SUU, Alm.del - 2017-18 - Endeligt svar på spørgsmål 469: Spm. om, at anmode Sundhedsstyrelsen om at redegøre for og forholde sig til det nye svenske studie om alkoholforbrug, som blev omtalt i P1 Morgen den 24. januar 2018, til sundhedsministeren

Sundheds- og Ældreudvalget 2017-18
SUU Alm.del
Offentligt

Research Article

Genetic and Metabolic Diseases

JOURNAL

OF HEPATOLOGY

Alcohol consumption in late adolescence is

associated with an increased risk of severe liver

disease later in life

Graphical abstract

8.00

6.00

Authors

Hannes Hagström, Tomas Hemmingsson,

Andrea Discacciati, Anna Andreasson

Correspondence

4.00

Hazard ratio

[email protected]

(H. Hagström)

Lay summary

2.00

1.00

Alcohol consumption (grams per day)

100

Highlights

Alcohol consumption early in life was associated with an

increased risk for development of severe liver disease after

39 years of follow-up.

The risk increased in a dose-response pattern, with no

evidence of a threshold effect.

Trend towards an increased risk of severe liver disease in

men consuming less than current recommendations for a

safe alcohol intake.

We investigated more than 43,000 Swed-

ish men in their late teens enlisted for

conscription in 1969–1970. After almost

40 years of follow-up, we found that

alcohol consumption was a significant

risk factor for developing severe liver

disease, independent of confounders. This

risk was dose-dependent, and was most

pronounced in men consuming two

drinks per day or more.

http://dx.doi.org/10.1016/j.jhep.2017.11.019

SUU, Alm.del - 2017-18 - Endeligt svar på spørgsmål 469: Spm. om, at anmode Sundhedsstyrelsen om at redegøre for og forholde sig til det nye svenske studie om alkoholforbrug, som blev omtalt i P1 Morgen den 24. januar 2018, til sundhedsministeren

Research Article

Genetic and Metabolic Diseases

JOURNAL

OF HEPATOLOGY

Alcohol consumption in late adolescence is associated with

an increased risk of severe liver disease later in life

Hannes Hagström

1,2,

⇑

, Tomas Hemmingsson

3,4

, Andrea Discacciati

, Anna Andreasson

6,7,8

Centre for Digestive Diseases, Division of Hepatology, Karolinska University Hospital, Stockholm, Sweden;

Clinical Epidemiology Unit,

Department of Medicine, Solna, Karolinska Institutet, Stockholm, Sweden;

Institute of Environmental Medicine, Karolinska Institutet,

Stockholm, Sweden;

Centre for Social Research on Alcohol and Drugs, Stockholm University, Stockholm, Sweden;

Unit of Biostatistics,

Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden;

Stress Research Institute, Stockholm University,

Stockholm, Sweden;

Department of Medicine, Solna, Karolinska Institutet, Stockholm, Sweden;

Department of Psychology,

Macquarie University, North Ryde, NSW, Australia

Background & Aims:

High alcohol consumption is associated

with an increased risk of severe liver disease. Current recom-

mendations suggest it is safe for men to consume 30 grams of

alcohol per day. We investigated the association between alco-

hol consumption early in life and later development of severe

liver disease.

Methods:

We used data on alcohol consumption at conscription

to military service from 43,296 men (18–20 years) in Sweden

between 1969 and 1970. Outcomes were deﬁned as incident

diagnoses of severe liver disease from systematic national regis-

tration of clinical events until the end of 2009. A Cox regression

model adjusted for body mass index, smoking, use of narcotics,

cognitive ability and cardiovascular capacity was applied.

Results:

During a mean follow-up of 37.8 years, 383 men devel-

oped severe liver disease. Alcohol consumption was associated

with an increased risk of development of severe liver disease

in a dose-response pattern (adjusted hazard ratio for every

one gram/day increase 1.02; 95% CI 1.01–1.02). No evidence of

a threshold effect was found. Importantly, a clear trend pointed

towards an increased risk of severe liver disease in men who

consumed less than 30 grams of alcohol per day.

Conclusion:

Alcohol consumption in young men is associated

with an increased risk of severe liver disease, up to 39 years

later in life. The risk was dose-dependent, with no sign of a

threshold effect. Current guidelines for safe alcohol intake in

men might have to be revised.

Lay summary:

We investigated more than 43,000 Swedish men

in their late teens enlisted for conscription in 1969–1970.

After almost 40 years of follow-up, we found that alcohol

consumption was a signiﬁcant risk factor for developing severe

liver disease, independent of confounders. This risk was dose-

dependent, and was most pronounced in men consuming two

drinks per day or more.

2017 European Association for the Study of the Liver. Published by

Introduction

Alcohol consumption is a known risk factor for the development

of cirrhosis.

1,2

Alcohol has been reported to account for 85,000

deaths per year in the US

and as many as 50% of all deaths from

liver cirrhosis on a global scale.

The exact amount of alcohol

needed to inﬂict liver damage is unclear and is affected by inter-

nal factors including genetics

and external factors including

drinking patterns, type of alcohol and diet.

Some evidence

points to a cut-off around 30 grams of pure alcohol per day

for men and 20 grams per day for women,

1,6–8

although data

from two meta-analyses indicate that the cut-off might be

lower at 20–25 grams per day.

9,10

This uncertainty is noted in

guidelines for alcoholic liver disease.

11,12

Much of the current evidence on the risk of liver disease

progression attributable to alcohol comes from studies with

selected populations, short follow-up periods or from cross-

sectional or case-control studies, where cases with manifest liver

disease might be prone to under report past alcohol consumption,

leading to misclassiﬁcation bias. Thus, the role of alcohol con-

sumption early in life needs to be investigated in studies where

alcohol consumption is measured before liver disease has devel-

oped, with adequate follow-up and data on possible confounders.

We examined if alcohol consumption in late adolescence in a

population of well-characterized adolescent men was associ-

ated with an increased risk of severe liver disease later in life,

and if a cut-off level could be identiﬁed.

Material and methods

Study population

We used data from a nationwide population-based study con-

ducted during 1969–1970 of all Swedish men compulsorily

enlisted for conscription. During that time, conscription was

mandatory in Sweden, and only 2–3% of men were exempted

from conscription, mostly due to severe disabilities or diseases.

This study was based on 49,321 Swedish men, age 18–20, con-

scripted during that period.

Variables

Baseline

All conscripts underwent an extensive health examination with

height and weight measurements, and personal interviews.

Keywords: Alcohol; Long-term follow-up; Epidemiology; Decompensated liver

disease; Cirrhosis.

Received 20 May 2017; received in revised form 30 October 2017; accepted 12 November

2017

⇑

Corresponding author. Address: Centre for Digestive Diseases, Unit of Hepatology,

Karolinska University Hospital, 141 86 Stockholm, Sweden. Tel.: +46 (0) 8 5858 2305;

fax: +46 (0) 8 5858 2335.

E-mail address:

[email protected]

(H. Hagström).

Journal of Hepatology

2018

vol. xxx

xxx–xxx

Please cite this article in press as: Hagström H et al. Alcohol consumption in late adolescence is associated with an increased risk of severe liver disease later in life. J Hepatol (2018),

https://doi.org/10.1016/j.jhep.2017.11.019

Research Article

They also ﬁlled in questionnaires on alcohol consumption,

smoking and use of narcotics.

Alcohol consumption

Questionnaires with items regarding the amounts consumed

(number of cans or bottles, or centiliters of beer, wine and

spirits), and frequency of consumption (daily, once a week, less

than once a week, never) were completed during the conscrip-

tion examination. Information on typical alcohol content for

the different kinds of beverages during the period when the

examinations were conducted was retrieved from the Swedish

alcohol retailing monopoly. Grams of 100% alcohol consumed

per day were estimated for each individual.

Body mass index

We used data on height and weight taken at the physical exam-

ination to calculate body mass index (BMI) (kg/m

Smoking and use of narcotics

Smoking at the time of conscription was classiﬁed as either 0

(non-smoker), 1–5, 6–10, 11–20, or more than 20 cigarettes

per day. Use of narcotics was deﬁned as having tried or actively

using any drugs, except for alcohol and tobacco, at the time of

conscription.

Cognitive ability and cardiovascular ﬁtness

A test of cognitive ability was performed as described else-

where.

14,15

Cardiovascular ﬁtness using an ergometer cycle

was tested at baseline when the men’s maximum work capacity

divided by body weight was assessed and transformed into a

numeric scale.

For both tests, the men received a score of

one to nine, a higher number indicates a better result. Both tests

have been found to be predictors of severe liver disease in the

cohort.

Follow-up

All Swedish citizens are assigned a personal identity number a

few days after immigration or birth,

this data was available

at the time of conscription.

The National Patient Register (NPR) was established in 1964.

It includes information on dates of hospital admissions, dis-

charges, and diagnoses classiﬁed according to International

Classiﬁcation of Diseases (ICD) codes, versions 7-10. The register

also includes information on hospital-based outpatient visits

since 2001. The coverage of the register is approximately 99%

of all somatic discharge diagnoses since 1987, and the validity

of hospital discharge diagnoses is between 85–95% depending

on diagnosis.

The Causes of Death Register (CDR) contains data from 1952

regarding the causes of death of all Swedish citizens, including if

the person died abroad. It is mandatory for the responsible

physician to report the underlying cause of death (e.g. stroke)

and any disease that could have contributed to the death of

the individual (e.g. atrial ﬁbrillation).

Severe liver disease

We used diagnoses of liver cirrhosis, decompensated liver

disease (hepatocellular carcinoma [HCC], ascites, esophageal

varices (bleeding or not bleeding), hepatorenal syndrome or

hepatic encephalopathy), speciﬁc coding for liver failure from

the NPR, or death from any of the above in the CDR as our

Genetic and Metabolic Diseases

primary end point variable

severe liver disease.

ICD codes for

the diagnoses used in the present study are listed (Table

S1).

Alcohol abuse and viral hepatitis

Additionally, we obtained ICD codes for any alcohol abuse

related diagnosis as well as all ICD codes for viral hepatitis.

ICD codes for these diagnoses are listed (Table

S1).

Statistical analysis

We excluded 6,025 men due to missing data regarding any of

the covariates, leaving a ﬁnal sample of 43,296 men. All analy-

ses were performed in STATA 13.0 (StataCorp, College Station,

Texas, USA) and a two-sided alpha value of 0.05 was used to test

for statistical signiﬁcance.

Descriptive data

Descriptive data are presented per alcohol consumption cate-

gory. Dichotomous and categorical variables are presented as

percentages, and continuous variables are presented as mean

values.

Survival analysis

Men were followed up from conscription until the ﬁrst regis-

tered diagnosis of severe liver disease. Follow-up times were

censored at the time of death due to any cause, emigration or

end of the follow-up period (31 December 2009).

Cox regression was used to assess the association between

alcohol consumption and the hazard of severe liver disease.

Alcohol consumption was modeled both as a categorical and

a continuous variable. Categories were deﬁned as 0 (abstainers,

reference), 1–5, 6–10, 11–15, 16–20, 21–25, 26–30, 31–40,

41–50, 51–60 and above 60 grams of pure alcohol per day.

Second-degree fractional polynomials

were used to explore

a potential nonlinear dose-response relationship between

alcohol consumption and severe linear disease. A

value for

nonlinearity was calculated as proposed in Royston and

Altman.

We constructed one crude univariate model and one multi-

variate model. The multivariate model was adjusted for BMI,

smoking, use of narcotics, cognitive ability and cardiovascular

ﬁtness at conscription.

Estimates of the ﬁnal models are presented as hazard ratios

(HRs) and 95% CIs. We checked the assumption of proportional-

ity of the hazards using Schoenfeld residuals. We observed no

evidence of departure from this assumption.

Sensitivity analyses

Firstly, a competing risk regression method

22,23

was used to

estimate the association with the risk of severe liver disease

using overall mortality as competing risk, while adjusting for

the confounders previously mentioned. Secondly, all men who

received an ICD-code for viral hepatitis during follow-up were

excluded from the analysis.

Ethical considerations

This study was approved by the regional ethics committee at

Karolinska Institutet (dnr 2004/5:9 – 639/5). Due to the charac-

ter of the database and the anonymization of all data, no written

informed consent was needed.

For further details regarding the materials used, please refer

to the

Supplementary material

and the

CTAT table.

Journal of Hepatology

2018

vol. xxx

xxx–xxx

Please cite this article in press as: Hagström H et al. Alcohol consumption in late adolescence is associated with an increased risk of severe liver disease later in life. J Hepatol (2018),

https://doi.org/10.1016/j.jhep.2017.11.019

JOURNAL

OF HEPATOLOGY

Results

Mean daily alcohol consumption among the 43,296 men

included in the analyses was 8.6 grams (SD 11.2). Categorized,

6.1% men were abstainers, a large proportion, 43.2%, reported

a consumption between 1–5 grams per day and 4.6% reported

a consumption of more than 60 grams per day (Table

1).

Smok-

ing was common at the time of the conscription with 58.4%

reporting being a smoker, with a signiﬁcantly correlation

between smoking and higher consumption of alcohol. Mean

BMI was 21.0 kg/m

, with 6.7% of individuals being overweight

or obese. Baseline descriptive data per alcohol consumption

category are presented (Table

1).

Severe liver disease

The men were followed for a mean period of 37.8 years (SD 4.9,

range 0.1–39) or 1,638,622 person-years. During this time, 3,024

men died and 419 men, 0.97%, emigrated and were censored. A

total of 383 men were diagnosed with severe liver disease. Out

of these 383 men, 208 (54.3%) died during the follow-up period.

Mean time from conscription to the ﬁrst diagnosis of severe liver

disease was 25.5 years (SD ± 11.2, range 3–39).

Alcohol abuse disorder and viral hepatitis

During follow-up, 2,661 men received an alcohol abuse diagno-

sis. Of these, 243 were subsequently diagnosed with severe liver

disease (9.1%). Alcohol consumption at baseline was a strong

predictor of a diagnosis of an alcohol abuse diagnosis during

follow-up. Compared to abstainers, the hazard of obtaining an

alcohol abuse diagnosis was moderately elevated among men

who reported a consumption of between 1–5 grams of alcohol

per day at baseline (HR 1.48; 95% CI 1.13–1.95;

= 0.005) and

was highly elevated among men who reported a consumption

of more than 60 grams per day at baseline (HR 5.22; 95% CI

3.72–7.32;

<0.001).

A total of 386 men were diagnosed with viral hepatitis

during follow-up, of which 22.3% were later diagnosed with

severe liver disease. The prevalence increased with alcohol

consumption, corresponding to an HR of 3.81 (95% CI 1.59–9.12;

= 0.003) for those who consumed more than 60 grams of alco-

hol per day at time of conscription compared to abstainers.

Adjusted HRs for alcohol abuse diagnosis and viral hepatitis

per alcohol consumption category are presented (Table

S3a

and S3b).

Alcohol consumption as a predictor of severe liver disease

In univariate analysis, alcohol consumption was associated with

development of severe liver disease later in life, both when

tested as a continuous variable (HR 1.03 for each additional

gram of alcohol per day, 95% CI 1.03–1.04), and as a categorical

variable. Here, a close association between alcohol consumption

and increased risk of severe liver disease was seen in a dose-

dependent pattern that ﬁrst became statistically signiﬁcant in

the category of 6–10 grams of alcohol per day (HR 2.01; 95%

CI 1.03–3.91), with the highest risk seen in men who consumed

more than 60 grams per day (HR 11.05; 95% CI 5.22–23.40)

compared to abstainers (Table

2).

When adjusting for body mass index, smoking, use of

narcotics, cognitive ability and cardiovascular capacity, a signif-

icant association was observed for alcohol as a continuous vari-

able (HR 1.017; 95% CI 1.010–1.023 for every one gram per day

increase). When we modeled alcohol consumption using frac-

tional polynomials, we observed no evidence of nonlinearity

(p = 0.82) (Fig.

1).

The association followed a dose-dependent

pattern similar to the crude analyses and became statistically

signiﬁcant beginning with the category of 31–40 grams of alco-

hol per day (HR 2.31; 95% CI 1.06–5.05, [Table

2]).

HRs for

development of severe liver disease for included covariates,

adjusted for alcohol consumption as a continuous variable are

presented (Table

S2).

A graph of the cumulative incidence for the development of

severe liver disease, stratiﬁed on alcohol consumption cate-

gories is presented (Fig.

2).

Sensitivity analyses

Alcohol consumption at baseline was signiﬁcantly associated

with development of severe liver disease in the competing risk

regression (subhazard ratio for alcohol consumption as a con-

tinuous variable 1.02; 95% CI 1.01–1.02;

<0.001).

Removing the 386 men diagnosed with viral hepatitis during

follow-up from the Cox regression analysis yielded lower but sig-

niﬁcant estimates on the risk of alcohol consumption (HR for

alcohol as a continuous variable 1.01; 95% CI 1.01–1.02;

= 0.001).

Table 1. Participant characteristics.

Daily alcohol

consumption,

n (%)

BMI (kg/m

)

Smoking

Non–smoking (%)

1–5 cig/day (%)

6–10 cig/day (%)

11–20 cig/day (%)

>20 cig/day (%)

Use of narcotics (%)

Cognitive ability (1–9)

Cardiovascular ﬁtness (1–9)

Severe liver disease

during follow–up, n (%)

2,630

(6.1)

21.0

85.7

3.8

5.6

4.1

0.8

1.7

5.4

6.1

11 (0.4)

1–5 g

18,717

(43.2)

20.9

52.2

12.8

19.1

14.5

1.4

4.8

5.4

6.2

128 (0.7)

6–10 g

9,872

(22.8)

21.0

33.9

12.6

25.7

25.3

2.5

11.0

5.5

6.1

88 (0.9)

11–15 g

4,826

(11.2)

21.0

26.2

11.0

25.5

33.4

4.0

17.4

5.4

6.1

44 (0.9)

16–20 g

2,642

(6.1)

21.1

22.8

9.3

23.2

39.0

5.7

23.2

5.4

6.0

33 (1.3)

21–25 g

1,676

(3.9)

21.2

18.6

9.3

22.9

41.7

7.6

23.6

5.1

5.9

26 (1.6)

26–30 g

954

(2.2)

21.2

18.5

6.8

19.9

44.4

10.4

30.4

5.0

5.9

17 (1.8)

31–40 g

883

(2.0)

21.1

17.4

7.47

17.8

43.8

13.5

36.8

5.0

5.7

18 (2.0)

41–50 g

402

(0.9)

21.4

14.2

5.0

14.4

47.3

19.2

41.3

4.8

5.7

12 (3.0)

51–60 g

267

(0.6)

21.1

13.5

7.1

16.5

42.3

20.6

42.0

4.6

5.7

12 (4.2)

>60 g

427

(1.0)

21.1

10.8

4.0

9.1

45.0

31.2

53.9

4.5

5.5

18 (4.2)

Characteristics of the cohort, stratiﬁed on alcohol consumption. Dichotomous and categorical variables are presented as percentage and continuous variables are presented

as mean values. Associations between variables and alcohol consumption were tested using Mann-Whitney rank sum test for dichotomous variables and Spearman rank

correlation for ordinal and continuous variables. BMI, body mass index. Cig, cigarettes.

Journal of Hepatology

2018

vol. xxx

xxx–xxx

Please cite this article in press as: Hagström H et al. Alcohol consumption in late adolescence is associated with an increased risk of severe liver disease later in life. J Hepatol (2018),

https://doi.org/10.1016/j.jhep.2017.11.019

Research Article

Table 2. Alcohol consumption as a predictor for severe liver disease.

Univariate

Alcohol consumption

Continuous (g/day)

Categorized (g/day)

1–5

6–10

11–15

16–20

21–25

26–30

31–40

41–50

51–60

>60

Cases

383

122

43,296

2,630

18,717

9,872

4,826

2,642

1,676

954

865

390

255

409

Crude HR

1.032 (1.026–1.037)

1.56 (0.84–2.90)

1.98 (1.05–3.72)

2.01 (1.03–3.91)

2.67 (1.34–5.35)

3.66 (1.80–7.43)

3.87 (1.78–8.42)

5.02 (2.37–10.62)

7.45 (3.29–16.89)

11.68 (5.15–26.46)

11.05 (5.22–23.40)

value

<0.001

(ref)

0.16

0.033

0.041

0.005

<0.001

0.001

<0.001

Genetic and Metabolic Diseases

Multivariate

Adjusted HR

1.017 (1.010–1.023)

1.39 (0.74–2.59)

1.50 (0.79–2.87)

1.33 (0.67–2.65)

1.58 (0.77–3.24)

2.03 (0.98–4.23)

1.95 (0.87–4.35)

2.32 (1.06–5.05)

2.99 (1.28–7.00)

4.83 (2.07–11.29)

3.66 (1.65–8.11)

value

<0.001

(ref)

0.31

0.22

0.42

0.21

0.058

0.10

0.034

0.011

<0.001

0.001

Crude and adjusted hazard ratios with 95% CIs for development of severe liver disease depending on alcohol consumption at time of conscription.

Adjusted for BMI, smoking, use of narcotics, cardiovascular ﬁtness and cognitive ability at conscription. BMI, body mass index. HR, hazard ratio.

Discussion

8.00

6.00

4.00

2.00

1.00

Alcohol consumption (grams per day)

100

Fig. 1. Hazard ratios (HRs) during 39 years of follow-up.

For individuals

with severe liver disease by alcohol consumption (grams per day) at the time

of conscription in 43,296 Swedish men in 1969–70. The solid line represents

HRs and the short-dashed lines the 95% CIs estimated from multivariable Cox

models modeling alcohol consumption using fractional polynomials. The

long-dashed lines represent HRs from linear-response Cox models.

Cumulative incidence of liver disease

Proportion with severe liver disease

0.05

<30 grams per day

≥30 grams per day

0.04

0.03

0.02

0.01

0.00

Years of follow-up

Fig. 2. Cumulative incidence graph for the development of severe liver

disease.

Stratiﬁed on a mean daily alcohol consumption (<30

vs.

≥30 grams

per day) in 43,296 Swedish men at the time of enlistment for conscription in

1969–70, and followed for up to 39 years.

Herein, we show that consumption of alcohol early in life is

associated with an increased risk of developing severe liver

disease in men after 39 years of follow-up, independent of

established confounders. This is, to the best of our knowledge,

the longest follow-up time ever documented in a population-

based cohort. The association increased in a dose-response

pattern, with no evidence of a threshold effect. Importantly,

consumption of lower levels of alcohol than the currently used

‘‘safe” cut-off of 30 grams of alcohol per day, down to 20 grams

of alcohol per day, were borderline statistical signiﬁcant in the

adjusted model. In the univariate model, men consuming as lit-

tle as one to ﬁve grams of alcohol per day had an increased risk

of severe liver disease compared to abstainers, indicating that

the increased risk of severe liver disease might be present even

at very low doses of alcohol. Importantly, we could ascertain

cases that during follow-up were diagnosed with alcohol use

disorder, as well as viral hepatitis. Men with a high consump-

tion of alcohol at baseline were more likely to receive an alcohol

use disorder diagnosis during follow-up, which has been shown

previously.

The association between alcohol consumption and liver dis-

ease remained after removing cases with viral hepatitis from

the model, although the estimates were slightly reduced, indi-

cating that the excess risk of a high consumption of alcohol

early in life cannot be explained by a risk-behaviour and later

contracting viral hepatitis.

Our results are consistent with previous large-scale epidemi-

ological studies, although the follow-up period in this study is

unmatched. Speciﬁcally, in the landmark Dionysos study, a

cut-off of around 30 grams of alcohol per day was found to dif-

ferentiate between high and low risk of cirrhosis.

In persons

who consumed more than 30 grams per day, 2.2% of men were

diagnosed with alcoholic cirrhosis, which is comparable to the

2.7% of men that consumed 30 to 60 grams of alcohol in our

study who later developed severe liver disease. In fact, our

estimates are higher, and include only register-captured cases

of severe liver disease, why we will likely have missed cases

of subclinical cirrhosis indicating that the estimates might be

falsely low. However, the long follow-up time in our study

can likely explain our somewhat higher estimates, as it allows

more time to experience an outcome. Although well designed

and executed, the cross-sectional methodology of the Dionysos

Hazard ratio

Journal of Hepatology

2018

vol. xxx

xxx–xxx

Please cite this article in press as: Hagström H et al. Alcohol consumption in late adolescence is associated with an increased risk of severe liver disease later in life. J Hepatol (2018),

https://doi.org/10.1016/j.jhep.2017.11.019

JOURNAL

OF HEPATOLOGY

study will have led to missing of cases that died from liver

disease prior to the examination, possibly underestimating the

detrimental effect of alcohol, a limitation not present in our

study.

The present study only included men. It is, however, known

that alcohol has detrimental effects on liver health in women

as well. In data from the Million Women Study on more than

1.3 million middle-aged women followed for a mean time of

6.2 years, a consumption of more than 150 grams of alcohol

per week increased the risk of a future diagnosis of cirrhosis

(RR 3.44; 95% CI 2.70–4.37) compared to consuming less than

70 grams per week.

These estimates are slightly higher than

ours, which could be explained by older age at baseline and an

increased susceptibility of liver damage in women from alcohol.

Indeed, in a study of 13,295 Danish men and women, the risk of

death or hospitalization due to liver cirrhosis increased from a

consumption of 7–13 units of alcohol per week in women, while

this risk was only signiﬁcant at 14–27 drinks per week for men.

In a study of 6,152 alcohol misusing men and women from

Denmark,

a threshold effect rather than a dose-response effect

was seen. This differs from our results, and could possibly be

explained by our much larger sample size that allows for cate-

gorization of narrower ranges of alcohol consumption, and the

fact that we examined the entire male population at the time,

not focusing on at-risk individuals.

Many liver diseases progress slowly

28–30

and outcomes are

relatively rare, and a long follow-up period in a large cohort is

needed to obtain accurate estimates. Given that risk factors

for severe liver disease such as diabetes type 2

and obesity

are increasing,

there might be more cases of severe liver dis-

ease in the near future. The current study provides long-term

estimates for the risk of alcohol consumption early in life on

the development of severe liver disease, and can be used to

guide future public health decisions.

The strengths of this study are the large population-based

cohort (n = 43,296), very long follow-up time (39 years) and

low (1%) loss to follow-up, which minimizes the risk of selection

bias. Also, investigating men of a relatively low age as in the

present study minimizes the risk for misclassiﬁcation bias, since

none of the men had experienced a liver-related event prior to

baseline.

We had access to detailed and credible baseline data regarding

the exposure status on almost the entire male population from

the study period, as well as on a multitude of possible con-

founders such as BMI

17,33

and use of narcotics. The national,

population-based registers used for ascertaining outcome status

are validated and a source of very high-quality data. The use of

liver decompensation and cirrhosis, which in almost all cases

leads to hospitalization at some point, and liver-related death

as a joint outcome variable allowed us to minimize bias regarding

the outcome status. However, there were too few disease-speciﬁc

cases, such as HCC, to allow for analyses of speciﬁc disease out-

comes. As discussed above, another limitation is that the present

study only includes men and the results may not be generalizable

to women.

Although the stability of alcohol consumption over time in

this cohort is uncertain, previous studies have shown that

alcohol consumption as reported at conscription is strongly

associated with alcohol problems many years later.

Neverthe-

less, there could be residual confounding in the current study,

most importantly regarding changes in alcohol consumption

during follow-up. Also, people with high consumption of alcohol

were more likely to have tried drugs, and are thus more likely to

contract viral hepatitis during follow-up, which would affect the

estimates. However, we controlled for this by excluding cases

with viral hepatitis during follow-up. It was also evident that

men that consumed high amounts of alcohol at baseline to a

greater proportion were subsequently diagnosed with an alco-

hol abuse disorder, suggesting that consuming high amounts

of alcohol early in life is associated with continued drinking,

and later a higher risk of developing severe liver disease.

Another limitation is that we did not have access to more

detailed data on alcohol consumption, such as relation to meals

and binge drinking, which may inﬂuence the effect of consump-

tion on the development of severe liver disease. Indeed, daily

drinking has been associated with an increased risk of alcoholic

cirrhosis compared to drinking less frequently, independent of

the total amount of consumption.

The current study suggests that the risk of alcohol consump-

tion on the development of severe liver disease later in life is

already present from an early age. It is likely that this increased

risk is caused by a longer exposure to alcohol, compared to

starting to drink later in life, and that individuals with a longer

history of alcohol consumption have increased risk of severe

liver disease. The risk was signiﬁcantly elevated for men who

consumed more than 30 grams per day, and importantly

approaching signiﬁcance for men who consumed more than 20

grams per day in the adjusted model. However, this model might

be over-adjusted, and we cannot exclude the possibility of a type

2 error.

Although national variances in the recommended level of safe

intake of alcohol differ, these data indicate that the risk of severe

liver disease might already be present at 20 grams of alcohol per

day or lower, and that some countries current cut-off levels

might be set too high. This has clear implications for public

health decision-making, and should be further explored in future

studies. Ideally, a prospective study on the risk of alcohol on liver

outcomes should include a large population-based cohort com-

posed of both women and men starting from a young age, with

repeated measurements of quantity, mode and type of alcohol

consumption. Further, a long enough follow-up time should be

used, as progression to cirrhosis is slow. Such studies, although

difﬁcult to orchestrate, are highly warranted.

Conclusion

Alcohol consumption in late adolescent men is associated with

an increased risk of developing severe liver disease later in

life, after 39 years of follow-up. The risk increased in a

dose-dependent pattern with no evidence of a threshold effect.

Current recommendations for safe alcohol consumption

regarding the risk for development of severe liver disease in

men might be set too high.

Financial support

HH was supported by grants from the Swedish Gastroenterol-

ogy fund. TH none. AD none. AA none.

Conﬂict of interest

The authors declare no conﬂicts of interest that pertain to this

work. Please refer to the accompanying ICMJE disclosure forms

for further details.

Journal of Hepatology

2018

vol. xxx

xxx–xxx

Please cite this article in press as: Hagström H et al. Alcohol consumption in late adolescence is associated with an increased risk of severe liver disease later in life. J Hepatol (2018),

https://doi.org/10.1016/j.jhep.2017.11.019

Research Article

Authors’ contributions

Study conception and design: HH, TH, AA. Acquisition of data:

TH. Statistical analysis: HH, AA, AD. Analysis and interpretation

of data: HH, TH, AA, AD. Drafting of manuscript: HH. Critical

revision: HH, TH, AA, AD. Guarantor of article: Hannes

Hagström. All authors approved the ﬁnal version of the article,

including the authorship list.

Genetic and Metabolic Diseases

[16]

Aberg MA, Nyberg J, Toren K, Sorberg A, Kuhn HG, Waern M. Cardio-

vascular ﬁtness in early adulthood and future suicidal behaviour in men

followed for up to 42 years. Psychol Med 2014;44:779–788.

[17]

Hagstrom H, Stal P, Hultcrantz R, Hemmingsson T, Andreasson A.

Overweight in late adolescence predicts development of severe liver

disease later in life: A 39years follow-up study. J Hepatol 2016;65:

363–368.

[18]

Ludvigsson JF, Otterblad-Olausson P, Pettersson BU, Ekbom A. The

Swedish personal identity number: possibilities and pitfalls in health-

care and medical research. Eur J Epidemiol 2009;24:659–667.

[19]

Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall C,

et al. External review and validation of the Swedish national inpatient

[20]

Royston P, Ambler G, Sauerbrei W. The use of fractional polynomials to

model continuous risk variables in epidemiology. Int J Epidemiol 1999;

28:964–974.

[21]

Royston P, Altman DG. Regression using fractional polynomials of

continuous covariates: parsimonious parametric modelling. J Roy Stat

Soc: Ser C (Appl Stat) 1994;43:429–467.

[22]

Jepsen P, Vilstrup H, Andersen PK. The clinical course of cirrhosis: The

importance of multistate models and competing risks analysis. Hepa-

tology 2015;62:292–302.

[23]

Fine JP, Gray RJ. A proportional hazards model for the subdistribution of

a competing risk. J Am Stat Assoc 1999;94:496–509.

[24]

Andreasson S, Allebeck P, Brandt L. Predictors of alcoholism in young

Swedish men. Am J Public Health 1993;83:845–850.

[25]

Liu B, Balkwill A, Reeves G, Beral VMillion Women Study Collaborators.

Body mass index and risk of liver cirrhosis in middle aged UK women:

prospective study. Br Med J 2010;340:7.

[26]

Becker U, Deis A, Sorensen TIA, Gronbaek M, BorchJohnsen K, Muller CF,

et al. Prediction of risk of liver disease by alcohol intake, sex, and age: A

prospective population study. Hepatology 1996;23:1025–1029.

[27]

Kamper-Jorgensen M, Gronbaek M, Tolstrup J, Becker U. Alcohol and

cirrhosis: dose–response or threshold effect? J Hepatol 2004;41:

25–30.

[28]

Singh S, Allen AM, Wang Z, Prokop LJ, Murad MH, Loomba R. Fibrosis

progression in nonalcoholic fatty liver vs nonalcoholic steatohepatitis: a

systematic review and meta-analysis of paired-biopsy studies. Clin

Gastroenterol Hepatol 2015;13:643–654, [e641–e649; quiz e639–e640].

[29]

Schwartz JM, Reinus JF. Prevalence and natural history of alcoholic liver

disease. Clin Liver Dis 2012;16:659–666.

[30]

Poynard T, Bedossa P, Opolon P. Natural history of liver ﬁbrosis

progression in patients with chronic hepatitis C. Lancet 1997;349:

825–832.

[31]

El-Serag HB, Tran T, Everhart JE. Diabetes increases the risk of chronic

liver disease and hepatocellular carcinoma. Gastroenterology 2004;126:

460–468.

[32]

Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, et al.

Global, regional, and national prevalence of overweight and obesity in

children and adults during 1980–2013: a systematic analysis for the

Global Burden of Disease Study 2013. Lancet 2014;384:766–781.

[33]

Hagstrom H, Tynelius P, Rasmussen F. High BMI in late adolescence

predicts future severe liver disease and hepatocellular carcinoma: a

national, population-based cohort study in 1.2 million men. Gut 2017.

[34]

Sjolund S, Hemmingsson T, Gustafsson JE, Allebeck P. IQ and alcohol-

related morbidity and mortality among Swedish men and women: the

importance of socioeconomic position. J Epidemiol Community Health

2015;69:858–864.

[35]

Askgaard G, Gronbaek M, Kjaer MS, Tjonneland A, Tolstrup JS. Alcohol

drinking pattern and risk of alcoholic liver cirrhosis: a prospective cohort

study. J Hepatol 2015;62:1061–1067.

Supplementary data

Supplementary data associated with this article can be found, in

the online version, at

https://doi.org/10.1016/j.jhep.2017.11.

019.

References

[1]

Sorensen TI, Orholm M, Bentsen KD, Hoybye G, Eghoje K, Christoffersen

P. Prospective evaluation of alcohol abuse and alcoholic liver injury in

men as predictors of development of cirrhosis. Lancet 1984;2:241–244.

[2]

Lieber CS, Jones DP, Decarli LM. Effects of prolonged ethanol intake:

production of fatty liver despite adequate diets. J Clin Invest 1965;44:

1009–1021.

[3]

Mokdad AH, Marks JS, Stroup DF, Gerberding JL. Actual causes of death in

the United States, 2000. JAMA 2004;291:1238–1245.

[4] World Health Organization. Global status report on alcohol and health –

2014 ed. 2014 [cited 2017; Available from:

http://www.who.int/sub-

stance_abuse/publications/global_alcohol_report/en/.

[5]

Tian C, Stokowski RP, Kershenobich D, Ballinger DG, Hinds DA. Variant in

PNPLA3 is associated with alcoholic liver disease. Nat Genet 2010;42:

21–23.

[6]

Bellentani S, Saccoccio G, Costa G, Tiribelli C, Manenti F, Sodde M, et al.

Drinking habits as cofactors of risk for alcohol induced liver damage. Gut

1997;41:845–850.

[7]

Leevy CM. Fatty liver: a study of 270 patients with biopsy proven fatty

liver and review of the literature. Medicine 1962;41:249–276.

[8]

Teli MR, Day CP, Burt AD, Bennett MK, James OF. Determinants of

progression to cirrhosis or ﬁbrosis in pure alcoholic fatty liver. Lancet

1995;346:987–990.

[9]

Corrao G, Bagnardi V, Zambon A, Torchio P. Meta-analysis of alcohol

intake in relation to risk of liver cirrhosis. Alcohol Alcoholism 1998;33:

381–392.

[10]

Rehm J, Taylor B, Mohapatra S, Irving H, Baliunas D, Patra J, et al. Alcohol

as a risk factor for liver cirrhosis: a systematic review and meta-analysis.

Drug Alcohol Rev 2010;29:437–445.

[11]

O’Shea RS, Dasarathy S, McCullough AJ. Alcoholic liver disease. Am J

Gastroenterol 2010;105:14–32.

[12]

EASL clinical practical guidelines: Management of alcoholic liver disease.

J Hepatol 2012;57:399–420.

[13]

Sjolund S, Hemmingsson T, Allebeck P. IQ and level of alcohol

consumption-ﬁndings from a national survey of Swedish conscripts.

Alcohol Clin Exp Res 2015;39:548–555.

[14]

Hemmingsson T, Melin B, Allebeck P, Lundberg I. Cognitive ability in

adolescence and mortality in middle age: a prospective life course study.

J Epidemiol Community Health 2009;63:697–702.

[15]

Hemmingsson T, Melin B, Allebeck P, Lundberg I. The association

between cognitive ability measured at ages 18–20 and mortality during

30 years of follow-up–a prospective observational study among Swedish

males born 1949–51. Int J Epidemiol 2006;35:665–670.

Journal of Hepatology

2018

vol. xxx

xxx–xxx

Please cite this article in press as: Hagström H et al. Alcohol consumption in late adolescence is associated with an increased risk of severe liver disease later in life. J Hepatol (2018),

https://doi.org/10.1016/j.jhep.2017.11.019