Extent of In Utero Transfer of Tenofovir From Mother to Fetus: A Paired Analysis of Hair Specimens Collected at Birth From a Cohort in the United States

Abstract

Understanding in utero transfer of antiretrovirals is critical for interpreting safety. Hair levels measure cumulative exposure. We measured tenofovir (TFV) concentrations in hair at delivery among women living with human immunodeficiency virus receiving TFV disoproxil fumarate-based treatment and their infants, using liquid chromatography–tandem mass spectrometry. Among 103 mother-infant pairs, the mean log₁₀ ratio of infant-to-maternal TFV levels was 1.08 (95% confidence interval, .97–1.20). TFV transfer was 60% lower from mothers who had preterm compared with term deliveries and 42% lower from mothers who had cesarean compared with vaginal deliveries. Like prior studies assessing transfer via short-term measures (plasma, cord blood, amniotic fluid), we found high cumulative transfer using hair.

We measured tenofovir concentrations in hair at delivery among women with human immunodeficiency virus receiving tenofovir disoproxil fumarate-based treatment and their infants. Like prior studies assessing transfer via short-term measures (plasma, cord blood, amniotic fluid), we found high cumulative transfer.

Despite significant strides in preventing vertical human immunodeficiency virus (HIV) transmission, 160 000 infants acquired HIV in 2018 worldwide, mainly in low- and middle-income countries [1]. Tenofovir (TFV)–containing combination antiretroviral (ARV) therapy (ART) is currently recommended as first-line HIV treatment [2, 3]. TFV disoproxil fumarate (TDF) in a fixed-dose combination with emtricitabine/TDF is recommended for preexposure prophylaxis to prevent HIV acquisition in populations at risk, including pregnant women in HIV high-burden settings [2]. As global efforts intensify to eliminate vertical HIV transmission and prevent HIV, the likelihood of women using TDF during pregnancy is increasing.

TDF exposure during pregnancy appears generally safe in terms of perinatal and maternal outcomes [4]. However, recent data suggest that in utero TFV exposure may stunt infant bone growth or trigger other adverse outcomes [4]. Understanding the pharmacokinetics of in utero TFV transfer is critical for interpreting safety. Limited prior studies have evaluated in utero TFV transfer by using measures—such as TFV levels in maternal plasma, cord blood, amniotic fluid, and meconium [5–7]—that reflect short-term TFV exposure in the infant. These studies all show high rates of maternal-to-infant TFV transfer.

As a monitoring matrix, neonatal hair can quantify cumulative prenatal exposure to maternal medications in utero [8]; neonatal hair begins to form at approximately 20 weeks gestation [9] and, similar to adult hair, incorporates drugs from the circulation over time [10]. To date, hair metrics have not been harnessed to quantify TFV exposure in utero. In the current study, using data from the Surveillance Monitoring for ART Toxicities (SMARTT) Dynamic Cohort, a prospective cohort of pregnant women living with HIV conducted by the Pediatric HIV/AIDS Cohort Study (PHACS) network in the United States, we sought to evaluate in utero TFV transfer by measuring maternal and infant TFV hair levels at birth.

METHODS

The PHACS SMARTT study was designed to assess safety of in utero exposure to ARVs and inform treatment guidelines for pregnant women living with HIV [11]. The SMARTT Dynamic study enrolled from a network of 18 clinical sites in the United States, including Puerto Rico, and all pregnancies with a delivery date between 1 June 2014 and 1 July 2016 were eligible for the hair substudy. Pregnancies with no reported maternal ART use during pregnancy, no maternal consent for hair collection, or lack of maternal scalp hair were ineligible. Pregnancies in which the woman had reported use of TDF-based regimens at some point during pregnancy were eligible for the current analysis. Within this group, pregnancies with available paired maternal and infant hair specimens were included.

The SMARTT protocol was approved by institutional review boards at each of the participating sites and the Harvard T.H. Chan School of Public Health; all participants provided written informed consent. Hair substudy procedures were also approved by the University of Washington and the University of California, San Francisco, Committees on Human Research.

At study entry, all infants were HIV exposed, yet uninfected per SMARRT eligibility criteria [11] and maternal ART data, including start and stop dates, were abstracted from clinical records. Birth characteristics (gestational age, mode of delivery) and maternal HIV parameters during pregnancy (plasma HIV RNA concentration and CD4 T-cell counts) were also abstracted. Unsuppressed viral load was defined as an HIV RNA level ≥400 copies/mL. Demographic characteristics were obtained by self-report.

Small hair samples (approximately 100 strands) were collected at a single time point at or shortly after childbirth among willing women living HIV and their infants. We assessed maternal and infant TFV exposure during the last month of pregnancy by cutting each hair sample down to the proximal 1 cm (reflecting 1 month’s worth of growth). Infant hair was washed with distilled water and air dried before analyses. TFV hair concentrations were analyzed using validated liquid chromatography/tandem mass spectrometry methods at the University of California, San Francisco, Hair Analytical Laboratory. The assay is validated from 0.002–0.400 ng of TFV per milligram of hair and was peer reviewed and approved by the National Institutes of Health’s Clinical Pharmacology and Quality Assurance program. We first assayed maternal hair specimens. Infant hair specimens were not assayed (and the pairs excluded from all analyses) if the mother’s level was below the lower limit of quantification (LLQ), because these are not compatible with any amount of transfer.

Descriptive statistics summarized characteristics of mother-infant pairs with TFV concentrations available. Weight-normalized TFV hair concentrations and the ratio of infant-to-maternal levels were log₁₀ transformed. Pairs with infant hair levels below the LLQ were treated as left-censored at log(LLQ/maternal concentration). For pairs with hair concentrations above the upper limit of quantification (ULQ), the ULQ for TFV (0.400 ng/mg) was used in analyses.

The distribution of TFV concentrations in maternal and infant hair, along with the infant/maternal hair concentration ratio, were summarized descriptively. We assessed the degree of in utero transfer of TFV from mother to infant by calculating ratios of infant to maternal hair concentrations for each mother-infant pair. Scatterplots and Spearman correlation coefficients were used to assess the extent of correlation between maternal-infant hair concentrations.

To explore covariates associated with transfer, we used univariable linear regression to estimate the absolute difference and percentage change in unadjusted mean log₁₀-transformed ratios per 1-unit increase for continuous measures or presence versus absence of the characteristic for categorical covariates. No multivariable analysis was conducted, given the limited sample size. We performed sensitivity analyses by quantifying in utero TFV transfer only among mother-infant pairs in which both mother and infant hair TFV levels were within the limits of quantification and in which mothers reported TDF use during all trimesters.

RESULTS

We measured TFV hair levels for 116 mother-infant pairs with TDF-based ART exposure during pregnancy; within this group, 103 (89%) mothers had TFV levels above the LLQ and had singleton births and were included in the analysis with their infants (Supplementary Figure 1). The median maternal age at delivery (interquartile range [IQR]) was 31.7 (26.2–36.4) years, and 70% of mothers self-identified as non-Hispanic black (Table 1). The median (IQR) gestational age at delivery was 38 (38–39) weeks, and 8% of mother’s had unsuppressed viral loads in late pregnancy. Commonly used anchors with TDF/emtricitabine–containing regimens included atazanavir-ritonavir (28%), rilpivirine (24%), and darunavir-ritonavir (17%). The median (IQR) duration of TDF use during pregnancy was 32.3 (22.4–38.3) weeks.

Table 1.

Correlates of Log₁₀-Transformed Ratios of Infant-to-Mother Tenofovir Concentrations in Hair Among Women With Tenofovir Disoproxil Fumarate Exposure During Pregnancy

				Univariable Linear Regressionc
Maternal Characteristics	Mother-Infant Pair, No.	No. (%),a	Log10 Ratio of Infant-to-Mother TFV Concentration, Mean (95% CI)b	Absolute Difference (95% CI)	Change (95% CI), %	P Value
Age at delivery, y	103
25–34		53 (51)	1.000 (.844–1.157)	−0.151 (−.484 to .183)	−29.3 (−67.2 to 52.4)	.37
≥35		35 (34)	1.182 (.989–1.375)	0.031 (−.321 to .383)	7.3 (−52.3 to 141.4)	.86
<25		15 (15)	1.151 (.857–1.446)	Reference	…	…
Race/ethnicity
White/other non-Hispanic		8 (8)	0.919 (.517–1.321)	−0.129 (−.553 to .295)	−25.7 (−72.0 to 97.1)	.55
Hispanic		23 (22)	1.253 (1.016–1.490)	0.205 (−.068 to .477)	60.3 (−14.4 to 200.2)	.14
Black non-Hispanic		72 (70)	1.048 (.914–1.182)	Reference	…	…
Achieved at least high school graduation	103
Yes		73 (71)	1.153 (1.021–1.285)	0.237 (−.007 to 0.482)	72.8 (−1.6 to 203.4)	.06
No		30 (29)	0.916 (.710–1.122)	Reference	…	…
Recreational drug use in pregnancyd	103
Yes		14 (14)	1.107 (.801–1.414)	0.027 (−.303 to .357)	6.4 (−50.2 to 127.6)	.87
No		89 (86)	1.080 (.959–1.202)	Reference	…	…
Pain medication use in pregnancy	103
Yes		4 (4)	1.038 (.463–1.612)	−0.048 (−.634 to .537)	−10.5 (−76.8 to 244.5)	.87
No		99 (96)	1.086 (.971–1.201)	Reference	…	…
Antidepressant use in pregnancy	103
Yes		4 (4)	0.869 (.296–1.442)	−0.224 (−.808 to .360)	−40.3 (−84.4 to 129.2)	.45
No		99 (96)	1.093 (.978–1.208)	Reference	…	…
Marijuana use in pregnancy	103
Yes		10 (10)	1.050 (.687–1.413)	−0.038 (−.420 to .344)	−8.3 (−62.0 to 121.1)	.85
No		93 (90)	1.088 (.969–1.207)	Reference	…	…
Tobacco use in pregnancy	103
Yes		16 (16)	0.988 (.701–1.274)	−0.114 (−.425 to .198)	−23.1 (−62.5 to 57.6)	.47
No		87 (84)	1.102 (.979–1.225)	Reference	…	…
Alcohol use in pregnancy	103
Yes		10 (10)	1.059 (.696–1.422)	−0.028 (−.410 to .354)	−6.3 (−61.1 to 125.9)	.88
No		93 (90)	1.087 (.968–1.206)	Reference	…	…
Gestational age at birth, median (IQR), wk	103	38 (38–39)		0.030 (−.062 to .122)	7.1 (−13.3 to 32.3)	.52
Preterm birth (<37 wk gestation)	103
Yes		10 (10)	0.721 (.366–1.077)	−0.402 (−.776 to −.028)	−60.3 (−83.2 to −6.2)	.04
No		93 (90)	1.123 (1.007–1.240)	Reference	…	…
Mode of delivery	102
Cesarean		55 (54)	0.976 (.824–1.129)	−0.236 (−.461 to −.012)	−42.0 (−65.4 to −2.7)	.04
Vaginal delivery		47 (46)	1.213 (1.048–1.377)	Reference	…	…
BMI at deliverye	68
25.0–29.9		20 (29)	1.246 (.971–1.521)	0.112 (−.402 to .626)	29.4 (−60.4 to 323.1)	.66
≥30		40 (59)	1.003 (.809–1.197)	−0.131 (−.607 to .345)	−26.0 (−75.3 to 121.4)	.58
18.5–24.9		8 (12)	1.134 (.699–1.569)	Reference	…	…
Latest CD4 T-cell count in pregnancy	100
<350/μL		27 (27)	1.105 (.883–1.327)	0.022 (−.238 to .282)	5.3 (−42.1 to 91.5)	.87
≥350/ μ L		73 (73)	1.082 (.947–1.217)	Reference	…	…
Latest RNA level during pregnancy	102
≥400 copies/mL		8 (8)	1.292 (.889–1.694)	0.217 (−.202 to .636)	64.7 (−37.2 to 332.1)	.31
<400 copies/mL		94 (92)	1.075 (.957–1.192)	Reference	…	…
Infant hair color	103
Black		85 (83)	1.137 (1.015–1.259)	0.304 (.013−.596)	101.6 (2.9–294.6)	.04
Other/mixed		18 (17)	0.833 (.568–1.098)	Reference	…	…
ART regimen containing ≥3 classes	100
Yes		9 (9)	0.946 (.560–1.332)	−0.146 (−.551 to .259)	−28.5 (−71.9 to 81.7)	.48
No		91 (91)	1.092 (.970–1.213)	Reference	…	…
NNRTI use during pregnancy	100
Yes		32 (32)	1.165 (.961–1.370)	0.128 (−.120 to .376)	34.2 (−24.2 to 137.5)	.31
No		68 (68)	1.038 (.897–1.178)	Reference	…	…
PI use during pregnancy (vs none)	100
Yes		57 (57)	1.019 (.866–1.172)	−0.139 (−.372 to .094)	−27.4 (−57.6 to 24.2)	.24
No		43 (43)	1.158 (.982–1.334)	Reference	…	…
INSTI use during pregnancy (vs none)	100
Yes		37 (37)	0.955 (.767–1.144)	−0.196 (−.434 to .041)	−36.3 (−63.2 to 10.0)	.10
No		63 (63)	1.151 (1.007–1.296)	Reference	…	…
Cumulative TDF use in pregnancy, median (IQR), wk	100	32.3 (22.4–38.3)	…	0.007 (−.005 to .018)	1.6 (−1.0 to 4.3)	.24
TDF use in first trimester only	103
Yes		2 (2)	−0.277 (−1.042 to .489)	−1.387 (−2.160 to −.615)	−95.9 (−99.3 to −75.7)	<.001
No		101 (98)	1.111 (1.003–1.219)	Reference	…	…

Abbreviations: ART, antiretroviral therapy; BMI, body mass index; CI, confidence interval; INSTI, integrase strand transfer inhibitor; IQR, interquartile range; NNRTI, nonnucleoside reverse-transcriptase inhibitors; PI, protease inhibitor; TDF, tenofovir disoproxil fumarate; TFV, tenofovir.

^aData represent no. (%) of mothers or infants, unless otherwise specified.

^bEstimated unadjusted mean log₁₀-transformed ratio of infant-to-mother TFV concentration for mother-infant pair with a specific characteristic.

^cModels were based on unique pregnancy with hair specimen tested for both mother and infant and with detectable maternal hair TFV concentration level; for pregnancies (n = 2) with undetectable infant hair concentration, the lower limit of quantification (LLQ; 0.002 ng/mg) was used in calculating the ratio; for pregnancies with hair concentration above the upper limit of quantification (ULQ) (1 maternal hair and 82 infant hair specimens), the ULQ (0.4 ng/mg) was used in calculating the ratio.

The estimated absolute difference in unadjusted mean log₁₀-transformed ratio of infant-to-mother TFV concentration (1) per 1-unit increase in continuous covariate measures or (2) between mother-infant pair with a specific characteristic and the reference group for categorical covariate measures. The estimated percentage change in unadjusted mean ratio of infant-to-mother TFV concentration (1) per 1-unit increase in continuous covariate measures or (2) between mother-infant pair with a specific characteristic and the reference group for categorical covariate measures.

^dRecreational drugs included marijuana, cocaine, heroin, sedatives, methamphetamines, phencyclidine, opium, stimulants, barbiturates, amphetamines, inhalants, lysergic acid diethylamide, and other hallucinogens.

^eBMI was calculated as weight in kilograms divided by height in meters squared.

The median (IQR) time from birth to hair collection was 3 (1–14) days. Eighty-two (80%) infants had TFV hair levels above the ULQ, and the frequency of infant levels this high varied by maternal TFV hair level categories [12] (Figure 1A); only 1 maternal hair level was above the ULQ. The median concentration of TFV (IQR; absolute range) was 0.02 (0.01–0.04; 0.002–0.40) ng/mg in maternal hair and 0.40 (0.40–0.40; 0.002–0.40) ng/mg in infant hair. The mean log₁₀ ratio of infant-to-maternal TFV levels was 1.08 (95% confidence interval, .97–1.20), and the median was 1.21 (IQR, 0.83–1.40; absolute range, −1.37 to 2.22). The Spearman coefficient between maternal and infant TFV levels was 0.221 (P = .02; Figure 1B), suggesting a weak correlation. Among pregnancies in which TDF was used across all 3 trimesters (n = 71), the median log₁₀ ratio of infant-to-maternal TFV levels was 1.19 (IQR, 0.83–1.40; absolute range −1.36 to 2.22). Among mother-infant pairs in which both mother and infant hair TFV levels were within the limits of quantification (n = 18), the median log₁₀ ratio of infant-to-maternal TFV levels was 1.21 (IQR, 0.09–1.33; absolute range −0.23 to 2.06).

Figure 1.

A, Frequency of infant tenofovir (TFV) hair concentrations above the upper limit of quantification (ULQ) by maternal TFV hair concentration categories based on dosage per week cutoffs (N = 103). B, Infant by maternal log₁₀ TFV hair concentrations plotted by maternal concentrations at delivery (N = 103). Calculations were based on unique pregnancies with hair specimens tested for both mother and infant and with maternal hair level above the lower limit of quantification (LLQ). For pregnancies with undetectable infant hair concentration (n = 2), the LLQ value (0.002 ng/mg) was used. For pregnancies with hair level above the ULQ (1 maternal and 82 infant hair specimens), the ULQ value (0.4 ng/mg) was used in calculating the ratio. Maternal TFV hair concentration cutoffs were defined according to benchmarks from prior directly observed studies among nonpregnant women [12].

TFV transfer was 60% lower from mothers who had preterm versus term deliveries (mean log₁₀ ratio, 0.72 vs 1.12; P = .04) (Table 1) and 42% lower from mothers with had cesarean versus vaginal deliveries (mean log₁₀ ratio, 0.98 vs 1.21; P = .04). TFV transfer was also lower from mothers who used TFV during the first trimester rather than in other trimesters (mean log₁₀ ratio, −0.28 vs 1.11; P < .001). There was a trend toward 36% lower TFV transfer among mother who received integrase strand transfer inhibitors (INSTs) than among those who did not (mean log₁₀ ratio, 0.96 vs 1.15; P = .10). Higher TFV transfer from mothers was marginally associated with maternal achievement of high school graduation and infant black hair color (Table 1). We did not observe associations (P < .10) between TFV transfer and other characteristics in this limited sample.

Discussion

To our knowledge, this is the first evaluation of mother-to-infant TFV transfer using hair concentrations. Prior studies assessed in utero TFV exposure via short-term metrics and similarly found high rates of TFV transfer, with our study confirming that such transfer is cumulative. In exploratory models, transfer was lower among mothers who had preterm delivery, those with only first-trimester TDF use, and those with cesarean deliveries. Our data contribute to ongoing discussions, regarding which ART regimens can minimize toxicity in infants while maximizing protection.

TFV may cross the placenta by both passive and active transport. While passive placental transport of TFV is limited due to its physical and chemical properties [13], both TDF and TFV are substrates for various efflux transporters expressed in the human placenta, which could explain the high levels of TFV exposure observed among infants. P-glycoprotein expression and activity decreases in the placenta in the third trimester [14], the period of exposure assessed in our study. TFV is predominantly excreted by the kidney as unmetabolized drug [13]. Once TFV crosses the placenta and enters fetal circulation, it may be excreted by the fetal kidneys into amniotic fluid and subsequently reabsorbed by the fetus from swallowed amniotic fluid [6].

Prior studies measuring TFV concentrations in plasma, cord blood, amniotic fluid, and meconium estimated infant-to-mother TFV ratios of 0.60–0.88 [5–7]. Drug concentrations in cord blood reflect maternal exposure over a short time period and do not accurately represent exposure in a newborn already capable of drug metabolism. Single levels of ARVs collected from infant plasma or meconium at delivery provide only a snapshot of recent exposure and levels demonstrate significant day-to-day variation in these matrices. Our study is the first to quantify cumulative TFV exposure in the infant during the last month of pregnancy. We found that transfer was lower among preterm deliveries, suggesting that transfer varies during gestation. Because P-glycoprotein activity is reduced in the last weeks of pregnancy, the higher transfer observed among term deliveries is expected [13].

Our method of measuring TFV in paired hair samples could be applied to future studies assessing maternal-to-infant transfer and safety of newer ARVs (eg, dolutegravir, TFV alafenamide, and cabotegravir). Hair collection has feasibility advantages over other matrices, because hair can be collected, stored, and shipped at room temperature without biohazardous precautions. The methods described here could be particularly useful for evaluating infant ARV exposure in low- and-middle-income settings with high HIV burden, where ARV exposure during pregnancy is most prevalent.

Our results contribute to the safety-monitoring field of TFV exposure during pregnancy. In the PROMISE (Promoting Maternal-Infant Survival Everywhere) trial, use of TDF-based ART among mothers resulted in more frequent severe adverse pregnancy outcomes than the use of zidovudine-based ART [15]. A potential mechanism for this finding could be a pharmacokinetic interaction between TDF and protease inhibitors (PIs), as PI dosages were increased during the third trimester trimester in PROMISE [15]. Unlike PROMISE, SMARTT is an observational study, and dosage was not measured. In our study, we found that receipt of PIs was not associated with in utero TFV transfer, and receipt of INSTIs trended toward a 36% decrease. INSTIs are affected by pharmacokinetic changes during pregnancy and do not require boosting in most formulations. Dolutegravir use during pregnancy was hindered in recent years owing to a possible association of maternal dolutegravir use and neural tube defects [16]. However, new data indicate that this risk is lower than previously reported. More studies are needed that evaluate objective measures of in utero exposure to ARVs in relationship to infant outcomes.

Our study has limitations. More than half of paired specimens could not be tested, mainly owing to insufficient amount of hair. Our sample size limited our ability to fit multivariable models or to detect statistical associations between covariates and ratios of infant-to-mother TFV concentrations. Results from our exploratory univariable models could be confounded by other factors. Larger studies powered to identify correlates of in utero TFV exposure, and the relationship between objective metrics of TFV exposure and infant outcomes, are needed.

In conclusion, we assessed in utero transfer rates of the most commonly used ARV worldwide via hair levels for the first time and found high rates of TFV transfer. This method can be used for other ART agents and other agents commonly used in pregnancy to help maximize their efficacy while minimizing safety concerns.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. We thank the women for their participation in the Pediatric HIV/AIDS Cohort Study (PHACS), and the individuals and institutions involved in the conduct of PHACS. The study was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cofunding from the National Institute on Drug Abuse, the National Institute of Allergy and Infectious Diseases, the National Institute of Mental Health, the National Institute of Neurological Disorders and Stroke, the National Institute on Deafness and Other Communication Disorders, the National Institute of Dental and Craniofacial Research, the National Cancer Institute, the National Institute on Alcohol Abuse and Alcoholism, the Office of AIDS Research, and the National Heart, Lung, and Blood Institute, through cooperative agreements with the Harvard T.H. Chan School of Public Health (principal investigator, George R Seage III; program director, Liz Salomon) and the Tulane University School of Medicine (principal investigator, Russell Van Dyke; co–principal investigator, Ellen Chadwick; project director, Patrick Davis). Data management services were provided by Frontier Science and Technology Research Foundation (principal investigator, Suzanne Siminski), and regulatory services and logistical support were provided by Westat (principal investigator, Julie Davidson).

The following institutions, clinical site investigators and staff participated in conducting PHACS SMARTT in 2018, in alphabetical order—Ann & Robert H. Lurie Children’s Hospital of Chicago: Ellen Chadwick, Margaret Ann Sanders, Kathleen Malee, Scott Hunter; Baylor College of Medicine: William Shearer, Mary Paul, Chivon McMullen-Jackson, Ruth Eser-Jose, Lynnette Harris; Bronx Lebanon Hospital Center: Murli Purswani, Mahoobullah Mirza Baig, Alma Villegas; Children’s Diagnostic & Treatment Center: Lisa Gaye-Robinson, Jawara Dia Cooley, James Blood, Patricia Garvie; New York University School of Medicine: William Borkowsky, Sandra Deygoo, Jennifer Lewis; Rutgers–New Jersey Medical School: Arry Dieudonne, Linda Bettica, Juliette Johnson, Karen Surowiec; St Jude Children’s Research Hospital: Katherine Knapp, Kim Allison, Megan Wilkins, Jamie Russell-Bell; San Juan Hospital/Department of Pediatrics: Nicolas Rosario, Lourdes Angeli-Nieves, Vivian Olivera; SUNY Downstate Medical Center: Stephan Kohlhoff, Ava Dennie, Jean Kaye; Tulane University School of Medicine: Russell Van Dyke, Karen Craig, Patricia Sirois; University of Alabama, Birmingham: Cecelia Hutto, Paige Hickman, Dan Marullo; University of California, San Diego: Stephen A. Spector, Veronica Figueroa, Megan Loughran, Sharon Nichols; University of Colorado, Denver: Elizabeth McFarland, Emily Barr, Christine Kwon, Carrie Glenny; University of Florida, Center for HIV/AIDS Research, Education and Service: Mobeen Rathore, Kristi Stowers, Saniyyah Mahmoudi, Nizar Maraqa, Rosita Almira; University of Illinois, Chicago: Karen Hayani, Lourdes Richardson, Renee Smith, Alina Miller; University of Miami: Gwendolyn Scott, Maria Mogollon, Gabriel Fernandez, Anai Cuadra; Keck Medicine of the University of Southern California: Toni Frederick, Mariam Davtyan, Jennifer Vinas, Guadalupe Morales-Avendano; and University of Puerto Rico School of Medicine, Medical Science Campus: Zoe M. Rodriguez, Lizmarie Torres, Nydia Scalley.

Disclaimer. The conclusions and opinions expressed in this article are those of the authors and do not necessarily reflect those of the National Institutes of Health or the US Department of Health and Human Services.

Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases (R21AI138618 and 2R01AI098472) and the Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01HD100201) with support from the University of Washington and University of California, San Francisco (UCSF) Centers for AIDS Research (P30 AI027757 and 2P30AI027763). PHACS was supported by the National Institute of Child Health and Human Development, with cofunding from the National Institute of Dental and Craniofacial Research, the National Institute of Allergy and Infectious Diseases, the National Institute of Neurological Disorders and Stroke, the National Institute on Deafness and Other Communication Disorders, the Office of AIDS Research, the National Institute of Mental Health, the National Institute on Drug Abuse, and the National Institute on Alcohol Abuse and Alcoholism, through cooperative agreements with the Harvard T.H. Chan School of Public Health (grant HD052102) and the Tulane University School of Medicine (grant HD052104).

Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.