MATERIALS AND METHODS

Patients

Between August 1988 and February 1999, a retrospective follow up study of a cohort of 240 vertically HIV-1 infected children from the Departments of Paediatrics of the General University Hospitals “Gregorio Marañón” and “12 de Octubre” Madrid, and “Virgen del Rocío” Hospital Sevilla, Spain, was performed. Of 240 HIV-1 infected children who had at least two different immunological and virological evaluations, 150 children who had adequate data on AIDS stage at entry and subsequent progression were selected for analysis of their progression to AIDS, while receiving various ART regimens given for up to 30 months (study A). Those children who were already in clinical category C, and those for whom we did not have enough evaluations were excluded from study A.

The children were divided into four groups according to the ART protocol given during the follow-up: (1) NT group (not treated): 58 ART naïve children; (2) MT group (on monotherapy): 36 children treated with a nucleoside analogue HIV-1 reverse transcriptase inhibitor (NRTI) alone, and who had previously been ART naive; (3) CT group (on combination therapy): 31 children treated with two NRTIs; and (4) HAART group (on potent antiretroviral therapy): 25 children treated with two NRTIs and at least one HIV protease inhibitor (PI) (table 1). No child in the dual ART or HAART groups was included in the NT or MT groups. Progression to AIDS was defined as progression to clinical category C. Children who entered the study earlier tended to have no therapy or monotherapy, compared with those who entered the study later.

Sixty one of the 240 children were enrolled into a longitudinal study to assess the evolution of the most relevant immunological and virological markers of progression to AIDS (study B). For inclusion in this second study group, patients had to meet the following selection criteria: (a) duration of follow up 12–24 months; (b) aged above 2 years when they entered the study; and (c) to have had at least four determinations of log₁₀ plasma VL and %CD4+ and %CD8+ lymphocytes, in order to estimate the slope as well as the overall trend of these variables. Of the 61 children, 13 were in the NT, 16 the MT, 15 the dual-ART, and 17 the HAART groups (table 2).

All infants were diagnosed as HIV-1 infected on the basis of positive results in both DNA polymerase chain reaction and virus culture assays, as described previously.¹⁷ Clinical classification of the children was based on the 1994 revised guidelines of the Centres for Disease Control and Prevention (CDCP).¹⁸ The study was conducted according to the declaration of Helsinki and approved by the ethical committee for all hospitals involved. Drugs were prescribed by the treating physician according to CDCP guidelines³ after obtaining written informed consent from legal guardians. Response to therapy was controlled at 0, 3, 6, 9, 12, 15, and 18 months by serial measurements of %CD4+ and %CD8+, VL, and by collecting clinical data, according to published guidelines.³

Quantification of T cell subsets in peripheral blood

T lymphocyte subsets in peripheral blood were quantified by direct immunofluorescence using monoclonal antibodies of the T series and multiparameter flow cytometry (FACScan, Becton-Dickinson, Heidelberg, Germany), as described previously.¹⁷

Quantitative HIV-1 RNA assay

Blood samples were collected in EDTA tubes, separated within four hours, and plasma was stored at −70°C. The VL was measured in 200 μl of plasma by reverse transcriptase polymerase chain reaction (Amplicor monitor kit, Roche Diagnostic Systems, Brandenburg, New Jersey, USA). The lower limit of detection was 200 HIV-1 RNA copies/ml.^19,20

Statistical analysis

CD4+ and CD8+ counts are expressed as percentages. In all analyses, HIV-1 RNA concentrations were transformed to log₁₀ scale in order to normalise their distribution. Slopes of %CD4+ and %CD8+ and log₁₀ VL within individuals were calculated, expressing them as %CD4+ and %CD8+ per year and log₁₀ VL per year. Slope calculations were then analysed by multiple linear regression analysis to assess the effects of antiretroviral combination therapy on the trends (slopes) of peripheral blood CD4+ T lymphocytes and plasma VL during the course of follow up. This multiple linear regression analysis was performed for all children collectively and adjusted for baseline %CD4+, %CD8+, log₁₀ VL, and age, to compensate for differences among groups at entry to the study. Estimated marginal means of the slopes of each group according to the ART protocol were obtained. Progression to AIDS was determined by the Greenwood method using Kaplan–Meier curves. Statistical significance was set at p < 0.05; probabilities were compared by the log rank test (Mantel–Haenzel). This analysis was performed for all children according to ART groups. The time to reach baseline and its relative risk (RR) were estimated by the proportional hazard Cox regression equation for VL and %CD4+. RR <1 indicates a protective effect.

RESULTS

Influence of ART protocol on the clinical evolution of HIV-1 disease

In study A, 150 HIV-1 infected children were followed for up to 30 months. Table 1 shows baseline immunological and virological characteristics. Table 2 shows drugs used in both groups.

Figure 1 shows progression to AIDS. No child in the HAART group progressed to AIDS and only two of the 31 children in the dual-ART group progressed to AIDS. The NT and MT groups progressed to AIDS faster (14 of 58 and seven of 36 children, respectively; p = 0.03 and p = 0.04 compared with the HAART group).

Longitudinal study of immunological and virological progression

In study B, two of 17 children in the HAART group were treated with an NRTI, and another two with two nucleoside analogues prior to enrolment; the remaining 13 children were ART naïve. Table 3 shows baseline clinical, virological, and immunological characteristics of the 61 children.¹⁸ When the baseline %CD4+ and %CD8+ were analysed, no statistically significant differences between treatment groups were observed, although there was a trend towards lowest %CD4+ in children receiving more complete treatment. However, significant differences were found when analysing normalised VL values (log₁₀) between the NT and HAART groups (p = 0.01).

The mean values of %CD4+ and %CD8+ in peripheral blood, and log₁₀ VL in all groups were analysed at 0, 3, 6, 9, 12, 15, and 18 months of follow up. In the NT group there was no reduction of mean VL, which remained high throughout the follow up period (fig 2). In contrast, %CD4+ decreased, %CD8+ increased, with inversion of the CD4:CD8 ratio over time, as expected from the known natural history of infection. In the MT group, %CD4+ and %CD8+ remained constant during the first 12 months, but after 15 months of therapy a clear fall in VL and a concomitant increase in %CD8+ were observed (fig 2). When analysing the HAART group, a completely different pattern of response to treatment from the other groups was observed: there was a dramatic decrease in VL after the third month of treatment, which remained low during the remaining follow up period, and an increase in %CD4+, with stable %CD8+ (fig 2).

In the HAART group there was a trend towards an increase in CD4+ T lymphocytes (9.89 (1.78) increase in % CD4 per year). Similar differences, although of the opposite sign, were observed with respect to log₁₀ VL in the HAART group. In this group of children, a clear downward trend in VL during the entire study period was observed (−0.66 (0.18) log₁₀ VL per year). There were significant differences between the HAART group and the other groups for %CD4+ trends and log₁₀ VL trends (p < 0.01). The mean increase of 8 units of %CD4+/year was greater in the HAART group than in the other groups. Likewise, when VL was analysed, highly significant differences were found between the HAART group and the NT and MT groups, with an estimated mean difference of −0.65 log₁₀ per year. The negative sign indicates that in the one year period, the decrease in VL was 0.65 log₁₀ VL, greater in the HAART group than in the NT and MT groups.

Kaplan–Meier and Cox regression analyses to define success versus failure of therapy

Time to return to baseline VL and %CD4+ (considered to be two end points representing failure of therapy), were assessed by the Kaplan–Meier method and Cox regression analysis, as well as the relative risk (RR) of reaching both parameters. As table 4 shows, virological differences were noted between the three groups of treated children in time and RR of reaching baseline VL. Similarly, with regard to time in reaching baseline %CD4+ T cells, immunologically important differences were noted between groups (table 4). A total of 42.8% of children in the MT group, 47.6% in the dual-ART group, and 50% in the HAART group achieved undetectable VL. Highly significant differences between the three groups of treated children compared with the NT group (p < 0.05) were observed. When analysing RR for reaching baseline %CD4+ values, children from the dual-ART (p = 0.025) and HAART groups (p = 0.002) showed highly significant differences compared with the NT group, indicating a greater protective effect.

DISCUSSION

We report an observational study to assess the progression to AIDS and the evolution of markers of HIV-1 infection progression (VL and %CD4+) in HIV-1 infected children receiving different ART regimens. It is important to consider that essential aspects of treatment such as nutritional management and prophylaxis against Pneumocystis carinii pneumonia, changed substantially over the course of the 11 year period of the study. Prophylaxis is probably at least as important as ART in preventing the onset of AIDS in children. As expected from the natural history of HIV-1 infection, in the NT group of children a decrease in CD4+ T cells was observed, not just as a consequence of infection, but also of increasing age.^1,21 In addition, an increase of the %CD8+ and VL in parallel with the decrease of %CD4+ was observed, in accordance with data published elsewhere.^1,22 Our results indicate that the HAART group had a significant difference in the rate of progression to AIDS compared to the MT group, but not to the dual-ART group, in agreement with previous results.²³ However, a better trend of evolution of both %CD4+ and VL is clearly seen in children on HAART, as has also been reported in the literature.^5,24

Monotherapy regimens for the management paediatric infection have been shown to be of benefit in symptomatic children, resulting in a clear improvement in quality of life, growth, and neurological development, as well as immunological and virological parameters,^25–28 as confirmed in this study. However, our data indicate that this beneficial effect is transitory and disappears with time, as is illustrated in the trends in VL and %CD4+ (fig 2), probably as a consequence of the development of drug resistance.

Dual therapy has been shown to be much more effective than monotherapy in adults; studies in ART naïve symptomatic children were therefore initiated, with similar conclusions.^13,29 Nevertheless, in our study combination therapy was not shown to be more effective than monotherapy: the decrease of VL and increase of %CD4+ were similar in both MT and dual-ART groups, and no significant differences in the longitudinal trends of both parameters were observed. These data contradict results obtained in clinical trials that included only naïve children.^12,13,27 However, in our study, among the group of children on combination therapy, 12 of 15 had previously received monotherapy and were changed to dual therapy after treatment had failed.

The benefits of HAART in infants and children previously treated with nucleoside analogues have also been shown previously.^14,30 In agreement with other studies, a dramatic decrease in VL after the third month of treatment was observed in children in the HAART group, with a concomitant increase of %CD4+ and %CD8+ that subsequently remained within the normal range for age in each child. These data support the greater effectiveness of HAART than other regimens, in accordance with clinical trials.³⁰ The increase in %CD4+ was significantly higher in the HAART group; inhibition of viral replication presumably allows recovery of CD4+ T cell numbers, in agreement with previous studies.^5,24 Undetectable VL was achieved in half of the children, and a dramatic reduction of VL occurred in the remainder. This incomplete reduction reflects differences between the evolution of infection in adults and children. Therefore, failure to reach undetectable VL in children should probably not be considered as treatment failure, at least with currently available drugs.^5,11,14 The time taken for %CD4+ and VL (two end points of therapy failure) to return to baseline, was longer in all three groups of children receiving ART than in untreated children. The time to reach undetectable VL was the same for all three groups of treated children, although a higher proportion in the HAART group reached it. Several factors may be operating; for example, some doses may not result in serum concentrations sufficient to inhibit viral replication completely, leading to the earlier occurrence of quasispecies resistant to administered drugs.

In this observational study, differences between ART treatment regimens appear less significant than in clinical trials. Adherence to treatment by the children was not monitored in this study, whereas in clinical trials measures are taken to ascertain adherence. However, the data presented here are derived from routine clinical practice, where children have a wide variety of social and other conditions (including adoption, being orphaned, living with grandparents or relatives, or in community centres). The higher the number of drugs administered, the more difficult it is to determine the efficacy of treatment.

Our study shows that in this area as in so many others, children are not simply small adults. HIV-1 infection in children is mostly acquired by vertical transmission, a special form of acute infection. Further studies are required to investigate more effective therapeutic strategies for suppressing VL to undetectable amounts and to increase %CD4+ rapidly and effectively.