Interlaboratory variation in a commercial bone mineral analyzer.

Measurements of bone mineral content (BMC) were made in 14 different laboratories in the U.S. and four in Europe using commercially produced instrumentation (Norland Bone Mineral Analyzer) for 125I absorptiometry. A three-chambered standard (dipotassium hydrogen phosphate) was measured in each laboratory following their own calibration. The values of BMC in the middle range (0.6 g/cm) all were adequate (within +/- 2%), but the BMC values were underestimated by 5% or more in five laboratories for the largest chamber and in three laboratories for the smallest chamber. Width values were accurate (+/- 3%) over 0.7-1.6 cm. The effect of underestimating large values in clinical studies is to reduce the difference between normals and abnormals. Calibration error also may be responsible for the variable normal values found in the U.S. and Europe by some users of this instrument.