Background:  HIV-infected children (HIV⁺) may be at risk for poor bone mineral accrual, but the etiological factors are ill-defined. We sought to characterize bone mineral density among HIV⁺ children enrolled in the US-based PHACS Adolescent Master Protocol (AMP) and to evaluate associations with ART as well as virologic and nutritional factors.

Methods: Total body and lumbar spine bone mineral density were measured by dual energy X-ray absorptiometry (DEXA) in 190 HIV⁺ children. Clinical (HIV stage and viral load, CD4 percentage, current ART), and anthropometric data were obtained simultaneously. Age- and sex-adjusted z-scores were calculated for height, body mass index, and bone mineral density (adjusted by bone age if <2 SD from chronological age) using population standards. Children with low bone mineral density (total body or lumbar spine z-score <–1.5), had calcium, parathyroid hormone, bone-specific alkaline phosphatase, and 25-OH vitamin D measurements taken. Multiple linear regression models were fit to identify predictors of lumbar spine z-scores, adjusted for race, height z, Tanner stage, and HIV viral load (<400 vs ≥400 copies/mL).

Results:  Mean age of the cohort was 12.3 years (range 7 to 16 years) with 45% male. The majority was African American (76%). Compared to population norms, the cohort was shorter (z = –0.374, p <0.0001), with greater body mass index (z = 0.300, p = 0.0002), and lower total body bone mineral density (z = –0.220, p = 0.015). Lumbar spine bone mineral density did not differ from normal (z = –0.053, p = 0.56). In an adjusted analysis, children on NNRTI therapy had higher lumbar spine z-scores (estimate 0.68, p = 0.0027) and those on PI therapy had lower scores (–0.49, p = 0.052). There was no association with NRTI therapy (estimate 0.55, p = 0.19). Children with viral load <400 copies/mL had lower lumbar spine scores (estimate –0.57, p = 0.003). The percentage of children with z <–1.5 on SP, TB, or either measure was 10.6%, 16.7%, and 20.0%, respectively. Of those with low lumbar spine or total body, 27 had triggered follow-up laboratory tests, 67% of which had low serum 25-OH D (<20 ng/mL). All 27 had normal calcium. Children with lower 25-OH D had higher parathyroid hormone compared to those with normal 25-OH D (mean 44 vs 30 pg/mL, p = 0.05), while there was no difference in bone-specific alkaline phosphatase.

Conclusions:  Low bone mineral density among HIV⁺ children is common. Limited data suggest that vitamin D deficiency may be an important contributing factor and should continue to be evaluated. NNRTI therapy may be protective and PI detrimental to bone health.