High prevalence of undiagnosed vertebral fractures in patients suffering from hip fracture at their hospital admission: weak concordance among observers.

Abstract

Vertebral fracture is often underdiagnosed. Patients with hip fracture may suffer from vertebral fracture without knowing it. The diagnosis of vertebral fracture is sometimes difficult because there is no consensus regarding the definition of osteoporotic vertebral fracture, and several indexes may be used to diagnose it and the concordance between several observers may not be optimal. To study the concordance in the diagnosis of vertebral fracture done by three different doctors: an orthopedic surgeon, a radiologist, and a bone mineral metabolism expert. A lateral thoracic-lumbar spine X-Ray was performed in 177 patients suffering from hip fracture to assess the presence or absence of vertebral fractures. Three different observers applied Genant's criteria for it. Concordance between observers was measured using Cohen's kappa coefficient. Patients suffering from hip fractures have undiagnosed vertebral fractures in a range that varies from 41.8 to 47.5% depending on the observer. The concordance in the diagnosis of vertebral fractures is quite low, ranging a Cohen's kappa coefficient from 0.43 to 0.55 and a percentage of concordance varying from 64 to 72%. The best concordance was found between observers 1 and 3. Depending on the observer who made the diagnosis, the prevalence of previously undiagnosed vertebral fractures in patients with HF varied widely. We selected three different observers to assess the possible differences in the diagnosis of vertebral fractures among these patients and using the same method (Genant's semi-quantitative assessment), surprisingly, there was little concordance among the three of them. Patients with hip fracture have high prevalence of undiagnosed vertebral fractures. The diagnosis of these fractures varies widely depending on the observers and the Cohen's kappa coefficient and percentage of concordance is rather low.