The Particular Story of Italians' Relation with Alcohol: Trends in Individuals' Consumption by Age and Beverage Type

Aims

In Italy, sales data show a dramatic fall in alcohol consumption between 1970 and 2010. The aim of our study is to provide updated information on trends, prevalence and determinants of alcohol drinking in Italy, using individual-level data.

Methods

Seven nationally representative cross-sectional surveys were conducted in Italy between 2006 and 2014, on a total sample of 21,416 participants aged ≥15 years, with available information on weekly consumption of wine, beer and spirits.

Results

Per capita alcohol consumption decreased by 23% between 2006 and 2014 (from 5.6 to 4.4 drinks/week), due to the fall in wine drinking. Overall, the prevalence of alcohol drinkers was 61.6%. Individuals predominantly drinking wine were 35.3%, beer 11.1% and spirits 6.4%. A direct trend of drinking with age was observed for total alcohol (multivariate odds ratio, OR = 1.75 for ≥65 vs. 15–24 years) and predominant wine drinking (OR = 8.05), while an inverse trend was observed for beer (OR = 0.17) and spirit drinkers (OR = 0.33). Women (OR = 0.24), obese individuals (OR = 0.67) and those from southern Italy (OR = 0.63) were less frequently, while those with high education (OR = 1.97), ex- (OR = 1.46) and current smokers (OR = 2.17) were more frequently alcohol drinkers.

Conclusion

On the basis of individual-level data we confirm and further update to 2014 the decreasing trend in alcohol consumption in Italy over the last few decades, mainly due to the fall in wine consumption. This may be due to anti-alcohol policies, programs and campaigns adopted and developed over the last decade. It also reflects socio-cultural changes in drinking, particularly in younger generations.

INTRODUCTION

Alcohol consumption has been identified as one of the major avoidable risk factors for several health conditions (WHO, 2014a; Smyth et al., 2015). According to the World Health Organization (WHO), around 40% of the worldwide adult population drinks alcohol, and the European Region is the world area with the highest per capita alcohol consumption (WHO, 2014a). The type of alcoholic beverage substantially varies by country, according to socio-cultural and territorial characteristics; beer is mainly consumed in North America and northern/central Europe, beer and spirits in South-East Asia (WHO, 2014a) and wine is mostly consumed in southern Europe. Patterns of drinking are related to the type of alcoholic beverages, with wine more frequently consumed on a regular basis and during meals (La Vecchia et al., 2014).

In Italy, official sales data (‘disappearance data’ in the terminology of the Food Agriculture Organization (FAO) monitoring of foods and beverages entering the market), suggested per capita alcohol consumption falling by over two thirds, from more than 19 l/year in 1970 to 6 l/year in 2010 (La Vecchia et al., 2014; WHO, 2014a,b; Gallus et al., 2015). This drop is mainly attributable to the dramatic reduction in wine consumption observed over the last few decades (La Vecchia et al., 2014). Although a fall in alcohol drinking was observed also in other southern European countries (including France and Spain), where wine is traditionally the most consumed alcoholic beverage (Pyorala, 1990; Gual and Colom, 1997; La Vecchia et al., 2014), in Italy the decline was more pronounced. Thus, after having been second only to France in terms of per capita alcohol consumption, Italy is now one of the high-income countries with the lowest levels of alcohol drinking worldwide (La Vecchia et al., 2014; WHO, 2014a; Gallus et al., 2015). Accordingly, in a 2009 pan-European Union survey, Italy and Portugal showed the lowest prevalence of alcohol drinkers (Eurobarometer, 2010).

There is a lack of individual-level data on prevalence, drinking patterns, socio-demographic and lifestyle context related to alcohol consumption in Italy. To fill this knowledge gap, we collected data from nationally representative surveys conducted among the adult Italian population between 2006 and 2014.

METHODS

Data were collected through face-to-face surveys annually conducted in Italy between 2006 and 2014 by DOXA, the Italian branch of the Worldwide Independent Network/Gallup International Association (WIN/GIA). Each year, the overall sample consisted of ∼3000 individuals aged ≥15 years, representative of the general Italian population, in terms of sex, age, geographic area and socioeconomic characteristics (Gallus et al., 2013). We combined data from each survey, excluding surveys conducted in 2011 and 2012, where no information on alcohol consumption was collected. Thus, for the present analysis, the total sample included 21,416 Italian individuals aged ≥15 years (10,299 men and 11,117 women).

Participants were selected through a representative multistage sampling. The first stage was used to select municipalities (the smallest Italian administrative division) in all of the 20 Italian regions (the largest Italian administrative division). Thus, taking as criteria two characteristics, region and size, we identified from 116 to 152 municipalities (according to different survey years), representative of the Italian universe of municipalities (method known as proportional stratified sample). In the second stage, an adequate number of electoral wards (each ward corresponding to a given district of the municipality) was randomly extracted from each municipality, so that the various types of more or less affluent areas of the municipality were represented in the right proportions (i.e., central and suburban districts, outskirts and isolated houses). In the third stage, individuals were randomly selected from electoral lists, within strata of sex and age group. Adolescents aged 15–17 years, not included in the electoral lists, were chosen by means of a ‘quota’ method (by sex and exact age). Field substitution was used as the preferred strategy to deal with non-responders. Unavailable participants were replaced by their neighbors (living in the same floor/building/street) within the same sex and age group. During data processing, statistical weights were generated to assure representativeness of the Italian population aged 15 years or over.

Ad hoc trained interviewers conducted interviews using a structured questionnaire in the context of a computer-assisted personal in-house interview (CAPI). Besides general information on socio-demographic characteristics, we collected data on anthropometric measures and smoking habits. A section of the questionnaire focused on alcohol consumption. For each type of alcoholic beverage considered (i.e., beer, wine, aperitifs/digestifs and spirits), participants were asked about their average intake per week, expressed as number of cans of beer (333 ml), glasses of wine (125 ml) and shot glasses of aperitifs/digestifs and spirits (30 ml).

Aperitifs/digestifs and spirits were grouped together in a unique category (generically named ‘spirits’). Abstainers were defined as participants reporting 0 drinks per week for any type of alcoholic beverage, while alcohol drinkers were those consuming at least one drink per week. According to the most consumed alcoholic beverage, drinkers were further categorized in predominantly wine, predominantly beer and predominantly spirit drinkers. Drinkers of multiple beverages were those consuming the same number of drinks/week of at least two different alcoholic beverages. Men consuming ≥15 drinks/week and women consuming ≥8 drinks/week were defined as heavy drinkers (U.S. Department of Agriculture and U.S. Department of Health and Human Services, 2010; CDC, 2014).

Education was categorized into low (no qualification up to middle school diploma), intermediate (high school) and high (university). Geographic area was categorized in northern (eight regions), central (four regions) and southern Italy (eight regions, including islands). Body mass index (BMI, kg/m²) was calculated from self-reported weight and height.

Statistical analysis

Odds ratios (OR), and the corresponding 95% confidence intervals (CI), for alcohol drinkers (total alcohol, predominantly wine, predominantly beer and predominantly spirit drinkers) vs. abstainers, and for heavy drinkers vs. light/moderate drinkers were computed using unconditional multiple logistic regression models including terms for sex, age, level of education, geographic area, BMI, smoking status and year of interview. All the analyses were performed with SAS version 9.2 statistical package (SAS Institute, Cary, NC, USA).

RESULTS

In Italy, over the period 2006–2014, the per capita consumption of total alcohol decreased by 23% (from 5.6 in 2006 to 4.4 drinks/week in 2014; Fig. 1). Wine consumption decreased by 31% (from 3.9 to 2.7 drinks/week), while beer (0.9 drinks/week in both 2006 and 2014) and spirit consumption (from 0.8 in 2006 to 0.7 drinks/week in 2014) did not substantially change over time.

Fig. 1.

Stacked area charts showing the average weekly consumption of wine, beer and spirits, in the Italian population aged ≥15 years. Italy, 2006–2014.

Men reduced alcohol consumption by 26% (from 9.0 in 2006 to 6.7 drinks/week in 2014) and women by 12% (from 2.5 in 2006 to 2.2 drinks/week in 2014; Fig. 2). A decrease in alcohol consumption was consistently observed in all the age groups: the fall was 43% in the young (15–24 years), 22% in participants aged 25–64 years, and 23% among the elderly (≥65 years).

Fig. 2.

Stacked area charts showing the average weekly consumption of wine, beer and spirits, in the Italian population aged ≥15 years, according to sex and age group. Italy, 2006–2014.

Over the period 2006–2014, the prevalence of alcohol drinkers in Italy was 61.6% (95% CI: 61.0–62.3%), while abstainers were 38.4% (95% CI: 37.7–39.0%). Predominantly wine, beer and spirits drinkers were 35.3% (95% CI: 34.7–35.9%), 11.1% (95% CI: 10.7–11.5%) and 6.4% (95% CI: 6.1–6.7%), respectively (Table 1).

Table 1.

Odds ratios^a (OR) and corresponding 95% confidence intervals (CI) of total alcohol, predominantly wine, beer and spirit drinkers vs. abstainers, according to selected characteristics. Italy, 2006–2014

	N	All alcohol drinkers		Predominantly wine drinkers		Predominantly beer drinkers		Predominantly spirits drinkers		Drinkers of multiple beverages	Abstainers
		%	OR (95% CI)	%	OR (95% CI)	%	OR (95% CI)	%	OR (95% CI)	%	%
Total	21,416	61.6	–	35.3	–	11.1	–	6.4	–	8.8	38.4
Sex
Men	10,299	78.4	1b	44.2	1b	14.6	1b	8.6	1b	11.0	21.6
Women	11,117	46.1	0.24 (0.23–0.26)	27.0	0.24 (0.22–0.26)	7.9	0.24 (0.22–0.27)	4.4	0.22 (0.19–0.25)	6.8	53.9
Age group (years)
15–24	2686	52.5	1b	9.1	1b	22.4	1b	10.6	1b	10.4	47.5
25–44	7440	66.9	1.81 (1.64–2.01)	28.7	4.45 (3.78–5.23)	16.1	1.07 (0.93–1.22)	9.5	1.31 (1.10–1.56)	12.6	33.1
45–64	6476	63.1	1.69 (1.52–1.88)	43.2	6.48 (5.50–7.64)	7.4	0.44 (0.37–0.52)	4.6	0.64 (0.52–0.79)	7.9	36.9
≥65	4814	56.5	1.75 (1.55–1.98)	49.5	8.05 (6.76–9.58)	2.2	0.17 (0.13–0.22)	1.6	0.33 (0.25–0.45)	3.2	43.5
P for trend			<0.001		<0.001		<0.001		<0.001
Geographic area
Northern Italy	9788	65.4	1b	39.7	1b	11.9	1b	5.7	1b	8.1	34.6
Central Italy	4218	65.7	0.97 (0.90–1.06)	38.6	0.99 (0.90–1.07)	9.7	0.78 (0.68–0.91)	8.2	1.30 (1.10–1.55)	9.2	34.3
Southern Italy	7410	54.3	0.63 (0.59–0.68)	27.5	0.55 (0.51–0.60)	11.0	0.67 (0.59–0.74)	6.3	0.80 (0.69–0.93)	9.5	45.7
Level of education
Low	8775	55.7	1b	39.7	1b	7.5	1b	2.9	1b	5.6	44.3
Intermediate	9586	63.9	1.25 (1.16–1.35)	31.0	1.06 (0.98–1.16)	14.0	1.34 (1.19–1.52)	8.9	2.51 (2.12–2.97)	10.0	36.1
High	3055	71.4	1.97 (1.77–2.19)	36.3	1.80 (1.59–2.04)	12.2	1.58 (1.33–1.87)	8.5	3.19 (2.58–3.96)	14.4	28.6
P for trend			<0.001		<0.001		<0.001		<0.001
BMI categoryc
Normal weight	12,079	61.7	1b	32.3	1b	12.7	1b	7.2	1b	9.5	38.3
Overweight	6330	66.1	0.99 (0.92–1.07)	42.0	0.99 (0.91–1.07)	9.4	0.96 (0.85–1.09)	5.8	1.05 (0.89–1.22)	8.9	33.9
Obesity	1733	52.9	0.67 (0.60–0.75)	36.4	0.66 (0.58–0.75)	7.3	0.68 (0.54–0.85)	4.0	0.67 (0.51–0.89)	5.2	47.1
P for trend			<0.001		<0.001		0.004		0.093
Smoking status
Never smoker	13,264	54.3	1b	30.4	1b	10.2	1b	5.4	1b	8.3	45.7
Ex-smoker	3276	70.8	1.46 (1.33–1.60)	50.2	1.48 (1.34–1.64)	9.0	1.64 (1.39–1.94)	4.3	0.67 (0.51–0.89)	7.3	29.2
Current smoker	4876	75.3	2.17 (2.00–2.35)	38.5	1.94 (1.77–2.14)	14.9	2.41 (2.13–2.74)	10.6	3.09 (2.66–3.59)	11.3	24.7

^aORs were estimated using unconditional multiple logistic regression models after adjustment for sex, age, geographic area, level of education, body mass index (BMI), smoking status and year of interview.

^bReference category.

^cThe sum does not add up to the total because of some missing values. Underweight and normal weight individuals combined.

Estimates in bold are statistically significant at 0.05 level.

Prevalence of alcohol drinkers was significantly lower in women (46.1%; OR = 0.24) than in men (78.4%). The same figure was also observed for predominantly wine, beer and spirit drinkers. The prevalence of alcohol drinkers increased with increasing age: as compared to those aged 15–24 years, ORs were 1.81 for 25–44, 1.69 for 45–64 and 1.75 for ≥65 years (P for trend <0.001). A direct trend with age was also observed for predominantly wine drinkers (P for trend <0.001), while an inverse trend was found for predominantly beer (P for trend <0.001) and predominantly spirit drinkers (P for trend <0.001). Compared to northern Italy, participants from central Italy were less frequently predominantly beer drinkers (OR = 0.78) and more frequently predominantly spirit drinkers (OR = 1.30), whereas participants from southern Italy less frequently consumed alcoholic beverages, both considering total alcohol (OR = 0.63) and all the specific beverages (OR = 0.55 for wine, OR = 0.67 for beer, and OR = 0.80 for spirit drinkers). Alcohol drinkers (any beverage) were significantly more frequent among intermediate (OR = 1.25) and high educated individuals (OR = 1.97), as compared to low educated ones (P for trend <0.001). This relationship was consistently observed according to various alcoholic beverages. Compared to normal weight participants, obese individuals were less frequently alcohol drinkers (OR = 0.67), as well as predominantly wine (OR = 0.66), beer (OR = 0.68) and spirit drinkers (OR = 0.67). Compared to never smokers, alcohol drinkers (any beverage) were more frequent among ex-smokers (OR = 1.46), and most frequent among current smokers (OR = 2.17); similar patterns were observed for predominantly wine drinkers (OR = 1.48 and 1.94, for ex- and current smokers, respectively) and predominantly beer drinkers (OR = 1.64 and 2.41, respectively). Spirit consumption was more frequent among current smokers (OR = 3.09), but less frequent among ex-smokers (OR = 0.67).

Out of 13,196 alcohol drinkers, 22.5% were heavy drinkers, 22.9% among men and 21.8% among women (Table 2). The proportion of heavy drinkers among all drinkers increased with increasing age (P for trend <0.001), and was significantly lower in central (OR = 0.83) and southern (OR = 0.55) as compared to northern Italy. No significant relationship was found between educational level and the proportion of heavy drinkers overall and among men, but women with a high level of education were more frequently heavy drinkers compared to those with a low education (OR = 1.44; 95% CI: 1.16–1.78; P for trend <0.001; data not shown in tables). Among drinkers, heavy drinking was significantly more frequent among overweight (OR = 1.18), as compared to normal weight participants, and in current (OR = 1.73) and ex-smokers (OR = 1.34), as compared to never smokers. As compared to predominantly wine drinkers, heavy drinkers were less frequent among predominantly beer (OR = 0.22), predominantly spirit drinkers (OR = 0.45), and drinkers of multiple beverages (OR = 0.15). Proportion of heavy drinkers significantly decreased over time (from 24.8% in 2006, to 20.1% in 2014; P for trend <0.001).

Table 2.

Prevalence of heavy drinkers (≥15 drinks/week for men and ≥8 drinks/week for women) among Italian alcohol drinkers aged ≥15 years. Odds ratios^a (OR) and corresponding 95% confidence intervals (CI) of heavy drinking vs. light drinking, according to selected characteristics. Italy, 2006–2014

	N drinkers	Heavy drinkers vs. ‘light’ drinkers
		%	OR (95% CI)
Total	13,196	22.5	–
Sex
Men	8074	22.9	1b
Women	5122	21.8	0.97 (0.88–1.06)
Age group (years)
15–24	1411	7.4	1b
25–44	4976	18.4	2.56 (2.07–3.17)
45–64	4090	27.6	4.29 (3.45–5.33)
≥65	2719	29.9	4.87 (3.86–6.14)
P for trend			<0.001
Geographic area
Northern Italy	6398	25.8	1b
Central Italy	2770	23.3	0.83 (0.74–0.92)
Southern Italy	4027	16.4	0.55 (0.50–0.61)
Level of education
Low	4890	25.5	1b
Intermediate	6124	20.4	0.98 (0.88–1.09)
High	2182	21.3	1.09 (0.95–1.25)
P for trend			0.328
BMI categoryc
Normal weight	7449	20.2	1b
Overweight	4183	26.4	1.18 (1.07–1.30)
Obesity	917	23.7	0.95 (0.80–1.13)
P for trend			0.151
Smoking status
Never smoker	7205	18.9	1b
Ex-smoker	2321	27.8	1.34 (1.19–1.50)
Current smoker	3670	26.1	1.73 (1.56–1.91)
Type of alcoholic beverage
Predominantly wine	7559	32.3	1b
Predominantly beer	2381	8.2	0.22 (0.18–0.26)
Predominantly spirit	1368	15.5	0.45 (0.38–0.53)
Multiple drinkers	1888	6.1	0.15 (0.12–0.18)

^aORs were estimated using unconditional multiple logistic regression models after adjustment for sex, age, geographic area, level of education, body mass index (BMI), smoking status and year of interview.

^bReference category.

^cThe sum does not add up to the total because of some missing values. Underweight and normal weight individuals combined.

Estimates in bold are statistically significant at 0.05 level.

DISCUSSION

In less than one decade, per capita alcohol consumption in Italy decreased by 23%, from 5.6 drinks per week in 2006 to 4.4 in 2014. This was essentially due to the reduction in wine consumption, which decreased by 31%, whereas beer and spirit consumption remained quite stable over the period. These findings are in line with those from sales data, showing a steady decline by 70% in alcohol consumption (mainly due to a fall in wine consumption) between 1970 and 2010 (La Vecchia et al., 2014; WHO, 2014b; Gallus et al., 2015). However, the ∼ 6 l of pure alcohol per adult per year from sales data in 2010 corresponds to an average consumption of about 9 drinks per adult per week, which is two times the average alcohol consumption reported in our study. There is therefore a substantial underreporting of alcohol drinking from interview-based surveys. Our data further update the Italian scenario, showing that wine consumption continues to decrease, leading to a fall in the prevalence of alcohol drinkers from 65% in 2006 to 59% in 2014.

These findings corroborate those from large cross-sectional household Italian surveys conducted by the Italian National Institute of Statistics (ISTAT), showing a decrease in the prevalence of alcohol drinkers from 69% in 2003 to 64% in 2013 (ISTAT, 2014). Italy is today one of the European countries with the lowest prevalence of alcohol drinking (Eurobarometer, 2010). It is not by chance that, according to WHO data, Italy has now decreasing liver cirrhosis death rates and an extremely low prevalence of alcoholism (1.0 vs. 7.5% in the WHO European Region) and of alcohol dependence (0.5 vs. 4.0% in the WHO European Region) (WHO, 2014a). A study conducted in north-eastern Italy showed how mortality rates for liver disease decreased by 30%—and for liver cirrhosis by 50%—between 1995 and 2010 (Fedeli et al., 2014). Liver cancer rates increased in Italy to the early 1990s, to appreciably decrease thereafter (La Vecchia et al., 2002; Bosetti et al., 2008; Bertuccio et al., 2013). Between 1985–1989 and 2000–2004, mortality from coronary heart disease has decreased by 32% in men (from 91.5 to 62.7/100,000) and by 26% in women (from 38.3 to 28.2), and mortality from cerebrovascular disease by 44% in men (from 70.4 to 39.6/100,000) and by 43% in women (from 53.2 to 30.4) (Levi et al., 2009). These falls were comparable to those observed in the European Union as a whole (Levi et al., 2009). Apart for other main factors influencing cardiovascular and cerebrovascular mortality, this is not surprising, given the possibility that lower doses of alcohol confer protection against thrombotic vascular diseases (Corrao et al., 2004).

The fall in wine consumption observed since 1970s is likely due to the combination of multiple factors. Changes in eating habits have played a key role; the Italian population, particularly younger generations, are drifting away from a Mediterranean dietary pattern (da Silva et al., 2009) and consequently from the Mediterranean way of drinking, which consist in a regular, moderate wine consumption during meals (Giacosa et al., 2014; La Vecchia et al., 2014). Moreover, lunch in Italy has lost its central position as a family meal, and wine is now essentially consumed during dinners, while in the past it was regularly consumed both at lunch and dinner (Gual and Colom, 1997; La Vecchia et al., 2014). Finally, the spread of non-alcoholic or low-alcohol drinks may have contributed to wine drinking fall (Pyorala, 1990; Gual and Colom, 1997).

More recently, the scientific research providing growing knowledge on the detrimental effects of alcohol for human's health (Roerecke and Rehm, 2012; WHO, 2014a) inspired the development and implementation of a number of national policies, mass-media campaigns and prevention programs in several European countries (de Looze et al., 2014; Rossow and Storvoll, 2014). In Italy, comprehensive legislation regulating alcohol purchasing, banning alcohol advertising targeted to minors, and restricting the maximum blood alcohol content (BAC) of car drivers from 0.8 to 0.5 g/l, came into force in 2001 (Anonymous, 2001). Further restrictions were introduced in 2008, through significant law changes, including more severe punishments for those driving in a state of inebriation (Anonymous, 2008).

Legislation introduced in 2012 further regulated age limits, banning sales for alcoholic beverages to minors (i.e., aged <18 years) (Anonymous, 2012). All these policies specifically targeted to the adolescents have at least in part contributed to trigger or maintain alcohol drinking reduction observed in the last decade. This is corroborated by our finding which shows the sharpest decline in alcohol drinking among the young (more than 40% decrease between 2006 and 2014) compared to other age groups (around 20% decrease). It appears therefore that those policies have been efficacious in reducing both alcohol consumption and heavy drinking.

In the whole period 2006–2014, 62% of Italian adults defined themselves as alcohol drinkers: prevalence of predominantly wine drinkers (35%) was more than three-fold higher than that observed for predominantly beer drinkers (11%), and five-fold higher than that observed for predominantly spirit drinkers (6%). Prevalence of drinkers was higher in men (78%) than in women (46%), confirming Italian estimates observed in men (78%) and women (51%) in 2013 by the ISTAT survey (ISTAT, 2014).

Although wine is still the most frequently consumed alcoholic beverage, the gap between the consumption of wine and that of beer and spirits is declining. Moreover, among the elderly, alcohol drinkers almost exclusively drank wine, whereas among the young both predominantly beer (22.4%) and spirit drinkers (10.6%) already exceeded predominantly wine drinking (9.1%). This suggests that the gap between wine consumption and beer and spirit consumption may further decline due to a simple cohort effect.

In our dataset, more educated subjects were more frequently alcohol drinkers. This confirms findings from several studies conducted in other high-income countries (Romelsjo, 1989; van Oers et al., 1999; Bonevski et al., 2014), and suggests that (moderate to heavy) alcohol drinking in higher socioeconomic groups is still widely accepted not only in men, but also in women (van Oers et al., 1999).

We observed that prevalence of alcohol consumption was higher among normal weight rather than obese individuals, both for total alcohol and various specific beverages. The cross-sectional design of our study precludes drawing any causal inference, but although alcohol is an important source of total energy intake (Jequier, 1999; Dennis et al., 2009), a large body of evidence has suggested an inverse association between alcohol intake and BMI (Liu et al., 1994; Kleiner et al., 2004; Rohrer et al., 2005; Gearhardt and Corbin, 2009; Sayon-Orea et al., 2011). This inverse association with BMI has been explained in terms of competition between food and alcohol for similar neurobiological pathways (Kleiner et al., 2004; Gearhardt and Corbin, 2009).

As expected (Grant, 1998; Falk et al., 2006), compared to never smokers, alcohol drinking was significantly more frequent in current smokers. This was observed for total alcohol and in predominantly drinkers of each specific alcoholic beverage. The joint consumption of tobacco and alcohol is a major public health problem, since their combined exposure substantially increases the risk of selected cancers, including in particular those of the upper aero-digestive tract (Franceschi et al., 1990; Hashibe et al., 2009). Overall, up to 75% of head and neck cancers are attributable to tobacco and alcohol (Negri et al., 1993; Hashibe et al., 2007, 2009), whose joint effect on these neoplasms has been shown to be multiplicative or even greater than multiplicative (Franceschi et al., 1990; Hashibe et al., 2009). In Italy, the reduction by almost 70% of alcohol consumption observed over the last 50 years (La Vecchia et al., 2014), and the concurrent steady decrease in smoking prevalence (Lugo et al., 2015), resulted in a substantial decline in oral cavity/pharynx, esophageal and laryngeal cancer mortality, which halved in Italy in only 25 years, between 1985 and 2010 (La Vecchia et al., 2014).

We found that 13.8% of Italian adults (22.5% among all alcohol drinkers) are heavy drinkers, according to the US guidelines for definition of heavy drinking (U.S. Department of Agriculture and U.S. Department of Health and Human Services, 2010; CDC, 2014). This prevalence is more than twice the one observed among adults in the USA in 2013 (6.2%) (CDC, 2013). The Italian heavy drinkers were mostly represented by middle-age/elderly people with a sharp preference for wine consumption. Thus, the subpopulation of heavy drinkers (i.e., drinking more than two drinks per day for men and more than one drink per day for women) in Italy is mostly represented by adults with a likely regular consumption of wine at meals, and therefore by those with a strong adherence to the traditional Mediterranean way of drinking (Giacosa et al., 2014; La Vecchia et al., 2014).

Limitations of the present study include those inherent to the cross-sectional study design, and the self-reported information on alcohol consumption. Strengths include the representativeness of the survey of the general adult Italian population, and the large sample size, which allowed us to obtain accurate prevalence estimates and to analyze differences in alcoholic patterns among various strata of the Italian adult population, after allowance for several covariates.

Analyzing a large individual-level dataset, we confirm and further update to 2014 the decreasing trend in alcohol consumption in Italy over recent decades, mainly due to fall in wine consumption. This decrease is reflected in the downward trend of head and neck (La Vecchia et al., 2014) and liver cancers, the falling death rates of cirrhosis and the extremely low prevalence of neurological disorders, including alcoholism and alcohol dependence, in Italy compared to other large European countries (WHO, 2014a).

FUNDING

The survey was conducted with the contribution of the Italian Ministry of Health. The work of C.L.V. and A.L. is partially supported by the Italian Foundation for Research on Cancer (FIRC) and by the Italian Ministry of Health, General Directorat of European and International Relations. The work of M.R. was supported by a scholarship from FIRC. The work of S.G. was partially supported by the Italian League Against Cancer (Milan).

CONFLICT OF INTEREST STATEMENT

None declared.