According to recent studies, there is insufficient information available to assess the human teratogenic risk for the majority of drugs introduced in the United States during the last 20 years.1,2 Although all drugs must undergo animal teratology testing before they are approved by the U.S. Food and Drug Administration (FDA), animal studies are not necessarily predictive of the teratogenic potential in humans.1 Furthermore, clinical trials in humans typically exclude pregnant women, and postmarketing safety surveillance systems often are ineffective in providing data on the potential risk of birth defects associated with drug use during pregnancy.1 Consequently, many drugs have few or no data available on the safety of use during pregnancy.
The human immunodeficiency virus (HIV) protease inhibitor Viracept (nelfinavir mesylate) was introduced in the United States in 1997 by Agouron Pharmaceuticals, Inc (La Jolla, CA), a Pfizer Company. Nelfinavir mesylate is an inhibitor of the HIV protease, which prevents cleavage of the gag-pol polyprotein, resulting in the production of immature, noninfectious virus. Since its introduction, nelfinavir has been a part of the Antiretroviral Pregnancy Registry, which is an active prospective registry of exposures to antiretroviral therapies designed to provide useful information to health care providers on the outcomes of pregnancy after exposure to antiretroviral therapy. The specific focus of the Antiretroviral Pregnancy Registry is to detect any major teratogenic effect associated with the antiretroviral therapies included in the registry.3
Information on the teratogenic risk of antiretroviral medications is important because antiretroviral therapy is recommended for all HIV-infected pregnant women to reduce maternal HIV ribonucleic acid levels and decrease the risk of mother-to-child transmission.4 Highly active antiretroviral therapy has been shown to reduce transmission to less than 2%.5 Protease inhibitors are an important component of highly active antiretroviral therapy and are being used increasingly during pregnancy. Although there are insufficient data to recommend a specific protease inhibitor during pregnancy, the Centers for Disease Control and Prevention (CDC) recommends that nelfinavir can be used as part of a highly active antiretroviral regimen during pregnancy.6 Furthermore, nelfinavir is the most commonly reported protease inhibitor in pregnancy.3
Nelfinavir is assigned FDA Pregnancy Category B status, which indicates that “animal reproduction studies have failed to demonstrate a risk to the fetus, and there are no adequate and well-controlled studies in pregnant women.”7 According to the prescribing information, nelfinavir was not shown to be associated with mutations in either genes or chromosomes in a battery of in vitro and in vivo tests, including microbial mutagenesis (Ames), mouse lymphoma, chromosome aberrations in human lymphocytes, and an in vivo rat micronucleus assay. Nelfinavir produced no effects on either male or female mating and fertility or embryo survival in rat studies at exposures comparable to human therapeutic exposure. There also were no effects on fetal development or maternal toxicity when nelfinavir was administered to pregnant rats at systemic exposures comparable with human exposure. Additional studies in rats indicated that exposure to nelfinavir from midpregnancy through lactation had no effect on the survival, growth, and development of the offspring to weaning. Subsequent reproductive performance of these offspring was also not affected by maternal exposure to nelfinavir. Nelfinavir produced no fetal development effects in rabbits up to a dose at which a slight decrease in maternal body weight was observed; however, even at the highest dose evaluated, systemic exposure in rabbits was significantly lower than human exposure.8
As part of its ongoing commitment to examining the safety of nelfinavir use in human pregnancy, Agouron Pharmaceuticals joined the Antiretroviral Pregnancy Registry in 1997. The primary objective of this work is to examine the data from the Antiretroviral Pregnancy Registry on the human teratogenic risk of nelfinavir mesylate.
MATERIALS AND METHODS
This study used a subset of the data from the Antiretroviral Pregnancy Registry, which was designed to monitor prenatal exposures to antiretroviral therapy and to detect an increase in the risk of major birth defects.3 The registry is international and began in 1989. It is managed by Inveresk under the sponsorship of Abbott Laboratories; Agouron Pharmaceuticals, Inc; Bristol-Myers Squibb Company; Boehringer Ingelheim Pharmaceuticals, Inc; Gilead Sciences, Inc; GlaxoSmithKline; F. Hoffmann-La Roche Ltd; and Merck & Co, Inc. The scientific conduct and analysis of the registry are overseen by an independent advisory committee consisting of members from the CDC, FDA, the National Institutes of Health, as well as the private sector. Members include specialists in maternal and fetal medicine, teratology, infectious disease, epidemiology, and biostatistics.
The Antiretroviral Pregnancy Registry uses a prospective exposure-registration cohort design that includes pregnant women with prenatal exposures to any antiretroviral therapy. Worldwide health care providers voluntarily register pregnant women by providing detailed information on antiretroviral exposure in pregnancy, demographic characteristics, severity of disease indicators, and prenatal history, including prenatal testing. This information is provided early in pregnancy before the outcome is known. Shortly after delivery, the health care providers submit data on additional antiretroviral exposures in pregnancy and details of the pregnancy outcome. If a birth defect is reported at outcome or as a result of prenatal testing, the health care provider submits detailed information on the birth defect, concomitant drug exposures during pregnancy, coexisting maternal conditions, and/or family history. The registry has Western Institutional Review Board approval and was granted a waiver of informed consent based on its process for protecting patient anonymity. See the Antiretroviral Pregnancy Registry Interim Report for more details.3
For this study, all records of pregnant women exposed to nelfinavir, used alone or in combination with other antiretroviral drugs, were extracted from the registry database and analyzed. The study period includes all cases closed from January 1997, when nelfinavir joined the registry, through July 31, 2002.
The prevalence of birth defects in the study group is compared with the prevalence derived from the CDC's Metropolitan Atlanta Congenital Defects Program. The CDC program is a population-based birth defects surveillance system that includes all live and stillbirths to mothers in the metropolitan area of Atlanta, Georgia.9 There are approximately 40,000 births in this 5-county area each year; approximately 40% are African American.9 The program actively searches for birth defects by abstracting medical records at all obstetric and pediatric referral hospitals in the area.9 The CDC surveillance system does not capture specific maternal drug exposure data.10 Therefore, infants exposed prenatally to nelfinavir and the other antiretroviral drugs in the registry may be included in the CDC data.
Key variables from the registry data include pregnancy outcome, gestational age, infant birth weight, birth defects, and earliest trimester of exposure to nelfinavir. Pregnancy outcomes are classified into 1 of the following mutually exclusive categories: spontaneous pregnancy loss, induced abortion, stillbirth, and live birth. A stillbirth is defined as a fetus born dead after 20 weeks of gestation or weighing more than 500 g. However, if gestational age and birth weight are not available for pregnancy losses, the registry accepts the health care provider's classification of spontaneous pregnancy loss or stillbirth. Birth weight is reported in grams. Low birth weight is defined as less than 2,500 g and very low birth weight as less than 1,500 g. Gestational weeks are determined by the first day of the last menstrual period and/or the estimated date of delivery based on ultrasonograms before 20 weeks of gestation. If there is any discrepancy between these 2 estimates, the ultrasound-obtained estimate is used. Premature delivery is defined as delivery occurring before 37 weeks of gestation and very premature delivery as before 32 weeks of gestation. Because of the strong association between prematurity and both multiple gestation and congenital anomalies,11 analyses of prematurity and low birth weight are restricted to singleton live births without reported birth defects. For the purpose of this study, earliest trimester of exposure to nelfinavir is defined as the first trimester in which the nelfinavir exposure occurred. Once the gestational weeks are calculated, trimesters are determined with the second trimester beginning at week 14 and the third trimester beginning at week 28. Because most structural defects have their origins in the first trimester of pregnancy, exposures occurring in the second and third trimesters are grouped in the analysis.12
A birth defect is defined as any major structural or chromosomal abnormality or any cluster of 2 or more minor abnormalities occurring in infants or fetuses of at least 20 weeks of gestational age.3 All birth defects reported to the registry are reviewed and classified by a physician specializing in genetics and dysmorphology and trained by the CDC to use the Metropolitan Atlanta Congenital Defects Program system.9 All birth defects are further reviewed by an independent scientific advisory committee comprising experts in maternal–fetal medicine, pediatrics, teratology, infectious diseases, epidemiology, and biostatistics from private practice, academia, government, and the pharmaceutical industry. The registry's definition of a birth defect is consistent with, but not restricted to, the CDC system. To be consistent with the CDC, the registry includes abnormalities reported in fetal losses that were 20 weeks or more of gestation. To increase the sensitivity of the registry, clusters of 2 or more minor abnormalities are included. However, the CDC system only includes minor defects in the presence of a major defect.9 To increase the possibility of identifying a potential signal, the registry has adopted an organ classification system, which classifies birth defects by affected organ and organ system.13 The registry also assesses the temporal association between the exposure to antiretroviral therapy and the stage of fetal development during which the defect is apt to occur.3
The prevalence of birth defects is calculated by dividing the number of infants with birth defects by the total number of live births. Pregnancy losses with or without defects are excluded from the denominator to be consistent with the calculation used by the CDC.9 This rate was compared with the rate reported by the CDC's population-based surveillance system. Additionally, first-trimester exposures (the period in which organogenesis occurs) are compared with the combined second/third-trimester exposures. The Fisher exact test was used to test the differences between the groups. SAS statistical software (SAS, Cary, NC) was used for all statistical analyses.
For all defects combined, a cohort of 300 newborns exposed in the first trimester is sufficient to detect a 2-fold increase in risk of birth defects compared with the CDC's expected prevalence (3.1 per 100 live births, 95% confidence interval [CI] 3.2, 3.2), with 80% power and a type I error rate of 5%. For specific defects, the power to detect an increased risk is low. For example, a cohort of 1,000 newborns exposed in the first trimester is needed to detect a 2-fold increase in risk of heart and circulation defects as well as genital and urinary tract defects.
Through July 2002, the registry has received complete registration and pregnancy outcome data for 2,966 prospective pregnancy exposures to 1 or more antiretroviral drugs. Of those, 944 pregnant women were exposed to nelfinavir, and an additional 85 women exposed to nelfinavir were lost to follow-up. In the 4 years since nelfinavir was added the registry, there has been a 9-fold increase in reports of pregnant women exposed to nelfinavir. More than 94% of the reports were from the United States. Table 1 shows the demographic characteristics and clinical status indicators of all cases reported to the registry with a nelfinavir exposure during pregnancy.
Of the 944 reported pregnant women exposed to nelfinavir, there were 960 pregnancy outcomes (16 multiple births). The majority of reported pregnancy exposures involved combination therapy with nucleoside analog reverse transcriptase inhibitors. Table 2 shows the pregnancy outcome and birth defect status by antiretroviral treatment regimen. Overall, there were 915 live births and 21 birth defects reported for a prevalence of 2.3% (95% CI 1.4, 3.5; Table 3). Among 301 live births with a first-trimester exposure to nelfinavir, there were 9 birth defects, for a prevalence of 3% (95% CI 1.4, 5.6). These proportions do not differ significantly from the CDC's birth defect surveillance system, which has a total prevalence of 3.1 per 100 live births (95% CI 3.1, 3.2; P = .99) and an early diagnosis prevalence of 2.2 (95% CI 2.1, 2.2; P = .32). Early diagnosis refers to birth defects identified either before birth or during the first day of life. Additionally, the prevalence of birth defects among offspring of women with first-trimester exposure to nelfinavir is not significantly different from the prevalence of second/third-trimester exposures to nelfinavir (risk ratio 1.7; 95% CI 0.7, 3.9).
Table 4 lists all reported birth defects among infants exposed prenatally to nelfinavir. All cases involved combination therapy with exposures to other antiretroviral therapies in addition to nelfinavir. One case with polycystic kidney disease resulted in an induced abortion at 19 weeks of gestation. This case was not included in the calculation of prevalence of birth defects based on the registry's and CDC's definition.3,9 Among the 22 infants exposed to nelfinavir prenatally with reported birth defects, there is no pattern that might suggest a common etiology.
Although the registry is prospective, data from retrospective reports (pregnancies with a known outcome at the time of reporting) are sometimes received. Retrospective reports tend to be biased toward reporting the severe and unusual cases and are not reflective of the general experience with the medication. Therefore, they are not included in the prospective analysis. However, all retrospective reports with birth defects are reviewed by the Scientific Advisory Committee to assist in the detection of any unusual patterns in birth defects. None has been detected to date.
Table 5 shows the gestational age at delivery and birth weight for singleton live births exposed to nelfinavir prenatally without reported birth defects. Overall, the prevalence of prematurity was 11.8%, and the prevalence of low birth weight was 15%. As expected, infants with earliest exposure to antiretroviral therapy in the first or second trimester were more likely to be born prematurely, to have low birth weight, or both than those with earliest exposure in the third trimester.
Although few drugs introduced in the United States in the last 20 years have sufficient information to assess human teratogenic risk, for nelfinavir, adequate numbers of first-trimester exposures have been monitored to detect a 2-fold increase in the risk of overall birth defects. No such increases have been detected when compared with the rate derived from the CDC's population-based birth defect surveillance system. Additionally, the prevalence is not significantly different when comparing first-trimester exposures when organogenesis occurs with second- and third-trimester exposures. Because nelfinavir is the most commonly reported protease inhibitor in pregnancy, information on the human teratogenic risk is important.3 Previous studies have reported on antiretroviral exposure in pregnancy, but none has had sufficient power to calculate risk of birth defects.11,14
Although the lost to follow-up rate in this study is relatively low (8%), there are some demographic and clinical differences between those who were lost to follow-up and those with complete information. It also is possible that the rate of birth defects differs between the 2 groups. The registry makes every effort to obtain follow-up information. Health care providers are sent a reminder letter followed by 3 phone calls and a final letter requesting follow-up data.
There are a number of limitations that should be considered when interpreting the results of our study, including sample size restrictions, potential classification and reporting bias, potential ascertainment bias, andgeneralizability to the population of pregnant women with antiretroviral exposure worldwide. Although a denominator of at least 300 first-trimester exposures to nelfinavir is sufficient to detect a doubling of risk for overall birth defects when compared with the CDC rates, the sample size is too small to detect a lower risk for overall birth defects and too small to assess the risk of individual birth defects.
The Antiretroviral Pregnancy Registry considers cases prospective if they are enrolled before any prenatal testing or if they had negative findings on a prenatal test as long as they were enrolled before delivery. Cases with evidence of a birth defect on a prenatal test before enrollment are considered retrospective and excluded from the analysis. This inclusion/exclusion policy could potentially create biased results by lowering the overall risk of birth defects.10 Additionally, because reporting of pregnancies is totally voluntary, it is possible that even in prospectively reported cases, potential bias could exist. For example, high-risk pregnancies or low-risk pregnancies may be more likely to be reported. In an attempt to limit these 2 potential sources of bias, the registry includes a group of health care providers who commit, in writing, to report every pregnancy with an antiretroviral exposure that is treated at their site. This subgroup does not have the potential differential reporting bias that the other prospective cases may have. However, this group in itself may be different from other cases (eg, patients from large centers or high-risk practices). Ascertainment of birth defects is likely incomplete because it is limited to those defects diagnosed at or soon after birth. In addition, reports are most often from obstetricians, who may not have complete information about the infant, rather than from pediatricians.10 Birth defects reported by obstetricians are usually limited to those apparent at delivery, which could lead to underascertainment of internal defects.10
Finally, the Antiretroviral Pregnancy Registry is limited to voluntary reports from health care providers who choose to report pregnancy exposures to antiretroviral therapy. The approximately 3,000 pregnancy exposures reported to the Antiretroviral Pregnancy Registry is clearly much lower than the actual number of exposures given the prevalence of HIV-infected pregnant women worldwide. Thus, the results of our study may not be generalizable to all pregnancy exposures to nelfinavir. The demographic characteristics, especially ethnic composition, of the study population also may differ from the general population as well as the CDC's population-based birth defects surveillance system population.
Emphasizing the limitations of voluntary prenatal drug exposure registries, the Antiretroviral Pregnancy Registry Advisory Committee requires that the following consensus statement be included with any presentation of the data from the registry.3
“The Registry's analytic approach is to evaluate drugs in specific classes of antiretroviral therapies (NRTIs [nucleoside analog reverse transcriptase inhibitor(s)], nnRTIs [non-nucleoside reverse transcriptase inhibitor(s)], NtRTIs [nucleotide reverse transcriptase inhibitors], and PIs [protease inhibitor(s)]). Currently there are five specific drugs with large enough groups of exposed women to warrant a separate analysis. These drugs are lamivudine, nelfinavir, nevirapine, stavudine, and zidovudine.
To date, the Registry has not demonstrated an increased prevalence of birth defects overall, or in the specific classes studied, or among women exposed to lamivudine, nelfinavir, nevirapine, stavudine, or zidovudine individually or in combination during the first trimester when compared with observed rates for ”early diagnoses“ in population-based birth defects surveillance systems. While the Registry to date has not detected a major teratogenic signal overall or within classes of drugs or the five individual drugs analyzed separately, the population exposed and monitored to date is only sufficient to detect a two-fold risk of relatively common defects, and cannot detect an increase in the risk of relatively rare defects.
These findings should provide some assurance when counseling patients.”
Because antiretroviral therapy is recommended for all HIV-infected pregnant women to reduce maternal HIV ribonucleic acid levels and decrease the risk of mother-to-child transmission, information on the effects of antiretroviral therapy exposure during pregnancy and fetal outcome is critical. The results of this study suggest that the overall prevalence of birth defects among infants exposed prenatally to nelfinavir is not significantly different compared with the prevalence derived from the CDC's population-based birth defects surveillance system. The number of first-trimester exposures in our study is only sufficient to detect a 2-fold increase in the overall risk of birth defects; the numbers are not sufficient to detect the risk of individual defects. Although nelfinavir should be used in pregnant women only if the benefits clearly outweigh the potential risk to the fetus, the findings from this study should provide some assurance to health care providers and their HIV-infected pregnant patients. The Antiretroviral Pregnancy Registry is ongoing; exposures to all antiretroviral therapies during pregnancy may be reported by calling 800-258-4263 in the United States and Canada and 910-256-0238 internationally or by e-mailing the registry at email@example.com.
1. Lo WY, Friedman JM. Teratogenicity of recently introduced medications in human pregnancy. Obstet Gynecol 2002;100:465–73.
2. Doering PL, Boothby LA, Cheok M. Review of labeling of prescription drugs: is the current system adequate to inform of risks [review]? Am J Obstet Gynecol 2002;187:333–9.
3. Antiretroviral Pregnancy Registry Steering Committee. Antiretroviral Pregnancy Registry International Interim Report for 1 January 1989 through 31 July 2003. Wilmington (NC): Registry Coordinating Center; 2002. Available at: http://www.apregistry.com
. Accessed March 15, 2004.
4. Watts DH. Management of human immunodeficiency virus infection in pregnancy [review]. N Engl J Med 2002;346:1879–91.
5. Cooper ER, Charurat M, Mofenson L, Hanson I, Pitt J, Diaz C, et al, for the Women and Infants’ Transmission Study Group. Combination antiretroviral strategies for the treatment of pregnant HIV-1 infected women and prevention of perinatal HIV-1 transmission. J Acquir Immune Defic Syndr 2002;29:484–94.
6. Centers for Disease Control and Prevention. Public Health Service task force recommendations for the use of antiretroviral drugs in pregnant women infected with HIV-1 for maternal health and for reducing perinatal HIV-1 transmission in the United States. MMWR 1998;47 (No. RR-2). Available at: http://www.cdc.gov/mmwr/PDF/RR/RR4702.pdf
. Retrieved April 20, 2004.
9. Centers for Disease Control and Prevention. Metropolitan Atlanta Congenital Defects Program. Surveillance procedure manual and guide to computerized anomaly record. Atlanta (GA): US Department of Health and Human Services, Public Health Service; 1998.
10. Honein MA, Paulozzi LJ, Cragan JD, Correa A. Evaluation of selected characteristics of pregnancy drug registries. Teratology 1999;60:356–64.
11. Kotelchuck MC, Wise P. Epidemiology of prematurity and goals for prevention. In: Taeusch HW, Yogman MW, editors. Follow-up management of the high-risk infant. Boston (MA): Little, Brown; 1987. p. 3–14.
12. Stevenson RE, Hall JG, Goodman RM, editors. Human malformations and related anomalies. Vol 1. Oxford monographs on medical genetics, no. 27. New York (NY): Oxford University Press; 1993.
13. Scheuerle A, Tilson H. Birth defect classification by organ system: a novel approach to heighten teratogenic signalling in a pregnancy registry. Pharmacoepidemiol Drug Saf 2002;11:465–75.
© 2004 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
14. Edwards SG, Larbalestier N, Hay P, de Ruiter A, Welch J, Taylor GP, Easterbrook P. Experience of nevirapine use in a London cohort of HIV-infected pregnant women. HIV Medicine 2001;2:89–91.