Pharmacology Conference

Clinical Practice
Studies advancing the understanding of PK/PD gender issues have begun to generate an understanding of the differences in clinical effects. For example, a study published in 1996 showed that a class of opiate drugs were more effective in women than men following dental surgery.13 It has been shown that the menstrual cycle changes the PK of only a few commonly prescribed medications.13 Treatment of women with HIV has sparked research interest in the relationship between new antiviral agents and effects on menstrual cycles. Similarly, principal investigators have begun to address questions about menstrual cycle effects while conducting clinical trials to evaluate specific drug effects on women. Significant advances have also been made in recognizing the importance of adverse effects occurring during drug therapy and/or following ingestion of dietary supplements.

. Gaps in Knowledge
Despite the impressive advances in policies regarding the participation of women as subjects in research and analysis of data for gender effects, gaps in our knowledge remain regarding the behavior of drugs in women. Areas where much additional research is needed include mechanisms of drug action, which include PK, PD, pharmacogenetics; biological/molecular basis of pharmacological effects such as ion channels and membrane transporters; chronopharmacology; and modulators such as sex steroid hormones that can influence PK and receptor sensitivity in target end organs. Similarly, progress is also needed for analysis of data by sex.

Research is also needed in identification and quantification of risks associated with use of pharmacologic agents at different points in the life span. For example, there are major gaps in our knowledge of basic mechanisms of action for drugs taken by pregnant women. Long-term safety data of therapeutic agents in regard to fetal outcome are also lacking. Studies are needed to explore the interaction of drugs that have parallel metabolic pathways with exogenous hormones (hormonal contraceptives and hormone replacement therapy).

Components of Pharmacologic Effects
Recent reviews describe the status of gender-based PK and PD research.11,12 Although significant research advances have been made, the gaps in our knowledge base in this area represent a serious impediment to understanding and assessing the benefits and risks of both current and proposed clinical interventions during a woman's life cycle. Research is needed to evaluate the altered susceptibility to pharmacological agents throughout the life span with special issues pertinent to:

Although gender differences have long been observed in many areas of biology, until relatively recently it has not been appreciated that these differences may contribute to the variability of drug response in humans. However, it is becoming increasingly clear that males and females have differences in drug response (traditionally thought to be related to changes in body mass, body fat, and muscle mass), for both PK and PD reasons. For example, fluvoxamine plasma concentrations are 40-50 percent lower in men than women, but the mechanism underlying this difference is unknown and gender-based PD data are not available.14 In a drug interaction study, isradipine was shown to significantly decrease the area under the curve (AUC) of lovastatin in male but not female subjects.15 A systematic study of the mechanisms for such gender-related variability in drug response has yet to be undertaken. Nor is there presently any systematic requirement to carry out PD studies whenever gender-related PK variability is observed.

Pharmacogenetics. Pharmacogenetics is defined as hereditary variations in response to drugs,16 where the majority of variability is caused by alteration in the functional activity of metabolic enzymes. The study of polymorphism (i.e., capacity to exist in many forms) of genes and their proteins (including enzymes involved in drug metabolism) is a rapidly expanding field. Families of enzymes including cytochrome P450s, UDP glycosyltransferases, sulfotransferases, and N-acetyltransferases catalyze a large number of metabolic reactions. Gender differences in the metabolism of certain drugs metabolized by specific isozymes of P450 have been found.17 However, for the most part gender differences in the amount of the enzymes are not well defined in vitro.18,19

Genetic differences in the activity of enzymes have been demonstrated for N-acetyltransferases and the cytochrome P450 isozymes, CYP 2D6 and CYP 2C19.20-22 There are definite ethnic/racial differences in the genetic polymorphism found. For example, although 5-10 percent of whites lack CYP 2D6, less than 2 percent of Asians and blacks are deficient. Similarly, CYP 2C19 deficiency occurs in 15-25 percent of Asians, but in only 2-5 percent of whites and blacks. However, there are few data available regarding the interaction between gender and genetic polymorphism. Only a few of the studies that have both (a) measured the incidence of genetic polymorphism and (b) also specifically mentioned the inclusion women in the study population have analyzed the data based on gender. For example, a review of the literature on the polymorphism of CYP 2C19 is given in Table 1. Of the 38 studies evaluated, only 9 studies performed any type of gender analysis and only 6 studies provided these data. In two of the studies, the incidence of the poor metabolizer phenotype/genotype was numerically smaller in women than in men; in a single study, the incidence was higher in women than in men, and in two studies no difference was noted. However, as demonstrated by Evans et al.,23 the gender effect may be race dependent. In Saudi Arabian women, the incidence of CYP 2C19 deficiency tended to be greater, and in Filipino women the incidence tended to be lower. Based on these studies, it appears that CYP 2C19 genetic polymorphism may not be strongly influenced by gender. This conclusion must be tempered, however, with the possibility that all the studies cited may have lacked the necessary statistical power to determine gender differences.

If most drugs had been developed in mixed-gender populations then the extent to which these issues contribute to gender-related variability in drug response would be clearer. However, women were systematically excluded from many clinical trials until relatively recently, and gender-related differences in drug response have only been sporadically reported.

On the other hand, there is some evidence that the biological mechanisms that contribute to PK and PD of drugs have gender-related differences. For example, among the P450 drug-metabolizing enzymes, women have been reported to express higher levels of CYP3A24,25 and lower levels of the several other P450 isozymes than do men,24 although with sparteine as a marker, there was reported to be no effect of gender on CYP 2D6 activity.26 In another study in which 300 mg of oral caffeine were administered, female subjects exhibited significant toxic effects. These women had markedly lower N-demethylation indices as well as lower levels of CYP1A2.27 Women also apparently express lower levels of glucuronyltransferases.24,26 However, among the other families of drug-metabolizing enzymes, including the sulfotransferases and the glutathione-S-transferases, less is known.

Biologic/Molecular Basis of Pharmacological Factors
Membrane Transporters. Besides differences in drug-metabolizing enzymes, other gender-specific processes may contribute to differences in absorption, distribution, and excretion of drugs. These include the many membrane transporters, for which therapeutic agents may be substrates, such as the drug efflux pumps p-glycoprotein and multidrug resistance protein (MRP). These transporters pump substrates out of the cell, and the overexpression of MRP has been shown to be linked with resistance to anticancer therapies. The metabolite of the female hormone estradiol, estradiol-17-beta-glucuronide, is pumped into the bile by MRP and is also pumped out of tumor cells that express MRP.28,29 In addition, female mice express a higher basal mdr2 p-glycoprotein than do male mice and thus are able to secrete phospholipids with greater activity.30 Although gender differences in the expressions of the efflux pumps could contribute to differences in drug pharmacokinetics and response, little work has been done in this area.

Ion Channels. In recent years, sex hormones have been shown to have effects on tissues other than those involved in reproduction. The heart has sex hormone receptors, and exposure to these hormones can alter the expression of ion channels responsible for the electrical activity of the heart. These relationships may help us understand clinical observations such as the relationship between the menstrual cycle and the occurrence of cardiac arrhythmias. Importantly, it may also explain gender differences in cardiac response to drugs. A large number of diverse drugs such as antiarrhythmic drugs, terfenadine, astemizole, cisapride, halofantrine, and erythromycin have the potential to induce a possibly lethal cardiac arrhythmia, torsades de pointes. The mechanism of this not uncommon adverse drug reaction is blockage of potassium channels. One remarkable feature of this adverse event is that the risk of developing this type of arrhythmia is far greater in women than in men who take these drugs.4-8 Most likely, this is because potassium channel expression and sensitivity to drugs are regulated by sex hormones.7 Further research in this area is urgently needed.

Increased sensitivity of cardiac sodium channels to bupivacaine has been demonstrated after treatment with progesterone.8 Such sensitivity is the most likely explanation for reported deaths among pregnant women given this drug for local anesthesia during delivery.

In an example of gender differences in receptors, the density of 5-HT2 receptors was found to be significantly up-regulated in platelets from depressed women.31 Therefore, a measure of the 5-HT2 density may, in the future, be used as a marker for depression in women, although not in men. However, further research is needed to verify this finding. When monitoring for changes in physiological levels of prolactin as a function of drug therapy, both buspirone and dl-fenfluramine were associated with greater prolactin responses in women than in men.32,33 Buspirone and dl-fenfluramine may cause greater changes in normal physiological functions in women than in men.

Modulators: Sex Hormones. The sex hormones are obviously expressed differently between the genders. Exogenously administered steroidal hormones, such as oral contraceptives, have been shown to have effects on drug metabolism.24 Rat data have shown that ethenyl steroids bind irreversibly to rat hepatic cytochrome P450, thereby blocking this pathway of metabolic activity.34 In vitro evidence suggests that progesterone affects CYP3A4 by either inhibition or activation.35 Such impairment of metabolic capability has significant implications for drug-drug interactions when another drug is also metabolized by the same P450 enzyme. Whether endogenous sex hormones are responsible for gender-related differences in drug response is unknown, but existing evidence is intriguing. There is evidence, for instance, that changes in hormonal levels during the menstrual cycle may influence the disposition of drugs. For example, among debrisoquine (a marker for CYP2D6) extensive metabolizers, the debrisoquine metabolic ratio (MR) was significantly lower in the luteal phase of the menstrual cycle than in the ovulatory phase or pre-ovulation.36 The metabolism of alfentanil and prednisolone, both metabolized by CYP3A4, are decreased after menopause during a time when estrogen and progesterone are greatly reduced.37,38

Pregnancy and its dramatic effect on sex hormones are known to alter drug metabolism. An example is the increased elimination of anti-epileptic drugs in pregnant women.24 Interestingly, women who have never given birth were reported to have decreased CYP1A2 levels, but those who had given birth had the same level of CYP1A2 as men.39 We are currently unable to predict whether these changes are generalizable. Until this is studied systematically, variability in drug response will continue to be a problem in therapeutic dosing.

Chronopharmacology. Chronopharmacology is the study of the time-dependent dosing of pharmacologic agents. Chronopharmacology is dependent upon inherited factors such as gender-related differences and genetic variations,40,41 age-related differences,42,43 inter-individual differences in chronoeffectiveness of drugs related to diseases such as cancer,44,45 as well as drug-dependent alteration due to biological rhythms. Chronopharmacology also takes into consideration specified chronobiotics, agents that can influence biologic rhythms.

The area of chronopharmacology is relatively new, with the establishment of a journal in that area in the 1980's. Therefore, chronopharmacology must take into consideration basic principles of medicine and individual patient characteristics. Minimal work has been done to examine the chronopharmacologic differences or similarities between women and men.

A chronopharmacokinetic gender-related difference in men and women was demonstrated for the cephalosporin antibiotic, cefodizine. When this drug was administered at four separate times over a 24-hour period, the AUCs for women were significantly higher (p<0.001) than those for men. The largest AUC occurred at midnight for both women and men, and the smallest occurred at 6:00 pm.Even when chronopharmacokinetic effects are observed, other factors may need to be considered. For example, in the case of indomethacin, a nonsteroidal anti-inflammatory drug, chronopharmacokinetic variation has been noted in young, healthy subjects and in patients with arthritis,42 but not in elderly patients.43

Summary Statement
Importance of Combining Pharmacokinetic and Pharmacodynamic Data: Consideration of the Interactive Nature of Factors Yielding Pharmacological Effects. Although it has become apparent that there are many gender differences in the PK of drugs, the existence or magnitude of PD differences cannot always be predicted accurately. This phenomenon is well illustrated in the case of the drug propranolol.

In the NIH-sponsored "Beta Blocker Heart Attack Trial," women seemed to have higher plasma concentrations of propranolol than men given equivalent dosages. A natural response to this information would have been to recommend that women be given lower dosages of propranolol to correct for their lower clearance of the drug. However, Flockhart et al. compared the actual degree of beta adrenoceptor blockade with propranolol in men and women using an isoproterenol challenge infusion and found that women had a lower sensitivity to propranolol that compensated for their higher plasma concentrations.9 This difference negated the need for lower doses, which could have reduced the overall efficacy of the drug. Examples such as these clearly demonstrate the need to examine PD research in conjunction with PK research in order to avoid erroneous conclusions. In addition, emerging information on other components of pharmacological effects should be considered in terms of their interactive effects. Figure 2 provides a new model for considering these interactive effects on overall drug safety/toxicity and efficacy.

Figure 2: Interrelationship of Pharmacological Parameters

Preclinical Studies
In keeping with newer approaches to the study of pharmacological effects, there is an urgent need to develop reliable preclinical (animal and in vitro) models that will provide information about mechanisms of action of drugs, pharmacokinetics and pharmacodynamics, and aid in assessing the role of hormones in drug metabolism. More detailed preclinical information is also needed to guide the planning and implementation of clinical development programs and to improve the efficiency of human testing.

Drug Use in Pregnancy
Pregnant women, like all other groups of individuals, may benefit from pharmacological agents. However, concern for potential fetal harm caused by drugs taken during pregnancy has resulted in exclusion of pregnant women from clinical research. This exclusion was also extended to women of childbearing potential to prevent fetal exposure to unknown effects of a new drug. In 1993, FDA issued a guideline that removed previous policies that had recommended that women of childbearing potential be excluded from phase I and II studies of non-life-threatening diseases.47 Protocols for drug testing have included criteria to minimize the risk of pregnancy in women of childbearing potential enrolled in clinical trials and taking the experimental agent. Thus, the number of pregnancies where there is exposure to experimental agents is limited. To further compound this problem, no systematic program exists to monitor new drugs for untoward effects on pregnant women once they are approved despite the widespread use of pharmaceutical agents by pregnant women. Although there may be unexpected adverse consequences when pregnant women take therapeutic agents necessary for treatment of a medical problem, the precise etiology of an adverse event is often difficult to determine. Background rates of 3-4 percent fetal anomalies, and even higher rates of spontaneous and therapeutic abortions, confound the clinical picture and may mask a drug-related outcome.

In addition to the lack of information on safety of therapeutics with regard to fetal outcome, data are not available to address safety or dosing considerations for the pregnant woman herself. The physiology of pregnancy may provide an environment in which some toxic effects (such as those involving hormone-sensitive organs) of particular drugs may be more likely to be manifest. Also, changes in renal and hepatic function, plasma protein binding, and volume of distribution may substantially alter the pharmacokinetics of drugs in pregnancy so that some drugs may need to be dosed very differently in pregnant women and nonpregnant women.

Pregnant women have great potential for exposure to medications,48,49 for pre-existing medical conditions (such as epilepsy), for conditions occurring during pregnancy (such as hypertension), and also through inadvertent exposure before becoming pregnant or during early-phase pregnancy (before a woman is aware she is pregnant; in the U.S., 50 percent of pregnancies are unplanned).50 Because most medications have little or no information on use during pregnancy or on the risk for the developing human fetus, scientifically valid and rigorous preclinical methods for evaluating potential risks in pregnant women must be developed. Current methods for collecting information on pregnancy exposure to drugs are limited. Some pharmaceutical companies may have a registry to receive reports on birth outcomes from women who have taken a particular drug while pregnant. However, the value of these registries is limited: these registries are not widely known, reporting is voluntary, the numbers of reports may be small, followup may be incomplete, and other factors related to pregnancy outcome may not be well described.

The FDA may receive an individual report when a baby is born with a major malformation and the mother remembers having taken a drug during her pregnancy. If the birth defect is very rare, cluster reports may be helpful. However, this technique is "numerator" oriented and is unable to determine the frequency of adverse event occurrence within the exposed population. This approach can be used for generating hypotheses, but not for providing definitive information. Additional studies are usually needed to determine the nature and strength of any association between an adverse event and exposure to a drug. The problem of establishing whether a true association exists is compounded by the bias introduced in any retrospective study. Whenever there is an adverse outcome, recall of preceding events is greater than after a normal outcome. Such recall bias can be eliminated in prospective studies. Attention should be devoted to devising systems for prospective monitoring of fetal outcomes and of performing longer follow-up studies monitoring therapeutic exposure of pregnant women. The benefits of conducting clinical trials with AZT in pregnant women was demonstrated in the 1990's. There have also been limited clinical trials of pharmacological agents in halting preterm labor or treating pre-eclampsia.

Another potential problem during pregnancy is use of dietary supplements. Even though many of these substances are not currently tested for safety and efficacy, they are being taken in increasing amounts by women, including pregnant women, who believe they are safe when in fact their safety is not known.

Breast-Feeding and Maternal Drug Use. Although the benefits of breast-feeding on newborn and infant health and for the maternal-infant relationship are well established, it is important to understand the potential for drug-related infant morbidity, via breast-feeding, following maternal drug intake. Because pharmacologic agents are rarely tested for use in lactating women, there are many uncertainties regarding their safety for breast-fed infants. Research is also needed to ensure the safety of breast-feeding women and their nursing infants following drug administration during delivery and puerperium. Potential risks to such infants are increased following multidose, high-dose, or long-standing therapy, particularly in highly vulnerable premature neonates. For example, the limited data regarding the safety of agents with CNS activity raise concerns about toxicity and abnormal neurological development in the breast-feeding infant.51 Similarly, there are safety concerns for antithyroid drugs, such as the thioureylenes, which are known to be secreted in breast milk.52

Factors that facilitate the transfer into milk, as well as the PK and PD properties of the drug in the mother and infant, must be evaluated. Drug properties that promote low milk concentrations include large volume of distribution in the serum, high protein binding, low lipid solubility, ionization at physiologic pH, and large molecular weight. Following transfer into breast milk, drugs with low bioavailability and short elimination half-lives in neonates provide increased safety margins.53
For drugs that may pose a potential concern to the newborn, steps can be taken to minimize risk. These include such interventions as selection of alternative effective drug regimens with safer profiles in breast-feeding, timing of drug dosing to minimize accumulation in the breast milk, surveillance for newborn or infant symptomatology that may be a sign of toxicity, or even the determination of drug levels in the infant's circulation.54

Research Recommendations and Conclusions
Recommendations are grouped into the following areas: Components of Pharmacologic Effects, Preclinical Studies, and Drug Use in Pregnancy.

Components of Pharmacologic Effects
Recommendations. Examine the underlying mechanisms that contribute to gender differences in drug effects and disposition encompassing topics including, but not limited to, membrane transport pumps, ion channels, chronobiology, metabolic enzyme differences, sex hormone receptors, drug interactions between hormonal contraceptives and hormone replacement therapy, and drugs affecting chemical compounds such as the neurotransmitters.

Research Goals. Carry out basic and applied research on the following:

1. Pharmacokinetics, pharmacodynamics, and the importance of combining PK and PD data.

Develop decision-analysis tools to help determine when, and under what circumstances, PK and PD studies should be performed. Assess gender-related PK and PD differences in multiple populations including women with childbearing potential, pregnant women, lactating women, perimenopausal women, menopausal women, and senior women. The examples cited earlier in this report clearly demonstrate that gender-specific biology, reflected by such modulators as the sex hormones, can have a dramatic influence on the physiological response to drugs. Further research into the mechanism of these differences will be essential before it will be possible to predict in advance, or even identify, gender differences that result in increased drug toxicity or lack of responsiveness in either sex.

Adequate research support will be required when PD studies are warranted.
Study the effects of drugs in women across the life span in relation to drug absorption, distribution, metabolism, and elimination, with special focus on postmenopausal women (with and without hormone replacement therapy) and senior women who experience changes in organ function. With regard to women of childbearing years, the pharmacology working group strongly supported continued emphasis on including these women in all phases of clinical trials as part of drug development. To maximize safety and effectiveness of drugs used by these women, interaction of drugs with hormonal contraceptives must be studied routinely as should pharmacological agents for diseases that are common for this age group (e.g., autoimmune disorders, depression, HIV, migraine). Other specific topics for research include:

Examine drug interactions, an area that includes two components: drug-drug and drug-nutrient interactions; women are underrepresented in studies for both. Drug-nutrient (food, vitamin supplements) interactions have not been well studied. Examination of the broad spectrum of possible drug-drug and drug-nutrient interactions is warranted both for available pharmacologic agents and for those currently in clinical trials.

2. Pharmacogenetics

Research studies are needed that have sufficient statistical power to determine the nature of the interaction between gender and genetic polymorphism. CYP 450 isoenzymes, such as CYP 3A4 and CYP 1A2, which are responsible for estrogen metabolism, may also differentially affect other substrates. Isoenzyme 1A2 is thought to influence the production of 16a-OH and 4-OH, whereas 3A4 is predominantly responsible for the less active 2-OH estrogen metabolite (55). There may be other examples of drug-metabolizing enzymes currently unknown that may be differentially expressed in women as a whole or in specific racial groups.

Drug action and disposition may be altered by endogenous and exogenous sex hormones. Estrogen has been shown to inhibit the metabolism of CYP 1A2 model compounds, but its effect on the action and disposition of other therapeutic agents is not well studied. The role of CYP 1A2 and 3A4 in gender differences of drug metabolism, as well as the interface between these isoenzymes and estrogen metabolism, induction, or inhibition merits further investigation with respect to oral contraceptive failure and cancer induction.

The importance of the cytochrome P450 isoenzymes has long been established. CYP 3A4 is quantitatively the most important isoenzyme, making up at least 60 percent of the total P450 hepatic content and is responsible for metabolizing the majority of medications (56). P450 metabolism is affected by steroid, dietary, and medication inhibition and induction (57). There is evidence that CYP 3A4 activity is greater in younger women compared with men and postmenopausal women (58). Demonstrated decreased metabolism in postmenopausal women (59, 60) may make them more liable to adverse effects of certain medications. Careful study of this possibility is needed.

CYP 1A2 is also thought to have gender-related differences in expression. Higher plasma levels of 1A2 drug substrates have been reported in women (61, 62). In addition, the activity of this isoenzyme has been found to be reduced in women receiving contraceptive medications and during pregnancy (63). The potential inhibition of CYP 1A2 by estrogen is demonstrated by decreased caffeine metabolism in women receiving estrogen replacement therapy (64).

3. Membrane transporters

Multidrug resistance may be caused by overexpression of either p-glycoprotein or the multidrug-resistance protein (MRP); such resistance is characterized by a decreased cellular drug accumulation due to an enhanced drug efflux. There may be gender-specific expression of p-glycoprotein and MRP related drug efflux pumps and other membrane transporters. Such gender-related differences have important value in the clinical setting, where decisions must be made regarding selection of drug type and dosage (e.g., in cancer chemotherapy).

4. Ion Channels

Differences in sodium and potassium channel sensitivities between men and women may explain differential adverse effects observed following intake of certain medications. Further pre-clinical and clinical research is needed to better define these differences.

5. Modulators

Sex Hormones as Possible Causes of Receptor Sensitivity. Differences in receptor sensitivity need to be studied. For example, women are known to have a longer QT interval, which may differentially predispose them to cardiac arrhythmias. In addition, women have been shown to have lower sensitivity to the beta-blocker, propranolol, which would not have been predicted by PK studies alone; therefore, this lower sensitivity is a PD issue.

6. Chronopharmacology

When examining treatment of diseases specific to women, the gender differences related to biologic rhythms and their application to efficacy of drugs needs to be studied. The chronopharmacology of disease needs to be studied because biologic rhythms may change during the disease process and affect the dosing requirements for a specific drug.

Preclinical Studies
Recommendation. Focus special attention on the continued development of animal and other models to predict with some accuracy the reproductive and developmental effects of drugs in humans. With these models in hand, progress may be made in understanding safety and efficacy profiles of drugs and predicting PK and PD with greater accuracy.

Research Goals

1. Evaluate therapeutic efficacy and potential side effects through whole animal studies. Thus, the effects of estrogens on the cardiovascular system must include information on the heart itself as well as the peripheral vasculature.

2. Place receptor and cell-signaling research in the context of the whole organism, including physiological and therapeutic considerations in animal research.

Drug Use in Pregnancy
Recommendation. Develop improved methods for collecting and analyzing data on the use of pharmacologic agents in pregnancy with respect to safety and efficacy for the mother and safety for the infant.

Research Goals

1. Women have a great potential for exposure to medications not only for conditions occurring during pregnancy but also inadvertent exposure prior to knowledge of the pregnancy. Since most medications have little or no information on risk for the developing human fetus, we must begin to develop scientifically valid and rigorous methods of evaluating potential risks in humans including frequency of structural and functional birth defects, and reproductive effects such as rates of spontaneous abortion. Some anomalies may not manifest themselves until later in neonatal development; thus it is necessary to conduct followup of birth outcomes at different times during early child development.

2. Critical research should be performed to determine the best systems to link data from a mother to her newborn. This should be followed by implementation of such systems for Medicaid data, at HMOs, and at academic centers.

3. For those drugs commonly used by pregnant women, clinical trials in pregnant women, particularly in the second and third trimester, should be performed to address pharmacokinetics and dosing recommendations. All trials of drugs in pregnant women should collect maternal safety data for the purpose of comparison with the safety profile of the drug in nonpregnant women.

4. Most reported birth anomalies are major and visible; research is needed on anomalies that are more difficult to detect, such as those that are neurodevelopmental or immunologic. Because of the large number of pharmacologic agents that may used and the potential for adverse effects, this could be a very large project. Thus, it is important to establish priorities based on the therapeutics taken most often by women. For example, what effect will anti-depressants, especially the selective serotonin reuptake inhibitors taken by pregnant women have on the cognitive function and development of the child? Because currently used classic animal teratology models do not examine immunologic markers, research, both animal and human, is needed to determine appropriate endpoints for assessment of immunologic function.

5. Research is needed on the effects of maternal medications on the breast-feeding infant. Drug concentration measurements in breast milk should be considered for inclusion in studies of the PK and PD of new drugs. Based on an understanding of mechanisms and principles of drug excretion into milk, methods should be developed to predict infant exposure levels.

Summary of Recommendations

Recommendations for Research

1. Examine underlying mechanisms that contribute to gender differences in drug action and disposition. Define the pharmacokinetics, pharmacogenetics, chronobiology, modulators, biologic/molecular factors, and pharmacodynamics of pharmacologic agents across the life span (prenatal, infants, children, adolescents, women of childbearing potential, pregnant women, lactating women, perimenopausal women, postmenopausal women, seniors) taking into consideration various racial and ethnic groups. Specifically, assess:

In addition, examine drug-drug interactions and drug-nutrient effects.

2. Perform preclinical research in the development and validation of in vitro and whole animal models to test and predict gender specific pharmacologic differences:

3. Incorporate research on potential adverse effects on the pregnant woman, the fetus, and the newborn in evaluating PK and PD of new drugs. Develop methods of assessing subtle effects, such as possible neurodevelopmental and immunologic impairment in the child.

4. Include pregnant women in clinical trials.

Recommendations for Promoting Implementation. Interagency and public-private collaboration is required in order to carry out the full range of research recommended for studying pharmacologic issues in women, including pregnant women. It is essential that NIH prioritize studies needed to fill the serious data gaps described in this report. NIH should increase funding allocation for research on this topic and encourage interdisciplinary collaboration across all health professions and the basic sciences, and development of longitudinal studies to cover both basic and translational research. Finally, attention to studying mechanisms of pharmacologic effects should become an integral component of the review process across Institutes and Offices whenever a pharmacologic agent is part of a study protocol for prevention or treatment of a disease or condition.

Thanks are due to all Workgroup members and especially to Gail Anderson, Marietta Anthony, Marianne Rollings, Rosalie Sagraves, Ana Szarfman, Terry Toiga, Diane Thompson, Catherine White, Ray Woosley, Sheila Weiss,and Cheryl Zimmerman.