Mother-to-child transmission is the main cause of HIV-1 infection in young children. According to the World Health Organization, there were about 700,000 new infections in children worldwide during the year 2004, most of them in resource-limited countries. In affluent countries, tests for HIV-1 RNA or DNA are used for identification of infected children. In those found infected, concentrations of CD4+ T cells and HIV-1 RNA are determined every 3 months to guide decisions about initiation of antiretroviral therapy (ART) and to monitor the efficacy of ART once started.1,2 Systems for cost-effective CD4+ counting applicable to resource-poor settings have recently become available.3,4 However, measurement of HIV-1 RNA using commercial kits remains expensive in most countries and, in addition, requires rapid transportation of the samples to the laboratory, complicated specimen pretreatment procedures, sophisticated laboratory equipment and infrastructure, and well-trained personnel. In resource-poor countries, these requirements are met only in few, mostly urban, treatment centers, thereby limiting the efficacy of area-wide ART programs. These considerations also apply largely to real-time polymerase chain reaction (PCR) procedures despite their considerably reduced reagent costs.
Alternative viral load tests have therefore been sought. HIV-1 p24 antigen (p24) measured by signal-amplified enzyme-linked immunosorbent assay (ELISA) of heat-denatured plasma and, more recently, particle-associated reverse transcriptase (RT) activity were proposed (reviewed by Schupbach5 and Crowe et al6). Like HIV-1 RNA, RT is a marker which in plasma samples is found only inside viral particles. As a consequence, there should be an excellent linear correlation between the concentrations of HIV-1 RNA and RT, at least if inhibitors or enzyme degradation during transportation in hot environments is absent.7,8 Although p24 as a structural HIV-1 protein is also present in the virion, ultracentrifugation experiments have demonstrated that the overwhelming proportion of p24 detectable in plasma of chronically infected patients is located outside particles and present in free or immune-complexed form.9 As a consequence, the correlation between HIV-1 RNA and p24, although highly significant in all studies, is less tight than that between the particle-specific markers, based on the individually different quantities of intraviral and extraviral p24 which, in addition, may also change over time in a given patient.8,10-20 However, this lesser correlation does not imply that p24 must be an inferior viral marker. Studies in adults have demonstrated that p24 is comparable with HIV-1 RNA as an independent predictor of CD4+ decline, clinical progression, and mortality in Swiss and US patients.12,15 In addition, some studies have now found a better correlation between CD4+ counts and p24 than between CD4+ counts and HIV-1 RNA.8,18,21
Based on own experience and numerous studies demonstrating a dependence of the fate of CD4+ T cells on the HIV-1 RNA load (reviewed by the Panel on Clinical Practices for Treatment of HIV Infection22), a physician will expect that a patient with a high viral RNA load at a given visit will more likely lose CD4+ cells during the time to the next visit than one with a low viral load. If an initially high viral load is reduced by ART, CD4+ cells are expected to increase. Conversely, if cessation of ART or if viral failure increases an initially low viral load, the CD4+ cells are likely to drop. In short, both the initial viral load and its successive change are likely to affect the CD4+ cell count in an inverse relationship.
The present study investigates whether this relationship can be demonstrated in pediatric HIV-1 infection. We investigated which viral parameter, HIV-1 RNA or p24, was better correlated with the short-term CD4+ T-cell changes observed between consecutive visits of HIV-1-infected children. Twenty-four children were followed for a median time of almost 4 years. Prospectively measured concentrations of p24 and HIV-1 RNA and their relative changes during consecutive intervals within children were related to the respective relative changes in CD4+ counts.
MATERIALS AND METHODS
HIV-1-infected children seen between September 1992 and August 2000 at the HIV outpatient clinic of the University Children s Hospital of Zurich were enrolled after informed consent. Most patients were participating in clinical trials of the Pediatric European Network for Treatment of AIDS or of the Pediatric AIDS Group of Switzerland. Blood samples were collected at time points dictated by the trial protocols, before change of treatment or at regular 3-month intervals.
Quantification of Viral Components
The Amplicor HIV Monitor version 1.0 kit (Roche, Molecular Systems, Basel, Switzerland) with a nominal lower quantification limit of 400 copies per milliliter was used according to the manufacturer s guidelines for testing of all samples before August 1997. This included 40 samples from the period before availability of the Monitor assay in 1996, which were tested retrospectively. From August to December 1997, the Amplicor HIV-1 Monitor version 1.0 kit was used in combination with Roche s mix-in primers to improve detection of HIV-1 subtypes A and E (now renamed CRF01_AE) or G, and from January 1998 onward, the kit version 1.5 was used with the 50-copies-per-milliliter quantification limit procedure. No increases of HIV-1 RNA unparalleled by increases in p24 were seen after the introduction of the mix-in primers. This indicated that virus strains that were not well detected by version 1.0 without the mix-in primers were not widely present in the studied cohort and that, in consequence, sequential data from samples before and after this modification can be merged for each child (see also "Discussion"). For evaluation, detectable HIV-1 concentrations below the categorical cutoffs of 400 or 50 copies per milliliter, respectively, were entered as measured, whereas undetectable concentrations were set to 1 copy per milliliter. All HIV-1 RNA measurements were conducted at the Swiss National Center for Retroviruses, and the laboratory participated regularly, and with continuously excellent results, in international quality control trials (EU Concerted Action Quality Control Panel HIV-1 RNA-PCR, CLB Amsterdam VQC Panel HIV-1 RNA-PCR, QCMD Proficiency Panel HIV-1).
All p24 measurements in heat-denatured plasma were conducted prospectively at the Swiss National Center for Retroviruses, as described earlier,14 and largely by the same experienced person. Briefly, 100 μL plasma was diluted with 500 μL of 0.5% Triton X-100 in 1.5-mL Eppendorf tubes, denatured by heating at 100oC for 5 minutes on a Techne (Cambridge, UK) dry heat block, and tested in duplicate with the NEN/DuPont s (now PerkinElmer s) HIV-1 Core Profile ELISA in combination with the ELAST ELISA amplification system (both purchased from NEN Life Science Products, Geneva). Absorbance was read using a Dynatech MR5000 ELISA reader (Microtech Produkte, Embrach, Switzerland). p24 was quantitated by means of combined kinetic and end-point analysis using the Quanti-Kin Detection System (DL3, Diagnostica Ligure srl, Genova, Italy). This permitted quantification in a range from about 500 to 6,250,000 fg/mL with a single sample dilution.14 Each ELISA plate carried a blank, 4 negative controls, and the full set of 9 quantitative standards. To maximize interrun precision, we used the same prediluted quantitative standards frozen in aliquots at −30°C. Intrarun quality control was based on the Quanti-Kin Detection System software. A cutoff corresponding to the mean of the 4 negative controls plus 3 standard deviations was used. This study did not yet use an improved virus lysis buffer developed more recently to further increase sensitivity.18
Absolute numbers and percentages of peripheral blood CD3+, CD4+, and CD8+ T cells were determined prospectively in the Clinical Immunology Laboratory of the Zurich University Children s Hospital, using a whole blood preparation method and 3-color flow cytometry as described23 and ''4-color dual platform flow cytometry on a FACScalibur (BD Biosciences, Basel, Switzerland) and a MultiTEST program (BD Biosciences). Gating was done on side scatter channel and CD45+ lymphocytes. Absolute lymphocyte values were calculated using a Sysmex XE-2100 counter (Sysmex, Kobe, Japan). The laboratory participated regularly in the international quality assessment program UK National External Quality Assessment Service.
Stata version 9.0 (StataCorp LP, College Station, TX) was used for statistical analyses. To account for the hierarchical data structure with repeated measurements within children, we used linear mixed models with child id (intercept) and time since the start of ART treatment (slope) as random effects. We thus assessed the population average relationship between the relative change in the 2 log-transformed HIV parameters, HIV-1 RNA and HIV-1 p24 [ie, log10 (second measurement) − log10 (first measurement)], and the associated relative change in CD4+ cell count, defined as the difference of the log-transformed CD4 counts of the second and first measurements [ie, log10 (second measurement) − log10 (first measurement)]. In the analyses, we controlled for gender, child s age at start of interval for successive measurements, the number of days between successive measurements, and the time since start of ART treatment. To evaluate the statistical support for the p24 model compared with that of the HIV-1 RNA model, we used the Akaike Information Criterion (AIC) to calculate AIC weights, which indicate the relative likelihood of the model given the data. We investigated the degree of agreement in the predictions from the 2 models, using the limits-of-agreement procedure of Bland and Altman and the concordance correlation coefficient of Lin.24,25
Twenty-four children (17 girls and 7 boys) infected by HIV-1 through maternal transmission were followed. Relevant baseline and follow-up information is presented in Table 1. The time window covered the period from 1992 to 2000, but 98.7% of all data were generated from 1996 to 2000.
At study entry, stages according to the Centers for Diseases Control classification were A2 (n = 3), A3 (n = 1), B1 (n = 1), B2 (n = 2), B3 (n = 2), C2 (n = 1), and C3 (n = 14).26 Eight children were receiving no treatment; 4 of these then directly received a 2-class triple-drug combination (highly active antiretroviral therapy; HAART), whereas the other 4 first received a single-class combination of at least 2 drugs before receiving HAART. Sixteen children were already taking antiretroviral drugs at study entry. Eight of the 16 were initially on single-class combinations before receiving HAART; 3 received HAART at all time points; and the remaining 5 had more complicated treatment courses with episodes of single-class bitherapy or tritherapy, no therapy, and HAART. Altogether, 312 (78.2%) of the 399 evaluable data sets were obtained from children while under HAART, 15 (3.8%) while untreated, 68 (17%) while receiving single-class treatment with at least 2 drugs, and 4 (1%) while receiving a '''2-drug 2-class combination. All children were receiving HAART at the end of the study period.
Concentrations of HIV-1 RNA and p24 were above the cutoff in 368 (92.2%) and 326 (81.7%) samples, respectively. Both markers were positive in 315 samples and negative in 20samples, resulting in a concordance of 84.0% (Fisher exact P value < 0.0001). Of the 64 discordant samples, 53 were positive for HIV-1 RNA but negative for p24, and 11 samples were positive for p24 but negative for HIV-1 RNA. The detectable concentrations of HIV-1 RNA ranged from 5 to 1.8 × 106 copies per milliliter and those of p24 ranged from 240 to 600,000 fg/mL.
Determinants of CD4+ Change Between Visits
Data on corresponding changes in concentrations of HIV-1 RNA and p24 were available for 375 intervals. Controlling for variation between individuals, there was a positive relationship between concentrations of HIV-1 RNA and p24 (P < 0.0001; Fig. 1). While controlling for initial CD4+ T-cell count, age, sex, days since start of therapy, and number of days between visits, the change in CD4+ T cells was negatively correlated to the corresponding change in HIV-1 RNA and independently so of the initial HIV-1 RNA concentration (Table 2, model 1). Thus, a change in HIV-1 RNA was associated with an inversely directed change in CD4+ T cells (Fig. 2A). Similarly, we found a negative relationship with the change in p24 over the interval (Fig. 2B), whereas there was also a trend for an additional negative effect of the initial concentration of p24. This indicates that an increase in p24 corresponds to a decline in CD4+ T cells and that this decline in CD4+ cells tended to increase with the initial p24 concentration. Indirect model comparison using AIC revealed a slightly higher statistical support for the p24 model than for the HIV-1 RNA model (AIC weights of 0.63 and 0.37, respectively). The mean difference in the predicted log10 change in CD4+ cell counts between the 2 models was 0.002 (SD = 0.052), and 89% of the data were within the Bland-Altman limits of agreement [95% confidence interval (CI): −0.100 to 0.104]. Best agreement in prediction was found for measurement intervals that showed large increases in CD4+ cell count (Fig. 3). The overall concordance correlation coefficient was 0.95 (95% CI: 0.94-0.96, P < 0.001).
This is the first study in which short-term changes in CD4+ counts in HIV-1-infected children were assessed for their dependence on the prospectively determined viral markers HIV-1 RNA and HIV-1 p24. The results show that the relative CD4+ changes observed between consecutive visits showed a trend for being related to the p24 concentration measured at the first of these visits and were clearly related to the relative change in p24 concentration between the 2 visits. Relative change of p24, and quite possibly also its initial level, is thus relevant for the short-term relative CD4+ cell change, whereas with respect to HIV-1 RNA, only its relative change was found relevant (Table 2; Fig. 2). Overall, p24 measurement "explained" the relative CD4+ cell change slightly better than did HIV-1 RNA. These results suggest treatment monitoring of connately infected children based on p24 as a potential alternative for the standard monitoring procedure based on HIV-1 RNA. Using HIV-1 p24 instead of HIV-1 RNA has the further advantage of higher environmental stability (no express transport to the laboratory needed;'storage at −20°C for many years); lower technical complexity of the test, which results in lower costs for equipment and technical infrastructure and lower demands regarding the education and training of the technical personnel performing the assay; and an attractive price that, based on country and purchase volume, is between US $5 and $10 per single-well test.7
HIV-1 RNA was originally chosen as a marker for viral monitoring based on its prediction of long-term disease progression and responsiveness to initiation of ART.22 The present study now suggests that, although it was less frequently positive than was HIV-1 RNA and it exhibited only a modest correlation with HIV-1 RNA (Fig. 1), p24 may be equally relevant to the changes in CD4+ cells, whose destruction is the central feature of AIDS in both children and adults. Most of the p24 present in plasma of chronically infected patients are located outside the particles and may thus represent virus expression in tissues, which is not reflected by the concentration of HIV-1 RNA or RT in plasma. Inasmuch as long-term CD4+ decline corresponds to the sum of all the small changes observed during successive short intervals such as those investigated here, this result also bears relevance to long-term CD4+ decline in HIV-1-infected children.
Regarding possible limitations of our findings, several versions of HIV-1 RNA measurement (see "Materials and Methods") were used during the study period, which might negatively affect the precision of HIV-1 RNA measurement. We examined this possibility but did not find significant differences in HIV-1 RNA concentration in dependence of the test version. When looking at the individual courses of the children and focusing on the transition points between the different Monitor versions, we found that 20 of the 375'interval data sets were associated with a switch from version 1.0 to mix-in primers, whereas another 20 interval data sets were'associated with a switch from mix-in primers to version 1.5. We did not find changes in HIV-1 RNA concentration that were not associated with similar changes in p24. That such changes, which might reflect an improved detection of non-B subtypes by the improved test versions, were not found is also'in'accordance with the fact that most children in this group are infected with HIV-1 subtype B, which caused more than 95% of adult HIV infections in Switzerland in the period between 1981 and 1994 when these children were born.27 Moreover, most of the children investigated here were born to mothers infected by intravenous drug use. In a representative'investigation of HIV-1 infections newly diagnosed in 1997-1998, 94% of the adults with an intravenous drug use risk still carried a subtype B virus.28 In accordance with these considerations, a subtype B virus was indeed identified in all of 9 children of the present study for whom genetic resistance testing'was'performed. Thus, there is no indication that the modifications in HIV-1 RNA quantification affected the study to a relevant degree.
We are also aware that our results are based on a sample of 24 children only. In addition, there was limited statistical power to associate changes in HIV-1 RNA and p24 with those in CD4+ cell count because ART was successful throughout the study period in nearly all children, as reflected in minimal to moderate changes in monitoring parameters over most successive visits. As a result, our predictive models are primarily driven by the infrequent, more extreme periodic alterations in viral load, p24, and CD4+ cell count. It especially remains to be shown whether the prognostic value of p24 indicated using Swiss data is transposable to children receiving ART in resource-poor countries. Therefore, to further validate the potential of p24 in pediatric HIV-treatment monitoring indicated here, there is an urgent need for additional comparative HIV-1 RNA and p24 data, preferably from a larger cohort of children in high- and low-income settings. As mentioned in the "Introduction," there is no direct, constant relationship between HIV-1 RNA and p24. It is therefore not possible to translate a p24 concentration into a corresponding HIV-1 RNA concentration. Neither is it possible to assess the risk of progression associated with a given p24 level by means of the risk associated with a corresponding HIV-1 RNA level. Future studies will thus have to define the p24 concentration below which a short-term CD4 decline does not normally occur and the slopes of CD4 decline associated with different higher p24 concentrations. Previous work has demonstrated that a p24 concentration above 5 pg/mL carried about the same risk of progression to AIDS as did HIV-1 RNA of greater than 30,000 copies or a CD4+ count of less than 350/μL.15
Furthermore, the application of these findings to resource-poor settings worldwide may be limited by a suboptimal detection of p24 of certain non-B subtypes by the PerkinElmer HIV-1 Core Profile ELISA. Importantly, however, several studies have now demonstrated good sensitivity of the test, at least if conducted with our improved virus lysis buffer,18 in infected children from Rwanda, South Africa, and Zimbabwe, that is, countries whose HIV-1 epidemics are dominated by subtype C.17,29,30
In conclusion, our study suggests that monitoring of p24 in HIV-1-infected children can fully replace the standard procedure based on measurement of HIV-1 RNA; this should now be investigated in larger studies. p24-based monitoring should currently be limited to subtypes B and C, whose p24 has already been shown to be detected well by the PerkinElmer test. Our findings also imply that a mere comparison of p24 to HIV-1 RNA (setting the latter as an infallible gold standard) is not appropriate for evaluation of alternative markers; both markers must be evaluated in parallel with respect to a defined task such as prediction of CD4+ decline, clinical progression, or death.
We thank Dr Matthias Egger for helpful discussions, and Antonietta Baumgartner and Lucia Bertodatto for excellent work in the laboratory.
1. Sharland M, Blanche S, Castelli G, et al. PENTA guidelines for the use of antiretroviral therapy, 2004. HIV Med
. 2004;5(Suppl 2):61-86.
2. Mofenson LM. US Public Health Service Task Force recommendations for use of antiretroviral drugs in pregnant HIV-1-infected women for maternal health and interventions to reduce perinatal HIV-1 transmission in the United States. MMWR Recomm Rep
. 2002;51:1-38. quiz CE1-4.
3. Cassens U, Gohde W, Kuling G, et al. Simplified volumetric flow cytometry allows feasible and accurate determination of CD4 T lymphocytes in immunodeficient patients worldwide. Antivir Ther
4. Janossy G, Jani IV, Bradley NJ, et al. Affordable CD4(+)-T-cell counting by flow cytometry: CD45 gating for volumetric analysis. Clin Diagn Lab Immunol
5. Schupbach J. Viral RNA and p24 antigen as markers of HIV disease and antiretroviral treatment success. Int Arch Allergy Immunol
6. Crowe S, Turnbull S, Oelrichs R, et al. Monitoring of human immunodeficiency virus infection in resource-constrained countries. Clin Infect Dis
7. Jennings CL, Fiscus SA, Danilovic AD, et al. Comparison of two human immunodeficiency virus (HIV) RNA surrogate assays to the standard HIV RNA assay. J Clin Microbiol
8. Stevens G, Rekhviashvili N, Scott LE, et al. Evaluation of two commercially available, inexpensive alternative assays used for assessing viral load in a cohort of human immunodeficiency virus type 1 subtype C-infected patients from South Africa. J Clin Microbiol
9. Schupbach J. Measurement of HIV-1 p24 antigen by signal-amplification- boosted ELISA of heat-denatured plasma is a simple and inexpensive alternative to tests for viral RNA. AIDS Rev
10. Schupbach J, Flepp M, Pontelli D, et al. Heat-mediated immune complex dissociation and enzyme-linked immunosorbent assay signal amplification render p24 antigen detection in plasma as sensitive as HIV-1 RNA detection by polymerase chain reaction. AIDS
11. Boni J, Opravil M, Tomasik Z, et al. Simple monitoring of antiretroviral therapy with a signal-amplification-boosted HIV-1 P24 antigen assay with heat-denatured plasma. AIDS
12. Ledergerber B, Flepp M, Boni J, et al. Human immunodeficiency virus type 1 p24 concentration measured by boosted ELISA of heat-denatured plasma correlates with decline in CD4 cells, progression to AIDS, and survival: comparison with viral RNA measurement. J Infect Dis
13. Bürgisser P, Vernazza P, Flepp M, et al. Performance of five different assays for the quantification of viral load in subjects infected with various subtypes of HIV-1. J Acquir Immune Defic Syndr
14. Nadal D, Böni J, Kind C, et al. Prospective evaluation of amplification-boosted ELISA for heat-denatured p24 antigen for diagnosis and monitoring of pediatric HIV-1 infection. J Infect Dis
15. Sterling TR, Hoover DR, Astemborski J, et al. Prognostic value of heat-denatured HIV-1 p24 antigen and correlation with plasma HIV-1 viral load and CD4+
T-lymphocyte level in adults. J Infect Dis
16. Ribas SG, Ondoa P, Schupbach J, et al. Performance of a quantitative human immunodeficiency virus type 1 p24 antigen assay on various HIV-1 subtypes for the follow-up of human immunodeficiency type 1 seropositive individuals. J Virol Methods
17. Pascual A, Cachafeiro A, Funk ML, et al. Comparison of an assay using signal amplification of the heat-dissociated p24 antigen with the Roche Monitor human immunodeficiency virus RNA assay. J Clin Microbiol
18. Schupbach J, Boni J, Bisset LR, et al. HIV-1 p24 antigen is a significant inverse correlate of CD4 T-cell change in patients with suppressed viremia under long-term antiretroviral therapy. J Acquir Immune Defic Syndr
19. Jennings CL, Danilovic AD, Burgess DL, et al. A Comparison of Two Non-RNA-Based Assays for the Quantitation of HIV. Poster # 03-A-3803-ASM. Presented at: 43rd ICAAC; Sept. 14-17, 2003; Chicago, Ill.
20. Respess RA, Cachafeiro A, Withum D, et al. Evaluation of an ultrasensitive p24 antigen assay as a potential alternative to human immunodeficiency virus type 1 RNA viral load assay in resource-limited settings. J Clin Microbiol
21. Schupbach J, Gunthard H, Joos B, et al. HIV-1 p24 may persist during long-term highly active antiretroviral therapy, increases little during short treatment breaks, and its rebound after treatment stop correlates with CD4+
T cell loss. J Acquir Immune Defic Syndr
. 2005;40: 250-256.
23. Anonymous. 1997 revised guidelines for performing CD4+
T-cell determinations in persons infected with human immunodeficiency virus (HIV). Centers for Disease Control and Prevention. MMWR Morb Mortal Wkly Rep
24. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res
25. Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics
26. 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep
27. Stoeckli TC, Steffen-Klopfstein I, Erb P, et al. Molecular epidemiology of HIV-1 in Switzerland: evidence for a silent mutation in the C2V3 region distinguishing intravenous drug users from homosexual men. Swiss HIV Cohort Study. J Acquir Immune Defic Syndr
28. Böni J, Pyra H, Gebhardt M, et al. High frequency of non-B subtypes in newly diagnosed HIV-1 infections in Switzerland. J Acquir Immune Defic Syndr Hum Retrovirol
29. Sherman GG, Stevens G, Stevens WS. Affordable diagnosis of human immunodeficiency virus infection in infants by p24 antigen detection. Pediatr Infect Dis J
30. Zijenah LS, Tobaiwa O, Rusakaniko S, et al. Signal-boosted qualitative ultrasensitive p24 antigen assay for diagnosis of subtype C HIV-1 infection in infants under the age of 2 years. J Acquir Immune Defic Syndr
Keywords:© 2006 Lippincott Williams & Wilkins, Inc.
CD4+ T lymphocyte; p24 antigen; pediatric HIV-1 infection; prognostic marker; resource-poor settings; treatment monitoring; viral load