Exercise during growth has a positive influence on bone mineral accrual, yet little is known about how bone geometry and strength adapt to loading during growth. Our primary objective was to compare changes in proximal femur bone geometry and strength between 31 prepubertal (Tanner Stage 1) boys who participated in a school-based, high-impact circuit intervention (12 min, three times a week) for 20 months and 33 maturity-matched controls. Our secondary objective was to compare changes in total body (TB), proximal femur (PF), and lumbar spine (LS) bone mineral content (BMC) and bone area (BA) in these groups. We assessed geometric variables and bone strength at the narrow neck (NN), intertrochanteric (TR) region, and femoral shaft regions by applying the Hip Structure Analysis program to proximal femur dual energy X-ray absorptiometry scans (DXA, Hologic QDR 4500). Further, we assessed total body, lumbar spine, and proximal femur BMC and BA by DXA and derived total body lean mass and fat mass from total body scans. Intervention (10.2 +/- 0.5 years) and control boys (10.1 +/- 0.5 years) had similar baseline height (140.8 vs. 141.3 cm) and weight (36.9 vs. 35.4 kg), and average 20-month physical activity scores (Physical Activity Questionnaire for Children, PAQ-C) and calcium intakes (861 vs. 852 mg/day, food frequency questionnaire). Twenty-month height and weight changes were not significantly different between groups; lean mass changed more (P < 0.05) in intervention boys (22.8%) than control boys (18.6%). At the NN region, intervention boys had greater bone expansion on both the periosteal (+2.6%, P = 0.1) and endosteal (+2.7%, P = 0.2) surfaces, resulting in significantly greater changes in section modulus (bone bending strength) (+7.5%, P = 0.02, ANCOVA, adjusting for height change, final Tanner Stage, and baseline bone values). Changes at the intertrochanteric and femoral shaft regions were not significantly different between groups. Femoral neck (FN) BMC changes were significantly greater in intervention boys (+4.3%, P < 0.01); changes in BA and BMC for other regions were not significantly different between groups. In summary, a school-based, high-impact exercise intervention implemented three times a week for 12 min is an effective strategy for site-specific gains in bone strength at the narrow neck region of the proximal femur.