*These results were presented in part at the European Society of Pediatric Infectious Diseases (ESPID) 2010 Meeting, Nice, France (Abstract 253A).
†See Appendix A.
‡See Appendix B.
Dr Klara M. Posfay-Barbe, Children's Hospital of Geneva, 6, Rue Willy-Donzé, 1211 Geneva 14, Switzerland. Tel: +41 22 382 5462; fax: +41 22 382 5490; e-mail: Klara.PosfayBarbe@hcuge.ch
HIV-infected children have impaired antibody responses after exposure to certain antigens. Our aim was to determine whether HIV-infected children had lower varicella zoster virus (VZV) antibody levels compared with HIV-infected adults or healthy children and, if so, whether this was attributable to an impaired primary response, accelerated antibody loss, or failure to reactivate the memory VZV response.
In a prospective, cross-sectional and retrospective longitudinal study, we compared antibody responses, measured by enzyme-linked immunosorbent assay (ELISA), elicited by VZV infection in 97 HIV-infected children and 78 HIV-infected adults treated with antiretroviral therapy, followed over 10 years, and 97 age-matched healthy children. We also tested antibody avidity in HIV-infected and healthy children.
Median anti-VZV immunoglobulin G (IgG) levels were lower in HIV-infected children than in adults (264 vs. 1535 IU/L; P<0.001) and levels became more frequently unprotective over time in the children [odds ratio (OR) 17.74; 95% confidence interval (CI) 4.36–72.25; P<0.001]. High HIV viral load was predictive of VZV antibody waning in HIV-infected children. Anti-VZV antibodies did not decline more rapidly in HIV-infected children than in adults. Antibody levels increased with age in healthy (P=0.004) but not in HIV-infected children. Thus, antibody levels were lower in HIV-infected than in healthy children (median 1151 IU/L; P<0.001). Antibody avidity was lower in HIV-infected than healthy children (P<0.001). A direct correlation between anti-VZV IgG level and avidity was present in HIV-infected children (P=0.001), but not in healthy children.
Failure to maintain anti-VZV IgG levels in HIV-infected children results from failure to reactivate memory responses. Further studies are required to investigate long-term protection and the potential benefits of immunization.
HIV-infected children are more susceptible to infectious diseases, including varicella zoster virus (VZV) infection. This is likely to result from impaired immune responses, as reflected in a higher rate of vaccine failure to most immunizations . Before highly active antiretroviral therapy (HAART) was available, chickenpox recurred frequently [2–4], and HIV-infected patients were more likely to have bacterial superinfections, pneumonia, cerebellitis and encephalitis following VZV infection [5,6]. More recently, Bekker et al.  reported the frequent loss of antibodies elicited by wild-type infections or immunizations in HAART-treated children. Similarly, several HIV-infected children of the Swiss Mother & Child HIV (MoCHiV) cohort had undetectable anti-VZV immunoglobulin G (IgG) levels despite previously confirmed VZV infection. This observation is intriguing: the persistence of VZV humoral immunity is generally life-long , as community re-exposure and endogenous viral reactivation both contribute to reactivate anti-VZV memory responses and maintain humoral immunity . This suggests limitations in the capacity of HIV-infected children to generate, maintain and/or reactivate immune memory.
In Switzerland, where VZV immunization is only recommended for nonimmune adolescents, VZV is endemic and seroprevalence reaches 95% before 15 years of age . Until 2008, a single dose of VZV vaccine was recommended; since then, two doses have been recommended . For HIV-infected children with normal CD4 cell counts, even before adolescence, immunization with VZV vaccine is recommended. However, this recommendation is mostly ignored. To determine whether the waning of anti-VZV antibodies in HIV-infected children resulted from impaired primary responses, accelerated antibody loss and/or failure to reactivate anti-VZV memory responses, we assessed anti-VZV IgG antibodies in sera prospectively collected over a 10-year period in HIV-infected children, compared with HIV-infected adults and age-matched noninfected healthy children. We also assessed the kinetics of anti-VZV antibodies over time, and measured their avidity, a useful marker of antigen-specific memory B cell maturation .
Blood samples from HIV-1-infected children were prospectively collected on a yearly basis between 1997 and 2008. All HIV-infected children of the Swiss MoCHiV cohort, in which almost all HIV-infected children in Switzerland are followed, were enrolled through six referral centres. Inclusion criteria were being HIV-positive, belonging to the MoCHiV cohort, and having at least two frozen serum samples ≥1 year apart. Exclusion criteria included age <1 year to avoid misinterpretation as a consequence of the presence of maternal antibodies, and serum samples drawn within 12 months of the administration of intravenous immunoglobulins. HIV-1-infected adults were enrolled in one centre. Medical history and demographic data, including birth date, gender, date of HIV diagnosis, age at sampling, yearly CD4 cell counts, yearly HIV RNA levels and treatment, were obtained from the cohort registry. Age- and gender-matched children undergoing minor elective surgery and without immunosuppression were recruited as healthy controls in one centre. They were distributed among four quartiles based on the age of the HIV-infected children (A1, <8.2 years; A2, 8.2–11.5 years; A3, 11.5–15.5 years; A4, >15.5 years).
Patients in the three groups (and/or their legal guardians) provided written consent for the use of these samples and their medical data. All data were analysed anonymously. Immunization against VZV was not recommended during the study period.
To identify risk factors for the waning of VZV antibodies, we compared initially VZV-positive HIV-infected children who had waning VZV antibodies with age-matched HIV-infected children who had protective VZV antibodies in all available samples.
This study was approved by the institutional Ethics Committee in all centres, and by the scientific boards of the Swiss HIV Cohort Study (SHCS) and MoCHiV.
All serum samples were obtained between January 1997 and October 2008. Measurement of anti-VZV IgG antibodies was performed in the Laboratoire de Vaccinologie (University Hospitals of Geneva) using an ‘in-house’ enzyme-linked immunosorbent assay (ELISA)  which compared favourably with the Virion® commercial kit (Virion Servion, Würzburg, Germany) (data not shown). To maximize the sensitivity of the assay, 96-well plates [Nunc Maxisorp (C), Nunc AS, Roskilde, Denmark] were coated with a lectin affinity purified VZV glycoprotein (East Coast Bio, North Berwick, ME, USA). Eight serial serum dilutions were incubated prior to the successive addition of biotin-conjugated goat anti-human IgG antibody (anti-human IgG biotin; Sigma, St Louis, MO), horseradish peroxidase streptavidin (HRP-streptavidin conjugate; Zymed, San Francisco, CA), and 2,2′-azino-bis-3-ethylbenzthiazoline-6-sulphonic acid (ABTS; Roche Diagnostics, Rotkreuz, Switzerland) substrate. Optical densities (ODs) were read at 405 nm and analysed by comparison to a standard curve included in each plate (SoftMaxPro software, version 5, Molecular Devices Inc, Sunnyvale, CA, USA). Results were compared with two reference sera: an National Institute for Biological Standards and Control (NIBSC) standard [World Health Organization (WHO) international standard; 50 IU/L] and a standard from Merck (Whitehouse Station, NJ, USA), calibrated in VZV glycoprotein (VZV-gp) units, previously used in vaccine efficacy studies . The cut-off of the assay (30 IU/L) was defined conservatively as the mean plus two standard deviations of 72 negative samples. Results below this cut-off were arbitrarily given a value of 15 IU/L. Including both standards in a large number of assays, we established that in our assay a titre of 5 VZV-gp units/mL (suggested as a putative protective threshold following immunization ) corresponded to 33.1 IU/L of the WHO international standard (not shown). We thus experimentally and conservatively defined 50 IU/L as the ‘minimal threshold’ for our assay. An anti-VZV booster response was experimentally defined as a >4-fold increase in anti-VZV IgG levels between two consecutive samples or a >2-fold increase resulting in an absolute increase of ≥1000 IU/L (not shown).
Antibody avidity increases during the maturation process of memory B cells, such that re-exposure to endogenous or exogenous antigen results in antibodies of higher avidity. Accordingly, antibody avidity is an indirect marker for the reactivation of memory responses . The avidity of anti-VZV antibodies was determined by adding various dilutions (0–3 M) of sodium thiocyanate to serum-containing antigen-coated wells, as previously described [16–18]. Results are expressed as the avidity index (AI), defined as the thiocyanate concentration at which 50% of the VZV-specific antibodies were eluted. As AI may fail to identify differences attributable to a small pool of high- or low-avidity antibodies, analyses were completed by calculating the percentage of antibodies dissociated at each thiocyanate concentration (AVISCAN) [19,20]. The Aviscan gives information about the distribution of different avidities within an antibody population of heterogenous avidities.
All P-values were two-tailed. P-values <0.05 were considered statistically significant. Continuous variables were assessed using parametric or nonparametric tests when appropriate, whereas categorical data were assessed using the χ2 or Fisher's exact test. Linear regression was used to analyse potential risk factors for low anti-VZV IgG levels and AI, whereas conditional logistic regression was used to identify potential risk factors for a complete loss of VZV antibodies. All variables were examined at the univariate level. Thereafter, only variables with a P-value <0.25 by univariate analysis were included in the multivariate model .
Change in anti-VZV IgG levels over time in HIV-infected children and adults were analysed using mixed linear models. This statistical model takes into account the repeated measurement of each individual across time. We included as predictors the group of patients (HIV-infected children or adults), the time of measure (linear trend) and the time of measure squared (quadratic trend) to account for a downward trend that could be faster for high VZV levels and slower for low levels. Finally, we adjusted for age, CD4 T-cell count and VZV serological reactivation.
Statistical analyses were performed using spss (v15.0; SPSS Inc., Chicago, IL), with the exception of longitudinal analyses, which were performed using the lme statistical package of the R software, v 2.9.2 .
Ninety-seven vertically HIV-infected children (541 samples) and 78 HIV-infected adults (440 samples) met the study inclusion criteria (Table 1). In 2008, the CD4 T-cell count and percentage (P<0.001 for both) and the HIV RNA level (P=0.007) were higher in HIV-infected children than adults. There were more female patients among the HIV-infected children than among the HIV-infected adults (P=0.002). More than 90% of the children had not been immunized against VZV: their anti-VZV IgG levels presumably resulted from wild-type infection. The median anti-VZV IgG titre was 264 IU/L [interquartile range (IQR) 747 IU/L] in HIV-infected children and 1535 IU/L (IQR 1600 IU/L; P<0.001) in the adults (Fig. 1), even after exclusion of VZV-seronegative, possibly unexposed individuals (P<0.001). Twenty-one per cent (20 of 97) of the HIV-infected children had undetectable VZV antibodies, compared with 3% (two of 78; P<0.001) of the adults.
Table 1. Demographics for HIV-infected adults and children, and HIV-negative children (2008)
HIV-positive adults (n=78)
HIV-positive children (n=97)
HIV-negative children (n=97)
ART, antiretroviral therapy; CDC, Centers for Disease Control and Prevention; HAART, highly active antiretroviral therapy; SD, standard deviation.
Female gender [% (n)]
Age (years) [mean (SD)]
Ethnic origin (% (n)]
Viraemia (copies/mL) [mean (SD)]
Patients with undetectable viraemia [% (n)]
CD4 T-cell percentage [mean (SD)]
CD4 T-cell count (cells/μL) [mean (SD)]
CD4 T-cell count category [% (n)]
CDC stage [% (n)]
ART category [% (n)]
Treated with HAART
Treated with ART only
Age at initiation of
ART (years) [mean (SD)]
HAART (years) [mean (SD)]
At baseline, differences in anti-VZV IgG level, HIV RNA level, CD4 cell count and CD4 percentage between HIV-infected children and adults were already significant (P<0.001, <0.001, <0.001 and 0.001, respectively) (data not shown). In this cross-sectional analysis, none of the following variables was predictive of lower anti-VZV IgG levels in HIV-infected children: age, gender, ethnicity, CD4 T-cell count and percentage, HIV RNA level, age at initiation of HAART, absence/presence of HAART and duration of HAART.
To determine whether anti-VZV antibodies declined more rapidly in HIV-infected children than adults, we assessed the change in antibody titres over time in all subjects who initially had negative VZV antibodies and then became positive following exposure (484 samples from 85 children and 435 samples from 77 adults). Twenty per cent (17 of 85) of previously VZV-positive children failed to maintain anti-VZV IgG levels >50 IU/L, compared with 2.6% (two of 77; P<0.001) of adults. The odds ratio for antibody waning in children, adjusted for the CD4 cell count, was 17.74 [P<0.001; 95% confidence interval (CI) 4.36–72.25]. These 17 HIV-infected children were compared with 54 randomly selected age-matched HIV-infected children who maintained anti-VZV IgG levels >50 IU/L throughout the study period. The two groups were comparable in terms of gender, age, CD4 T-cell count and duration of HAART. Univariate analyses demonstrated that higher HIV RNA level (P=0.001), absence of HAART (P=0.037) and lower CD4 percentage (P=0.027) were significantly associated with failure to maintain VZV antibodies. In the multivariate analysis, only higher HIV RNA level remained significant (P=0.011). Longitudinal analyses showed that the trend of anti-VZV IgG level over time was not significant in adults (Fig. 2). Anti-VZV IgG levels were lower in children at all time-points (P<0.001), but did not decline more rapidly than in adults and even slightly increased over time (P=0.01). This remained true after adjusting for age. Thus, the failure of 20% of HIV-infected children to maintain anti-VZV antibodies did not reflect a general pattern of antibody loss in HIV-infected children.
The lower anti-VZV IgG levels in HIV-infected children could result from weak primary anti-VZV responses. We thus compared the anti-VZV IgG levels of the 97 HIV-infected children with those of 97 gender- and age-matched healthy children (Table 1). The median anti-VZV IgG titre was lower in HIV-infected than healthy children (1151 IU/L; IQR 1535; P<0.001) (Fig. 1), even after exclusion of VZV-seronegative children (P<0.001). Anti-VZV antibodies were undetectable in only 5% (five of 97) of healthy children, compared with 21% (20 of 97) of HIV-infected children (P=0.001).
Anti-VZV antibody levels increased with age in healthy children (P=0.004) but not in HIV-infected children (Fig. 3). Accordingly, anti-VZV IgG levels were lower in HIV-infected children in all age quartiles except for A1. This difference persisted after exclusion of VZV-seronegative patients (data not shown). This suggested that weaker anti-VZV primary responses are elicited when VZV infection occurs in older HIV-infected children, or that anti-VZV IgG levels fail to increase with age in HIV-infected children.
To distinguish between the induction of weaker primary responses and the failure of secondary anti-VZV responses in HIV-infected children, we compared the avidity of anti-VZV antibodies in HIV-infected and healthy children. The mean AI of anti-VZV antibodies was lower in the 77 VZV-positive, HIV-infected children than in the 92 VZV-positive, healthy children (mean AI 2.12 ± 0.69 vs. 2.52 ± 0.67, respectively; P<0.001). This was true for all age quartiles (A1, P=0.078; A2, P=0.025; A3, P=0.003; A4, P=0.784). The proportion of low-avidity anti-VZV antibodies was higher in HIV-infected than in healthy children (28% vs. 21%, respectively; P<0.001), whereas that of high-avidity antibodies was lower in HIV-infected than in healthy children (29% vs. 37%, respectively; P<0.001). We identified no influence of age, gender, CD4 T-cell count or percentage, HIV RNA level, duration of HAART, or age at initiation of HAART on avidity.
A lower avidity of anti-VZV antibodies in HIV-infected than healthy children could result from limitations of the primary induction of high-affinity antibodies, as observed in HIV-infected infants , and/or from a less effective reactivation of VZV-specific memory B cells. We thus compared anti-VZV IgG levels and avidity in the first and last available serum samples of 63 HIV-infected children with two VZV-positive samples ≥1 year apart (median interval 4.08 years; range 1.17-9.42 years). The mean AI increased from 1.93 ± 0.58 to 2.14 ± 0.66 between the two series of samples (P=0.039). In 36 of 63 children (57%) with no evidence of serological booster responses, mean AI (first sample of 36/63 HIV-infected children without serological booster response: 1.93 vs. last sample of the same patients: 1.95; P=0.817) remained low, and it even declined in 12 of these 36 children (33%). Twenty-seven children had evidence of anti-VZV booster responses. This was associated with a significant increase in the anti-VZV AI (from mean 1.94 ± 0.64 to 2.39 ± 0.82; P=0.014) and a decline in the proportion of low-avidity antibodies (from 31% to 24%; P=0.006). Remarkably, this avidity maturation was only observed in 48% (13 of 27) of these children, who were younger (mean age 5.58 vs. 8.14 years, respectively; P=0.037) than the 14 patients with no anti-VZV avidity maturation.
In healthy children, we observed no correlation between anti-VZV IgG level and AI: some children maintained low levels of high-avidity antibodies, indicating successful avidity maturation. In contrast, a significant correlation between anti-VZV IgG level and AI was present in HIV-infected children (P=0.001): anti-VZV IgG levels were significantly lower in children with a lower AI, i.e. no evidence of successful memory B-cell maturation/reactivation. Thus, the waning of anti-VZV antibodies in a significant proportion of HIV-infected children resulted from the failure to maintain and/or reactivate anti-VZV memory responses.
This study showed that the waning of anti-VZV antibodies in HIV-infected children, compared with HIV-infected adults and healthy children, was associated with lower antibody avidity, reflecting the failure to generate, maintain or reactivate memory B-cell responses.
Rapid antibody decline was previously reported following immunization of HIV-infected patients . This may also affect humoral responses elicited by natural infection and results in absent or low antibody levels . The lower anti-VZV IgG levels were not explained by differences in age, gender, or ethnicity. A lower exposure rate to chickenpox is unlikely, as chickenpox is endemic, and HIV-infected patients have regular peer contact. HIV-infected children had higher CD4 T-cell counts than HIV-infected adults, as expected . The HIV RNA level was higher in children than in adults, because of lower HAART rates (88% vs. 99%) and suboptimally controlled infection [26,27]. Yet, HIV-infected children were almost 18 times more likely than adults to lose anti-VZV antibodies. Our longitudinal analysis indicated that high HIV RNA level, absence of HAART and low CD4 percentage were associated with the waning of VZV-specific antibodies.
Lower anti-VZV IgG levels were not attributable to a universally accelerated antibody loss: HIV-infected children had lower levels than adults throughout the 10-year study period and their antibody levels even increased slightly over time. These lower levels could reflect impaired primary responses [1,24]. However, anti-VZV IgG levels were lower in VZV-positive, HIV-infected children than in healthy children in all age quartiles except the youngest: this suggests that primary responses to VZV exposure were only impaired in older children, possibly as a result of HIV disease progression, and/or that some HIV-infected children failed to maintain/reactivate anti-VZV immunity.
To define whether the failure to reactivate anti-VZV memory responses may explain the lower anti-VZV IgG levels, we compared anti-VZV IgG levels in HIV-infected and healthy children. Anti-VZV IgG levels increased with age in healthy children but not in HIV-infected children, suggesting that endogenous reactivation and/or exogenous exposure led to the reactivation and differentiation of memory B cells into antibody-secreting-cells in healthy children more frequently or efficiently than in HIV-infected children. As avidity increases during the immune response and after re-exposure to an antigen [16,28–31], we next assessed the avidity of anti-VZV antibodies: the lower avidity of anti-VZV antibodies in HIV-infected than healthy children confirmed the impairment of their anti-VZV memory responses. This is in accordance with the recent observation that HIV-1 infection impairs the induction and avidity maturation of immunization-induced measles antibodies . How HIV infection impairs avidity maturation has not yet been elucidated. Although somatic mutation of immunoglobulin genes is a T-cell-dependent phenomenon, we observed no correlation between anti-VZV IgG level, avidity maturation and CD4 T-cell count. However, HIV has multiple direct effects on B-cell responses  and the percentage of memory B cells was even suggested as a marker of HIV disease progression . Lastly, HIV uptake by follicular dendritic cells affects germinal centres  in which affinity maturation is initiated.
Remarkably, anti-VZV IgG level and avidity correlated in HIV-infected children, in contrast to healthy children, in whom low concentrations of high-affinity antibodies were not rare. This indicates that healthy children maintain immune memory cells over a prolonged period, producing high-avidity antibodies even in the absence of boosting by antigen exposure, whereas immune memory only persists in HIV-infected children with high anti-VZV IgG levels. That these children with high anti-VZV IgG levels of high-avidity antibodies may have benefited from earlier/more frequent VZV exposure, thus reactivating and maintaining their memory B cells more efficiently, is an interesting possibility. In contrast, almost 25% of our HIV-infected children experienced a decline in anti-VZV antibody avidity over time, which was associated with a decline in their anti-VZV IgG levels. We couldn't identify predictors to explain why these patients had a different response. They had obviously not successfully maintained functional memory cells and therefore had to generate a ‘new primary response’ of low magnitude and avidity at the time of repeat exposure.
This study has some limitations. Precise information about chickenpox history was lacking: some children who lost their antibodies after exposure may have been considered “unexposed”, and we could not assess possible correlations between age at VZV infection and immune responses. Specific risk factors for the loss of anti-VZV immunity could have been missed, although we examined many factors commonly used as markers of HIV disease and management. Finally, we obtained a single sample from healthy children and could not therefore compare the kinetics of their anti-VZV IgG levels over time with that of HIV-infected patients.
This study has several clinical implications. Physicians caring for HIV-infected children should be aware that a history of chickenpox or VZV immunization does not provide lifelong humoral immunity [24,36], unlike in healthy children [8,24,36,37]. Cell-mediated immunity (CMI) may contribute to the persistence of protection and/or reduce disease severity even in the absence of antibodies [6,38]. However, CMI may remain appropriate [24,39,40] or be altered even in HAART-treated children [36,39], such that its contribution to protection may not be predicted for a given patient. As a consequence, it may be useful to obtain VZV serology at the time of exposure, especially in children with delayed and/or partly effective treatment and persistent HIV RNA levels – identified here as a determinant of antibody loss. As a consequence of our study design, we could not evaluate the risk of VZV disease recurrence in patients who lost anti-VZV humoral immunity nor determine whether booster VZV immunization reactivates immune memory cells. Finally, although VZV immunization is effective in HIV-infected children , its long-term efficacy should be repeatedly assessed through serologies as vaccine-induced responses are significantly weaker than those elicited by natural infection.
The Pediatric Infectious Diseases Group of Switzerland (PIGS): C. Aebi, W. Bär, Ch. Berger (Chair), F. Besson, U. Bühlmann, J.-J. Cheseaux, D. Desgrandchamps, A. Diana, A. Duppenthaler, A. Gervaix, H. P. Gnehm, U. Heininger, U.A. Hunzikerr, C. Kahlert, C. Kind, H. Kuchler, A. Loher, V. Masserey-Spicher, C. Myers, D. Nadal, K. Posfay-Barbe, C. Rudin, U. B. Schaad, C.-A. Siegrist, J. Stähelin, B. Vaudaux, C.-A. Wyler-Lazarevic and W. Zingg.
The Swiss HIV Cohort Study (SHCS) and the Swiss Mother & Child HIV Cohort Study (MoCHiV): C. Aebi, M. Battegay, E. Bernasconi, J. Böni, P. Brazzola, H. C. Bucher, Ph. Bürgisser, A. Calmy, S. Cattacin, M. Cavassini, J.-J. Cheseaux, G. Drack, R. Dubs, M. Egger, L. Elzi, M. Fischer, M. Flepp, A. Fontana, P. Francioli (President of the SHCS, Centre Hospitalier Universitaire Vaudois, Lausanne), H. J. Furrer, C. Fux, A. Gayet-Ageron, S. Gerber, M. Gorgievski, H. Günthard, Th. Gyr, H. Hirsch, B. Hirschel, I. Hösli, M. Hüsler, L. Kaiser, Ch. Kahlert, U. Karrer, C. Kind, Th. Klimkait, B. Ledergerber, G. Martinetti, B. Martinez, N. Müller, D. Nadal, F. Paccaud, G. Pantaleo, L. Raio, A. Rauch, S. Regenass, M. Rickenbach, C. Rudin (Chairman of the MoCHiV Substudy, Basel UKBB, Basel), P. Schmid, D. Schultze, J. Schüpbach, R. Speck, P. Taffé, A. Telenti, A. Trkola, P. Vernazza, R. Weber, C.-A. Wyler-Lazarevic and S. Yerly.
1Obaro SK, PugatchD, LuzuriagaK.Immunogenicity and efficacy of childhood vaccines in HIV-1-infected children. Lancet Infect Dis2004; 4: 510–518.
2Gershon AA, MervishN, LaRussaPet al. Varicella-zoster virus infection in children with underlying human immunodeficiency virus infection. J Infect Dis1997; 176: 1496–1500.
3Srugo I, IsraeleV, WittekAE, CourvilleT, VimalVM, BrunellPA.Clinical manifestations of varicella-zoster virus infections in human immunodeficiency virus-infected children. Am J Dis Child1993; 147: 742–745.
4von Seidlein L, GilletteSG, BrysonYet al. Frequent recurrence and persistence of varicella-zoster virus infections in children infected with human immunodeficiency virus type 1. J Pediatr1996; 128: 52–57.
5Liese JG, GroteV, RosenfeldE, FischerR, BelohradskyBH, v KriesR.The burden of varicella complications before the introduction of routine varicella vaccination in Germany. Pediatr Infect Dis J2008; 27: 119–124.
6Wood SM, ShahSS, SteenhoffAP, RutsteinRM.Primary varicella and herpes zoster among HIV-infected children from 1989 to 2006. Pediatrics2008; 121: e150–e156.
7Bekker V, ScherpbierH, PajkrtD, JurriaansS, ZaaijerH, KuijpersTW.Persistent humoral immune defect in highly active antiretroviral therapy-treated children with HIV-1 infection: loss of specific antibodies against attenuated vaccine strains and natural viral infection. Pediatrics 2006; 118: e315–e322.
8Amanna IJ, CarlsonNE, SlifkaMK.Duration of humoral immunity to common viral and vaccine antigens. N Engl J Med2007; 357: 1903–1915.
9Edmunds WJ, BrissonM.The effect of vaccination on the epidemiology of varicella zoster virus. J Infect2002; 44: 211–219.
10Heininger U, Braun-FahrlanderC, DesgrandchampsDet al. Seroprevalence of varicella-zoster virus immunoglobulin G antibodies in Swiss adolescents and risk factor analysis for seronegativity. Pediatr Infect Dis J2001; 20: 775–778.
11 Office fédéral de la santé publique. Commission fédérale pour la vaccination. Plan de vaccination suisse 2008. Directives et recommandations (précédemment Supplement VIII). Bulletin OFSP 2008:1–20.
12Shivarov V, ShinkuraR, DoiTet al. Molecular mechanism for generation of antibody memory. Philos Trans R Soc Lond B Biol Sci2009; 364: 569–575.
13Posfay-Barbe K, SottasC, DelhumeauCet al. New assays to quantify varicella immunity in pediatric liver transplant recipients. 48th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC)/IDSA. Washington, DC, October 2008 [Abstract G1-429].
14Li S, ChanIS, MatthewsHet al. Inverse relationship between six week postvaccination varicella antibody response to vaccine and likelihood of long term breakthrough infection. Pediatr Infect Dis J2002; 21: 337–342.
16Goldblatt D, VazAR, MillerE.Antibody avidity as a surrogate marker of successful priming by Haemophilus influenzae type b conjugate vaccines following infant immunization. J Infect Dis1998; 177: 1112–1115.
17Macdonald RA, HoskingCS, JonesCL.The measurement of relative antibody affinity by ELISA using thiocyanate elution. J Immunol Methods1988; 106: 191–194.
18Pullen GR, FitzgeraldMG, HoskingCS.Antibody avidity determination by ELISA using thiocyanate elution. J Immunol Methods1986; 86: 83–87.
19Schallert N, PihlgrenM, KovarikJet al. Generation of adult-like antibody avidity profiles after early-life immunization with protein vaccines. Eur J Immunol2002; 32: 752–760.
20Siegrist CA, PihlgrenM, TougneCet al. Co-administration of CpG oligonucleotides enhances the late affinity maturation process of human anti-hepatitis B vaccine response. Vaccine2004; 23: 615–622.
21Hosmer DW, LemeshowS.Applied Logistic Regression. New York: Wiley, 2000.
22Team RDC. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2009. Available at www.R-project.org.
23Madhi SA, KuwandaL, SaarinenLet al. Immunogenicity and effectiveness of Haemophilus influenzae type b conjugate vaccine in HIV infected and uninfected African children. Vaccine2005; 23: 5517–5525.
24Bekker V, WesterlakenGH, ScherpbierHet al. Varicella vaccination in HIV-1-infected children after immune reconstitution. AIDS2006; 20: 2321–2329.
25Denny T, YogevR, GelmanRet al. Lymphocyte subsets in healthy children during the first 5 years of life. JAMA1992; 267: 1484–1488.
26Mellors JW, MunozA, GiorgiJVet al. Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection. Ann Intern Med1997; 126: 946–954.
27Mellors JW, RinaldoCRJr, GuptaP, WhiteRM, ToddJA, KingsleyLA.Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science1996; 272: 1167–1170.
28Kangro HO, ManzoorS, HarperDR.Antibody avidity following varicella-zoster virus infections. J Med Virol1991; 33: 100–105.
29Goldblatt D, RichmondP, MillardE, ThorntonC, MillerE.The induction of immunologic memory after vaccination with Haemophilus influenzae type b conjugate and acellular pertussis-containing diphtheria, tetanus, and pertussis vaccine combination. J Infect Dis1999; 180: 538–541.
30Joseph H, MillerE, DawsonM, AndrewsN, FeaversI, BorrowR.Meningococcal serogroup a avidity indices as a surrogate marker of priming for the induction of immunologic memory after vaccination with a meningococcal A/C conjugate vaccine in infants in the United Kingdom. J Infect Dis2001; 184: 661–662.
31Richmond P, BorrowR, GoldblattDet al. Ability of 3 different meningococcal C conjugate vaccines to induce immunologic memory after a single dose in UK toddlers. J Infect Dis2001; 183: 160–163.
32Nair N, MossWJ, ScottSet al. HIV-1 infection in Zambian children impairs the development and avidity maturation of measles virus-specific immunoglobulin G after vaccination and infection. J Infect Dis2009; 200: 1031–1038.
33Moir S, FauciAS.B cells in HIV infection and disease. Nat Rev Immunol2009; 9: 235–245.
34Titanji K, De MilitoA, CagigiAet al. Loss of memory B cells impairs maintenance of long-term serologic memory during HIV-1 infection. Blood2006; 108: 1580–1587.
35Taruishi M, TerashimaK, DewanZet al. Role of follicular dendritic cells in the early HIV-1 infection: in vitro model without specific antibody. Microbiol Immunol 2004; 48: 693–702.
36Levin MJ, GershonAA, WeinbergAet al. Immunization of HIV-infected children with varicella vaccine. J Pediatr2001; 139: 305–310.
37Wutzler P, FarberI, WagenpfeilS, BisanzH, TischerA.Seroprevalence of varicella-zoster virus in the German population. Vaccine2001; 20: 121–124.
38Ampofo K, SaimanL, LaRussaP, SteinbergS, AnnunziatoP, GershonA.Persistence of immunity to live attenuated varicella vaccine in healthy adults. Clin Infect Dis2002; 34: 774–779.
39Weinberg A, WizniaAA, LaFleurBJ, ShahS, LevinMJ.Varicella-Zoster virus-specific cell-mediated immunity in HIV-infected children receiving highly active antiretroviral therapy. J Infect Dis2004; 190: 267–270.
40Weinberg A, LazarAA, ZerbeGOet al. Influence of age and nature of primary infection on varicella-zoster virus-specific cell-mediated immune responses. J Infect Dis2010; 201: 1024–1030.
41Son M, ShapiroED, LaRussaPet al. Effectiveness of varicella vaccine in children infected with HIV. J Infect Dis2010; 201: 1806–1810.