Cunningham, Coleen K. MD*; Balasubramanian, Rajalakshmi ScD†; Delke, Isaac MD‡; Maupin, Robert MD§; Mofenson, Lynne MD¶; Dorenbaum, Alex MD**; Sullivan, John L. MD††; Gonzalez-Garcia, Adolfo MD‡‡; Thorpe, Edwin MD§§; Rathore, Mobeen MD¶¶; Gelber, Richard D. PhD†
Marked changes in perinatal HIV transmission have been observed in the United States and other developed countries since 1994, when Pediatric AIDS Clinical Trials Group (PACTG) protocol 076 showed that administration of zidovudine (ZDV) to the HIV-infected woman during pregnancy and labor and to her newborn decreased the risk of perinatal HIV transmission. 1 With the implementation of recommendations for universal prenatal HIV counseling and testing, antiretroviral (ARV) prophylaxis and elective cesarean delivery, and avoidance of breast-feeding, the rate of perinatal HIV infection has diminished to <2%. 2,3 Consistent with this, in a recent international, randomized, controlled trial of nevirapine compared with placebo in nonbreast-feeding women receiving “standard ARV therapy” (PACTG 316), the overall transmission rate was only 1.5%, without a significant difference between the nevirapine and placebo groups. 4 In PACTG 316, the most significant factor correlated with the risk of transmission was maternal viral load at delivery. It was noted that transmission appeared rare in women identified as white non-Hispanic (white) (1/280; 0.4%) compared with those identified as black non-Hispanic (black) (15/724; 2.0%) or Hispanic (3/228; 1.3%); although this finding did not reach statistical significance, the overall transmission rate in PACTG 316 was low, and the number of infant infections in the study was small (19 infants). The specific objective of the current analysis was to further explore the relationship of race/ethnicity and perinatal transmission in PACTG 316 using data from sites in the mainland United States and Puerto Rico and to determine whether other factors that might be associated with race/ethnicity (such as time of initiation of prenatal care or type of therapy received) might influence this relationship.
PACTG 316 was an international, multicenter, randomized, blinded trial of a 2-dose nevirapine regimen (single 200-mg dose to the woman in labor and single 2-mg/kg dose to the neonate at 48–72 hours of life) compared with placebo to reduce perinatal transmission of HIV in women on standard ARV therapy and who were not breast-feeding. A detailed description of the trial design and primary analysis results has been published. 4 In brief, enrollment was offered to HIV-infected pregnant women receiving “standard ARV therapy” as defined by their care provider, except that women receiving nonnucleoside reverse transcriptase inhibitors were excluded. The research protocol was approved by each site's institutional review board and all participants provided written, informed consent. Overall, 73% of women received antenatal combination ARV therapy and 23% received ZDV alone; restricted to only US sites, 83% received antenatal combination therapy and 17% ZDV alone.
The primary analysis included women enrolled in sites throughout the United States and Puerto Rico as well as Europe, Brazil, and the Bahamas and included all deliveries in which the woman received intrapartum study drug and infant infection status was known. For the purpose of the present analysis, only PACTG sites (continental United States and Puerto Rico) were included and all deliveries for which maternal RNA and infant infection status were known were included in the analysis, irrespective of whether study drug was delivered.
Baseline demographic, clinical, and treatment data were collected at enrollment and at delivery and 6 weeks postpartum for all women in the study. Infants had clinical and laboratory data collected at birth and at protocol-defined intervals through the first 6 months of life. Sites were instructed to record race/ethnicity as assigned by the study participants. The following categories were used: white non-Hispanic (white); black non-Hispanic (black); Hispanic/Latino or Spanish culture or origin (Hispanic), regardless of color; and other. Suppression of viral replication was defined as having HIV RNA levels of <400 copies/mL at the time of delivery.
The distribution of selected maternal disease and treatment characteristics was assessed by race/ethnicity category. Differences between characteristics for each race/ethnicity group were evaluated by Fisher-Freeman-Halton exact test 5 for categorical variables and by Kruskal-Wallis test for continuous variables. Logistic regression models were fit to evaluate possible association of factors with HIV transmission and with viral load at delivery. Models for viral load at delivery included treatment factors such as type and time of initiation of maternal ARV therapy, time in pregnancy when prenatal care was initiated and mode of delivery, as well as factors such as race/ethnicity, tobacco, alcohol, and drug use. Models for HIV transmission included these factors as well as other maternal disease characteristics such as viral load at delivery, duration of membrane rupture, CD4 at entry, Centers for Disease Control (CDC) disease category, and presence of sexually transmitted infection (STI). Variables associated with the outcome at P < 0.05 level (based on likelihood ratio test) were retained in the final models.
Of the 1052 women randomized at the US PACTG sites, 1002 delivered at a study site, 978 had infant infection status determined, and 905 had maternal delivery HIV RNA levels available. Only 14 of these 905 women were characterized as “other” for race/ethnicity and there were no HIV transmissions in this small group. These 14 women were excluded from all remaining analyses. Therefore, the final analysis group included 891 women from US sites who had known race/ethnicity, infant infection status, and maternal delivery HIV RNA levels. Of these 891 women, 572 (64%) were black, 206 (23%) were Hispanic, and 113 (13%) were white. All women who had infected infants were black or Hispanic (11/572 and 3/206, respectively), whereas none of the women identified as white had an infected infant (0/113). However, this difference was not statistically significant (P = 0.54, Fisher-Freeman-Halton exact test for differences among the 3 race/ethnicity groups).
Although HIV transmission was not significantly different between racial groups, there were important differences in baseline demographic, clinical, and laboratory parameters between groups (Table 1). White subjects had higher entry CD4 cell counts and lower delivery HIV RNA levels than did women of other race/ethnicity. White subjects were significantly more likely to have used illegal drugs, smoked cigarettes, and used alcohol during pregnancy. The occurrence of STIs differed by race/ethnicity (Table 2). Bacterial vaginosis was more frequent in black women than white and Hispanic women (P = 0.0002); syphilis occurred more frequently in black and Hispanic women than white women (P = 0.01); and herpes simplex virus infection was more common in white women than women of black or Hispanic race/ethnicity (P = 0.05). Overall, more black women than women of white or Hispanic race/ethnicity had ≥1 STI (P = 0.03). However, neither STIs overall nor individual STIs were significantly associated with risk of perinatal HIV transmission.
Most women entered prenatal care during the first trimester of pregnancy and the time of entry into care was not significantly different among women of the 3 racial and ethnic groups (Table 1). To determine whether treatment provided during pregnancy varied between groups, the time of initiation of ARV, type of ARV, and percent of women undergoing elective cesarean delivery were compared (Table 3). In this cohort, white women were more likely to initiate ARV prior to the first trimester; ie, women were already receiving ARV therapy at the time they became pregnant. Forty-two percent of white women started therapy prior to pregnancy compared with 29% of Hispanic and 27% of black women. Approximately two-thirds of women who started on treatment prior to pregnancy received combination therapy with a protease inhibitor, whereas one-third received combination therapy without a protease inhibitor, regardless of race/ethnicity (Table 4).
For women who started therapy during pregnancy, approximately one-third received combination therapy including a protease inhibitor, and this did not differ by race/ethnicity (Table 4). However, for women in this group who did not receive combination protease inhibitor therapy, black and Hispanic women were more likely to receive monotherapy (25% each for black and Hispanic) than were white women (12%), although this trend was not statistically significant. Women who were not receiving therapy prior to pregnancy most often cited “not knowing their HIV status” as the reason for not being on treatment; this occurred in 251 black (44%), 85 Hispanic (41%), and 36 white women (32%). Infant ARV treatment did not vary by racial/ethnic group (Table 3).
In the present analysis cohort, the incidence of elective cesarean delivery differed according to race/ethnicity (P = 0.04;Table 3) with black women having the lowest rate. This difference, however, was restricted to women with HIV RNA ≤400 copies/mL at delivery, who have a low risk of transmission; there was no difference in the rate of elective cesarean delivery by race/ethnicity for women with detectable HIV RNA levels at delivery (HIV RNA > 400) (P = 0.62).
Other factors potentially associated with transmission risk, including duration of membrane rupture, intrapartum maternal antibiotics (as a surrogate for potential chorioamnionitis), and fetal scalp electrode use, were compared by race/ethnicity among the 447 women with active viral replication (HIV RNA > 400 copies/mL) at the time of delivery. Among the 48 white women with HIV RNA > 400 copies/mL at delivery, mean and median duration of membrane rupture (8.0 hours and 1.6 hours, respectively) were longer than in the 308 black women (5.4 hours and 1.1 hours) and 91 Hispanic women (3.0 hours and 0 hours) (P = 0.002). Other intrapartum procedures and interventions including antibiotic administration and invasive fetal monitoring did not differ among racial/ethnicity groups.
In univariate analysis for modeling/predicting the probability of HIV transmission from mother to child in this subset of women in PACTG 316, the only significant predictor of HIV transmission (P < 0.05) was maternal HIV RNA at the time of delivery. In multivariate modeling, when including all variables satisfying P value ≤0.25 in the univariate model, race/ethnicity and CDC disease category did not improve the model (data not shown).
Because HIV RNA at delivery is the primary factor associated with transmission risk, we sought to determine which demographic and treatment/intervention variables would predict successful suppression of viral replication at delivery (ie, HIV RNA ≤ 400 copies/mL at delivery). Univariate analysis demonstrated that type of ARV therapy (combination with and without protease inhibitor, monotherapy, no ARV), time of initiation of ARV (before vs. during or after the second trimester), time of initiation of prenatal care (before or during third trimester), race (white vs. other race/ethnicity), mode of delivery (spontaneous vaginal vs. other mode) were individually significantly associated with successful viral suppression at delivery (Table 5). In bivariate models that included ARV type and individual variables that had values P ≤0.25 in univariate analysis, time of ARV initiation, time of prenatal care initiation, and race/ethnicity each retained significance in predicting successful viral suppression at delivery. Additionally, race/ethnicity remained predictive of successful viral suppression at delivery in a multivariate model incorporating all of these variables (type of ARV, time of ARV initiation, and time of prenatal care initiation) (P = 0.01).
The transmission rate of 1.5% in PACTG 316 is reflective of success achieved in developed countries in reducing perinatal HIV transmission. In this study, maternal HIV RNA level at delivery was the single most important factor in predicting risk of transmission, and the small number of infected infants (19 infants) limits the ability to evaluate other potential correlates of transmission risk and larger data sets may be needed to further evaluate the impact of race on perinatal HIV transmission. The differences in transmission by race/ethnicity observed in PACTG 316 were of concern to the study team and led to a further investigation of the relationship between race/ethnicity and other risk factors for transmission.
Because HIV RNA level is so critical in determining transmission, factors predicting successful suppression of viral replication at delivery, defined as delivery HIV RNA levels ≤400 copies/mL, were explored. A number of factors, including race/ethnicity, type of ARV therapy, and time in pregnancy when prenatal care was initiated were significant predictors of viral suppression at delivery. Time of initiation and type of ARV used as well as access into medical care might be expected to impact on HIV RNA; without access to ARV treatment, a decrease in viral load would not occur. However, race/ethnicity as an independent predictor of viral suppression during pregnancy has not been previously described.
Maternal disease status differed by race/ethnicity in PACTG 316, with pregnant white women having significantly higher entry CD4 cell counts and lower delivery HIV RNA levels than either black or Hispanic women. We found that a higher proportion of white women had started therapy prior to pregnancy compared with women of other race/ethnicity and that women who started therapy before pregnancy, regardless of race/ethnicity, were more likely to have received highly active antiretroviral therapy (HAART) with a protease inhibitor. Thus, the overall greater use of ARV combinations that included a protease inhibitor among white women appears to be attributable to a larger proportion of white women being on ARV therapy prior to pregnancy compared with the other race/ethnic groups.
The more advanced disease status, as demonstrated by lower CD4 counts, of the black and Hispanic women might be expected to lead to a greater use of HAART among those who started treatment during pregnancy; however, this is not what was observed. For women who initiated therapy during pregnancy, there were similar rates of HAART use in all 3 racial/ethnicity groups, although among women who did not receive a protease inhibitor, there was greater use of monotherapy in the black and Hispanic women. The present study is limited by the fact that clinical information such as CD4 count and HIV RNA was collected during the periods of study but was not available for times prior to pregnancy or earlier during pregnancy when many of the decisions about maternal ARV therapy were being made.
The most frequent reason cited by study subjects for not receiving ARV therapy prior to pregnancy was that they had not known their HIV status. In addition, there was a significantly longer interval from the time of diagnosis of HIV infection to delivery in white women compared with the black and Hispanic women in the cohort. These findings suggest that some of the differences in prescribing ARV medication between racial/ethnic groups may be due to a greater number of white women in the present study having their HIV infection diagnosed prior to pregnancy. The white women in this cohort have a high prevalence of illegal drug use and therefore may have been identified as being at risk for HIV. In contrast, black and Hispanic women in the present cohort were not likely to report illegal drug use and therefore may not have been perceived to be at risk for HIV infection. It is also possible that the black and Hispanic women in this study may have less access to health care or may be less likely to be offered or agree to an HIV test. These findings suggest that all women of childbearing age should be offered HIV testing when they access medical care, not just those with risk factors such as drug abuse or pregnancy.
In contrast to the difference in antenatal ARV therapy, the intrapartum approach to care for those women with detectable HIV RNA and for infants after delivery was identical for all racial/ethnic groups. Elective cesarean delivery has been shown to reduce the risk of perinatal transmission in both a clinical trial and meta-analysis. 6,7 In the overall study cohort there was less frequent use of elective cesarean delivery in black women. However, when restricted to women with delivery HIV RNA levels of >400 copies/mL—those for whom elective cesarean delivery would likely provide a benefit—there is no difference in incidence of elective cesarean among racial/ethnic groups.
Other investigators have described racial/ethnic differences in practices for treatment and other care in HIV-infected adults in the United States. 8,9 In a review of the literature, Palacio et al 10 reported racial/ethnic disparities in utilization of highly active ARV therapies. In studies of HIV-infected women who were not pregnant, factors such as care by an HIV specialist, maternal age >25 years, education, absence of drug use, and not being African American have been shown to predict increased use of highly active ARV combinations. 11–13
Differences by race/ethnicity in quality of care received or access to care have also been described for other medical conditions including congestive heart failure, pneumonia, and coronary bypass surgery. 14,15 We were not able to document a difference in highly active treatment administered in those treated prior to pregnancy or those starting therapy during pregnancy. Participants who reported combination therapy without a protease inhibitor were not receiving highly active treatment: in all cases they were receiving combinations of nucleoside analogues. In fact, stratifying the analysis according to when ARV therapy started relative to pregnancy (Table 4), the percent of women who received “highly active” protease inhibitor–containing combinations was virtually identical across racial/ethnic groups.
Even after accounting for type of ARV and time of initiation of ARV, race/ethnicity remained significant in predicting viral suppression at the time of delivery. Other investigators have not found a difference in response to ARV therapy related to race. Jensen-Fangel et al 16 reported identical rates of viral suppression and CD4 increases in whites and nonwhites in a Danish cohort with free access to medication. Further, in a cohort of US patients, race/ethnicity was not found to impact on CD4 response in patients with suppressed virus on HAART. 17 These reports suggest that the differences in viral suppression seen in this study are not related to genetic factors but may be related to treatment differences that were not measured in our study or to other differences including other infections, adherence, or pharmacokinetics. The present study was not designed to evaluate all of these factors.
The implications of higher RNA and lower CD4 cells in pregnant black and Hispanic women in this trial are important both for the woman's health and for her child's outcome. Further prospective studies should be designed to explore the factors that are considered when determining the type of ARV combinations initiated during pregnancy and whether race/ethnicity, language barriers, access to care, health care provider, or socioeconomic factors affect health care providers' recommendation for therapy or the patients' willingness to initiate therapy.
The International PACTG study team includes Brigitte Bazin (ANRS France), Yvonne Bryson (University of California School of Medicine, Los Angeles), Paula Britto (Statistical and Data Analysis Center, Harvard School of Public Health, Boston, MA), Mary Culnane and Scharla Estep (National Institute of Allergy and Infectious Diseases, Bethesda, MD), Jean Francois Delfraissy (ANRS France), Marie-Louise Newell and Simona Fiore (ECS Institute of Child Health, London England), Maria Gigliotti (Boehringer Ingelheim), Mark Mirochnick (Department of Pediatrics, Boston University School of Medicine, Boston, MA), Claire Rekacewicz (ANRS France), Maureen Shannon (San Francisco General Hospital, San Francisco, CA), and Heather Watts (National Institute of Child Health and Human Development, Bethesda, MD). Participating sites are listed in the primary publication from the study (Dorenbaum A, Cunningham CK, Gelber RD, et al. Two-dose intrapartum/newborn nevirapine and standard ARV therapy to reduce perinatal HIV-1 transmission: a randomized trial. JAMA. 2002;288:189–1998). In addition to the site participants listed in the primary manuscript, one additional site also enrolled patients in this study: Donna Sandifer and Rene Smit at Children's National Medical Center, Washington, DC.
1. Connor EM, Sperling RS, Gelber R, et al. Reduction of maternal-infant transmission of human immunodeficiency virus type 1 with zidovudine treatment. N Engl J Med. 1994;331:1173–1180.
2. Mofenson LM and the Committee on Pediatric AIDS. Technical report: perinatal human immunodeficiency virus testing and prevention of transmission. Pediatrics
. 2000;106 (6)e88. Available at: http://www.pediatrics.org/cgi/contents/full/106/6/e88
. Accessed November 10, 2003.
3. Cooper ER, Charurat M, Mofenson L, et al. Combination antiretroviral strategies for treatment of pregnant HIV-1-infected women and prevention of perinatal HIV-1 transmission. J Acquir Immune Defic Syndr. 2002;29:484–494.
4. Dorenbaum A, Cunningham CK, Gelber RD, et al. Two-dose intrapartum/newborn nevirapine and standard antiretroviral therapy to reduce perinatal HIV-1 transmission: a randomized trial. JAMA. 2002;288:189–198.
5. Freeman GH, Halton JH. Note on an exact treatment of contingency, goodness of fit and other problems of significance. Biometrika. 1951;38:141–149.
6. Anonymous. Elective caesarean-section versus vaginal delivery in prevention of vertical HIV-1 transmission: a randomised clinical trial. Lancet. 1999;353:1035–1039.
7. Anonymous. The mode of delivery and the risk of vertical transmission of human immunodeficiency virus type 1: a meta-analysis of 15 prospective cohort studies. The International Perinatal HIV Group. N Engl J Med. 1999;340:977–987.
8. Moore RD, Stanton D, Gopalan R, et al. Racial differences in the use of drug therapy for HIV disease in an urban community. N Engl J Med. 1994;330:763–768.
9. Shapiro MF, Morton SC, McCaffery DF, et al. Variations in the care of HIV-infected adults in the United States: results from the HIV cost and services utilization study. JAMA. 1999;281:2305–2315.
10. Palacio H, Kahn JG, Richards TA, et al. Effect of race and/or ethnicity in use of antiretrovirals and prophylaxis for opportunistic infection: a review of the literature. Public Health Rep. 2002;117:233–251.
11. Gardner LI, Holmberg SD, Moore J, et al. Use of highly active antiretro-viral therapy in HIV-infected women: impact of HIV specialist care. J Acquir Immune Defic Syndr. 2002;29:69–75.
12. Turner BJ, Zhang D, Laine C, et al. Association of provider and patient characteristics with HIV-infected women's antiretroviral therapy regimen. J Acquir Immune Defic Syndr. 2000;27:20–29.
13. Cook JA, Cohen MH, Grey D, et al. Use of highly active antiretroviral therapy in a cohort of HIV-seropositive women. Am J Public Health. 2002;92:82–87.
14. Ayanian JZ, Weissman JS, Chasan-Taber S, et al. Quality of care by race and gender for congestive heart failure and pneumonia. Med Care. 1999;37:1260–1269.
15. Mukamel DB, Murthy AS, Weimer DL. Racial differences in access to high-quality cardiac surgeons. Am J Public Health. 2000;90:1774–1777.
16. Jensen-Fangel S, Pedersen L, Pedersen C, et al. The effect of race/ethnicity on the outcome of highly active therapy for human immunodeficiency virus type 1-infected patients. Clin Infect Dis. 2002;35:1541–1548.
17. Giordano TP, Wright JA, Hasan MQ, et al. Do sex and race/ethnicity influence CD4 cell response in patients who achieve virologic suppression during antiretroviral therapy. Clin Infect Dis. 2003;37:433–437.
© 2004 Lippincott Williams & Wilkins, Inc.