The prevalence odds ratios (95% confidence intervals [CIs]) are adjusted for univariate risk factors (eTable 1). The sum of birth defects from the subgroups is larger than the number of birth defects in overall analyses in which only the first birth defect was counted because some participants had birth defects from more than 1 subgroup. Data are plotted on a log scale.
Pasternak B, Hviid A. Use of Acyclovir, Valacyclovir, and Famciclovir in the First Trimester of Pregnancy and the Risk of Birth Defects. JAMA. 2010;304(8):859-866. doi:10.1001/jama.2010.1206
Author Affiliations: Department of Epidemiology Research, Statens Serum Institut, Copenhagen, Denmark.
Context Herpes simplex and herpes zoster infections are common and often treated with antiviral drugs including acyclovir, valacyclovir, and famciclovir. Safety of these antivirals when used in the first trimester of pregnancy is insufficiently documented.
Objective To investigate associations between exposure to acyclovir, valacyclovir, and famciclovir in the first trimester of pregnancy and risk of major birth defects.
Design, Setting, and Participants Population-based historical cohort study of 837 795 live-born infants in Denmark from January 1, 1996, to September 30, 2008. Participants had no diagnoses of chromosomal aberrations, genetic syndromes, birth defect syndromes with known causes, or congenital viral infections. Nationwide registries were used to ascertain individual-level information on dispensed antiviral drugs, birth defect diagnoses (categorized according to a standardized classification scheme), and potential confounders.
Main Outcome Measure Prevalence odds ratios (PORs) of any major birth defect diagnosed within the first year of life by exposure to antiviral drugs.
Results Among 1804 pregnancies exposed to acyclovir, valacyclovir, or famciclovir in the first trimester, 40 infants (2.2%) were diagnosed with a major birth defect compared with 19 920 (2.4%) among the unexposed (adjusted POR, 0.89; 95% confidence interval [CI], 0.65-1.22). For individual antivirals, a major birth defect was diagnosed in 32 of 1561 infants (2.0%) with first-trimester exposure to acyclovir (adjusted POR, 0.82; 95% CI, 0.57-1.17) and in 7 of 229 infants (3.1%) with first-trimester exposure to valacyclovir (adjusted POR, 1.21; 95% CI, 0.56-2.62). Famciclovir exposure was uncommon (n = 26), with 1 infant (3.8%) diagnosed with a birth defect. Exploratory analyses revealed no associations between antiviral drug exposure and 13 different subgroups of birth defects, but the number of exposed cases in each subgroup was small.
Conclusion In this large nationwide cohort, exposure to acyclovir or valacyclovir in the first trimester of pregnancy was not associated with an increased risk of major birth defects.
Acyclovir, valacyclovir, and famciclovir are antiviral agents used in the treatment of herpes simplex and herpes zoster infections. For genital and labial herpes, these antivirals are used as short-course treatment in primary infections and as episodic or chronic suppressive therapy for frequently recurring disease.1,2 In herpes zoster, antivirals effectively reduce acute symptoms as well as the risk for postherpetic neuralgia.3
The prevalence of herpes simplex is high4,5 and almost 40% of individuals with genital herpes experience at least 6 recurrences in the first year after disease onset,6 which is often the threshold for episodic or chronic suppressive treatment.1 More than 1% of susceptible women acquire herpes simplex during the first trimester of pregnancy7 and the incidence of herpes zoster is 1.5 to 2 per 1000 person-years in the third and fourth decades of life.8 Given this background, antiviral treatment will be indicated for a significant number of women in pregnancy.
Although the safety of acyclovir, valacyclovir, and famciclovir in general has been well established,9,10 data on the use of these antivirals in early pregnancy are limited. Animal studies, although of uncertain applicability to humans,11 did not demonstrate teratogenic effects initially, whereas later studies suggested that multiple defects may be induced with very high doses of acyclovir.12,13 The US Food and Drug Administration has classified acyclovir, valacyclovir, and famciclovir as category B drugs in pregnancy.14- 16 Information on the safety of acyclovir, with regard to teratogenicity in humans, is mainly based on data from a pregnancy registry managed by the drug manufacturer.17 This study reported that the rate of major birth defects in 596 pregnancies exposed in the first trimester was 3.2% and compared it to the rate expected (3.2%) in the general population. In addition to the lack of a valid control group, recruitment to this study relied on spontaneous reporting. Two other acyclovir studies were very limited in size.18,19 There are no published data on valacyclovir in early pregnancy and experience with famciclovir is limited to 4 pregnancies.19
We conducted a nationwide registry-based cohort study to assess associations between acyclovir, valacyclovir, and famciclovir use in the first trimester of pregnancy and major birth defects. Our primary objective was to investigate the risk of any major birth defects. In secondary explorative analyses, we examined risks in subgroups of major birth defects by organ system.
We used data from nationwide registries to conduct a historical cohort study including all infants born alive in Denmark from January 1, 1996, to September 30, 2008, and evaluated associations between exposure to oral acyclovir, valacyclovir, and famciclovir in the first trimester of pregnancy and major birth defects diagnosed within the first year of life. In supplementary analyses, we investigated associations between use of dermatological acyclovir and penciclovir creams and the risk of birth defects. Individual-level data were linked between registries using the unique personal identification number assigned to all inhabitants in Denmark. The study was approved by the Danish Data Protection Agency. Ethics approval is not required for registry-based research in Denmark.
The Medical Birth Register (MBR) contains individual-level information on all deliveries by women living in Denmark,20 including the personal identification numbers of the parents and the newborn, date of birth, multiple births, gestational age, and various physical characteristics of the newborn. On the basis of the MBR, we identified a cohort of all live births in Denmark from January 1, 1996, to September 30, 2008. The date of conception was calculated by subtracting gestational age from date of birth. For births in which gestational age was missing (0.9%), we imputed the cohort median of 280 days. The MBR records gestational age based on the last menstrual period and corrected by ultrasonographic measurements in most women.21 A validation study of gestational age registration in the MBR compared registered data with medical records and found that 87% of registrations were correct if agreement was defined as within 1 week.22
The Prescription Drug Register,23 established in 1995, contains individual-level information on all prescriptions filled at all Danish pharmacies. Each record contains the personal identification number of the patient, date of filling the prescription, anatomic therapeutic chemical (ATC) code, number of packages, package size, and number of daily defined doses in the prescription. We obtained information (ATC code) on oral acyclovir (J05AB01), valacyclovir (J05AB11), famciclovir (J05AB09), and dermatologic acyclovir (D06BB03) and penciclovir (D06BB06) prescriptions filled by cohort mothers from 4 weeks before conception until birth. Systemic acyclovir, valacyclovir, and famciclovir are prescription-only drugs in Denmark, while dermatological acyclovir and penciclovir creams have also been available over the counter since 1993 and 1999, respectively. We lacked information on inpatient antiviral drug treatment.
Cases of birth defects were identified through the National Patient Register (NPR),24 which contains individual-level information on hospital visits (emergency department and outpatient) and admissions including diagnoses assigned by clinicians according to the International Classification of Disease (ICD). The NPR does not contain diagnostic data from the primary care setting, which limits the detection of diagnoses to those that have been diagnosed in hospitals. We accessed NPR data covering January 1, 1996, to March 31, 2009. For multiple births, any child was considered as a potential case. Major birth defects were defined according to the EUROCAT (a European network for surveillance of congenital anomalies) classification for subgroups of major congenital anomalies.25 We modified the EUROCAT protocol for the purpose of our study: infants with diagnoses of chromosomal aberrations, genetic disorders, and birth defect syndromes with known causes (n = 2944), and congenital viral infections possibly associated with birth defects (n = 259) were identified and excluded (ICD codes are available in eAppendix). Minor defects were excluded from evaluation according to the EUROCAT exclusion list.26
From the MBR, the Central Person Register,27 and Statistics Denmark, we obtained information on birth year and the mother's parity, age at conception, smoking status during pregnancy, country or continent of origin, place of residence at the time of conception, and educational level and socioeconomic class in the year of conception. From the NPR and the Prescription Drug Register, we obtained information on maternal diseases and drug exposures that may be associated both with herpes simplex or herpes zoster and birth defects; infectious diseases in the first trimester; history of sexually transmitted infections, diabetes mellitus, and immunodeficiency; and filled prescriptions for antineoplastic and immunomodulating agents from 3 months before conception throughout the first trimester, oral glucocorticoids in the first trimester, and oral antibiotics in the first trimester (ICD and ATC codes in eAppendix). We also identified high-risk pregnancies and history of birth defects in siblings back to 1977. We did not have data on folic acid exposure.
We used logistic regression to estimate prevalence odds ratios (PORs) with 95% confidence intervals (CIs) comparing prevalence odds of major birth defects in infants from pregnancies exposed to antivirals and in infants from unexposed pregnancies (SAS software version 9.1, SAS Institute Inc, Cary, North Carolina). Results were considered statistically significant when the 95% CI did not overlap 1.0 in either direction. Potential confounders were included in regression models if they were significantly (P < .05; 2-sided) associated with major birth defects in univariate analyses. P values in univariate analyses were estimated with missing values excluded. In adjusted models, we applied multiple imputation for variables with missing values using the Markov Chain Monte Carlo method.
Children were observed for the first registered diagnosis of a major defect within 1 year after birth. For the primary outcome measure of all major birth defects, we combined all subgroups of malformations. In preplanned exploratory analyses we evaluated subgroups by organ system without correction for multiple comparisons. Any filling of an acyclovir, valacyclovir, or famciclovir prescription was considered as exposure. The main exposure time window comprised the first trimester (12 weeks). Analysis of exposure within 4 weeks before conception and in the second and third trimesters was performed for comparison. The timing of exposure was defined by the date of filling the prescription, and any pregnancy could contribute to any exposure time window. In case a pregnancy was exposed in more than 1 time window, the corresponding PORs (except for crude estimates) were adjusted for effects of each other. The study had 80% power to detect a 47% relative increase (POR, 1.47) in the risk of birth defects in those exposed to any antiviral (n = 1804; 5% 2-sided α level; EpiInfo, version 3.5.1, http://www.cdc.gov/epiinfo).
The cohort included 837 795 live births (34 787 multiple births), among whom 19 960 (2.4%) were diagnosed with a major birth defect during the first year of life. Table 1 and Table 2 present descriptive characteristics of cohort participants.
Among 1804 pregnancies exposed to acyclovir, valacyclovir, or famciclovir at any time in the first trimester, 40 infants (2.2%) had a diagnosis of a major birth defect, compared with 19 920 of 835 991 infants (2.4%) among the unexposed pregnancies (crude POR, 0.93; 95% CI, 0.68-1.27). Risk estimates for those potential confounders that were statistically significant risk factors for birth defects in univariate analyses are shown in eTable 1. Adjusting for these variables in a multivariate model, acyclovir, valacyclovir, or famciclovir exposure at any time in the first trimester was not associated with increased risk of major birth defects (POR, 0.89; 95% CI, 0.65-1.22), as compared with unexposed pregnancies. Table 3 presents risk estimates for the association between use of antivirals and major birth defects for the 3 antivirals together and individually in the different exposure time windows. First-trimester use of acyclovir, the most commonly prescribed antiviral, was not associated with major birth defects (32 cases among 1561 exposed [2.0%] vs 2.4% in the unexposed; adjusted POR, 0.82; 95% CI, 0.57-1.17). Neither valacyclovir nor famciclovir were associated with major birth defects, although use of the latter was very uncommon.
The Figure shows exploratory analyses of associations between use of antivirals and major birth defect subgroups by organ system. There was no significant increase in the prevalence of any major birth defect subgroup among mothers exposed to antivirals in the first trimester of pregnancy. Results were similar when restricted to acyclovir. However, these analyses were based on a small number of cases among the exposed and should be interpreted with caution. Exposure to any antiviral within 4 weeks before conception (eTable 2) was associated with increased risk of major heart defects (POR, 1.71; 95% CI, 1.05-2.79), while major eye defects (POR, 2.76; 95% CI, 0.98-7.76) and nervous system defects (POR, 2.37; 95% CI, 0.87-6.50) had relatively high PORs but were not statistically significant. There were no statistically significant associations between antiviral exposure and any other subgroup of major birth defects among those exposed within 4 weeks before conception or any subgroup of birth defects among those exposed in the second and third trimesters (eTable 2).
Supplementary analyses of acyclovir and penciclovir dermatological creams found no significant associations between exposure to these drugs and the risk of major birth defects (Table 4).
We conducted alternative analyses to test the robustness of our results. First, to account for the possibility that women who had started taking antivirals before conception may have continued use beyond conception, we performed analyses that accounted for the number of daily defined doses in individual antiviral drug prescriptions (daily drug intake and consumption of the entire drug package were assumed). Women who had filled antiviral prescriptions within 4 weeks before conception and received enough doses to have a theoretical chance of continued exposure beyond the day of conception were grouped with women who had filled prescriptions in the first trimester. The adjusted POR for having offspring with major birth defects in this group was almost identical to that of the main analysis (0.89; 95% CI, 0.65-1.21; 42 cases among 1896 exposed [2.2%] vs 2.4% unexposed). Mothers who had filled prescriptions within 4 weeks before conception but had not received enough doses to have a theoretical chance of continued exposure beyond conception had an adjusted POR similar to that of the main analysis (1.20; 95% CI, 0.87-1.67; 38 cases among 1339 exposed [2.8%] vs 2.4% unexposed). An analogous analysis of the heart defects subgroup, also accounting for daily doses in prescriptions, showed that an increased risk remained among those exposed within 4 weeks before conception (18 cases among 1339 exposed [1.3%] vs 0.8% in the unexposed; adjusted POR, 1.83 [95% CI, 1.14-2.95]), while exposure in the first trimester was not associated with heart defects (13 cases among 1896 exposed [0.7%] vs 0.8% in the unexposed; adjusted POR, 0.79 [95% CI, 0.45-1.39]).
Second, in analyses restricted to those who used antivirals exclusively in the first trimester (no prescriptions filled in the other 2 exposure time windows), antiviral exposure was not associated with major birth defects (31 cases among 1339 exposed [2.3%]; adjusted POR, 0.92 [95% CI, 0.65-1.32]).
Third, filling an antiviral prescription at any time in the period of maximal susceptibility to teratogenic agents, 2 to 8 weeks after conception,11 was not associated with increased risk of major birth defects (15 cases among 997 exposed [1.5%]; adjusted POR, 0.58; 95% CI, 0.35-0.98). Analyses restricted to acyclovir produced similar estimates (13 cases among 857 exposed [1.5%]; adjusted POR, 0.60; 95% CI, 0.34-1.04).
Fourth, as correlation among infants from the same pregnancy could affect risk estimates, we performed analyses excluding multiple births. Among 1738 singleton pregnancies exposed to any antiviral in the first trimester, 36 (2.1%) were diagnosed with a major birth defect (adjusted POR, 0.86; 95% CI, 0.62-1.21).
Fifth, because prepregnancy antiviral use may be associated with undetected reuse in pregnancy, we conducted an analysis restricted to mothers without a history of antiviral use (prescription history available until 9 months before conception on average) and without a history of sexually transmitted infections including anogenital herpes. There were 16 cases among 717 exposed to any antiviral in the first trimester (2.2%), compared with 18 940 cases among 797 922 unexposed (2.4%; adjusted POR, 0.89; 95% CI, 0.54-1.47).
Finally, as multiple imputation was used for variables with missing values, we wanted to address the potential for bias introduced by missing data. We therefore performed analyses restricted to cohort participants without any missing values among covariates included in the multivariate model. Adjusted PORs were very similar to those of the main analysis: 0.86 for any antiviral (95% CI, 0.62-1.21; 36 cases among 1657 exposed [2.2%]) and 0.83 for acyclovir alone (95% CI, 0.57-1.19; 30 cases among 1441 exposed [2.1%]).
This large nationwide historical cohort study found no association between exposure to acyclovir, valacyclovir, or famciclovir in the first trimester of pregnancy and the risk of any major birth defect. In analyses of individual antivirals, risk estimates were similar for acyclovir and valacyclovir. While analysis of valacyclovir was based on a limited number of exposed cases, results at least indicate that it is not a major human teratogen. Analyses of famciclovir were based on only 26 exposed pregnancies and should therefore not be viewed as evidence of safety of this drug. Exploratory analyses of subgroups of major birth defects by organ system showed no statistically significant associations with first-trimester exposure to antiviral drugs or acyclovir alone. Finally, supplementary analyses showed that exposure to dermatological acyclovir cream was not associated with major birth defects, whereas the number exposed to penciclovir cream was insufficient to draw firm conclusions.
Our results are in concert with previously published reports. A study using the acyclovir pregnancy registry found that the rate of major birth defects among 596 pregnancies exposed in the first trimester was 3.2%.17 This was compared with the expected rate in the general population (3.2%). However, the study lacked statistical testing and a proper control group. Further limitations included recruitment of exposed pregnancies by spontaneous reporting, which may have introduced selection bias, and a loss to follow-up of 27%. Surveillance for birth defects was restricted to the immediate postnatal period so that malformations diagnosed later would have been missed. A regional Danish study applied a methodology similar to ours, but was limited by a small sample size (n = 72) with respect to oral acyclovir.18 This study also included dermatological acyclovir exposure in early pregnancy (n = 474) and found no association with birth defects.18 A study using prescription data and questionnaire information on pregnancy outcomes found no birth defects in 22 pregnancies exposed to acyclovir or famciclovir.19 There are no published data on valacyclovir. Thus, previous studies have reported on fewer than 700 pregnancies exposed in the first trimester or just 72 if only analytical studies are considered, as compared with 1804 in our study.
Several strengths and limitations of our study deserve particular mention. The registry-based study design allowed nationwide coverage over 13 years, independent ascertainment of dispensed prescriptions and birth defect diagnoses, and a complete 1-year follow-up. Defects diagnosed after 1 year of age would have been missed. A series of sensitivity analyses demonstrated the robustness of our results for variable exposure definitions, including an analysis restricted to the period of maximal susceptibility to teratogenic agents 2 to 8 weeks after conception.11 Diagnostic information in the NPR has high validity, ie, registrations were correct for 88% of birth defect diagnoses overall28 and 89% of cardiac malformations29 while the completeness of registration was 90%.28 Our study did not include abortions. If drug-induced birth defects were associated with increased risk of planned or spontaneous abortion, results of the study would have been biased toward the null.
We included a wide range of potential confounders but evaluation of maternal comorbidity was incomplete because the NPR is restricted to the hospital setting. Furthermore, unmeasured confounding remains a possibility. Because of the size of the cohort, a factor masking a birth defect risk associated with antiviral use would have to be either common or the associations with both antiviral use and a reduced risk of birth defects would have to be very strong.
Use of filled prescriptions as a measure of drug exposure eliminates recall bias and increases the precision of the information on the type of drug used, as compared with interview data. A major limitation is, however, that nonadherence to the dispensed drugs would bias results toward no effect and obscure teratogenic effects, if present, among women who were adherent.
Exposure within 4 weeks before conception was included mainly for comparison. The observations of increased risk of heart defects and elevated, but nonsignificant, PORs for eye and nervous system defects in those exposed within 4 weeks before conception lack a biological explanation. Maximal vulnerability to teratogenic agents is typically confined to the period of organogenesis (2-8 weeks postconception) and heart defects most often arise between 6.5 and 8 weeks of gestation.11 The observations are therefore probably chance findings partly attributable to multiple comparisons of 13 subgroups in the different time periods or the result of unmeasured confounding.
The exploratory analyses of subgroups of major defects involved few exposed cases in each subgroup and therefore cannot exclude teratogenic effects with certainty. Risks of birth defects by subgroup and of specific defects should be investigated further in larger studies.
Acyclovir and penciclovir creams are also available as over-the-counter drugs and nonprescription use in those classified as unexposed in our study would therefore bias results toward no effect. Such a bias might also be conferred by inpatient antiviral exposure, which was not detected in the registries. Given the size of the group classified as unexposed, however, effects of exposure misclassification would be minimal.
Our study, to our knowledge the largest of its kind, found no significant association between first-trimester exposure to antiherpetic antiviral drugs and major birth defects. Consequently, it has immediate clinical implications and may support informed decisions on safety when prescribing antivirals for herpes infections in early pregnancy. Acyclovir is the most extensively documented antiviral and should therefore be the drug of choice in early pregnancy, while data on valacyclovir and famciclovir are still insufficient. Future research on antiherpetic antivirals and mother-child health should include safety studies with regard to spontaneous abortion and preterm birth, and during breastfeeding.
Corresponding Author: Björn Pasternak, MD, PhD, Department of Epidemiology Research, Statens Serum Institut, Artillerivej 5, 2300 Copenhagen S, Denmark (email@example.com).
Author Contributions: Dr Pasternak had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Pasternak, Hviid.
Acquisition of data: Hviid.
Analysis and interpretation of data: Pasternak, Hviid.
Drafting of the manuscript: Pasternak.
Critical revision of the manuscript for important intellectual content: Pasternak, Hviid.
Statistical analysis: Hviid.
Obtained funding: Hviid.
Study supervision: Hviid.
Financial Disclosures: None reported.
Funding/Support: This work was supported by grants from the Danish Medical Research Council and the Lundbeck Foundation.
Role of the Sponsor: The funding agencies had no role in the design and conduct of the study; in the collection, management, analysis, and interpretation of the data; or in the preparation, review, and approval of the manuscript.