Antioxidants And Aging

Although Linus Pauling did not discuss the possibility that v itamin C might protect against mental deterioration associated with aging in Vitamin C and the Common Cold, this possibility has become one of the most popular reasons for taking large daily doses of v itamin C and other antioxidant s such as vitamin E and beta-carotene (p-carotene). The rationale is rooted in the chemical react ions of these compounds with reactive oxygen species. Cells of living organisms contain energy factories called mitochondria, which carry out the final steps in the conversion of energy to a form that can be used to do the work of the cells. These steps require oxygen, which we obtain by breathing. The process is quite efficient, but small amount s of oxygen are converted to byproduct s such as hydrogen peroxide, ozone, nitric oxide, superoxide radicals, and hydroxyl radicals. These substances are called reactive oxygen species because they are derived from oxygen and react readily with essential cell constituents such as DNA, proteins, and lipids. By altering these constituents, reactive oxygen species may cause mutations that lead to unrestrained cell division and growth, that is, cancer, or they may cause cell death and thus contribute to disease (Evans and Hall iwell 2001).

Cells have various nat ural mechanisms to protect them from these effect s, but damage from reactive oxygen species may accumulate over time, contributing to the gradual deterioration called aging (see Chapter 7 for a discussion of this and other hypotheses about aging). Vitamins C and E and p-carotene are called ant ioxidants because they interact with react ive oxygen species, thus reducing the potential for damaging reactions of these substances with DNA, proteins, and lipids in cells. Because the brain has a high rate of oxygen use and abundant lipids in the cell membranes that form connections bet ween cells, brain cells may be part icularly vulnerable to the damaging effects of reactive oxygen species. Therefore, it is quite reasonable to hypothesize that increased amounts of antioxidant s in the diet might protect against the mental deterioration of aging.

One common way of designing medical and nutritional research is called a prospective design. This is illustrated by a long-term study of residents of Basel, Switzerland, begun in 1960. Suppose we hypothesize that antioxidants protect against memor y loss with aging. One way to test this hypothesis would be to initiate a study of young to middle-aged indiv iduals who differ in their use of antioxidants or their blood levels of antioxidants before decreased memory ability associated with aging is expected to occur. Ideally, this group would represent a random sample of a population of interest and few subjects would drop out of the st udy before it s conclusion, which might be many years later. This is a prospective study because we are look ing forward in time to make a predict ion about a future outcome (increased memory loss) that may result from a current condit ion (low intake of ant ioxidant compounds). A group of Swiss researchers led by W. J. Perrig tested the memor y abilities of 442 participants in the Basel study in 1993, when they were between 65 and 94 years old (Perrig et al. 1997). Antioxidant levels had been measured in the blood of these indiv iduals in 1971, independent ly of any assessment of their memory ability. Therefore, this qualifies as a prospective study.

Researchers at the National Cancer Inst itute (NCI) did a classic prospective study of the health effects of smoking in the 1960s (Giere 1997), which provides a model for other researchers, such as the Sw iss group that st udied antioxidant s and memory. From more than 400,000 men who volunteered for the smoking study, the researchers selected about 37,000 smokers and an equivalent number of nonsmokers for comparison. The key strength of this large-scale st udy was that each smoker was paired with a nonsmoker, and members of these pairs were similar in obvious characteristics such as age and ethnic group, as well as a host of other characteristics that might influence health, ranging from religion to the average amount of sleep per night. Not surprisingly, the death rate of smokers after only 3 years was t wice that of nonsmokers.

There are several pitfalls of prospective studies compared to the kinds of randomized experiments that have been used to study the effects of vitamin C on the common cold, as described later in this chapter. One of the most significant risks in a prospective study is that individuals being compared may differ in other factors besides those hypothesized to produce an effect. If the frequency of the effect differs between two groups, such as smokers and non-smokers, the reason for this difference may not be the factor the st udy was designed to test but another factor that also differs between groups, often called a confounding factor. This is why the extensive matching of smokers and nonsmokers was so important in the st udy by the NCI: it was an attempt to minimize the possibility that differences in death rates could be due to confounding factors. This matching required a massive effort to recruit 400,000 volunteers to get 74,000 subjects for intensive study. Of course, in principle, it's impossible to measure and control for all conceivable confounding factors. However, it would obv iously be both unethical and impract ical to study the health effect s of smok ing experimentally, by randomly assigning individuals to smoke or not smoke for many years and then measuring their survival rates.

The st udy of ant ioxidants and aging by Perrig's group differed from the NCI study because the subject s in the former were not div ided into two groups but varied along a cont inuous scale in their blood concentrations of vitamin C, vitamin E, and p-carotene. This variation was evidently not due to differences in the use of v itamin supplement s by the subject s because only about 6% stated that they used supplements, and these individuals did not necessarily have high levels of the v itamins circulat ing in their blood. One of the most interesting results of this study was that blood plasma levels of each of these vitamins in 1971 were strongly correlated with plasma levels in the same individuals in 1993. This might reflect dietary differences among individuals that remained consistent for more than 20 years or genetic differences that affected how v itamins are processed and stored in the body.

Perrig's group used five standard tests of memory performance. One of these was fairly obscure and had no relat ionship with blood levels of ant ioxi-dants, but the other four were more useful in testing this relationship. Implicit memory, or priming, was tested by showing subjects a picture containing several familiar objects on a computer screen,1 then showing the subjects individual pictures of 15 of these familiar objects randomly interspersed with 15 new objects. On average, subjects named familiar objects 17% faster than new ones simply because of the priming effect. Free recall was tested by asking the subjects to name as many of the object s in the initial picture as possible after a 20-minute delay. On average, subjects recalled 8.2 objects. Recognition was tested after the free-recall assignment by showing a pict ure containing some old objects from the initial picture and some of the new object s that were used in the priming phase. Subject s were asked to ident ify the object s that were in the initial pict ure, and the researchers calculated an index of recognition ability based on the number of correct choices and errors. Fi nally, semantic memory was tested by asking subjects to define 32 words; the average subject got 19.6 definitions correct.

The essential results of this study can be summarized in a set of correlation coefficients, which express the relationships between each of the antioxidants (and a few other physiological measurements of the subjects) and each of the memory tests. There are 45 of these correlation coefficients because there were five memory tests done in 1993 and nine physiological measurements, which included levels of vitamin C, vitamin E, and p-carotene, measured in 1971 and 1993, and blood pressure, cholesterol, and ferritin (an iron-containing compound that may reverse the antioxidant effects of vitamin C ), measured in 1993.

If we wish to make our own interpretation of the results of the study, rather than simply rely on the authors' conclusion that "ascorbic acid [vitamin C] and p-carotene plasma level are associated with better memory performance" (Perrig et al. 1997:718), we need to consider what can really be learned from correlation analysis, as well as the limitations of this fundamental statistical tool. The correlat ion coefficient measures the relat ionship between two variables, such as blood plasma level of p-carotene and ability to recognize objects seen recently, on a scale from —1 to +1. The best way to think about the meaning of this coefficient is to visualize the relat ionship graphically. If two variables are completely unrelated, then the correlat ion coefficient equals 0. Figure 2.1A illustrates this pattern (more accurately, lack of pattern). In other words, Figure 2.1A shows what the relationship between plasma level of p-carotene and recognition ability would look like if there were no relat ionship between these t wo variables.

By contrast to the random cloud of points shown in Figure 2.1A, which corresponds to a correlat ion coefficient close to 0, a perfect linear relat ion-ship between two variables would have a correlation coefficient of + 1 if larger values of one variable corresponded to larger values of the other (Figure 2.1C) or — 1 if the reverse were true.

What does the relationship between the plasma level of p-carotene and the recognition ability of the 4 42 elderly subjects actually look like in comparison to the reference patterns shown in Figure 2.1A and 2.1C? The researchers didn't provide their raw data, but they stated that the correlat ion coefficient for p-carotene measured in 1993 and recognition ability measured in 1993 was 0.22. Figure 2.1B illustrates a relationship between two variables that are correlated at a level of 0.22.2

Does Figure 2.1B make you skeptical of the association between p-carotene and memory ability claimed by Perrig's group? Although I made these graphs to help you think crit ically about the result s of this study, I hesitate to encourage undue skepticism because, as an ecologist, I frequently work with relationships that are just as messy. Chemists, molecular biologists, and even physiologists are used to working with relat ionships that look like Figure 2.1C, and they regularly scoff at data that look l ike Figure 2.1B. On the other hand, ecologists, psychologists, and medical scient ists have learned how to deal with patterns like Figure 2.1B because they are interested in phenomena

A: No Correlation

5 Easy Ways To Stop Smoking

5 Easy Ways To Stop Smoking

Your first day without cigarettes can be difficult, but having a plan will make it easier! Learn what steps to take on the day you quit smoking.

Get My Free Ebook

Post a comment