PEDIATRICS Vol. 104 No. 5 November 1999, p. e56

ELECTRONIC ARTICLE:
Natural History of Human Immunodefiency Virus Type 1 Infection in Children: A Five-Year Prospective Study in Rwanda

Rosemary Spira, MD^*, Philippe Lepage, MD, PhD, Philippe Msellati, MD, PhD^§, Philippe Van de Perre, MD, PhD, Valeriane Leroy, MD, PhD^*, Arlette Simonon, PhD, Etienne Karita, MD, François Dabis, MD, PhD^*, and for the Mother-to-Child HIV-1 Transmission Study Group

From the ^* Unité INSERM U 330, Université Victor Segalen Bordeaux 2, Bordeaux, France; the  Department of Pediatrics, Centre Hospitalier de Kigali, Kigali, Rwanda; the ^§ Programme SIDA ORSTOM, Abidjan, Côte d'Ivoire; and the   AIDS Reference Laboratory, National AIDS Control Program, Kigali, Rwanda.

	ABSTRACT

Top Abstract Methods Results Discussion Conclusion References

Objective.  To compare morbidity and mortality of human immunodeficiency virus type 1 (HIV-1)-infected and HIV-1-uninfected children and to identify predictors of acquired immunodeficiency syndrome (AIDS) and death among HIV-1-infected children in the context of a developing country.

Design.  Prospective cohort study.

Setting.  Maternal and child health clinic of the Centre Hospitalier de Kigali, Rwanda.

Participants.  Two hundred eighteen children born to HIV-1-seropositive mothers and 218 born to seronegative mothers of the same age and parity were enrolled at birth.

Outcome Measures.  Deaths, clinical AIDS, nonspecific HIV-related manifestations, and use of health care services.

Results.  Fifty-four infected and 347 uninfected children were followed up for a median of 27 and 51 months, respectively. With the exception of chronic cough, the risk of occurrence of nonspecific HIV-related conditions was 3 to 13 times higher in infected than in uninfected children. The recurrence rate and severity of these findings were increased systematically in infected infants. Estimated cumulative risk of developing AIDS was 28% and 35% at 2 and 5 years of age, respectively. Estimated risk of death among infected children at 2 and 5 years of age was 45% and 62%, respectively, a rate 21 times higher than in uninfected children. Median survival time after estimated infection was 12.4 months. Early infection, early onset of HIV-related conditions, failure to thrive, and generalized lymphadenopathy were associated with subsequent risk of death and/or AIDS, whereas lymphoid interstitial pneumonitis was predictive of a milder disease.

Conclusions.  In Africa, HIV-1-infected children develop disease manifestations early in life. Specific clinical findings are predictive of HIV-1 disease, AIDS stage, and death. Bimodal expression of HIV-1 pediatric disease is encountered in Africa, as in industrialized countries, but prognosis is poorer. human immunodeficiency virus infection, children, vertical transmission, natural history, Africa.

According to the World Health Organization (WHO), 1.1 million of children were living with human immunodeficiency virus (HIV) infection worldwide at the end of 1997.¹ Of these, the great majority live in sub-Saharan Africa and were infected by their mother during pregnancy, delivery, or breastfeeding. Data on the natural history of human immunodeficiency virus type 1 (HIV-1) infection in children followed prospectively from birth and prognostic studies are scarce in developing countries. Available reports show a high variability in mortality and morbidity indicators according to the study settings and a probable bimodal evolution of pediatric HIV disease.^2-5 A better understanding of the clinical course and of the predictors of disease progression is needed to improve diagnosis and treatment of HIV-1 perinatally infected children in developing countries, an area vastly unexplored.⁵

A prospective cohort study on the perinatal transmission of HIV and the natural history of HIV-1 infection was conducted in Kigali, the capital of Rwanda, from 1988 to 1994.^6,7 This report describes the clinical and immunologic findings in the first 5 years of life of the 436 children enrolled in this cohort.

	METHODS

Top Abstract Methods Results Discussion Conclusion References

Study Site, Population, and Follow-up

Details about enrollment and follow-up procedures in the cohort have been described elsewhere.⁶ Briefly, 218 infants born to 215 HIV-seropositive mothers were enrolled at birth between November 1988 and June 1989 at the maternity ward of the Centre Hospitalier de Kigali. A comparison group of 218 children born to HIV-seronegative mothers of the same age and parity were enrolled simultaneously.

The children and their mothers were followed every 2 weeks during the first 2 years of life and every 4 weeks thereafter. Thus, during 5 years, they were visited regularly at home by a social worker or they attended the outpatient clinic organized within the Mother and Child Health Clinic of the Centre Hospitalier de Kigali. In addition, the children were systematically examined by a pediatrician every 3 months and assessed for HIV-related signs and symptoms. The following conditions were regarded as HIV-related: chronic diarrhea (14 days); chronic cough (14 days); chronic fever (14 days); severe or recurrent pneumonia; hepatomegaly; splenomegaly; generalized dermatitis; lymphoid interstitial pneumonitis; oral candidiasis (beyond the neonatal period); chronic parotitis; persistent generalized lymphadenopathy (lymph nodes measuring 0.5 cm and present in two or more extra inguinal sites); and failure to thrive (<80% of weight for age). When necessary, children were seen in the outpatient department or hospitalized in the pediatrics department and treated free of charge. All of the children were immunized during the first 15 months of life, following the recommendations of the WHO Expanded Program of Immunization with one exception: a high dose Edmonston-Zagreb measles vaccine was given at 6 months of age in place of the Schwarz vaccine given at 9 months of age.⁸ The physicians and social workers were blinded to the HIV infection status of the children up to 18 months of age.

Laboratory Methods

At delivery and at 3-month intervals thereafter, the child and mother's sera were tested for HIV-1 antibodies by a commercial enzyme-linked immunosorbent assay, further confirmed by a commercial Western Blot technique if the test was positive. As maternal seroconversion may have occurred during follow-up, children born to HIV-negative mothers may have become exposed to vertically acquired infection through breastfeeding and become subsequently infected. Polymerase chain reaction (PCR) testing for diagnosis of HIV-1 infection was performed on all available samples, collected routinely from birth using a home-made combination of primers of the gag and pol regions of the genome. Detailed description and information on performances of this PCR technique, which reached sensitivity and specificity levels of 100% and 98%, respectively, have been described elsewhere.^9,10 Blood for total lymphocyte and CD4/CD8 count by a manual indirect immunofluorescent method was obtained 15 days after delivery in mothers and at 6 months of age in children.

Definition of Pediatric HIV Infection and Classification of the Children

Children were considered HIV-1-infected (group 1) if they had at least one positive PCR result during follow-up, regardless of clinical and serologic criteria. Children who had a positive PCR test at birth were considered infected from their first day of life, whereas for other children, the infection date was estimated as the median date of the interval between the last negative and the first positive PCR test result. All infected children had a seropositive mother, diagnosed either at delivery or after seroconversion during follow-up.

Children were considered HIV-uninfected if they had no positive PCR test result and either 1) they were negative for HIV antibody at 15 months of age and did not fulfill the modified WHO clinical case definition of pediatric acquired immunodeficiency syndrome (AIDS), in which severe and/or recurrent pulmonary infection is considered a major sign in place of chronic cough as a minor sign^11,12 during the follow-up, or 2) they died before 15 months of age, they were HIV-seronegative, and their death was not considered HIV-related (death was defined as HIV-related either in a child fulfilling the modified WHO clinical definition of pediatric AIDS or in a child with at least one HIV-related condition when last seen and dying from severe infection or persistent diarrhea beyond the first 4 weeks of life). Uninfected children were either born to seropositive (group 2) or seronegative mothers (group 3). The HIV infection status of children was considered indeterminate (group 4) otherwise.

Statistical Analysis

HIV infection was treated as a time-dependent variable so that infected children, estimated to have been contaminated after birth, were considered uninfected until the last quarterly examination preceding their infection and were considered infected thereafter. Morbidity and mortality outcomes were described using the Kaplan-Meier method. The two groups of uninfected children (groups 2 and 3) were first compared. When appropriate, all uninfected children were compared with the infected ones (group 1). Mortality, nonspecific HIV-related morbidity, and use of health care services (outpatient clinic, hospital ward, and medical injections) were compared by univariate analyses, using the Log-rank test, and by multivariate analyses, using the Cox proportional hazards model with HIV infection considered to be time varying. Entry time was the date of birth for all children. Multivariate analyses were performed using descending stepwise procedures after inclusion of all variables associated to the variable of interest in univariate analysis with a significance level <20%. We defined the recurrence rate of each type of HIV-related condition and medical service provided as the proportion of 3-month periods or quarters with the event present in each group, normalized to rates per 100 child-quarters. Risk ratios then were computed by dividing recurrence rates of the different groups compared. A prognostic analysis restricted to HIV-infected children was performed by Cox proportional hazard models to identify predictive factors of death and of the development of AIDS.

	RESULTS

Top Abstract Methods Results Discussion Conclusion References

Of the 436 eligible children, 54 were diagnosed as infected (group 1); 45 of these children were born to HIV-infected mothers, and the remaining 9 children were born to women who were uninfected at delivery but who became HIV-seropositive during the follow-up, and thus became themselves infected too.¹² A total of 347 children were considered uninfected: 138 were born to HIV-infected women (group 2) and 209 to HIV-uninfected women (group 3). The remaining 35 children, born to seropositive mothers, had an indeterminate infection status and were excluded from the following analysis.

Among the 54 infected children, 39 were estimated to have been contaminated in utero, intrapartum, or during their first 3 months of life and thus were considered as infected throughout their follow-up. The remaining 15 children, who were thought to have been contaminated later in life, were considered infected only from the 3-month period of their estimated contamination: second (2), fourth (9), fifth (2), sixth (1), and seventh quarters (1), respectively. The median estimated age at infection among the 54 infected children was 1.6 months.

Overall, the median follow-up time was 27.4 months in group 1 (range: 4.5-58.2 months), 51.5 months in group 2 (range: 8.9-58.1 months), and 51.4 months in group 3 (range: 0.03-58.6 months). At the end of the study period, 6 children had been lost to follow-up from group 1 (11.1%), 17 from group 2 (12.3%), and 25 from group 3 (12.0%).

Male to female sex ratio was 1 overall, without difference among the three groups combined. All but 4 children were breastfed for a median duration of 18 months (range: 1.2-38.9 months). None of the infected children received antiretroviral therapy or opportunistic infection prophylaxis.

Among the 215 HIV-infected women, only 1 presented clinical AIDS at inclusion. The median CD4 cell count at 2 weeks after delivery was 757/mm³ (range: 179-2992).

Comparison of the Two Groups of Uninfected Children

Comparisons between uninfected children born to seropositive women (group 2) and uninfected children born to seronegative women (group 3) showed that these two groups did not differ with respect to mortality during the 5 years of follow-up (Cox hazard ratio [HR] of death for children of group 2 compared with children of group 3: 0.4, 95% confidence interval [CI]: 0.1-1.6; P = .20). Comparison of morbidity between these two groups showed that the risk of onset during follow-up of most of the HIV-related conditions studied was strictly comparable between children of group 2 and children of group 3 (data not shown). The risk of occurrence of three clinical manifestations tended to increase among uninfected children born to seropositive mothers, compared with unexposed children: severe pneumonia (HR for children of group 2, compared with children of group 3: 1.4; 95% CI: 0.9-2.0; P = .12), generalized dermatitis (HR: 1.7; 95% CI: 0.9-2.5; P = .11), and chronic parotitis (HR: 3.3; 95% CI: 0.7-9.9; P = .12). Children of group 2 also tended to have an increased risk of hospitalization during follow-up (HR: 1.3; 95% CI: 0.9-1.4; P = .10). Considering that these differences were not statistically significant, the remainder of the analysis considered uninfected children of groups 2 and 3 as a whole and consisted in their comparison to infected children of group 1.

Mortality

Twenty-eight (52%) infected children and 13 (4%) uninfected children died during follow-up. Figure 1 shows the Kaplan-Meier estimate of survival curves of the children, according to their infection status. The probability of death in infected children was 0.26 (95% CI: 0.16-0.41) at first birthday, 0.45 (95% CI: 0.32-0.60) at 2 years of age, and 0.62 (95% CI: 0.47-0.78) at 5 years of age. The estimated cumulative probability of death in uninfected children was 0.04 (95% CI: 0.02-0.07) at 5 years of age. The univariate Cox proportional hazards model estimated that the overall risk of death was 20.7 times higher in infected children than in uninfected children (95% CI: 10.7-40.0). Median age at death was 13.4 months for infected (n = 28) and 7.7 months for uninfected children (n = 13). The median time of survival was 12.4 months after estimated infection and 6.1 months after occurrence of the first HIV-related condition in the 28 infected children who died. Risk of death among children infected in the first 3 months of life was 5.1 times higher than among those who acquired infection later (95% CI: 1.5-17.0). The most common causes of death among infected children were pulmonary infections (9 cases) and diarrhea (8 cases).

View larger version (8K): [in this window] [in a new window]   Fig. 1.   Time to death, according to Kaplan-Meier estimate in 54 HIV-infected and 347 uninfected children. The Mother-to-Child HIV-1 Transmission Study cohort, Kigali, Rwanda, 1988-1994.

AIDS Diagnosis

Of the 54 HIV-infected children, 14 (26%) developed AIDS, according to the modified WHO clinical definition. The cumulated probabilities of the development of AIDS at 1, 2, and 5 years of age were 0.17 (95% CI: 0.09-0.32), 0.28 (95% CI: 0.17-0.45), and 0.35 (95% CI: 0.22-0.53), respectively. The median duration of HIV infection without AIDS was 13.9 months from the estimated date of contamination. Among the 28 infected children who died, 9 met the case definition of AIDS. The median time of survival after the occurrence of AIDS was 9 months (range: 4-21 months).

HIV-related Signs and Symptoms

All 54 infected children presented at least one nonspecific HIV-related condition during follow-up. From the estimated date of contamination, median time of infection without any HIV-related condition was 5.3 months (range: 0-13.3 months). Estimated age at onset of the first HIV-related condition was 8.9 months. The initial clinical signs that occurred most frequently were failure to thrive (51.9%) and persistent lymphadenopathy (44.4%). These conditions occurred either alone or in combination with other clinical manifestations: 34 children (63%) first presented with one sign or symptom, 11 (20%) with two, and 9 (17%) with three to six conditions. Morbidity was not different between the 39 children infected perinatally and the 15 children infected after 3 months of age (data not shown).

The occurrence of nonspecific HIV-related signs and symptoms in the first 5 years of life is described for infected and uninfected children in Table 1. Study time was first divided into two periods (before and after 2 years of life), and comparisons between these two periods within each group of children showed that risks of onset of the different manifestations were approximately constant over time (data not shown). The most frequent signs among infected children were chronic cough, persistent generalized lymphadenopathy, and failure to thrive, which were estimated to occur at least once in 97%, 93%, and 92% of the infected children, respectively. All of the HIV symptomatic conditions studied also occurred among uninfected children, although the risk of onset among uninfected children was significantly lower (3-13 times) than among infected children, except for chronic cough, which occurred in almost all of the children, regardless of their HIV infection status. The clinical features for which recurrence had the greatest ability to discriminate infected from uninfected children were oral candidiasis and chronic parotitis: they were found almost 20 times more frequently in infected than in uninfected children. All of the other conditions, including chronic cough, occurred significantly more often in infected than in uninfected children but were less specific of HIV infection. When present, chronic parotitis, generalized dermatitis, splenomegaly, and persistent lymphadenopathy tended to occur earlier in infected than in uninfected children.

Contacts With the Health Care System

Table 2 shows the estimated proportions of infected and uninfected children using health care services at least once during follow-up and the average frequency of attendance in both groups. The risks of being admitted to a hospital ward and of receiving an injection at least once were 2.8 and 2.5 times higher, respectively, in infected than in uninfected children. However, the proportion of children seen at least once in an outpatient clinic did not differ between the two groups. Infected children had significantly more recurrent use of each of the three types of services (1.5-3.5 times) than did uninfected ones.

Factors Associated With Survival and AIDS Occurrence

An early HIV infection (before 3 months of age) and a short interval (<6 months) between HIV infection and occurrence of the first HIV-related conditions were associated with a poor prognosis (relative hazards of death: 10.3 and 5.4, respectively; Table 3; Fig 2A and 2B). Failure to thrive as the initial pattern of clinical HIV disease tended to be associated with an increased risk of death (P = .07). The occurrence of chronic diarrhea, splenomegaly, or chronic parotitis at least once during follow-up was associated with a decrease in the subsequent risk of death (HR: 0.2, 0.3, and 0.1, respectively). The risk of death was not significantly higher in children who had developed AIDS in comparison to the other children. No specific combination of clinical manifestations was associated with differences in survival (data not shown). Biologically, neither the maternal CD4 cell count at day 15 nor the child's CD4/CD8 ratio at 6 months of age was predictive of death (Table 3).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Kaplan-Meier curves of time to death and AIDS, according to selected covariates in 54 HIV-1-infected children. The Mother-to-Child HIV-1 Transmission Study cohort, Kigali, Rwanda, 1988-1994.

Clinical AIDS occurred 15 times more frequently in children whose initial pattern of HIV disease included persistent generalized lymphadenopathy (95% CI: 3.1-73.9) and 4.7 times more often in those who presented failure to thrive as the first HIV-related symptom (95% CI: 1.4-16.1; Table 4; Fig 2C). Occurrence of lymphoid interstitial pneumonitis during the course of the disease was associated with a decreased risk of developing AIDS (HR: 0.04; 95% CI: 0.01-0.4) (Table 4; Fig 2D). Age at infection, interval between infection and onset of the first HIV-related conditions, and maternal and children biological characteristics were not associated with the subsequent development of clinical AIDS (Table 4).

	DISCUSSION

Top Abstract Methods Results Discussion Conclusion References

This study is the first report describing the long-term natural history of HIV-1 infection from the estimated date of contamination and the prognostic factors in African children followed up prospectively from birth to 5 years of age, compared with uninfected children. The cumulative rate of children lost to follow-up in our cohort was <12% in 5 years and seems to be particularly low in this African context, strengthening our results. Most published data concerning pediatric HIV infection in developing countries have been either prospective studies that focused primarily on mother-to-child HIV transmission and thus reported limited follow-up of children born to HIV-infected women^13-21 or nonprospective studies focusing on HIV-infected children.^22-26 Thus, despite the large body of information that has been accumulated on mother-to-child HIV transmission, only a few prospective studies, focusing specifically on natural history in HIV-infected children, have been published in developing countries.^27,28 However, such comparative and prospective studies represent the only approach to identify and evaluate the frequency of early and nonspecific clinical or laboratory findings in infected children. This knowledge of early symptomatology may be useful for diagnosis and case management in areas where laboratory facilities are limited, as is often the case in developing countries.⁵

Although clinical patterns of pediatric HIV infection may differ according to the timing of mother-to-child transmission, all infected children of the cohort were included in our analysis regardless of their estimated date of infection (antenatal, neonatal, or postnatal). Thus, the whole spectrum of vertically acquired pediatric HIV cases, which corresponds with the different patterns of disease routinely encountered in breastfeeding African populations, is considered in this report. In our cohort, morbidity was not different between children infected early in life and other children, but mortality was higher in children in the former group, and analyses were adjusted systematically on timing of transmission to take this variable into account.

The choice of a positive PCR test result as criterion for HIV infection in infants of this cohort was motivated by the ability to estimate the timing of infection by the date of their first positive PCR test result. This may have had an influence on our results: a negative PCR test result at birth may have been compatible with an infection during the neonatal period if viral transmission occurred in the few hours preceding (intrapartum transmission) or following (early postnatal transmission) the test. Therefore, some false-negative results may have occurred leading to the misclassification, at least transient, of infected children in the uninfected group. Thus, differences between infected and uninfected children may have been underestimated in our analysis.

Our findings first suggest that uninfected children born to seropositive mothers and those born to seronegative mothers do not differ significantly in terms of mortality, morbidity, and use of health care services, although the former children tend to have an increased risk both to develop some defined conditions (ie, severe pneumonia, generalized dermatitis, and chronic parotitis) and to be admitted to the hospital ward. These trends may reflect a consequence of HIV exposure in uninfected children born to seropositive mothers. Considering that our data did not formally prove this potential effect of perinatal HIV exposure, we pooled the two groups of uninfected children in all analyses and compared them as a whole to infected children. We had shown previously, in the same cohort, that HIV-1-uninfected children had similar neonatal anthropometric measurements,⁶ the same growth patterns,²⁹ and comparable neurodevelopmental performances³⁰ whether they were born to infected or uninfected mothers.

All infected children presented at least one nonspecific HIV-related condition in the first 13 months of HIV-1 infection. Although nonspecific findings occurred in both infected and uninfected infants, they were much more common and recurrent among the infected infants. The risk of initial occurrence of chronic cough was not higher in infected than in uninfected children, probably because of its lack of specificity, but chronic cough occurred more recurrently in infected infants. Moreover, these nonspecific features seemed to be more severe in HIV-1-infected children, assessed by the rates of hospital admission and of medical injections that were increased dramatically, compared with uninfected children throughout follow-up. The probability of using outpatient clinic services at least once was very high (nearly 100% for all children), and there was no difference between infected and uninfected subjects (most likely because all of these children were treated free of charge and because even clinic attendance for minor complaints was included). This very high rate of use of medical care services reflects the high quality of follow-up within the cohort.

In HIV-1-infected children, the most frequently observed clinical signs were chronic cough, failure to thrive, and generalized lymphadenopathy, which also were reported among the most frequent HIV-related conditions in other developing countries.^15,16,24,25 Chronic cough and failure to thrive were present in almost half of the initial patterns of symptomatic disease in our series. However, these three conditions were also common in uninfected children. In our cohort, the most specific findings of HIV-1 infection were oral candidiasis and chronic parotitis, a pattern similar to the one observed in the New York City cohort.³¹ But in the New York cohort, failure to thrive and splenomegaly also were reported to be specific symptoms of HIV-1 infection, a difference most likely explained by the higher occurrence of these two symptoms among uninfected children in developing countries.

This study of the progression to AIDS and death among HIV-infected children confirms the probable bimodal clinical expression of the disease. A subgroup of short-term survivors developed severe conditions and died within the first 2 years of life, whereas the rates of mortality and progression to AIDS were lower between 2 and 5 years, suggesting a slower disease progression in children who survived at least 2 years. This bimodal pattern of HIV-1-related disease, which has been described previously in American³² and European³³ HIV-1-infected children, most likely reflects the time when HIV transmission occurs, the infectious dose transmitted by the mother to her offspring, and/or the virulence of the HIV-1 strain.² In this report, the association observed between an early estimated age at infection and a shorter survival supports the first of these hypotheses.

The rates of the occurrence of death in HIV-1-infected children (26% at first birthday, 45% at 2 years of age, and 62% at 5 years of age) are close to the rates observed in Uganda and in Zaïre in the first 2 years of life^27,28 but are dramatically higher than the rates reported in Brazil and Western countries, where death is estimated to occur in 9% to 16% of the infected children at 1 year of age and 25% to 36% at 5 years of age.^5,3134-36 The increased mortality in African populations may be explained, at least partly, by the difference in the use of antiretroviral and prophylactic treatments, which have been widespread in industrialized countries and are nonexistent in developing countries. But, because these treatments usually have been prescribed after the child's first birthday, early increased mortality in African infants is unlikely to reflect the effect of these therapies. The more rapid disease progression in the first year of life also may be caused by a more advanced stage of maternal HIV disease in Africa. However, infected mothers of our cohort did not present clinical or biological signs of advanced disease. Thus, the increased mortality in African HIV-infected children could rather be the witness of a more severe pattern of disease in developing countries, perhaps in relation to multiple and early infectious exposures. The observed rates of progression to AIDS (17% at 1 year of age and 35% at 5 years of age) are close to the rates reported in other cohorts in developing⁵ and industrialized countries,^31,34 but the heterogeneity among the studies in the clinical AIDS definitions used (WHO, modified WHO, and Centers for Disease Control and Prevention definitions³⁷) makes comparisons difficult to interpret.⁵

Several studies in developing countries²⁸ and many in industrialized countries^31-33,38,39 have found factors associated with prognosis in HIV-1-infected children. In fact, early onset of HIV-related conditions, failure to thrive, and generalized lymphadenopathy were associated with a poor prognosis, whereas lymphoid interstitial pneumonitis was associated with a mild disease. Our data support the significant prognostic value of these conditions, regarding death and/or AIDS occurrence. The failure to thrive and generalized lymphadenopathy had prognostic value, if they occurred as first manifestations of HIV disease, and the occurrence of lymphoid interstitial pneumonitis, at any time during follow-up, was associated with a good prognosis. Chronic diarrhea, splenomegaly, and chronic parotitis also seemed to be associated with a longer survival in this report. These three signs were most likely the witnesses of a chronic HIV disease rather than real prognostic determinants. The absence of the prognostic value of clinical AIDS on survival reflects the inadequacy between the arbitrary definition of AIDS and the clinical disease observed in this African population of HIV-1-infected children, in which no other combination of clinical outcomes was found to have prognostic value. This was probably attributable to the limited sample size.

	CONCLUSIONS

Top Abstract Methods Results Discussion Conclusion References

In summary, we found that African HIV-1-infected infants develop disease manifestations early in life and that specific clinical findings are predictive of HIV-1 infection, AIDS, and death. Even if the bimodal evolution that is observed in industrialized countries is confirmed by these data, HIV disease seems to be more severe in developing countries, despite a higher frequency of late acquisition of infection that is related to breastfeeding. This suggests that in addition to the timing of infection other mechanisms may be involved in disease progression. Other cohort data may be useful to confirm these findings and to document morbidity and mortality more thoroughly, especially in the era of antiretroviral therapy. Evaluation of different strategies of prophylaxis of opportunistic infections and case management of the most frequent pediatric HIV-related conditions, adapted to the context of developing countries, are needed to offer adapted and large scale care for HIV-infected children in these countries.

	ACKNOWLEDGMENTS

This study was funded in part by the AIDS Task Force of the European Commission (EC/DGVIII), the Agence Nationale de Recherches sur le SIDA (France) and the Belgo-Rwandan Medical Cooperation.

We thank Lieve Fransen (EC/DGVIII) and the Program National de Lutte contre le SIDA (Ministry of Health of Rwanda) for having encouraged this study. We also thank the nurses of the Department of Pediatrics and the social workers of the Mother-to-Child HIV-1 Transmission Study for their active collaboration. We also thank Marie-Louise Newell for advice and criticism in the preparation of this manuscript.

This article is dedicated to the memory of the mothers, children, and personnel of the project and of the members of the Department of Pediatrics who were murdered in 1994 during the genocide in Rwanda.

	FOOTNOTES

The Mother-to-Child HIV-1 Transmission Study group includes the following members: Pediatrics, Drs Bazubagira, Hitimana, Lepage, Nsengumuremyi, and Goethem; Microbiology, Drs Bogaerts, Karita, Simonon, and Van de Perre; and Epidemiology, Drs Dabis, Leroy, Msellati, Salamon, and Spira.

The current affiliation of Dr Philippe Lepage is the Centres Hospitaliers Inter-Universitaires Ambroise Paré and Tivoli, Mons La Louvière, Belgium.

The current affiliation of Dr Philippe Van de Perre is the Centre Muraz, Bobo Dioulasso, Burkina Faso.

Deceased.

Received for publication Jan 11, 1999; accepted May 3, 1999.

Reprint requests to (R.S.) Unité INSERM 330, Université Victor Segalen Bordeaux 2, Bordeaux, France. E-mail: rosemary.spira{at}dim.u-bordeaux2.fr

	ABBREVIATIONS

WHO, World Health Organization; HIV, human immunodeficiency virus; HIV-1, human immunodeficiency virus type 1; AIDS, acquired immunodeficiency syndrome; PCR, polymerase chain reaction; HR, hazard ratio; CI, confidence interval.

	REFERENCES

Top Abstract Methods Results Discussion Conclusion References

1. The Joint United Nations Programme on HIV/AID/World Health Organization. Report on the Global HIV/AIDS EpidemicJune 1998. Geneva, Switzerland: The Joint United Nations Programme on HIV/AID; 1998:75
2. Lepage P, Hitimana DG. Natural history and clinical expression of HIV-1 infection in children. AIDS. 1991;5:113-121. Supplement
3. Ryder RW, Temmerman M. The effect of HIV-1 infection during pregnancy and the perinatal period on maternal and child health in Africa. AIDS . 1991;5:75-85. Supplement
4. Coulter J. HIV infection in African children. Ann Trop Paediatr. 1993, 13:205-215
5. Lepage P, Spira R, Kalibala S, Care of human immunodeficiency virus-infected children in developing countries: report of a workshop for clinical research. Pediatr Infect Dis J 1998; 17:581-586 [CrossRef][Medline]
6. Lepage P, Dabis F, Hitimana DG, Perinatal transmission of HIV-1: lack of impact of maternal HIV infection on characteristics of livebirths and on neonatal mortality in Kigali, Rwanda. AIDS 1991; 5:295-300 [Medline]
7. Lepage P, Van de Perre P, Msellati P, Mother-to-child transmission of HIV-1 and its determinants: a cohort study in Kigali, Rwanda. Am J Epidemiol 1993; 137:589-599 [Abstract/Free Full Text]
8. Lepage P, Dabis F, Msellati P, Safety and immunogenicity of high-dose Edmonston-Zagreb measles vaccine in children with HIV-1 infection: a cohort study in Kigali, Rwanda. Am J Dis Child 1992; 146:550-555 [Abstract]
9. Simonon A, Lepage P, Karita E, 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 Hum Retrovirol 1994; 7:952-957
10. Van de Perre P, Simonon A, Msellati P, Postnatal transmission of human immunodeficiency virus type 1 from mother to infanta prospective cohort study in Kigali, Rwanda. N Engl J Med 1991; 325:593-598 [Abstract]
11. World Health Organization Acquired immunodeficiency syndrome (AIDS). Wkly Epidemiol Rec 1986; 61:69-73
12. World Health Organization. Report on the Meeting of the Technical Working Group on HIV/AIDS in Childhood. Geneva, Switzerland: World Health Organization; 1989. WHO/GPA/SFI/89.2AF
13. Ryder RW, Nsa W, Hassig SE, Perinatal transmission of the human immunodeficiency virus type 1 to infants of seropositive women in Zaire. N Engl J Med 1989; 320:1637-1642 [Abstract]
14. Ryder RW, Nsuami M, Nsa W, Mortality in HIV-1-seropositive women, their spouses and their newly born children during 36 months of follow-up in Kinshasa, Zaire. AIDS 1994; 8:667-672 [Medline]
15. Hira SK, Kamanga J, Bhat GJ, Perinatal transmission of HIV-1 in Zambia. Br Med J 1989; 299:1250-1252
16. Lallemant M, Le Coeur S, Samba L, Mother-to-child transmission of HIV-1 in Congo, central Africa. AIDS 1994; 8:1451-1456 [Medline]
17. Halsey NA, Boulos R, Holt E, Transmission of HIV-1 infections from mothers to infants in Haitiimpact on childhood mortality and malnutrition. JAMA 1990; 264:2088-2092 [Abstract]
18. Taha TET, Dallabetta GA, Canner JK, The effect of human immunodeficiency virus infection on birthweight, and infant and child mortality in urban Malawi. Int J Epidemiol 1995; 24:1022-1029 [Abstract/Free Full Text]
19. Datta P, Embree JE, Kreiss JK, Mother-to-child transmission of human immunodeficiency virus type 1: report from the Nairobi study. J Infect Dis 1994; 170:1134-1140 [Medline]
20. Adjorlolo-Johnson G, De Cock KM, Ekpini E, Prospective comparison of mother-to-child transmission of HIV-1 and HIV-2 in Abidjan, Ivory Coast. JAMA 1994; 272:462-466 [Abstract]
21. Zijenah L, Mbizvo MT, Kasule J, Mortality in the first 2 years among infants born to human immunodeficiency virus-infected women in Harare, Zimbabwe. J Infect Dis 1998; 178:109-113 [Medline]
22. Vetter KM, Djomand G, Zadi F, Clinical spectrum of human immunodeficiency virus disease in children in a West African city. Pediatr Infect Dis J 1996; 15:438-442 [CrossRef][Medline]
23. Lucas SB, Peacock CS, Hounnou A, Disease in children infected with HIV in Abidjan, Côte d'Ivoire. Br Med J 1996; 312:335-338 [Abstract/Free Full Text]
24. Gayle HD, Gnaore E, Adjorlolo G, HIV-1 and HIV-2 infection in children in Abidjan, Côte d'Ivoire. J Acquir Immune Defic Syndr Hum Retrovirol 1992; 5:513-517
25. Lepage P, Van de Perre P, Van Vliet G, Clinical and endocrinologic manifestations in perinatally human immunodeficiency virus type 1-infected children aged 5 years or older. Am J Dis Child 1991; 145:1248-1251 [Abstract]
26. De Cock KM, Zadi F, Adjorlolo G, Retrospective study of maternal HIV-1 and HIV-2 infections and child survival in Abidjan, Côte d'Ivoire. Br Med J 1994; 308:441-443 [Abstract/Free Full Text]
27. Thea DM, St Louis ME, Atido U, et al. A prospective study of diarrhea and HIV-1 infection among 429 Zairian infants. N Engl J Med. 1993,329:1696-1702
28. Berhane R, Bagenda D, Marum L, et al. Growth failure as a prognostic indicator of mortality in pediatric HIV infection. Pediatrics. 1997;100(1). URL: http://www.pediatrics.org/cgi/content/full/100/1/e7
29. Lepage P, Msellati P, Hitimana DG, Growth of human immunodeficiency type 1-infected and uninfected children: a prospective cohort study in Kigali, Rwanda, 1988 to 1993. Pediatr Infect Dis J 1996; 15:479-485 [CrossRef][Medline]
30. Msellati P, Lepage P, Hitimana DG, Van Goethem C, Van de Perre P, Dabis F. Neurodevelopmental testing of children born to human immunodeficiency virus type 1 seropositive and seronegative mothers: a prospective cohort study in Kigali, Rwanda. Pediatrics. 1993,92:843-848
31. Bamji M, Thea DM, Weedon J, Prospective study of human immunodeficiency virus 1-related disease among 512 infants born to infected women in New York City. Pediatr Infect Dis J 1996; 15:891-898 [CrossRef][Medline]
32. Scott GB, Hutto C, Makuch RW, Survival in children with perinatally acquired human immunodeficiency virus type 1 infection. N Engl J Med 1989; 321:1791-1796 [Abstract]
33. Blanche S, Tardieu M, Duliege AM, Longitudinal study of 94 symptomatic infants with perinatally acquired human immunodeficiency virus infection -Evidence for a bimodal expression of clinical and biological symptoms. Am J Dis Child 1990; 144:1210-1215 [Abstract]
34. The European Collaborative Study Natural history of vertically acquired human immunodeficiency virus-1 infection. Pediatrics 1994; 94:815-819 [Abstract/Free Full Text]
35. Barnhart HX, Caldwell MB, Thomas P, Natural history of human immunodeficiency virus disease in perinatally infected children: an analysis from the pediatric spectrum of disease project. Pediatrics 1996; 97:710-716 [Abstract/Free Full Text]
36. Tovo PA, De Martino M, Gabiano C, Prognostic factors and survival in children with perinatal HIV-1 infection. Lancet 1992; 339:1249-1253 [CrossRef][Medline]
37. Centers for Disease Control and Prevention Classification system for human immunodeficiency virus (HIV) infection in children under 13 years of age. Mor Mortal Wkly Rep CDC Surveill Summ 1987; 36:225-236
38. Pizzo PA, Wilfert CM, and the Pediatric AIDS Siena Workshop II Markers and determinants of disease progression in children with HIV infection. J Acquir Immune Defic Syndr Hum Retrovirol 1995; 8:30-44 [Medline]
39. Diaz C, Hanson C, Cooper ER, Disease progression in a cohort of infants with vertically acquired HIV infection observed from birth: the women and infants transmission study (WITS). J Acquir Immune Defic Syndr Hum Retrovirol 1998; 18:221-228 [Medline]