Spiking of the tibial tubercles--a radiological feature of osteoarthritis?

OBJECTIVE To determine whether 'spiking' or angulation of the tibial tubercle is associated with other radiographic markers of osteoarthritis (OA) or pain in the knee joint, and could be taken as a reliable marker for early OA, in a large general population sample. METHODS A total of 950 women from the Chingford general population survey underwent anteroposterior extended weight bearing radiography of the knees. Angulation of the tip of the medial and lateral tubercles, and height of the tubercles above the tibial plateau were measured. These measures were compared with standard radiographic indices including qualitative Kellgren and Lawrence global score, individual scores of osteophytes and joint space narrowing, and pain score. Intraobserver and interobserver reproducibility for assessment of spiking was tested in a subgroup of 50 films using two observers and two readings. Tibial spiking (angulation and height) was defined for this study as the top 10th centile for the whole population. Patients with normal radiographs (Kellgren and Lawrence grade 0) were allocated to quartile groups on the basis of spiking to define severity. Odds ratios were then calculated for the association of spiking and knee pain. RESULTS The majority of the measures of tibial spiking were highly reproducible. There was a significant correlation between tibial spike angulation and the presence of osteophytes, but not joint space narrowing. The correlations for spike height with osteophytes and joint space narrowing were poor. There was an association between spike angulation at the lateral tubercle and reported knee pain (odds ratio 1.45 (95% confidence interval 1.03 to 2.03)) after adjustment for age, body mass index, and Kellgren and Lawrence score. There was no association between medial spike angulation or spike height and pain. Among the 950 women, 683 (72%) had normal radiographs (Kellgren and Lawrence = 0); in this group there was a similar association between pain and lateral spike angulation, but not medial spike angulation or spike height. CONCLUSIONS Tibial spiking is associated with the presence of knee osteophytes and is reproducible, but does not have a strong independent relationship with knee pain. In patients with normal radiographs there is no useful correlation between tibial spiking and pain. Isolated tibial spiking is not a reliable sign of early knee OA, and should not routinely be reported.