Abstract
OBJECTIVE. The goal was to describe the frequency, characteristics, and correlates of infectious disease morbidity during the first 6 months of life among HIV-1–exposed but uninfected infants.
METHODS. The study population consisted of infants enrolled in the National Institute of Child Health and Human Development International Site Development Initiative Perinatal Study who were HIV-1 uninfected and had follow-up data through the 6-month study visit. Definitive and presumed infections were recorded at study visits (birth, 6–12 weeks, and 6 months).
RESULTS. Of 462 HIV-1–uninfected infants with 11644 child-weeks of observation, 283 experienced ≥1 infection. These 283 infants experienced 522 infections (1.8 infections per infant). The overall incidence rate of infections was 4.5 cases per 100 child-weeks of observation. Overall, the most common infections were skin or mucous membrane infections (1.9 cases per 100 child-weeks) and respiratory tract infections (1.7 cases per 100 child-weeks). Thirty-six percent of infants had >1 respiratory tract infection (1.8 cases per 100 child-weeks). Incidence rates of upper and lower respiratory tract infections were similar (0.89 cases per 100 child-weeks and 0.9 cases per 100 child-weeks, respectively). Cutaneous and/or oral candidiasis occurred in 48 neonates (10.3%) and 92 older infants (19.3%). Early neonatal sepsis was diagnosed in 12 infants (26.0 cases per 1000 infants). Overall, 81 of 462 (17.5%) infants were hospitalized with an infection. Infants with lower respiratory tract infections were hospitalized frequently (40.7%). The occurrence of ≥1 neonatal infection was associated with more-advanced maternal HIV-1 disease, tobacco use during pregnancy, infant anemia, and crowding. Lower maternal CD4+ cell counts, receipt of intrapartum antibiotic treatment, and country of residence were associated with postneonatal infections.
CONCLUSIONS. Close monitoring of HIV-1–exposed infants, especially those who are anemic at birth or whose mothers have more-advanced HIV-1 disease or who smoked during pregnancy, remains important.
Infectious diseases account for nearly 90% of early childhood deaths in developing countries. A substantial proportion of this morbidity and death occurs among young infants.1 Protection against viral and bacterial infections during early infancy (when innate and specific host defenses are maturing) is conferred primarily through breastfeeding2 and placental transfer of specific antibodies.3
Because current prophylaxis regimens have been successful in decreasing mother-to-child transmission of HIV-1, uninfected infants born to HIV-1–infected mothers represent a growing population.4 In developing countries, such infants are vulnerable to acquiring infectious diseases if they are not themselves HIV-1 infected. Poor sanitation, limited maternal education, avoidance of breastfeeding, deficient transplacental transfer of IgG antibodies,5 and household crowding or close contact with immunodeficient individuals colonized with diverse pathogens6 may play a role. However, little information is available regarding patterns of infectious disease morbidity among HIV-1–exposed infants.7–10 To our knowledge, such information for infants from Latin America and the Caribbean region is not available.
The National Institute of Child Health and Human Development (NICHD) International Site Development Initiative (NISDI) Perinatal Study, a prospective cohort study of HIV-1–infected mothers and their infants in Latin America and the Caribbean region,11 represents a unique opportunity for analyzing the pattern of infectious morbidity among infants in these countries. The objective of this analysis was to describe the frequency, characteristics, and correlates of infectious disease morbidity occurring during the first 6 months of life among HIV-1–exposed infants.
METHODS
NISDI Perinatal Study Protocol
The primary objectives of the NISDI Perinatal Study include characterizing adverse events during pregnancy, the postpartum period, and early infancy. As requirements for enrollment of HIV-1–infected women into this protocol, antiretroviral prophylaxis and alternatives to breast milk needed to be available. Enrollment began in September 2002, and is ongoing. All subjects provide signed informed consent before enrollment into the study. The protocol was approved by the ethical review board at each clinical site where subjects were enrolled, as well as by institutional review boards at the sponsoring institution (NICHD) and the data coordinating center (Westat).
HIV-1–infected women are enrolled in the study during pregnancy and are monitored until 6 months after delivery. A medical history is obtained and a physical examination is conducted at each visit. Laboratory studies are conducted at most visits. The infants are examined at 3 study visits, namely, before hospital discharge after birth and at 6 to 12 weeks and 6 months of age. A medical history is obtained, a physical examination is performed (with assessments of growth and morbidity), and laboratory studies (flow cytometry studies and HIV-1 diagnostic assays, hematologic assays, and biochemical assays) are performed at each study visit. As part of the interval history, information is collected regarding the infant’s symptoms, diagnoses, medications, and hospitalizations.
When infants became ill, mothers were advised to seek medical attention at the clinical site; if medical attention was sought, then clinical, radiographic, and/or laboratory documentation would be available at the site. Therefore, in addition to data collected at NISDI Perinatal Study visits, clinical, radiographic, and/or laboratory data collected as part of routine medical care at the clinical site (and included in the subject’s medical charts) could be obtained. However, for mild events such as upper respiratory tract infections (eg, rhinitis or conjunctivitis) or mild diarrhea for which the mother did not seek medical attention, there would be no clinical, radiographic, or laboratory data available. Infants were considered HIV-1 uninfected if (1) they had ≥2 negative HIV-1 virologic assays (eg, HIV-1 culture or HIV-1 DNA polymerase chain reaction assays), with 1 test performed at ≥1 month of age and 1 performed at ≥4 months of age, and no positive virologic tests; (2) they had 1 positive HIV-1 virologic assay and 2 later HIV-1 virologic tests were negative (≥1 of which was performed at ≥4 months of age); or (3) they had 2 negative HIV-1 antibody tests (≥1 of which was performed at ≥6 months of age).
Study Population and Definitions
The study population for this analysis included infants enrolled in the NISDI Perinatal Study as of October 1, 2004, who completed follow-up monitoring in the study until 6 months of age and who were known to be HIV-1 uninfected. An infection was defined as having occurred when a definitive or presumptive diagnosis of infection was recorded. Specific standardized guidelines regarding the recording of diagnoses were followed by the sites. Definitive diagnosis required, in the majority of cases, that an organism be identified or serologic and/or antigenic evidence be found, unless the clinical findings revealed the causative agent (eg, varicella zoster infection). Otherwise, the diagnosis was classified as presumed. An infection was considered incident if the infant had been free of any symptoms consistent with infection for the previous week.
Infections were categorized as congenital, systemic, central nervous system, upper respiratory tract, lower respiratory tract, gastrointestinal, renal and urinary system, or skin and mucous membranes. Incident infections were tallied at the date of initial diagnosis. Specific diagnoses included in each category, as well as the protocol definitions for each, are listed in the Appendix⇓. Collection of microbiologic specimens was at the discretion of the subject’s physician and was not mandated by the study protocol. Because of the possible overlap in the clinical diagnoses of bronchiolitis and bronchitis in this age group, the 2 categories were combined for analysis as bronchiolitis.
Infant gestational age at birth (in completed weeks) was determined through either obstetrical estimation or pediatric newborn examination.12 Centers for Disease Control and Prevention growth standard reference curves were used for categorization of each subject’s weight for age at birth and postnatally.13 Infant hemoglobin values and absolute neutrophil counts and maternal CD4+ cell counts were classified as normal or abnormal (grade ≥1) according to the Division of AIDS grading system.14 Infant absolute lymphocyte counts were classified as normal or abnormal by using ranges of normal laboratory values appropriate for age.15 Maternal clinical disease staging was performed with the use of the 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults.16 If antiretroviral drugs were received during pregnancy, then the reason for receipt of antiretroviral drugs was categorized as prophylaxis or treatment (prophylaxis: the patient was not receiving antiretroviral drugs when she became pregnant but administration of ≥1 antiretroviral drug was initiated during pregnancy and discontinued at or before the 6-12-week postpartum visit; treatment: the patient was receiving antiretroviral drugs before pregnancy and/or continued antiretroviral drug therapy after the 6-12-week postpartum visit).
Statistical Analyses
Frequencies of infections were calculated according to the age at the initial diagnosis, that is, early neonatal (0–6 days), late neonatal (7–27 days), or postneonatal (≥28 days). Incidence rates were calculated by using person-weeks as the denominator. Proportions were calculated and 95% confidence intervals (CIs) were calculated by using the exact binomial method. Associations of categorical variables with neonatal and postneonatal infections were evaluated by using the Fisher-Freeman-Halton exact test.17 Variables at least marginally associated with outcome (P ≤ .20) were considered candidates for the multivariate logistic regression modeling. Both stepwise selection and backward elimination strategies were applied, to determine whether the 2 selection procedures arrived at the same parsimonious model (using a 5% significance level).
RESULTS
Derivation of the Study Population
As of October 2004, 700 infants had been born to HIV-1–infected women enrolled in the NISDI Perinatal Study. Of those 700 infants, 487 had completed follow-up monitoring through 6 months of age, 204 were still undergoing follow-up monitoring in the study, 6 had died, and 3 had been lost to follow-up monitoring. Of the 487 infants who had completed the protocol, 462 were HIV-1 uninfected, 4 were HIV-1 infected, and 21 were of indeterminate HIV-1 infection status. Of the 6 infants who died during the first 6 months of life, all were of indeterminate HIV-1 infection status. Three of those 6 deaths occurred during the neonatal period (2 attributable to necrotizing enterocolitis and 1 attributable to perinatal asphyxia), and 3 deaths occurred after the first 1 month of life (attributable to sepsis, pneumonia, and bronchoaspiration of the gastric contents). Therefore, the study population included 462 infants who were known to be HIV-1 uninfected and who completed the 6-month study visit.
Characteristics of the Study Population and Occurrence of Infections
Characteristics of the study population are shown in Table 1. None of the infants was breastfed. All except 1 infant received zidovudine prophylaxis (median duration: 6.0 weeks; 5th percentile: 6.0 weeks; 95th percentile: 8.0 weeks). The remaining infant received nevirapine prophylaxis. Of 462 infants with 11644 child-weeks of observation, 283 infants (61%) had ≥1 infection; 121 (43%) of 283 infants had ≥1 infection with onset during the neonatal period and 218 (77%) of 283 children had ≥1 infection with onset during the postneonatal period. The 283 infants experienced 526 infections (mean: 1.9 events per infant). Few infants had >2 infections (1.3% during the early neonatal period, 4.7% during the late neonatal period, and 14.5% during the postneonatal period). The overall incidence rate of infections was 4.5 infections per 100 child-weeks of observation (95% CI: 4.1–4.7 infections per 100 child-weeks of observation).
Characteristics of the Study Population (N = 462) and Risk of Neonatal and Postneonatal Infections
Types of Infections
Overall, the most common infections were skin or mucous membrane infections and lower respiratory tract infections (Table 2). Infections of the respiratory tract (upper and lower) occurred among 150 infants (32.5%) (data not shown). The most infrequent types of infections were systemic infections, congenital infections, and infections of the central nervous system, gastrointestinal tract, and renal/urinary tract (Table 2). Early neonatal sepsis was diagnosed for 12 infants (26.0 cases per 1000 infants; 95% CI: 14.1–46.2 cases per 1000 infants).
Numbers and Types of Infections, Age at Onset of Infection, and Incidence Rate of Infections for Infants With ≥1 Infection
Hospitalizations
Table 3 summarizes hospitalizations and delayed hospital discharge (>3 days after birth) attributable to infection. Overall, 81 (17.5%) of 462 infants were hospitalized at least once with an infection. The proportions of infants with hospitalization or delayed hospital discharge attributable to an infection were as follows: early neonatal, 25 infants (5.4%); late neonatal, 14 infants (3.0%); postneonatal, 44 infants (9.5%). None of the infants with upper respiratory tract infections were hospitalized. In contrast, infants with lower respiratory tract infections were hospitalized frequently (40.7%). Overall, of the 108 episodes of lower respiratory tract infections, 20 (18.5%) were attributable to pneumonia (12 hospitalized; 60.0%) and 88 (81.5%) were attributable to bronchiolitis (33 hospitalized; 37.5%). In contrast, there were only 4 hospitalizations (11.8%) among 34 episodes of acute gastroenteritis.
Number of Hospitalizations or Delayed (>3 Days) Hospital Discharges After Birth Attributable to Infections, According to Diagnostic Category and Age
Analyses of Infections in the Neonatal Period (Onset at <28 Days of Age)
Variables that were associated in univariate analyses (P ≤ .2) with an increased risk of infections beginning in the neonatal period and were included initially in the multivariate modeling were number of persons living in the household, maternal tobacco use during pregnancy, maternal alcohol use during pregnancy, reason for receipt of antiretroviral drugs during pregnancy, maternal HIV-1 clinical classification at hospital discharge after delivery, maternal receipt of intrapartum antibiotic treatment, infant gender, infant gestational age, infant birth weight, infant weight-for-age percentile at birth, and infant hemoglobin level at hospital discharge (Table 1). Estimated odds ratios (ORs) and CIs, both unadjusted and adjusted, for the variables included in the final model for neonatal infections are shown in Table 4. After adjustment, 4 variables remained in the model, namely, maternal HIV-1 clinical classification, maternal tobacco use during pregnancy, infant anemia at birth, and number of persons living in the household (crowding) at birth.
Unadjusted and Adjusted ORs and 95% CIs for Neonatal Infections
Analyses of Infections in the Postneonatal Period (Onset at ≥28 Days of Age)
Variables that were associated in univariate analyses (P ≤ .2) with an increased risk of infections beginning in the postneonatal period and were included initially in the multivariate modeling were maternal country of residence, maternal education, maternal employment, reason for receipt of antiretroviral drugs during pregnancy, maternal HIV-1 classification at hospital discharge after delivery, maternal CD4+ cell count at hospital discharge after delivery, mode of delivery, maternal receipt of intrapartum antibiotic treatment, infant gestational age, infant weight-for-age percentile at birth, and infant hemoglobin level at hospital discharge (Table 1). Estimated ORs and CIs, both unadjusted and adjusted, for the variables included in the final model for postneonatal infections are shown in Table 5. After adjustment, 3 variables remained in the model, namely, lower maternal CD4+ cell count at hospital discharge, maternal receipt of intrapartum antibiotic treatment, and maternal country of residence.
Unadjusted and Adjusted ORs and 95% CIs for Postneonatal Infections
Infectious Disease Diagnoses Included in Each Category and Protocol Definitions for Each Diagnosis
DISCUSSION
In this large cohort of HIV-1–exposed but uninfected infants in Latin America and the Caribbean region, ∼60% of infants experienced infectious disease morbidity during the first 6 months of life. Lower respiratory tract and systemic infections frequently were severe, because they were the cause of hospitalization for many infants. More-advanced maternal HIV-1 disease, maternal tobacco use during pregnancy, infant anemia at birth, and number of persons living in the household were associated with neonatal infections, whereas lower maternal CD4+ cell counts soon after delivery, maternal receipt of intrapartum antibiotics, and country of residence were associated with postneonatal infections.
Although there have been studies of the natural history of pediatric HIV-1 infections and associated morbidity,18–20 only a few prospective studies in North America,8,9 Europe,10 and Africa7 have assessed morbidity among HIV-1–exposed but uninfected infants, compared with HIV-1–infected infants. However, none of those studies focused specifically on infectious disease morbidity, which potentially can be preventable.
We found a much higher incidence of neonatal infections (119 of 468 infants [26%] experienced ≥1 infection) than did a large North American study, in which 4.2% of infants <30 days of age had ≥1 infection.21 The higher incidence of presumed early neonatal sepsis (26.0 cases per 1000 infants), compared with that found for the North American population (1.0–8.1 cases per 1000 infants),22 is also concerning.
Our observed rate of presumed oral candidiasis (which represented approximately two thirds of the skin and mucous membranes infections) was comparable to that observed among HIV-1–uninfected African children <18 months of age.23 In addition, although the incidence of diarrhea and gastroenteritis (7.1%) among HIV-1–uninfected infants in our study, who did not have the benefits of breastfeeding,24,25 was similar to the estimates for Brazilian infants born to HIV-1–uninfected mothers,26 it was much lower than that (35%) found for infants <6 months of age in Brazilian urban slums27 and that (30%) for HIV-1–uninfected infants from South Africa.28 All HIV-1–exposed infants enrolled in the NISDI Perinatal Study had replacement feeding ensured and special attention regarding hygienic measures for formula preparation, which might have affected positively the incidence and severity of diarrhea but not oral candidiasis among these infants.
Our findings of any respiratory infection for 150 of the studied infants (32.5%) was higher than that in a Brazilian community-based study (20%).26 Although there is the possibility of recall bias with respect to upper respiratory tract infections, this is much less likely for lower respiratory tract infections. Recent data from community based-studies regarding lower respiratory tract infection rates in normal, HIV-1–unexposed infants are not available. Existing estimates based on large sample sizes for infants <6 months of age in the general population vary widely, from 0.08 to 0.57 cases per 100 child-weeks at risk in the United States29,30 to 426.4 to 1591.2 cases per 100 child-weeks at risk in developing countries.31 Our findings (0.7–1.1 cases per 100 child-weeks) are closer to those of developed countries. Ideally, a comparison group of infants born to HIV-1–uninfected mothers from a similar sociodemographic background would be included in studies to evaluate whether HIV-1–uninfected infants born to HIV-1–infected mothers are more susceptible to lower respiratory tract infections than are infants without HIV-1–infected mothers. Other authors evaluating HIV-1–uninfected infants have not included a control group of normal infants.8,9,23,28 HIV-1–uninfected infants in our study had a rate of lower respiratory tract infections similar to that of South African, HIV-1–infected infants <18 months of age (15.3%).28 As in a North American study,8 approximately two thirds of the lower respiratory tract infections observed in our study were diagnosed as acute bronchiolitis and the rest as pneumonia. Considering the young age, it is likely that most of these infections were caused by viruses, mainly respiratory syncytial virus (as confirmed for 1 infant), one of the most common and troublesome viruses of infancy associated with acute bronchiolitis.32
Infants with lower respiratory tract infections had a very high (41%) hospitalization rate. In a Brazilian cohort of 5304 infants, only 2% were hospitalized because of acute bronchiolitis.33 In general, high hospitalization rates have been reported only for infants with underlying conditions, such as prematurity, congenital heart disease, or bronchopulmonary dysplasia.34,35 The highest hospitalization rate for infants with bronchopulmonary dysplasia in a developed country was 39%.36 It is possible that the very high hospitalization rate observed in our study of term infants without underlying conditions could be explained by the hospital admission practices for HIV-1–uninfected infants at the research study sites. However, because higher hospitalization rates may correspond to more-severe disease, with the possibility of long-term sequelae,36 it is important to evaluate further whether HIV-1–exposed but uninfected infants are at high risk for bronchiolitis, as well as to determine what factors, besides lack of breastfeeding,32 are related to infection and disease severity.
In terms of risk factors for infections among these HIV-1–exposed but uninfected infants, the association of more-advanced maternal HIV-1 clinical disease stage with neonatal infections leads to the hypothesis that high rates of maternal genital colonization with pathogens or subclinical chorioamnionitis among HIV-1–infected mothers because of immunologic deterioration may play a role in early-onset sepsis. In addition, HIV-1–infected women with advanced HIV-1 disease may have lower antibody titers for common pathogens and/or a lower transfer of these protective antibodies across the placenta.5 Similarly, the association of low maternal CD4+ cells and postneonatal infections may reflect lower protective titers of passively acquired antibodies and exposure to mothers highly colonized with pathogens because of immunodeficiency.6 Indeed, a higher risk of morbidity during the first few months of life among African HIV-1–exposed but uninfected infants born to HIV-1–infected mothers with advanced disease has been described.37
Infant anemia has been associated with infectious gastrointestinal and respiratory morbidity in young infants from a developing country, even after controlling for environmental and socioeconomic factors.38 The identification of anemia at birth as a risk factor for neonatal infections reinforces the need for close monitoring of hemoglobin values among infants of HIV-1–infected mothers, especially those whose mothers received antiretroviral drugs during pregnancy.
It is well known that maternal smoking during pregnancy affects the fetus in a number of ways that may result in chronic hypoxia, metabolic changes, and low birth weight.39 Aside from adverse effects on pulmonary function in the neonatal period,39 there have been no reports of maternal smoking during pregnancy and neonatal infections (as has been shown for alcohol use during pregnancy).40
It is well recognized that postneonatal infections are associated with the socioeconomic status of the mother (and this may vary substantially from country to country). We do not know the mechanism for the association between maternal receipt of intrapartum antibiotic treatment and postneonatal infections
The strengths of this analysis include the use of prospectively collected data beginning during pregnancy. To avoid underestimation of the incidence of infections, both presumed and confirmed infections were analyzed. We decreased the likelihood of misclassification bias by avoiding classification of any early-onset respiratory condition without systemic signs or typical radiologic findings as newborn pneumonia, although this approach could have underestimated the rates of pneumonia and/or early-onset sepsis.
A limitation of our study, along with most or all other studies in this area, is that definitive comparisons with infants born to HIV-1–uninfected mothers from the same population could not be made because we did not have a control group (which ideally would be composed of nonbreastfed infants). Selection bias is possible, because of enrollment of mothers who were willing to enroll in a research study and were willing to be monitored, with their infants, for several months. Finally, in our cohort, most systemic neonatal infections were clinical sepsis. Taking into account the fact that clinical sepsis can be also caused by noninfectious perturbations of homeostasis, the rate of bloodstream infections found in this study could be an overestimate.
CONCLUSIONS
We have provided the first characterization of infectious disease morbidity in a large cohort of HIV-1–exposed infants in Latin America and the Caribbean region. Of particular concern is the frequent incidence of early neonatal sepsis and mucocutaneous candidiasis and hospitalizations for acute bronchiolitis. Close monitoring of HIV-1–exposed infants, especially those who are anemic at birth or whose mothers have more-advanced HIV-1 disease or who smoke during pregnancy, remains important. Knowledge of such risk factors for infectious disease morbidity during the neonatal and postneonatal periods may facilitate development of appropriate clinical interventions to decrease the frequency and severity of infectious diseases among these infants.
Acknowledgments
This work was supported by NICHD contract N01-HD-3-3345.
The NISDI Perinatal Study Group principal investigators, study coordinators, coordinating center representatives, and NICHD staff members include the following: Marcelo H. Losso, Adriana S. Durán, Silvina Ivalo, Alejandro Hakim (Hospital General de Agudos José María Ramos Mejía, Buenos Aires, Argentina); Pedro Cahn, Maria Rolon (Hospital Juan Fernandez, Buenos Aires, Argentina); Edgardo Szyld, Eduardo Warley (Hospital Diego Paroissien, Buenos Aires, Argentina); Mariana Ceriotto, Susana Luciano, Maria Laura Collins (Hospital de Agudos Dra Cecilia Grierson, Buenos Aires, Argentina); Perry Gomez, Percival McNeil, Marva Jervis, Chanelle Diggiss, Rosamae Bain (Princess Margaret Hospital, Nassau, Bahamas); Jorge Pinto, Victor Melo, Fabiana Kakehasi (Universidade Federal de Minas Gerais, Belo Horizonte, Brazil); Ricardo de Souza, Jose Mauro Madi (Universidade de Caxias do Sul/Hospital Geral de Caxias do Sul, Caxias do Sul, Brazil); Ricardo de Souza, Rosangela Boff, Ruti Pipi (Universidade de Caxias do Sul/Ambulatório Municipal de doencas sexualmente transmissiveis/AIDS, Caxias do Sul, Brazil); Ricardo de Souza, Breno Riegel Santos, Rita Lira (Universidade de Caxias do Sul/Hospital Conceicao, Porto Alegre, Brazil); Ricardo de Souza, Rosana da Fonseca, Mario Peixoto, Rita Lira (Universidade de Caxias do Sul/Hospital Femina, Porto Alegre, Brazil); Marisa M. Mussi-Pinhata, Geraldo Duarte, Alessandra C. Marcolin (Hospital das Clinicas da Faculade de Medicina de Ribeirao Preto da Universidade de Sao Paulo, Ribeirão Preto, Brazil); Marcos Machado D’Ippolito, Esau Custodio Joao, Jacqueline Menezes, Guilherme Amaral Calvet (Hospital dos Servidores do Estado, Rio de Janeiro, Brazil); Regina Celia de Menezes Succi, Prescilla Chow Lindsey (Federal University of Sao Paulo, Sao Paulo, Brazil); Javier Ortiz Ibarra, Ricardo Figueroa-Damian, Guadalupe Noemi Plazola-Camacho (Instituto Nacional de Perinatología, Mexico City, Mexico); Data Management and Statistical Center: René Gonin, James Korelitz, Susan Truitt, Roslyn Hennessey, Yolanda Bertucci, Laura Freimanis, D. Robert Harris, Julianne Byrne (Westat); NICHD: Lynne Mofenson, Jack Moye, Jennifer S. Read, Leslie Serchuck, Heather Watts.
Footnotes
- Accepted September 25, 2006.
- Address correspondence to Marisa M. Mussi-Pinhata, MD, Departamento de Puericultura e Pediatria da FMRP-USP, Avenida Bandeirantes 3900, 14049-900 Ribeirão Preto, SP, Brazil. E-mail: mmmpinha{at}fmrp.usp.br
The authors have indicated they have no financial relationships relevant to this article to disclose.
This work was presented in part at the 12th Conference on Retroviruses and Opportunistic Infections; February 22–25, 2005; Boston, MA.
REFERENCES
- Copyright © 2007 by the American Academy of Pediatrics