OBJECTIVE: To better describe the safety profile of pregnancy exposures to the human papillomavirus (HPV) type 6/11/16/18 vaccine by acquiring and analyzing postmarketing data on pregnancy outcomes (ie, live births, abortions, fetal deaths, and congenital anomalies).
METHODS: Enrollment criteria included an identifiable patient and health care provider from the United States, France, or Canada and exposure within 1 month before the date of onset of the last menstrual period or at any time during pregnancy. Outcomes of interest were pregnancy outcomes and birth defects. Prospectively reported cases (reported before the outcome of the pregnancy was known) were used for rate calculations.
RESULTS: For the 517 prospective reports with known outcome, 451 (87.2%) were live births, including three sets of twins. Of 454 neonates, 439 (96.7%) were normal. The overall rate of spontaneous abortion was 6.9 per 100 outcomes (95% confidence interval [CI] 4.8–9.6). The prevalence of major birth defects was 2.2 per 100 liveborn neonates (95% CI 1.05–4.05). There were seven fetal deaths (1.5 per 100 outcomes, 95% CI 0.60–3.09).
CONCLUSION: Rates of spontaneous abortions and major birth defects were not greater than the unexposed population rates. Although no adverse signals have been identified to date, the HPV6/11/16/18 vaccine is not recommended for use in pregnant women.
LEVEL OF EVIDENCE: III
To date, data collected in the pregnancy registry for the human papillomavirus 6/11/16/ 18 vaccine do not support a relationship between vaccine exposure and adverse fetal outcomes.
From Merck Research Laboratories, West Point, Pennsylvania; and the Department of Pediatrics, University of Washington, Seattle, Washington.
See related editorial on page 1168.
Funding of the registry is provided by Merck Research Laboratories.
The authors thank David Radley, MS (statistician, Merck and Co., Inc.), for his expert review of the manuscript. Writing support was provided by Heather L. Sings, PhD (Merck and Co., Inc.).
Corresponding author: Adrian Dana, MD, Merck Research Laboratories, PO Box 1000, UG3C-54, North Wales, PA 19454-1099; e-mail: email@example.com.
Financial Disclosure Drs. Dana, Korn, and Haupt and Mss. Buchanan, Goss, Seminack, and Shields are employees of Merck & Co., Inc. (Whitehouse Station, NJ) and hold stock or stock options. Dr. Cunningham has received consultancy fees from Merck & Co., Inc.
Human papillomavirus (HPV) causes cervical cancer, a large portion of anal and other genital cancers, genital warts, recurrent respiratory papillomatosis, and cancers of the head and neck.1 A quadrivalent HPV6/11/16/18 vaccine (Gardasil/Silgard; Merck and Co., Inc., Whitehouse Station, NJ) was licensed in the United States in June 2006 and subsequently in the European Union. The vaccine is a three-dose regimen and is indicated in girls and women 9–26 years of age for the prevention of cervical, vulvar, and vaginal cancer; their respective precancers; and genital warts caused by HPV types 6, 11, 16, and 18. The vaccine is labeled Food and Drug Administration Pregnancy Category B, because animal reproductive toxicology studies revealed no evidence of impaired female fertility or harm to the fetus due to the vaccine. However, because there are no adequate and well-controlled studies in pregnant women, and animal reproduction studies are not always predictive of human response, the vaccine is not recommended for use in pregnant women.
In the five phase III clinical trials, all pregnancies were followed for outcome, including those that occurred before, during, or after vaccine exposure. The range of follow-up for the individual studies was 0.6–3.7 years. In total, there were 2,008 and 2,029 pregnancies with known outcome among 10,126 vaccine and 10,425 placebo recipients aged 16–45 years, respectively (see accompanying article by Garland et al). No significant differences were noted for the proportions of pregnancies resulting in live birth, fetal loss, spontaneous abortion, or congenital anomalies. The vaccine was well tolerated among women who became pregnant.
Because the vaccine is recommended for women of childbearing age, it is acknowledged that inadvertent pregnancy exposures are likely to occur in the postmarketing setting. As with any drug or vaccine, rare and potentially serious adverse events may not emerge during clinical trials involving a limited number of people. To collect further data regarding the safety profile of the vaccine, Merck and Co., Inc. has established the Pregnancy Registry for Gardasil. The goal of the registry is to acquire and analyze information on pregnancy outcomes (ie, live births, abortions, fetal deaths, and congenital anomalies) to better describe the safety profile of pregnancy exposures. Here, we provide the first 2 years of postlicensure surveillance.
MATERIALS AND METHODS
The pregnancy registry is a company-run registry whose purpose is to collect data regarding the safety profile of HPV6/11/16/18 vaccine in pregnancy. Because the vaccine is not indicated in pregnancy, all pregnancy exposures are either inadvertent or due to off-label use. The registry is, therefore, based on voluntary postmarketing reports. The pregnancy registry is conducted in the United States, Canada, and France. However, regardless of the country of origin, all reports of exposure during pregnancy that are received by Merck and Co., Inc. or Sanofi Pasteur MSD (Lyon, France), are entered into the Merck safety database termed the New Worldwide Adverse Experience Systems and are reviewed by registry personnel for pregnancy outcomes.2 External experts are consulted as needed to evaluate reported outcomes. An annual report including cumulative data is prepared and then reviewed by a consultant teratologist. The annual report is distributed to health care providers who report cases of exposure to the registry. This article summarizes the content in the second annual report, which covers the first 2 years after licensure (from June 1, 2006, through May 31, 2008).
The registry enrollment criteria include 1) an identifiable patient (presence of a unique patient identifier), 2) exposure within 1 month before the date of onset of the last menstrual period or at any time during pregnancy, 3) identification of a health care provider, and 4) residency in a country where the registry is active. A patient or provider may decline to participate in the registry at any time. A report is considered lost to follow-up when no further information about the outcome of the pregnancy is available despite repeated attempts to obtain information.
Reports are classified by the timing of registry notification in relation to pregnancy outcome. Prospective reports are those received before the outcome of the pregnancy is known. Retrospective reports are those received after the outcome of the pregnancy is known and include reports made after prenatal testing has identified an abnormality, even if the pregnancy outcome has not yet occurred.
The primary outcomes of interest are pregnancy outcomes including live births, fetal deaths (defined as nonviable conceptuses in pregnancies 20 weeks or later after the last menstrual period), elective terminations, spontaneous abortions (defined as spontaneous loss of conceptuses before the 20th week after the last menstrual period), and birth defects. To identify defects that were not detected at birth, pediatric medical records are requested for up to 2 years after birth for all newborns with written maternal consent. Birth defects are defined according to the Center for Disease Control and Prevention Metropolitan Atlanta Congenital Defects Program guidelines.3,4 Gestational age and birth weights are categorized according to definitions from the March of Dimes Foundation.5 A full-term delivery is defined as 37–42 weeks from the last menstrual period, preterm births are defined as less than 37 weeks of gestation, and very preterm births are defined as those that occurred before completion of the 32nd week. Low birth weight is defined as less than 2,500 g (less than 5.5 lb) at term, and very low birth weight is defined as less than 1,500 g (less than 3.3 lb) at term.
Prospective reports are analyzed separately from retrospective reports to minimize the reporting bias toward abnormal outcomes inherent in retrospective reporting. The enrolled prospective reports with an estimated date of delivery before May 31, 2008 (the cutoff date for this report), comprise the primary analysis cohort and are used for rate calculations. Reports for which an estimated date of delivery was not provided were assigned an estimated date of delivery, which was calculated to be 7 months after the initial date that the report of exposure during pregnancy was made to the company. Birth defect frequencies were calculated on prospective reports using Metropolitan Atlanta Congenital Defects Program methodology: prevalence at birth, expressed as the number of affected cases (liveborn, fetal deaths, and terminations at 20 weeks or later) per 100 liveborn neonates.6 The overall rate of spontaneous abortion was calculated as the number of spontaneous abortions per the total number of natural pregnancy outcomes. Spontaneous abortions involving multiple fetuses (ie, twins) are counted as a single spontaneous abortion.6–8 Because the natural outcome of a pregnancy that was terminated is unknown, elective abortions are not included in rate calculations. The rate of fetal death was calculated as the number of fetal deaths per the number of live births plus fetal deaths.7 Confidence intervals (CIs) for the rates were calculated following the procedures of Lilienfeld and Stolley.9
We estimated the timing of exposure to HPV6/11/16/18 vaccine in relation to the first day of the last menstrual period. Time of exposure was calculated for all doses that fell within the defined exposure period (within 1 month before the last menstrual period or at any time during pregnancy). The first trimester was defined as 0–13 weeks from the last menstrual period, the second trimester was defined as 14–27 weeks from the last menstrual period, and the third trimester was defined as 28–42 weeks from the last menstrual period.
The pregnancy registry for the HPV6/11/16/18 vaccine is enhanced postmarketing surveillance implemented at the request of the U.S. Food and Drug Administration. The concept sheet for the registry was reviewed by the Food and Drug Administration. Patient consent is not required for a health care provider to report exposures of HPV6/11/16/18 vaccine during pregnancy because the registry procedures are part of the routine postmarketing surveillance activities performed for voluntarily reported adverse events. Because the registry does not have direct access to the patients, the health care providers are asked to discuss the registry and to obtain consent from the patient. If the provider does not obtain informed consent but continues to provide information to the pregnancy registry, that information is accepted, in compliance with the Code of Federal Regulations and Health Insurance Portability and Accountability Act rules. Because the data in the registry are collected via postmarketing surveillance and are Health Insurance Portability and Accountability Act exempt, there is no institutional review board review.
Between June 1, 2006, and May 31, 2008, 2,003 reports of exposure to HPV6/11/16/18 vaccine during pregnancy were received from France, the United States, and Canada, and 1,440 met the enrollment criteria (Fig. 1). Of these, 577 were not included in this analysis because the estimated date of delivery was after the cutoff date (May 31, 2008). Data from these reports will be included in the next annual report.
Of the 863 reports with estimated date of deliveries before May 31, 2008, 787 were reported prospectively, ie, before the outcome of the pregnancy was known (Fig. 1). Of the 787 prospective reports, 140 (17.8%) were lost to follow-up despite repeated attempts to obtain information. For 130 of 787 reports (16.5%), attempts are still being made to obtain outcome information, and these outcomes will be included in the next annual report. The 517 prospective reports with known outcome are shown in Table 1.
For 21 of the 26 cases of elective abortion, the reason for the termination was not provided. The remaining five reported that the termination was due to nonmedical reasons, citing social grounds or unplanned pregnancy. One report of elective abortion involved a fetus diagnosed with anencephaly and a hypoplastic heart defect. Of the 34 spontaneous abortions, including one triplet abortion, none reported a congenital anomaly. When considering maternal age, no spontaneous abortions were reported in girls aged 9–15 years, 13 were reported in girls and women aged 16–18 years, 18 were reported in women aged 19–26 years, and three were reported in women older than 26 years. The overall rate of spontaneous abortion was 6.9 per 100 outcomes (95% CI 4.8–9.6).
There were seven fetal deaths (Table 2). Two had attributable causes: 1) placental abnormalities and severe intrauterine growth restriction and 2) cord accident. Three had other risk factors: 1) mother CDE (Rh) negative, father CDE positive, 2) endometriosis, and 3) smoking/bipolar disorder/history of narcotic overdose. One had no other information reported. The seventh case of fetal death was reported as a “stillborn baby delivered early.”
There were 451 exposed pregnancies that resulted in the live birth of 454 newborns (this number includes four twin pregnancies resulting in seven liveborn neonates). The duration of gestation was reported in 218 of 451 pregnancies, with 190 of 218 (87%) going to term. Twenty-eight pregnancies (including all of the twin pregnancies) resulted in preterm deliveries ranging from 22 to 36 weeks of gestation. Nine of these were before the start of the 33rd week of gestation (very preterm). Birth weights were reported for 256 of 454 neonates (56.4%). Five had very low birth weights ranging from 700 g (1.54 lb) to 1,477 g (3.25 lb); 24 (including two sets of twins) had low birth weights ranging from 1,506 g (3.3 lb) to 2,471 g (5.4 lb); and 227 had birth weights of 2,500 g (5.5 lb) or greater, including one set of twins. The highest birth weight reported was 4,602 g (10.1 lb).
Of 454 neonates in the prospective reports, 439 (96.7%) were normal. There were 10 prospective reports of neonates with major birth defects, for a prevalence at birth of 2.2 per 100 liveborn neonates (95% CI 1.05–4.05). These birth defects are listed in Table 3 (prospective reports), by system organ class. Of these 10 major birth defects, three were identified prenatally, five were identified at birth or sometime thereafter, and two did not specify when the defect was identified.
There were four additional reports of neonatal abnormalities that did not meet the Metropolitan Atlanta Congenital Defects Program classification as major birth defects as follows: 1) A female neonate born at 39 weeks of gestation was found to have a complete agglutination of the labia minor at her 4-month examination; 2) an otherwise healthy female neonate at term with mild metatarsus adductus deformities and mild curly toe deformities was born to a woman with diet-controlled gestational diabetes; 3) a male neonate with bilateral hydrocele was delivered at 33 2/7 weeks by cesarean secondary to hemolysis, elevated liver enzymes, low platelets syndrome; and 4) a full-term healthy male neonate was born with a congenital brown nevus on the forehead.
There was one live birth that subsequently resulted in an early neonatal death: a preterm neonate with intrauterine growth restriction and maternal risk factors including antiphospholipid syndrome and placental insufficiency.
The timing of vaccine exposure was available for 395 pregnancies with known outcomes. Forty-four of these 395 women experienced a multiple exposure, ie, more than one vaccination within 1 month before the date of the last menstrual period or during pregnancy (43 double and one triple exposure). Forty-three (98%) of the multiple exposures (42 double, one triple) resulted in live births, and one double exposure in the first trimester ended in an elective abortion (reason not given). Of the 90 exposures that were determined to have occurred within 30 days before the last menstrual period, 84 (93%) were live births, three (3%) were spontaneous abortions, two (2%) were elective abortions, and one (1%) resulted in fetal death. Of the 84 live births with an exposure within 30 days before the last menstrual period, 28 (33%) had a second vaccine exposure (25 during the first trimester and three during the second trimester).
Seventy-six retrospective reports of exposure to HPV6/11/16/18 vaccine during pregnancy were received (Table 1). Six retrospective reports included pregnancy outcomes with major birth defects (Table 3, retrospective reports). This included three live births, one spontaneous abortion, and two fetal deaths. One additional report initially included prenatal testing (maternal serum α-fetoprotein and a sonogram) which indicated an increased risk of Down syndrome, but the neonate was normal at birth. The remaining two fetal deaths were described as follows: 1) fetal death at approximately 37 weeks of gestation—the cord was wrapped around the fetus' arm and shoulder four times; and 2) neonate was born at approximately 25 weeks of gestation (via cesarean delivery due to preterm labor) and then died—the autopsy report was not available, but the perinatologist stated that the most marked abnormality was reversal of end-diastolic flow in the umbilical artery and intrauterine growth restriction.
The pregnancy registry for the HPV6/11/16/18 vaccine was established at the time of licensure as part of a multifaceted approach to monitoring exposures to the vaccine during pregnancy. Although the numbers of exposures are not sufficient to rule out a low risk, data collected from the first 2 years of the pregnancy registry for the HPV6/11/16/18 vaccine do not support a relationship between vaccine exposure and adverse pregnancy or fetal outcomes; however, the vaccine is not recommended for use during pregnancy because there are no well-controlled data in pregnant women.
The HPV6/11/16/18 vaccine is licensed for use in women and girls at an age of childbearing potential. Inadvertent exposures during pregnancy occur, and these exposures may lead to anxiety on the part of the exposed woman and her health care professional. Several studies have shown that women overestimate the risk of fetal malformation after exposure to medication, with some attributing the same risk to nonteratogenic drugs as to known human teratogens, such as thalidomide.10,11 In a study of Canadian health care professionals, Pole et al12 found that even health care professionals who reviewed the prescribing information of drugs with low risk rated them as unsafe. Because of the ethical restrictions on testing drugs in pregnant women, there are usually few human pregnancy data to inform drug labeling at the time of product approval. To encourage the collection and analysis of pregnancy exposure data, the U.S. Food and Drug Administration has published guidance for industry on establishing pregnancy exposure registries to provide information for the label, to inform clinical practice, and to improve patient care.13 The information from the registry adds to our knowledge of the safety profile of the HPV6/11/16/18 vaccine and supplements the animal toxicology studies and clinical trial data.
For the prospective reports, the rate of spontaneous abortion (6.9 per 100 outcomes, 95% CI 4.8–9.6) was comparable to that reported in the literature. Among clinically recognized pregnancies (ie, known pregnancy based on late menstrual period, symptoms of early pregnancy, or test results), 15% result in spontaneous abortion.14 In the general population, the rate of fetal deaths is approximately 0.62–1 per 100 outcomes,7,15 which is slightly lower than our findings of 1.5 per 100 outcomes (95% CI 0.60–3.09). However, the number of fetal deaths is small (seven reports), and the CI is wide. One fetal death was attributed to a cord accident, and one involved a placental abnormality and severe intrauterine growth restriction. Three other reports contain possible risk factors, including concomitant medication (methylphenidate hydrochloride), CDE incompatibility, and maternal drug use.
The major birth defects reported to the pregnancy registry demonstrate no specific pattern or target organ, and the timing of vaccine exposure and biologic plausibility do not support a causal relationship between the HPV vaccine and these defects. In the Metropolitan Atlanta Congenital Defects Program, the expected prevalence of major malformations at birth is 2.67%, although the rates of specific malformations such as chromosomal and neural tube defects differ according to maternal age.6 The rate of major malformations in the pregnancy registry was 2.2 per 100 liveborn neonates, with a 95% CI of 1.05–4.05.
In the pregnancy registry, there were two reports of schizencephaly (one retrospective and one prospective), a rare central nervous system malformation thought to result from an early focal destruction of the germinal matrix. The estimated population prevalence rate for schizencephaly is 1.54 in 100,000 births.16 There were two reports of anencephaly (one retrospective and one prospective). Anencephaly is thought to be a severe form of anterior neural tube defect within the spectrum of myelomeningocele and spina bifida. The estimated population prevalence rate is 1.1 in 10,000 births.17 The exact cause of schizencephaly and anencephaly are unknown. Anencephaly is thought to be an anterior neural tube defect, whereas schizencephaly is thought to result from an early destructive process of the germinal matrix after neural tube closure. Because these two entities have different pathogenic mechanisms and differing critical periods during development, they are not considered to be the same class of malformation. However, it should be noted that one of the cases with anencephaly also had a hypoplastic left heart. Although these could represent the co-occurrence of two independent congenital malformations, it suggests that it may represent an undefined multiple malformation syndrome. No comment can be made about the incidence of either anencephaly or schizencephaly with only a single case of each in the prospective and retrospective reports. We will continue to monitor for the occurrence of these defects in the ongoing registry.
There were two reports of talipes (one prospective, exposure at 4–5 weeks after the last menstrual period, and one retrospective, exposure at 17–18 weeks after the last menstrual period) and a single prospective report of cleft lip (exposure at 2.7 weeks). The pathogenesis of talipes is multifactorial, with genetic and environmental risk factors. The prevalence at birth of talipes in the general population is 1–3 in 1,000 live births.18 The etiology of cleft lip and palate is multifactorial, with many known environmental, nutritional, and genetic factors related to increased risk.19 The estimated risk of cleft lip with or without cleft palate is 1 in 1,000 births.18
There are limitations to the data collected in the pregnancy registry. Voluntary reporting is subject to bias, such as underreporting of birth defects if the malformations are not manifest at delivery or if there is variable use of diagnostic tests. There is a bias toward reporting abnormal outcomes, especially in retrospective reports. For this reason, retrospective reports are analyzed separately from prospective reports, and only prospective reports are used for rate calculations. Because there is no comparator group, event rates must be compared with population background rates. It is also difficult to interpret the significance of specific malformations that occur rarely in the general population. Finally, bias may be introduced by the high rate of loss to follow-up in spontaneously reported data whose lag time to outcome is a period of many months. Although the total number of outcomes at present is small, the data from this pregnancy registry are comparatively large when considered against other vaccine exposure in pregnancy data sources.20,21
In summary, the evaluation of pregnancy outcomes among women exposed to HPV6/11/16/18 during pregnancy has generated no safety concerns after 2 years. Because of the limited number of reports with known outcomes to date, no definitive conclusions can be made about the potential effects of exposure to HPV6/11/16/18 vaccine during pregnancy. The HPV6/11/16/18 vaccine is not recommended for use in pregnant women. Monitoring of the safety of the vaccine in pregnancy continues and the approach is multifaceted. The registry is an important part of the monitoring effort and is ongoing; the next annual report is anticipated in the last quarter of 2009. Physicians and other health care providers should remember to question all women of childbearing age about the possibility of pregnancy before administering any vaccination. Health care professionals are encouraged to enroll patients who inadvertently receive the vaccine into the pregnancy registry. Information on reporting to the registry can be found in the prescribing information in countries where the registry is active.
1.Barr E, Sings HL. Prophylactic HPV vaccines: new interventions for cancer control. Vaccine 2008;26:6244–57.
2.Lehman HP, Benson JO, Beninger PR, Anderson CA, Blumenthal SJ, Sharrar RG. A five-year evaluation of reports of overdose with indinavir sulfate. Pharmacoepidemiol Drug Saf 2003;12:449–57.
3.Centers for Disease Control and Prevention, Metropolitan Atlanta Congenital Defects Program. Surveillance procedure manual and guide to computerized anomaly record. Atlanta (GA): Centers for Disease Control and Prevention; 1998.
6.Correa A, Cragan JD, Kucik JE, Alverson CJ, Gilboa SM, Balakrishnan R, et al. Reporting birth defects surveillance data 1968–2003 [published erratum appears in Birth Defects Res A Clin Mol Teratol 2008;82:41–62]. Birth Defects Res A Clin Mol Teratol 2007;79:65–186.
7.MacDorman MF, Kirmeyer S. Fetal and perinatal mortality, United States, 2005. Natl Vital Stat Rep 2009;57(8):1–19.
8.Ventura SJ, Abma JC, Mosher WD, Henshaw S. Estimated pregnancy rates for the United States, 1990–2000: an update. Natl Vital Stat Rep 2004;52(23):1–9.
9.Lilienfeld DE, Stolley PD. Foundations of epidemiology. 3rd ed. New York (NY): Oxford Press; 1994.
11.Koren G, Bologa M, Long D, Feldman Y, Shear NH. Perception of teratogenic risk by pregnant women exposed to drugs and chemicals during the first trimester. Am J Obstet Gynecol 1989;160:1190–4.
12.Pole M, Einarson A, Pairaudeau N, Einarson T, Koren G. Drug labeling and risk perceptions of teratogenicity: a survey of pregnant Canadian women and their health professionals. J Clin Pharmacol 2000;40:573–7.
14.Scott JR. Early pregnancy loss. In: Scott JR, DiSaia PJ, Hammond CB, Spellacy WN, editors. Danforth's obstetrics and gynecology. Philadelphia (PA): Lippincott Williams & Wilkins; 1999. p. 143–53.
15.Fox R, Pillai M, Porter H, Gill G. The management of late fetal death: a guide to comprehensive care. Br J Obstet Gynaecol 1997;104:4–10.
16.Curry CJ, Lammer EJ, Nelson V, Shaw GM. Schizencephaly: heterogeneous etiologies in a population of 4 million California births. Am J Med Genet A 2005;137:181–9.
17.Mathews TJ, Honein MA, Erickson JD. Spina bifida and anencephaly prevalence: United States, 1991–2001. MMWR Recomm Rep 2002;51(RR-13):9–11.
18.Wynbrandt J, Ludman MD. The encyclopedia of genetic disorders and birth defects. 3rd ed. New York (NY): Fact on File, Inc.; 2008.
19.Krapels IP, Vermeij-Keers C, Muller M, de Klein A, Steegers-Theunissen RP. Nutrition and genes in the development of orofacial clefting. Nutr Rev 2006;64:280–8.
20.Wilson E, Goss MA, Marin M, Shields KE, Seward JF, Rasmussen SA, et al. Varicella vaccine exposure during pregnancy: data from 10 years of the pregnancy registry. J Infect Dis 2008;197(suppl 2):S178–84.
© 2009 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
21.Ryan MA, Seward JF; Smallpox Vaccine in Pregnancy Registry Team. Pregnancy, birth, and infant health outcomes from the National Smallpox Vaccine in Pregnancy Registry, 2003–2006. Clin Infect Dis 2008;46(suppl 3):S221–6.