Systematic review of interventions for low bone mineral density in children with cerebral palsy.

AIM

To systematically review the efficacy of interventions to improve low bone mineral density (LBMD) in children and adolescents with cerebral palsy (CP).

METHODS

We performed a systematic search for published randomized, controlled trials (RCTs) and controlled clinical trials (CCTs) of children with CP (aged 0-20 years, all Gross Motor Function Classification System [GMFCS] levels) who received various medical and physical interventions for LBMD compared with no intervention or standard care. Study validity was evaluated by using the Physiotherapy Evidence Database (PEDro) scale. Standardized mean differences (SMDs) were calculated when data were sufficient.

RESULTS

Eight of 2034 articles met the inclusion criteria (6 RCTs, 2 CCTs). There were 3 trials of bisphosphonates, one of which (Henderson RC, Lark RK, Kecskemethy HH, Miller F, Harcke HT, Bachrach SJ. J Pediatr. 2002;141[5]:644-651) revealed a large and significant effect on BMD in 1 of 3 sites in the distal femur (SMD: 1.88 [95% confidence interval (CI): 0.52-3.24]). There were 3 trials of weight-bearing through varying approaches, one of which (Caulton JM, Ward KA, Alsop CW, Dunn G, Adams JE, Mughal MZ. Arch Dis Child. 2004;89[2]:131-135) showed a large and significant effect on the lumbar spine when increasing static standing time (SMD: 1.03 [95% CI: 0.21-1.85]). The administration of vitamin D and calcium produced a large, nonsignificant effect on BMD in the lumbar spine (Jekovec-Vrhovsek M, Kocijancic A, Prezelj J. Dev Med Child Neurol. 2000;42[6]:403-405) (SMD: 0.88 [95% CI: -0.07 to 1.84). Growth hormone combined with vitamin D and/or calcium resulted in effects comparable with vitamin D and/or calcium on BMD in the lumbar spine (Ali O, Shim M, Fowler E, et al. J Clin Endocrinol Metab. 2007;92[3]:932-937) (SMD 0 [95% CI: -1.24 to 1.24]).

CONCLUSIONS

Important effects on LBMD have been observed in small and diverse cohorts of children with CP. It is unclear whether small sample sizes or variable treatment responses account for nonsignificant findings. Additional large RCTs are needed of both physical and medical approaches.