A longitudinal study of supine lateral DXA of the lumbar spine: A comparison with posteroanterior spine, hip and total-body DXA

We report a study to assess whether supine lateral dual-energy X-ray absorptiometry (DXA) scans of the lumbar spine provide better data for monitoring response to treatment than alternative measurement sites such as the posteroanterior (PA) spine, hip and total body. The study population was 152 women enrolled in a placebo-controlled clinical trial of cyclical etidronate therapy. All subjects were 1-10 years after the menopause with bone mineral density (BMD) between 0 and -2 SD of age-matched normal women. Paired PA and lateral spine, left hip and total-body DXA scans were performed at baseline, 1 year and 2 years on a Hologic QDR-2000. One hundred and thirty-one subjects completed the study. Mean percentage change from baseline at 2 years in the treated (n = 61) and control (n = 70) groups was calculated for vertebral body, width-adjusted (WA) vertebral body, midvertebral body and WA mid-vertebral body BMD measurements on the lateral scans and compared with the percentage changes in PA spine, femoral neck, trochanter, Ward's triangle and total-body BMD. The long-term precision for each BMD measurement site was obtained by linear regression analysis in subjects taking placebo. Overall treatment effect, defined as the difference in the percentage change in BMD in the two treatment groups at 2 years, was divided by long-term precision to give an index of the ability of each site to monitor response to treatment. Results (and standard errors) normalized to the ratio of treatment effect/precision for PA spine BMD were as follows: PA spine, 1.00; vertebral body, 0.89 (0.14); WA vertebral body, 0.78 (0.14); mid-vertebral body, 0.65 (0.14); WA midvertebral body, 0.60 (0.13); femoral neck, 0.35 (0.15); trochanter, 0.45 (0.15); Ward's triangle, 0.59 (0.22); total body, 0.52 (0.19). Although treatment effect was larger for lateral than for PA spine BMD, this advantage was offset by the greater precision errors. PA spine BMD remains the optimum measurement for longitudinal studies in recently postmenopausal women.