Breast milk alpha-defensins are associated with HIV type 1 RNA and CC chemokines in breast milk but not vertical HIV type 1 transmission

Alpha-defensins are proteins exhibiting in vitro anti-HIV-1 activity that may protect against mother-to-child transmission of HIV-1 via breast milk. Correlates of alpha-defensins in breast milk and transmission risk were determined in a cohort of HIV-1-infected pregnant women in Nairobi followed for 12 months postpartum with their infants. Maternal blood was collected antenatally and at delivery for HIV-1 viral load and infant HIV-1 infection status was determined < 48 h after birth and at months 1, 3, 6, 9, and 12. Breast milk specimens collected at month 1 were assayed for alpha-defensins, HIV-1 RNA, subclinical mastitis, and CC and CXC chemokines. We detected alpha-defensins in breast milk specimens from 108 (42%) of 260 HIV-1-infected women. Women with detectable alpha-defensins (> or =50 pg/ml) had a median concentration of 320 pg/ml and significantly higher mean breast milk HIV-1 RNA levels than women with undetectable alpha-defensins (2.9 log(10) copies/ml versus 2.5 log(10) copies/ml, p = 0.003). Increased alpha-defensins concentrations in breast milk were also associated with subclinical mastitis (Na (+)/K(+) ratio > 1) and increased breast milk chemokine levels. Overall, 40 (15%) infants were HIV-1 uninfected at birth and subsequently acquired HIV-1. There was no significant association between month 1 alpha-defensins and risk of HIV-1 transmission. In conclusion, alpha-defensins were associated with breast milk HIV-1 viral load, chemokine levels, and subclinical mastitis, all of which may alter risk of infant HIV-1 acquisition. Despite these associations there was no significant relationship between breast milk alpha-defensins and mother-to-child transmission, suggesting a complex interplay between breast milk HIV-1, inflammation, and antiinfective factors.     

High frequency of rapid immunological progression in African infants infected in the era of perinatal HIV prophylaxis

OBJECTIVES: To determine the natural history of HIV infection following peripartum single-dose nevirapine (sd-NVP) prophylaxis in a resource-limited country, and to assess implications for antiretroviral therapy (ART) roll-out programmes. METHODS: Infants of HIV-infected mothers in KwaZulu-Natal, South Africa, were tested on days 1 and 28 to detect intrauterine (IU) and intrapartum (IP) infection. Infant follow-up included monthly viral load and CD4 cell measurement. ART was initiated at infant CD4 cell% < or = 20%. RESULTS: In 740 infants born to 719 HIV-infected women, mother-to-child transmission (MTCT) was 10.3% (69% IU, 31% IP). Median viral load was higher in mothers of infants infected IP than IU (279 000 versus 86 600 copies/ml; P = 0.039) and lower in mothers of uninfected infants (median 26 750 copies/ml; P < 0.001). Peak viraemia was higher in infants infected IP than IU (5 160 000 versus 984 000 copies/ml; P < 0.001). Median viral load at birth in IU-infected infants (155 000 copies/ml) fell 1.4 log to 6510 copies/ml by day 5 and was beneath the detection limit using dried blood spot analysis in 38% of infants. CD4 cell% declined rapidly, to < or = 20% in 70% and < or = 25% in 85% [current World Health Organization (WHO) criteria for initiating ART] of infants by 6 months. CONCLUSIONS: MTCT was reduced by sd-NVP through an effect on IP transmission. Where MTCT occurred despite NVP, two-thirds of transmissions arose IU; IP-infected babies were born to mothers with very high viral load. Disease progression was particularly rapid, 85% infants meeting WHO criteria for ART within 6 months. These findings argue for more effective MTCT-prevention programmes in resource-limited countries.     

Quantification of RNA in HIV type 1 subtypes D and G by NucliSens and Amplicor assays in Abidjan, Ivory Coast

We have compared the performance of the NucliSens and the standard and modified HIV Monitor assays to quantify HIV-1 RNA plasma viral load in 12 tuberculosis patients infected with HIV-1 env subtype D (n = 3) and env subtype G (n = 9) in Ivory Coast. RNA was quantified in all nine subtype G specimens by the modified Amplicor HIV Monitor (mean, 4.6 log10 copies/ml; range, 3.1-6.3 log10/ml), in seven specimens by NucliSens (mean, 4.4 log10 copies/ml; range, 2.7-5.5 log10 copies/ml), and in 6 specimens by the standard Amplicor HIV Monitor assay (mean, 4.2 log10 copies/ml; range, 3.5-5.0 log10 copies/ml). All three subtype D samples were amplified by both the modified Amplicor HIV Monitor (mean, 4.5 log10 copies/ml; range, 3.8-5.1 log10 copies/ml) and NucliSens (mean, 3.8 log10 copies/ml; range, 2.8-5.0 log10 copies/ml); two samples were quantified by the standard Amplicor HIV Monitor assay (mean, 3.0 log10 copies/ml; range, 2.4-3.6 log10 copies/ml). Our preliminary results suggest that the modified Amplicor HIV Monitor can accurately quantify HIV-1 RNA viral load in persons infected with subtype D and G strains.     

Dual infection with human immunodeficiency virus type 1 and type 2: impact on HIV type 1 viral load and immune activation markers in HIV-seropositive female sex workers in Abidjan, Ivory Coast

To determine the impact of dual infection with HIV-1 and HIV-2 on HIV-1 viral load and markers of immune activation among HIV-seropositive FSWs in Abidjan, we analyzed blood samples obtained from consenting HIV-seropositive FSWs attending a confidential clinic between September 1996 and June 1997 in Abidjan. Among HIV-1 and HIV-2 dually seropositive FSWs, polymerase chain reaction (PCR) testing with HIV-1 and HIV-2 primers was used to differentiate between FSWs who were PCR positive only for HIV-1 and those positive for both HIV-1 and HIV-2 (dually infected). Of the 203 FSWs, 151 (74%) were HIV-1 seropositive only (median age, 26 years), 4 (2%) were HIV-2 seropositive, and 48 (24%) were dually seropositive (median age, 30 years). Of the 48 dually seropositive FSWs, 33 (69%) were dually infected and 15 (31%) were dually seropositive. Median CD4+ T cell counts per microliter were not significantly different among the three groups (525 for HIV-1 positive only, 502 for dually infected, and 416 for dually seropositive) (p = 0.14). Median viral load (log10 copies/ml) was not significantly different among the HIV-1-only FSWs (4.8 log10 copies/ml) compared with the 32 dually infected FSWs (4.6 log10 copies/ml) and 14 dually seropositive FSWs (4.7 log10 copies/ml; p = 0.95). Median levels of HLA-DR immune activation were increased in both CD4+ and CD8+ T cells for the dually infected (n = 27) FSWs compared with those infected with HIV-1 only (n = 123) (p = 0.019 and p = 0.01, respectively). Dual infection does not appear to influence levels of HIV-1 viral load in vivo. However, levels of HLA-DR are higher among FSWs dually infected with HIV-1 and HIV-2 than among those infected with HIV-1 only.     

Maternal viral load and timing of mother-to-child HIV transmission, Bangkok, Thailand. Bangkok Collaborative Perinatal HIV Transmission Study Group

OBJECTIVES: To determine the proportion of HIV-1-infected infants infected in utero and intrapartum, the relationship between transmission risk factors and time of transmission, and the population-attributable fractions for maternal viral load. DESIGN: Prospective cohort study of 218 formula-fed infants of HIV-1-infected untreated mothers with known infection outcome and a birth HIV-1-positive DNA PCR test result. METHODS: Transmission in utero was presumed to have occurred if the birth sample (within 72 h of birth) was HIV-1-positive by PCR; intrapartum transmission was presumed if the birth sample tested negative and a later sample was HIV-1-positive. Two comparisons were carried out for selected risk factors for mother-to-child transmission: infants infected in utero versus all infants with a HIV-1-negative birth PCR test result, and infants infected intrapartum versus uninfected infants. RESULTS: Of 49 infected infants with an HIV-1 birth PCR result, 12 (24.5%) [95% confidence interval (CI), 14 -38] were presumed to have been infected in utero and 37 (75.5%) were presumed to have been infected intrapartum. The estimated absolute overall transmission rate was 22.5%; this comprised 5.5% (95% CI, 3-9) in utero transmission and 18% (95% CI, 13-24) intrapartum transmission. Intrapartum transmission accounted for 75.5% of infections. High maternal HIV-1 viral load (> median) was a strong risk factor for both in utero [adjusted odds ratio (AOR) 5.8 (95% CI, 1.4-38.8] and intrapartum transmission (AOR, 4.4; 95% CI, 1.9-11.2). Low birth-weight was associated with in utero transmission, whereas low maternal natural killer cell and CD4(+) T-lymphocyte percentages were associated with intrapartum transmission. The population-attributable fraction for intrapartum transmission associated with viral load > 10 000 copies/ml was 69%. CONCLUSIONS: Our results provide further evidence that most perinatal HIV-1 transmission occurs during labor and delivery, and that risk factors may differ according to time of transmission. Interventions to reduce maternal viral load should be effective in reducing both in utero and intrapartum transmission.     

Lower levels of HIV RNA in semen in HIV-2 compared with HIV-1 infection: implications for differences in transmission

BACKGROUND AND OBJECTIVES: HIV-2 infection, in comparison with HIV-1, is characterized by lower plasma viral loads, slower CD4 cell count decline, decreased AIDS-related mortality, and lower rates of mother-to-child and sexual transmission. To gain further insight into why HIV-1 is more readily transmitted as compared with HIV-2, we analyzed semen and plasma HIV RNA levels in HIV-1 and HIV-2-positive men from Senegal. DESIGN AND METHODS: Twenty-two HIV-1 and 10 HIV-2-infected subjects from the University of Dakar donated semen and blood samples for this analysis. HIV-1 and HIV-2 viral loads in semen and plasma were quantified using type-specific polymerase chain reaction assays. RESULTS: The mean age of the subjects was 37 and 40 years; mean CD4 cell count was 222 and 276 cells/microl and the mean plasma viral load was 4.7 and 3.0 log10 copies/ml for HIV-1 and HIV-2, respectively (P = 0.002). HIV RNA was detected in semen in 21 of 22 (95%) of HIV-1 and seven of 10 (70%) of HIV-2-infected subjects; P = 0.07). However, the levels of HIV RNA present in semen were markedly different between those with HIV-1 and HIV-2, with a mean of 4.4 log10 copies/ml among those with HIV-1 and a mean of 2.6 log10 copies/ml among those with HIV-2 (P < 0.001). In multivariate analysis, plasma viral load and HIV type, but not CD4 cell count, were independently predictive of semen viral load (P = 0.03, 0.05, 0.48, respectively) CONCLUSIONS: These data suggest that differences in semen viral load between HIV-1 and HIV-2 may be in part responsible for the markedly different transmission rates of these two viruses. In addition, risk of male genital tract shedding strongly correlates with plasma viral loads. Interventions that decrease viral load may help decrease transmission of both HIV-1 and HIV-2.     

Maternal HIV-1 DNA load and mother-to-child transmission

While many factors contribute to mother-to-child transmission (MTCT) of HIV-1, maternal plasma HIV-1 RNA viral load (RNA-VL) has been consistently found as the main risk factor, including when antiretroviral prophylaxis was used to prevent MTCT. However the predictive value of RNA-VL is poor. A recent study of HIV-1-positive pregnant women who did not receive antiretroviral prophylaxis reported an association between HIV-1 DNA viral load (DNA-VL) and MTCT that was stronger than the association between RNA-VL and MTCT. We sought to determine if HIV-1 DNA-VL was independently associated with MTCT of HIV in a population of women who received zidovudine prophylaxis during pregnancy and whose infants received zidovudine after birth. Patients were 33 non-breastfeeding transmitting (TR) and 33 nontransmitting mothers (NTR) from Perinatal HIV Prevention Trial (PHPT-1), a multicenter clinical trial conducted in Thailand comparing zidovudine prophylaxis durations to prevent MTCT. TR and NTR mothers were matched according to baseline RNA-VL. Maternal peripheral blood mononuclear cell (PBMC)-associated HIV-1 DNA was extracted from whole blood, and DNA-VL was established by quantitative real-time polymerase chain reaction. We found that TR had a significantly higher cell-associated HIV-1 DNA viral load than did NTR. Median TR DNA-VL was 2.54 log(10) copies per microgram PBMC DNA, while it was 2.28 log(10) copies per microgram PBMC DNA in NTR (Wilcoxon p = 0.02). In summary, HIV-1 DNA viral load was associated with MTCT in a population of women who received antiretroviral prophylaxis during pregnancy, independently from RNA viral load.     

Early diagnosis of paediatric HIV-1 infection among African breast-fed children using a quantitative plasma HIV RNA assay

OBJECTIVE: To evaluate the performance of a quantitative plasma HIV-1 RNA assay for HIV infection diagnosis among African breast-fed children. METHODS: Serial plasma specimens collected in the first week, at day 45-90, 6 months and 9-12 months of age from HIV-exposed children born to HIV-1-infected women enrolled in the DITRAME ANRS 049a perinatal intervention trial (Abidjan, C?te d'Ivoire) were tested for HIV-1 plasma RNA using a branched DNA (bDNA) assay. Sensitivity and specificity of this RNA test were assessed in comparison with a qualitative DNA polymerase chain reaction (PCR) performed on the same blood samples and allowing a reliable detection of the predominant subtype A. RESULTS: Among 91 samples from 53 infected children which tested positive by DNA PCR, the sensitivity of the bDNA test was 100% [95% confidence interval (CI), 96.0-100.0] at < or = 8 days (n = 19), 6-12 weeks (n = 43), 6 months (n = 26), and 9-12 months (n = 3). The median plasma HIV-1 RNA viral load ranged from 242 000 copies/ml at < or = 8 days to more than 500 000 copies/ml at day 45-90 and at 6 months. Of 106 specimens from 106 uninfected children who were DNA PCR- negative at month 3 or 6 of age, HIV-1 RNA was undetectable in 103, yielding an overall specificity for the bDNA test of 97.2% (95% CI, 92.0-99.4). The viral load in the three remaining samples with false-positive results was low (410, 937 and 3752 copies/ml, respectively). CONCLUSIONS: The quantitative bDNA assay appears a suitable tool for early, reliable and easy diagnosis of paediatric HIV-1 infection among a population of African breast-fed children.     

广西地区未感染HIV婴儿T淋巴细胞及亚群的研究

目的探讨广西地区未感染艾滋病病毒（HIV）婴儿T淋巴细胞及亚群的基本特征,为初步排除并辅助诊断HIV阳性产妇所生婴儿是否感染HIV提供依据。方法收集68例未感染HIV婴儿（未感染组）和13例感染HIV婴儿（对照组）的抗凝全血,用四色荧光流式细胞术检测外周血CD3＋、CD4＋、CD8＋T淋巴细胞绝对数及其百分比,并计算CD4/CD8比值;用Nuclisens EasyQ方法检测血浆HIV-1病毒载量。结果 （1）未感染组外周血CD3＋、CD4＋、CD8＋T淋巴细胞绝对数（细胞/mm3）的中位数分别为4 5702、8181、410;CD3、CD4、CD8的百分比中位数分别为61.42%、39.39%、18.40%;CD4/CD8比值的中位数为2.25;在未感染组中,男性婴儿与女性婴儿的T淋巴细胞及亚群的比较,各检测指标差异没有统计学意义（P〉0.05）;血浆HIV-1病毒载量均为未检测出水平。（2）对照组外周血CD3＋、CD4＋、CD8＋T淋巴细胞绝对数的中位数分别为4 930、1 4603、326（细胞/mm3）;CD3、CD4、CD8百分比的中位数分别为69.95%、13.77%、54.69%;CD4/CD8比值的中位数为0.25;血浆HIV-1病毒载量的中位数为800 000（拷贝/ml）。（3）除CD3＋T淋巴细胞绝对数之外,未感染组与对照组的CD4＋、CD8＋T淋巴细胞绝对数、CD3、CD4、CD8百分比及CD4/CD8比值的比较,差异有统计学意义（P〈0.05）。结论在资源有限的情况下,参考婴儿T淋巴细胞及亚群的异常变化,可初步判断HIV阳性产妇所生婴儿是否感染HIV,T淋巴细胞及亚群的检测,可以作为一种辅助诊断HIV感染的手段。     

Whole blood versus plasma spots for measurement of HIV-1 viral load in HIV-infected African patients

The increasing availability of antiretroviral drugs to HIV-infected populations in developing countries highlights the need to develop field-friendly practical methods for HIV-1 viral load measurements to monitor the effects of treatment. We compared use of whole-blood spots versus plasma dried on filter paper to quantify HIV-1 viral load in 51 African patients with HIV-1. The mean log10 HIV-1 viral loads were 4.22 for dried plasma spots (DPS) and 4.20 for dried whole-blood spots (DBS). The difference between the pairs of log10 viral load for DPS and DBS were significantly correlated with their mean (Spearman's r=0.31, p=0.03). This correlation between the difference and mean of viral load was no longer evident for values of log10 DPS that were less than 5 (r=0.01, p=0.93). For the 38 paired values with log10DPS of less than 5, the mean difference (log10DPS-log10DBS) was -0.04 (SD 0.29). Dried whole blood stored on filter paper at room temperature shows potential as a field-friendly alternative to plasma for measurement of HIV-1 viral load.     

Maternal plasma human immunodeficiency virus type 1 RNA level: a determinant and projected threshold for mother-to-child transmission.

To prevent mother-to-child human immunodeficiency virus type 1 (HIV-1) transmission, it is important to identify its determinants. Because HIV-1 RNA levels can be reduced by antiviral therapy, we examined the role of maternal plasma HIV-1 RNA level in mother-to-child transmission. We used quantitative competitive PCR to measure HIV-RNA in 30 infected pregnant women and then followed their infants prospectively; 27% of the women transmitted HIV-1 to their infants and maternal plasma HIV-1 RNA level correlated strikingly with transmission. Eight of the 10 women with the highest HIV-1 RNA levels at delivery (190,400-1,664,100 copies per ml of plasma) transmitted, while none of the 20 women with lower levels (500-155,800 copies per ml) did (P = 0.0002). Statistical analysis of the distribution of HIV-1 RNA loads in these 30 women projected a threshold for mother-to-child transmission in a larger population; the probability of a woman with a viral RNA level of < or = 100,000 copies per ml not transmitting is predicted to be 97%. Examination of serial HIV-1 RNA levels during pregnancy showed that viral load was stable in women who did not initiate or change antiviral therapy. These data identify maternal plasma HIV-1-RNA level as a major determinant of mother-to-child transmission and suggest that quantitation of HIV-1 RNA may predict the risk of transmission.     

Plasma viral load threshold for sustaining intrahost HIV type 1 evolution

The objective of the present study was to determine if natural suppression of plasma viremia below the detection limit of commercial assays (50-80 copies HIV-1 RNA/ml) can contain the HIV-1 evolution. HIV-1 quasispecies complexity in PBMC DNA was assessed in the env gene at two time points in 14 long-term nonprogressors (LTNPs). Sequence changes consistent with viral evolution was found in all patients with a median plasma RNA viral load >100 copies/ml. Evidence of low-level viral evolution was detected in two of four patients with intermittent viremia and a median plasma HIV-1 RNA load of >80 copies/ml. No significant evolution was observed in the three LTNPs with persistent viral suppression below the detection limit. Overall, a significant positive correlation (p < 0.001) was observed between viral evolution and plasma RNA viral load in the LTNPs analyzed. These results suggest that the detection limit of ultrasensitive viremia assays could represent an important threshold below which intrahost HIV-1 evolution does not occur.     

Differences in HIV-2 plasma viral load and immune activation in HIV-1 and HIV-2 dually infected persons and those infected with HIV-2 only in Abidjan, Côte D'Ivoire

OBJECTIVE: To determine whether blood plasma levels of HIV-2 RNA viral loads and immune activation markers differ between persons infected with HIV-2 only and those dually infected with HIV-1 and HIV-2. METHODS: Between September 1996 and February 2000, we collected, analyzed and compared levels of HIV-2 RNA in plasma and immune activation markers among 52 persons infected with HIV-2 alone and 75 with confirmed dual infection. We also compared viral load and immune activation in patients who were infected with HIV-1 only and those who were dually infected. RESULTS: When we conducted a CD4 T-cell count-stratified multivariate analysis of HIV-2 viral load, controlling for difference in CD4 T-cell counts, age and sex: at < 200 x 10 CD4 T cells/l, HIV-2 viral load was 2.0 log10 copies/ml lower in dually infected patients than in HIV-2 only patients (P < 0.0001). At CD4 T-cell counts between 200 x 10 and 500 x 10/l, HIV-2 viral load was 0.3 log10 copies/ml lower in dually infected patients (P = 0.45). However, at CD4 T-cells counts > 500 x 10/l, HIV-2 viral load was 0.9 log10 copies/ml higher in dually infected patients (P < 0.0001). Dually infected persons with undetectable HIV-2 viral loads had significantly higher median levels of CD8 T cells expressing CD38 (P < 0.001) and HLA-DR (P = 0.01) than HIV-2 only infected patients. CONCLUSION: These results suggest that in dual infection, the level of HIV-2 replication depends on the immune status of the patients, with HIV-1 out-replicating HIV-2 as disease progress.     

Primary HIV-1 infection in African children infected through breastfeeding

OBJECTIVES: To describe acute retroviral syndrome and associated primary viraemia in African children infected with HIV-1 through breastfeeding. DESIGN: Matched case-control study performed retrospectively within the ANRS 049 DITRAME project conducted in 1995-1998 in Abidjan, C?te d'Ivoire. METHODS: Cases were children infected by HIV-1 postnatally through breastfeeding. All were HIV-1 negative by DNA PCR at least 45 days of age, but positive on a subsequent sample. This period was considered as surrounding the estimated date of postnatal contamination. Signs/symptoms occurring within this period were recorded in cases and compared with those occurring during the same time period in uninfected breastfed children (controls). For cases, plasma specimens were tested for HIV-1 plasma RNA using the branched DNA assay. RESULTS: Of 22 infants infected postnatally (median age at first positive sample, 185 days; range, 87-373 days), 21 (95.5%) exhibited at least one clinical sign, compared with only 27 of the 44 (61.4%) uninfected children (P = 0.003). Three independent factors were associated with primary HIV-1 infection: mononucleosis-like syndrome [odds ratio (OR), 8.3; 95% confidence interval (CI), 1.4-47.8], dermatitis (OR, 6.0; CI, 1.1-31.9), and generalized lymphadenopathy (OR, 26.5; CI, 2.0-348.4). Among cases, initial median plasma HIV-1 RNA viral load was 5.92 log10 copies/ml; this declined to 4.96 log10 12 months after the first positive viral load. CONCLUSIONS: These results may be useful for the recognition of early paediatric cases of postnatal transmission in Africa and could enable targeting of those who should benefit from HIV RNA or DNA testing for primary HIV-1 infection and their subsequent care.     

Quantitative detection of human immunodeficiency virus type 1 (HIV-1) viral load by real-time RT-PCR assay using self-quenched fluorogenic primers

HIV-1 viral load is a basic marker to evaluate the severity of HIV-1 related diseases and to monitor the effectiveness of treatment. A method based on real-time RT-PCR technology has been developed to quantify HIV-1 RNA using self-quenched fluorogenic primers known as LUX primers. They were used in this study to recognize a low variable gag region of subtype E and B consensus sequences. Specificity was verified by amplicon melting temperatures. An external standard curve was constructed with 10 fold serial dilutions of synthetic HIV-gag RNA. A broad range linear relationship (10 to 10(6) copies/ml) was observed between the number of PCR cycles needed to detect a fluorescent signal and the number of RNA copies. The intra- and inter-assay coefficients of variation were 0.72 to 2.54% and 3.14 to 8.83%, respectively, thus indicating good reproducibility. Thirty out of fifty HIV-infected individual plasma samples were quantified by this method and compared with the AMPLICOR HIV-1 Monitor assay, which is widely considered the reference technique for HIV-RNA viral load measurement. The results indicate that the AMPLICOR HIV-1 Monitor assay and real-time RT-PCR using LUX primers are in good agreement (mean difference in log10 copies/ml+/-2 standard deviations = 0.21+/-1.34).     

Maternal HIV-1 and HIV-2 infection and child survival in The Gambia

OBJECTIVE: To compare the survival of children born to HIV-1 or HIV-2 seropositive mothers with that of children born to HIV-seronegative mothers and to evaluate risk factors for mortality. DESIGN: Physician-blinded prospective study. METHODS: One hundred and one HIV-1-seropositive, 243 HIV-2-seropositive pregnant women, and 468 HIV-seronegative women (control group) matched by age, parity, and health centre, were followed up in a study of mother-to-child transmission of HIV. Mothers and children were seen at 2 and 6 months of age and subsequently followed at 3-monthly intervals up to 18 months of age. HIV infection in children was diagnosed by polymerase chain reaction at 2, 9 or 18 months and by antibody assays at 18 months. RESULTS: Fifteen per cent of children born to HIV-1-infected mothers died compared with 7% of children born to HIV-2-infected mothers [hazard ratio, 2.3; 95% confidence interval (CI), 1.1-4.7; P = 0.02], and 6% of HIV-seronegative mothers (hazard ratio, 2.6; 95% CI, 1.4-5.0; P = 0.003). Six of the 17 children known to be HIV-1 infected died compared with none among the eight HIV-2-infected children (P = 0.13). High proviral load in the babies, high antenatal maternal RNA plasma viral load, and maternal death increased child mortality significantly. CONCLUSIONS: More children born to HIV-1-infected mothers died in comparison with those born to HIV-2-infected mothers or to mothers from the control group. This effect was due to excess death in HIV-1-infected infants which was associated with a high viral load in the affected mother and child.