Pharmacokinetic studies in pregnant women

ETHICS nature publishing group Pharmacokinetic Studies in Pregnant Women GJ Anger1 and M Piquette-Miller1 Prescription and over-the-counter drug use during pregnancy is necessary for many women today. A study of US and Canadian women found that, on average, 2.3 drugs were used during pregnancy; however, 28% reported using more than 4.1 For some women, this is because they become pregnant with preexisting conditions that require ongoing or intermittent pharmacotherapy. For others, this is because pregnancy itself can give rise to new medical conditions such as gestational diabetes and preeclampsia. The principal concern of prescribing physicians is whether or not agents will harm the fetus (i.e., have teratogenic effects). This concern rose to prominence primarily as a result of the thalidomide disaster. Marketed for use in morning sickness, thalidomide was found to be a potent teratogen capable of producing a variety of birth defects relating to development.2 Consequently, determining the teratogenicity of new drugs currently dominates the objectives of pregnancy-relevant experiments conducted throughout drug development. This often comes at the expense of valuable pharmacokinetic (PK) studies, which are seldom performed pre-market. Sex differences in PK parameters have been demonstrated in animals and humans since the 1930s.3,4 It is, therefore, not surprising that differences also arise in pregnancy. A wide array of physiological and hormonal changes occur during pregnancy; most begin early in the first trimester and increase linearly until parturition.5 Physicians lacking adequate PK information typically prescribe the standard adult dose in pregnancy, and this can be either inadequate or excessive depending on a variety of factors. The purpose of this report is to highlight this issue and illustrate how current methods used to obtain PK data in pregnancy are insufficient. The steps that are being taken to address this issue will also be discussed. PHARMACOKINETIC CONSIDERATIONS Absorption An increased level of progesterone reduces gastric emptying and intestinal motility with corresponding alterations to bioavailability parameters. These effects are most pronounced in the third trimester when progesterone levels are at their peak.6 Although there is less of a concern with repeated dosing regimens, there is reason to believe that progesterone level would affect the efficacy of oral drugs taken as a single dose, such as anti-emetic and analgesic drugs like paracetamol (Table 1).5 Distribution There is an increase in both total body weight and fat distribution during pregnancy. Much of the increase in weight is due to increased water (6–8 l) in intravascular (i.e., plasma) and extravascular (i.e., tissues such as the breasts and uterus) compartments.5,7 Increased water within the body creates a larger volume of distribution (Vd) for drugs that are hydrophilic. In addition, an increase in body fat percentage means that there is also a larger Vd for lipophilic drugs. For some drugs, a larger Vd could necessitate a higher initial and maintenance dose to obtain therapeutic plasma concentrations. Conversely, pregnant women exhibit decreased plasma protein binding because of decreased levels of the protein albumin, due in part to hemodilution.5,8 This increases the concentration of free drug in plasma for highly bound drugs and thereby necessitates dosage adjustment. Metabolism Drug metabolism, via phase I oxidative drug metabolism and phase II conjugation, is also altered in pregnancy. The CYP450 enzyme CYP3A4 represents a major route of drug metabolism for many drugs used in pregnant women, including nifedipine, carbamazepine, midazolam, and antiretroviral drugs.9,10 In the case of anti-retroviral drugs, subtherapeutic plasma concentrations during pregnancy not only have a negative impact on the mother’s health (i.e., increased viral load and resistance formation) but also increase the chance of vertical HIV (human immunodeficiency virus) transmission.11 The activity and amount of CYP3A4 are increased in pregnancy and corresponding increases have been demonstrated in the clearance of its 1 Department of Pharmaceutical Sciences, Faculty of Pharmacy, University of Toronto, Toronto, Ontario, Canada. Correspondence: M Piquette-Miller ([email protected]) Published online 19 September 2007. doi:10.1038/sj.clpt.6100377 184 VOLUME 83 NUMBER 1 | JANUARY 2008 | www.nature.com/cpt ETHICS Table 1 Clinical examples of selected drugs that demonstrate altered plasma drug concentrations during pregnancy Drug Altered physiological parameter Plasma concentration References Caffeine k CYP1A2 and NAT2 activity in liver m 16 Cefatrizine m Volume of distribution k 26 Lamotrigine m UGT1A4 activity in liver k 13, 14 Morphine m UGT2B7-mediated glucuronidation k 27, 28 Nelfinavir m CYP3A4 activity in liver k 29 Oxazepam m UGT2B7-mediated glucuronidation k 27, 28 Paracetamol k Gastric emptying and intestinal motility k 30 substrates.12 The levels of CYP2D6 also increase in pregnancy with corresponding increases in the metabolism of substrates, including dextromethorphan, fluoxetine, and nortriptyline.9 With regard to phase II, the activity of important conjugating enzymes such as UGT1A4 is altered during pregnancy. Oral clearance of the UGT1A4 substrate lamotrigine during pregnancy, particularly when taken as monotherapy, is significantly higher than pre-pregnancy levels.13,14 The metabolic capacity of phase I and II enzymes in pregnancy is further enhanced by a significant increase in hepatic blood flow, which serves to increase the amount of drug available to the liver for metabolism.15 It is important to note that not all alterations to drug metabolism in pregnancy result in increased metabolism. For example, caffeine plasma levels increase across the duration of gestation (Table 1).16 Renal elimination During pregnancy, there is a 50–80% increase in renal blood flow, which results in a corresponding 40–65% increase in glomerular filtration.17 This increase in renal clearance can have notable effects on drugs that are eliminated by the kidneys. Increases in elimination rates that range from 20 to 60% have been reported for ampicillin, cefuroxime, cephradine, cefazolin, atenolol, digoxin, lithium, and many others.9 ORIGINS OF THE KNOWLEDGE GAP Drugs are often approved by regulatory agencies on the basis of clinical trials that are devoid of pregnant participants. In the United States, standard reproductive toxicology studies are performed in animals and are used to assign a category in a labeling subsection concerned with birth defects and other effects on reproduction and pregnancy. Instituted by the US Food and Drug Administration (FDA) in 1979, the information provided by category assignment is generally considered vague and hard to apply.18 On the basis of demonstrated safety and efficacy in the general public and reproductive toxicity studies, physicians typically prescribe these drugs to pregnant patients based on their own relative assessment of risks versus benefits. To err on the side of caution, physicians tend not to deviate from the standard adult dose even when efficacy is questioned. A major difficulty in establishing PK information in pregnancy is due to a lack of well-controlled clinical trials.19 CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 83 NUMBER 1 | JANUARY 2008 Post-market data collection generally requires a longer period of time to generate information that can be used to make meaningful recommendations, as drugs are carefully tested during pregnancy only after first being tested in the general public. Often, safety information is acquired from clinical reports of atypical drug actions in pregnant patients (e.g., poor efficacy, adverse drug reactions (ADRs), and so on). In most cases, collaborative teams of clinicians and academic scientists will then work to explain the phenomenon in a process that generates invaluable PK data. This can, however, take too long to generate the kind of information that will aid in dosing recommendations. Moreover, the ethics of this process are questionable, as pregnant women are theoretically required to undergo an atypical drug experience before clinicians are alerted. Furthermore, this assumes that clinicians will then take the initiative to either inform the scientific community or strike the collaborations required to examine the drug further. The most commonly cited reasons for drug developers adopting the above approach are linked to ethical and legal considerations. Clinical research in Europe and North America has not always been monitored as closely as it is today, and examples of unethical conduct are all too abundant. One example is the infamous Tuskegee Syphilis Study conducted by the US Public Health Service from 1932 to 1972. In this study, the effects of tertiary syphilis were monitored in a group of African Americans who were not informed of the purpose of the study and were not encouraged to take penicillin once it was proved to be an effective treatment in 1945. Regulations such as those established by the World Medical Association’s Declaration of Helsinki Principles in 1964 were born of unethical practices such as these and included guidelines for clinical research in vulnerable populations such as children, mentally disabled persons, prisoners, and pregnant women. With pregnant women, however, the vulnerable entity was the unborn fetus, and drug developers and regulatory agencies responded by excluding not only pregnant women but also all women of childbearing age. This response was codified in 1977 when the FDA formally restricted this group from participation in phase I and II clinical trials.20 One argument against exclusion is that by excluding pregnant mothers from clinical trials, an information gap is 185 ETHICS created that makes prescribing medications much more dangerous for both the mother and the vulnerable fetus. For example, failure to properly manage seizures in epilepsy not only results in harm to the mother but also increases the risk of miscarriage. A review of studies published on anticonvulsant use in pregnancy found that 30–50% of epileptic women on anticonvulsants experience an increase in seizure frequency while pregnant.21 Indeed, alterations to PK have been linked to the decreased efficacy of carbamazepine and phenytoin during pregnancy.22 Evidence-based dosage adjustments are now commonly implemented with anticonvulsants. The fear of litigation is another factor in the pharmaceutical industry’s reluctance to conduct well-controlled clinical trials in pregnant women for the purposes of obtaining PK data. Given the preponderance of lawsuits that follow most drug-related injuries, this fear may be warranted. From a drug developer’s point of view, introducing a drug into pregnant women is risky and is often avoided when possible. What should be noted, however, is that pregnant women who experience post-market injuries can still sue for damages. Consequently, it stands to reason that performing studies in the context of a clinical trial and in a relatively small group of pregnant women is less risky than leaving it in the hands of physicians. A fear of litigation should not be the sole basis for excluding pregnant women in clinical trials if the appropriate reproductive toxicology studies have been performed, and there is no reason to suspect teratogenicity. CURRENT INITIATIVES AND FUTURE DIRECTIONS Steps have been taken in recent years by various regulatory agencies and the scientific community to address the issue of inadequate drug information in special populations. In 1993, the FDA acknowledged the need to obtain detailed information for drugs that could be taken by pregnant women. This was done first by removing the 1977 restrictions as well as the publication of specific guidelines.23 Shortly thereafter, the FDA established the Office of Women’s Health and has continued to formalize its new stance by way of inclusions in the Modernization Act (FDAMA) of 1997 and several additions to titles 21 and 45 of the US Code of Federal Regulations beginning in the late 1990s. Several guidance documents have been released on topics such as exposure registries and, of relevance to this report, PK studies in pregnancy. However, the problem with many of these initiatives, with respect to pregnant women, is that they make inclusion in clinical trials compulsory only for drug candidates who are specifically targeted at this demographic or for whom use in pregnancy will likely be prevalent. The vast majority of the FDA guidance on the issue comes in the form of recommendations. As delaying the release of a blockbuster drug can translate into lost sales on the order of millions of dollars per day, studies to refine dosage recommendations in pregnancy are unlikely to be performed frequently in pre-marketing if they are not made compulsory or incentives provided. One possible incentive could be the 186 extension of patent privileges in exchange for detailed PK data in pregnancy, as is done already for certain drugs in pediatric populations via the FDAMA’s Pediatric Exclusivity Provision. Steps to address this issue would ideally occur in premarketing, but post-marketing trials represent an option that could attract greater participation from developers while still increasing the speed at which information becomes available. This option is noted in the FDA guidance document, ‘‘Pharmacokinetics in Pregnancy—Study Design, Data Analysis, and Impact on Dosing and Labeling.’’ Although this document focuses on the inclusion of pregnant women in phase III, it anticipates that the majority of PK studies in pregnant women will likely occur in the post-marketing period with pregnant women who have already been prescribed the drug.24 Drugs that lack PK data in pre-marketing could be flagged for immediate investigation in the first wave of pregnant patients to consume it. As opposed to the current standard of monitoring adverse drug reactions through registries, developers could initiate PK studies in this population and then compare data to postpartum measurements. One final step that is being taken to improve our understanding of PK in pregnancy involves a more proactive approach in which known alterations to physiology that occur during pregnancy are considered throughout the drug development process. A drug developer’s decision to include or exclude pregnant women in trials is then based on the properties of the drug examined. For example, as stated, drugs that are substrates for certain CYP450s are metabolized faster during pregnancy. The inclusion of pregnant women in clinical trials for drugs that are substrates of significantly altered CYP450 enzymes should therefore be considered. This same approach could be used to predict the likelihood of teratogenic effects such as fetal accumulation.25 Two potential hurdles to this approach are the bioinformatics expertise that is required and the fact that its predictive validity has not yet been demonstrated. SUMMARY The efficacy of pharmacotherapy during pregnancy is altered in many cases because of changes in physiology that can significantly affect PK. Steps currently being taken to address this issue were discussed. Common themes involve a need to use incentives and binding guidelines and a need to promote more phase IV trials in pregnancy when data are not available pre-market. Known modifications to PK could also be considered when deciding whether PK studies in pregnancy are warranted. With drugs for which PK data are sparse, evidence-based dosing decisions of benefit to both mother and fetus are difficult, and options in the pharmacopoeia are effectively limited because of this. Future efforts must encourage the production of PK information for as wide an array of drugs as possible if women are to benefit from the same therapeutic effects when pregnant as they do when not. VOLUME 83 NUMBER 1 | JANUARY 2008 | www.nature.com/cpt ETHICS ACKNOWLEDGMENTS GJA is supported by a studentship from the Ontario HIV Treatment Network. MP-M is supported by an operating grant from the Canadian Institutes of Health Research. 15. CONFLICT OF INTEREST The authors declared no conflict of interest. 17. 16. 18. & 2008 American Society for Clinical Pharmacology and Therapeutics 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Mitchell, A.A., Hernandez-Diaz, S., Louik, C. & Werler, M.M. Medication use in pregnancy: 1976–2000. Pharmacoepidemiol. Drug Saf. 10, S146 (2001). McBride, W. Thalidomide and congenital abnormalities. Lancet 278, 1358 (1961). Czerniak, R. Gender-based differences in pharmacokinetics in laboratory animals. Int. J. Toxicol. 20, 161–163 (2001). Curry, B. Animal models used in identifying gender-related differences. Int. J. Toxicol. 20, 153–160 (2001). Dawes, M. & Chowienczyk, P.J. Pharmacokinetics in pregnancy. Best Pract. Res. Clin. Obstet. Gynaecol. 15, 819–826 (2001). Dawood, M.Y. Circulating maternal serum progesterone in high-risk pregnancies. Am. J. Obstet. Gynecol. 125, 832–840 (1976). Reynolds, F. Pharmacokinetics. In Clinical Physiology in Obstetrics 3rd edn. (eds. Chamberlain, G. & Pipkin, F.B.) 239–260 (Blackwell Sciences, London, 1998). Vickers, M. & Brackley, K. Drugs in pregnancy. Curr. Obstet. Gynaecol. 12, 131–137 (2002). Mattison, D. & Zajicek, A. Gaps in knowledge in treating pregnant women. Gend. Med. 3, 169–182 (2006). Schwartz, J.B. The influence of sex on pharmacokinetics. Clin. Pharmacokinet. 42, 107–121 (2003). de Maat, M.M. et al. Subtherapeutic antiretroviral plasma concentrations in routine clinical outpatient HIV care. Ther. Drug Monit. 25, 367–373 (2003). Little, B.B. Pharmacokinetics during pregnancy: evidence-based maternal dose formulation. Obstet. Gynecol. 93, 858–868 (1999). de Haan, G.J. et al. Gestation-induced changes in lamotrigine pharmacokinetics: a monotherapy study. Neurology 63, 571–573 (2004). Pennell, P.B. Antiepileptic drug pharmacokinetics during pregnancy and lactation. Neurology 61, S35–S42 (2003). CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 83 NUMBER 1 | JANUARY 2008 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Nakai, A., Sekiya, I., Oya, A., Koshino, T. & Araki, T. Assessment of the hepatic arterial and portal venous blood flows during pregnancy with Doppler ultrasonography. Arch. Gynecol. Obstet. 266, 25–29 (2002). Tsutsumi, K. et al. The effect of pregnancy on cytochrome P4501A2, xanthine oxidase, and N-acetyltransferase activities in humans. Clin. Pharmacol. Ther. 70, 121–125 (2001). Conrad, K.P. Mechanisms of renal vasodilation and hyperfiltration during pregnancy. J. Soc. Gynecol. Investig. 11, 438–448 (2004). Frederiksen, M.C. The drug development process and the pregnant woman. J. Midwifery Womens Health 47, 422–425 (2002). Faden, R. Ethics of studying pregnant women /http://www.fda.gov/ cder/present/clinpharm2000/1205preg.txtS (2000). FDA. General considerations for the clinical evaluation of drugs ohttp://www.fda.gov/CDER/GUIDANCE/old034fn.pdf4 (Health and Human Services, 1977). Sawle, G. Epilepsy and anticonvulsant drugs. In Prescribing in Pregnancy 3rd edn. (ed. Rubin, P.) 112–126 (BMJ Books, London, 2000). Yerby, M.S. et al. Pharmacokinetics of anticonvulsants in pregnancy: alterations in plasma protein binding. Epilepsy Res. 5, 223–228 (1990). FDA study and evaluation of gender differences in the clinical evaluation of drugs ohttp://www.fda.gov/CDER/GUIDANCE/ old036fn.pdf4 (Health and Human Services, 1993). FDA pharmacokinetics in pregnancy—study design, data analysis, and impact on dosing and labeling. ohttp://www.fda.gov/CDER/ GUIDANCE/5917dft.pdf4 (Health and Human Services, 2004). Gedeon, C. & Koren, G. Designing pregnancy centered medications: drugs which do not cross the human placenta. Placenta 27, 861–868 (2006). Papantoniou, N. et al. Pharmacokinetics of oral cefatrizine in pregnant and non-pregnant women with reference to fetal distribution. Fetal Diagn. Ther. 22, 100–106 (2007). Tomson, G., Lunell, N.O., Sundwall, A. & Rane, A. Placental passage of oxazepam and its metabolism in mother and newborn. Clin. Pharmacol. Ther. 25, 74–81 (1979). Gerdin, E., Salmonson, T., Lindberg, B. & Rane, A. Maternal kinetics of morphine during labour. J. Perinat. Med. 18, 479–487 (1990). Villani, P. et al. Pharmacokinetics of nelfinavir in HIV-1-infected pregnant and nonpregnant women. Br. J. Clin. Pharmacol. 62, 309–315 (2006). Simpson, K.H., Stakes, A.F. & Miller, M. Pregnancy delays paracetamol absorption and gastric emptying in patients undergoing surgery. Br. J. Anaesth. 60, 24–27 (1988). 187