A live, attenuated viral vaccine for prevention of varicella was licensed in March 1995 and recommended for routine childhood immunization and vaccination of susceptible older children and adults. From March 17, 1995 to March 16, 2000, approximately 21.2 million doses of VARIVAX (Varicella virus vaccine live [Oka/Merck], Merck & Co., Inc., West Point, PA) have been distributed. The vaccine is considered contraindicated during pregnancy, and it is recommended that pregnancy be avoided for 3 months (Merck & Co., Inc.)1 or 1 month (Advisory Committee on Immunization Practices2 and American Academy of Pediatrics)3 after vaccination because the effects, if any, of the vaccine on fetal development are not known.
Wild-type varicella zoster virus infection occurs in approximately 0.05–0.07% of pregnancies.4 Fetal or embryonic infections with wild-type varicella zoster virus may result in a pattern of anomalies known as congenital varicella syndrome5 that may be manifested by cutaneous scarring in a dermatome distribution and/or hypoplasia of an extremity. Additional manifestations may include low birth weight, microcephaly, localized muscular atrophy, ocular anomalies, and neurologic abnormalities.6 In a prospective study conducted in the United Kingdom and West Germany, the risk of congenital varicella syndrome in approximately 1700 cases of gestational varicella was estimated to be 0.4% for maternal infection occurring between 0 and 12 weeks after the first day of the last menstrual period (LMP) and 2% for maternal infection occurring between 13 and 20 weeks after the LMP.7 Serosurveys estimate that approximately 5% of U.S. women of childbearing age are seronegative for varicella zoster virus, with higher rates of susceptibility occurring in women from Puerto Rico and other tropical areas.8 Women of childbearing age susceptible to chickenpox are at risk of contracting varicella during pregnancy and should be vaccinated.2,9
A pregnancy registry for varicella vaccine was established as a collaborative effort between the Centers for Disease Control and Prevention and Merck & Co., Inc. to identify exposed women and to obtain information on the outcome of their pregnancies. The objective of the registry is to estimate whether exposure to varicella vaccine in pregnancy is associated with congenital varicella syndrome or other birth defects. We report the findings from the first 5 years of the pregnancy registry.
MATERIALS AND METHODS
The pregnancy registry is an observational post-marketing surveillance effort that collects information on the offspring of women exposed to varicella vaccine during pregnancy or within 3 months before conception. Women with an expected date of delivery up to March 16, 2000 were identified from voluntary reports received from health care providers or the public. The enrollment criteria included an identifiable health care provider, an identifiable patient (by name, initials, chart number, or code), an exposure occurring in the United States or Puerto Rico (or Canada as of January 1999), and receipt of vaccine within 3 months before, or at any time during, pregnancy. The pregnancies were monitored and the outcomes ascertained from questionnaires completed voluntarily by the health care providers. Providers who did not complete the mailed questionnaires were contacted by telephone and prompted to submit the forms. If the forms were not received or were incomplete, the outcome information was collected in a telephone interview with the provider. The primary outcome of interest was congenital varicella syndrome; secondary outcomes included congenital malformations and pregnancy outcomes (ie, live births, spontaneous abortions, late fetal deaths [stillbirths], and elective abortions). Information on the analysis of the products of conception from fetal losses was requested but was rarely available; thus, the risks of congenital anomalies for fetal losses were not calculated.
The reports were classified by the timing of registry notification in relation to the pregnancy outcome (prospective or retrospective), by the serologic status of the woman at the time of vaccination (seronegative or unknown), and by the timing of exposure to the first dose of vaccine in relation to the gestational week of pregnancy from the LMP. Prospective reports are those received after vaccination but before the outcome of pregnancy was known. Retrospective reports are those received after the outcome of pregnancy was known or if a probable outcome was identified before delivery by a diagnostic test (eg, a congenital anomaly identified by ultrasound scanning). To avoid the selection bias toward abnormal outcomes from voluntary retrospective reports,10 only prospective reports were used to calculate the outcome rates. Retrospective reports were used to identify the potential types of adverse experiences.
Serology testing is not a requirement before vaccination; thus, most women enrolled in the registry lacked serologic evidence of susceptibility and received the vaccine on the basis of a negative history of the disease. Since 71–93% of adults who report a negative or uncertain history of varicella are immune on testing,2 most women with unknown serologic status were likely to have been immune to varicella at the time of the vaccination. For this reason, the outcome rates for congenital varicella syndrome and congenital anomalies were calculated using only the women shown to be seronegative before vaccination. Confidence intervals (CI) were calculated using the Haenszel et al11 method for Poisson-distributed variables.
On the basis of the known risks from wild-type varicella zoster virus and biologic plausibility, we defined three potential risk periods for exposure: high risk, exposure during the first or second trimester of pregnancy; moderate risk, exposure during either the third trimester or the month before the LMP; and low risk, exposure more than 1 month before the LMP. We defined exposure on the basis of the first dose of vaccine received.
From March 17, 1995 through March 16, 2000, the registry received 701 reports of inadvertent exposure to varicella vaccine from 3 months before or at any time during pregnancy. Twenty-one retrospective reports and 362 prospective reports were available for analysis (the remainder were excluded as shown in Figure 1). Of the 362 prospectively reported women with known pregnancy outcomes, 92 (25%) were known to be seronegative before vaccination and 270 (75%) were of unknown serologic status. Most women (66%) had received the first dose of vaccine in the month immediately preceding or immediately after their LMP.
Fifty-eight (63%) of the 92 seronegative women reported prospectively were vaccinated during the first or second trimester of pregnancy, and their pregnancies resulted in 56 live births (97%) and two first-trimester spontaneous abortions (3%). No infants had congenital varicella syndrome (rate 0%, 95% confidence interval [CI] 0, 15.6) (Table 1). Two term infants were born with congenital anomalies—postaxial polydactyly and a variant of tetralogy of Fallot (Table 2). One additional pregnancy exposed in the high-risk period was terminated electively after prenatal diagnosis of trisomy 21. Sixteen seronegative women (17%) were vaccinated in the moderate risk period—two in the third trimester, resulting in two live births, and 14 in the month before the LMP, resulting in 13 live births and one spontaneous abortion. No congenital varicella syndrome or congenital anomalies were reported. Eighteen susceptible women (20%) were vaccinated in the low-risk period—2 and 3 months before the LMP; these pregnancies resulted in 14 live births and four spontaneous abortions. No major congenital anomalies were reported. One minor congenital anomaly was diagnosed—small preauricular sinuses. No cases of congenital varicella syndrome were reported in any risk group (rate 0%, 95% CI 0, 10.2).
Of the 270 women reported prospectively whose serologic status was unknown, 234 pregnancies resulted in the birth of 240 live born infants (including four sets of twins and one set of triplets). No cases of congenital varicella syndrome were reported. Two live born infants were born with hypospadias. One pregnancy was aborted electively after the prenatal diagnosis of trisomy 18. The gestational week of spontaneous abortion was reported in 29 of 35 reports; 26 (90%) occurred during the first trimester (0–13 weeks), and three (10%) occurred in the second trimester (at 14, 17, and 20 weeks' gestation). One fetal death occurred at approximately 23 weeks' gestation; the autopsy revealed no evidence of varicella-related lesions, cultures for varicella virus were negative, and the cause of death was not determined.
Five retrospective reports of congenital anomalies were received (Table 3). Two induced abortions (for holoprosencephaly and trisomy 21) were reported. One report described a case of cystic hygroma and anasarca identified by ultrasound scanning before spontaneous abortion. Another report described a delivery at 35 weeks' gestation of a male infant who, at age 2 years, was diagnosed with left kidney dysplasia, compensatory hypertrophy of the right kidney, and modest renal insufficiency. His mother had received varicella vaccine at 28 weeks after her LMP. The final retrospective report of a congenital anomaly described a 2-month-old infant who was brought to the emergency department for seizure activity. The infant had been born with left anophthalmia. A computed tomography scan revealed a left orbital roof defect, a left orbital mass with encephalocele, a left suprasellar arachnoid cyst, and microcephaly. The infant did not have varicella antibodies by fluorescent antibody to membrane antigen testing. The Centers for Disease Control and Prevention reviewers who analyzed this report confirmed that the ocular anomaly was unlike those associated with congenital varicella syndrome. They concluded that the anomaly was not likely to be due to an infectious disease process but was a neural tube defect.
As of March 16, 2000, the registry had received 21 reports of inadvertent administration of varicella vaccine instead of varicella zoster immune globulin to pregnant women who had been exposed to chickenpox. The timing of the exposures ranged from gestational weeks 4 to 34. All 17 live born infants reported prospectively were reportedly normal and without congenital anomalies (eight were born to seronegative mothers). One woman of unknown serologic status received both varicella vaccine and varicella zoster immune globulin at 29 weeks' gestation and subsequently delivered healthy triplets at 34 weeks' gestation. One retrospective report described a woman who received varicella vaccine instead of varicella zoster immune globulin at 28 weeks' gestation and gave birth to a child with renal anomalies (as described earlier).
We report here the birth outcomes from a cohort of women reported to the Pregnancy Registry for VARIVAX who received varicella vaccine during, or shortly before pregnancy. Although no cases of congenital varicella syndrome occurred among any of the live birth offspring, the precision of this study was limited by small numbers, so continued surveillance and enrollment of exposed pregnant women is warranted. Most of the first doses of vaccine were administered in the month immediately preceding or immediately following the LMP, shortly before conception or before the women realized they were pregnant. Data reported to the registry identified a problem of product confusion; in 21 cases, pregnant women were administered varicella vaccine instead of varicella zoster immune globulin after an exposure to wild-type varicella.
No reports of congenital varicella syndrome were received in the susceptible or unknown immunity cohorts. The background rate for major congenital anomalies in the U.S. population is 2.2 to 3.2 per 100 live born infants,12 with an additional 14.2 to 39.9 per 100 live born infants with minor malformations.13 The 2.2% rate is obtained from diagnoses made within 24 hours of birth, which would include most, but not all, of the congenital anomalies identified in the pregnancy registry's prospective reports. Both major and minor anomalies are included in the pregnancy registry database. The occurrence of three major or minor congenital anomalies in 85 live born outcomes from prospective seronegative reports (3.5%; 95% CI 0.72, 10.2) does not exceed what is reported in the general U.S. population. These three anomalies (postaxial polydactyly, tetralogy of Fallot variant, and preauricular sinuses) are not related by morphology or by critical period of exposure and thus, do not appear to be associated with exposure to varicella vaccine. One additional congenital anomaly, trisomy 21, identified by prenatal testing in the fetus of a seronegative mother and electively terminated, appears unrelated to the anomalies identified in the live born infants.
Whether reported prospectively or retrospectively, the congenital anomalies identified showed no specific pattern in either the susceptible cohort or the sample population as a whole. Two cases of hypospadias from prospective reports may be attributed to background occurrence and one prospective and one retrospective case of Down syndrome may reflect the age of the exposed mothers (ages 30 and 40). The timing of vaccine administration and biologic plausibility do not support the assumption of a causal relationship between varicella vaccine exposure and these anomalies. The reports of these specific anomalies will be monitored closely in the ongoing registry.
Ascertaining the serologic status of a pregnant woman with a negative or uncertain history of varicella is an important component of prenatal care. After testing, seronegative women should be counseled to avoid exposure to varicella and, if exposure occurs, to report the exposure immediately so as to receive varicella zoster immune globulin in a timely fashion. Those who remain seronegative throughout pregnancy should be vaccinated after delivery.
One finding of concern was that product confusion between varicella vaccine and varicella zoster immune globulin continues to occur despite publication of the first seven reports in 1996.14 The Advisory Committee on Immunization Practices recommends that varicella zoster immune globulin be administered to susceptible pregnant women within 96 hours of varicella zoster virus exposure to lessen the risk and severity of disease.2 Since licensure of the vaccine, 21 reports of accidental vaccination of pregnant women have been made. These errors may result in delayed or no provision of varicella zoster immune globulin to women at risk of severe varicella disease. Health care providers need to be more cognizant of the distinct indications and uses of these two different products.
There are a number of limitations to these data. Post-licensure surveillance relies on the voluntary reporting of individuals and health care professionals. Voluntary reporting is known to be subject to various types of bias, including inaccuracy and underreporting.15 Because outcome information is provided primarily by prenatal health care providers, malformations that are not manifest at delivery may not be detected. Procedures to obtain newborn and pediatric records to corroborate outcomes have been implemented but were not in place during most of this study period. Spontaneous abortions and elective terminations may be underrepresented if they occurred before the exposure was reported. The timing of the first dose of vaccine was used to stratify women into risk groups because most (75–94%)1 adults seroconvert after their first dose. Finally, because most of the women enrolled in the study were of unknown serologic status but likely to have been immune to varicella before vaccination, we cannot draw conclusions from the outcomes of their pregnancies and must rely on the smaller number of women known to have been susceptible before vaccination.
Reports of vaccination before and during pregnancy suggest that most instances of inadvertent vaccination with the live attenuated varicella vaccine occur before women are aware they are pregnant. Vaccination with the live attenuated varicella vaccine remains contraindicated in pregnancy, although the decision to terminate a pregnancy should not be based on whether the vaccine was administered during pregnancy.2 Physicians and other health care providers should remember to question all women of childbearing age about the possibility of pregnancy before administering any vaccine. At the time of varicella vaccine administration, women should be reminded to avoid pregnancy after vaccination for an interval of 1 to 3 months. It is reassuring that no cases of congenital varicella syndrome were reported among the offspring of women with inadvertent exposure to the vaccine during pregnancy. Because of the small numbers, this study has limited precision to define the risk precisely, if any, associated with exposure. Health care professionals are encouraged to continue to enroll patients in the Pregnancy Registry by calling (800) 986–8999.
1. Physicians' Desk Reference, 54th ed. VARIVAX [varicella virus vaccine live (Oka/Merck)]. Montvale, NJ: Medical Economics, 2000:1901–3.
2. Centers for Disease Control and Prevention. Prevention of varicella: Recommendations of the Advisory Committee on Immunization Practices. MMWR Morb Mortal Wkly Rep 1996;45(RR-11):1–36.
3. American Academy of Pediatrics. Varicella zoster infections. In: Peter G, ed. 1997 Red book: Report of the Committee on Infectious Diseases, 24th ed. Elk Grove Village, IL: American Academy of Pediatrics, 1997: 573–85.
4. Sever J, White LR. Intrauterine viral infections. Annu Rev Med 1968;19:471–86.
5. Laforet EG, Lynch CL. Multiple congenital defects following maternal varicella: Report of a case. N Engl J Med 1947;236:534–7.
6. Roberts RM. Fetal effects from varicella-zoster. In: Buyse ML, ed. Birth defects encyclopedia. Cambridge, MA: Blackwell Scientific Publications, 1990:708–10.
7. Enders G, Miller E, Cradock-Watson J, Bolley I, Ridehaigh M. Consequences of varicella zoster in pregnancy: Prospective study of 1739 cases. Lancet 1994;343:1547–50.
8. Gershon AA, Raker R, Steinberg S, Topf-Olstein B, Drusin LM. Antibody to varicella-zoster virus in parturient women and their offspring during the first year of life. Pediatrics 1976;58:692–6.
9. Centers for Disease Control and Prevention. Prevention of varicella: Updated recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep 1999;48(RR06);1–5.
10. Mitchell A. Special considerations in studies of drug-induced birth defects. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons; 1994:595–608.
11. Lilienfeld DE, Stolley PD. Foundations of epidemiology, 3rd ed. New York: Oxford University Press, 1994:302–3.
12. Honein MA, Paulozzi LJ, Cragan JD, Correa, A. Evaluation of selected characteristics of pregnancy drug registries. Teratology 1999;60:356–64.
13. Leppig KA, Werler MM, Cann CI, Cook CA, Holmes LB. Predictive value of minor anomalies. I. Association with major malformations. J Pediatr 1987;110:531–7.
14. Centers for Disease Control and Prevention. Unintentional administration of varicella virus vaccine—United States, 1996. MMWR Morb Mortal Wkly Rep 1996;45:1017–8.
© 2001 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
15. Hagler L, Luscombe F, Siegfried J. A primer on postmarketing surveillance. Drug Information J 1987;21:71–107.