Risk Factors for Opportunistic Illnesses in Children With Human Immunodeficiency Virus in the Era of Highly Active Antiretroviral Therapy FREE

The introduction of highly active antiretroviral treatment (HAART) has resulted in a declining incidence of opportunistic illnesses (OIs) in individuals with human immunodeficiency virus (HIV).^1- 8 However, OIs among some individuals treated with HAART continue to occur,^3- 4 which may suggest a clinical failure of HAART. Generally, HAART causes a progressive improvement in immune function demonstrated by an increase in CD4⁺ T-lymphocyte (CD4) counts and a decrease in HIV-1 viral load.^9- 10 In children, HAART has been associated with a similar recovery in immunologic status,^11- 14 although viral suppression is usually less evident in children compared with adults.¹⁵ A combination of CD4 percentage or counts and HIV-1 viral load have proven to be the best predictors of future disease progression and mortality in both adults and children.^16- 20

Recently, the long-term effects of HAART on changes in the CD4 percentage were evaluated prospectively in the Pediatric AIDS Clinical Trials Group (PACTG) 219 cohort of children and adolescents with HIV in the United States.² The investigators reported a varied response to therapy according to the level of immunosuppression at the time of treatment initiation; children with the most severe immunosuppression experienced the greatest increase in CD4 percentage over time. However, only about one third of these children achieved a CD4 percentage within the normal range (CD4 percentage ≥25%) after 3 years. Furthermore, 16% of the children who initiated HAART while their CD4 percentages were in the normal range had decreases in their CD4 percentages to less than 25% during follow-up.²

Few studies have examined the occurrence of OIs in relation to HAART use in pediatric HIV infection. Thus, we used data from a large US cohort of children perinatally infected with HIV who initiated HAART at different ages and levels of immunosuppression to examine the relationship between HAART use and other clinical, demographic, and virologic factors and the occurrence of OIs in the pediatric HIV population.

The source population for this study was the PACTG 219C cohort, which began enrollment in September 2000 to assess the long-term effects of in utero, perinatal, and childhood exposure to antiretroviral therapy (ART) in children with HIV across the United States. The PACTG 219C protocol is a revised version of PACTG 219, a protocol initiated in 1993 and described in detail elsewhere.²¹ Children eligible for the current study must have been perinatally infected with HIV and must have received HAART (as defined later) prior to or at enrollment in PACTG 219C. Institutional review board approval was obtained at all of the participating clinical centers, and the children's parents or guardians provided written informed consent. There were 2318 subjects perinatally infected with HIV enrolled in PACTG 219C between September 15, 2000, and August 31, 2003, and 1323 (57.1%) of them were previously enrolled in PACTG 219. Of the 2318 subjects, 12 (0.5%) were never treated with ART, 276 (11.9%) never received HAART, 101 (4.4%) initiated HAART after enrollment in PACTG 219C, and 2 (0.1%) had no follow-up after enrollment and thus were excluded. The remaining 1927 subjects (83.1%) met the eligibility criteria and were subsequently followed up for the occurrence of OIs until August 31, 2003.

At enrollment in PACTG 219C and every 3 months thereafter, participants underwent a physical examination and a medical history was obtained, including information on signs, symptoms, or diagnoses of any OIs (defined according to the Centers for Disease Control and Prevention [CDC] disease classification of pediatric HIV infection).^22- 23 Information on treatment (including prophylactic treatment) or hospitalization for any OIs was also recorded. Measures of HIV-1 viral load and CD4 counts and percentages were also collected at each visit.

Detailed information on lifetime ART use, including start and stop dates of each drug, was obtained at enrollment in PACTG 219C. Use of HAART was defined as combination ART with at least 3 drugs, of which at least 1 was either a protease inhibitor (PI) or a nonnucleoside reverse transcriptase inhibitor.^8- 10 Duration of HAART use was defined as time from HAART initiation to enrollment in PACTG 219C and was categorized into 3 groups: 0 to less than 1 year, 1 to less than 4 years, and 4 or more years. These categories were chosen to represent short-term, medium-term, and long-term HAART use. Subjects were further classified according to age at HAART initiation (ages 0 to <2 years, 2 to <10 years, and ≥10 years). In addition, treatment experience prior to initiating HAART, excluding zidovudine prophylaxis, was considered. Any change from one ART drug or combination of drugs to another was considered a change in regimen. Thus, subjects were categorized as treatment naïve, receiving 1 to 2 ART regimens, or receiving 3 or more different ART regimens prior to HAART. Subjects were also classified according to the number of different HAART regimens they received from HAART initiation to enrollment in PACTG 219C (1 HAART regimen and treatment naïve at HAART initiation, 1-2 different HAART regimens, and ≥3 different HAART regimens).

The main outcome was the first occurrence of an OI during follow-up in PACTG 219C (regardless of OI events before enrollment). All of the OI events were classified according to the CDC's criteria for pediatric HIV infection^22- 23 as category B (OI-B) or category C (OI-C) events based on the type of OI and the subject's history of OIs. If a subject had a first OI-B event and a subsequent OI-C event, both the OI-B and OI-C events were considered in the respective incidence and regression analyses. However, if a subject had a first OI-C event and a subsequent OI-B event, only the OI-C event was considered. For pneumonias, only presumed events (with radiological and clinical evidence of pneumonia) or proven events (that were culture confirmed) were considered.

Incidence rates (per 100 person-years) and exact Poisson 95% confidence intervals (CIs) for the first OI (either OI-B or OI-C) event, the first OI-B event, and the first OI-C event were calculated for the period between September 15, 2000, and August 31, 2003. For these analyses, end of follow-up was defined as the date of the first OI event, date of loss to follow-up, or August 31, 2003, whichever came first. In addition, incidence rates (per 100 person-years) and exact Poisson 95% CIs for specific OIs (both OI-B and OI-C events) were calculated for the follow-up period.

The median CD4 percentage and HIV-1 viral load at enrollment and at the end of follow-up in PACTG 219C were determined for all of the study subjects. Subanalyses were performed among 741 study subjects previously enrolled in PACTG 219 for whom we had information on CD4 percentage at the time of HAART initiation. The HIV-1 RNA measurements were not recorded in the PACTG 219 protocol and thus could not be determined at HAART initiation. In addition, the proportion of subjects in different categories of CD4 percentage and HIV-1 viral load were determined by OI occurrence during follow-up.

The relationship between age at HAART initiation, duration of HAART use, and the risk of a first OI (either OI-B or OI-C) event during cohort follow-up was assessed using Cox proportional hazards regression while adjusting for the following potential confounders: age at enrollment, number of ART regimens prior to initiating HAART, number of HAART regimens before enrollment, adherence to HAART at enrollment, sex, and ethnicity. In a subanalysis including the 741 subjects with information on CD4 percentage at HAART initiation, CD4 percentage and CDC disease category at HAART initiation were added to the regression model to adjust for severity of illness at HAART initiation. The validity of the proportional hazards assumptions was assessed by examining residual plots to determine whether the hazards were approximately constant over time.²⁴ Hazard ratios (HRs), 95% CIs, and P values were calculated using the likelihood ratio test.²⁴ All of the analyses were performed using SAS version 8.2 software (SAS Institute, Inc, Cary, NC).

The characteristics of the 1927 subjects in the study population by OI occurrence during follow-up are summarized in Table 1. The risk of an OI during cohort follow-up was higher in children who were older, were Hispanic or non-Hispanic black, were in CDC category C, had a CD4 percentage of less than 15%, had a higher HIV-1 viral load at enrollment, initiated HAART at an older age, and were treated with at least 3 different ART regimens before HAART initiation or at least 3 different HAART regimens before enrollment in PACTG 219C.

Median follow-up time was 28 months (range, 6 days to 35 months), and 316 subjects (16.4%) were censored prior to August 31, 2003. Reasons for censoring included site closure (n = 227), subjects moving away from the site (n = 41), death unrelated to OI (n = 7), investigators being unable to contact the subjects (n = 15), parents withdrawing consent (n = 14), subjects being unable to adhere to the study protocol (n = 8), subjects being too ill (n = 3), and imprisonment (n = 1).

Subjects were treated with a median of 2 (range, 0-12) different ART regimens prior to HAART initiation followed by a median of 1 (range, 1-17) different HAART regimen before PACTG 219C enrollment. Of the 387 subjects who were treatment naïve when initiating HAART, 250 (64.6%) continued to receive the same HAART regimen through enrollment in PACTG 219C. The first HAART regimen consisted of 2 nucleoside reverse transcriptase inhibitors and 1 PI for 1224 subjects (63.5%), 2 nucleoside reverse transcriptase inhibitors and 1 nonnucleoside reverse transcriptase inhibitor for 279 subjects (14.5%), 1 or 2 nucleoside reverse transcriptase inhibitors, 1 PI, and 1 nonnucleoside reverse transcriptase inhibitor for 395 subjects (20.5%), and other nucleoside reverse transcriptase inhibitor, PI, and nonnucleoside reverse transcriptase inhibitor combinations for 29 subjects (1.5%).

The OI incidence rates per 100 person-years of follow-up in PACTG 219C are shown in Table 2. In total, 185 first OI-B events and 57 first OI-C events occurred among 226 subjects during follow-up. The incidence rate was 6.21 (95% CI, 5.42-7.07) per 100 person-years for the first OI regardless of category, 4.99 (95% CI, 4.30-5.76) per 100 person-years for the first OI-B event, and 1.47 (95% CI, 1.12-1.91) per 100 person-years for the first OI-C event. The most frequently occurring OI-B event was bacterial pneumonia, and the most frequently occurring OI-C event was esophageal candidiasis. In 7 children, the OI-C event (2 Pneumocystis jiroveci pneumonia events, 3 culture-confirmed sepsis events, and 2 HIV encephalopathy events) led to death within a month after diagnosis. Ten additional children died from AIDS-related OIs after having had a prior OI-C event.

During follow-up, 721 (37.4%) of the participants received prophylaxis against P jiroveci pneumonia and 166 (8.6%) received prophylaxis against Mycobacterium avium complex infection. Children receiving P jiroveci pneumonia or M avium complex prophylaxis were significantly older at enrollment and HAART initiation and were more likely to be in CDC category C, have a lower CD4 percentage (<15%), and have a higher HIV-1 viral load (>10 000 copies/mL) at enrollment compared with subjects receiving no prophylaxis during follow-up (all P<.001).

The median CD4 percentage improved over time, from 22% (95% CI, 21%-23%) just prior to HAART initiation (among 741 subjects previously enrolled in PACTG 219) to 30% (95% CI, 29%-30%) at enrollment in PACTG 219C, and it persisted at 30% (95% CI, 29%-30%) until the end of follow-up. This increase in median CD4 percentage over time was observed across different subgroups, including sex, ethnic origin, age at HAART initiation, and age at enrollment. However, there were differences in median CD4 percentage levels within subgroups; girls had a significantly higher median CD4 percentage than boys when initiating HAART (25% vs 20%, respectively), at enrollment (31% vs 28%, respectively), and at the end of follow-up (31% vs 29%, respectively). Furthermore, subjects who initiated HAART when they were younger than 2 years had significantly higher median CD4 percentages compared with older children (aged >10 years) at HAART initiation (27% vs 17%, respectively), at enrollment (35% vs 23%, respectively), and at the end of follow-up (34% vs 23%, respectively). The HIV-1 viral load for the study population did not change substantially between enrollment (median viral load, 1081 copies/mL; 95% CI, 809-1397 copies/mL) and the end of follow-up (median viral load, 818 copies/mL; 95% CI, 574-1140 copies/mL).

The proportions of subjects in different categories of CD4 percentage at HAART initiation, at enrollment in PACTG 219C, and at the end of follow-up are shown by OI occurrence during follow-up in Figure, A. More subjects with OIs than without OIs were in the low CD4 percentage (<15%) category at HAART initiation (49.6% of subjects with OIs vs 23.7% of subjects without OIs), at enrollment (41.2% of subjects with OIs vs 7.7% of subjects without OIs), and at the end of follow-up (41.2% of subjects with OIs vs 8.3% of subjects without OIs). In addition, the proportion of subjects with normal CD4 percentages (≥25%) was more pronounced over time among subjects without an OI (from 48% to 71%) compared with those experiencing OIs (from 23% to 38%). The proportion of subjects in different categories of HIV-1 viral load at enrollment in PACTG 219C and at the end of follow-up did not change much over time (Figure, B). However, more subjects with OIs than without OIs were in the high viral load categories (≥100 000 copies/mL) both at enrollment (26.6% vs 5.7%, respectively) and at the end of follow-up (28.8% vs 6.1%, respectively). Also, the proportion of subjects with good viral suppression (<400 copies/mL) was lower among subjects with OIs compared with those without OIs at enrollment (12.8% vs 28.0%, respectively) and at the end of follow-up (16.4% vs 32.8%, respectively).

The risk of having a first OI (regardless of category) during follow-up in relation to HAART exposure was examined using Cox proportional hazards regression (Table 3). In the multivariate regression model, there was a significantly higher risk of a first OI during follow-up for non-Hispanic black subjects (HR, 2.09; 95% CI, 1.29-3.38) and Hispanic subjects (HR, 1.94; 95% CI, 1.16-3.23) compared with non-Hispanic white subjects. Subjects who enrolled in PACTG 219C when they were aged 10 years or older had a significantly lower risk of a first OI compared with subjects younger than 2 years at enrollment (HR, 0.37; 95% CI, 0.15-0.92). In contrast, subjects who initiated HAART when they were aged 10 years or older had a significantly higher risk of a first OI compared with those initiating HAART when they were younger than 2 years (HR, 2.48; 95% CI, 1.23-5.00). Subjects who switched between 3 or more HAART regimens had a significantly increased risk of a first OI compared with subjects who were treatment naïve at HAART initiation and continued to receive the same HAART regimen after HAART initiation (HR, 2.73; 95% CI, 1.40-5.32). There was no relationship between sex, duration of HAART use, number of ART regimens before initiating HAART, or short-term adherence to HAART and the risk of a first OI (Table 3). Regression analysis performed for OI-B and OI-C events separately gave similar results except for boys having a marginally significant lower risk of a first OI-B event during follow-up compared with girls (HR, 0.75; 95% CI, 0.56-1.00).

To control for severity of illness at HAART initiation, we performed separate Cox proportional hazards regression analyses for the 741 subjects previously enrolled in PACTG 219 for whom we had information on CD4 percentage at the time of initiating HAART. Initially, we ran the same multivariate model as shown in Table 3. The results were materially unchanged except for wider CIs, indicating no significant risk of OI according to race or ethnicity, age at HAART initiation, and age at enrollment. Subsequently, CD4 percentage and CDC disease category at HAART initiation were added to the multivariate model as indicators of severity of illness at the time of initiating combination therapy (Table 4). In this adjusted model, only CDC disease category and CD4 percentage at enrollment were related to OI risk during follow-up. Both CDC category B (HR, 3.21; 95% CI, 1.89-5.45) and category C (HR, 3.35; 95% CI, 1.77-6.35) at HAART initiation were associated with a higher risk of a first OI compared with being asymptomatic or in category A. A CD4 percentage of less than 15% at HAART initiation was associated with a 2-fold increased risk of a first OI compared with a normal CD4 percentage (≥25%) (HR, 2.04; 95% CI, 1.18-3.53).

Given the fact that there are about 2.1 million children living with HIV or AIDS in the world and that combination therapies have become increasingly available and recommended, it is of greatest importance to define the optimal time to initiate HAART. Our findings that HAART initiation at an older age (age >10 years) was associated with an increased risk of OIs may have important clinical and policy implications. Despite the obvious risk of fast progression in young children, the question of when HAART should be initiated in children remains controversial. According to the US guidelines,²⁵ treatment intervention as early as possible regardless of the child's immune function is recommended. In contrast, European²⁶ and World Health Organization²⁷ guidelines recommend postponing treatment in children until the CD4 percentages fall below 15% or 20%.^25- 26 Although some investigators have found that recovery to normal CD4 cell counts is independent of age,²⁸ the recovery of naïve CD4 cells is usually more rapid and effective in young children,^2,14,29 which may favor early intervention with HAART before severe immunosuppression has occurred.¹⁴ Furthermore, from a long-term perspective, only a small fraction of children who start therapy when their CD4 percentages have fallen below 15% will reach normal CD4 percentages (≥25%). Again, this supports the view that HAART should be initiated as early as possible.²

The CD4 cell count at HAART initiation has previously been demonstrated as a strong predictor for subsequent OI risk.⁴ In accordance, the significant association we found between older age (age >10 years) when initiating HAART and the risk of an OI during follow-up (Table 3) diminished when we controlled for the children's immunologic and clinical status (CD4 percentage and CDC disease category, respectively) at HAART initiation (Table 4). Subjects who already were in CDC category B or C at HAART initiation had a higher risk of additional OIs during follow-up compared with subjects who initiated HAART before development of clinical disease, even after adjustment for their CD4 percentages at HAART initiation (Table 4). In addition, more children with OIs than without OIs had a CD4 percentage of less than 15% at HAART initiation (49.6% of children with OIs vs 23.7% of children without OIs), at enrollment (41.2% of children with OIs vs 7.7% of children without OIs), and at the end of follow-up (41.2% of children with OIs vs 8.3% of children without OIs) (Figure, A). These results suggest that HAART initiation should not be delayed until the CD4 percentage has fallen below 15% or the child has already experienced an AIDS-defining illness.

Our findings of an increased risk of OIs during follow-up among non-Hispanic black and Hispanic children (Table 3) compared with non-Hispanic white children have been previously observed yet not fully explained in the literature.³⁰ Genetic variation in immunologic status between ethnic groups has been suggested, especially since social disparities (such as access to treatment or adherence) have not been significantly associated with ethnic groups.^2,31 In accordance, we found lower median CD4 percentages among both non-Hispanic black and Hispanic subjects at HAART initiation, at enrollment, and at the OI event compared with non-Hispanic white subjects, and race or ethnicity was no longer predictive of risk in the model adjusted for the CD4 percentage at HAART initiation (Table 4). The relationship between sex, immunologic parameters, and clinical outcomes is less clear.^32- 33 We found that girls had higher median CD4 percentages at all of the time points during follow-up whereas their viral loads were similar to those of boys. In contrast, the CD4 percentage–adjusted analyses showed that boys had a slightly lower risk of a first OI-B event compared with girls, a finding that has previously been reported for mortality in our pediatric cohort.²¹

This study has a number of potential limitations, including confounding by severity of illness. Previous analyses of this study population indicated that the most severely ill children were the first to start PI-based treatment.²¹ To adjust for severity of illness, we included measures of immunologic (CD4 percentage) and clinical (CDC disease category) status at HAART initiation. These 2 measures, together with sex and age-adjusted weight, have recently been used in a Pediatric AIDS Severity System score and have proven to be highly predictive of mortality in children with HIV (W.M.D., G.R.S., Kate Buchacz, PhD, Geoffrey A. Weinberg, MD, Kenneth McIntosh, MD, unpublished data, December 31, 1996). Indeed, the increased risk of a first OI associated with an older age at initiating HAART disappeared when we adjusted for CD4 percentage and CDC disease category at the time of HAART initiation (Table 4). Unfortunately, HIV-1 RNA measurements at HAART initiation were unavailable for our subjects, which precluded adjustment for this variable in our analyses.

Other potential limitations include our measure of duration of HAART use, which was based on lifetime experience at enrollment and may not have represented HAART use at the actual time of OI occurrence. In addition, we did not take into account any interruptions or discontinuation of HAART after enrollment in PACTG 219C but rather used an intention-to-continue-treatment approach. However, previous data from the PACTG 219 cohort indicate that very few children (<5%) discontinue HAART once it has been initiated.²¹ Furthermore, our risk estimates were based on the first OI to occur in PACTG 219C follow-up and may not have been the first OI while receiving HAART. Despite accrual of our study population from institutions across the United States, our results may not be generalizable to the underlying base population of children with HIV, especially since our study was conducted within a self-selective cohort of children who may have participated in previous ART clinical trials in the United States. Nevertheless, our study included 1927 children, thus representing a relatively large sample of children with HIV receiving care in the United States.

We conclude that OIs are still occurring in the pediatric HIV population after the introduction of HAART, but at lower incidence rates than in the pre-HAART era.³⁴ Low CD4 percentage (<15%) and CDC disease category B or C at the time of HAART initiation remained the strongest predictors of OIs in the pediatric HIV population in the era of HAART. A larger proportion of children with OIs than without OIs had a CD4 percentage of less than 15% at both HAART initiation and OI occurrence. Presumably, it is not the duration of HAART use or the age at HAART initiation but rather the sustained response to HAART that predicts the risk of OIs after HAART initiation. The optimal time to initiate HAART in children can only be, and needs to be, assessed in future randomized studies.

A related article was published in JAMA (Gona P, Van Dyke RB, Williams PL, et al. Incidence of opportunistic and other infections in HIV-infected children in the HAART era. JAMA. 2006;296;292-300).

Corresponding Author: Nathalie Ylitalo, MD, PhD, Department of Epidemiology, Harvard School of Public Health, 655 Huntington Ave, Boston, MA 02155 (nathalie.ylitalo@oncology.gu.se).

Accepted for Publication: February 8, 2006.

Author Contributions:Study concept and design: Ylitalo, Hughes, Dankner, Van Dyke, and Seage. Acquisition of data: Ylitalo, Hughes, Nachman, Dankner, Van Dyke, and Seage. Analysis and interpretation of data: Ylitalo, Brogly, Hughes, Nachman, Dankner, and Seage. Drafting of the manuscript: Ylitalo. Critical revision of the manuscript for important intellectual content: Brogly, Hughes, Nachman, Dankner, Van Dyke, and Seage. Statistical analysis: Ylitalo, Hughes, and Seage. Obtained funding: Hughes and Van Dyke. Administrative, technical, and material support: Ylitalo, Brogly, and Van Dyke. Study supervision: Hughes and Seage. Dr Ylitalo had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Funding/Support: This study was funded by the US National Institute of Allergy and Infectious Diseases, the National Institute of Child Health and Human Development, and the Center for Biostatistics in AIDS Research at the Harvard School of Public Health (the statistical and data analysis center of the Pediatric AIDS Clinical Trials Group) under cooperative agreement 5 U01 AI41110 with the US National Institute of Allergy and Infectious Diseases.

Role of the Sponsors: The US National Institute of Allergy and Infectious Diseases and the National Institute of Child Health and Human Development were involved in the design, data collection, and conduct of protocol 219C but were not involved in the present analysis, the interpretation of the data, the writing of the manuscript, or the decision to submit the manuscript for publication.

Acknowledgment: We acknowledge Shirley Traite and Matthew Manne at the Center for Biostatistics in AIDS Research at the Harvard School of Public Health for their invaluable programming support. We also thank all of the children, adolescents, and families for their participation in Pediatric AIDS Clinical Trials Group 219C.

Contributing Pediatric AIDS Clinical Trials Group 219C Site Personnel and Pediatric AIDS Clinical Trials Group Centers

Contributing Pediatric AIDS Clinical Trials Group 219C Site Personnel

University of Medicine and Dentistry of New Jersey, Newark: P. Palumbo, P. Andrew, A. Dieudonne, B. Dashefsky; Robert Wood Johnson Medical School, Piscataway, NJ: S. Gaur, P. Whitley-Williams, A. Malhotra, L. Cerracchio; Harbor-UCLA Medical Center, Torrance, Calif: M. Keller, J. Hayes, A. Gagajena, C. Mink; Department of Pediatrics, Johns Hopkins University, Baltimore, Md: N. Hutton, B. Griffith, M. Joyner, C. Kiefner; Baylor Texas Children's Hospital, Houston, Tex: F. Minglana, M. E. Paul, W. T. Shearer, C. D. Jackson; Sinai Children's Hospital, Chicago, Ill: D. C. Johnson, D. Kowalski, B. Wolfe, D. Ryan; The Columbia Presbyterian Medical Center and Cornell University New York Presbyterian Hospital, New York: A. Higgins, M. Foca, P. LaRussa, A. Gershon; University of Miami, Coral Gables, Fla: G. B. Scott, C. D. Mitchell, L. Taybo, C. Gamber; Children's Hospital and Research Center, Oakland, Calif: A. Petru, T. Courville, K. Gold, L. Johnson; Phoenix Children's Hospital, Phoenix, Ariz: J. P. Piatt, J. Foti, L. Clarke-Steffen; University of North Carolina at Chapel Hill: T. Belho, B. Pitkin, J. Eddleman; Schneider Children's Hospital, New Hyde Park, NY: V. R. Bonagura, S. J. Schuval, C. Colter; Harlem Hospital, New York: E. J. Abrams, M. Frere, D. Calo, S. Champion; Children's Hospital at Downstate, Brooklyn, NY: E. Handelsman, H. J. Moallem, D. M. Swindell, J. M. Kaye; Jacobi Medical Center, New York: M. Chin, K. Dorio, A. Wiznia, M. Donovan; San Juan Hospital, San Juan, Puerto Rico: M. Acevedo, M. Gonzalez, L. Fabregas, M. E. Texidor; Yale University School of Medicine, New Haven, Conn: W. A. Andiman, S. Romano, L. Hurst, J. de Jesus; SUNY Upstate Medical University, Syracuse, NY: L. B. Weiner, K. A. Contello, W. A. Holz, M. J. Famiglietti; SUNY Stony Brook, Stony Brook, NY: S. Nachman, D. Nikolic-Djokic, D. Ferraro, J. Perillo; Howard University, Washington, DC: S. Rana, H. Finke-Castro, P. H. Yu, J. C. Roa; University of Florida Health Science Center, Jacksonville: M. H. Rathore, A. Khayat, K. Champion, S. Cusic; St Jude Children's Research Hospital, Mitalicis, Tenn: P. M. Flynn, K. Knapp, N. Patel; Vanderbilt University Medical Center, Nashville, Tenn: G. Wilson; Department of Pediatrics, Washington University School of Medicine, St Louis Children's Hospital, St Louis, Mo: K. A. McGann, L. Pickering, G. A. Storch; The Children's Hospital of Philadelphia, Philadelphia, Pa: S. D. Douglas, G. Koutsoubis, R. M. Rutstein, C. A. Vincent; Charity Hospital of New Orleans and Earl K. Long Early Intervention Clinic, New Orleans, La: M. Silio, T. Alchediak, C. Boe, M. Cowie; Baystate Medical Center Children's Hospital, Springfield, Mass: B. W. Stechenberg, D. J. Fisher, A. M. Johnston, M. Toye; Medical College of Georgia, Augusta: C. S. Mani, S. Foshee, B. Kiean, S. Cobb; University of Maryland Medical Center, Baltimore: J. Farley, K. Klipner; University of California, San Diego: S. Spector, R. Viani, L. Stangl.

Pediatric AIDS Clinical Trials Group Centers

Cooper Hospital–University Medical Center, Camden, NJ; Children's Hospital Boston, Boston, Mass; Boston Medical Center, Boston; University of California, Los Angeles, Medical Center, Los Angeles; Children's Hospital of Los Angeles, Los Angeles; Long Beach Memorial, Long Beach, Calif; Chicago Children's Memorial Hospital, Chicago; Cook County Hospital, Chicago; The University of Chicago Children's Hospital, Chicago; Women's and Children's HIV Program, Moffitt Hospital, University of California, San Francisco; Mother, Child, and Adolescent HIV Program, University of California, San Diego; Phoenix Children's Hospital; Duke University, Durham, NC; New York University School of Medicine Bellevue Hospital, New York; Children's National Medical Center, Washington; Children's Hospital and Regional Medical Center, Washington, DC; University of South Florida, Tampa; Oregon Health and Science University, Portland; Children's Hospital of the King's Daughters, Norfolk, Va; Lincoln Medical and Mental Health Center, New York; Mount Sinai Medical Center, Miami Beach, Fla; University of Illinois, Chicago; Children's Hospital of Michigan, Detroit; Children's Medical Center of Dallas, Dallas, Tex; Los Angeles County–University of Southern California Medical Center, Los Angeles; Children's Hospital, University of Colorado, Denver; North Broward Hospital District, Fort Lauderdale, Fla; University of Florida, Gainesville; University of Rochester Medical Center, Rochester, NY; University of Mississippi Medical Center, Jackson; Medical College of Virginia, Virginia Commonwealth University, Richmond, Va; University Children's Hospital AIDS Program, University of Puerto Rico, San Juan; St Christopher's Hospital for Children, Philadelphia; Bronx Lebanon Hospital Center, New York; St Luke’s-Roosevelt Hospital Center, New York; Montefiore Medical, Albert Einstein College of Medicine, New York; Metropolitan Hospital Center, New York; University of Massachusetts Medical School, Worcester; Connecticut Children's Medical Center, Hartford; University of Alabama at Birmingham; University of South Alabama, Mobile; The Medical Center, Columbus, Ga; Incarnation Children's Center, New York; St Joseph's Hospital and Medical Center, Paterson, NJ; Wyler Hospital, Chicago; Children's Hospital of Oakland, Oakland; Emory University Hospital, Atlanta, Ga; Ruiz Arnau University Hospital, Bayamon, Puerto Rico; Medical University of South Carolina, Charleston; Children's Hospital at Albany Medical Center, Albany, NY; Columbus Children's Hospital, Columbus, Ohio; Public Health Unit, Palm Beach County, West Palm Beach, Fla.