A pilot study of the reproducibility and validity of measuring knee subchondral bone density in the tibia

To describe the reproducibility and validity of six different measurement techniques for knee subchondral bone mineral density (sBMD).

A consecutive sample of 50 male and female participants from a population-based longitudinal study had sBMD assessed using dual energy X-ray absorptiometry scans. Anthropometric, knee pain, cartilage and bone measures by magnetic resonance imaging and radiographic osteoarthritis (OA) were assessed. The six methods were defined as: (1) the midpoint of one intercondylar spine, across the tibial surface and descending 10mm; from the midpoint of the two intercondylar spines (2) the top of the spine descending 20mm, (3) 10-20mm beneath the top of the spine; from the tibial surface descending, (4) 10mm, (5) 15 mm, and (6) 20mm.

All six methods had excellent reproducibility (intra-class correlation coefficient 0.98-1.00). sBMD was higher in males (methods 2-4) and higher in those with medial tibial osteophytes (methods 1, 3 and 4). Medial tibial cartilage defects and overall cartilage defects correlated with sBMD (methods 3 and 4). Method 2, which includes the intercondylar spine, correlated with medial tibial bone size. Measuring sBMD using methods 3 and 4 produced the greatest number of associations with joint features of OA.

These preliminary results need confirmation in larger longitudinal samples but suggest that sBMD can be accurately measured and plays a role in knee OA. Methods 3 and 4 had the best concurrent validity; however, method 2 adds additional information on tibial bone size, suggesting that two measures are necessary in clinical studies.