Modeling the impact of alcohol dependence on mortality burden and ...

5 downloads 200 Views 548KB Size Report
Received 29 May 2012; accepted 2 August 2012. KEYWORDS ... The net burden caused by alcohol consumption was 1 in 7 deaths in men and. 1 in 13 deaths ...
European Neuropsychopharmacology (2013) 23, 89–97

www.elsevier.com/locate/euroneuro

Modeling the impact of alcohol dependence on mortality burden and the effect of available treatment interventions in the European Union J. Rehma,b,c,d,e,n, K.D. Shielda,e, G. Gmela, M.X. Rehmf, U. Frickg a

Centre for Addiction and Mental Health (CAMH), 33 Russell Street, Toronto, ON M5S 2S1, Canada Dalla Lana School of Public Health (DLSPH), University of Toronto, 155 College Street, Toronto, ON M5T 3M7, Canada c Department of Psychiatry, University of Toronto, 27 King’s College Circle, Toronto, ON M5S 1A1, Canada d Institute for Clinical Psychology and Psychotherapy, TU Dresden, Chemnitzer Str. 46, D-01187 Dresden, Germany e Institute of Medical Science (IMS), University of Toronto, 27 King’s College Circle, Toronto, ON M5S 1A1, Canada f Faculty of Arts and Sciences—Politics and Governance, Ryerson University, 350 Victoria Street, Toronto, ON M5B 2K3, Canada g Psychiatric University Hospital, University of Regensburg, Universit¨ atsstr. 42, D 93051 Regensburg, Germany b

Received 29 May 2012; accepted 2 August 2012

KEYWORDS

Abstract

Alcohol consumption; Alcohol dependence; Treatment; Mortality; Europe

Alcohol consumption is a major risk factor for the burden of disease, and Alcohol Dependence (AD) is the most important disorder attributable to this behavior. The objective of this study was to quantify mortality associated with AD and the potential impact of treatment. For the EU countries, for the age group 15–64 years, mortality attributable to alcohol consumption in general, to heavy drinking, and to AD were estimated based on the latest data on exposure and mortality. Potential effects of AD treatment were modeled based on Cochrane and other systematic reviews of the effectiveness of the best known and most effective interventions. In the EU 88.9% of men and 82.1% of women aged 15–64 years were current drinkers; and 15.3% of men and 3.4% of women in this age group were heavy drinkers. AD affected 5.4% of men and 1.5% of women. The net burden caused by alcohol consumption was 1 in 7 deaths in men and 1 in 13 deaths in women. The majority of this burden was due to heavy drinking (77%), and 71% of this burden was due to AD. Increasing treatment coverage for the most effective treatments to 40% of all people with AD was estimated to reduce alcohol-attributable mortality by 13% for men and 9% for women (annually 10,000 male and 1700 female deaths avoided). Increasing treatment rates for AD was identified as an important issue for future public health strategies to reduce alcohol-attributable harm and to complement the current focus of alcohol policy. & 2012 Elsevier B.V. and ECNP. Open access under CC BY-NC-ND license.

n

Corresponding author at: Centre for Addiction and Mental Health, 33 Russell Street, Room 2035B, Toronto, ON M5S 2S1, Canada. E-mail address: [email protected] (J. Rehm).

0924-977X & 2012 Elsevier B.V. and ECNP. Open access under CC BY-NC-ND license. http://dx.doi.org/10.1016/j.euroneuro.2012.08.001

90

1.

J. Rehm et al.

Introduction

Alcohol use disorders, commonly defined as comprising alcohol dependence (AD) and alcohol abuse (American Psychiatric Association, 1994), are globally one of the most prevalent mental disorders, affecting an estimated 3.6% of the population between 15 and 64 years of age worldwide (men: 6.3%; women 0.9%) (Rehm et al., 2009). There are marked regional variations in prevalence, with the World Health Organization’s European region showing the highest prevalence in this age group (5.5% total; men: 9.1%; women 2.0%). Within the European region, Russia and surrounding countries have the highest prevalence, but the European Union (EU) also shows a high prevalence (Rehm et al., 2005; Wittchen et al., 2011). The public health impact of alcohol use disorders, as with most other mental disorders, has been mainly seen as disabling rather than fatal, especially in comparison to cardiovascular disorders and cancer (Samokhvalov et al., 2010; World Health Organization, 2008). The disabling effects manifest themselves in absenteeism from work, failure to fulfill one’s social roles, and other interpersonal problems and functional problems in daily life (Ormel et al., 1994; Samokhvalov et al., 2010). As most of the public health impact of alcohol use disorders stems from AD (Langlois et al., 2011; Samokhvalov et al., 2010), we will restrict ourselves to this disorder. The underlying evidence for the conclusion that the consequences of AD are disabling rather than fatal stems from the Global Burden of Disease (GBD) studies (Lopez et al., 2006; World Health Organization, 2008). The GBD estimates with respect to mortality associated with AD in turn have been based on the review of Harris and Barraclough (Harris and Barraclough, 1998), indicating a Standardized Mortality Ratio (SMR) of 1.80 (95% confidence interval (CI): 1.76–1.84) for men and of 3.84 (95% CI: 3.54–4.15) for women. In other words, people with AD have a 1.8 or 3.8 fold higher ratio of observed deaths to expected deaths for men and women respectively, where expected deaths are calculated based on the age- and sex-comparable general population. However, the Harris and Barraclough review is quite out of date, with more than 30 new studies appearing since 1998 (Roerecke and Rehm, 2012). Also, most of the more recent estimates of mortality risks associated with AD have been markedly higher than those reported in the Harris and Barraclough review, especially those estimates based on treated samples (for two examples see: (Campos et al., 2011; Hayes et al., 2011)). Studies on mortality of AD in general populations showed lower risk than treatment, but still elevated risk compared to the general population without AD (Dawson, 2000; Fichter et al., 2011). It is the aim of this contribution to look into the mortality impact of AD in the EU based on the latest available information on exposure, risk relations and mortality. More specifically, we estimate the impact of AD on mortality in each of the member states of the EU, and aggregate the results for the EU as a whole. In a second step, we compare estimates of AD-attributable mortality to estimates of the impact of alcohol consumption overall and of the impact of heavy drinking in the EU. The approach to estimate on the country level and then aggregate is necessary as drinking

cultures vary markedly between regions in Europe, and sometimes vary from country to country or even within countries, based on product traditions, drinking patterns, and social reactions to alcohol (Iontchev, 1998; Popova et al., 2007; Room, 2010). AD seems to have a marked effect on mortality, and one way to reduce this effect could be by treatment. There are effective treatment options for this condition, leading either to abstinence, or to a reduction in heavy drinking days and overall consumption (Magill and Ray, 2009; R¨ osner et al., 2010a, 2010b; Smedslund et al., 2011). However, less than 10% of all people with AD are currently treated in the EU, which is the lowest treatment coverage of any mental disorder (Alonso et al., 2004; Rehm et al., 2012). Thus, as a last step, we estimate the effects of increased treatment interventions on the associated burden of AD.

2. 2.1.

Experimental procedures Definition of population

Population was restricted to the 27 countries of the EU. We restricted age to 15–64 years, as death certificates are less reliable with respect to cause of death in older age (Harteloh et al., 2010),  and especially for the very old (Alperovitch et al., 2009). Also, the relative risks for alcohol-attributable causes tend to decrease with age (Klatsky and Udaltsova, 2007). As a result, the consequences of consumption, both detrimental and beneficial, tend to be exaggerated in the older age groups when age-unspecific relative risks from meta-analyses are used as is commonly the case. A second reason to restrict ourselves to people less than 65 years of age was our aim to concentrate on premature mortality, as this age threshold has often been used in the European context to define premature mortality (Lapostolle et al., 2008). The under 15 year-olds were excluded since alcohol-attributable deaths in this age group are very rare, except as the result of the impact of someone else’s drinking (e.g., traffic fatalities caused by drunk drivers).

2.2.

Definition of exposure categories

Alcohol consumption status was measured in three categories: lifetime abstainers, former and current drinkers. For current drinkers, average alcohol consumption in grams of pure alcohol per day was used as the main exposure variable. Heavy drinking was defined as the average daily consumption of alcohol at or above 3.33 standard drinks for women (40+ g of pure alcohol) and 5 standard drinks for men (60+ g of pure alcohol). The heavy drinking categories were taken from the European Medicines Agency guideline on the development of medicinal products for the treatment of AD (European Medicines Agency, 2010) based on the monitoring guide of the World Health Organization (World Health Organization, 2000; see Web Appendix A—2.1 for additional details). Alcohol dependence was operationalized by standardized assessment based on either ICD or DSM (Rehm et al., 2005). Exposure data for alcohol consumption on a country basis were obtained from the World Health Organization Regional Office for Europe which was processed by the routines implemented in the Global Information System on Alcohol and Health (for more details see Shield et al., 2012b and http://www.who.int/gho/alcohol/en/ index.html). Following the methodology developed for the Com parative Risk Assessment for alcohol of the GBD 2005/2010 study (Kehoe et al., 2012; Rehm et al., 2010b), information from general population surveys was triangulated with adult per capita consump tion to correct for undercoverage (Rehm et al., 2007).

Impact of AD on mortality burden Data on the prevalence of AD per country were taken from Rehm and colleagues (Rehm et al., 2012). These data are summarised with sources in Web Appendix A—1. Data on risk relations for alcohol or heavy drinking as exposure were obtained from the above-mentioned ongoing GBD Comparative Risk Assessment for alcohol (Rehm et al., 2010a). Data on risk relations for AD are described below in the modeling strategy. Data on mortality were obtained from the 2004 update of the GBD study (World Health Organization, 2008). We relied on these data rather than on country statistics for reasons of comparability between countries (Lopez et al., 2006).

2.3.

Modeling strategy

2.3.1. Estimating mortality attributable to alcohol and to heavy drinking In order to estimate alcohol-attributable mortality, we calculated the sex-, age and disease-specific Alcohol-Attributable Fraction (AAF) for all alcohol-attributable diseases and applied these AAFs to respective mortality cases as obtained from the 2004 GBD study (Shield et al., 2012b; World Health Organization, 2008). An AAF represents the proportion of each cause of death that is attributable to alcohol consumption, on the basis of there being the counterfactual scenario of no alcohol consumption (i.e., everyone is a lifetime abstainer; for definitions see (Benichou, 2001; Walter, 1976, 1980)). Similarly, attributable fractions for heavy drinking were computed. Overall, the same algorithms as the current GBD study were used (Kehoe et al., 2012; Rehm et al., 2007, 2010a, 2010b; Shield et al., 2012b). 2.3.2. Estimating mortality attributable to AD The mortality attributable to AD can be estimated by a similar methodology in principle as was used for alcohol-attributable mortality (Walter, 1976, 1980), i.e., by calculating the proportion of deaths attributable to AD based on prevalence of exposure and sex- and age-specific relative risk estimates for AD. In addition, two subpopulations of people with AD were distinguished based on their mortality risk:





For 20% of the people with AD in any given country, we utilized the risk relations from Hayes et al. (2011), assuming that 20% reflected the proportion of all people with AD who were in treatment or with similarly severe dependence (Alonso et al., 2004; Kohn et al., 2004). For the remaining 80% of the people with AD, we assumed a relative risk of 2.25 for men and 2.7 for women up to age 44, and of 2 for men and 2.4 for women aged 45 and older, thus reflecting the average level of severity of untreated AD in the general population (for men there are three general population studies of mortality risks, two involving a very long follow-up period (Murphy et al., 2008; Ojesj¨ o et al., 1998; Vaillant, 1996); in addition there are several studies without separation of sex (Dawson, 2000; Fichter et al., 2011)).

2.3.3. Estimating the effect of treatment interventions The effects of treatment of AD have traditionally been measured as a downward shift of alcohol consumption (either to abstinence or to reduced drinking (Barbosa et al., 2010)), or by a decrease in the relative risk of dying (as a result of reduced and more controlled use of alcohol (McQueen et al., 2011)). In order to estimate the effects of treatment interventions on the mortality in a given country, the drinking population of the country was simulated with a set of 100,000 points, each representing one drinker. The average consumption of each ‘‘drinker’’ was simulated using the known drinking

91 prevalence distribution for each country. It was assumed that people with AD in need of treatment would be among the heavier drinkers in the AD population (men consuming 72 g of pure alcohol or more per day, and women consuming 48 g of pure alcohol or more per day). People with AD who were also heavy drinkers were thus randomly selected from the heavy drinking populations, and the treatment effects were applied to each of them. We simulated the effects of treating 10%, 20%, 30% and 40% of all people with AD with five different interventions which were selected based on the seminal review of Hester and Miller (2003), for a detailed summary of assumptions see Table 1). Each drinker’s risk of dying was then evaluated in this treated drinking population and the AAFs were then calculated in the same fashion as described above. Comparing the newly derived AAFs to the original AAFs allowed us to estimate the lives saved within one year by a given treatment intervention. Additional details on the methodology can be found in Web Appendix A—2.2. 2.3.4. Estimating uncertainty Uncertainty of the AAFs and the mortality attributable to alcohol consumption and heavy drinking were computed using Monte Carlo methods (Ross, 2006). Each parameter of the function was randomly sampled 40,000 times based on their distribution, and these random sets of variables were used to calculate the variance of the AAFs (Gmel et al., 2011b; see Web Appendix A—2.1). The more samples there are, the more closely the estimated variance approaches the true variance. The uncertainty of the estimates of mortality attributable to AD was calculated using derivatives (Gmel et al., 2011a for a similar methodology). The variance of the estimates for the intervention is described in Web Appendix A—2.2.

3.

Results

The prevalence of alcohol consumption among adults 15–64 years of age was high in the EU: 88.9% of men and 82.1% of women were current drinkers (total 85.5%), and 9.4% of men and women combined, quite a large proportion, drank above the thresholds set for heavy drinking (see above for definition: 15.3% of men and 3.4% of women). There are some regional differences: the average level of alcohol consumption is highest in the Central and Eastern parts of the EU (Rehm et al., 2012; Shield et al., 2012b). Regional differences are more pronounced in patterns of drinking, especially heavy drinking occasions, which are more detrimental in Nordic countries, the Central-East and Eastern regions, and in the British Isles (Rehm et al., 2012; Shield et al., 2012b), and less detrimental in Mediterranean Southern Europe. The prevalence of AD was estimated to be 5.4% for men and 1.5% for women (overall 3.5%). Again, there are notable differences between regions: countries located in Southern Europe (primarily along the Mediterranean Sea) had the lowest AD rates, with 0.6% for women and 1.7% for men (see Web Appendix A—1). These primarily wine drinking countries not only have had the lowest level of average consumption, but consistently have displayed the least detrimental drinking patterns in the EU and even globally (Rehm et al., 2003, 2004). As both average level of consumption and frequency of heavy drinking occasions have been linked to AD (Caetano et al., 1997; Dawson et al., 2008), it is no surprise that the Southern European countries exhibited the lowest prevalence rates of AD. However, other possible reasons for these observed AD prevalence rates are discussed below.

92

Table 1

Overview of assumptions for modeling interventions.

Interventions

Main results (effects assumed to be stable for Risk relations 1 year)

MI: Smedslund et al. (2011) and CBT: Magill & Ray, (2009)

The usual dose-dependent risk relations between average consumption of alcohol and disease outcomes were used, multiplied by 2 to account for the overall higher mortality risk of people with AD (Harris & Barraclough, 1998). For injury, the RR from Harris & Barraclough, 1998 was used for AD and the RR from Corrao et al. (2004) was used for non-dependent people.

Smedslund et al. (2011)

McQueen et al. (2011) and Room et al. (2005) McQueen et al. (2011)

Pooled estimates of two reviews of R¨ osner et al. (2010a). For this simulation we are concerned about the differences in consumption between baseline and followup in the group receiving medications only.

J. Rehm et al.

Motivational Interviewing (MI) and For MI, an average drop of 15.8 g of pure Cognitive Behavioral Therapy (CBT) alcohol per day was assumed (measured against 1 no intervention; 95% CI from 9.6 g to 21.8 g of pure alcohol). The effect after one year was very small and not significant (average: 1.2 g of pure alcohol reduction per day), the average effect over the year was a 3.2 g reduction of pure alcohol per day (95% CI: 1.2 g to 5.2 g of pure alcohol per day).For CBT, almost the same effect was found in studies with a notreatment control as the comparison condition (15.9 g of pure alcohol per day). In addition, Project MATCH did not find any significant differences (Project MATCH Research Group, 1997) between MI and CBT.We modeled the results based on a drop of15.8 g per day over the year. Motivational Interviewing and An average drop of 21.8 g of pure alcohol per Cognitive Behavioral Therapy 2 day was assumed as the upper limit of the CI for MI/CBT (see above). We assumed proportional CIs compared to the first MI/CBT scenario. Brief Intervention (BI) 1 An average drop of 13.5 g of pure alcohol per day with a 95% CI from 2.7 g to 24.5 g of pure alcohol. Brief Intervention 2 An average reduction of the RR for mortality by 0.6 (95% CI: from 0.40 to 0.91). This scenario represents the ‘‘best case’’ for BI, as hospitalization is linked to mortality, and AD plays an important role in mediating and moderating premature mortality (e.g. De Lorenze et al., 2011; O’Brien et al., 2007). However, similar effects were obtained in an meta-analyses on all BIs (Cuijpers et al., 2004). For 55.0% of the patient population there was a Pharmacological therapy (for reduction in drinking of 13% on average; for simulation, the effects of 18.1% of the patient population a reduction in Randomized Controlled Trials of acamprosate and opioid antagonist drinking of 50%; and for 26.8% of the population treatment were combined) abstinence resulted.

Source

Impact of AD on mortality burden

93

Nordic countries with overall relatively low adult per capita consumption, such as Sweden, have a proportionally much higher percentage of AD. The Central East and Eastern European countries, and in particular the Baltic countries, have the highest AD prevalence, stemming primarily from the combination of high overall consumption and detrimental drinking patterns, including, but not limited to, heavy binge drinking (see Rehm et al., 2005 for further explanation). Figure 1 summarizes these different mortality burdens for the EU (again for the population aged 15–64 years): the overall mortality burden from alcohol consumption (only the detrimental effects on mortality without considering any beneficial effects of moderate drinking on diabetes (Baliunas et al., 2009) and ischemic disease (Puddey et al., 1999; Rehm and Roerecke, 2011)); the net burden, calculated by subtracting the beneficial effects from the detrimental effects; the burden of heavy drinking using the European Medicines Agency risk categories of 40+ g of pure alcohol for women and 60+ g of pure alcohol for men described above; and the mortality burden due to AD. Overall, alcohol consumption contributes greatly to mortality. Based on the net burden, 1 in 7 deaths in men and 1 in 13 deaths in women in the EU are estimated to be caused by alcohol consumption. As can be seen clearly, most of the mortality due to alcohol consumption stems from heavy drinking (about 77% of the net burden, 67% of the overall burden). AD comprises 71% of the net burden and 62% of all harmful alcohol-attributable burden. In other words, the majority of alcohol-attributable burden is due to AD, presumably by means of heavy drinking. In the EU as a whole, and in nearly every country individually (detailed results available from the authors but not shown), pharmacotherapy is the most effective treatment, closely followed by brief interventions in the hospital. This order of effectiveness of treatment is reversed if only men are considered. The other three treatment scenarios show significantly less effectiveness in reducing mortality. See Web Appendices A—3 and 4. Increasing treatment coverage to 40% of all people with AD could lead to a reduced burden of alcohol-attributable mortality of about 13% for men and 9% for women. This represents a little less than 12,000 lives saved in the EU within 12 months (10,000 men and 1700 women). Moreover, this analysis only considers diseases where alcohol has only a direct effect within one year; i.e., it excludes latent effects, such as

the effects of alcohol on cancers (International Agency for Research on Cancer, 2010). It can be hypothesised that the number of deaths avoided would be significantly higher in the longer term. Although five times more deaths would be avoided in men compared to women (see Figures 2 and 3), this apparent gap is put into perspective when looking at the proportion of alcohol-related deaths avoided, which is much more similar between the sexes.

4.

Discussion

We have shown that alcohol consumption is the cause of a large premature mortality burden in the EU, that the majority of this burden seems to be due to AD, and that more than 10% of this burden could potentially be removed by increasing treatment capacity and coverage for AD. Before we discuss implications, the potential limitations of the key results should be examined. With respect to the mortality caused by alcohol consumption and heavy drinking in general, our results confirm the results of the past GBD risk assessments (Rehm et al., 2009; World Health Organization, 2009), but, of course, the same limitations mentioned there apply here. Bias may be introduced by the modeling of exposure with the triangulation of surveys and adult per capita consumption including unrecorded consumption where 80% of the adult per capita consumption was assumed to be consumed (Rehm et al., 2010b). Unrecorded consumption, by its very nature, is hard to estimate, and these estimates contain measurement error (World Health Organization, 2011). Finally, the risk estimates taken from the meta-analyses may not apply to all EU countries, especially with respect to injury where the risk may be dependent on societal and socio-technical contexts. With respect to the triangulation, various sensitivity analyses show that the burden remains substantial under different methodologies (Rehm et al., 2006; Rey et al., 2010; Shield et al., 2012a). Unrecorded consumption is less important in Europe, where only 16% of the overall consumption falls into this category (Shield et al., 2012b), and, thus, its impact on the overall results is not significant. Finally, most of the underlying studies in the meta-analyses used are from the EU or from high income countries with similar conditions as the EU, so that the relative risks seem to be comparable.

25.0% Men

Women

Total

20.0% 15.0% 10.0% 5.0% 0.0%

Figure 1

Alcohol-attributable

Alcohol-attributable (net)

Heavy drinking

Alcohol dependence

Men

16.1%

13.9%

11.1%

10.7%

Women

8.5%

7.7%

5.3%

3.7%

Total

13.6%

11.8%

9.2%

8.4%

Proportion of mortality attributable to alcohol consumption and AD for the EU in 2004; 15–64 year olds.

94

J. Rehm et al.

Figure 2 Number of deaths avoided over one year in men by treatment for AD in the EU in 2004 by five different treatment modalities.

Figure 3 Number of deaths avoided over one year in women by treatment for AD in the EU in 2004 by five different treatment modalities.

Thus, despite these limitations, the overall conclusion that alcohol consumption is responsible for a large degree of premature mortality seems to be justified. With respect to the impact of AD, the above conclusions rely on the underlying assumptions, i.e., the validity of the prevalence data and the validity of the relative risks. With respect to prevalence, it should be mentioned that there may be additional cultural factors which explain the differences between regions. While alcohol consumption is deeply culturally embedded in Southern Europe, alcohol problems and dependence are tabooed in this region, and, thus, these countries may also have a higher tendency to deny dependence symptoms for reasons of stigma (Room and M¨ akel¨ a, 2000). In the rest of Europe, and especially in the Nordic countries and Central-East and Eastern European countries, where AD prevalence is notably higher (see Web Appendix A—1), drinking to intoxication and showing its effects are

much less of a taboo. As a result, there may be less denial of AD symptoms explaining the relatively higher prevalence. While there have been problems with measuring the prevalence of AD in the general population of the different countries, there are sufficient, well-conducted studies to indicate that the estimated overall prevalence of AD of about 3.5% is not an overestimate for Europe. However, some of the single country prevalence estimates may be under- or overestimated. The overall estimate for Europe in our study is also quite close to other estimates for Europe based on different methodologies (Rehm et al., 2005; Wittchen et al., 2011) as well as to estimates from other high income countries such as the US (3.8% cf. Hasin et al., 2007). The second underlying assumption concerns the validity of the relative risks. The mortality risks used above depend on the pooled estimates for general population studies and the use of one of the largest recent studies for the

Impact of AD on mortality burden treatment samples. Methodologically rigorous metaanalyses are necessary to quantify the risks for both the general population and for treatment samples. However, while such formal analyses may result in slightly different estimates, the overall conclusion appears to be valid. Finally, with respect to the potential impact of treatment, while we used the most rigorously controlled results mainly from Cochrane reviews, there is always the question to what degree such studies can be generalized into everyday practice (Carroll and Rounsaville, 2007; Hollon, 1996; Nathan et al., 2000). This discussion is usually subsumed under the headings of ‘‘efficacy’’ versus ‘‘effectiveness’’ (Streiner, 2002). Randomised clinical trials have comprehensive exclusion criteria, often excluding co-morbidities which are typical for people with AD who are seeking treatment (Blanco et al., 2008). There are also social characteristics which have been shown to differ between community samples of people with AD and participants in randomised clinical trials (Grant et al., 2007). As our simulation study in some scenarios included up to 40% of all people with AD in the general population, these above-noted differences may become relevant. However, it is hard to predict the direction of the bias. We have clear examples of treatment where effectiveness was better than efficacy, such as in modern methods of treating depression, as well as for the opposite case where efficacy was better than effectiveness. Overall, we modeled interventions based on literature which was not exclusively on people with AD, but included problem drinkers as well. Many of the underlying trials did not formally measure AD. Assuming the same effect size for AD, treatment seems justified, as motivational interviewing and cognitive behavioral therapy are widely used in AD treatment with good effects (Hester & Miller 2003; Morgenstern et al., 2001). Brief interventions are usually used for problem drinkers or mild dependence only. We included brief interventions only in hospital settings, as a best case scenario, where the effect on mortality had been assessed in a situation where patients experience higher risk due to other health conditions. If alcohol consumption is reduced in such a situation, the effects are more immediate. On the other hand, the effects of pharmacotherapy, which overall was found most effective, were mainly tested in people with more severe dependence. An important question is the feasibility of treating with pharmacotherapy 40% of people with AD, and we have very little experience with respect to pharmacological AD treatment in the community. Currently, pharmacological treatment is used only in a minority of treatment situations in the EU (Alonso et al., 2004), and most often not in community or general practitioner settings, but in specialised treatment centers. Studies examining the applicability of pharmacological approaches in non-specialised settings should be a priority for future research. Overall, increasing treatment availability was shown to have the potential for increasing public health in Europe. In fact, the order of magnitude of potential gains in mortality shown above is comparable to potential gains from other interventions (Rehm et al., 2011) usually recommended as best policies to reduce alcohol-attributable harms and to increase public health (World Health Organization, 2010) for Europe (see the recently adopted WHO action plan for 2012–2020 to reduce the harmful use of alcohol in the

95 European region http://www.euro.who.int/__data/assets/ pdf_file/0006/147732/wd13E_Alcohol_111372.pdf). However, increasing treatment availability should not replace current alcohol policies in Europe, but rather should be implemented in addition to proven, cost-effective alcohol policy measures such as increasing price, banning advertising, or restricting availability of alcoholic beverages (Anderson et al., 2009; Chisholm et al., 2004), as different measures are directed at different sub-populations. Prevention and treatment are both important to reduce the alcohol-attributable burden in Europe.

Role of the funding source This study was supported by an unrestricted educational Grant from Lundbeck, a contract from the WHO Regional Office for Europe, and by salary support to the first author from the Ontario Ministry of Health and Long-Term Care. None of the funders had any role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. The authors have full access to all the data in the study and had final responsibility for the decision to submit the paper for publication.

Contributors JR conceptualized the whole study, obtained funding, co-developed the methodology of the modeling, co-wrote the draft, and contributed to writing the final version of the article. KDS and GG co-developed the methodology, carried out the statistical calculations and main modeling, wrote the draft description of the methodology and part of Web Appendices A, and contributed to writing the final version of the article. MXR helped in writing the final draft, obtained information and data sources, carried out descriptive statistics and participated in writing the final version of the article. UF co-developed the modeling and contributed to writing the final version of the article.

Conflicts of interest The first author is on the Medical Advisory Board of Nalmefene, a pharmaceutical which has been submitted for alcohol dependence treatment. The various funding sources have been declared separately, but none had any impact on the design of the study. In addition to this no conflicts of interests are declared.

Acknowledgments We would like to thank Peter Anderson and Uli Wittchen for comments on the overall study, as well as Michelle Tortolo for referencing various versions of this manuscript.

Appendix A.

Supplementary material

Supplementary data associated with this article can be found in the online version at http://10.1016/j.euroneuro.2012.08.001.

References Alonso, J., Angermeyer, M.C., Bernert, S., Bruffaerts, R., Brugha, T.S., 2004. ESEMeD/MHEDEA 2000 Investigators, 2004. Use of

96 mental health services in Europe: results from the European Study of the Epidemiology of Mental Disorders (ESMeD) project. Acta Psychiatr. Scand. 109, 47–54.  Alpe rovitch, A., Bertrand, M., Jougla, E., Vidal, J.S., Ducimetiere, P., Helmer, C., 2009. Do we really know the cause of death of the very old? Comparison between official mortality statistics and cohort study classification. Eur. J. Epidemiol. 24, 669–675. American Psychiatric Association, 1994. Diagnostic and statistical manual of mental disorders. American Psychiatric Association, Washington, DC. Anderson, P., Chisholm, D., Fuhr, D., 2009. Effectiveness and costeffectiveness of policies and programmes to reduce the harm caused by alcohol. Lancet 373, 2234–2246. Baliunas, D., Taylor, B., Irving, H., Roerecke, M., Patra, J., Mohapatra, S., Rehm, J., 2009. Alcohol as a risk factor for type 2 diabetes—a systematic review and meta-analysis. Diabetes Care 32, 2123–2132. Barbosa, C., Taylor, B., Godfrey, C., Rehm, J., Parrott, S., Drummond, C., 2010. Modelling lifetime QALYs and health care costs from different drinking patterns over time: a Markov model. Int. J. Methods Psychiatr. Res. 19, 97–109. Benichou, J., 2001. A review of adjusted estimators of attributable risk. Stat. Methods Med. Res. 10, 195–216. Blanco, C., Olfson, M., Okuda, M., Nunes, E.V., Liu, S.M., Hasin, D.S., 2008. Generalizability of clinical trials for alcohol dependence to community samples. Drug Alcohol Depend. 98, 123–128. Caetano, R., Tam, T., Greenfield, T., Cherpitel, C., Midanik, L., 1997. DSM-IV alcohol dependence and drinking in the U.S. population: a risk analysis. Ann. Epidemiol. 7, 542–549. Campos, J., Roca, L., Gude, F., Gonzalez-Quintela, A., 2011. Longterm mortality of patients admitted to the hospital with alcohol withdrawal syndrome. Alcohol Clin. Exp. Res. 35, 1180–1186. Carroll, K.M., Rounsaville, B.J., 2007. A vision of the next generation of behavioral therapies research in the addictions. Addiction 102, 850–862. Chisholm, D., Rehm, J., van Ommeren, M., Monteiro, M., 2004. Reducing the global burden of hazardous alcohol use: a comparative cost-effectiveness analysis. J. Stud. Alcohol 65, 782–793. Corrao, G., Bagnardi, V., Zambon, A., La Vecchia, C., 2004. A metaanalysis of alcohol consumption and the risk of 15 diseases. Prev. Med. 38, 613–619. Cuijpers, P., Riper, H., Lemmers, L., 2004. The effects on mortality of brief interventions for problem drinking: a meta-analysis. Addiction 99, 839–845. Dawson, D.A., 2000. Alcohol consumption, alcohol dependence, and all-cause mortality. Alcohol Clin. Exp. Res. 24, 72–81. Dawson, D.A., Li, T.K., Grant, B.F., 2008. A prospective study of risk drinking: at risk for what? Drug Alcohol Depend. 95, 62–72. De Lorenze, G.N., Weisner, C., Tsai, A.L., Satre, D.D., Quesenberry, C.P.J., 2011. Excess mortality among HIV-infected patients diagnosed with substance use dependence or abuse receiving care in a fully integrated medical care program. Alcohol Clin. Exp. Res. 35, 203–210. European Medicines Agency, 2010. Guideline on the Development of Medicinal Products for the Treatment of Alcohol Dependence. European Medicines Agency, London, UK. Fichter, M.M., Quadflieg, N., Fischer, U.C., 2011. Severity of alcohol-related problems and mortality: results from a 20-year prospective epidemiological community study. Eur. J. Psychiatry Clin. Neurosci. 261, 293–302. Gmel, G., Shield, K., Rehm, J., 2011a. Developing a methodology to derive alcohol-attributable fractions for HIV/AIDS mortality based on alcohol’s impact on adherence to antiretroviral medication. Popul. Health Metr. 9, 5. Gmel Jr., G, Shield, K., Frick, H., Kehoe, T., Gmel Sr., G, Rehm, J., 2011b. Estimating uncertainty for alcohol-attributable fractions for infectious and chronic disease. BMC Med. Res. Methodol. 11, 48.

J. Rehm et al. Grant, B.F., Harford, T.C., Muthen, B.O., Yi, H.Y., Hasin, D.S., Stinson, F.S., 2007. DSM-IV alcohol dependence and abuse: further evidence of validity in the general population. Drug Alcohol Depend. 86, 154–166. Harris, E.C., Barraclough, B., 1998. Excess mortality of mental disorder. Br. J. Psychiatry 173, 11–53. Harteloh, P., de Bruin, K., Kardaun, J., 2010. The reliability of cause-of-death coding in the Netherlands. Eur. J. Epidemiol. 25, 531–538. Hasin, D.S., Stinson, F.S., Grant, B.F., 2007. Prevalence, correlates, disability and comorbidity of DSM-IV alcohol abuse and dependence in the United States: results from the National Epidemiologic Survey on Alcohol and Related Conditions. Arch. Gen. Psychiatry 64, 830–842. Hayes, R.D., Chang, C.K., Fernandes, A., Broadbent, M., Lee, W., Hotopf, M., Stewart, R., 2011. Associations between substance use disorder sub-groups, life expectancy and all-cause mortality in a large British specialist mental healthcare service. Drug Alcohol Depend. 118, 56–61. Hester, R.K., Miller, W.R., 2003. Handbook of alcoholism treatment approaches. Effective alternatives, third edition Allyn and Bacon, Boston, USA. Hollon, S.D., 1996. The efficacy and effectiveness of psychotherapy relative to medications. Am. Psychol. 51, 1025–1030. International Agency for Research on Cancer, 2010. Alcoholic Beverage Consumption and Ethyl Carbamate (Urethane). IARC Monograph 96 on the Evaluation of Carcinogenic Risks to Humans. International Agency for Research on Cancer (IARC), Lyon, France. Iontchev, A., 1998. Central and Eastern Europe. In: Grant, M. (Ed.), Alcohol and Emerging Markets: Patterns, Problems, and Responses. International Center for Alcohol Policies, Washington, DC, pp. 177–201. Kehoe, T., Gmel, G.j., Shield, K., Gmel, G.s., Rehm, J., 2012. Modelling alcohol consumption as a distribution and determining the impact of the distribution on estimated alcohol-attributable harms. Popul. Health Metr. 10(1), 6. Klatsky, A.L., Udaltsova, N., 2007. Alcohol drinking and total mortality risk. Ann. Epidemiol. 17, S63–S67. Kohn, R., Saxena, S., Levav, I., Saraceno, B., 2004. The treatment gap in mental health care. Bull. World Health Organ 82, 858–866. Langlois, K.A., Samokhvalov, A.V., Rehm, J., Spence, S.T., Gorber, S.K., 2011. Health state descriptions for Canadians: mental illnesses. Statistics Canada, Ottawa, Canada. Lapostolle, A., Lefranc, A., Gremy, I., Spira, A., 2008. Measure of premature mortality: comparison of deaths before age 65 and  emiologie  expected years of life lost. Rev. d’epid et de sante publique 56, 245–252. Lopez, A.D., Mathers, C.D., Ezzati, M., Jamison, D.T., Murray, C.J.L., 2006. Global burden of disease and risk factors. The World Bank and Oxford University Press, New York and Washington. Magill, M., Ray, L.A., 2009. Cognitive-behavioral treatment with adult alcohol and illicit drug users: a meta-analysis of randomized controlled trials. J. Stud. Alcohol Drugs 70, 516–527. McQueen, J., Howe, T.E., Allan, L., Mains, D., Hardy, V., 2011. Brief interventions for heavy alcohol users admitted to general hospital wards. Cochrane Database Syst. Rev. 8, CD005191. Morgenstern, J., Blanchard, K.A., Morgan, T.J., Labouvie, E., Hayaki, J., 2001. Testing the effectiveness of cognitivebehavioral treatment for substance abuse in a community setting: within treatment and posttreatment findings. J. Consult. Clin. Psychol. 69, 1007–1017. Murphy, J.M., Burke Jr., J.D., Monson, R.R., 2008. Mortality associated with depression—a forty-year perspective from the Stirling County Study. Soc. Psychiatry Psychiatr. Epidemiol. 43, 594–601. Nathan, P.E., Scott, S.P., Dolan, S.L., 2000. Research on psychotherapy efficacy and effectiveness: between Scylla and Charybdis? Psychol. Bull. 126, 964–981.

Impact of AD on mortality burden O’Brien, J.M.J., Lu, B., Ali, N.A., Martin, G.S., Aberegg, S.K., Marsh, C.B., Lemeshow, S., Douglas, I.S., 2007. Alcohol dependence is independently associated with sepsis, septic shock, and hospital mortality among adult intensive care unit patients. Crit. Care Med. 35, 345–350. Ojesj¨ o, L., Hagnell, O., Otterbeck, L., 1998. Mortality in alcoholism among men in the Lundby Community Cohort, Sweden: a fortyyear follow-up. J. Stud. Alcohol 59, 140–145. ¨ st¨ Ormel, J., VonKorff, M., U un, T.B., Pini, S., Korten, A., Oldehinkel, T., 1994. Common mental disorders and disability across cultures. Results from the WHO Collaborative Study on psychological problems in general health care. J. Am. Med. Assoc. 272, 1741–1748. Popova, S., Rehm, J., Patra, J., Zatonski, W., 2007. Comparing alcohol consumption in central and eastern Europe to other European countries. Alcohol Alcohol 42, 465–473. Project MATCH Research Group, 1997. Matching alcoholism treatment to client heterogeneity: project MATCH posttreatment drinking outcomes. J. Stud. Alcohol 58, 7–30. Puddey, I.B., Rakic, V., Dimmitt, S.B., Beilin, L.J., 1999. Influence of pattern of drinking on cardiovascular disease and cardiovascular risk factors—a review. Addiction 94, 649–663. Rehm, J., Baliunas, D., Borges, G.L.G., Graham, K., Irving, H.M., Kehoe, T., Parry, C.D., Patra, J., Popova, L., Poznyak, V., Roerecke, M., Room, R., Samokhvalov, A.V., Taylor, B., 2010a. The relation between different dimensions of alcohol consumption and burden of disease—an overview. Addiction 105, 817–843. Rehm, J., Kehoe, T., Gmel, G., Stinson, F., Grant, B., Gmel, G., 2010b. Statistical modeling of volume of alcohol exposure for epidemiological studies of population health: the example of the US. Popul. Health Metr. 8, 3. Rehm, J., Klotsche, J., Patra, J., 2007. Comparative quantification of alcohol exposure as risk factor for global burden of disease. Int. J. Methods Psychiatr. Res. 16, 66–76. Rehm, J., Mathers, C., Popova, S., Thavorncharoensap, M., Teerawattananon, Y., Patra, J., 2009. Global burden of disease and injury and economic cost attributable to alcohol use and alcohol use disorders. Lancet 373, 2223–2233. Rehm, J., Patra, J., Gnam, W., Sarnocinska-Hart, A., Popova, S., 2011. Avoidable cost of alcohol abuse in Canada. Eur. Addict. Res. 17, 72–79. Rehm, J., Patra, J., Popova, S., 2006. Alcohol-attributable mortality and potential years of life lost in Canada 2001: implications for prevention and policy. Addiction 101, 373–384. Rehm, J., Rehn, N., Room, R., Monteiro, M., Gmel, G., Jernigan, D., Frick, U., 2003. The global distribution of average volume of alcohol consumption and patterns of drinking. Eur. Addict. Res. 9, 147–156. Rehm, J., Roerecke, M., 2011. Alcohol, the heart and the cardiovascular system—what do we know and where should we go? Drug Alcohol Rev. 30, 335–337. Rehm, J., Room, R., Monteiro, M., Gmel, G., Graham, K., Rehn, N., Sempos, C.T., Frick, U., Jernigan, D., 2004. Alcohol use. In: Ezzati, M., Lopez, A.D., Rodgers, A., Murray, C.J.L. (Eds.), Comparative Quantification of Health Risks: Global and Regional Burden of Disease Attributable to Selected Major Risk Factors. World Health Organization, Geneva, Switzerland, pp. 959–1109. Rehm, J., Room, R., Van den Brink, W., Jacobi, F., 2005. Alcohol use disorders in EU countries and Norway: an overview of the epidemiology. Eur. Neuropsychopharmacol. 15, 377–388. Rehm, J., Shield, K., Rehm, M., Gmel, G.j., Frick, U., 2012. Alcohol consumption, alcohol dependence, and attributable burden of disease: potential gains from effective interventions for alcohol dependence. Centre for Addiction and Mental Health, Toronto, ON. Rey, G., Boniol, M., Jougla, E., 2010. Estimating the number of alcohol-attributable deaths: methodological issues and illustration with French data for 2006. Addiction 105, 1018–1029. Roerecke, M., Rehm, J., 2012. Alcohol use disorders and mortality—a systematic review and meta-analysis. Centre for Addiction and Mental Health, Toronto, ON.

97 Room, R., Babor, T., Rehm, J., 2005. Alcohol and public health: a review. Lancet 365, 519–530. Room, R., 2010. Alkogol’naya politika: polozhenie del i problemy v Evrope i Severnoi Azii (Alcohol politics: status of the problem in Europe and North Asia). In: Khalturnia, D.A., Korotaev, A.V. (Eds.), Alkogol’naya katastrofa i vozmozhnosti gosudarstvennoi politiki v preodolenii alkogol’noi sverkhsmertnosit v Rossii (The Alcohol Catastrophe and the Possibilities of Public Policy in Overcoming Extreme Alcohol Mortality in Russia) 2nd ed. URSS, Moscow, pp. 232–262. Room, R., M¨ akel¨ a, K., 2000. Typologies of the cultural position of drinking. J. Stud. Alcohol 61, 475–483. R¨ osner, S., Hackl-Herrwerth, A., Leucht, S., Lehert, P., Vecchi, S., Soyka, M., 2010a. Acamprosate for alcohol dependence. Cochrane Database Syst. Rev., 9.CD004332. R¨ osner, S., Hackl-Herrwerth, A., Leucht, S., Vecchi, S., Srisurapanont, M., Soyka, M., 2010b Opioid antagonists for alcohol dependence. Cochrane Database Syst. Rev., 12. CD001867. Ross, S.M., 2006. Simulation, statistical modeling and decision science, 4th edition Elsevier Academic Press, Amsterdam, NL. Samokhvalov, A.V., Popova, S., Room, R., Ramonas, M., Rehm, J., 2010. Disability associated with alcohol abuse and dependence. Alcohol Clin. Exp. Res. 34, 1871–1878. Shield, K., Kehoe, T., Taylor, B., Patra, J., Rehm, J., 2012a. Mortality and potential years of life lost attributable to alcohol consumption for individuals 0 to 64 years of age in Canada in 2005. BMC Public Health 12, 91. Shield, K.D., Kehoe, T., Gmel, G.j., Rehm, M.X., Rehm, J., 2012b. Societal burden of alcohol. In: Anderson, P., Møller, L., Galea, G. (Eds.), Alcohol in the European Union. Consumption, Harm and Policy Approaches. World Health Organization Regional Office for Europe, Copenhagen, Denmark, pp. 10–28. Smedslund, G., Berg, R.C., Hammerstrom, K.T., Steiro, A., Leiknes, K.A., Dahl, H.M., Karlsen, K., 2011. Motivational interviewing for substance abuse. Cochrane Database Syst. Rev. 5, CD008063. Streiner, D.L., 2002. The 2 ‘‘Es’’ of research: efficacy and effectiveness trials. Can. J. Psychiatry 47, 552–556. Vaillant, G.E., 1996. A long-term follow-up of male alcohol abuse. Arch. Gen. Psychiatry 53, 243–249. Walter, S.D., 1976. The estimation and interpretation of attributable risk in health research. Biometrics 32, 829–849. Walter, S.D., 1980. Prevention of multifactorial disease. Am. J. Epidemiol. 112, 409–416. Wittchen, H.U., Jacobi, F., Rehm, J., Gustavsson, A., Svensson, M., Jonsson, B., Olesen, J., Allgulander, C., Alonso, J., Faravelli, C., Fratiglioni, L., Jennum, P., Lieb, R., Maercker, A., van Os, J., Preisig, M., Salvador-Carulla, L., Simon, R., Steinhausen, H.C., 2011. The size and burden of mental disorders and other disorders of the brain in Europe 2010. Eur. Neuropsychopharmacol. 21, 655–679. World Health Organization, 2000. International Guide for Monitoring Alcohol Consumption and Related Harm. World Health Organization, Department of Mental Health and Substance Dependence. Noncommunicable Diseases and Mental Health Cluster, Geneva, Switzerland. World Health Organization, 2008. The Global Burden of Disease: 2004 Update. World Health Organization, Geneva, Switzerland. World Health Organization, 2009. Global Health Risks. Mortality and Burden of Disease Attributable to Selected Major Risks. World Health Organization, Geneva, Switzerland. World Health Organization, 2010. Global Strategy to Reduce the Harmful Use of Alcohol. World Health Organization, Geneva, Switzerland. World Health Organization, 2011. Global Status Report on Alcohol and Health. World Health Organization, Geneva, Switzerland.