Product safety is of paramount concern to the general public, manufacturers, retailers and the government. In a white paper sponsored by the Nutraceutical Institute partnership of Rutgers University and Saint Joseph's University Center for Food Marketing (1999), it was reported that safety is the first concern of those in the dietary supplement industry followed, sequentially, by commercial interests and consumer education. Despite this priority, discussion of safety assessment has received relatively little attention in public forums within the industry or in industry related publications.
This could be attributed, most notably, to the lack of official safety testing guidelines and to limited regulatory control under the Dietary Supplement Health and Education Act (DSHEA). Unlike pharmaceuticals and food additives, where, under the Food, Drug and Cosmetic Act, it is the responsibility of industry to demonstrate product safety, under DSHEA, it is the burden of the Food and Drug Administration (FDA) to demonstrate toxicity of dietary supplements. This, coupled with a prevailing assumption that "natural" is safe, a long history of "traditional" use and economic factors such as cost, time and lack of proprietary protection have led to lack of motivation and a laissez-faire passive attitude towards safety assessment.
Nevertheless, as the industry continues to grow, and as the use of botanicals and other dietary supplements accelerates, there is an increasing likelihood that these products will come under greater scrutiny by consumer advocates, legislators, regulators and the public at large.
In September 1998, an international workshop was conducted to address questions about the availability and quality of safety data on medicinal herbs1. The workshop was sponsored by the National Toxicology Program (NTP), the National Institute of Health (NIH) Office of Dietary Supplements, the NIH Office on Women's Health, the Department of Health and Human Services (DHHS) Office of Disease Prevention and Health Promotion, the Food and Drug Administration (FDA) Office of Special Nutrition and the Society for the Advancement of Women's Health Research. As an outgrowth of this workshop, the NTP concluded that there is inadequate toxicity information available on botanicals and called for safety assessment studies of several popular herbs including aloe vera, kava kava, milk thistle extract and ginseng2. Other trade related scientific bodies are currently collaborating with these agencies and are also collecting and evaluating safety data on botanicals and their derivatives.
It is probable that the convergence of growing public exposure and the associated susceptibility to potential risk will soon lead to increased regulatory scrutiny and the development of standards and guidelines for testing the safety of dietary supplements including botanicals. This article presents an overview of safety assessment with particular emphasis on biological safety and proposes a practical, economically feasible scheme for the safety assessment of botanicals.
Quality Control Issues
Before determining the safety or toxicity of any specific botanical, certain basic quality control systems should be addressed. These include: (i) standardization of production procedures, including extraction and processing; (ii) establishment of specifications for raw materials and finished product; (iii) development and validation of test methods to certify the purity of the finished product; (iv) implementation of a reporting system to trace each batch of product and materials used in its production; (v) determination of stability characteristics and proper storage conditions and (vi) documentation of the foregoing.
Chemical and microbiological considerations. The first step in evaluating the safety of a botanical product should be an analysis for the presence of potentially adverse chemical or microbiological contaminants. Official and non-official, validated, peer reviewed chemical and microbiological procedures are available for qualitative and quantitative determinations of most contaminants typically found in plant materials. These include synthetic chemical residues such as pesticides, herbicides, fertilizers, processing chemicals, cleaning solutions, disinfectants, heavy metals and naturally occurring toxicants such as aminonitriles, amines, oxalates, phytates, estrogens and phenolic derivatives.
Potentially harmful microorganisms are also relatively easy to detect using official or peer reviewed methods for known food pathogens, spoilage organisms, mycotoxins, filth and foreign matter.
The Environmental Protection Agency (EPA) maintains safety criteria and enforcement programs for a broad spectrum of agricultural and industrial chemicals. The FDA and Department of Agriculture (USDA) have well-established microbiological safety standards and enforcement programs.
Assuming the chemical analyses and microbiological assays substantiate the purity of the botanical and satisfy the available standards of safety, a toxicology test program can be initiated.
Biological safety. The main objective in assessing the biological safety of a botanical product is to demonstrate, with a high degree of certainty, that the product in question is safe under the recommended conditions of use to achieve the intended effect,. i.e. that it does not produce any adverse effects in humans. This objective can be achieved using sound scientific test procedures that are time and cost effective.
Recommended Procedures
In the absence of any official guidelines, the following approach is proposed. It is an adaptation of FDA's traditional tiered approach to establish the safety of direct food additives3. It uses proven scientific test procedures and provides thresholds for characterizing safety. Its flexibility enables decision-making at various stages in the scheme, with the potential for reducing testing time and costs, depending on results. It is intended as an interim recommendation to fill the void in safety testing of botanicals until official guidelines are promulgated.
Step 1. Literature Review
The selection of tests is based on a thorough review of the scientific literature to obtain the latest safety information on the botanical and its active constituents. The review should seek information on structure/activity relationships, expected mode of action, target organ(s)/tissue(s), metabolites and end products, especially under the intended conditions of use and exposure. Records of "traditional" use and case histories should be included. Special attention should be given to the component of the botanical to be evaluated and its characterization compared to the product component proposed for testing. This information may then be used to determine the threshold level for test initiation (see below).
Step 2. Test Systems
In general, a combination of genetic toxicity and oral toxicity tests are recommended. They should be conducted ideally under Good Laboratory Practices (GLP) in a USDA-accredited facility that abides by the standards of the Federal Animal Health and Welfare Act and that maintains an active Institutional Animal Care and Use Committee (IACUC).
Three test thresholds for decision making are suggested. Each threshold is based on the available information in the literature, comparison of the structural characteristics of the active components of the botanical to similar compounds with known biological activity and prior testing, if any. Each study is designed to provide safety information for deciding whether to continue testing or to discontinue product development, based on benefit-risk and economic considerations. In all cases, a qualified toxicologist should advise management on whether additional testing is warranted.
Threshold 1: Preliminary Screen
A botanical product in this category satisfies the regulatory safety criteria previously described, but has little or no information available on its biological safety or chemical characterization. In such cases, minimal testing is recommended, to estimate a level of safety and to determine whether additional studies are warranted.
Initially, such a product should be screened for acute oral toxicity and mutagenicity. An oral "limit test" is suggested if there is absolutely no information available on its safety or if the literature indicates that it may be safe for ingestion. The LD50 test is recommended if the literature suggests possible toxicity at high levels of ingestion or if the results of the "limit test" indicate potential oral toxicity.
Both the "limit test" and the LD50 are single dose tests with rodents, where the test substance is administered by oral intubation and the test subjects are observed for 14 days. A single dose level, usually 5000 mg/kg body weight, is administered in the "limit test." The LD50 requires at least three dose levels. The levels can be selected from literature information, a preliminary range finding test or the results of the "limit test." The LD50 is useful for bracketing the safety limits of the botanical, predicting its toxicity and estimating the dose levels for subsequent tests.
While the "limit test" and LD50 provide important safety information rapidly and relatively inexpensively, they are "single dose" tests and therefore not likely to represent conditions of prolonged use. They should be regarded only as screening tests, useful for eliminating the need for further testing, if substantial toxicity is observed and/or for establishing dose levels for subsequent repeated dose studies. In cases where low levels of toxicity are observed, additional repeated dose testing (threshold 2 testing) may still be warranted, especially if extrapolations suggest the observed toxic levels to be above the anticipated use level and where there seems to be a favorable benefit to risk ratio.
In addition to the acute oral toxicity tests, a short-term genetic toxicity assay should be conducted before embarking on a repeated dose study (threshold 2 or 3).
The single point mutation (Ames test) is suggested. Although it sometimes gives false negative results, it is rapid, relatively inexpensive and is widely used as a predictor of carcinogenicity. If the results of this test are ambiguous or positive it may be useful to run a chromosome aberration and/or DNA damage assay. These are of longer duration and are more expensive than the point mutation assay, but could provide valuable information in deciding whether or not to terminate a product.
In cases where there is concern that the botanical may cause antigenicity because of its composition, a simple screening test with rodents may be included as part of threshold 1 testing. This involves an induction period, usually of two weeks duration, a 4-5 week rest period, followed by a challenge. Immediately after challenge the group administered the botanical is observed for anaphylaxis and is compared to the response of a positive control group, induced with a known allergen.
If the results of the Ames test, the single dose acute oral tests and antigenicity screen are negative, initiation of a repeated dose study (threshold 2 or 3) is warranted.
Threshold 2: Short Term Repeated Dose Testing
This threshold calls for repeated dosing over 28 days. It is useful for assessing the safety of a botanical under conditions of frequent use. In some cases, threshold 1 testing can be bypassed if appropriate dose levels can be established from the literature.
The 28 day study should employ at least three dose levels plus a negative (zero dose) control group. Doses are selected to establish a Lowest Observed Adverse Effect Level (LOAEL) and a No Observed Adverse Effect Level (NOAEL). The test product is administered daily, either by oral intubation or in the diet, over 28 consecutive days, during which time signs of gross toxicity, behavioral abnormalities and mortalities are noted. Gross necropsies are performed on all decedents and at test termination. Immediately prior to termination, blood samples are taken from all test groups for hematology and clinical chemistry. At termination, selected organs and tissues are removed from these groups, weighed and preserved for histological examination. Initially, tissue specimens from the high dose and negative control groups are examined. If there are no blood or histological abnormalities, additional testing may not be necessary. If the results from the high dose group are abnormal or ambiguous, the remaining test groups may be evaluated.
Depending on the results of the in-life phase and the post- test evaluations, a LOAEL and NOAEL may be determined and used to estimate the parameters of safety. By dividing the NOAEL by a safety factor, an acceptable level for safe use can be estimated. Additional toxicity tests are optional, unless otherwise indicated by the findings.
If not previously performed, an Ames test to detect carcinogenic potential is recommended. Confirmatory data may be obtained with a chromosome aberration and/or DNA damage assay.
Threshold 3: Longer Term Repeated Dose Testing
These studies are designed for botanicals that have passed the rigors of threshold 2 testing but where there are concerns about safety, if the botanical is regularly consumed for an extended period of time. In some cases, knowledge about the mode of action of the botanical and the time required to achieve its desired health benefit may justify initiating a 90- day test in lieu of the threshold 2, 28-day study.
The dosing, measurements, termination and other procedures are virtually the same as for the 28-day repeated dose study. Additional options include interim blood hematology and chemistry and interim sacrifice with necropsy and histology of selected tissues to assess the progression of any toxicity. An untreated control group plus at least three dose levels are tested. Dose levels are based on literature values, the results of an LD50 or the 28-day repeated dose study.
Ideally, the results of a 90-day repeated dose study enable identification of specific target organs and the establishment of No Adverse Effect (NOAEL) and Lowest Adverse Effect (LOAEL) levels. By dividing the NOAEL by a reasonable safety factor, an acceptable use level can be estimated with a high degree of confidence.
Additional Tests
In some cases there may be reason to conduct tests for reproductive toxicity, developmental toxicity, neurotoxicity and immunotoxicity. These can be incorporated into one of the threshold procedures or can be conducted independently.
Reproduction and developmental toxicity. If the literature or test results suggest there is potential for reproductive or developmental toxicity due to the ingestion of a botanical product, a two generation study with a developmental (teratology) phase may be considered. These are logistically complex studies that can elucidate any adverse effects on the reproductive systems of males and females as well as the postnatal maturation and reproductive health of the offspring over two generations. The developmental phase measures the incidence of death of the developing organism, anatomical anomalies, growth and functional abnormalities of the offspring. Because these studies are complex, it is suggested that they be separate from any of the proposed threshold studies. A minimum of three dose levels plus a negative control are recommended.
Neurotoxicity. If it is anticipated that the botanical may cause adverse neurological effects, a preliminary neurotoxicity screen should be considered. The screen may be conducted independently or it may be incorporated into an appropriate threshold study. The screen is a Functional Observation Battery (FOB) designed to quantitate neurological, behavioral and physiological dysfunctions. Evaluation at all dose levels is recommended to determine if there is a dose-response relationship. If there are adverse findings, histological examination of tissues from the brain, spinal cord and the peripheral nervous system may be performed to enable a benefit:risk assessment.
Immunotoxicity. The primary indicators of immunotoxicity can be determined from the hematology, clinical chemistry, gross necropsy and histopathology evaluations made in the 28 day or 90 day repeated dose threshold study. If suppression of the immune system is suspected, a single dose limit test to assess the functional responsiveness of major components of the immune system to a T-cell antigen may be warranted. In this test, rodents are exposed to the test botanical or a positive control material for at least four weeks. During the last four days, they are immunized. The serum is then assayed to determine the effects on antibody (immunoglobulin) levels.
The Decision Tree (Figure 1)
The foregoing proposal is a practical, economically feasible approach, using thresholds to assess the safety of botanical plants and their derivatives. It is not intended to elucidate mechanisms of action, but rather to provide meaningful end points that a manufacturer can use to determine whether a product is safe for the marketplace and to establish an acceptable use level. Each threshold provides a decision making point that may enable determination of the Lowest Observed Adverse Effect Level (LOAEL) and the No Observed Adverse Effect Level (NOAEL) of a botanical. Ancillary tests can complement the threshold tests when there is special concern beyond general toxicity.
Using the results from the proposed procedures, a decision tree can be established that will enable the manufacturer to make a safety assessment (risk analysis) of the botanical under study in a stepwise, logical fashion. At each step, the manufacturer, with the guidance of a qualified toxicologist, can decide whether to proceed with testing or terminate testing based on safety, benefit:risk and economic considerations.
This stepwise approach is designed to reduce the cost and time required to determine whether a product is safe for the market place with a high level of confidence. Until official guidelines are published, this process can serve as a beginning in providing credible safety data on botanical plants and their derivatives, thereby increasing consumer, regulatory, legislative and supplier confidence. It is also important for establishing safety parameters for clinical trials.
NW
References
1Journal of Environmental Health Perspectives:1999, v. 106:A590-A592.
2NTP Update, July 1999:13-14.
3Toxicological Principles for the Safety Assessment of Direct Food Additives and Color Additives Used on Food. 1993.