Comparison of nuclear magnetic resonance spectroscopy with dual-photon absorptiometry and dual-energy X-ray absorptiometry in the measurement of thoracic vertebral bone mineral density: compressive force versus bone mineral.

31P nuclear magnetic resonance spectroscopy (NMRS) measurements were made on human T2 and T3 vertebral bodies. The bone mineral content (BMC) of isolated vertebral bodies minus the posterior elements and disks was measured using (1) NMRS on a 3.5 T, 85 mm bore GE Medical Systems NT-150 superconducting spectrometer, (2) a Lunar Corporation DPX-L dual-energy X-ray absorptiometry (DXA) scanner in an anterior-posterior (AP) orientation, (3) a Norland Corporation XR26 DXA scanner, also in an AP direction, and (4) a Norland Corporation model 2600 dual-photon absorptiometry (DPA) densitometer in both the AP and superior-inferior (SI) directions. Vertebral body volumes were measured using a water displacement technique to determine volume bone mineral densities (VBMD). They were then compressed to failure using an electrohydraulic testing device, followed by ashing in a muffle furnace at 700 degrees C for 18 h. Correlations of BMC between NMRS and DPA, DXA and ashing were excellent (0.96 < or = r < or = 0.99); in a one-way analysis of variance (ANOVA) test, means were not statistically different at a p level of 0.757. The correlations of VBMD between NMRS and the other methods were not as good (0.83 < or = r < or = 0.95); in a one-way ANOVA test, means were not statistically different at a p level of 0.089. BMC was a better predictor of ultimate compressive failure than VBMD for all six methods. For NMRS, the regression coefficient for BMC was r2 = 0.806, compared with r2 = 0.505 for VBMD. NMRS may prove an alternative to present methods of determining bone mineral.