Mother-to-infant HIV transmission can occur during the intrauterine period, the intrapartum period, or postnatally through breast-feeding. The evidence for intrauterine infection is based on published observations of HIV in aborted fetuses from as early as 8 weeks of gestation(1), reports of increased risk of transmission with placental membrane inflammation(2), and positive blood cultures or polymerase chain reaction (PCR) assays at birth(3,4). However, only 20% to 50% of HIV-infected infants have positive HIV cultures or positive DNA PCR assays in the first few days of life(5) despite the fact that, at least in newly infected adults, it takes a short time for cultures to become positive after infection(6,7). Because more than 85% of HIV-infected infants have positive cultures or PCR assays by 1 month of age and because that proportion changes little over the next 5 months of life, it has been postulated that transmission in the children who were culture negative at birth occurred during the intrapartum period(3,8).
Bryson et al.(9) proposed a definition for differentiating in utero from intrapartum transmission of HIV-1. They hypothesized that early positive HIV cultures or PCR assays (<48 hours) are the result of intrauterine infection and that positive assays only after 7 days of age reflect infection occurring during the intrapartum period. However, because it is not possible to directly ascertain the time of transmission, it is necessary to resort to indirect observations to support this hypothesis. For example, differences in clinical state or immunologic status between those who are culture or PCR positive at birth compared with those positive at more than 7 days of age would lend credibility to defining in utero versus intrapartum transmission based on timing of the first positive culture.
In this communication, we present immunologic data from infants born of HIV-infected women enrolled in the Women and Infants Transmission Study (WITS) to support the hypothesis. Because the initial study visit in WITS could occur through the first 7 days of life, we defined early events as occurring at ≤7 days of age andlate events as >7 days of age. We chose a subset of CD8+ lymphocytes as the marker of immunologic change. CD8+ lymphocytes increase during the course of HIV infection(10). The human leukocyte DR antigen on CD8+ lymphocytes is a marker that appears on lymphocytes shortly after activation(11). We postulated that infants infected in utero have positive HIV cultures at≤7 days of age and also have elevated CD8+DR+ lymphocytes. We postulated that infants infected during the intrapartum period do not have positive HIV cultures or CD8+ activation until infection is established at some time after birth. The findings contribute to the developing picture of the pathogenesis of perinatal HIV transmission.
The infants who participated in this study were enrolled in WITS, a multisite natural history study of perinatal HIV infection in the United States. HIV-infected women were enrolled at one of the study centers(i.e., Boston and Worcester, Massachusetts; New York City, New York; San Juan, Puerto Rico; and Chicago, Illinois). The appropriate Institutional Review Boards approved participation in the project at all sites. Mothers of the infants were enrolled in WITS as early as possible during the pregnancy. Their infants were followed prospectively with visits at ≤7 days of age as well as at 1, 2, 4, 6, 9, 12, 15, and 18 months and every 6 months thereafter. Because this report is focused on the early consequences of HIV infection, the data reported herein are limited to study visits during the first 6 months of life. The visit windows were 0 to 7 days for the first visit ±2 weeks for the 1- and 2-month visits, and ±4 weeks for the 4- and 6-month visits. The analysis was done using the nominal age of these visit windows.
The WITS cohort of first-born singleton infants of HIV-infected mothers included 599 infants as of March 1994. The analysis for this report included only infants who had culture results available from two or more visits (working definition of HIV infection is described later), had a culture result from ≤7 days of life, and had attained sufficient age to define the infection status. Of 455 infants with known infection status, 351 were uninfected infants, and 104 were infected. Of the 144 infants not included in the analysis, 30 were infected infants who were missing one or more data points for classification into the early or late culture positive group, and the remaining were infants of indeterminate infection status, principally because of not having reached an age at which infected status could be determined. Among the infected infants, 19 had positive cultures at ≤7 days of age, and 55 had negative cultures at ≤7 days of age but positive cultures at 1 month of age or later. Not all infants had flow cytometry results at all time points. The actual number of infants reported in each age group therefore varied and is indicated in the results.
Determining HIV Infection Status
Infection status was determined by HIV co-culture. Blood for the culture was obtained from peripheral veins (not the umbilical cord) at≤7 days and at each of the later time points. The assay was performed using the National Institute of Allergy and Infectious Diseases (NIAID) AIDS Clinical Treatment Group (ACTG) consensus quantitative coculture technique(12). Each of the sites participated in the NIAID virology quality assurance program and was fully certified by that program. The details and overall results of HIV culture in this cohort are described elsewhere(3). DNA PCR studies were also done on some of the specimens. DNA PCR was performed on EDTA-anticoagulated blood as described previously(13).
For the purposes of this analysis, an infant was defined as infected when two or more peripheral blood cultures were positive and as uninfected when two or more cultures, all obtained at 1 month of age or older and one that was obtained at or after 6 months of age, were negative, and there were no positive cultures. Infants not meeting these criteria were defined as having indeterminate infection and were excluded from analysis.
The timing of a positive culture was defined, for the purposes of this communication, as early if the culture drawn at birth through 7 days of age was positive and late if the ≤7-day culture was negative and the culture at the nominal age of 1 month or older was positive.
Samples were prepared from EDTA-anticoagulated blood using a whole blood lysis system according to the Division of AIDS flow cytometry guidelines(14). Identical reagents (Becton-Dickinson, San Jose, CA, U.S.A.) were used at all sites and consisted of an isotype control (IgG1, IgG2a), CD45/CD14 (gating reagent), CD4/CD8 (compensation reagent), CD3/CD4 (CD4% cells), and CD8/HLA-DR (activated CD8 subset) that were conjugated to fluorescein isothiocyanate and phycoerythrin, respectively. All sites used FACScan (Becton-Dickinson) instruments. All sites participated in the ACTG flow cytometry proficiency testing program and in a special WITS-wide quality assurance program. The specific methods and the results of the quality assurance program have been published(15). For presentation of results, only the relative (%) counts were used because of the improved precision of the results when the blood count and differential were not required and the fact that on occasion, the blood count was not obtained. which resulted in an increased number of missing data points in the analysis.
A fixed effects multiple regression model with repeated measurements(16) was used to estimate the association between early versus late culture positivity and CD8+DR+ lymphocytes. The predictors of the regression model included an indicator variable of early culture positivity, indicator variables for each nominal age, interaction terms between early culture positivity and age indicator variables, an indicator variable of zidovudine therapy in the mother during pregnancy, and time-varying covariate indicating cumulative zidovudine therapy in the infant (nonuse versus any prior used. Because each infant contributed data at more than one time point, the repeated measurement approach was used to account for the within-subject correlation of the measurements. The regression coefficients were estimated using an unstructured within subject correlation model. The PROC MIXED procedure in the statistical computer package SAS(17) was used for analysis. This procedure does not require infants to have data at all time points, and therefore the use of each infant's data was maximized.
The values of the percentages of CD8+ and HLA-DR+ lymphocytes(CD8+DR+%) were logarithm transformed to approximate a normal distribution, which is an assumption for the regression model we used. Because of the logarithmic transformation for the values of CD8+DR+, the estimate of the regression coefficient of the indicator variable of early positive cultures can be transformed to the estimate of the mean ratio of early versus late positive cultures. When the mean ratios of early positive cultures versus late positive cultures were significantly different at different ages, as indicated by the interaction term between early culture positivity and age, we further tested the difference between early and late positive cultures at each age.
A similar regression model was used to compare CD4+% in infants with early versus late positive cultures. CD4+% values were not transformed because the distribution of values was close to a normal distribution. The estimate of the regression coefficient of the indicator variable of early positive cultures in this model estimated the mean difference in CD4+% between infants with early and late positive cultures.
We believed that the skewed distribution of CD8+DR+% at ≤7 days of age suggested the possibility of a subpopulation of patients embedded in the whole group who were physiologically different from those with lower values. For some analyses, the infants were stratified on the basis of having CD8+DR+% at ≤7 days of age in the highest quartile or in the lower quartiles. The highest quartile CD8+DR+% included those with counts of 7% or higher. The median CD8+DR+% of infected infants in the highest quartile was 9%, and it was 2% for infected infants in the lower quartiles.
A multiple regression model with repeated measurements was also used to estimate the difference in CD4+% between infected infants who had elevated CD8+DR+% at ≤7 days and those who did not. The indicator variables of zidovudine therapy in the mother during pregnancy and in the infant were also included as predictors in the regression model.
The proportion of lymphocytes from uninfected infants at ≤7 days of age bearing the CD8+DR+ marker was modest (median 2.0%, mean 2.6 ± 4.4% [mean ± standard deviation]) and increased slowly during the first 6 months of age (median at 6 months 4.0%, mean 5.0 ± 4.1%) (Fig. 1). Lymphocytes from infants with late positive cultures had similar CD8+DR+% at ≤7 days as the uninfected (median 2.0%, mean 3.0 ± 3.2%). However, CD8+DR+% increased rapidly (median at 1 month of 6.0%, mean 8.5 ± 7.6%) and reached a plateau between 2 and 6 months of age (median at 6 months of 12.0%, mean 13.5 ± 7.4%). In contrast, CD8+DR+ lymphocytes from infants with early positive cultures was higher at≤7 days of age (median 5.0%, mean 6.6 ± 5.8%) than the uninfected or the late culture positive infants and increased to a plateau by 1 month (median 11.5%, mean 12.9 ± 7.8%) that persisted through 6 months (median 12.0%, mean 13.2 ± 7.6%).
The differences between early and late positive cultures were assessed by fitting a fixed effects multiple regression model, as described previously. The regression analysis showed that the mean ratios of CD8+DR+% for early versus late culture positive infants varied at different ages (F4,254 = 4.09, p = 0.0031). We therefore compared the mean CD8+DR+% of infants with early versus late positive cultures at each month of age separately. Adjusting for zidovudine therapy in the mother and infant in the regression model, the mean values of CD8+DR+% were more than twice as high in infants with early compared with late positive cultures at ≤7 days of age (mean ratio of early to late positive cultures was 2.43, 95% CI = 1.47 to 4.03). Significant differences were also observed at 1 month (mean ratio 1.86, 95% CI= 1.14 to 3.03) and 4 months (mean ratio 1.86, 95% CI= 1.11 to 3.09) but not at 2 and 6 months of age. The significant result at 4 months of age is not a robust finding, because unlike the other results in this report, the significance varied with the statistical methods. When we did not adjust for zidovudine therapy in the mother and infant, the significance level was 0.05 instead of 0.018. When we used nonparametric univariate analysis, the difference was not significant (p = 0.13). However, overall, the activated lymphocytes, as identified by the percentage of cells bearing the CD8+DR+ marker, were significantly more common in those who were culture positive in the first week of life than in those who were not.
We also investigated in a preliminary manner whether the conclusions would be altered if DNA PCR were used for defining infection status. Nineteen infants had early positive HIV-1 cultures or early positive DNA PCR test results. Fifteen (88%) of 17 who had both tests performed had concordant results. The discrepant results were for one infant who had a negative early and positive late HIV culture but had an early positive DNA PCR and for one infant who had an early positive HIV culture but an early negative DNA PCR assay. The CD8+DR+% was lower in the latter patient than the median of infants overall who had early positive cultures.
We investigated whether the values in the right tail of the distribution of CD8+DR+% (highest quartile) came from infected or uninfected infants, as described in the Methods section. When the CD8+DR+% at ≤7 days of age was stratified on the basis of being in the highest quartile (CD8+DR+ ≥7%) versus the lower quartiles, a significantly greater proportion of infected infants was in the highest (12 [26%] of 47) than uninfected (1 [8%] of 163) (χ2 = 10.7, p = 0.001).
Because one of the nearly universal findings in HIV infection is an eventual decline in the percent of CD4+ lymphocytes over time, we examined the difference in CD4+% in infants with early positive cultures compared with late. There was no significant difference between the two infected groups in CD4+% at any age from ≤7 days through 6 months of age. The mean CD4+% at ≤7 days of age in infants with early and late positive cultures was 49% and 51%, respectively. The mean CD4+% in infants with early versus late positive cultures at later ages also were not significantly different from each other(40% versus 44% at 1 month, 38% versus 37% at 2 months, 32% versus 31% at 4 months, and 33% and 34% at 6 months). The CD4+% in uninfected infants at ≤7 days (52% ± 11%, n = 237) was similar to the infected infants but declined at a much slower rate than the infected and, by 6 months of age, was 47 ± 8% (n = 276).
In contrast to the lack of effect of early compared with late culture positivity on CD4+%, the CD8+DR+% from the first week of life did have an effect. Stratification of infected infants by CD8+DR+% at ≤7 days of age revealed that the CD4+% was substantially lower during the first 4 months of life in the high CD8+DR+% group compared with the low CD8+DR+% group(Fig. 2). The mean difference of CD4+% was significantly different from 0 in these groups at ≤7 days through 4 months of age (≤7 days p = 0.035, 1-month p = 0.048, 2-monthp = 0.042, and 4-month p = < 0.0005). Activation of CD8+ lymphocytes, presumably by intrauterine infection, was associated with lower CD4+ lymphocyte percentages during the first 4 months of life than with later infection. However, by 6 months of age, there was no longer a significant difference.
Infants with high CD8+DR+% at birth did not have different rates of low birth weight or shorter gestations than those with lower CD8+DR+% (p = 1.00 and p = 0.79, respectively). However, overall, more infected infants had low birth weight (<2500 g) than those uninfected but not shorter gestations (<37 weeks) than those uninfected (p = 0.013 and p = 0.26 byχ2, respectively).
The timing of the transmission of HIV from infected mother to her infant is critically important in understanding the pathogenesis of perinatal HIV infection and in targeting intervention strategies. In an attempt to provide a uniform structure to studies of timing and pathogenesis, it has been proposed that early culture positivity is related to intrauterine transmission and late positivity with intrapartum transmission(9).
Our definition of early infection (i.e., positive culture at ≤7 days) differed from that proposed by Bryson et al.(i.e., positive test at ≤48 hours)(9). The proposal that early or intrauterine infection be defined as virologic positivity ≤48 hours was not based on experimental data, but rather was a hypothesis for further evaluation. The WITS protocol was established before Bryson's proposal. The WITS protocol allowed the initial HIV-1 culture to be drawn up to 7 days of age, a more liberal definition of early test positivity than that of Bryson. However, in a metanalysis by Dunn et al., the estimated sensitivity of PCR for diagnosis was relatively constant from birth to day 8, increased rapidly from 8 to 14 days of age, and reached a plateau thereafter(18). Similarly, in an analysis of WITS data on use of HIV-1 culture for diagnosis, the sensitivity of a positive culture was stable during the first week of life and increased rapidly thereafter(19). These findings are consistent with our use of an early infection definition that includes a positive virologic test between 0 and 7 days of age. For this article, we therefore defined early culture positivity as occurring at ≤7 days of age. We then examined the relation between early culture positivity and changes in an immunologic activation marker, CD8+DR+, that is known to rise shortly after antigenic stimulation.
The key observation was that the proportion of activated CD8+ lymphocytes (CD8+DR+) at ≤7 days of life was significantly greater in those infants who had early positive HIV cultures. The demonstration of an association of early culture positivity with an immunologic marker, CD8+DR+, that rises after a delay of a number of days is consistent with the hypothesis that intrauterine infection correlates with early culture positivity. An example is measles infection, in which the levels of activation markers rise 1 to 2 weeks after the beginning of the infection(20). We suggest that those who presumably are infected around delivery do not have evidence of activation and that those who have been infected for some period do have evidence of stimulation. Also consistent with our findings is the report by Gesner et al.(21) that soluble CD8, probably shed from CD8+DR+ lymphocytes on stimulation, is higher in infants who are culture positive at less than 1 month of age than those who are culture positive at 1 month to 1 year of age.
The difference in the CD8+DR+% between early culture-positive patients and late culture-positive patients was also statistically significant at 1 and 4 months of age but not at 2 months (p = 0.20) (Fig. 1). The differences observed at different ages may be a consequence of the relatively small sample, or it may be the sum of the effects of a "wave" of progressive immunologic changes from intrauterine infection being superimposed on the effects of peripartum infection. There may also be other physiologic effects that cannot be teased out without a much larger cohort. By 6 months, it is likely that any effect of the time of infection has been overcome by the many other factors that affect the viral load and its immunologic consequences.
We also examined DNA PCR as a technique for establishing the diagnosis of infection within the first week of life in a preliminary manner and compared it with culture. The DNA PCR results were concordant in 88% of cases in which both techniques were examined. However, because of the limited number of subjects who had DNA PCR assays, the conclusions must be limited.
We also found that CD4+ counts in infected infants, as we have reported previously(22), begin to diverge from those of uninfected infants shortly after birth. The CD4+% of the infected group as a whole does not differ significantly between those who had presumed early compared with late infection. However, when the CD4+% of the infected infants showing the greatest degree of antigen stimulation (manifested by a CD8+DR+ increase) were compared with those with less stimulation, significant differences emerged. The CD4+ counts fell more in those with the greatest elevation of CD8+DR+ during the first 4 months of life. The disparity was gone by 6 months of age.
We questioned why there is not a significant relation between early positive cultures and later CD4+ counts but there is a significant relation between CD8+DR+% and later CD4+ counts. It is likely that there are differences in the timing of intrauterine infection such that some fetuses are infected early and have time to mount a response to the initial exuberant burst of unchecked virus production that results in infection and destruction of many CD4+ cells. Others are infected later in gestation and have had little time to mount a significant response. However, after the first few months of an infection, the immune response, including the rise in the number of CD8+DR+ lymphocytes, tempers the infection and the rate of destruction of virus infected cells resulting in stabilization of CD4+ counts and reduced disparity between early- and late-infected infants. The lack of association of early culture positivity with CD4+% early in life may reflect the heterogeneity of infection timing and viral load among that group of infants who are infected in utero.
There may be other explanations for the finding of immune activation in association with early culture positivity besides the postulate that it is a marker for intrauterine infection. For example, early elevation of CD8+DR+ lymphocyte counts may be the consequence of a larger viral inoculum than those who had the elevation later. This explanation seems unlikely because the initial studies were drawn within the first week of life, which would allow little time for the neonatal immune system to be activated. Even at ≤7 days, there was a difference in CD4+ lymphocyte concentration in those with the highest CD8+ DR+ concentration. Therefore, it is likely that the process had been underway for some time before delivery.
Some studies have shown premature birth to be associated with increased risk of transmission, and some investigators have hypothesized that in utero HIV infection might be associated with intrauterine growth retardation(23,24). We therefore examined whether the birth weight and duration of gestation were different in the infants who had elevated CD8+DR+% compared with those with lower CD8+DR+% at ≤7 days of age. We found no significant difference in the birth weight or duration of gestation in infants with elevated or lower CD8+DR+%.
Previous studies have been hampered by the small number of patients available for study in a standardized manner. Prospective, longitudinal, multisite perinatal cohort studies, such as WITS, offer the opportunity to examine the earliest events in HIV infection because, unlike most studies with adults, the patients are identified prospectively and the events studied before they have passed. Our findings support the hypothesis that early culture positivity in infants is likely the result of intrauterine infection and late culture positivity the result of peripartum infection.
Acknowledgments: The authors wish to thank Donald Brambilla, Ph.D., for the initial statistical analysis and Les Kalish, Sc.D., for thoughtful statistical comments and review of the manuscript. The material was presented in part (abstract 131) at the 34th Interscience Conference on Antimicrobial Agents and Chemotherapy, Orlando, FL, October 4-7, 1994. Financial support was provided by the National Institutes of Health (AI-82507, AI-82506, AI-85005, AI-34841, and AI-05072; HD-8-2913). The principal investigators, study coordinators, and program officers include Ruth Tuomala, Ellen Cooper, Donna Mesthene (Brigham and Women's Hospital and Boston City Hospital, Boston, MA), Harold Fox, Jane Pitt, Alice Higgins (Columbia Presbyterian Hospital, New York, NY), Sheldon Landesman, Hermann Mendez, Gail Moroso(State University of New York, Brooklyn, NY), Clemente Diaz, Edna Pacheco-Acosta (University of Puerto Rico, San Juan), Kenneth Rich, Geraldine Alexander (University of Illinois at Chicago, IL), Mary Glenn Fowler, Judy Lew (National Institute of Allergy and Infectious Diseases, Bethesda, MD), Lynne Mofenson, Jack Moye (National Institute of Child Health and Human Development, Bethesda, MD), and Sonja McKinlay, Kathy Sherrieb (New England Research Institute, Watertown, MA).
1. Lewis SH, Reynolds-Kohler C, Fox HE, Nelson JA. HIV-1 in trophoblastic and villous Hotbauer cells, and haematological precursors in eight-week fetuses. Lancet
2. St. Louis ME, Kamenga M, Brown C, et al. Risk for perinatal HIV-1 transmission according to maternal immunologic, virologic, and placental factors. JAMA
3. McIntosh K, Pitt K, Brambilla D, et al. Blood culture in the first 6 months of life for the diagnosis of vertically transmitted human immunodeficiency virus infection. J Infect Dis
4. Simonon A, Lepage P, Karita E, et al. An assessment of the timing of mother-to-child transmission of human immunodeficiency virus type 1 by means of polymerase chain reaction. J Acquir Immune Defic Syndr
5. Mofenson L. Epidemiology and determinants of vertical HIV transmission. Semin Pediatr Infect Dis
6. Nielsen C, Pedersen C, Lundgren JD, Gerstoft J. Biological properties of HIV isolates in primary HIV infection: consequences for the subsequent course of infection. AIDS
7. Joubert L, Meyohas M-C, Oliver I, et al. Viral load in symptomatic primary HIV infection. AIDS
8. De Rossi A, Ometto L, Mammano F, Zanotto C, Giaquinto C, Chieco-Bianchi L. Vertical transmission of HIV-1: lack of detectable virus in peripheral blood cells of infected children at birth. AIDS
9. Bryson YJ, Luzuriaga K, Sullivan JL, Wara DW. Proposed definitions for in utero versus intrapartum transmission of HIV-1. N Engl J Med
10. Plaeger-Marshall S, Hultin P, Bertolli J, et al. Activation and differentiation antigens on T cells of healthy, at-risk, and HIV-infected children. J Acquir Immune Defic Syndr
11. Agostini C, Pizzolo G, Zambello R, et al. Shedding of the soluble form of the CD8 complex by CD8+
cells in HIV-1-infected patients. AIDS
12. Hollinger FB, ed. ACTG virology manual for HIV laboratories
, version 2.1. Bethesda, MD: Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institute of Health, September 1993.
13. Bremer JW, Lew J, Cooper E, et al. Diagnosis of infection with human immunodeficiency virus type 1 by a DNA polymerase chain reaction assay among infants enrolled in the Women and Infants Transmission Study. J Pediatr
14. Calvelli T, Denny TN, Paxton H, Gelman R, Kagan J. Guideline for flow cytometric immunophenotyping: a report from the National Institute of Allergy and Infectious Diseases, Division of AIDS. Cytometry
15. Landay AL, Brambilla D, Pitt J, Hillyer G, et al. Interlaboratory variability of CD8 subset measurements by flow cytometry and its application to multicenter clinical trials. Clin Diagn Lab Immunol
16. Wara JH. Linear models for the analysis of longitudinal studies. Am Statistician
17. SAS Technical Report P-229. SAS/STAT Software: changes and enhancements. Release 6.07. Cary, NC: SAS Institute, 1992.
18. Dunn DT, Brandt CD, Krivine A, et al. The sensitivity of HIV-1 DNA polymerase chain reaction in the neonatal period and the relative contributions of inter-uterine and intra-partum transmission. AIDS
19. Kalish LA, Pitt J, Lew J, et al. Defining the time of fetal or perinatal acquisition of human immune deficiency virus type 1 infection on the basis of age at first positive culture. J Infect Dis
20. Griffin D, Ward B, Jauregui E, Johnson R, Vaisberg A. Immune activation in measles. N Engl J Med
21. Gesner M, DiJohn D, Krasinski K, Borkowsky W. Increased soluble CD8 (sCD8) in human immunodeficiency virus 1-infected children in the first month and year of life. Pediatr Infect Dis J
22. Brambilla DJ, Rich K, Landay A, et al. Early differences in lymphocyte subsets between HIV+ and HIV− infants [abstract POB01-0865]. The 9th International Conference on AIDS, Berlin, June 6-11, 1993.
23. European Collaborative Study: Risk factors for mother-to-child transmission of HIV-1. Lancet
24. Nair P, Alger L, Hines S, Seiden S, Hebel R, Johnson JP. Maternal and neonatal characteristics associated with HIV infection in infants of seropositive women. J Acquir Immune Defic Syndr