Scientific Standards for Studies on Modified Risk Tobacco Products (2012)

Summary

Smoking is the leading cause of preventable morbidity and mortality in the United States, contributing to approximately 443,000 premature deaths each year nationally (CDC, 2008). Smoking-related disease causes more deaths than alcohol, illicit drug use, homicide, and suicide combined (Mokdad et al., 2004). Another 8.6 million smokers in the United States live with a smoking-attributable illness (CDC, 2009a). In total, tobacco-related mortality amounts to approximately 5.1 million years of potential life lost per year (CDC, 2008). Smoking also imposes enormous costs on the U.S. health care system and economy, with an estimated $193 billion in losses due to health care costs and productivity losses per year (CDC, 2008).

The current prevalence of cigarette use is 20.6 percent among adults and 19.5 percent in youth (CDC, 2010, 2011). After substantial declines in adult smoking rates through the 1980s and 1990s, the rate of U.S. adult smokers has remained relatively static from 20.9 percent in 2004 to 20.6 percent in 2009 (CDC, 2010). Between 1997 and 2003, smoking prevalence among high school students declined substantially from 36.4 percent to 21.9 percent; this decline slowed from a 21.9 percent youth smoking rate in 2003 to 19.5 percent in 2009 (CDC, 2011). Of the 46 million adult smokers in the United States, an estimated 70 percent of smokers wish to quit completely, and approximately 45 percent of smokers attempt to quit each year (CDC, 2002, 2009b). However, only approximately 6 percent of the smokers who attempt to quit are successful for 1 month or more (HHS, 2000).

THE FAMILY SMOKING PREVENTION AND TOBACCO CONTROL ACT

The Family Smoking Prevention and Tobacco Control Act of 2009 (FSPTCA)¹ grants the Food and Drug Administration (FDA) broad authority to regulate the manufacturing, distribution, and marketing of tobacco products, including “modified risk tobacco products” (MRTPs). Generally, an MRTP is defined by the law as any tobacco product that is sold or distributed for use to reduce harm or the risk of tobacco-related disease.

Under the FSPTCA, no MRTP may be marketed without an order for sale from the U.S. Department of Health and Human Services (HHS). To be marketed, the product must meet one of two public health standards: either (1) an empirically demonstrated Modified Risk claim or (2) a Special Rule for Certain Products claim, specifying a reduced-exposure product.

To meet the Modified Risk standard, the applicant must prove with scientific evidence that the product, as actually used by consumers, will (1) significantly reduce harm and the risk of tobacco-related disease to individual users and (2) benefit the health of the population as a whole, taking into account both users and nonusers of the product.

Under the Special Rule for Certain Products, the Secretary of HHS may issue an order for the sale of a reduced-exposure product for which there is inadequate long-term epidemiologic data to support a finding under the Modified Risk standard but where the available evidence demonstrates that a substantial reduction in morbidity and mortality is “reasonably likely.”

In regard to both standards, the law further specifies that the Secretary should also take into account how the marketing of the MRTP affects the likelihood of current users continuing tobacco product use with an MRTP who otherwise would have quit, nonusers initiating tobacco use with an MRTP, and the risks and benefits compared to other smoking-cessation products.

The concept of harm reduction informs the public health rationale for permitting the development and potential marketing of modified risk tobacco products. The basic premise of harm reduction is the continuation of a potentially hazardous or dangerous behavior, with the aim of decreasing the potentially adverse consequences of these behaviors (Marlatt, 2002). In the context of tobacco harm reduction, “a product is harm reducing if it lowers total tobacco-related morbidity and mortality, even though use of that product may involve continued exposure to tobacco-related toxicants” (IOM, 2001).

_______________

¹ Family Smoking Prevention and Tobacco Control Act of 2009, Public Law 111-31, 123 Stat. 1776 (June 22, 2009).

Modification of the risk profiles of tobacco products is only one component of a comprehensive, multifaceted strategy to minimize the negative health effects of tobacco use. Tobacco harm reduction efforts specifically target users that are unwilling or unable to quit. In conjunction with tactics to prevent initiation of tobacco use and to promote immediate cessation, MRTPs with reduced risk profiles may potentially lessen the harm of tobacco for the substantial portion of U.S. smokers who are unable or unwilling to abstain.

REQUIREMENT FOR REGULATIONS, GUIDANCE, AND CONSULTATION WITH THE INSTITUTE OF MEDICINE

The FSPTCA requires the FDA to issue guidance or regulation on the scientific evidence required for the assessment and ongoing review of an MRTP applicant. The law also specifically requires the FDA to consult with the Institute of Medicine (IOM) in developing guidance and regulation on the “design and conduct of such studies and surveillance.” Box S-1 provides the statement of task.

COMMITTEE APPROACH

The IOM convened a multidisciplinary committee of 15 experts with backgrounds in addiction, cardiology, pulmonology, oncology, epidemiology, study design methodology, biostatistics, risk perception, adolescent behavior, drug or device regulation and law, population health, tobacco initiation and cessation, and toxicology. Over the course of 10 months, the committee held five meetings; extensively reviewed literature; heard representatives from the tobacco industry, public health advocacy groups, and regulatory agencies; and consulted with external subject area experts.

To fully grasp the nature of the task, the committee sought guidance not only from the statement of task, but also from the enabling statutory language of the FSPTCA Section 911(l)(2). While it is essential to address

minimum standards for scientific studies, the committee interpreted the task more broadly than the simple rearticulation of basic scientific principles or the review of current scientific methods. The committee was particularly wary of making “perishable” recommendations that may lose relevance as time passes and scientific methods and technologies evolve. Rather, the committee sought to provide enduring insights into what constitutes credible and meaningful evidence of the effect of tobacco products on public health. The committee’s insights can be generally organized into three categories:

1.   Types of studies and evidence on MRTPs

2.   Design of studies on MRTPs and decision making

3.   Governance

EVIDENCE AND STUDIES

Generally, the evidence to support the marketing approval of an MRTP will come from three categories: health effects of the MRTP, addictive potential of the MRTP, and perceptions about the MRTP.

Evidence and Studies of Health Effects

Laboratory analysis of the performance and of the constituents of tobacco products will be the first step in the evaluation of any new product. These analyses involve standard methods of extraction, sample cleanup, analyte identification, and quantitation. There are important limitations to laboratory analysis of product performance and composition. First, laboratory analysis of constituents may not reflect constituent uptake under conditions of use. In particular, smoking machines do not replicate human smoking conditions. There is currently no proven way to replicate the many ways humans use tobacco. As such, it is crucial to describe the smoking regimen or other extraction methods employed. Second, there may be other unidentified compounds in tobacco that contribute in important ways to adverse health effects. Also, seemingly innocuous compounds can exacerbate the effects of toxicants.

The second step in the evaluation of an MRTP will be preclinical studies of toxicity. These assays are essential in identifying particularly risky or toxic products that should not be tested in humans, and to identify products that have reasonable potential to reduce risk and harm and therefore should proceed to clinical evaluation. In vitro assays for cytotoxicity, genotoxicity, apoptosis and cell proliferation, oxidative stress, inflammation, mucus production, and endothelial activation are a standard step in evaluations of all combusted and noncombusted products.

Evaluation of products in vitro should precede in vivo assays. Furthermore, assays in animal models should precede human assays. Although it is not possible to make laboratory animals use tobacco products the way humans do, and there are inherent interspecies differences that prevent meaningful extrapolation of human effects, it is still informative to observe the effect of tobacco products in live animal models. Assays of toxicity in humans will also be essential, in particular assays of urinary mutagenicity and sister chromatid exchange in peripheral lymphocytes.

Biomarkers of exposure measure human exposure to constituents of tobacco. Biomarkers of human exposure to specific constituents of tobacco include the constituents themselves, their metabolites, or protein- or DNA-binding products of the constituents or their metabolites. These biomarkers have the potential to bypass many of the uncertainties in product composition analysis and provide a realistic and direct assessment of carcinogen and toxicant dose in an individual. The first step in employing biomarkers of exposure is analytical validation. The second step is validation with respect to product use. Finally, biomarkers can be validated with respect to disease risk; however, there is no proof that any individual constituent or group of constituents is responsible for a given disease. For a biomarker of exposure to be accepted as a biomarker of risk or a surrogate endpoint of disease, there should be a strong biological rational as well as compelling data from clinical and epidemiologic studies.

Experimental designs, in particular randomized controlled trials (RCTs), provide data that can support the strong inferences about the effect of an MRTP on human health relative to conventional tobacco products. The use of appropriately designed clinical trials will be important to establish whether the use of the MRTP reduces exposure to toxicants or induces positive changes in surrogate markers as claimed by the manufacturer. An RCT is an effective means of examining acceptability and use of the MRTP, the ability of the MRTP to increase cessation in users of conventional tobacco products, and the likelihood that availability of the MRTP will lead to dual use. RCT methods can also produce evidence on whether and how much individuals use an MRTP after they have used it to help them quit conventional products, changes in perception of the MRTP with its continued use, and the MRTP’s ability to suppress tobacco withdrawal symptoms. It is important to recognize that no single RCT can address all of these areas, and each study should have a focused objective with a primary endpoint.

Observational epidemiologic studies play a critical and central role in the evaluation of MRTPs. Although they will rarely, if ever, have the compelling scientific credibility of experimental designs, these methods form the basis for most evaluation studies of regulated products in the community. Long, intensive, and robust observational studies of actual health

outcomes may be required to fully evaluate the net effects of MRTPs relative to conventional tobacco products.

Prospective cohort studies are obvious candidates for the evaluation of MRTPs, and will also be an essential tool to validating anticipated or claimed effects of marketed MRTPs on both individuals and on the public’s health. Cohort studies allow assessment of overall health status and outcomes, as well as offering the following important strengths:

•   Biochemical tobacco and MRTP exposure assessment can be made at baseline, offering unbiased exposure assessment before health outcomes occur.

•   There is less of a problem with retrospective recall of product use, because this information can be summarized at the start of the study, and followed prospectively.

•   Changing product use habits can be monitored as the study progresses.

•   Outcomes can be documented as they occur, and verification becomes more efficient.

A wide variety of outcomes can be evaluated in the same study, including both intermediate and clinical outcomes. In addition, other epidemiological study designs will be necessary to evaluate MRTPs and provide evidence on the public health effects of marketed MRTPs; these include retrospective cohort studies, case-control studies, crossover or case-crossover designs, and comparative effectiveness research methods. Case-control studies are commonly used because of their efficiency in assembling study participants, including when the disease outcomes are not common in general populations (e.g., varying levels of biomarkers). When the outcomes are short term and/or recurrent (particularly when using intermediate endpoints), an observational crossover or casecrossover design becomes feasible and informative. Comparative effectiveness research methods more critically inform health care and policy decisions, but these methods can also sharpen or extend observational studies comparing health outcomes associated with use of conventional tobacco products and use of certain MRTPs. Overall, different study designs will be necessary depending on the circumstances and the research question.

Evidence and Studies of Addictive Potential

Evaluation of the likelihood of initiation, maintenance, and persistence of use in both conventional tobacco users and nonusers is critical to estimating the public health effect of marketing an MRTP. Specifically, evaluation of the MRTP’s ability to promote initiation and continuation

of its regular use, switching to its use and cessation of the consumption of more harmful products, dual use, and to promote relapse back to more harmful tobacco use are all essential. All of these outcomes are logically related to the reinforcing value of the MRTP (that is, how rewarding it is).

There is a continuum of reinforcement value. In theory, the MRTP should be somewhat more reinforcing than nicotine replacement therapies but perhaps less reinforcing than conventional cigarettes. Ideally, an MRTP would be sufficiently reinforcing so as to attract smokers away from conventional cigarettes but not enough to encourage the widespread dependent use of the product by individuals who were previously nonusers, or who would have quit smoking.

Evaluation of the abuse and addiction potential of a product can be accomplished with a variety of experimental designs and in a variety of contexts, including subjective evaluations in laboratory contexts, acute self-administration studies in laboratory contexts, use in extended residence facilities, and natural environment contexts where long-term use can be studied in real-world circumstances via RCTs, cross-sectional survey studies, and longitudinal cohort studies.

Evaluation of reinforcement value in a laboratory setting is particularly important because the results of these studies reliably correspond to an agent’s addictive potential in real-world use. A standard with regard to human abuse liability drug testing is acute dose-effect comparison studies, because of the correspondence between subjective ratings of drug effects and real-world abuse potential. Behavioral economic self-administration studies will also be important in evaluating the reinforcement potency of a product. The usefulness of all studies in forecasting the risk for initiation and abuse of a product depends on study design factors. Important design considerations include the size of the sample, the nature of the sample (whether the sample includes heavy smokers or light smokers, smokers who want to quit, and nonsmokers), the characterization of the sample (age, sex, gender, ethnicity, educational attainment, socioeconomic status, etc.), and the nature of the comparison product.

Evidence and Studies of Risk Perception and Communication

Judgments about risk, otherwise known as risk perceptions, are a fundamental element to most theoretical models of health behavior and behavioral decision making. In general, these models argue that individuals’ perceptions about the value and likelihood of behavior-related positive and negative consequences and their vulnerability to those consequences play a key role in behavioral choices. As such, understanding individuals’ perceptions of tobacco-related products, including MRTPs, whether such perceptions change over time, and whether such perceptions

play a role in tobacco behavior, is critical. It will be important to identify consumers’ perceptions of disease risk, likelihood of addiction, likelihood of reducing or increasing others’ exposure to potentially hazardous compounds, and perceptions of risk compared to other products already on the market. It is also important to assess intentions of using the product. It is essential that the industry carefully crafts messages about risks and benefits of any MRTP and demonstrates through rigorous testing that people correctly understand and interpret the risks.

Studies evaluating risk perceptions and risk communication should be performed both before the marketing of an MRTP and after the MRTP has been marketed. Premarket research will play an essential role in developing the messages the tobacco industry can use to communicate information about MRTPs to consumers. This research will determine consumers’ ability to accurately understand messages that communicate information about the risks, benefits, and conditions of using an MRTP compared to existing tobacco products. Studies should also test how these messages influence consumers’ perceptions of the risks, benefits, and likelihood of addiction related to an MRTP. The first stage of premarket research will involve formative work using focus groups. The second stage should include discussions with groups of similar individuals to assess how the messages that were developed in the first stage are received by consumers. Finally, the effects of these messages on consumer perceptions should be tested. It will be important to evaluate consumer understanding and to compare consumer perceptions of an MRTP to conventional products. After the product is released on the market, it is vital to continue monitoring consumer perceptions and behavior related to that product. Conducting nationally representative cohort-sequential longitudinal surveys will be essential.

Table S-1 presents the evidence domains and example considerations for using evidence from the different domains.

STUDY DESIGN AND DECISION MAKING

Study Design

Studies should be designed appropriately to create an evidence base that can support a finding of public health benefit. The ultimate goal of studying the effect of an MRTP on human health and behavior is to be able to accurately predict the public health effects of allowing an MRTP to be marketed. In other words, the ultimate goal of scientific studies is to produce generalizable data. The “generalizability” of data, or the reliability of predictions that can be made about the real world based on scientific observations, will depend on the design of the studies.

TABLE S-1 Evidence Domains Relevant to an MRTP Application and Examples of Types of Findings

NOTE: This table is not comprehensive and is not intended to be a guideline or framework for the evaluation of MRTP applications.

Elements of study design that should be carefully evaluated include the size of the sample and the nature of the sample. Sample sizes should be carefully determined and tailored to the study design and the effects that are being studied. Statistically underpowered studies cannot support inferences or projections about the effects of a product. The nature of the study sample is critical to the usefulness of study results. Results from studies conducted in one population may not be applicable to other populations because the characteristics that define the study population either are related to or cause the responses to the product. As such, it is important to study a wide range of populations. It is particularly important to include populations that have a high risk of using tobacco and populations that will be affected by the marketing of the product.

Study designs should also carefully consider the degree of control imposed on experimental designs. Internal and external validity should be balanced not only within studies, but also across studies of the same product. Highly controlled experimental designs can eliminate many variables and confounders and support strong inferences, but simultaneously they can lose relevance to the real world as the conditions of product use do not reflect real-world circumstances and behaviors. Experimental designs that are less controlled can emulate circumstances that reflect real-world conditions and behaviors, and therefore they may be more relevant in predicting real-world effects, but uncontrolled variables may confound meaningful associations or inferences. Multiple complementary study designs will be necessary to provide the evidence necessary to meet the statutory public health standards.

Decision Making

It is clear that no single class of evidence (e.g., preclinical, RCTs, consumer perception, epidemiologic) in itself will be sufficient to support an MRTP application. The portfolio of evidence brought to the FDA to justify a modified risk or modified exposure claim will be substantial. To inform regulatory decision making, the FDA will need to process the evidence at a higher level, beyond merely amassing the evidence in support of the MRTP claims.

A key challenge facing the FDA will be integrating the various domains and levels of evidence provided by sponsors in support of an MRTP application. It would be helpful to have a systematic, explicit approach that weighs outcomes in terms of their public health importance, identifies the measures and data most relevant to those outcomes, and combines the available evidence in a manner that is psychometrically sound, objective, and reproducible. The approach to data integration that the FDA takes will be highly influential in determining whether an

MRTP is marketed, and the approach should be transparent, objective, and reproducible.

It is anticipated that modeling and simulation methods will play a role in integrating evidence to inform regulatory decisions. For modeling and simulation to be transparent, detailed information on all aspects of model structure, sources of evidence used, computational approach, construction of summaries, and reporting of the results should be available. Such information is essential not only for a proper scientific understanding of the modeling, but also for allowing researchers and other stakeholders in the regulatory process to critique and validate the model. It is important to ensure that the methods for data integration that inform decision making are neither arbitrary nor flawed.

Another critical factor in deciding whether to issue an order for the marketing of an MRTP is the amount of harm reduction claimed by an MRTP sponsor in an application. Harm reduction is inherently relative; a reduction claim is, by definition, relative to a comparison product. Selection of an appropriate comparison product is essential for informed and accurate decision making. The FSPTCA recognizes this, giving the Secretary of HHS authority to require product sponsors to compare their product to a commercially marketed representative product. The choice of appropriate comparison products will be driven by the type of MRTP being tested, the anticipated claim, and the study design. The comparison products may even differ between different classes of evidence. Two reference products come to the forefront in terms of integration and synthesis of evidence: leading brands and smoking-cessation products.

“Leading brands” represent a set of products that accounts for a significant portion of the market and could capture subgroups of interest (e.g., those of low socioeconomic status, who tend to use discount brands, and certain racial/ethnic minorities, who have higher rates of menthol cigarette use). Using leading brands increases the likelihood that the findings will have broader applicability to the population, which is crucial given the public health standard against which MRTPs are evaluated. Using leading brands as a comparator also avoids potential mischief in allowing comparisons between an MRTP and a product that is little used but inflates the apparent benefit of the MRTP.

Smoking-cessation products represent a standard (or tobaccocessation products in the case of smokeless tobacco users) as a comparison product, because these products pose very few, if any risks to health. These products provide an aspirational goal for risk and exposure from MRTPs. In principle, the closer the risks and exposures from the MRTP are to cessation products, the more confident a regulator can be in the chances for net public health benefit. Note that the use of this comparison product is not the same as studying whether the MRTP acts as an aid to

smoking cessation. Rather, the goal is to compare how the risk or exposure reduction attained with use of the MRTP compares to smoking-cessation product use of similar duration.

GOVERNANCE

The role of governance is to ensure the proper conduct of research. In addition to the essential role of protecting the interests of human research participants, governance of research is critical to the production of credible and reliable evidence. Governance and oversight of research conduct can prevent unethical behavior such as the falsification and manipulation of research data. Over time, the proper conduct of research can also build credibility and public trust.

There is profound distrust of the tobacco industry and of research supported by the tobacco industry. This distrust is the direct result of the tobacco industry’s history of improperly influencing or manipulating scientific findings and messaging about the health effects of tobacco. This history and the lack of trust may prevent independent experts from participating in research on tobacco products and therefore may impede the production of data on MRTPs necessary to assess public health impact. Particularly important illustrations of the lack of public trust include the fact that many major universities have bans on the acceptance of tobacco industry funding, and that many journals will not accept or publish research supported by the tobacco industry. Establishing the tobacco industry as a legitimate participant in tobacco research is an important consideration in the overall goal of producing evidence on the effects of MRTPs.

The need for academic institutions or other experts to conduct research on tobacco products is particularly important in the current regulatory environment. First, the tobacco industry currently lacks the infrastructure and expertise to independently produce the necessary evidence to support an application to market an MRTP. The FSPTCA now requires tobacco products to undergo a premarket approval process similar to drugs and devices. Prior to the passage of the FSPTCA, the tobacco industry lacked robust regulation, and as a consequence, the industry may lack the institutional and organizational capacity to assemble a complete application to meet the requirements of the law. In addition, there are significant domains of evidence that should be addressed in an application to market an MRTP wherein the tobacco industry either lacks the expertise or the willingness to independently conduct the research. In particular, research involving populations with a high risk for tobacco use such as behavioral research, studies of adolescents, research on abuse liability, and observational studies of health effects will be very challenging for the tobacco industry.

The committee recommends several strategies to create an environment conducive to the production of reliable and credible evidence, in spite of the tobacco industry’s reputation and currently limited infrastructure and expertise. The first strategy is to create a mechanism to distance the reputation and influence of the tobacco industry from experts, researchers, and institutions that will be critical to the production of evidence on MRTPs. Fear, either real or perceived, of being influenced by or aiding the tobacco industry prevents many institutions, researchers, and journals from having any association with the tobacco industry. Providing independence, autonomy, and separation from the industry addresses these fears. An independent third party that conducts research, provides oversight of research, distributes funding for research or manages research contracts, or otherwise provides governance of research may be a useful mechanism for reengaging the experts and institutions necessary to producing high-quality evidence on the effects of MRTPs. Relevant examples of third-party partnerships between industry and government include the Health Effects Institute and the Reagan-Udall Foundation. Currently, there are no independent entities that fulfill these roles for the tobacco industry.

The second strategy is to require that the conduct of research in support of MRTPs conform to ethical standards and that study information be made publicly available. Transparency and the proper conduct of research not only protect the interests of research participants, but also improve data quality. Requiring transparency and ethical conduct of research may also help change public perceptions of the tobacco industry, and subsequently engagement and support from key stakeholders may be more likely. Over time, requiring adherence to codes of ethics, and requiring the publication of study information and results, will improve the quality and availability of evidence about the effects of MRTPs on health.

FINDINGS AND RECOMMENDATIONS

In the committee’s view, the fundamental problem that confronts the FDA is a shortage of credible and reliable evidence about the effects of MRTPs on both individual and public health. The history of deceptive behavior by the tobacco industry undermined the trust of the public as well as the public’s confidence in the industry’s ability to rigorously conduct studies that will generate the data needed to evaluate these products. Therefore, the committee’s recommendations are designed to articulate the minimum standards for producing credible and reliable evidence to demonstrate that the marketing of an MRTP is consistent with the protection of public health. The committee articulates a strategy for the production of scientific evidence by making recommendations in three areas:

1.   Types of evidence and studies

2.   Design and integration

3.   Governance of studies

Types of Evidence and Studies

Finding 1: Types of Evidence. The public health standard articulated by the FSPTCA requires collection of scientific evidence from a wide range of disciplines and research domains. Although the committee respects the FDA’s independence and discretion in regulating MRTPs, the committee maintains there is a minimum range of research domains required to evaluate the effect of MRTPs on individuals and public health. Individual methods may change as the technology or state of the science may evolve, but the minimum standards for the domains of evidence will be relevant regardless of the state of the science in the future.

Recommendation 1: The FDA should require that studies submitted in support of an MRTP application address all key research domains needed to forecast and monitor the product’s public health impact, including

• product composition and performance;

• addiction potential and likelihood for initiation or persistence of use;

• human exposure to harmful and potentially harmful constituents;

• perceptions about the product’s effects and likelihood of addiction; and

• effects of the product on human health and surrogates of human health.

Finding 2: Phased Approach to New MRTPs. Many novel MRTPs are likely to be developed for marketing in the near future. There are inherent uncertainties and risks with new products that should be addressed. Risks should be minimized before new products are tested in humans. To address the risk of new products, a phased approach, similar to the New Drug Application framework for the regulation and control of new drugs, is appropriate for the evaluation of new MRTPs. A phased approach will help the FDA ensure that only products that are unlikely to be unsafe and have a reasonable expectation of reducing harm relative to conventional tobacco products will be used in human studies.

Recommendation 2: The FDA should establish guidance that conveys an expected sequencing of studies, such that preclinical work is completed

and submitted to the FDA before clinical (human subjects) work commences, and that there is a reasonable expectation based on preclinical work that a reduction or lack of harm will be seen in humans.

Finding 3: Clinical Trial Studies. Although the use of randomized controlled trial methods will be constrained for a number of reasons (including the practical limitations of study cost, size, and follow-up, and ethical constraints on randomizing study participants to harmful exposures), they will continue to play an essential role in creating an evidence base on the public health effects of MRTPs. Randomized controlled trial methods can provide highly reliable data on the likelihood of addiction and initiation or cessation of product use. Also, these methods can provide reliable evidence on human exposure.

Recommendation 3: The FDA should require randomized controlled trials in the following domains:

• Exposure reduction

• Self-administration of the MRTP

• Effects on use of conventional tobacco products

These randomized controlled trials should include multiple comparison products (such as nicotine replacement products, conventional cigarettes or smokeless tobacco, placebo preparations, and alternative nicotine delivery systems). These trials should also assess the effect of the MRTP on human exposure and on human health and surrogates of human health.

Finding 4: Requirement for Postmarket, Prospective Epidemiologic Studies. Postmarket studies of MRTPs will be critical to evaluating the effect of MRTPs on both individuals and the public’s health. In particular, the prospective cohort design will be an essential tool to validating anticipated or claimed effects of marketed MRTPs. These studies have several important strengths: (1) biochemical tobacco and MRTP exposure can be assessed at baseline, offering “unbiased” exposure assessment before health outcomes occur; (2) there is less of a problem with retrospective recall of product use, because this information can be summarized at the start of the study and followed prospectively; (3) changing product use habits can be monitored as the study progresses; (4) outcomes can be documented as they occur, and verification is more efficient; and (5) a wide variety of outcomes can be evaluated in the same study, particularly outcomes that are more common. Furthermore, cohort studies allow assessment of overall health status and outcomes.

Recommendation 4: The FDA should require prospective epidemiologic studies to commence upon issuance of a marketing order to confirm reduced exposure and reduced risk claims, and to examine effects of MRTP availability on the population as a whole, including the likelihood of initiation and cessation. The FDA should issue guidance on the design, conduct, and analysis of such studies.

Finding 5: Modeling of Public Health Outcomes. Mathematical modeling and simulation analysis provides a complementary approach to the conduct of empirical studies that can be useful at each stage of the regulatory process for MRTPs. Model-based analyses can (1) synthesize the available information from empirical studies of MRTPs; (2) enable researchers and decision makers to explore complex interactions and systems that may be impractical to evaluate in empirical studies; (3) allow researchers and decisions makers to explore “what if” questions relevant to decision making, which would not be practical to assess in empirical studies; and (4) be used to make projections about the short- and long- term effects of the introduction of MRTPs.

Recommendation 5: The FDA should issue guidance on the development and use of simulation and modeling approaches to predict public health impact through the systematic integration of information about relevant assumptions and influences. Such approaches should be tested for robustness with regard to results and assumptions, they should be public and transparent, and they should be validated against postmarketing epidemiologic research.

Design and Integration of Studies

Finding 6: Standards for Sampling in MRTP Studies. To have regulatory usefulness, studies of MRTPs must be generalizable to the overall population of interest and to specific populations, including populations at high risk for tobacco use. Failure to include relevant populations in studies will result in incomplete evidence on the effect of an MRTP on the public’s health and, therefore, will be inadequate to support regulatory decisions about the marketing of MRTPs.

Recommendation 6: The FDA should require studies to include populations of special relevance, including (but not limited to)

• users of tobacco products, including users who are and are not interested in quitting;

• in certain circumstances, nonusers of tobacco products;

• former smokers;

• beginning smokers;

• adolescents; and

• populations at a high risk for tobacco use, including, but not limited to, those low in socioeconomic status and educational attainment, and certain ethnic minorities.

Finding 7: Quality of Studies. The usefulness of a study to inform a regulatory decision hinges on the quality and appropriateness of the design. In many cases, complementary studies might be needed to provide a breadth of evidence for an informed regulatory decision with appropriate control of confounders and internal and external validity.

Recommendation 7: For all studies of the effects of MRTPs on human health and behavior, the FDA should require a range of designs that are properly powered, balance internal and external validity, and comprise multiple populations appropriate to the experimental questions being addressed.

Finding 8: Standards for Good Research Practice. A significant amount of guidance on minimum standards for scientific studies directly relevant to the evaluation of MRTPs has already been developed. Guidelines for formatting, design, conduct, and reporting of science are articulated in consensus statements, such as the Consolidated Standards of Reporting Trials (CONSORT) reporting criteria for clinical trials, the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement for observational studies, the publication criteria of the International Council of Medical Journal Editors, and the reporting criteria of the International Conference on Harmonization. These existing guidelines represent robust standards for the conduct of science across many of the research domains relevant to the evaluation of MRTPs.

Recommendation 8: The FDA should issue guidance to the industry regarding the format, design, conduct, and reporting of studies in support of MRTP applications that is based upon current generally accepted principles for scientific investigation.

Finding 9: Standards for Integration of Evidence. Regulatory decisions regarding MRTPs will be based on a wide range and variety of scientific evidence, and the integration of scientific evidence will play a pivotal role in that decision making. The assessment of MRTPs will typically require the evaluation and integration of evidence on risks and benefits across multiple diverse outcomes, such as measures of toxicity, biomarkers,

addictiveness, and disease endpoints. Modeling and simulation approaches are relevant to estimating public health effects of tobacco and, therefore, the FDA will likely engage in various methods of data integration, synthesis, and analysis, including, but not limited to, simulation and modeling. It is critical that these approaches are transparent and reproducible.

Recommendation 9: The FDA should develop and use an approach to data integration that is explicit and transparent with regard to the importance of the different outcomes, that uses optimal available evidence, and that employs objective and reproducible methods for data integration.

Governance of Studies

Finding 10: Independent Oversight and Conduct of Studies. It has been established in public records and as a matter of law that the tobacco industry has engaged in illegal and improper practices, including the destruction and manipulation of scientific data. As a result, the tobacco industry is profoundly isolated from the mainstream scientific community. Many major universities have policies against acceptance of tobacco funding, and many high-impact scientific and medical journals will not accept tobacco industry-supported manuscripts. The consequence of this isolation is a lack of the expertise and the resources necessary to produce high-quality science across the range of disciplines to support an application to market an MRTP. Use of a trusted third party, particularly for products developed by the tobacco industry, could provide an avenue for the production of credible evidence needed by the FDA to evaluate tobacco products. Ultimately, such a research structure could encourage and support the production and dissemination of credible and reliable evidence about the effects of tobacco products on the public’s health.

Recommendation 10: MRTP sponsors should consider use of independent third parties to undertake one or more key functions, including the design and conduct of research, the oversight of specific studies, and the distribution of sponsor funds for research. Such independent third parties should be approved by the FDA in advance of the research.

Finding 11: Public Disclosure of Research. Public availability of data not only builds credibility and public trust, but also benefits the public because it allows for independent analysis of study methods and data. The model of Clinicaltrials.gov is particularly compelling and relevant,

and a similar model of public accounting and open disclosure should be expected of the tobacco industry.

Recommendation 11: The FDA should require all MRTP sponsors to place all data generated in the development and marketing of the MRTP in a public repository selected by the FDA.

Finding 12: Proper Conduct of Research. Standards for the conduct of science and the protection of human research participants have been established for biomedical research enterprises not only in academics but also in commercial research. The FDA has the tools to ensure studies adhere to established standards in the drug development framework, which can be applied to the development of MRTPs. Those standards not only protect human participants, but they also build credibility into any data that are provided to the FDA, particularly by the tobacco industry. Institutional credibility and trustworthiness is particularly relevant in this context, given the history of unethical and illegal practices of the tobacco industry.

Recommendation 12: The FDA should require studies offered in support of an MRTP application to adhere to establiShed standards and principles of good research governance, including appropriately qualified investigators, transparency, independent institutional review board or ethical review, and adherence to the Common Rule (21 CFR parts 50 and 56).

REFERENCES

CDC (Centers for Disease Control and Prevention). 2002. Cigarette smoking among adults—United States, 2000. Morbidity and Mortality Weekly Report 51(29):642.

CDC. 2008. Smoking-attributable mortality, years of potential life lost, and productivity losses—United States, 2000-2004. Morbidity and Mortality Weekly Report 57(45):1226-1228.

CDC. 2009a. Tobacco use: Targeting the nation’s leading killer—at a glance 2009. National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, Department of Health and Human Services.

CDC. 2009b. Cigarette smoking among adults and trends in smoking cessation—United States, 2008. Morbidity and Mortality Weekly Report 58(44):1227-1232.

CDC. 2010. Vital signs: Current cigarette smoking among adults aged ≥ 18 years—United States, 2009. Morbidity and Mortality Weekly Report 59(35):1135-1140.

CDC. 2011. Ten great public health achievements—United States, 2001-2010. Morbidity and Mortality Weekly Report 60(19):619-623.

HHS (U.S. Department of Health and Human Services). 2000. Reducing tobacco use: A report of the surgeon general. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health.

IOM (Institute of Medicine). 2001. Clearing the smoke: Assessing the science base for tobacco harm reduction. Washington, DC: National Academy Press.

Marlatt, G. A. 2002. Harm reduction: Pragmatic strategies for managing high-risk behaviors. New York: The Guilford Press.

Mokdad, A. H., J. S. Marks, D. F. Stroup, and J. L. Gerberding. 2004. Actual causes of death in the United States, 2000. JAMA 291(10):1238-1245.