近端股骨Hounsfield单位预测十年骨折风险
本研究基于2004至2007年接受CT结肠成像(CTC)的40岁以上患者,测量近端股骨Hounsfield单位(HU)并随访至少10年,发现骨折患者髋部HU更低;以112 HU为阈值(对应80%灵敏度)可区分骨折风险,其阴性预测值超过97%。
要点
- CTC测量的近端股骨HU可机会性筛查骨质疏松相关骨折风险。
- 112 HU阈值对应80%灵敏度,整体骨折和髋部骨折阴性预测值均超过97%。
- 低于112 HU者10年整体骨折和髋部骨折发生率均显著高于高于阈值者。
核心结论
- CTC后10年内发生骨折的患者其髋部HU较低,HU下降可能是骨折风险的早期预警信号。
研究设计与样本
- 回顾性队列研究,纳入2004至2007年接受CTC且年龄40岁以上、随访至少10年的患者3711例,测量近端股骨HU并通过全球军事记录识别骨折。
主要发现
- 共183例骨折,骨折组近端股骨HU低于无骨折组。
- 112 HU阈值对应80%灵敏度,整体骨折和髋部骨折的阴性预测值均超过97%。
- 低于112 HU者10年整体骨折发生率为6.3%,高于112 HU者的1.7%。
- 低于112 HU者10年髋部骨折发生率为2.7%,高于112 HU者的0.07%。
- 112 HU阈值对应整体骨折的比值比为2.5(95%CI 1.7-3.6),髋部骨折为24.5(95%CI 3.3-175.5)。
关键数值与推荐
- 以112 HU为阈值,对应80%灵敏度筛查骨折风险。
- 该阈值下整体骨折和髋部骨折的阴性预测值均超过97%。
- 低于112 HU者10年整体骨折发生率为6.3%,髋部骨折为2.7%。
- 高于112 HU者10年整体骨折发生率为1.7%,髋部骨折为0.07%。
临床怎么用
- 可利用CTC机会性测量近端股骨HU,对低于112 HU的患者提示骨折风险并考虑进一步评估。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
UNLABELLED: Colon cancer screening occurs at younger ages than osteoporosis screening. Bone density measurements using virtual colonoscopy performed for colon cancer screening can provide an early warning sign of patients at potential risk for osteoporosis-related fractures. Earlier identification may improve treatment and potentially fracture prevention. INTRODUCTION: Opportunistic osteoporosis screening with computed tomography colonography (CTC) offers an opportunity to capitalize on earlier colorectal cancer screening to identify patients at risk of future fractures. The purpose of this study is to evaluate 10-year fracture and specifically hip fracture risk based on Hounsfield units (HU) obtained from CTC. METHODS: We identified all CTC scans between 2004 and 2007 of patients 40 years and older with 10 years minimum follow-up. Hounsfield units were measured within the proximal femur and fractures identified via worldwide military records. Patients were stratified into two cohorts based on the presence or lack of a fracture in the wrist, spine, hip, or proximal humerus. Hounsfield unit measurements were compared between groups using Student's t test and the HU threshold was calculated that best approximated an 80% sensitivity to optimally screen patients for fracture risk. The odds ratio, negative predictive value, 10-year incidence of fracture, and survival curves were calculated. RESULTS: We identified 3711 patients with 183 fractures over 10 years. The HU threshold that corresponded with an 80% sensitivity to identify fractures was 112 HU. The negative predictive value (NPV) for overall fractures and hip fractures was over 97%. The 10-year fracture incidence was higher in patients below 112 HU compared to those above for both overall fractures (6.3% vs 1.7%) and hip fractures (2.7% vs 0.07%). The 112 HU threshold corresponds with an odds ratio for overall fracture and hip fractures of 2.5 (95% confidence interval (CI), 1.7-3.6) and 24.5 (95% CI, 3.3-175.5), respectively. CONCLUSION: In the 10 years following CTC, patients who experienced a fracture had lower hip HU. Decreasing HU on CTC may be an early warning sign of fracture potential.
引用
- Christensen DL, Nappo KE, Wolfe JA, et al. Ten-year fracture risk predicted by proximal femur Hounsfield units. Osteoporos Int. 2020. PMID: 32594205. doi:10.1007/s00198-020-05477-y
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。