In 2007, the Food and Drug Administration approved 16 new molecular entities and several new biologics. None of these agents have human pregnancy experience, but some will be prescribed to women of reproductive age and exposure in early gestation is inevitable.
There also are situations when a woman's condition requires drug therapy, regardless of pregnancy. New antineoplastics are indicated when other therapies have failed to fit into this category, such as ixabepilone (Ixempra) and lapatinib (Tykerb), for breast cancer; nilotinib (Tasigna), for leukemia; and temsirolimus (Torisel), for advanced renal cancer.
Regardless of the circumstances, clinicians caring for women of reproductive age will be faced with the dilemma of how to counsel patients when there are few or no human pregnancy data. One method, using some of the drugs approved in 2007 as examples, is described here. When an exposure has already occurred, or when the maternal benefit for starting therapy clearly exceeds the fetal risk and there are no other alternatives, the estimation of fetal risk can be based on four questions:
▸ Is there human pregnancy experience for the drug?
▸ Is there human pregnancy experience with other drugs in the same class or with similar mechanisms of action?
▸ Does the drug cross the human placenta?
▸ Does the drug cause developmental toxicity in animals at doses less than or equal to 10 times the human dose?
Timing of the exposure is critical, and must be included in any estimation. Although organogenesis (5–10 weeks) is usually the most vulnerable period and exposure at that time should be avoided if possible, drugs can cause developmental toxicity throughout gestation.
For the new drugs described, the answer to the first question is no.
There are several examples that fit the second question. Aliskiren (Tekturna) is an antihypertensive that acts as a renin inhibitor. Two other classes of drugs acting on the renin-angiotensin system, angiotensin-converting-enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs), are known to cause marked fetal toxicity in the second and third trimesters; a similar effect might occur with aliskiren. Nebivolol (Bystolic) is a β-blocker used for hypertension. It has no intrinsic sympathomimetic activity (ISA), so, as with other β-blockers without ISA, its effects may include restricted placental and fetal weights when used during the second half of gestation.
Levocetirizine (Xyzal), the L-isomer of cetirizine (Zyrtec), is an antihistamine. There is no evidence of fetal harm from antihistamines, so the risk from exposure to levocetirizine is probably low. On the other hand, lisdexamfetamine (Vyvanse), which is indicated for attention-deficit/hyperactivity disorder, is a prodrug of dextroamphetamine, a drug known to cause human developmental toxicity.
Package inserts typically provide at least four important factors that help answer the third question: maternal concentration, elimination half-life, plasma protein binding, and molecular weight. For example, a small percentage of nonpregnant adults had measurable plasma concentrations of retapamulin (Altabax), a topical antibiotic, but the levels were very low (less than 1 ng/mL). The elimination half-life is unknown, but the medium molecular weight (518) suggests that the drug will cross the placenta. However, the amounts available for transfer appear to be clinically insignificant.
Although there are no definite methods to interpret animal studies, nearly all drugs known to cause human developmental toxicity also cause such toxicity in at least one animal species
The dose that causes toxicity in animals is critical, as is its relationship to the maximum recommend human dose (MRHD). Guidelines released by the Environmental Protection Agency in 1991 stated that if a drug did not cause developmental toxicity at doses less than or equal to 10 times the human dose based on body surface area (BSA) or systemic exposure (AUC), then the drug could be considered low risk for human fetal toxicity. Conversely, if a drug did cause toxicity at doses less than or equal to 10 times the human dose (in the absence of maternal toxicity), then it could be classified as having risk, but the risk magnitude would be unknown. These conclusions were similar to those reached by a panel convened in 2004 (Birth Defects Res. 2004;70[Part A]:7-12).
The strength of any risk estimation increases if two or more of the responses concur. For example, animal studies with retapamulin found no fetal toxicity after high systemic doses. These results, combined with the low systemic levels, reinforce the estimation that this is a low-risk drug. Conversely, animal studies with aliskiren observed fetal growth restriction. Thus, the estimation that this drug may cause fetal growth restriction is strengthened.