Risk and protective factors for mortality in a 35year follow up of 50,000 young men

1
Risk and protective factors for mortality in a
35-year follow up of 50,000 young men

• Anders Romelsjö, professor of alcohol-and drug
  research, SoRAD, Stockholms universitet
  affiliated professor, Department of Public Health
  Sciences,Karolinska Institutet
• Ingrid Davstad, PhD Student
• Marlene Stenbacka, associate professor
• Anders Leifman, statistician

2
Research questions

• Does the impact of early social and behavioural
  factors on mortality decline over time?
• Does moderate alcohol use reduce the risk of
  fatal myocardial infarction?

3
Material and methods

• 49,411 men aged 18-20 years, took part in
  mandatory conscription 1.7 1969-30.6 1970
• Data from questionnaire, health examination and
  psychological assessment at conscription linked
  to register data on mortality (and inpatient
  care) 1970-2004.
• Bivariate and multivariate Cox regression
  analyses were done

4
Outcome and analysis

• All-cause mortality 1970-2004 (n2, 671)
• 1970-1987 (n868)
  
• 1988-2004 (n1803)
• Myocardial infarction (n137)
• Alcohol-related (alcohol diagnoses) (n215)

5
Causes of death (n2,671). A majority avoidable?

• Injuries 455 18
• Suicide 443 17
• Suicide, undetermined 106 4
• Alcohol 215 8
• Drugs 49 2
• Liver cirrhosis 87 3
• Tumours 540 21
• Circulatory 416 16

6
Total mortality 1969-2004, multivariate Cox
regression. Statist. significant variables (HR)
13 november 2009
Namn Efternamn
6
7
Mortality 1969-87 1988-2004, multivariate Cox
regression.

• About the same variables predicted death during
  1969-1987 (n868) and 1988-2004 (n1,803), with
  about the same HR
• Thus, factors at young age (18-20 years) have a
  long-lasting impact on mortaility, in spite of
  possible positive changes in habits, which reduce
  increased HR
• These wereAlcohol use gt280 grams/week (4 bottles
  of wine)
• BMI gt30 (gt25 1988-2004) Smoking
• Run away from home Truancy Lowest IQ-quartile
• NOT drug use, poor health, social class, divorced
  parents, good emotional control, good social
  maturity.

8
Does alcohol protect from fatal myocardial
infarction? Bivariate analysis (HR)
9
Fatal myocardal infarction multivariate Cox
regression, HR 95 CI
10
Alcohol-related death bivariate Cox regression,
HR 95 CI
11
Summary

• Early social and behavioural factors (incl. low
  social class, smoking, hazardous alcohol use,
  overweight, drug use) predict death also late in
  life among men.
• No protective effect of alcohol for death or
  myocardial infarction death up to the age of 55
  years
• The majority of deaths are potentially avoidable,
  related to life-style, social and behavioural
  factors.
• Limitation No information about changes in
  habits.
• Causal pathways (Fathers social class)
• Prevention early in life (alcohol, obesity,
  behaviour, IQ etc)

12
Global burden of disease

• Alcohol causes 3.2 of 55 million deaths/year and
  4 (58 million) DALYs year 2000 1.5 and 3.5 in
  year 1990. Higher figures for Europe (Rehm et al,
  WHO, 2004)
• Reduction of alcohol-related problem is possible,
  mainly with restrictive availability policy (high
  taxes, limitation of outlets, home alcohol
  policy) and secondary prevention in health care)
  (Babor et al, WHO, 2003).

13
EU-perspective

• 55 million heavy users (gt40 grams/day for men and
  20 grams for women
• 195.000 alcohol-related deathsl
• More than 25 of fatal trafic accidents are
  alcohol-related (about 10.000/year)
• More than 25 of all deaths among young men are
  and 10 among young women are alcohol-related
  (injuries, suicide, homicide, 60 diseses)
• Alcohol is the third most important life style
  factors for mortality and morbidity, after after
  tobacco and high blood pressure
• Turkey has increased alcohol consumption with 20
  from year 2007 to 2008 (News 29 April, 2009).

14
Other studies

• Drug use and mortality risk Other studies are on
  selected clinical samples
• Predictors of positive social outcome
  (occupation, own income, education level and
  family situation 1990-2006) among those with
  certain risk factors (drug use, high alcohol
  consumption, deviant behaviour, poor social
  background etc.) Based on individual linkage
  with a register with these variables
• Cumulative risk size with number of risk, and
  protective factors, and their interaction.

15
Independent variables, ages 18-20 years

• RISK FACTORS
• Alcohol (0, (reference), 0.1-168 g 100
  alc/week, 168-280,280)
• Narcotic use (yes/no)
• Smoking (0,(reference),1-10 cig/day,11/day
• Body mass index (lt18.5, 18,5-25 (reference),
  25-30,gt30)
• Self-reported health (rather poor vs. good)
• Fathers social class (working class vs. others)
• Divorced parents (yes/no)
• Truancy ever
• Run away from home
• Low IQ (lowest quartile
• PROTECTIVE FACTORS
• Emotional control (good versus others)
• Social maturity

16
Alcohol-related death (n215) multivariate Cox
regression, HR 95 CI
17
Sweden

• Northern Europe, nation since about 1000 years
• Member of EU since 1995 (firm supporter of Turkey
  membership)
• Population 9 million people
• Population density about 20
• Public elections since 1922
• Market economy and social welfare, high living
  standard, cheap public health care and public
  reimbursement when sick, unemployed, poor. High
  taxes
• Two political blocks Social democrats (market
  economymore social welfare) vs more right wing
  parties ( less
  
• Unemployment 8.3

18
Total mortality 1969-2004, multivariate Cox
regression. Statist. significant variables (HR)
19
Does alcohol protect from fatal myocardial
infarction? Bivariate analysis (HR)
13 november 2009
Namn Efternamn
19