Striking it Rich

October 16, 2007

As nutritional supplements are subject to more and more scrutiny, the demand that products are supported by sound and unbiased research is essential. There are many different types of data that can support a product's safety and efficacy. The easiest, least expensive, and least accepted are in vitro studies (test tube experiments). However, these are probably the best approach when first developing a new product, since often the results may help guide the researcher into the best animal or human study.

As nutritional supplements are subject to more and more scrutiny, the demand that products are supported by sound and unbiased research is essential. There are many different types of data that can support a product's safety and efficacy. The easiest, least expensive, and least accepted are in vitro studies (test tube experiments). However, these are probably the best approach when first developing a new product, since often the results may help guide the researcher into the best animal or human study.

However, in vitro studies bypass potentially serious issues, such as absorption and transport to the desired area in the body. Animal studies are much better in that they avoid the above issues, but are subject to the question of whether the animal model used accurately reflects the way in which a human would absorb, distribute, metabolize, and excrete the product. Therefore, assuming sufficient safety data exists for a product, and enough preclinical data have been gathered to indicate the potential for efficacy, the next critical step is to run well-designed human clinical trials.

There is a wide variety of design choices when considering a human clinical trial. There is not one type of human clinical trial that fits all needs; although many would claim that a randomized control trial (RCT) is the only way to go. Admittedly, an RCT is the gold standard of human clinical designs, but prematurely rushing to this model could be very costly and provide negative results. Even if a product is efficacious, if wrong clinical endpoints are chosen, inappropriate or unaccepted methodology is used, or the trial is underpowered (too few subjects), the RCT could prove an expensive fiasco.

OPTIONS

Before jumping to the expensive, well-powered, double-blinded placebo-controlled trial, it may be advisable to look at less-expensive trials that will provide essential information needed to develop the expensive RCT. These can take a variety of forms-often an RCT is termed a pilot study if it is purposely small in order to control costs and look at a variety of potential clinical endpoints. 'Success' in this case may be a strong trend toward efficacy (p<0.2) for one or more endpoints instead of statistical significance (p<.05).

Another approach might be an open-label, in-house human clinical. In this case, costs are controlled because of simplification and in-house labor. Without a placebo, a baseline can be determined for a group of people and then they can be given the test product and observed over time. While these types of clinicals would not be accepted by the scientific community as proof of a product's efficacy, nevertheless the data gained can be invaluable in developing a properly designed and sufficiently powered RCT.

When developing the protocol for a clinical trial, it is especially important to consider what claim(s) you wish the trial to substantiate. It is highly recommended that the researcher work closely with the company's regulatory expert as well as the company's marketing group. Avoiding trials designed to substantiate nonallowed drug claims, or allowed but unwanted marketing claims, can be a very important early planning component of a study.

Once the desired claims are determined, the protocol can be developed. Deciding critical clinical endpoints as well as the correct population to study should be well thought out. For example, if the product is marketed to healthy individuals, it could be problematic if the study population is diseased. Extrapolation of results from a diseased population to healthy individuals could be considered improper by federal regulatory agencies. Additionally, diseased populations are inherently at greater risk, and therefore place the company at greater liability. In a diseased population, the odds are increased that a serious adverse event could occur that could incorrectly be attributed to the nutritional supplement.

Another type of study, the equivalence trial, is often used in drug trials, but rarely in the nutraceutical industry. With the money at stake in attempting to replace the drug of choice for a major disease, it is well worth the expense, and may even be required. However, such a scenario doesn't exist in the nutraceutical industry, and with the variety of products available in each supplement category, this is normally not a viable option.

Research Studies: How Diverse is Your Portfolio?

On November 1, 2008, the World Cancer Research Fund (WCRF; London) and the American Institute for Cancer Research (AICR; Washington, DC) will release an expert panel report, Food, Nutrition, Physical Activity, and the Prevention of Cancer.

The report, which examines the effects of diet and exercise on the incidence of 17 different types of cancer, took more than six years to complete and is based on an analysis of 7000 pieces of research.

To help healthcare professionals and consumers understand the report, this summer AICR published a guide to scientific research, Solving the Diet-Cancer Mystery.

The guide advocates a portfolio approach to research that considers the collective strengths and weaknesses of each type of study. The expert panel considered at least four different types of research: epidemiological studies, laboratory studies, controlled trials, and metaanalyses.

Epidemiological studies examine the risk of disease within specific populations or groups of people. Common types of epidemiological studies include ecologic, cohort, and case-control studies.

Epidemiological studies examine the risk of disease within specific populations or groups of people. Common types of epidemiological studies include ecologic, cohort, and case-control studies.

Ecologic studies usually focus on entire populations. In nutrition research, these studies may explore how the general diet of a particular region, such as the Mediterranean or Asia, is associated with health at a particular point in time. Ecologic studies aren't very helpful, however, at discerning which components are responsible for a diet's salutary effects.

Cohort studies follow a group of healthy people for a period of years. Cohort study participants typically complete food-frequency questionnaires or other kinds of surveys that help researchers detect associations between diet and health. One particular advantage of cohort studies is that they can use one cohort to measure many different outcomes.

Case-control studies involve at least two groups of people: a case group of participants diagnosed with a specific health condition and a control group of healthy volunteers.

Laboratory studies don't answer the big questions about nutrition and health, but they do provide scientists with useful data about how nutrients interact with each other or with living cells. In vitro lab studies take place in carefully controlled environments, which helps researchers eliminate confounding variables. In vivo studies, on the other hand, show how nutrients interact with living organisms.

Controlled trials, often called the gold standard of scientific research, require specific interventions in diet or lifestyle. Randomized, double-blinded trials are usually the study of choice for pharmaceutical researchers. However, these trials may not be as well suited for studying how well vitamins and minerals prevent or delay the onset of health conditions like cancer that manifest themselves over a period of decades.

Metaanalyses use statistical methods to pool the results of many different studies. While metaanalyses may provide a clearer picture of the association between a nutrient and disease, they can also produce misleading or skewed results by including data collected from different types of volunteers (e.g., sick and healthy populations) or by excluding data from particular trials.

Because each kind of study yields different insights, the WCRF/AICR expert panel relied on a portfolio approach that took into account nearly all kinds of existing research. Just as a diverse array of stocks is key to a good investor's stock portfolio, a diverse array of studies is the key to a good researcher's scientific portfolio.

'The portfolio approach allows researchers a comprehensive overview of how the relevant evidence is stacking up,' according to AICR. 'It takes time and patience. Ultimately, researchers are able to draw conclusions through well-informed discussion and debate that balances all the evidence from a vast array of sources.'

 

There are other types of studies that are not experimental human clinicals, but are considered observational human clinicals since there is no way to establish cause-and-effect types of relationships. These studies are not normally associated with nutraceutical products. Among these observational studies are the following commonly used designs:

 

 

 

 

  • Case-control studies: Here the researcher finds a population with a
        certain outcome in common (such as a certain disease) and
        compares them to a matched control group without that outcome. In
        this case, the researcher is looking backwards in time to determine
        any association of the outcome with various risk factors over time.

 

 

 

  • Cross-sectional studies: This is similar to case control, but here the
        researcher is looking at a single point in time. Therefore, the relative
        prevalence of a particular risk factor(s) in both groups is noted.

 

 

 

  • Cohort studies: These are also observational studies, but ones that
        look forward in time as opposed to backward (case-control) or at a
        single time point (cross-sectional). Also known as a prospective or
        longitudinal study, this study type is arguably stronger than the above
        study types, but still lacks the strength of an RCT.

Therefore, the strongest type of human clinical trial is the RCT. However, an equally respected variation of the RCT is the randomized crossover clinical trial. In this case, instead of having similar subjects randomly assigned to two or more groups, the subjects remain in a single group. This type of study can also be double blinded and placebo controlled. The participants are subjected to a treatment (or placebo) for the designated period of time, then taken off the treatment for a designated period of time (the so-called 'washout period' in which any residual effects of the treatment are dissipated), and then subjected to the next treatment, or placebo. The advantage of this design is that the subjects act as their own controls, thereby in theory reducing variability. The danger is that there could be carryover of treatment effects.

Finally, there is a way to combine the results of many different studies into something called a metaanalysis. In principle, only studies that meet certain predefined requirements would be included in the analysis. By combining the results of several studies, the statistical power is increased and often effects can be found where no significant effects were found in the original studies. While this may be seen as a good way to look at the total evidence published for a particular treatment or product, a poorly done metaanalysis can result from researcher bias or error. Typically, this occurs due to the exclusion or inclusion criteria (reasons the researcher chose to either exclude or include a particular study), as well as variations in the statistics used to analyze the data. Additionally, since methodologies and populations often differ between studies, combining data may not always be proper, even if predetermined criteria are met.

To summarize, the double-blinded, placebo-controlled, randomized trial is still commonly thought of as the gold standard of human clinical designs, but the randomized, double-blinded, crossover trial is also widely accepted. Though less accepted, large randomized open-label trials are becoming more common, since the number of subjects can be increased in a cost-effective manner. This type of study is reasonable when the endpoint is a laboratory test, such as cholesterol level, as opposed to a more subjective measurement such as pain or range of motion in a joint health study. Most observational studies aren't very relevant to nutraceutical research, and are less accepted since cause-and-effect relationships cannot be proven.