BackgroundKnee arthroplasty is a well-established surgery with good clinical outcomes. However, periprosthetic fractures and aseptic loosening negatively impact clinical outcomes, and osteoporosis is one of the causes of such complication. Osteoporosis is usually evaluated by bone mineral density of the lumbar spine and hip using dual-energy X-ray absorptiometry (DXA). However, the prevalence of this disease in patients with knee osteoarthritis scheduled for knee arthroplasty may be underestimated due to differences in the measurement sites. This study aimed to determine the appropriate measurement site for DXA that would not miss osteoporosis in female patients with knee osteoarthritis undergoing knee arthroplasty. MethodsWe measured bone mineral density preoperatively in the consecutive 50 female patients with knee osteoarthritis scheduled for knee arthroplasty by dual-energy X-ray absorptiometry at five sites: the lumbar spine, bilateral-total hip, and femoral neck. We then compared the prevalence of osteoporosis among the four combinations of the lumbar spine and single hip site (ipsilateral or contralateral total hip or femoral neck). ResultsOsteoporosis prevalence in the combination of the lumbar spine and ipsilateral or contralateral total hip was 32%, and that in the combination of the lumbar spine and contralateral femoral neck was 44%. Notably, the disease's prevalence in the combination of the lumbar spine and ipsilateral femoral neck was 50%, which was significantly higher than that in the other combinations. ConclusionOsteoporosis should be evaluated by bone mineral density in the combination of the lumbar spine and ipsilateral femoral neck in female patients with knee osteoarthritis scheduled for knee arthroplasty.