Objective To explore the clinical effect of bone grafting in the injured vertebrae combined with pedicle screw-rod system assisted with drugs in treating lumbar vertebral compression fracture in the elderly.Methods A retrospective case-control study was carried out to analyze the clinical data of 48 cases of lumbar vertebral compression fracture in the elderly admitted to the Union Hospital between October 2015 and October 2018.There were 34 males and 14 females,aged 60-78 years [(62.8 ± 2.5) years].Segment of injury was L1 in 37 cases,L2 in 7,L3 in 2,and L4 in 4.There were 20 cases in bone grafting group involving 15 males and five females,aged from 60 to 78 years [(63.7 ± 2.1)years].There were 28 cases in non-bone grafting group involving 19 males and nine females,aged from 60 to 75 years [(62.4 ± 2.9) years].The Frankel scale in bone grafting group was grade D in 15 cases and grade E in 5,and in non-bone grafting group was grade D in 18 cases and grade E in 10.All the cases were treated by pedicle screw fixation combined with antiosteoporosis drugs,and the bone grafting group were treated by bone grafting via the injured vertebrae,but the other group were not.The intraoperative blood loss and operation time for each segmental vertebrae were recorded.The visual analogue scale (VAS) before operation and one year after operation,the bone healing at three months and one year postoperatively were recorded.The compression rate of the injured vertebrae at operation,at one day and one-year postoperatively as well as the height loss of the injured vertebrae at postoperative one year were measured.Spinal cord injury was also evaluated by Frankel scale.Wound healing,lower limb thrombosis,lung infection and ulcer were observed.Results All the cases were followed up,with duration for 12-25 months [(16.2 ±3.4) months] in bone grafting group and 15-24 months [(17.5 ± 5.4) months] in non-bone grafting group (P ＞ 0.05).The blood loss was (240 ± 70) ml in bone grafting group and (210 ± 65) ml in non-bone grafting group (P ＞ 0.05).Screw implantation time for each segment was (38.5 ±9.6)minutes in bone grafting group and (30.5 ±5.4)minutes in non-bone grafting group (P ＜ 0.05).The VAS was (7.5 ± 1.8)points preoperatively and (1.5 ± 0.7)points at one year postoperatively in bone grafting group,while (7.8 ± 1.4) points and (2.9 ± 1.2) points in non-bone grafting group,with significant difference between the two groups at one year postoperatively (P ＜ 0.05).The fracture healing rate was 85％ (17/20) and 95％ (19/20) at three months and one year postoperatively in bone grafting group,while 64％ (18/28) and 86％ (24/28) in non-bone grafting group,with no statistical difference between the two groups (P ＞ 0.05).No significant difference was found in compression rate of the injured vertebral height between the two groups before and after operation (P ＞ 0.05).The height loss rate of the injured vertebrae was 3％ (2％,4％) at one year postoperatively in bone grafting group and 6％ (5％,8％) in non-bone grafting group (P ＜0.05).The Frankel scale was all Grade E postoperatively.During the follow-up,no lower limb thrombosis,lung infection or ulcer occurred.One case had bone non-union in bone grafting group and four in non-bone grafting group (P ＞ 0.05).Conclusion For lumbar vertebral compression fracture in the elderly,bone grafting in the injured vertebrae combined with pedicle screw-rod system assisted with drugs is significantly better in recovering and maintaining the injured vertebrae height,relieving the pain and promoting the bone healing,although the screw implantation time is prolonged.