Pubertal Onset in Children With Perinatal HIV Infection in the Era of Combination Antiretroviral Treatment

Paige L. Williams; Mark J. Abzug; Denise L. Jacobson; Jiajia Wang; Russell B. Van Dyke; Rohan Hazra; Kunjal Patel; Linda A. Dimeglio; Elizabeth J. Mcfarland; Margarita Silio; William Borkowsky; George R. Seage III; James M. Oleske; Mitchell E. Geffner

Disclosures

AIDS. 2013;27(12):1959–1970 

In This Article

Results

Characteristics of the Study Population

We evaluated 3006 children from 219C or AMP who were born to HIV-infected women and who had pubertal staging at age 7 years or older. We excluded 467 youth who had pubertal onset prior to study entry and were at Tanner stage 3 or higher at their first visit, leaving 2539 in analyses of pubertal onset (1253 girls and 1286 boys). Demographic characteristics are shown in Table 1 by HIV infection status; 56% were Black non-Hispanic and 29% were Hispanic.

The 2086 PHIV youth more often were born in earlier years than the 453 HEU children and had lower mean Z-scores for both BMI and height at the time of their first pubertal assessment. Among HIV-infected youth, 12% had CD4% below 15%, and 27% had viral load above 10 000 copies/ml proximate to the first pubertal assessment (Table 1). Males more often had low CD4% and high viral load than females (Supplemental Digital Content 1, http://links.lww.com/QAD/A339). At the time of the first pubertal assessment, 64% were on combination antiretroviral therapy (55% including a protease inhibitor and 9% without a protease inhibitor); the median duration of prior combination treatment among those with any prior use was 3.24 years. Of the 752 children not on combination treatment, 82% received mono-agent or dual-agent therapy, 6% were on three or more NRTIs, and 12% were not on any antiretroviral treatment.

Age at Pubertal Onset by HIV Status, Race/Ethnicity, and Birth Cohort

HIV-exposed but uninfected girls had pubertal onset at mean ages of 9.6 and 10.0 years according to breast and pubic hair stages, respectively, whereas boys had onset at mean ages of 10.4 and 10.7 years based on genitalia and pubic hair, respectively (Table 2). Consistent with most studies in girls, thelarche (reflected by breast staging) occurred earlier on average than pubarche (reflected by pubic hair growth).[40–43] Mean ages at pubertal onset by race/ethnicity were similar for HEU youth to those based on NHANES III data for girls,[40–43] but were later than NHANES III among both Black non-Hispanic (10.0 vs. 9.2–9.5 years) and white non-Hispanic boys (11.6 vs. 10.0 years).[3–5,40]

Compared to the HEU youth, PHIV children had significantly later estimated mean ages at onset, representing delays of about 6–11 months (Table 2). The mean age at pubertal onset for PHIV as compared to HEU was 10.3 vs. 9.6 years, 10.5 vs. 10.0 years, 11.3 vs. 10.4 years, and 11.5 vs. 10.7 years according to female breast, female pubic hair, male genitalia, and male pubic hair staging, respectively (all P < 0.001). Estimated mean ages by race/ethnicity indicated significantly earlier pubertal onset for black non-Hispanic youth (by 9–12 months) and for Hispanics (by 3–4 months) than for white non-Hispanic youth, consistent with previous studies.[3–5,40–43] Striking trends for earlier pubertal onset with more recent birth year were observed in both PHIV and HEU youth; however, the mean ages at onset were typically later for PHIV than HEU within each birth cohort (Fig. 1). In interval-censored models, this shift was attenuated after adjustment for race/ethnicity and birth cohort, with 4–6 months later onset for PHIV youth on average (Table 3) according to breast, genitalia, and male pubic hair staging measures. Higher Z-scores for prepubertal height and BMI were associated with significantly earlier pubertal onset, but further adjustment for these measures had little effect on differences by HIV infection status. In sensitivity analyses excluding youth at Tanner stage 2 at their first pubertal assessment (10% HEU, 11% PHIV), estimated mean ages at onset were slightly later (~2 months), but differences by HIV infection status were similar to those reported in Table 3.

Figure 1.

Mean age at pubertal onset by Tanner measure and birth cohort, for HIV-infected youth (unshaded bars) and HIV-exposed uninfected youth (shaded bars). P-values for statistical significance based on interval-censored models for the effect of HIV infection status fit separately within each birth cohort.

The percentage of PHIV youth with delay in pubertal onset was 4.1% overall, but showed substantial decreases over time (11.2, 3.1, 1.7, and 0.4% for those born before 1990, 1990–1992, 1993–1996, and 1997 or later, respectively; trend test P-value <0.001). In contrast, delay in pubertal onset was rare among HEU youth (2/453, 0.4%).

Association of HIV Disease Severity With Age at Pubertal Onset

Mean ages at pubertal onset were significantly later for youth with more advanced HIV disease status at the first pubertal assessment (Table 3). For both girls and boys, there was a significant association of low CD4 (as reflected by CD4% <15% and CD4 cell count <200 cells/μl) and high viral load (>10 000 copies/ml) with later pubertal onset, both with and without adjustment for race/ethnicity and birth cohort, with the exception of female pubic hair (Table 3). Associations remained significant for most Tanner measures after further adjustment for BMI and height Z-scores (Supplemental Digital Content 2, http://links.lww.com/QAD/A340). Measures of past HIV disease severity showed stronger associations with timing of pubertal onset in boys than in girls. Boys with CDC class C (prior AIDS-defining condition), low nadir CD4%, or higher peak viral load had significantly later pubertal onset than those with milder classifications. Cumulative viral burden reflected by copy-years viremia was associated with significantly later age at pubertal onset in boys after adjustment for race/ethnicity and birth cohort, but not in girls. No association with peak viral load or CDC class was observed among girls.

Association of Combination Treatment and Antiretroviral Drug Classes With Age at Pubertal Onset

In unadjusted models, youth on combination treatment with a protease inhibitor at the first pubertal assessment had a significantly earlier mean age at pubertal onset (by 2.7–3.7 months) than did youth not on combination treatment, with the exception of male genitalia (Table 3). In contrast, there was no significant difference in age at onset for those on combination regimens without protease inhibitors as compared to those youth unexposed to combination regimens for any Tanner measure. Associations for combination treatment with protease inhibitor were nonsignificant in females and reversed direction in males when further adjusted for birth cohort. Examination of differences by birth cohort revealed that mean ages at pubertal onset were earlier for those exposed to combination treatment among those born since 1997 for all measures except male genitalia, but later for those exposed to combination treatment among those born prior to 1990 for all four staging measures (Fig. 2); however, statistical tests for interaction were not significant.

Figure 2.

Mean age at pubertal onset by Tanner measure and birth cohort, for HIV-infected youth on combination antiretroviral treatment (cART, unshaded bars) as compared to those not on cART (non-cART, shaded bars) at the time of the first pubertal assessment. P-values for statistical significance based on interval-censored models for the effect of cART fit separately within each birth cohort.

Duration of prior combination treatment, overall or with a protease inhibitor, at the time of the first pubertal assessment also showed significant associations with earlier age at pubertal onset based on all four staging measures in unadjusted analyses (Table 3). Mean ages at onset ranged from 0.6 to 1.2 months earlier for each additional year on combination treatment, depending on staging measure. However, after adjustment for race/ethnicity and birth cohort, there was no association of duration of combination treatment (overall or with protease inhibitor) in girls, and the direction of effect reversed for boys. Similar associations were observed for prior years of use of protease inhibitors and NNRTIs (Table 3). Further adjustment for BMI and height Z-scores and for HIV-disease status as reflected by CD4% below 15% had little effect on estimated mean ages of pubertal onset (Supplemental Digital Content 2, http://links.lww.com/QAD/A340).

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....