Ransom, Carla E. MD*; Huo, Yanling MS†; Patel, Kunjal DSc, MPH†,‡; Scott, Gwendolyn B. MD§; Watts, Heather D. MD‖; Williams, Paige PhD†; Siberry, George K. MD, MPH‖; Livingston, Elizabeth G. MD¶; for the P1025 Team of the International Maternal Pediatric Adolescent AIDS Clinical Trials Group
Routine use of combination antiretroviral (ARV) drug regimens in pregnancy has resulted in a decline in the rate of maternal to child transmission of HIV from >20% to <1%.1,2 Current US guidelines recommend that HIV-infected pregnant women receive a 3-drug regimen of 2 nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs) and either a non-NRTI (NNRTI) or a protease inhibitor (PI).3 Current World Health Organization (WHO) guidelines are similar, recommending a 3-drug regimen of 2 NRTIs and an NNRTI.4 The preferred NRTIs in pregnancy are zidovudine and lamivudine. Use of tenofovir disoproxil fumarate (TDF), an NRTI and preferred drug in nonpregnant adults,5 has been increasing in pregnancy6 despite its recommendation as an alternative drug in pregnancy due to concerns about its potential adverse effects in the infant.
Human data on perinatal exposure to tenofovir include small case series, retrospective datasets, and prospective data from the Antiretroviral Pregnancy Registry and from the Pediatric HIV/AIDS Cohort Study network. In a small cohort study, Nurutdinova did not find any congenital malformations or growth abnormalities in 14 babies followed to 12 months who were exposed to tenofovir.7 In a study by Eastwood, TDF-exposed pregnancies were associated with efficacious viral suppression and lower risk for cesarean delivery for HIV viremia compared with those not exposed to TDF, with no increased risk for adverse neonatal outcome, including no difference in birth weight, preterm birth, or neonatal morbidity.8 Among prospective cases reported to the Antiretroviral Pregnancy Registry, there was no overall increase in birth defects among infants exposed to tenofovir (n = 1092) in the first trimester compared with later exposure (n = 782), and compared with baseline population risk.9,10 Pharmacokinetics studies demonstrate about 60% placental transfer of tenofovir.11
Long-term follow-up data on the effect of tenofovir on neonatal and infant growth are sparse and conflicting. Analysis from the Safety Monitoring for ART Toxicity (SMARTT) study of the Pediatric HIV/AIDS Cohort Study reported that children born to women who had received TDF as part of their highly active antiretroviral therapy (HAART) regimen during pregnancy were more likely to have lower length-for age and head-circumference-for-age z-scores at 1 year,6 despite no observed difference in growth measurements at birth. A single case report suggested fetal growth restriction with exposure.12 This is in contrast to the Development of AntiRetroviral Therapy in Africa trial, which in a subgroup analysis showed no effect of intrauterine tenofovir exposure on growth outcomes at birth through infancy.13 Overall, the literature is limited in scope on neonatal and infant growth outcomes after tenofovir exposure in utero. In short, the question on whether maternal use of TDF in pregnancy has adverse effects on infant growth remains unsolved. After initial reports from the SMARTT study suggested impaired infant growth (not confirmed in the final analysis),6 we began our study on a similar population. We hypothesized that maternal use of TDF during pregnancy would predict impaired growth during infancy. Our primary objective was to evaluate the impact of in utero tenofovir exposures on growth during infancy by determining whether tenofovir exposure during pregnancy is an independent predictor of birth weight and infant weight at 6 months of age, representing a delayed effect on infant growth. This is an important question to examine. With the increasing use of TDF in pregnancy despite limited safety data in the literature, large studies like ours are desperately needed to either confirm the safety profile in pregnancy or illuminate safety concerns, which may temper its use.
The study population included women enrolled in the International Maternal Pediatric Adolescent AIDS Clinical Trials Group protocol 1025, a prospective observational cohort study designed to assess the use and safety of ARVs in HIV-infected pregnant women and their infants. Approximately 37% of P1025 participants overall and 45% of TDF-exposed infants were also co-enrolled in the SMARTT study.6 Beginning in October 2002, HIV-infected pregnant women were enrolled if they were at least 13 years old and between 14-week gestation and 14-day postpartum. Since December 2007, enrollment was allowed as early as 8 weeks gestation. Institutional review boards approved the protocol at all 56 clinical sites located in the United States and Puerto Rico, and written informed consent was obtained from those enrolled. The population eligible for this study consisted of all liveborn singleton infants with an estimated date of confinement on or before September 15, 2011. Infants found to be HIV infected were excluded. For this analysis, only the first eligible pregnancy with a live-birth > 23 weeks of gestation was included for women enrolled for more than 1 pregnancy. This was done to avoid the need to account for correlation between multiple infants born to the same mother and simplify the analysis and interpretation of data. Only infants whose mothers used a combination ARV regimen (triple ARV regimen or greater) during pregnancy were included. Birth weight and infant weight at 6 months (±1 month) were utilized for this analysis.
ARV Classification and Covariates
Sociodemographic characteristics, obstetrical history, substance use information, maternal ARV use, and laboratory testing results were collected through medical chart abstraction or self-administered questionnaires. Potential covariates of interest included maternal viral load, CD4+ lymphocyte count, and Centers for Disease Control and Prevention (CDC) class at delivery, pregestational and gestational diabetes and hypertension, smoking during pregnancy, body mass index closest to delivery, age, race, ethnicity, education, and year of delivery.
Outcomes of Interest
Infant birth weight was defined in 4 different ways: absolute weight (continuous); gestational age- and sex-adjusted weight z-score (continuous); gestational age- and sex-adjusted weight z-score <fifth percentile (<−1.645), and small for gestational age (gestational age- and sex-adjusted weight z-score < 10th percentile). The Olsen growth curves14 were used to calculate the z-score for all the infants (preterm and term) per completed gestational age and sex. Because of the absence of growth curves at 42 weeks of gestation, the CDC standards for the full term infants were used for the infants born at 42 weeks.
Infant weight at 6 months of age was defined in 3 different ways: absolute weight (continuous); age- and sex-adjusted weight z-score (continuous); and age- and sex-adjusted weight z-score <fifth percentile (<−1.645), chosen because we felt this would be a clinically significant measure of weight at 6 months. The CDC growth charts were employed to calculate the z-score for term infants at their exact age at measurement (gestational age ≥37 weeks).15 For preterm infants (gestational age < 37 weeks), the exact age at the weight measurement was adjusted by subtracting the difference between 40 weeks and the gestational age at birth. The CDC growth standards were then applied to calculate the z-score based on the adjusted age. For weight at both birth and 6 months of age, several outcome definitions were chosen in an attempt to detect any signal that maternal TDF use in pregnancy would have toxic effects on infant growth.
Maternal TDF use during pregnancy was defined in 3 ways: any TDF use during pregnancy, duration of TDF use during pregnancy (no use vs. < 4 weeks vs. 4–12 weeks vs. > 12 weeks of use), and the earliest trimester in which TDF was used during pregnancy (no use vs. first trimester vs. second trimester vs. third trimester). The categories for TDF duration were chosen a priori based on clinical experience on what constituted short, medium, and prolonged exposure to TDF.
Multivariable regression models were built to evaluate whether maternal TDF use predicted infant weight measurements (general linear model for continuous weight outcomes and logistic regression model for binary weight outcomes), independent of potential covariates. The potential covariates were identified separately for each outcome from bivariable analyses with P < 0.10. For each weight outcome, a separate regression model was fit including each of the 3 TDF exposures and the same set of potential covariates. Other covariates included in any of the multivariable regression models: birth weight: race, ethnicity, mode of delivery, parity, last CD4 count during pregnancy, last RNA during pregnancy, last CDC classification during pregnancy, maternal obesity, pregestational diabetes, gestational diabetes, chronic hypertension, use of antihypertensive medication during pregnancy, use of diabetes medication during pregnancy, preeclampsia, smoking during pregnancy; weight at 6 months of age: race, ethnicity, education, mode of delivery, preeclampsia. SAS Version 9.2 (SAS Institute Inc, Cary, NC) was used to conduct all statistical analyses, and 2-sided P < 0.05 was considered statistically significant.
As of October 1, 2011, 2477 live-born infants with estimated date of confinement at or before September 15, 2011, enrolled in the P1025 study. After excluding infants who were HIV infected (n = 14), had HIV tests that were pending (n = 10), were born at or before 23 weeks of gestation (n = 1), or not from a singleton birth (n = 44 sets of twins; n = 2 sets of triplets), the remaining 2358 infants were further reduced to only include the infants from their mothers' first enrollments on study for a total of 2161 eligible infants in this analysis. Of the 2145 infants for whom maternal TDF use data were available, 2099 were exposed to a combined ARV regimen during pregnancy. The analyses were performed on the 2025 infants with birth weight available and 1496 infants with weight at 6 months of age available (Fig. 1).
Demographics, Growth, and TDF Exposure
Of 2025 infants with birth weight data available, 630 (31%) were exposed to tenofovir (Table 1). The median duration of TDF exposure among those exposed was 22.9 weeks. Compared with women unexposed to TDF, women exposed to TDF in pregnancy were more likely to be on a cARV regimen containing a PI (91% vs. 79%, P < 0.001), more often had CDC classification C (18% vs. 8%, P < 0.001), lower last CD4+ counts (CD4 < 200 copies/mm3: 15% vs. 8%, P < 0.001), and higher last RNA (RNA > 400 copies/mL: 19% vs. 15%, P = 0.03) before delivery, and were older (age at delivery ≥35 years: 21% vs. 15%, P < 0.001), but were otherwise similar. For infants exposed to tenofovir, 66% were exposed in the first trimester and 74% were exposed for at least 12-week duration (Table 2). Of the 98 and 225 infants with gestational age- and sex-adjusted birth weight z-scores <fifth percentile and <10th percentile, respectively, approximately 30% were exposed to TDF in utero (Table 2). Other characteristics of the study population are provided in Table 1.
The adjusted associations of maternal TDF use with birth weight are presented in Table 3 (weight in kg and z-scores) and in Table 4 (low birth weight). No significant associations were observed between TDF exposures (any TDF exposure, trimester of the first reported TDF use, and duration of TDF exposure) and the weight outcomes of interest. As expected, women who had a higher last CD4 count during pregnancy (CD4 <200 vs. ≥350 cells/mm3), women defined as obese (body mass index >40) during pregnancy, those who had at least 1 previous birth with >20 weeks of gestation, and those who had a pregestational or gestational diabetes diagnosis, were more likely to deliver a heavier baby as measured by 1 or more of the weight outcomes of interest compared with women who did not have these characteristics. In contrast, black women, women who used antihypertensive medications during pregnancy, women who had a preeclampsia diagnosis, and those who smoked cigarettes during pregnancy were more likely to deliver a baby with lower weight as measured by 1 or more of the weight outcomes of interest compared with women who did not have these characteristics.
Of 1496 infants with weight data available at 6 months of age, 457 (31%) were exposed to tenofovir (Table 2). The median duration of TDF exposure among those exposed was 21.3 weeks. For infants exposed to TDF in this group, 66% were exposed in the first trimester and 72% were exposed for >12-week duration. Of the 61 infants with age- and sex-adjusted 6 month weight z-scores <fifth percentile, 38% were exposed to TDF in utero. The group of women with infants followed for 6 months was not different than the entire group with infant birth weight available with respect to measured demographic and clinical information.
The multivariable associations of TDF exposures with weight at 6 months are presented in Tables 3 and 4. In comparison to mothers with no TDF exposure, maternal initiation of TDF during the second/third trimester was predictive of low infant weight at 6 months of age based on age- and sex-adjusted weight z-score <fifth percentile. No other significant associations were observed between TDF exposures (any TDF exposure, trimester of the first reported TDF use, and duration of TDF exposure) and other weight measures (absolute weight, age- and sex-adjusted weight z-score). Women who completed at least a high school education were more likely to have a heavier baby in terms of absolute weight and age- and sex-adjusted weight z-score at 6 months of age; babies whose mother had a preeclampsia diagnosis were more likely to have a lower mean absolute weight at 6 months of age than those whose mother did not have a preeclampsia diagnosis. Hispanic women were less likely to have a baby with age- and sex-adjusted weight z-score below the fifth percentile at 6 months of age, compared with non-Hispanic women.
No significant associations were observed between the TDF exposures and the preterm birth outcome. Of the 2099 infants exposed in utero to a combined ARV regimen, there were no differences in preterm birth among infants whose mothers were exposed to a TDF regimen compared with those who were not [117 (18%) vs. 232 (16%), P = 0.26]. Nonobese women, women on antihypertensive therapy during pregnancy, women with preeclampsia, and women with a last reported CD4+ lymphocyte count <200 cells per cubic millimeter during pregnancy were more like to deliver a premature baby.
With the increasing use of TDF by HIV-infected pregnant women, studies examining the safety profile of this drug in exposed neonates are crucial. In this large cohort of infants born to HIV-infected women receiving combination ARV regimens during pregnancy, in utero exposure to tenofovir does not seem to be associated with either infant birth weight or infant growth through 6 months of age. Although there was a marginal association with being underweight at 6 months of age in women who initiated TDF in the second or third trimester, there was no association between duration of maternal TDF use and 6-month weight outcomes. In addition, no association was found with absolute weight or overall age- and sex-adjusted z-score. As expected, women who were obese, diabetic, parous, and without hypertensive diseases of pregnancy tended to have larger babies, both at birth and at 6 months of age.
Although we feel it may not be clinically significant, the isolated finding of lower age- and sex-adjusted weight z-score of <fifth percentile at 6 months of age in infants of women who initiated TDF in the second or third trimester should not be completely overlooked. There are several studies in humans that suggest a delay in effect from ARV exposure not occurring until several months to years after exposure. For instance, febrile seizures were significantly more common in ARV-exposed compared with ARV-unexposed HIV-exposed infants; this effect did not seem until 6–12 months of age.16 In the Women and Infants Transmission Study, the significant difference in CD8+ cell counts by ARV exposure did not seem until 6–24 months.17 In the SMARTTs trial, data at 1 year but not at birth demonstrated lower mean length-for-age and head-circumference-for-age z-scores associated with maternal TDF use.6 More encouraging are results from the Development of AntiRetroviral Therapy in Africa trial,13 which recently reported no evidence that TDF effects growth of infants up to 2 years of age, in a cohort of 226 live births. The clinical relevance and underlying biologic mechanisms of this finding in our study are uncertain. Long-term studies are needed to determine whether this short-term effect on growth has any long lasting impact in childhood or adult life.
Our study has several strengths. We had a large sample size, with large (31%) proportion of infants having tenofovir exposure. In addition, we were able to examine growth in several different ways (absolute weight, underweight, and small for gestational age). We were also able to examine tenofovir exposure by several different measures, to determine if exposure in the first trimester had an effect or if cumulative dose of tenofovir had an effect. Neither seems to have affected fetal or infant growth.
There are, however, some limitations to our study. Tenofovir is eliminated by the kidneys. Renal toxicity with Fanconi syndrome and nephrogenic diabetes insipidus has been reported in children18 and adults19–22 exposed to TDF. There have also been reports of bone toxicity in infant rhesus macaques.23,24 Because of the rarity of these outcomes in infants and reporting bias, our study was unable to examine the effects of tenofovir exposure on either the kidneys or bones. It is reassuring that long-term safety data in rhesus macaques show the renal toxicity to be dose related, with no increased risk of congenital anomalies and good long-term safety data for the offspring.25 However, examining the renal effects of tenofovir exposure in utero is an unmet research need.
Like any cohort study, there is potential for selection bias in the nonrandom allocation of women to TDF as part of their multiagent HAART. In addition, more women exposed to TDF were on a PI containing HAART regimen than those not exposed to TDF, which may lead to confounding in results. Approximately 26% of the infants with birth weight available do not have available 6-month weight values to study. In addition, we were not able to examine growth beyond 6 months. Another limitation is that we only examined growth as a function of weight and not length or head circumference. We recognize this is an important limitation, especially in light of findings from the SMARTT study demonstrating adverse effects on length-for age and head-circumference-for-age z-scores at 1 year,6 despite no observed difference in growth measurements at birth. We also do not have data on maternal adherence to drug therapy, and thus true fetal exposure. Current guidelines from the Panel on Treatment of HIV-Infected Pregnant Women and Prevention of Perinatal Transmission assign TDF as an alternate agent (not first line) for the treatment of acute HIV infection in pregnant women, whereas it is considered first line treatment in other groups. The need for long-term data on TDF is vital given the rising use in a pregnant population. Long-term growth and developmental outcomes are still needed in children exposed to tenofovir in future large cohorts. There are many challenges to obtaining long-term safety data on drugs used to treat HIV in pregnancy. This population has many socioeconomic and cultural barriers to seeking and maintaining care, there is difficulty in monitoring drug adherence, and this group is often exposed to multiple drug regimens. Although we acknowledge that 45% of TDF-exposed infants were also co-enrolled in the SMARTT study,6 we feel our study still significantly contributes to the literature on this population given the sparse literature on the topic, the challenges to obtaining this information, and the growing use of TDF in clinical obstetric practice. The results of our study present an overall reassuring picture for both fetal growth and neonatal growth as reflected by weight in neonates exposed to TDF in pregnancy. Future studies should be designed to focus on rare neonatal outcomes (fracture, neutropenia, etc.) and long-term delayed effects on infant growth.
P1025 Team Acknowledgment: R. Tuomala, MD, Brigham and Women's Hospital, Boston, MA; E. Smith, MD, National Institute of Allergy and Infectious Diseases Division of AIDS, Pediatric Medicine Branch, Bethesda, MD; K. M. Oden, MHS, International Maternal Pediatric Adolescent AIDS Clinical Trials Group, Silver Spring, MD; D. Kacanek, ScD, Harvard School of Public Health, Boston, MA; E. Leister, MS, Harvard School of Public Health, Boston, MA; D. E. Shapiro, PhD, Harvard School of Public Health, Boston, MA; E. A. Barr, CPNP, CNM, MSN, University of Colorado Denver, The Children's Hospital, Denver, CO; D. W. Wara, MD, University of California at San Francisco, San Francisco, CA; A. Bardeguez, MD, MPH, FACOG, University of Medicine and Dentistry of New Jersey, Newark, NJ; S. K. Burchett, MD, MSc, Harvard Medical School, Boston, MA; J. Guiterrez, MD, Bronx-Lebanon Hospital, Bronx, NY; K. Malee, PhD, Ann and Robert H. Lurie Children's Hospital of Chicago, Chicago, IL; A. M. Stek, MD, Keck School of Medicine, University of Southern California, Los Angeles, CA; P. Tanjutco, MD, Washington Hospital Center, WA, DC; Y. Bryson, MD, David Geffen School of Medicine, University of California, Los Angeles, CA; M. T. Basar, BS, Frontier Science & Technology Research Foundation, Inc, Amherst, NY; A. Hernandez, MA, Frontier Science & Technology Research Foundation, Inc, Amherst, NY; A. Jennings, BS, Frontier Science & Technology Research Foundation, Inc, Amherst, NY; T. R. Cressey, PhD, BSc, Program for HIV Prevention & Treatment, Chang Mai, Thailand; J. Bryant, MPA, Westat, Rockville, MD.
Obstetrics Site Support: 2802 NJ Medical School Clinical Research Site (CRS) (A. D. Bardeguez, MD, MPH; L. Bettica, RN; C. Calilap-Bernardo, RN); 3601 University of California Los Angeles (UCLA)-Los Angeles, CA/Brazil AIDS Consortium (LABAC) CRS; 3801 Texas Children's Hospital CRS; 4001 Chicago Children's CRS; 4101 Columbia IMPAACT CRS (A. Higgins, RN; G. Silva, RN; S. Gaddipati, MD); 4201 University of Miami Pediatric/Perinatal HIV/AIDS CRS (S. Yasin, MD, C. Mitchell, MD, S. Lo Wong, MD, P. Bryan, RN); 4601 University California, San Diego Maternal, Child, and Adolescent HIV CRS (S. A. Spector, MD; A. Hull, MD; M. Caffery, RN, MSN; J. Manning, RN, BSN); 4701 Duke University Medical Center Pediatric CRS (E. Livingston, MD; M. Donnelly, PA; J. Wilson, RN; J. Giner, RN); 5003 Metropolitan Hospital NICHD CRS; 5009 Children's Hospital of Boston NICHD CRS (N. Karthas, RN, MS, CPNP; L. Tucker, BFA; A. Buck, RN; C. Kneut, RN, MS, CPNP); 5011 Boston Medical Center Pediatric HIV Program NICHD CRS; 5012 New York University NICHD CRS (S. Deygoo, MS; A. Kaul, MD; M. Minter, RN; S. Akleh, RN; Supported in part by Grant UL1 TR000038 from the National Center for the Advancement of Translational Science (NCATS), National Institutes of Health); 5013 Jacobi Medical Center Bronx NICHD CRS; 5015 Children's National Medical Center WA DC NICHD CRS; 5017 Seattle Children's Hospital CRS (A. Robson; J. Hitti, MD; C. Venema-Weiss, CNM; A. Klastorin, MSW); 5018 University of South Florida, Tampa NICHD CRS (K. L. Bruder, MD; G. Lewis, RN; D. Casey, RN); 5023 Washington Hospital Center NICHD CRS (S. Parker, MD; R. Scott, MD; P. Tanjutco, MD; V. Emmanuel, BA); 5031 San Juan City Hospital PR NICHD CRS (A. Mimoso, MD; R. Diaz, MD; E. Perez; O. Pereira); 5040 State University of New York Stony Brook NICHD CRS (J. Griffin, NP; P. Ogburn, MD); 5041 Children's Hospital of Michigan NICHD CRS; 5044 Howard University Washington, DC, NICHD CRS; 5045 Harbor UCLA Medical Center NICHD CRS; 5048 University of Southern California, Los Angeles, CA, NICHD CRS (A. Stek, MD; F. Kramer, MD; L. Spencer, MD; A. Kovacs, MD); 5051 University of Florida College of Medicine, Jacksonville NICHD CRS (M. Rathore, MD; I. Delke, MD; G. Thomas, RN; B. Millwood, RN); 5052 University of Colorado, Denver NICHD CRS (A. Katai, MHA; T. Kennedy, FNP-BC; K. Kinzie, FNP-BC; J. Wallace, MSW; Supported by NIH/NCATS Colorado CTSI Grant UL1 TR000154); 5055 South Florida CDC, Ft Lauderdale NICHD CRS; 5083 Rush University Cook County Hospital, Chicago NICHD CRS (J. Schmidt, MD; H. Cejtin, MD; M. McNichols, RN, MSN, CCRC; J. Senka, RN); 5091 University of California, San Francisco NICHD CRS (D. Cohan, MD; This publication was supported by NIH/NCRR UCSF-CTSI Grant UL1 RR024131. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH); 5092 Johns Hopkins University, Baltimore NCHD CRS (J. Anderson, MD; E. Sheridan-Malone, RN); 5093 Miller Children's Hospital Long Beach, CA, NICHD CRS (C. Tolentino-Balbridge, RN; Janielle Jackson-Alvarez, RN; D. Michalik, DO; J. S. Batra, MD); 5094 University of Maryland Baltimore NICHD CRS (D. Watson, MD; M. Johnson, DDS; C. Hilyard); 5095 Tulane University, New Orleans NICHD CRS (R. Maupin, MD; C. Dola, MD; Y. Luster, RN; S. Bradford, RN); 5096 University of Alabama Birmingham NICHD CRS (A. Tita, MD; M. Parks, CRNP; S. Robbins, BA); 6501 St Jude/UTHSC CRS (E. Thorpe, Jr, MD; K. Knapp, MD; P. Finnie, MSN; N. Sublette, RN, PhD); 6601 University of Puerto Rico Pediatric HIV/AIDS Research Program CRS (C. D. Zorrilla, MD; V. Tamayo-Agrait, MD); 6701 The Children's Hospital of Philadelphia IMPAACT CRS; 6901 Bronx-Lebanon Hospital IMPAACT CRS (R. Wright, MD); 7301 WNE Maternal Pediatric Adolescent AIDS CRS (S. Cormier, RN; K. Luzuriaga, MD; Supported by CTSA Grant 8UL1TR000161).