 |
|
|
| |
Prevalence of Congenital Anomalies in Infants With In Utero Exposure to Antiretrovirals
|
| |
| |
Download the PDF here
Download the PDF here
Pediatric Infectious Disease Journal Feb 2012
Knapp, Katherine M. MD*; Brogly, Susan B. PhD; Muenz, Daniel G. BA; Spiegel, Hans M. L. MD; Conway, Daniel H. MD; Scott, Gwendolyn B. MD; Talbot, Jeffrey T. MPH; Shapiro, David E. PhD,**; Read, Jennifer S. MD
From the *Department of Infectious Diseases, St. Jude Children's Research Hospital, Memphis, TN; Department of Epidemiology, Boston University School of Public Health, Boston, MA; Center for Biostatistics in AIDS Research, Harvard School of Public Health, Boston, MA; Henry M. Jackson Foundation for the Advancement of Military Medicine, Prevention Sciences Program, Division of AIDS (DAIDS), National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD; Department of Pediatrics, St. Christopher's Hospital for Children, Drexel University College of Medicine, Philadelphia, PA; Division of Pediatric Infectious Disease and Immunology, University of Miami Miller School of Medicine, Miami, FL; **Department of Biostatistics, Harvard School of Public Health, Boston, MA; and Pediatric, Adolescent, and Maternal AIDS (PAMA) Branch, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD.
Accepted for publication September 2, 2011.
Abstract
Background: Although use of efficacious interventions, including antiretrovirals (ARVs), has dramatically reduced the rate of mother-to-child transmission of human immunodeficiency virus, the safety of in utero ARV exposure remains of concern.
Methods: Data regarding 1112 infants enrolled in the International Maternal Pediatric Adolescent AIDS Clinical Trials Group protocol P1025 born between 2002 and 2007 were analyzed for this study. Congenital anomalies were classified based on the Metropolitan Atlanta Congenital Defects Program guidelines. Associations between congenital anomalies and timing of first in utero exposure to ARVs were evaluated by logistic regression analysis.
Results: Congenital anomalies were identified and confirmed in 61 of the 1112 infants, resulting in a prevalence of 5.49/100 live births (95% confidence interval, 4.22-6.99). Among the 80 anomalies identified, the organ systems involved included cardiovascular (n = 33), musculoskeletal (n = 15), renal (n = 9), genitourinary (n = 6), craniofacial (n = 4), and central nervous system (n = 2). First trimester exposure to efavirenz was associated with a significantly increased risk of congenital anomalies (odds ratio, 2.84; 95% confidence interval, 1.13-7.16). No significant associations were observed between exposure to other individual ARVs or classes of ARVs started at any time during pregnancy and infant congenital anomalies.
Conclusions: The observed rate of congenital anomalies in this cohort is higher than previously reported for the general population, but it is consistent with rates observed in other recent studies of children born to human immunodeficiency virus-infected women. Cardiovascular anomalies occurred most frequently. With the exception of a known teratogen (efavirenz), no statistically significant associations between in utero exposure to ARVs and congenital anomalies were identified.
Use of antiretrovirals (ARVs) for prevention of mother-to-child transmission (pMTCT) of human immunodeficiency virus (HIV) has been advocated since 1994,1 when the results of the groundbreaking Pediatric AIDS Clinical Trials Group (PACTG) protocol 0762 were made available. PACTG 076 demonstrated a reduction in MTCT transmission from 22.6% in the placebo group to 7.6% with use of a 3-part regimen of zidovudine.2,3 Zidovudine and other nucleoside reverse transcriptase inhibitors have putative fetal safety concerns based on the incorporation of these ARVs into human nuclear and mitochondrial DNA and the depletion of mitochondrial DNA in laboratory and animal studies.4-8 With the subsequent availability of additional ARVs, use of combination ARV regimens during pregnancy, whether for treatment of the mother's own HIV infection or for pMTCT, has been associated with lower rates of MTCT.9-11 It is now recommended that HIV-infected pregnant women receive combination ARV regimens with at least 3 agents, 1 of which should be a non-nucleoside reverse transcriptase inhibitor or a protease inhibitor.12 Although zidovudine has been used for years for pMTCT, there is relatively little experience with the newer ARVs, and concerns exist regarding potential adverse effects of in utero ARV exposure. In particular, efavirenz, a frequently-used non-nucleoside reverse transcriptase inhibitor, is not recommended for use by pregnant women (Food and Drug Administration Pregnancy Category D, positive evidence of fetal risk) based on evidence from nonhuman primate data and case reports of neural tube defects.12-16 As newer ARV classes and agents become available, and ARV regimens become more complex, continued surveillance of congenital anomalies of infants exposed to ARVs is vital.
International Maternal Pediatric Adolescent AIDS Clinical Trials Group protocol P1025 prospectively collected data on pregnant HIV-infected women and their infants. The objectives of this study were to estimate the prevalence of congenital anomalies in this population and to assess the association between in utero exposure to ARVs and congenital anomalies.
METHODS
P1025 Protocol
The International Maternal Pediatric Adolescent AIDS Clinical Trials Group protocol P1025 is a prospective, observational study designed to assess use and outcomes of ARVs during pregnancy and interventions for pMTCT (including ARV prophylaxis). Enrollment into P1025 began in October 2002 and is ongoing. HIV-infected women ³13 years of age were eligible for enrollment after the eighth week of pregnancy up to 14 days after delivery. Women diagnosed with HIV infection at the time of delivery or within 14 days after delivery were eligible to enroll up to 28 days after delivery. All infants born to enrolled mothers were eligible for enrollment. Infant protocol visits including physical examination were scheduled at birth, within the first 7 days of life, and at 2, 6, 16, 24, 36, and 48 weeks of age. Potential congenital anomalies were identified by physical examination at the study sites and/or through review of prenatal and neonatal records. Case report forms completed at each visit asked whether any congenital anomalies had been identified.
Study Population
The study population consisted of children born to HIV-infected women enrolled in P1025. Data available as of October 26, 2007 were analyzed. Eligible infants were those with an estimated delivery date on or before September 10, 2007, and for whom the congenital anomaly case report form had been submitted. To ensure that infant outcomes reasonably could be considered independent of one another, only infants who were singleton births and whose mothers had not participated in the protocol during previous pregnancies were included.
Outcome
Data for infants identified through computerized screening as having congenital anomalies recorded on clinical case report forms were reviewed by a panel of clinicians who were blinded to the mother's ARV exposure during pregnancy. Definitive classification of a congenital anomaly was made by clinician consensus, using the Metropolitan Atlanta Congenital Defects Program guidelines.17
ARV Exposure and Possible Confounders
Any ARV to which at least 1 infant with an identified congenital anomaly was exposed in utero was evaluated for this study. Because of known reports of adverse infant outcomes associated with in utero exposure to folate antagonists18-21 and selective serotonin-reuptake inhibitor (SSRI) antidepressants,22,23 and maternal substance use (including alcohol, tobacco, marijuana, cocaine, heroin, and methadone) during pregnancy,24,25 we assessed the occurrence of anomalies in relation to in utero exposure to these drugs.
Statistical Analysis
Group comparisons were performed using the χ2 test or Fisher exact test (when expected cell frequencies were small) for nominal variables and the Cochran-Mantel-Haenszel χ2 test for ordinal variables. A 2-sided P < 0.05 was defined as statistically significant. Logistic regression was used to estimate associations between the timing of initial in utero exposure to ARVs and the presence of an anomaly. For each ARV evaluated, exposure was categorized as either starting in the first trimester, starting in the second or third trimester, or unexposed. A priori, it was decided to consider all covariates with a univariate P ≤ 0.25 and folate antagonist use (regardless of P value) for inclusion in the multivariate model for each ARV using a stepwise selection algorithm. This algorithm selected only those covariates with a P ≤ 0.15 when entered into a model that also included the ARV exposure variable (and other covariates selected at earlier steps). Analyses were conducted using SAS 9.1 (SAS Institute, Cary, NC) and LogXact 8 (Cytel Inc, Cambridge, MA) software.
RESULTS
Derivation of the Study Population and Congenital Anomalies in This Population
Of 1306 enrolled infants, 1112 met inclusion criteria for this analysis (Fig., Supplemental Digital Content 1, http://links.lww.com/INF/A974). Congenital anomalies were identified in 61 of the 1112 infants. The prevalence of congenital anomalies in this population was 5.49/100 live births (95% confidence interval [CI], 4.22-6.99). There were 80 anomalies identified in these 61 infants, the most common of which were cardiovascular (33 of 80, 41.3%), followed by musculoskeletal anomalies (15 of 80, 18.8%) (Table, Supplemental Digital Content 2, http://links.lww.com/INF/A975).
Prevalence of Congenital Anomalies According to Maternal and Infant Characteristics
Table 1 shows the prevalence of congenital anomalies according to maternal and infant characteristics. There were no statistically significant differences in the presence or absence of congenital anomalies according to any maternal characteristic shown. The only infant characteristic associated with congenital anomalies was birth weight; those with low birth weight (<2500 g) were statistically significantly more likely to have congenital anomalies.
Association Between ARV Exposure and Presence or Absence of Congenital Anomalies
Table 2 shows the prevalence of congenital anomalies for ARV classes and individual drugs within each class by timing of first in utero exposure. Earliest ARV use during pregnancy was divided approximately evenly between first trimester and second/third trimester. There were only 12 children unexposed to ARVs in utero, and none of these was diagnosed with a congenital anomaly. Of those with first ARV exposure during the first trimester or second/third trimester, 6.4% and 4.8%, respectively, were reported to have congenital anomalies.
Table 2 also shows the results of the multivariate logistic regression analyses. Those covariates in Table 1 with a P ≤ 0.25 (ie, maternal characteristics: age, alcohol use, and SSRI use; infant characteristics: year of birth, birth weight, and HIV infection status) and folate antagonist use were entered into the stepwise selection algorithm. The same 4 covariates (maternal age at enrollment, infant year of birth, birth weight, and infant HIV infection status) were selected for the model for each ARV. In the multivariate logistic regression analyses, the only statistically significant association of an ARV with risk of congenital anomalies was for efavirenz exposure during the first trimester (adjusted odds ratio [OR], 2.84; 95% CI, 1.13-7.16). Of the 47 infants with first trimester exposure to efavirenz, 6 (12.8%) had congenital anomalies, which were identified as patent foramen ovale (N = 1), gastroschisis (N = 1), postaxial polydactyly, type A (N = 1), spina bifida cystica, Arnold-Chiari malformation (N = 1), talipes equinovarus (N = 1), and plagiocephaly (N = 1).
DISCUSSION
In this prospective study of outcomes in HIV-infected pregnant women and their infants, 80 congenital anomalies were identified in 61 of 1112 infants, yielding a congenital anomaly rate of 5.49/100 live births. First trimester efavirenz exposure was associated with a significantly increased risk of congenital anomalies (6 of 47, 12.8%).
Our findings support previous reports associating efavirenz exposure during the first trimester with congenital anomalies.26-28 Since efavirenz was first approved by the Food and Drug Administration in 1998, the package labeling has carried a warning about its use in pregnancy, based on animal data indicating an increased risk of developmental anomalies.26 Of the 20 infant cynomolgus monkeys exposed to efavirenz throughout gestation at levels equivalent to those seen in humans, 3 had congenital anomalies (anencephaly and unilateral anophthalmia, microphthalmia, and cleft palate), compared with none of the 20 monkeys in the control group. In 2005, Bristol-Myers Squibb changed the pregnancy category of efavirenz from category C (risk of fetal harm cannot be ruled out) to category D (positive evidence of fetal risk), based on 4 retrospective reports of neural tube defects (3 cases of meningomyelocele and 1 case of Dandy-Walker syndrome) in infants exposed to efavirenz during the first trimester.27 In a review of data from PACTG protocols 219 and 219C, investigators identified congenital anomalies in 5 of 32 infants (15.6%) exposed to efavirenz during the first trimester (adjusted OR, 4.31; 95% CI, 1.56-11.86), only 1 of which was a neural tube defect.28 In the 219/219C study, there was a congenital anomaly rate of 5.3/100 live births, very similar to the rate observed in our study. The data from 219/219C provide the first evidence of an association of efavirenz with congenital anomalies in a population-based study, although the number of infants with first trimester efavirenz exposure was small. It should be noted that some participants in PACTG 219C were also enrolled in P1025. However, only 15 cases overall were common to both studies, including 2 children exposed to efavirenz (a case of talipes equinovarus and a neural tube anomaly that had been previously reported).15
The prevalence of congenital anomalies in the P1025 study population (5.49/100 live births) is higher than that reported for the general population of the United States (approximately 3/100 live births).29 Some studies of congenital anomalies among HIV-exposed infants with intrauterine exposure to ARVs have shown prevalences more similar to that of the general population. (Table 3). For example, the prevalence of congenital anomalies among infants with in utero exposure to ARVs was 2.9% for the Antiretroviral Pregnancy Registry,34 2.8% to 3.1% for the National Study of HIV in Pregnancy and Childhood in the United Kingdom and Ireland,31 and 3.56% for the Women and Infants Transmission Study.32 Even lower prevalence of congenital anomalies has been reported from the European Collaborative Study (1.6%).30 However, other studies of HIV-infected women and their infants have documented higher rates of congenital anomalies.28,33 Our observed prevalence is similar to that of the more recent reports from the 219/219C study28 and from the National Institute of Child Health and Human Development International Site Development Initiative (NISDI) analysis of congenital anomalies among HIV-infected women in Argentina and Brazil.33 In the NISDI cohort, the overall rate of congenital anomalies for infants born to HIV-infected women was 6.16/100 live births (95% CI, 4.6-7.7). As in the current study, approximately 4% of the women in the NISDI cohort had first trimester exposure to efavirenz. In contrast to the P1025 and 219/219C data, there was no evidence of an association of efavirenz with congenital anomalies in the NISDI cohort.33
There are differences among the various studies of congenital anomalies in children exposed to ARVs in utero that may account for the varying prevalence rates reported. Of the studies mentioned, ours demonstrated the highest percentage of infants (46.5%) with earliest in utero exposure to any ARV during the first trimester. Congenital anomalies may be underreported in some of the studies. Differential ascertainment across the studies may explain differences in prevalence. In the 219/219C study, for which the cardiovascular system was the most common site for congenital anomalies, echocardiograms, most of which were required for clinical research, had been performed in 29% of all infants.28 Non-ARV drug exposure could account for differences in prevalence: for example, folate antagonist use was evaluated in the 219/219C study,28 the NISDI study,33 and in the current study, but was not specifically addressed in the other studies. The studies also differ in the number of infants evaluated.
As with the data from 219/219C,28 the most common anomalies in our study were cardiovascular, followed by musculoskeletal. In 219/219C, first trimester exposure to zidovudine was associated with a higher risk of cardiovascular anomalies and a lower risk of musculoskeletal anomalies. Because of limited power, we were not able to assess for such associations in this review.
To address possible confounding factors, multivariate models were adjusted for year of birth, maternal age, gestational age at birth, birth weight, and infant HIV infection status. Maternal substance use (alcohol, tobacco, marijuana, cocaine, heroin, and methadone) and in utero exposure to folate antagonists and SSRI antidepressants did not enter the multivariate models and were not significantly associated with presence of congenital anomalies in univariate analyses. Combination ARV regimens, whether used for treatment or for pMTCT, are recommended during pregnancy. It is not possible to adjust for all other ARV agents when assessing the association of a given ARV with infant congenital anomalies.
The results of this study are consistent with that of other recent cohort studies suggesting higher prevalences of congenital anomalies among infants born to HIV-infected women than have been reported for the general population. With the exception of a known teratogen (efavirenz), no statistically significant associations between in utero exposure to ARVs and congenital anomalies were identified. Our inability to identify other significant associations in this analysis may be due to lack of power; the limits of the 95% CIs for the adjusted ORs indicate the magnitude of associations that are unlikely based on our results (ie, those associations lying outside the CI). Further research is indicated to evaluate for possible confounding factors. Information available from voluntary retrospective reporting of anomalies is limited by selection bias, in that adverse outcomes may be overreported. The Antiretroviral Pregnancy Registry collects both prospective and retrospective data, and analyzes them separately. Clinicians caring for HIV-infected pregnant women are encouraged to report prospectively all women receiving ARVs during pregnancy to the Antiretroviral Pregnancy Registry (available at: www.APRegistry.com), whether anomalies are identified or not.
|
| |
|
|
|
|
|
