The Meta-analysis: The Meta-analysis: A noon A noon conference presentationconference presentation

Kendall Moseley, MD Kendall Moseley, MD

Kevin Woods, MDKevin Woods, MD

With commentary by Hunter Young, MD MHSWith commentary by Hunter Young, MD MHS

Alcohol Dosing and Total Alcohol Dosing and Total Mortality in Men and Mortality in Men and

WomenWomenAn Updated Meta-analysis of 34 Prospective Studies

Augusto Di Castelnuovo, ScD; Simona Costanzo, ScD; Vincenzo Bagnardi, ScD;Maria Benedetta Donati, MD, PhD; Licia Iacoviello, MD, PhD; Giovanni de

Gaetano, MD, PhD

How much alcohol is healthy for How much alcohol is healthy for you?you?

BackgroundBackground

Previous studies have shown that moderate Previous studies have shown that moderate amounts of alcohol have helped prevent amounts of alcohol have helped prevent coronary artery disease (CAD).coronary artery disease (CAD).

Alcohol abuse can be harmful.Alcohol abuse can be harmful.

Some studies indicate that alcohol may have Some studies indicate that alcohol may have different effects in men and women.different effects in men and women.

How is ETOH good for you?How is ETOH good for you?

Increased HDL cholesterol Increased HDL cholesterol Increased fibrinolysisIncreased fibrinolysis Decreased platelet aggregation and Decreased platelet aggregation and

coagulation factorscoagulation factors Possible beneficial effects on endothelial Possible beneficial effects on endothelial

function and inflammationfunction and inflammation

Clinical QuestionsClinical Questions

What is the relationship between alcohol What is the relationship between alcohol dosing and all-cause mortality?dosing and all-cause mortality?

Is there a statistical difference between men Is there a statistical difference between men and women? and women?

MethodsMethods

Pubmed search Pubmed search Limits: HumanLimits: Human Dates: (all studies – 12/2005)Dates: (all studies – 12/2005) Searched titles and abstracts containing: Searched titles and abstracts containing:

alcohol, beer, wine, spirits AND mortality or death

Exclusion CriteriaExclusion Criteria

only 1 category of risk (n=4) did not report mortality separately for the

sexes (n=5) considered mortality for specific causes (n=3) comprised multiple reports (n=9) reference category was not the one with the

lowest alcohol intake (n=4) relative risks or numbers of cases and person-

years were not available (n=14)

Data ExtractionData Extraction

73 studies identified

34 studies

56 studies total**

37 Male

19 Female

Exclusion criteria applied

** 14 reported results separately for the sexes; 1 study reported data for 2 age groups; 1 study for wine and beer. (These studies contributed 2 dose-response curves each.)

•Two studies contributed 4 curves (1 study reported results separately for 2 ethnic groups and sexes, and another for age groups and sexes.)

Define a DrinkDefine a Drink

The amount of a drink was taken as quantified by each author whenever possible; otherwise (7 studies) it was considered equivalent to 10 g of ethanol.

Considering a drink equivalent to either 12 or 14 g of ethanol did not change our results (data not shown).

A Standard Drink A Standard Drink

•U.S.  DEPARTMENT OF HEALTH & HUMAN SERVICES, National Institutes of Health National Institute on Alcohol Abuse and Alcoholism, Helping Patients Who Drink Too Much : A CLINICIAN’S GUIDE 2005 Edition

Helpful ConversionsHelpful Conversions

•U.S.  DEPARTMENT OF HEALTH & HUMAN SERVICES, National Institutes of Health National Institute on Alcohol Abuse and Alcoholism, Helping Patients Who Drink Too Much : A CLINICIAN’S GUIDE 2005 Edition

SUMMARY

•Over 1 million Subjects

•Studies representative of 5 countries

•Follow up years ranged from 5.5 – 26 years

•Level of Adjustment identified for over 20 factors.

Deriving the J-curveDeriving the J-curve

The regression models were log (relative risk [RR])=1xp2xq ; exponents p and q were selected among the following set:

{−2.0, −1.0, −0.5, 0.0, 0.5, 1.0, 2.0}.

When p=0, xp is replaced by log(x). When p=q, the model becomes

log(RR|x) =1xp2xq log(x).

Data AnalysisData Analysis

(1) the value x of alcohol intake (measured in grams per day) assigned as the midpoint of the reported ranges; x was defined as 1.2 times the lower boundary for the open-ended upper categories.

(2) frequency counts, adjusted relative risks, and 95% CIs for each x level

(3) covariates describing the characteristics of the study. Inverse variance–weighted methods, taking into account the correlation between estimates within each study, were used.

Subgroup AnalysisSubgroup Analysis

Level of AdjustmentLevel of Adjustment SexSex CountryCountry Sample sizeSample size Duration of follow upDuration of follow up

Subgroup AnalysisSubgroup Analysis

Level of AdjustmentLevel of Adjustment SexSex CountryCountry Sample sizeSample size Duration of follow upDuration of follow up

__________________________________________

Reversion pointReversion point: the dose of : the dose of alcohol at which the protection alcohol at which the protection against total mortality is no against total mortality is no longer statistically significant longer statistically significant at the 99% confidence level.at the 99% confidence level.

42 g/d6 g/d

•Mortality benefit could be associated with up to 42 g/d of alcohol consumption (approx 4 drinks/ day)

•Lowest risk of to total mortality associated with 6g/d (approx 0.5 drinks/day)

•Dose response curves are similar for both sexes when alcohol intake is light, but differs with heavier alcohol consumption

•Protection apparent up to 3 drinks/day in MEN

•Protection apparent up to 2 drinks/day in WOMEN

•Obvious differences between women from different countries. (p >.54 for differences between countries)

•However the differences between men from different countries were significant (p >.003)

•Maximum risk reduction for European Men (20 -28%)

•Maximum risk reduction for US Men (14-19%)

•** no explanation was offered to explain this relationship

Why is there a difference between Why is there a difference between the Sexes?the Sexes?

Key FindingsKey Findings

Low levels of ETOH intake are significantly Low levels of ETOH intake are significantly associated with reduced total mortality. associated with reduced total mortality.

While both sexes benefit from moderate While both sexes benefit from moderate ETOH intake, the dose of ETOH at which this ETOH intake, the dose of ETOH at which this benefit is realized is different for men (3 benefit is realized is different for men (3 drinks/day) and women (2 drinks/day) drinks/day) and women (2 drinks/day)

The relationship of ETOH intake and The relationship of ETOH intake and decreased mortality is lower in US-based decreased mortality is lower in US-based studies as opposed to European (men only)studies as opposed to European (men only)

DiscussionDiscussion

The Big PictureThe Big Picture

Why a meta-analysis? (and by the way, what is Why a meta-analysis? (and by the way, what is a meta-analysis?)a meta-analysis?)

How should this type of study be applied to How should this type of study be applied to your clinical decision-making?your clinical decision-making?

Other Related ResourcsOther Related Resourcs

National Institute on Alcohol Abuse and National Institute on Alcohol Abuse and Alcoholism Alcoholism