Human immunodeficiency virus type 1 (HIV-1) disease in the pediatric population progresses more rapidly than in adults and has a bimodal distribution of disease presentation.^1,2 The group of children who develop early and severe disease, frequently with the development of acquired immunodeficiency syndrome (AIDS)-defining illnesses and with a higher likelihood of death in the first years of life, have been identified as rapid progressors.² There is, however, a group of children who have survived for many years after the diagnosis of HIV infection. In this group, two populations have emerged: the long-term nonprogressive survivors (LTNS), who have remained asymptomatic or only mildly symptomatic over a period of years, and those who have survived despite clinical and laboratory evidence of disease progression, the long-term progressive survivors (LTPS). The availability of antiretroviral therapy and early medical intervention with prophylaxis and treatment of infections has altered to some degree the natural history of HIV infection and increased survival time of infected patients.³ Long-term survivors provide a unique opportunity to investigate the immunologic, virologic, and genetic characteristics of HIV-infected children to determine what factors may slow progression to disease and death.

In October 1993, the Pediatric AIDS Foundation sponsored a workshop on HIV-infected pediatric long-term survivors with the participation of investigators from seven research institutions. The workshop focused on a discussion of virological, immunological, and genetic factors that potentially could contribute to long-term survival. As a result of this workshop, investigators representing five large university affiliated medical centers nationwide volunteered to cooperate in a collaborative study of their patient populations. We summarize the results of an initial survey conducted at these sites of children with perinatal HIV infection or with transfusion-associated HIV-acquired infection describing their demographics, clinical findings, immunological status, and antiretroviral therapy.

METHODS

Design

Study Population

Laboratory Evaluation

Statistical Analysis

RESULTS

The racial distribution among perinatally HIV-infected children and those children with transfusion associated infection differed as seen in Table 1. The majority of perinatally infected children were predominantly black, followed by relatively equal proportions of white and Hispanic children. The proportion of black patients predominated in all absolute CD4 count groups. Children infected by transfusion had a relatively uniform racial distribution overall although there were small differences in the proportion of children within each CD4 strata. Asians were a minority in both groups.

Table 1. Racial Distribution of Long-term Survivors by Mode of Infection Within Each CD4 Strata [View Table]

As seen in Table 2, of 197 children enrolled in the survey, 54 (27%) had absolute CD4 counts 500 cells/mm³ (group 1), 54 (27%) had between 200 and 500 cells/mm³ (group 2), and 89 (35%) had CD4 counts <200 cells/mm³ (group 3). Ninety-five children (48%) had had AIDS-defining conditions (whether one or multiple episodes throughout the course of their disease) according to Centers for Disease Control and Prevention guidelines. Of those, 72 were perinatally infected and 23 were infected via blood products. There were no differences in the proportion of children with AIDS (50% in perinatal group vs 43% in transfusion group, P > .05), between mean absolute CD4 counts (379 cells/mm³ vs 269 cells/mm³, P > .05) and in the proportion of children within each CD4 strata (30% vs 20% in group 1, 28% vs 26% in group 2, and 42% vs 54% in group 3, P > .05) by mode of acquisition of infection. There was, however, a difference in age distribution between the two groups as seen in Fig 1. Perinatally infected children had a skewed distribution toward the younger years although children infected by blood transfusion had an age distribution clustering around 11 years (P < .001). CD4 counts tended to decrease with increasing age (500 cells/mm³: mean age 9.8 years, 200 to 500 cells/mm³: mean age 10.1 years, 200 cells/mm³: mean age 10.4 years, P = .03). Most children in the younger age distribution were perinatally HIV-infected.

Table 2. Pediatric Cohort of Long-term Survivors [View Table]

As seen in Table 3, among the 95 long-term survivors who had developed AIDS-defining conditions, 14 (15%) had absolute CD4 counts 500 cells/mm³ 26 (27%) had values between 200 and 500 cells/mm³, and 55 (58%) had <200 cells/mm³ (26% of group 1, 48% of group 2, and 62% of group 3 children). Among children with CD4 counts 500 cells/mm³, lymphoid interstitial pneumonitis was the most common AIDS-defining condition followed by recurrent bacterial infections. Other AIDS-defining conditions present in very small numbers throughout all ranges of CD4 counts were HIV encephalopathy, disseminated herpes simplex infection, lymphoma, and tuberculosis. Children with intermediate and lower CD4 counts had a higher prevalence of frequent recurrent bacterial infections followed by lymphoid interstitial pneumonitis. Only children in the lower CD4 count range developed disseminated cytomegalovirus infection, Mycobacterium avium intracellulare infections and toxoplasmosis. Therefore, among children in the cohort who had had an AIDS-defining condition, lymphoid interstitial pneumonitis (LIP) was the most common finding in the 500 cells/mm³ strata with recurrent bacterial infections being the most common cause of AIDS in the other CD4 groups. There were no differences in the prevalence of AIDS-defining conditions between perinatally infected children and children infected by transfusion of blood products (data not shown).

Table 3. AIDS-defining Conditions Among Pediatric Long-term Survivors [View Table]

Because children with a history of lymphoid interestitial pneumonitis are known to have a better prognosis than those with other AIDS-defining conditions, the prevalence of an AIDS diagnosis within each cohort was evaluated excluding patients with only LIP as their AIDS-defining condition (as seen in Table 4). The proportion of children with AIDS decreased from 26% to 19% in the >500 CD4 group, from 48% to 44% in the 200 to 500 CD4 cohort and from 62% to 56% in the <200 CD4 group. Therefore, by excluding children who had LIP as their only AIDS-defining condition from the AIDS cohort the number of children with significant clinical immunosuppression within each CD4 group did not change substantially. However, few children had only LIP as an AIDS-defining condition. When patients were further stratified according to presence or absence of LIP and absolute CD4 count range as seen in Table 5, children with lower absolute CD4 counts and a positive history for this disease had a much lower prevalence of other AIDS-defining conditions (29%) than the group of patients who had never had LIP (71%) but also had very low CD4 counts. This finding, however, was not observed within the cohorts of children with higher and intermediate absolute CD4 counts; among these patients the proportion of other AIDS-defining conditions was equally distributed regardless of history of LIP.

Table 4. Proportion of Patients With AIDS-defining Conditions [View Table]

Table 5. Prevalence of Other AIDS-defining Conditions Among Long-term Survivors Stratified According to History of Lymphoid Interstitial Pneumonitis (LIP) [View Table]

Lymphadenopathy was the most common clinical finding in all children (n = 160/197), and within each strata of CD4 count (present in 71% to 86% of patients in each category). Hepatomegaly was the second most frequent finding (n = 117) overall. Oral thrush and recurrent episodes of otitis media were the third most frequent clinical finding (n = 86). Splenomegaly and herpes zoster infection were present in 44% of patients; HIV dermatitis was present in 25%, parotid enlargement in 21%, thrombocytopenia in 17%, intermittent diarrhea in 16%, HIV cardiomyopathy in 10%, and HIV nephropathy in 5%. The prevalence of clinical findings overall increased as CD4 counts decreased with the exception of generalized lymphadenopathy, recurrent otitis media, parotid enlargement, and HIV nephropathy that were found less often in the lower CD4 strata than in the intermediate or higher CD4 count range. There were no differences noted by mode of acquisition of infection.

Overall, 179/197 children (92%) had received antiretroviral therapy (zidovudine, didanosine, and/or zalcitabine). As seen in Table 2, the use of antiretroviral therapy was highly prevalent in this cohort of children regardless of CD4 count or the presence of an AIDS-defining condition.

When subjects were categorized by presence of an AIDS-defining condition and CD4 count, as shown in Fig 2, the use of antiretrovirals, oral candidiasis, recurrent diarrhea, HIV cardiomyopathy, hepatomegaly, herpes zoster infections, HIV nephropathy, parotid enlargement, splenomegaly, and recurrent otitis media were findings more prevalent among children who had had AIDS-defining conditions. Lymphadenopathy and thrombocytopenia had an equal prevalence in both AIDS and non-AIDS groups although HIV dermatitis was a more frequent finding among children without AIDS although in both groups the prevalence increased as CD4 counts decreased.

Forty long-term survivors (20%) who had never had an AIDS-defining condition and had an absolute CD4 count 500 cells/mm³ were identified as LTNS. This group was comprised of 31 perinatally infected children and 9 children infected by neonatal transfusion. Their mean age was 10.0 years (range 8 to 14 years), mean absolute CD4 count was 823 cells/mm³ (range 504 to 3030 cells/mm³); 26 children (65%) were black, 8 (20%) were white, and 6 (15%) were Hispanic. Eighty-two percent of these children had received antiretroviral therapy, and as with the other children in the study, lymphadenopathy was present in a large number of patients (85%). Hepatomegaly was present in 37% of patients, splenomegaly in 25%, 35% had had recurrent episodes of otitis media, 20% had had herpes zoster infections and/or oral thrush, 15% had had thrombocytopenia, with diarrhea and HIV cardiomyopathy present in 10% and 7% of patients, respectively. In the LTNS children, the findings of parotid enlargement and HIV dermatitis were found more frequently than in the LTPS children (25% vs 21% and 22% vs 7%, respectively). These differences, however, were not statistically significant (P > .05).

DISCUSSION

The age distribution of long-term survivors was skewed toward the younger age group (8 years) among perinatally HIV-infected children which is in accordance with reports on the natural history of HIV infection^1,5,6 and also the timing of the pediatric epidemic (historically HIV acquired via blood transfusion occurred earlier in the AIDS epidemic whereas the perinatally acquired HIV epidemic peaked years later). The cohort surveyed only included living children and not children who might have died before or after 8 years because this survey was conducted in anticipation of future long-term survivor studies. Children infected by transfusion of blood products who were long-term survivors tended to be 10 years or older primarily due to the mode of acquisition of infection that became unlikely after 1985.⁷ Because there were fewer children infected via transfusion of blood products after this year, older children were more prevalent within this cohort. The mean age of long-term survivors tended to increase although moving from a higher CD4 count to a lower CD4 count category (9.8 years vs 10.1 years vs 10.4 years). These findings are similar to those reported by the Italian Register in which 21% of perinatally HIV infected children 5 years or older had less than 200 CD4 cells/mm³ at 61 to 72 months with the proportion rising to 41% at more than 96 months.⁸ Grubman et al⁹ also observed that 50% of children older than 9 years of age had CD4 counts <200 with 57% having an AIDS-defining condition.

The population of long-term survivors did not seem to differ significantly when categorized by mode of infection (perinatal vs transfusion). Racial distribution was an exception, with patients of African descent constituting the majority (64%) of vertically HIV-infected children and one-third of the children infected via transfusion of blood products. In both groups a similar proportion of patients had developed AIDS. In addition, when patients were categorized by CD4 strata, there were similar proportions of patients within each group. There were also no differences in clinical findings between perinatally infected children and those infected via blood products. The age distribution between groups differed and this finding was statistically significant (P < .001). The age of long-term survivors infected via blood products, however, was affected by the implementation of HIV blood screening programs as mentioned.

The study design precludes any comparisons regarding the length of survival time after initial exposure to HIV among vertically infected children and neonatal blood transfusion recipients. Recent studies have indicated that the median symptom-free survival time from birth to symptomatic infection may be longer for neonatal transfusion acquired HIV than for perinatally infected.¹⁰ Jones et al⁷ also estimated a longer incubation period from time of infection to diagnosis of AIDS among pediatric transfusion acquired cases when compared to perinatally acquired cases using a nonparametric procedure for truncated data. In adults, studies have attempted to find an association between the route of exposure to HIV-1 and clinical prognosis¹¹ with the rate of progression to AIDS among transfusion recipients being reportedly associated with the disease status of the donors at the time.¹² In perinatally infected infants, the rate of maternal transmission and disease progression appears to correlate with maternal virus burden as well as the timing of infection.¹³ In this study, both groups of HIV-infected children who survived until at least 8 years of age had relatively similar clinical profiles.

Overall, the majority of AIDS-defining conditions within the cohort were either recurrent bacterial infections (n = 51) and/or lymphoid interstital pneumonitis (n = 46). The prevalence of AIDS defining conditions traditionally associated with a poor prognosis such as Pneumocystis carinii pneumonia (n = 13) and disseminated atypical mycobacterial (n = 14) infections was very low with the majority of cases occurring in the group with lower CD4 counts. Among the group of children with higher CD4 counts, the majority of AIDS diagnosis was attributable to lymphoid interstital pneumonitis (n = 9) followed by recurrent bacterial infections (n = 6). The proportion of patients with AIDS in each CD4 cohort did not change significantly when children with LIP as the only AIDS-defining condition were excluded from the AIDS category. Interestingly, however, when children were stratified as to presence or absence of an LIP diagnosis within each CD4 count range, children who had had LIP had fewer additional AIDS-defining conditions than those with a negative history for this disease within the lower CD4 count strata. This finding is in accordance with reports associating LIP with a better disease prognosis.

The majority of clinical findings described occurred more frequently in the group of children with AIDS, with the exception of generalized lymphadenopathy that was highly prevalent in the entire cohort, thrombocytopenia that had a similar prevalence in both groups, and HIV dermatitis that was found more frequently in the children without an AIDS-defining diagnosis. As CD4 counts decreased within the cohort, the prevalence of most clinical findings increased in accordance with the natural history of HIV infection. Exceptions were the prevalence of lymphadenopathy that again did not change as well as recurrent otitis media, parotid enlargement, and HIV nephropathy that were not present as frequently in children with lower CD4 cell counts in our survey.

Among this cohort of pediatric long-term survivors, 20% of infected children (n = 40) had an absolute CD4 count >500 cells/mm³ and no AIDS-defining conditions. The demographics of this population was not different from the overall profile of the cohort. The most prevalent clinical finding was generalized lymphadenopathy (85%) followed by hepatomegaly (37%). In addition, 35% of these children had recurrent episodes of otitis media in the past. This population of children probably included a highly selected group of patients who had minimal to no progression of HIV disease such as the cohort of adult nonprogressors described in recent studies.^16,17 All patients in the survey were long-term survivors based exclusively on their age. However, it is apparent that long-term survivors constitute a diverse patient population within the dynamic process of disease progression. Studies have demonstrated that the median survival time of perinatally HIV-infected children is approximately 8 to 9 years.^8,10 Therefore, children who have lived until 8 years of age have survived longer than the median length of time in 50% of cases. As time passes, it is likely that median survival will be longer than 8 years particularly if intervention therapies are effective. In this context, true nonprogressors constitute a very unique group of patients and must be selected for the possibility of future investigational studies. Children who maintain a relatively normal CD4 count and remain free of significant clinical symptoms are long-term survivors who have managed to maintain their immunological status. A detailed investigation of this subpopulation will potentially add to the understanding and management of HIV disease.