The Tromsø Study: artifacts in forearm bone densitometry--prevalence and effect

Suboptimal performance of bone densitometer, operator and/or subject may cause artifacts of consequence both for individual patient management and research. The prevalence and effects of such artifacts are largely unknown in densitometry. A cross-sectional population-based study was carried out of artifacts in forearm bone densitometry with single X-ray Absorptiometry (SXA) of the nondominant hand (distal and ultradistal site). After the screening, all scans were reviewed for artifact detection and reanalysis. The effect on the bone mineral density (BMD) result was found by comparing artifactual scans with a reanalyzed version or with normal repeat scans. All women aged 50-74 years, all men aged 55-74 years and 5-10% samples of other age groups aged >/=25 years attending the fourth Tromso health study were invited to have bone densitometry. The response rate from the background population was 80% (n = 7948). Fourteen percent of subjects had a movement artifact at either the distal or ultradistal site. The individual BMD variation was twice as large in scans with a movement artifact (0.94%) compared with normal scans (0.58%) (p = 0.0027). The radial endplate was inaccurately detected in 74% of the scans. Reanalysis of these scans led to a mean 3.8% decrease in the BMD value and an increase in the prevalence of osteoporosis of 10%. Artifacts were thus common, and their effects were clinically relevant in forearm bone densitometry. Artifacts and their effects need to be characterized in other bone densitometry settings also.