ObjectiveTo explore the safety and effectiveness of graded infusion of bone cement in the unipedicular percutaneous vertebroplasty (PVP) for Kummell's disease. MethodsEighteen patients with Kummell's disease were treated by unipedicular PVP with graded infusion of bone cement between January 2012 and January 2014. Of 18 cases, 6 were male and 12 were female, aged from 65 to 88 years (mean, 75 years), with a disease duration from 3 to 32 months (mean, 11.6 months). The bone mineral density was measured by dual-energy X-ray absorptiometry; the T value ranged from -4.0 to -2.8 (mean, -3.4). Affected segments included T11 in 3 cases, T12 in 10 cases, L1 in 3 cases, and L2 in 2 cases. X-ray films were taken after operation to observe bone cement leakage and anterior height changes of affected vertebrae. Visual analogue scale (VAS) and Oswestry disability index (ODI) were used to assess pain status and functional activity. ResultsAll cases underwent smoothly unipedicular PVP and were followed up 12-26 months (mean, 14 months). Cement leakage occurred in 4 patients, including 1 case of anterior paravertebral soft tissue leakage, 2 cases of intervertebral disc leakage, and 1 case of canal venous leakage, but there was no other complications. The anterior height of affected vertebrae were significantly improved (P<0.05) from 29.1%±6.7% at preoperation to 68.1%±7.3% at 3 days after operation and 67.8%±5.9% at last follow-up; the VAS scores were significantly decreased (P<0.05) from 8.11±1.32 at preoperation to 2.14±0.78 at 3 days and 1.97± 0.50 at last follow-up; and ODI were significantly decreased (P<0.05) from 84.6%±8.5% to 24.1%±9.7% and 23.8%±10.2%; but no significant difference was found between at 3 days and at last follow-up (P>0.05). ConclusionGraded infusion of bone cement in unipedicular PVP is a safe and effective procedure for Kummell's disease and this technique could decrease the incidence of bone cement leakage.