Main Menu     Health Library     Clinical and Observation Studies                                                                                                                                            

ANTIOXIDANTS and DEATH from ALL CAUSES - A GLOBAL ANALYSIS

Abstract:  Seven risk ratios from six research studies on healthy persons found that use of combinations of antioxidants C and/or E reduced risk of all-cause death by 36%   This reduction in risk parallels the benefits found for use of the individual antioxidants in reducing risks of cardiovascular diseases and cancer. But as per results on risks of diseases a useful reduction in this risk may require 10 or more years duration of vitamin use.   Also antioxidants have not been found to benefit risks of person's suffering heart or other major diseases.  Also, use of Beta-Carotene or Vitamin A in amounts exceeding 15000 and 5000 IU respectively may be unhealthful, and use of Vitamin E in amounts much larger than 400 IU may not be beneficial.

The research study results of useful significance found in Medline searches for antioxidants and death from all causes follow in the appended table.  As expected from biological mechanisms, antioxidants reduce risk of disease as a function of time of exposure. But as will follow, there may be limits to the amount of antioxidants that are effective. And using excess amounts of them may negate their potential benefits.  The effective amounts of antioxidants are shown as equivalent IU values of Vitamin E.  Average Vitamin E equivalents of 0.36 per mg of Vitamin C, 0.009 per IU of Beta Carotene, and 0.74 per mcg of Selenium were derived from formulas shown in the accompanying reports of antioxidants and cardiovascular( CV) diseases. Selenium provides higher Vitamin E equivalent benefits for cancer than for CV disease.  See other papers of antioxidants and CV disease for Vitamin E;  Vitamin A;  Vitamin C; and Selenium and papers of antioxidants and cancer for Vitamin E;  Vitamin A;  Vitamin C; and Selenium,

Results on Healthy Persons:  The first displayed results of six observation studies OH1-6 show consistent and usually substantial benefits on death from all causes of healthy persons for all combinations of antioxidants researched.  An average overall risk factor for these 6 studies was about 0.65, or a 35% reduction in risk of death from all causes. This is quite similar to the reductions in risk found for the individual antioxidants on cardiovascular diseases and cancer, and the cumulative results from all studies is of highest significance.  This is a very major reduction in risk of death from all causes that could lead to significantly increased length of life.  But identifying via present health research a useful benefit requires a minimum usual measured exposure time of about 7-10 or more years. Expressed as reduction in risk per year of exposure the risk factor becomes about 0.95 or 5% per year. And as is explained in an accompanying article  Clinical and observations studies - Why do their Results Differ, the true average exposure time in most observation studies may have been closer to 17-20 years than to the measured study time of 10 years. This would suggest an annual risk factor of about 0.97 or a risk reduction closer to about 3% per year of antioxidant use.

Five useful clinical studies were found published on the effect of antioxidants on all-cause death for presumably healthy populations of individuals. These are listed as CH1-5.  Interestingly, the three studies using 390 IU or less Vitamin E equivalent showed an average risk benefit of 0.96 or 4% per year of measured exposure. But those of higher total antioxidant showed lesser or no benefit.  But most of these studies and especially study CH2 included amounts of beta carotene well above the 5000 and 15000 IU of Vitamin A and Beta Carotene indicated to be maximums to avoid potential harm in the accompanying individual summary reports on this nutrient. Thus these studies are far from definitive in identifying dose related effects of antioxidants or of Vitamin E per se.

Results on Survivors of Major Disease:  Results of the eleven studies CD1-11 show little or no benefit for use of antioxidants on patients having and presumably being treated for heart or other diseases.  Another study in 1902, Lancet 360:23 not included in the table studied in a clinical trial the use of 2300 IU of Vitamin E plus 250 mg of Vitamin C and 33,000 IU of beta-carotene vs. placebo for 5 years on 20,000 people high risk that mostly had previous heart disease or diabetes. No effect of heart disease, cancer, or all-cause death was found from this enormous dosage of antioxidants.. This again is consistent with the results found from the studies of individual antioxidant effects on risk that usually found little or no benefit of the individual antioxidants on heart disease patients. Such individuals usually would be taking multiple medications of more powerful medicines that could offset or replace the slow acting potential benefits of antioxidants. An exception here is study CD3 of Clark on benefits of selenium for those having skin cancer that did indicate a benefit, and these individuals probably would not have be taking heart disease type medications.

A recent meta analysis of 19 studies (2004, Miller III ER, Ann Int Med 142:-) suggested that very high dosage Vitamin E might increase death from all causes. These specific studies were included in the previous Life Ahead analysis. Analyses of the data of CD1-11 used in both this and the Life Ahead analyses.  But problems with this meta analysis are noted.  First, and as was acknowledged by the authors, nearly all of the studies used were on the patients of major disease that were found similarly in this Life Ahead study to obtain little if any benefit.  And second, the amounts of antioxidant cited  were only those of included Vitamin E, and 10 of the 19 studies included much larger effective amounts of other antioxidants. Thus the studies really were not of Vitamin E per se, and 7 of these studies included amounts of Beta Carotene that were well into the potential  problem range.  The three clinical studies on healthy persons of 205 to 390 IU equivalents of Vitamin E all showed an average benefit per year of use of about 4%, similar to that shown by the observation studies.  An approximate re-do of the analysis used in this Miller study using antioxidant equivalents rather than only the amounts of Vitamin E included suggested that antioxidant levels well above 400 IU equivalents but perhaps below the 1000 IU equivalent probably will be beneficial. But again this is highly qualified because most of the studies were not done on healthy people. See more on why survivors of heart disease may not obtain benefits from antioxidants in the paper on Cardiovascular Disease.

Another meta analysis (2007, Bjelakoovic, G, Jama 297:842) provided results of 68 clinical studies that included information on deaths from antioxidants from 68 clinical studies.  This study claimed no advantage for Vitamin E and implied little benefit for antioxidants. This study was misleading.  An analysis of it is provided at Antioxidantsupplement.

We really have no concise peer reviewed research data on Vitamin E use by healthy people of the 400 IU that today is popular. (The Passwater survey, (V9 in the herein report on Vitamin E and CHD) did report a strikingly consistent benefit of time of use and a 5% per year accumulating risk benefit on heart disease for use of 400 IU of Vitamin E for up to 25 years.)  And available  research does not confirm that amounts higher than this are of  further benefit.  This suggests that users may wish to limit use to a maximum of 400 IU, and this maximum amount for benefit is used in Life Ahead. The maximum amounts credited for benefit in Life Ahead are 200 IU for Vitamin E,750 total mg for Vitamin C and a 100 mg supplement for selenium, and as above, 5000 and 15000 IU of Vitamin A and Beta Carotene respectively..

The extensive research on Vitamin E and other antioxidants may provide our population the most effective, simple and inexpensive known way for improving long range health and length of life.  Unfortunately the confusion caused by poorly understood and inadequate analysis of the results of the clinical studies has seriously limited the effective communication of this from the research community to doctors and the public. Serious science requires a full understanding of why results for different kinds of population studies differ that moves beyond what today has been speculation.  Published study of this key problem appears near non-existent.  The extensive analyses developed in the Life Ahead project now reconcile nearly all observation and clinical study results on wellness factors with recognition of time of nutrient use.  See as noted above and on this site:  Clinical and Observation Studies -Why do their Results Differ?

THE MAJOR RESEARCH on ANTIOXIDANTS and DEATH from ALL CAUSES FOUND PUBLISHED

Understanding the Research:  The tables following provide most of the important useful research in on risks found published to latest date listed. Most researchers and health writers base their ideas on just a few most recent studies because it can take months of time to produce this type of a more complete review. But these comprehensive listings are an essential need to obtain a most up-to-date and correct answer to a most likely risk.  The important finding of research is the risk ratio, or the risk of disease of those with a better health factor vs. those with a base or usually average or poorest health factor.  A risk ratio of 0.5 means risk of those with better health factor have half the risk of those with the poorer one.  A ratio of 0.25 means they have only 1/4th the risk; a ratio above 1.0 means a negative result from the factor. 

Most risk ratios are followed by an error margin, as limits of 5% to 95% probability.  If a ratio is 0.75, and the limits are 0.5 to 0.9, the study really produces a likely result of a risk lying somewhere between 0.5 and 0.9.  The usually quoted risk simply of 0.75 is quite inaccurate. But because the upper limit is still below the null value of 1.00 this result usually will qualify in the study publicity as "Statistically Significant"  If the error margin is say 0.4 to 1.15, and the upper level is above 1.00, the individual study usually will be "Not significant."  Researchers sometimes say incorrectly that "We found no effect" when a value is "Not 95% significant"   The produces Statistical Confusion that can be very misleading. 

Because most health studies have high error margins, it is necessary to have results of multiple studies to obtain a probable truth.  Two studies, each finding a risk of 0.75, and each with say error margins individually of 0.45 to 1.10 can define in combination a quite significant result even though each individually did not include sufficient data and was "Not significant"   A best answer thus is revealed by what ALL of the RESEARCH shows in combination.  This is what the tables following attempt to show.

Look down at the risk ratios found.  If risks from most or all studies show values lower than 1.0 you can be assured that the effect being measured is of extremely high significance. The fact that some individual studies do not reach 95% significance individually can be meaningless when viewing multiple results.   5-95% limits that each are below 1.00 at usually confirm an individual study as significant in its own right. If most studies have risks well below 1.00 and error limits also below 1.0, the truth may be established at levels of millions to one.  Some studies include 3, 4 or 5 results at differing "dose" or amounts of factor levels.  If these show steadily declining ratios with dose, significance becomes even larger.  Some studies show "p" ratios of significance.  A p value of 0.05 or denotes 95% probability.  A lower p value of say 0.01 is 99% and a p of 0.001 can define significance at a thousand to one..