The reliability of diagnosing osteoporosis from spinal radiographs

Age and Ageing, July, 1994 by A.C. Scane, T. Masud, F.J. Johnson, R.M. Francis

Summary

In order to assess the value of spine radiographs in the diagnosis of osteoporosis, we have measured lumbar-spine bone mineral density (BMD) in 33 women aged 33-74 years whose radiographs were reported as showing 'osteopenia' only and 41 women aged 39-78 with radiological evidence of vertebral deformation. These results were compared with the normal range for young women and with control groups of 41 normal elderly women aged 65-70 and 15 normal younger women aged 50-60. Women with radiological osteopenia had a significantly lower BMD than the younger (p < 0.001) but not the older control group. The BMD in women with vertebral deformation was significantly lower than in both control groups (p < 0.001) and women with radiological osteopenia (p < 0.01). Only 67% of women with radiological osteopenia had a bone density below the normal range for young women, compared with 81% of those with vertebral deformation, though the latter may be an underestimate because of the artefactual elevation of bone mass caused by vertebral fractures in the lumbar spine. Our study suggests that a radiological report of reduced bone density may be misleading. Bone density measurements should be performed to confirm that patients so identified require treatment for osteoporosis. For patients with vertebral deformation after minimal or no trauma, bone density measurements are probably not required.

Introduction

Osteoporosis is one of the major age-related disorders, and is associated with an increased risk of fracture of the forearm, vertebral body and femoral neck(1)(2). The lifetime risk of these fractures for a 50-year-old white woman is 16% for the forearm, 15.6% for the vertebrae and 17.5% for the femoral neck(3). These fractures result in considerable morbidity, mortality and health and social service expenditure(1)(2)(4).

One of the major predictors of fracture risk is low bone mass, and a one standard deviation reduction in vertebral bone density is associated with an increase in fracture incidence of between 50% and 100%(5). By the eighth decade of life, 44.2% of women have a lumbar-spine bone density below the normal range for young women, and 36.7% have a low femoral-neck bone density(3). A number of treatments for osteoporosis are now available, which prevent bone loss and reduce the risk of fracture(6)(7)(8). The wider use of these treatments depends on the accurate diagnosis of osteoporosis. Osteoporotic fractures may not always be detected clinically, as only about 35% of individuals with vertebral crush fracture present for medical attention at the time of fracture(9). Facilities for measuring bone mineral density (BMD) are confined to few centres in the UK at the present time and it is therefore difficult to identify women who are likely to benefit from treatment for osteoporosis.

Many patients have spine radiography performed because of back pain, and in some cases these show a vertebral fracture without preceding trauma, which is usually attributed to osteoporosis. Other spine radiographs may be reported as showing apparent reduction of bone density, without evidence of vertebral fracture or deformation. It is generally accepted that up to 50% of bone must be lost before reduced bone density is apparent on spinal radiographs. If this is indeed the case, then it may be possible to initiate treatment for osteoporosis on the basis of radiographs alone. We have therefore attempted to assess the value of spine radiography in the diagnosis of osteoporosis by comparison of the reports of spinal radiographs with bone density measurements.

Patients and Methods

Over a 12-month period in 1990-91, we studied a group of 74 women who had been consecutively referred to the Bone Clinic at Newcastle General Hospital with a radiological diagnosis of vertebral osteoporosis. Each had undergone spinal radiography at one of a number of hospitals in the Northern Region prior to referral. On the basis of the radiological reports they were divided into two groups. There were 33 women whose radiographs were reported as showing reduced bone density only (radiological 'osteopenia'), whilst the other 41 women had evidence of vertebral deformation. All radiographs were reviewed and the findings confirmed, but as this investigation was designed to study current radiological practice no attempt was made to quantify the findings or to use standard criteria for vertebral deformation or fracture. Nevertheless, many of the deformed vertebrae fulfilled the criteria for grade 2 vertebral deformity as described by Melton and colleagues(10).

Dual-photon absorptiometry (DPA) of the lumbar spine was performed using a Novo BMC-Lab 22a densitometer with a gadolinium source. The measurements were averaged over L2-4, a site containing predominantly trabecular bone, and the results expressed as BMD in g/[cm.sup.2]. The precision of measurement was 4%, having been determined previously by repeated estimation in a group of normal subjects.

The BMDs in these two groups of women with a radiological diagnosis of osteoporosis were compared with the normal range for young women aged 20-40 and with two control groups. The first was a group of 41 normal elderly women aged 65-70 recruited from the age-sex registers of local GPs, whilst the second was a group of 15 normal younger women aged 50-60 who had undergone bone densitometry to establish a reference range for that age group. The results were analysed using analysis of variance followed by the multiple range test to distinguish significant differences between groups.