In order to realize this second selection is continued using the criteria of selection previously indicated. In this case it is come to exclude 23 works, that represent a 9.8% almost (table 3) because either they turn on investigations realized in experimentation animal or will talk about methodology (standard drink unit)

After this second selection publication references are included in estudio210. The following step consists of asking for primary documents, that is to say complete articles regarding these references.

These articles are characterized to present/display a great heterogeneity, and although in them problems of health and consumption of beer are approached, is necessary again to exclude a total of 162 articles, almost a 77.32%, due to the following reasons:

• Ecological studies that do not allow to release the data
• Investigations that lack control group
• “With regard to a case”
• Hiponatremias, relation with IMC (Index of Corporal Mass), deficit of Faith, etc. very diverse Investigations to each other and with different objectives that do not allow to draw global conclusions
• Consumption and immunology
• Relation of the consumption beer/alcohol with the behavior landlords
• Consumption of diverse types of alcohol, but released nonbeer

Systematic revision

The 48 selected articles, (22.86%) study more in detail and from the first moment it is shown that exists, taking care of the dependant variable or effect derived from the beer consumption, two dominant approaches in this type of investigations, that are, on the one hand the cardiovascular processes as as much cerebrovascular, and on the other hand the cancer.

An additional problem arises when, after realizing numerous claims we did not obtain that they send 24 articles to us (mainly regarding cancer and some to cardiovascular) reason why we are ourselves forced to continue only with the 24 articles on which we counted at this moment.

Table 4 shows the distribution of 24 analyzed articles and it is possible to be observed that practically 80% were original data and of them the greater percentage corresponds to effects on the morbidity/mortality by cardiovascular processes (47%), followed of the relation with diverse tumorlike processes (32%) and finally other effects (21%)

21% that there are including in the section of others correspond to 4 original articles that analyze the effect of the beer on the risk of maculopatía, the risk of hip fracture, the risks derived from the alcohol consumption generally and the beer in particular and a cross-sectional study that analyzes the relation with the subjective health respectively.

As far as revisions only one of them approach did not have relation with dominant them before mentioned, since it approaches the effect of the moderate alcohol consumption and the estrogen levels.

Systematic and quantitative revision

In view of the mentioned thing it is decided to continue from two perspective that they have as objective common to realize “a priori” a systematic and quantitative revision of the effects of the beer consumption:

1. On morbidity/mortality by ECV
2. On morbidity/mortality by Cancer

Reference level: The “drinkers of beer” are not always teetotallers, can consume other types of alcohol. This can give rise to a slant of bad classification.

Factors of confusion: They can be many and very varied, from consumption of tobacco, basic diseases, lipidic landlords….

Design of the study: In its great majority they are cohorte studies. There is one of cases and controls. In Table 6 are those studies of which it was possible to be extracted the data to try to calculate “a global” effect.

It is to stand out that the study of Bobak and cabbages is of cases and controls and the one of Gronback of 1995 talks about to the same cohorte studied by Truelsen in 1998. The last mentioned study is only included since it talks about morbidity.

Next they are to the obtained results when analyzing the combination of the data before mentioned with program META-ANALYST, using the method of fixed effects of PETO.

In the first place the analysis of the three studies of Cohorte is realized, and we can see that the global effect obtained to combine the populations coming from the three studies (cohorte of 34,345 subjects) is a OR of 0.89 (0,81-0,97) that less indicates that the not exposed cohorte of exposed ill than the one of, but c ² of heterogeneity is significant, which indicates to us that they are heterogenous studies and that therefore they are not due to combine. (1. Analysis cohortes). The graphical representation sample of expressive form the obtained global OR.

Later it was come to analyze the four studies, the three of previous cohortes and the one of cases and controls, in this case the global OR is very similar to obtained previously the 0.87 (0,80-0,95), heterogeneity c ² is significant also in this case since now the degrees of freedom are three since 4 studies are combined. (2. Analysis cohortes and cases and controls)

In this graphical representation the smaller precision of the study of cases and controls is observed of evident form (its ampler limits of confidence)