAM cases are due to alcohol consumption as well as smoking. The mortality rates are highest in the age range 35–64 years with 25% of the total mortality in males, and 13% in females. The median age at death in AAM exclusively based on alcohol is 15 years below the median age at death in the male and 24 years below that in the female population. Conclusions: Early intervention should be stressed.


    INTRODUCTION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The alcohol-attributable mortality (AAM) is a main indicator of the health burden on a nation caused by harmful alcohol consumption. Prevention and treatment efforts might be based on such figures. In principle, there are three approaches to estimate the AAM: (1) use of the mortality data of alcohol dependents who are known from treatment; (2) use of longitudinal data including information about causes of death from individuals detected as hazardous drinkers, alcohol abusers or dependents in population surveys; (3) use of data from mortality statistics combined with known alcohol-related relative risks of death. Mortality data of alcohol dependents followed-up after treatment in Germany clearly reveal that, among formerly treated alcohol dependents, the annual mortality is 2%, which is higher compared to the annual mortality rate in the population of the same age range of 0.4%. The follow-up time spans are 2 (John et al., 2000Go), 4 (Feuerlein et al., 1994Go) and 10 years (Längle et al., 1993Go). Similarly, elevated death rates (Banks et al., 2000Go) are shown by international evidence: of 113 patients, 20 had died 10 years after alcoholism treatment (Finney and Moos, 1992Go); a cumulative mortality of 20.7% from 659 alcoholism patients after 10 years (Hurt et al., 1996Go); of 357 male alcoholics 10–14 years after treatment 26.6% were deceased (Liskow et al., 2000Go); of 99 married male problem drinkers from specialized alcohol problem treatment, 44 were deceased 20 years thereafter (Marshall et al., 1994Go); of 383 alcoholics who had been admitted to a psychiatric clinic, 124 were deceased 12–32 years later (Berglund, 1988Go). The findings from seven large multisite studies in the USA show that 1–6.8% of patients in alcoholism treatment died during the first year after treatment (Miller et al., 2001Go). Taken together, roughly 2% of treated alcoholics die annually after treatment according to follow-up data for up to 20 years. Little is known about the number of years from the onset of disease until death. Of 41 male alcoholics who had been initially diagnosed as alcoholics from 1957 to 1972, 18 were deceased by 1993 (Öjesjö et al., 1998Go).

The second approach provides longitudinal data based on population-survey studies including information about causes of death. They reveal that, with increasing alcohol consumption, the risk of death from alcohol-attributable disease is elevated (Fuchs et al., 1995Go; Thun et al., 1997Go). Among individuals formerly detected as alcohol misusers or dependents, 26% were deceased 14 years thereafter (Neumark et al., 2000Go). The third approach uses data from national mortality statistics and alcohol-related relative risks of death from long-term cohort studies. Evidence shows a U- or J-shaped relationship between the amount of drinking and mortality (Andreasson, 1998Go; Rehm et al., 2001Go). Mortality statistics have the advantage of not suffering from three main origins of selection bias inherent in the first two approaches in which: (1) the broad scope of alcohol-attributable disease is not sufficiently reflected; (2) the reduction of the mortality risk by treatment outcome is not estimated; (3) there is data attrition by cases lost to follow-up. The mortality statistics approach is particularly suitable when all death cases are needed, including rare ones.

AAM rates have been published based on a formula by Shultz et al. (1990)Go (cf. Stinson and DeBakey, 1992Go; Stout et al., 1993Go; Sutocky et al., 1993Go; Landen, 1997Go). It has been amplified in recent studies (Single et al., 1999Go; World Health Organization, 2000Go; Ridolfo and Stevenson, 2001Go). According to this approach, alcohol-attributable aetiologic fractions (AF; the proportion of cases that may be attributed to the use of alcohol) are derived based on relative risk estimates for specific causes of death from large prospective population studies in which data on alcohol consumption and smoking were collected and data on causes of death were collected in the following time period (Boffetta and Garfinkel, 1990Go; Corrao et al., 1999Go; Ridolfo and Stevenson, 2001Go). On this basis, for the USA, 4.5% of the total mortality was estimated to be alcohol-attributable (McGinnis and Foege, 1999Go), and for Canada the estimate was 3% of the total mortality (Single et al., 1999Go). Based on data from Europe, increases or decreases of 1 l of pure alcohol in the per capita consumption were estimated to be associated with increases or decreases of 1.3% in all-cause mortality rates (Her and Rehm, 1998Go). In Italy, 4.6% among all male and 2.0% among all female death cases have been attributed to alcohol use. Restricted to middle adult age (36–64 years), in France, the rates are 19.1% of all male and 13.0% of all female death cases (Zureik and Ducimetière, 1996Go). Similar to France, Germany, which showed 10.8 l of pure alcohol per capita consumption in 1997, belongs to the five countries with the highest use according to alcohol beverage industries statistics (Produktschap Voor Gedistilleerde Dranken, 2000Go). In 1991, 5.4% of females and 22.0% of males (age 18–64 years) were risk drinkers (Meyer et al., 1998Go) according to the criteria of the British Medical Association (1995)Go.

One limitation of the estimates provided so far is that age groups, gender, as well as confounding caused by tobacco smoking, have not been considered to a satisfactory degree, although such knowledge might help to clarify whether female death cases occur at a younger age than male death cases. Therefore, the goal of this contribution was to estimate the AAM specified by the AF, gender and age at death, and at the same time, to estimate how large a part of this fraction is explained by tobacco smoking.


    SUBJECTS AND METHODS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Population
The data about alcohol risk drinking were taken from the German population [National Health Survey carried out in 1991, n = 7466 in the age range 25–69 years, calculations of grams of pure alcohol according to Meyer et al. (1998)Go]. The tobacco-smoking prevalence rate was estimated based on the largest data set available for this purpose, the 1995 microcensus (n = 132 194 smokers aged 18 years or older; John and Hanke, 2002). The mortality data were taken from the vital statistics provided by the Federal Statistical Office for the year 1997 which included causes of death according to International Classification of Diseases (9th revision). There were 462 072 female and 398 317 male death cases with information about age at death. The median age was 71 years in males and 81 years in females. This is close to the life expectancy of 74 years in males and 80 years in females (German Federal Statistics Office, 1999aGo). All data were analysed in 2001.

For the calculation of the AF, we followed the recent literature (Single et al., 1999Go; World Health Organization, 2000Go; Ridolfo and Stevenson, 2001Go), derived using the following formula:


where RRi,j,k = relative risk of mortality in exposed groups compared with the unexposed group (RR0), Pi,j,k = proportion of the population exposed in each group. From Corrao et al. (1999)Go the relative risks for 25/50/100 g of pure alcohol per day were used because they were provided based on a meta-analysis of 200 studies which included a large variety of diseases. We calculated the AF. If the cause of death was attributable to tobacco smoking as well as alcohol consumption, we used the higher AF, even when results showed lives were expected to be saved by alcohol consumption.

Alcohol-related mortality (AM)
Three main groups of AM were distinguished: (1) AAM with an AF of 1 (AAM1), i.e. diseases exclusively dependent on alcohol consumption (Shultz et al., 1991Go): alcohol psychosis (ICD-9 291), alcohol dependence (ICD-9 303), alcohol misuse (ICD-9 305.0), polyneuropathy from alcohol misuse (ICD-9 357.5), gastritis from alcoholism (ICD-9 535.3), alcoholic liver disease (ICD-9 571.0–571.3), elevated blood alcohol concentration (ICD-9 790.3), alcoholic poisonings (ICD-9 E860.0, E860.1); (2) AAM with an AF < 1 (AAM<1), i.e. diseases which show an elevated relative mortality risk due to alcohol consumption, but not to smoking, which included unspecified liver cirrhosis, liver cancer and traffic accidents under the influence of alcohol (ICD-9 155, 174, 531, 532, 571.4–571.9, 577.1, E810–E825; Corrao et al., 1999Go). Traffic accidents were calculated based on German relative risk data (16.9% alcohol-attributable; German Federal Statistics Office, 1999bGo); (3) tobacco and alcohol-attributable mortality (TAAM), i.e. further diseases which are attributable to alcohol consumption as well as tobacco smoking include several cancers and diseases of the heart and circulation (ICD-9 153, 154 according to Chao et al., 2000Go; ICD-9 140–150, 161, 401, 410–414, 425.5, 430–438 according to Shultz et al., 1991Go; John and Hanke, 2002). The three groups are specified for two reasons: First, the AF of 1 (AAM1) has the highest validity and may be seen as separate from death cases with an AF estimated from longitudinal data from population samples (AAM<1). Secondly, there is a considerable synergistic effect from tobacco smoking and alcohol risk drinking (John and Hanke, 2002). Therefore, a potential overlap has to be made obvious. In total, we used 187 single diseases and a further 162 single types of accidents or poisonings. Burn deaths were not counted as part of the tobacco-attributable mortality, because no adequate data were available for Germany. Age at death and AM rates in single age groups are presented.


    RESULTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
In males, there were 55 698 alcohol-related cases of death (Table 1Go), the median age at death was 64 years. Of all alcohol-attributable cases of death, 73.8% were TAAM cases, a further 22.3% were AAM1 cases and a further 3.8% were AAM<1 cases. In the age range 35–64 years, the total AM rate (24.7%) was highest due to a high rate of AAM1 below 45 years of age, and from 45 to 64 years a high rate of TAAM. The median age at death was lowest in both AAM groups at 56 years, in TAAM it was 68 years. Half of the total alcohol-attributable death cases occurred below 65 years of age, in AAM1 at this age 77.8% of all cases had died.


View this table:
[in this window]
[in a new window]
 
Table 1. Alcohol-related mortality by age in males
 
In females, there were 18 019 cases of death (Table 2Go), the median age at death was 71 years. Of all alcohol-attributable cases of death, 72.5% were TAAM cases, a further 23.5% were AAM1 cases and a further 4.0% were AAM<1 cases. In the age range 35–64 years, the total AM rate (13.1%) was highest and at 65 years of age or older became markedly low, more so than in males. The AAM1 death cases contributed more than TAAM to the total AM in the age range 35–44 years, and TAAM contributed more to the total AM in the age range 55–64 years. The median age at death was lowest in AAM1 at 57 years, it was 62 years in AAM<1 and 77 years in TAAM. As in males, the highest AM rates occurred below 65 years of age, of all AAM1 cases 72.5% were deceased by then. However, 31.0% of all attributable death cases in males occurred before 65 years of age compared to 12.8% of all attributable death cases in females.


View this table:
[in this window]
[in a new window]
 
Table 2. Alcohol-related mortality by age in females
 

    DISCUSSION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
This age- and gender-specific analysis of AM data shows, first, that in males as well as females the highest AM rates occur in the age range 35–64 years. Thus, alcohol-related disease, accounting for 25% of the total mortality in males and 13% in females, is a substantial part of the total mortality in the most productive part of adult life. Secondly, in males as well as in females about 75% of all AAM1 cases occur below 65 years of age. Thirdly, the male total AM rate is roughly double the female one below 65 years of age. From 65 years on, there is a more pronounced decrease in the rate in females compared to males. One explanation might be that females develop alcohol-related disease in a shorter time than do males (Brady and Randall, 1999Go). On the other hand, it may be assumed that due to the higher life expectancy of females, in general, the mortality is more evenly distributed over the life-span.

In males, the median age at death in all alcohol-attributable cases is 7 years, and in AAM1 cases it is 15 years, below that of the male general population. In females, the median age at death in all alcohol-attributable cases is 10 years, and in AAM1 it is 24 years, below that of the female general population. Thus, the alcohol-attributable reduction of life expectancy is more drastic in females. In order to give a complete picture of mortality, the large number of TAAM cases must be taken into account, although evidence is not clear as to how large a part is due to tobacco and how large a part is due to alcohol.

The results represent the AM in a high per capita consumption country, and, as such, they are equal to the rates shown for 35–64-year-olds in France for females, but not for males. The German rates are slightly lower than those found for Italy and are similar to those of Canada and the USA. These two countries show a lower per capita consumption. Thus, there are some inconsistencies compared to the other countries. These inconsistencies may be due to the fact that, with recent developments in the estimation procedure, further causes of death detected as alcohol-related are included in our analysis, which in addition, makes transparent the confounding caused by tobacco-related causes of death.

There are several methodological limitations of this study (John and Hanke, 2002aGo). The time span between the collection of the alcohol risk drinking data and the mortality data is only 6 years. However, besides some short-term death cases of risk drinking, such as accidents, many diseases take more years to manifest themselves as alcohol-attributable causes of death. In Germany, alcohol consumption levels have been rather stable (Produktschap Voor Gedistilleerde Dranken, 2000Go). Therefore, the drinking amounts estimated for this study seem to provide valid data. The smoking- and the alcohol risk drinking-attributable mortality is not adjusted according to confounders other than those analysed by Thun et al. (2000)Go. The studies according to the relative mortality risks carried out in the USA would not show the same results in European countries. We changed the alcohol-attributable death rates by traffic accidents as there are data available for Germany. Estimation methods in the area of AAM are developing continuously and are becoming more and more differentiated, e.g. in more chronic and more short-term disturbances caused by alcohol risk drinking. Up to now, in most causes of death, the international data about relative risks are by far the best, mainly because of the large sample sizes which allow the inclusion of rare diseases, and because of the long time span included. There are no comparable data from Germany. It may be suspected that the inclusion of alcohol consumption from population surveys produces large under-estimates. We did not include years of potential life lost (John and Hanke, 2002bGo) and life years with disability attributable to alcohol consumption. The data shown are from a high tobacco as well as a high alcohol consumption country. Taken together, several origins of potential bias exist. But, in spite of these limitations, the data reveal higher AAM1 rates below 65 years of age than the treatment follow-up data would suggest. This is probably due to the fact that somatic diseases are included in the mortality statistics data, which are neither in the foreground of treatment follow-up nor of survey-based data. Thus, the mortality statistics results give a more complete view. Therefore, it may be concluded that AM in both genders occurs at a young adult age to a substantial degree and that prevention and early intervention are strongly suggested by these data.


    ACKNOWLEDGEMENTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The work was supported by grant no. IX311a 406.68.43.05 from the Social Ministry of the Federal State of Mecklenburg-West Pomerania, Germany.


    FOOTNOTES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
* Author to whom correspondence should be addressed. Back


    REFERENCES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 SUBJECTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Andreasson, S. (1998) Alcohol and J-shaped curves. Alcoholism: Clinical and Experimental Research 22, 359S–364S.[ISI][Medline]

Banks, S. M., Pandiani, J. A., Schacht, L. M. and Gauvin, L. M. (2000) Age and mortality among white male problem drinkers. Addiction 95, 1249–1254.[ISI][Medline]

Berglund, M. (1988) Alcoholics committed to treatment: a prospective long-term study of behavioral characteristics, mortality, and social adjustment. Alcoholism: Clinical and Experimental Research 12, 19–24.[ISI][Medline]

Boffetta, P. and Garfinkel, L. (1990) Alcohol drinking and mortality among men enrolled in an American Cancer Society Prospective Study. Epidemiology 1, 342–348.[Medline]

Brady, K. T. and Randall, C. L. (1999) Gender differences in substance use disorders. Psychiatric Clinics of North America 22, 241–252.[ISI][Medline]

British Medical Association (1995) Guidelines on Sensible Drinking. British Medical Association, London.

Chao, A., Thun, M. J., Jacobs, E. J., Henley, S. J., Rodriguez, C. and Calle, E. E. (2000) Cigarette smoking and colorectal cancer mortality in the Cancer Prevention Study II. Journal of the National Cancer Institute 92, 1888–1896.[Abstract/Free Full Text]

Corrao, G., Bagnardi, V., Zambon, A. and Arico, S. (1999) Exploring the dose–response relationship between alcohol consumption and the risk of several alcohol-related conditions: a meta-analysis. Addiction 94, 1551–1573.[ISI][Medline]

Feuerlein, W., Küfner, H. and Flohrschütz, T. (1994) Mortality in alcoholic patients given inpatient treatment. Addiction 89, 841–849.[ISI][Medline]

Finney, J. W. and Moos, R. H. (1992) The long-term course of treated alcoholism: II. Predictors and correlates of 10-year functioning and mortality. Journal of Studies on Alcohol 53, 142–153.[ISI][Medline]

Fuchs, C. S., Stampfer, M. J., Colditz, G. A., Giovannucci, E. L., Manson, J. E., Kawachi, I. et al. (1995) Alcohol consumption and mortality among women. The New England Journal of Medicine 332, 1245–1250.[Abstract/Free Full Text]

German Federal Statistics Office (1999a) Alkoholunfälle im Straßenverkehr 1998 [Alcohol-related Traffic Accidents 1998]. Statistisches Bundesamt, Wiesbaden.

German Federal Statistics Office (1999b) Statistical Yearbook 1999 [Statistisches Jahrbuch 1999]. Metzler-Poeschel, Stuttgart.

Her, M. and Rehm, J. (1998) Alcohol and all-cause mortality in Europe 1982–1990: a pooled cross-section time-series analysis. Addiction 93, 1335–1340.[ISI][Medline]

Hurt, R. D., Offord, K. P., Croghan, I. T., Gomez-Dahl, L., Kottke, T. E., Morse, R. M. et al. (1996) Mortality following inpatient addictions treatment. Role of tobacco use in a community-based cohort. Journal of the American Medical Association 275, 1097–1103.[Abstract]

John, U. and Hanke, M. (2002a) Tobacco smoking- and alcohol drinking-attributable cancer mortality in Germany. European Journal of Cancer Prevention 11, 11–17.[ISI][Medline]

John, U. and Hanke, M. (2002b) Tobacco- and alcohol-attributable mortality and potential life years lost in Germany. European Journal of Public Health 12, in press.

John, U., Veltrup, C., Driessen, M., Wetterling, C. and Dilling, H. (2000) Motivationsarbeit mit Alkoholabhängigen [Motivational Treatment of Alcohol Dependents]. Lambertus, Freiburg.

Landen, M. G. (1997) Alcohol-related mortality and tribal alcohol legislation. Journal of Rural Health 13, 38–44.[ISI][Medline]

Längle, B., Mann, K., Mundle, G. and Schied, H. (1993) Ten years after — the posttreatment course of alcoholism. European Psychiatry 8, 95–100.

Liskow, B. I., Powell, B. J., Penick, E. C., Nickel, E. J., Wallace, D., Landon, J. F. et al. (2000) Mortality in male alcoholics after ten to fourteen years. Journal of Studies on Alcohol 61, 853–861.[ISI][Medline]

Marshall, E. J., Edwards, G. and Taylor, C. (1994) Mortality in men with drinking problems: a 20-year follow-up. Addiction 89, 1293–1298.[ISI][Medline]

McGinnis, J. M. and Foege, W. H. (1999) Mortality and morbidity attributable to use of addictive substances in the United States. Proceedings of the Association of American Physicians 111, 109–118.[ISI][Medline]

Meyer, C., Rumpf, H.-J., Hapke, U. and John, U. (1998) Regionale Unterschiede in der Prävalenz riskanten Alkoholkonsums: Sekundäranalyse des Gesundheitssurveys Ost-West [Variation by region in the prevalence of risk drinking: Analysis of the health survey East and West Germany]. Gesundheitswesen 60, 486–492.[Medline]

Miller, W. R., Walters, S. T. and Bennett, M. E. (2001) How effective is alcoholism treatment in the United States? Journal of Studies on Alcohol 62, 211–220.[ISI][Medline]

Neumark, Y. D., Van Etten, M. L. and Anthony, J. C. (2000) ‘Alcohol dependence’ and death: survival analysis of the Baltimore ECA sample from 1981 to 1995. Substance Use and Misuse 35, 533–549.[Medline]

Öjesjö, L., Hagnell, O. and Otterbeck, L. (1998) Mortality in alcoholism among men in the Lundby community cohort, Sweden: a forty-year follow-up. Journal of Studies on Alcohol 59, 140–145.[ISI][Medline]

Produktschap Voor Gedistilleerde Dranken (2000) World Drink Trends 2000. NTC Publications, Oxfordshire.

Rehm, J., Greenfield, T. K. and Rogers, J. D. (2001) Average volume of alcohol consumption, patterns of drinking, and all-cause mortality: results from the US National Alcohol Survey. American Journal of Epidemiology 153, 64–71.[Abstract/Free Full Text]

Ridolfo, B. and Stevenson, C. (2001) The Quantification of Drug-caused Mortality and Morbidity in Australia, 1998, Report 7. Australian Institute of Health and Welfare, Canberra.

Shultz, J. M., Rice, D. P. and Parker, D. L. (1990) Alcohol-related mortality and years of potential life lost — United States, 1987. Morbidity and Mortality Weekly Report 39, 173–178.[Medline]

Shultz, J. M., Rice, D. P., Parker, D. L., Goodman, R. A., Stroh, G. and Chalmers, N. (1991) Quantifying the disease impact of alcohol with ARDI software. Public Health Reports 106, 443–450.[ISI][Medline]

Single, E., Robson, L., Rehm, J. and Xi, X. (1999) Morbidity and mortality attributable to alcohol, tobacco, and illicit drug use in Canada. American Journal of Public Health 89, 385–390.[Abstract]

Stinson, F. S. and DeBakey, S. F. (1992) Alcohol-related mortality in the United States, 1979–1988. British Journal of Addiction 87, 777–783.[ISI][Medline]

Stout, R. W., Parkinson, M. D. and Wolfe, W. H. (1993) Alcohol-related mortality in the U.S. Air Force, 1990. American Journal of Preventive Medicine 9, 220–223.[ISI][Medline]

Sutocky, J. W., Shultz, J. M. and Kizer, K. W. (1993) Alcohol-related mortality in California, 1980 to 1989. American Journal of Public Health 83, 817–823.[Abstract]

Thun, M. J., Peto, R., Lopez, A. D., Monaco, J. H., Henley, S. J., Heath, C. W. et al. (1997) Alcohol consumption and mortality among middle-aged and elderly U.S. adults. New England Journal of Medicine 337, 1705–1714.[Abstract/Free Full Text]

Thun, M. J., Apicella, L. F. and Henley, S. J. (2000) Smoking vs other risk factors as the cause of smoking-attributable deaths: confounding in the courtroom. Journal of the American Medical Association 284, 706–712.[Abstract/Free Full Text]

World Health Organization (2000) International Guide for Monitoring Alcohol Consumption and Related Harm. World Health Organization, Geneva.

Zureik, M. and Ducimetière, P. (1996) High alcohol-related premature mortality in France: concordant estimates from a prospective cohort study and national mortality statistics. Alcoholism: Clinical and Experimental Research 20, 428–433.[ISI][Medline]