A total of 8224 eligible women were identified. Of these, 1144 were not invited to take part by their GP as it was felt to be inappropriate (medical, psychological, or social reasons) (see ). Of the 7080 women invited, 3200 were enrolled (38.9% of eligible or 45.2% of invited) between October 2007 and May 2009, and were assigned to the control arm (n = 2138) or screening (n = 1062). No data is available on those who declined to take part. Baseline characteristics of participants in the two arms are shown in . No differences were seen. Median age of participants was 72.1 years (interquartile range [IQR], 69.0–76.1).
Demographic and Baseline Characteristics of Participants Randomized to Control and Screening
Trial compliance was generally good, with 92.6% of the screening arm attending the assessment (see ). However, of those identified to be at high-risk of VFs (n = 401), 22.7% (n = 91) did not attend for spinal radiographs (see ). The main reason given for nonattendance was the distance needed to travel to the hospital. Of the 310 thoracolumbar radiographs performed 230 (74.2%) were reported as showing no evidence of VFs, 52 (16.8%) were reported as possible VFs (using wording such as “probably a little loss of height,” “possible wedging of a couple of vertebrae,” and “minor depression of the superior endplate”), and 28 (9.0%) were reported as showing definite VFs. For the primary outcome of new prescriptions of osteoporosis medications 92.2% (n = 979) of the screening arm and 90 · 8% (n = 1942) of the control arm returned the 6-month follow-up questionnaire. The 12-month follow-up was completed by 89.8% (n = 1920) and 90.3% (n = 959), for the control and screening arms, respectively.
Fig. 2 Flow diagram of participants in the intervention arm showing numbers assessed as being at high risk, numbers who had thoracolumbar radiographs, numbers identified with vertebral fractures (VFs), and those started on osteoporosis medication at 6 months. (more ...)
At the 6-month follow-up, allocation to screening increased the odds of prescription of osteoporosis medications by 124% (odds ratio [OR] for prescription 2.24; 95% confidence interval [CI], 1.16–4.33; p = 0.016) (see ). Between 6 and 12 months no differences were seen in prescription of osteoporosis medications. There was an association between the presence of VF on the radiograph report (see and ) and prescription of osteoporosis medication. However, only 26.3% (n = 21) of those with definite or possible VFs had been prescribed osteoporosis medication by 6 months: 17.3% of possible and 42.9% of definite. Reporting of definite VF was associated with an increased proportion of new prescriptions for osteoporosis medications at 6 months compared to possible reporting of a VF, or reporting of no VF (p = 0.004). Of those who were high risk and had radiographs performed, 64.3% of women prescribed osteoporosis medications within 6 months of starting the study had been identified with definite or possible osteoporotic VF according to the NHS radiograph report. However, some new prescriptions were for women in the intervention arm identified as not at high-risk during the screening, and some prescriptions were for those women for whom no VF was identified on radiograph (see ). The number needed to screen to produce one additional prescription for osteoporosis medications within 6 months was 92.
Osteoporosis Medication Prescription and Fractures in Those in the Control and Screening Arms at 6-Month and 12-Month Follow-Up
Prescription of Osteoporosis Medications and Presence of Vertebral Fracture on Thoracolumbar Radiograph Report
At 12 months no difference was seen in new fractures between participants in the control group and those in the screening arm (OR for fracture 0.60; 95% CI, 0.35–1.03; p = 0.063). A total of 3.5% of the control arm (n = 75) and 2.0% (n = 21) of the screening arm fractured between 0 and 12 months. However, between 6 and 12 months of follow-up allocation to screening reduced the odds of occurrence of new fractures by 72% (OR for fracture 0.28; 95% CI, 0.12–0.67; p = 0.004). Assessment of site of fracture in control and intervention arms (see ) suggests most of the reduction were forearm fractures.
Observed Fractures in the Control Arm Compared With Observed and Expected Fractures in the Screening Arm in the First 12 Months