Diagnostic agreement at the total hip using different DXA systems and the NHANES III database

In 1997, the National Health and Nutrition Examination Survey III (NHANES III) total hip reference database was adopted for T-score derivation in an effort to optimize diagnostic agreement among densitometers from different manufacturers. Our study was undertaken to evaluate the effectiveness of the NHANES III standardized database at achieving agreement in diagnostic classification (normal, osteopenia, or osteoporosis) based on total hip T-scores comparing 2 different dual-energy X-ray absorptiometry (DXA) systems. This was a retrospective analysis of standard bilateral hip and lumbar spine scans done in duplicate for 60 women scanned on both a GE Lunar Prodigy and Hologic Delphi DXA system. Classification based on lumbar spine T-scores using manufacturer-specific databases was also compared as no standardized lumbar spine reference database exists. Subject's mean age was 62 yr (range: 47-83 yr). There was no statistically significant difference in diagnostic classification between DXA systems (Prodigy vs Delphi), with agreement (same women classified same way) of 92% at the left total hip. Agreement was 100% when T-scores were greater than or equal to -0.8 and less than or equal to -1.2. There was 90% agreement between DXA systems at the lumbar spine. For both hip and spine, all diagnostic discrepancies occurred when the T-scores were at or near transition thresholds between normal and osteopenia or osteopenia and osteoporosis. The difference in mean T-scores between DXA systems at left total hip was 0.11 vs 0.32 for lumbar spine (p less than 0.001). Use of the NHANES III standardized database results in good diagnostic agreement at total hip between Prodigy and Delphi.