Editorial For reprint orders, please contact [email protected]
Assessing sex differences: methodological considerations Expert Rev. Clin. Pharmacol. 1(5), 585–587 (2008)
Kathleen Uhl, MD Author for correspondence
Assistant Commissioner for Women’s Health, US Food and Drug Administration, Office of Women’s Health, 5600 Fishers Lane, HF-8, Rockville, MD 20857, USA Tel.: +1 301 827 0350 Fax: +1 301 827 0926 [email protected]
Sherry Marts, PhD Society for Women’s Health Research Director, Scientific Affairs, 1025 Connecticut Ave, NW, Washington, DC 20036, USA Tel.: +1 202 496 5019 Fax: +1 202 833 3472 sherry@ womenshealthresearch.org
www.expert-reviews.com
“The importance of evaluating sex differences in response cannot be overstated, and knowledge and understanding of sex differences is a basic component for the development of personalized medicine.” Medical and scientific journals report a variety of findings owing to biological differences in the study participants, including those due to sex. Sex differences occur in disease pathophysiology, diagnosis, prognosis, treatment, prevention and health outcomes involving nearly all medical disciplines, therapeutic areas and healthcare professionals. Following administration of a pharmacologic agent, differences in safety and efficacy attributed to demographic variability are frequently explored and reported in the medical literature. Review articles of pharmacokinetic and/or pharmacodynamic differences observed between the sexes have been published [1] . In this article, we take a more provocative approach and explore the concept of sex differences in clinical study design from a methodological perspective and challenge some innate assumptions. Before a study can truly conclude that the data demonstrate a difference (or no difference) in response to treatment between men and women, several critical concepts must be considered.
federal agencies, as well as medical literature, document the historic and ongoing under-representation of women in clinical studies. Since the early 1990s, advances in public policy, specifically at the US FDA and NIH, have prohibited excluding women from clinical studies except when scientifically justified. Despite this progress, adequate representation is still lacking in particular areas of drug development, such as Phase I dose-finding studies, which drive subsequent dose selection and in specific therapeutic areas, such as cardiovascular disease (CVD). Even in the largest clinical studies for CVD, the level of female participation is typically 30% or less [2] – substantially lower than the prevalence of the disease in the general adult population. The inclusion of low numbers of women in trials is defended with the argument that increasing the number of study participants is prohibitive, citing concerns over difficulties in the recruitment and retention of women subjects and the number of study participants needed for valid subgroup analysis, resulting in increased time to completion of the study and increased expense. However, more “The appropriate inclusion of than 10 years of experience, including women in clinical research the successful enrollment and retention of studies is required to allow subjects in the Women’s Health Initiative, investigation of sex differences have demonstrated that recruitment and in treatment response…” retention strategies that are adapted to The appropriate inclusion of women women (and other subpopulations) can in clinical research studies is required be developed and implemented and can to allow investigation of sex differences also increase the pace of enrollment of in treatment response – whether it be men. Specific obstacles that may discourfor safety or efficacy. Numerous reports age or prevent women from participating from the Institute of Medicine (IOM) in trials include transportation costs, the and Government Accounting Office, need for child or elder care and the time
10.1586/17512433.1.5.585
© 2008 Expert Reviews Ltd
ISSN 1751-2433
585
Editorial
Uhl & Marts
commitment required. Consideration must be given to minimizing the inconvenience of participating in studies, such as minimizing the number of clinic visits by coordinating procedures and tests, extending clinic hours to evenings and weekends and, if possible, avoiding overnight stays in a clinical research unit. The growing experience with pediatric clinical studies demonstrates that appropriate study design adaptations have led to successful study participant recruitment and retention [3] . The question of sex differences in disease progression or therapeutic response is not necessarily served by simply ensuring an adequate sample size of both sexes. It requires a rational and scientific approach to detect, explore and understand the differences that emerge. Frequently, these types of analyses are performed post hoc, are descriptive in nature and look at trends in the data; they are not robust preplanned statistical analyses. Unfortunately, even these basic subgroup analyses are often not performed [4] . Methodological consideration must be given to study design and data analysis. At a minimum, a prospective plan for achieving adequate statistical power for valid subgroup analyses is required. The question of whether to design the study to have adequate statistical power to detect sex differences is crucial, and consideration should be given on how best to report results, positive or negative, if statistical power is inadequate. The editors of The New England Journal of Medicine have set forth guidelines for the use of subgroup analyses in their journal, recognizing the analytical challenges inherent to such analyses and the ensuing overstated and misleading results [5] . Results from exploratory post hoc analyses should not be reported as definitive or conclusive findings, but rather should be used to generate hypotheses for further testing. A few assumptions in clinical study design impact the ability to detect sex differences if, in fact, they are present. The assumption that the disease is the same in both men and women may be faulty. This assumption has proven erroneous, as we are now realizing with CVD, the most common cause of mortality in men and, importantly, women. The long-held belief that men and women were the same with respect to CVD has compromised the health of numerous women by missed diagnoses and poorly managed healthcare [6] . Not only does heart disease present differently between men and women, but the disease itself differs, and we are just beginning to understand the underlying mechanisms behind these differences [7] . Therefore, the results seen in a pharmacologic study in patients with CVD may be confounded by differences in the disease itself between the sexes and not merely response or outcome differences or similarities attributed solely to the drug. For any disease, the pathophysiology, the molecular basis for the disease, the genetic, genomic and proteomic manifestations and their inter-relationships may not be the same for men and women, and advances in clinical pharmacology can better discern these differences and advance knowledge in this area. Another consideration in study design is the assumption that men and women with the same signs and symptoms of disease are starting at the same point, that is, they have similar baselines; however, this assumption is not always true. For example, 586
women diagnosed with HIV infection were found to have a lower viral load at comparable CD4 + cell counts compared with men, which led to women being undertreated when physicians followed early HIV treatment protocols that were based on viralload data from studies conducted solely in men [8] . Sex differences in baseline electrocardiogram measurements confounded the interpretation of data to assess drug-related prolongation of the QT interval. Pain perception and tolerance vary between the sexes [9] , making it challenging to ensure similar baseline in pain-treatment studies. Another potential source of inaccurate reporting of sex differences is that the tool or instrument used to measure response may not have been independently validated in women and men. While it may be safely assumed that certain physicochemical measurements, such as blood pressure or blood chemistries, are valid in each sex, this assumption may be problematic with tools or instruments based on patient reports. Again, the corollary to pediatric and adolescent patient studies that required the development of assessment methodologies specific for that patient population is illustrative [3] . Thus, studies that fail to detect differences between the sexes may, in fact, do so because the measurement methodology may not be equally valid in both sexes.
“New and developing clinical pharmacology tools
that include pharmacogenomic, proteomic and metabolomic technologies … can further advance our understanding of sex differences.” Until the last quarter of the 20th Century, sex was not recognized as a variable in health research nor was it believed to be a factor that could affect health and illness. A 2001 IOM committee report concluded that sex should be considered when designing and analyzing studies in all areas and at all levels of biomedical and health-related research [101] . The importance of evaluating sex differences in response cannot be overstated, and knowledge and understanding of sex differences is a basic component for the development of personalized medicine. An improved understanding of biological sex differences will help us to better understand risk for disease, disease progression and differential response to treatment, also helping to identify patient populations at risk for adverse events, delayed complications or lack of efficacy. New and developing clinical pharmacology tools that include pharmacogenomic, proteomic and metabolomic technologies, as well as advances in information technologies to pool and mine data, can further advance our understanding of sex differences. The end result will be optimal diagnostics and therapeutics to improve the health of both sexes. Financial & competing interests disclosure
The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript. Expert Rev. Clin. Pharmacol. 1(5), (2008)
Assessing sex differences: methodological considerations
References 1
Franconi F, Brunelleschi S, Steardo L, Cuomo V. Gender differences in drug responses. Pharmacol. Res. 55(2), 81–95 (2007).
2
Harris, DJ, Douglas PS. Enrollment of women in cardiovascular clinical trials funded by the National Heart, Lung, and Blood Institute. N. Engl. J. Med. 343, 475–480 (2000).
3
Abdel-Rahman SM, Reed MD, Wells TG, Kearns GL. Considerations in the rational design and conduct of Phase I/II pediatric clinical trials: avoiding the problems and pitfalls. Clin. Pharmacol. Ther. 81(4), 483–494 (2007).
www.expert-reviews.com
4
Diagnosis and treatment of coronary heart disease in women: systematic reviews of evidence on selected topics. Evidence report/technology assessment: Number 81. Agency for Healthcare Research Quality Publication Number 03-E035 (2003).
5
Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in medicine – reporting of subgroup analyses in clinical trials. N. Engl. J. Med. 357(21), 2189–2194 (2007).
6
Blauwet LA, Redberg RF. The role of sex-specific results reporting in cardiovascular disease. Cardiol. Rev. 15(6), 275–278 (2007).
7
Wegner NK, Shaw LJ, Vaccarino V. Coronary heart disease in women: update 2008. Clin. Pharmacol. Ther. 83(1), 37–51 (2008).
Editorial
8
Sterling TR, Vlahov D, Astemborski J, Hoover DR, Margolick JB, Quinn TC. Initial plasma HIV-1 RNA levels and progression to AIDS in women and men. N. Engl. J. Med. 344, 720–725 (2001).
9
Wiesenfeld-Hallin Z. Sex differences in pain perception. Gend. Med. 2(3), 137–145 (2005).
Website 101
Institute of Medicine Committee on Understanding the Biology of Sex and Gender Differences. Exploring the Biological Contributions to Human Health: Does Sex Matter? Wizemann TM, Pardue ML (Eds). National Academies Press, Washington, DC, USA (2001) www.nap.edu/books/0309072816/html
587
Assessing sex differences: methodological considerations Expert Rev. Clin. Pharmacol. 1(5), 585–587 (2008)
Kathleen Uhl, MD Author for correspondence
Assistant Commissioner for Women’s Health, US Food and Drug Administration, Office of Women’s Health, 5600 Fishers Lane, HF-8, Rockville, MD 20857, USA Tel.: +1 301 827 0350 Fax: +1 301 827 0926 [email protected]
Sherry Marts, PhD Society for Women’s Health Research Director, Scientific Affairs, 1025 Connecticut Ave, NW, Washington, DC 20036, USA Tel.: +1 202 496 5019 Fax: +1 202 833 3472 sherry@ womenshealthresearch.org
www.expert-reviews.com
“The importance of evaluating sex differences in response cannot be overstated, and knowledge and understanding of sex differences is a basic component for the development of personalized medicine.” Medical and scientific journals report a variety of findings owing to biological differences in the study participants, including those due to sex. Sex differences occur in disease pathophysiology, diagnosis, prognosis, treatment, prevention and health outcomes involving nearly all medical disciplines, therapeutic areas and healthcare professionals. Following administration of a pharmacologic agent, differences in safety and efficacy attributed to demographic variability are frequently explored and reported in the medical literature. Review articles of pharmacokinetic and/or pharmacodynamic differences observed between the sexes have been published [1] . In this article, we take a more provocative approach and explore the concept of sex differences in clinical study design from a methodological perspective and challenge some innate assumptions. Before a study can truly conclude that the data demonstrate a difference (or no difference) in response to treatment between men and women, several critical concepts must be considered.
federal agencies, as well as medical literature, document the historic and ongoing under-representation of women in clinical studies. Since the early 1990s, advances in public policy, specifically at the US FDA and NIH, have prohibited excluding women from clinical studies except when scientifically justified. Despite this progress, adequate representation is still lacking in particular areas of drug development, such as Phase I dose-finding studies, which drive subsequent dose selection and in specific therapeutic areas, such as cardiovascular disease (CVD). Even in the largest clinical studies for CVD, the level of female participation is typically 30% or less [2] – substantially lower than the prevalence of the disease in the general adult population. The inclusion of low numbers of women in trials is defended with the argument that increasing the number of study participants is prohibitive, citing concerns over difficulties in the recruitment and retention of women subjects and the number of study participants needed for valid subgroup analysis, resulting in increased time to completion of the study and increased expense. However, more “The appropriate inclusion of than 10 years of experience, including women in clinical research the successful enrollment and retention of studies is required to allow subjects in the Women’s Health Initiative, investigation of sex differences have demonstrated that recruitment and in treatment response…” retention strategies that are adapted to The appropriate inclusion of women women (and other subpopulations) can in clinical research studies is required be developed and implemented and can to allow investigation of sex differences also increase the pace of enrollment of in treatment response – whether it be men. Specific obstacles that may discourfor safety or efficacy. Numerous reports age or prevent women from participating from the Institute of Medicine (IOM) in trials include transportation costs, the and Government Accounting Office, need for child or elder care and the time
10.1586/17512433.1.5.585
© 2008 Expert Reviews Ltd
ISSN 1751-2433
585
Editorial
Uhl & Marts
commitment required. Consideration must be given to minimizing the inconvenience of participating in studies, such as minimizing the number of clinic visits by coordinating procedures and tests, extending clinic hours to evenings and weekends and, if possible, avoiding overnight stays in a clinical research unit. The growing experience with pediatric clinical studies demonstrates that appropriate study design adaptations have led to successful study participant recruitment and retention [3] . The question of sex differences in disease progression or therapeutic response is not necessarily served by simply ensuring an adequate sample size of both sexes. It requires a rational and scientific approach to detect, explore and understand the differences that emerge. Frequently, these types of analyses are performed post hoc, are descriptive in nature and look at trends in the data; they are not robust preplanned statistical analyses. Unfortunately, even these basic subgroup analyses are often not performed [4] . Methodological consideration must be given to study design and data analysis. At a minimum, a prospective plan for achieving adequate statistical power for valid subgroup analyses is required. The question of whether to design the study to have adequate statistical power to detect sex differences is crucial, and consideration should be given on how best to report results, positive or negative, if statistical power is inadequate. The editors of The New England Journal of Medicine have set forth guidelines for the use of subgroup analyses in their journal, recognizing the analytical challenges inherent to such analyses and the ensuing overstated and misleading results [5] . Results from exploratory post hoc analyses should not be reported as definitive or conclusive findings, but rather should be used to generate hypotheses for further testing. A few assumptions in clinical study design impact the ability to detect sex differences if, in fact, they are present. The assumption that the disease is the same in both men and women may be faulty. This assumption has proven erroneous, as we are now realizing with CVD, the most common cause of mortality in men and, importantly, women. The long-held belief that men and women were the same with respect to CVD has compromised the health of numerous women by missed diagnoses and poorly managed healthcare [6] . Not only does heart disease present differently between men and women, but the disease itself differs, and we are just beginning to understand the underlying mechanisms behind these differences [7] . Therefore, the results seen in a pharmacologic study in patients with CVD may be confounded by differences in the disease itself between the sexes and not merely response or outcome differences or similarities attributed solely to the drug. For any disease, the pathophysiology, the molecular basis for the disease, the genetic, genomic and proteomic manifestations and their inter-relationships may not be the same for men and women, and advances in clinical pharmacology can better discern these differences and advance knowledge in this area. Another consideration in study design is the assumption that men and women with the same signs and symptoms of disease are starting at the same point, that is, they have similar baselines; however, this assumption is not always true. For example, 586
women diagnosed with HIV infection were found to have a lower viral load at comparable CD4 + cell counts compared with men, which led to women being undertreated when physicians followed early HIV treatment protocols that were based on viralload data from studies conducted solely in men [8] . Sex differences in baseline electrocardiogram measurements confounded the interpretation of data to assess drug-related prolongation of the QT interval. Pain perception and tolerance vary between the sexes [9] , making it challenging to ensure similar baseline in pain-treatment studies. Another potential source of inaccurate reporting of sex differences is that the tool or instrument used to measure response may not have been independently validated in women and men. While it may be safely assumed that certain physicochemical measurements, such as blood pressure or blood chemistries, are valid in each sex, this assumption may be problematic with tools or instruments based on patient reports. Again, the corollary to pediatric and adolescent patient studies that required the development of assessment methodologies specific for that patient population is illustrative [3] . Thus, studies that fail to detect differences between the sexes may, in fact, do so because the measurement methodology may not be equally valid in both sexes.
“New and developing clinical pharmacology tools
that include pharmacogenomic, proteomic and metabolomic technologies … can further advance our understanding of sex differences.” Until the last quarter of the 20th Century, sex was not recognized as a variable in health research nor was it believed to be a factor that could affect health and illness. A 2001 IOM committee report concluded that sex should be considered when designing and analyzing studies in all areas and at all levels of biomedical and health-related research [101] . The importance of evaluating sex differences in response cannot be overstated, and knowledge and understanding of sex differences is a basic component for the development of personalized medicine. An improved understanding of biological sex differences will help us to better understand risk for disease, disease progression and differential response to treatment, also helping to identify patient populations at risk for adverse events, delayed complications or lack of efficacy. New and developing clinical pharmacology tools that include pharmacogenomic, proteomic and metabolomic technologies, as well as advances in information technologies to pool and mine data, can further advance our understanding of sex differences. The end result will be optimal diagnostics and therapeutics to improve the health of both sexes. Financial & competing interests disclosure
The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript. Expert Rev. Clin. Pharmacol. 1(5), (2008)
Assessing sex differences: methodological considerations
References 1
Franconi F, Brunelleschi S, Steardo L, Cuomo V. Gender differences in drug responses. Pharmacol. Res. 55(2), 81–95 (2007).
2
Harris, DJ, Douglas PS. Enrollment of women in cardiovascular clinical trials funded by the National Heart, Lung, and Blood Institute. N. Engl. J. Med. 343, 475–480 (2000).
3
Abdel-Rahman SM, Reed MD, Wells TG, Kearns GL. Considerations in the rational design and conduct of Phase I/II pediatric clinical trials: avoiding the problems and pitfalls. Clin. Pharmacol. Ther. 81(4), 483–494 (2007).
www.expert-reviews.com
4
Diagnosis and treatment of coronary heart disease in women: systematic reviews of evidence on selected topics. Evidence report/technology assessment: Number 81. Agency for Healthcare Research Quality Publication Number 03-E035 (2003).
5
Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in medicine – reporting of subgroup analyses in clinical trials. N. Engl. J. Med. 357(21), 2189–2194 (2007).
6
Blauwet LA, Redberg RF. The role of sex-specific results reporting in cardiovascular disease. Cardiol. Rev. 15(6), 275–278 (2007).
7
Wegner NK, Shaw LJ, Vaccarino V. Coronary heart disease in women: update 2008. Clin. Pharmacol. Ther. 83(1), 37–51 (2008).
Editorial
8
Sterling TR, Vlahov D, Astemborski J, Hoover DR, Margolick JB, Quinn TC. Initial plasma HIV-1 RNA levels and progression to AIDS in women and men. N. Engl. J. Med. 344, 720–725 (2001).
9
Wiesenfeld-Hallin Z. Sex differences in pain perception. Gend. Med. 2(3), 137–145 (2005).
Website 101
Institute of Medicine Committee on Understanding the Biology of Sex and Gender Differences. Exploring the Biological Contributions to Human Health: Does Sex Matter? Wizemann TM, Pardue ML (Eds). National Academies Press, Washington, DC, USA (2001) www.nap.edu/books/0309072816/html
587
