Table 2. Association of Varus Thrust at 12 months with Medial Joint Space Narrowing in Knees with OA (at enrollment or 12 months) over up to 7 years Subsequent Follow-up.

In the first 2 rows, the table shows the frequency of knees with each outcome for the dichotomous outcomes (row percentage) and mean (SD) of change for the continuous outcome among knees with (first row) and without (second row) varus thrust. The third and fourth rows show adjusted odds ratios and 95% confidence intervals [OR (95% CI)] for thrust for each of the 2 dichotomous outcomes (first 2 columns) and the adjusted regression coefficients (95% CI) for the continuous outcome (third column).

	Frequencies (%) and Adjusted Odds Ratios (95% CI) n = 3421 knees [975 (28.5%) with thrust]		Means (SD) and Adjusted Regression Coefficients (95% CI) n = 3650 knees [1085 (29.7%) with thrust]
	Dichotomous outcomes		Continuous outcome
Models and Covariates	Medial joint space narrowing, whole grade	Medial joint space narrowing, partial grade	Annualized loss of medial joint space width (mm), measured at x = 0.250 location
	n (%) with outcome		Mean (SD) annualized loss of medial joint space width
Varus thrust present (975 knees)	192 (19.7%)	329 (33.7%)	Varus thrust present (1085 knees)	0.173 (0.290)
Varus thrust absent (2446 knees)	345 (14.1%)	525 (21.5%)	Varus thrust absent (2565 knees)	0.117 (0.230)
Varus thrust, adj. for age, gender, BMI	1.50 (1.22, 1.85)	1.81 (1.51, 2.17)	0.053 (0.032, 0.074)
Varus thrust, adj. for age, gender, BMI, WOMAC Pain	1.49 (1.20, 1.83)a	1.80 (1.50, 2.16)b	0.049 (0.028, 0.069)c

^aAlso significant: BMI, WOMAC Pain

^bAlso significant: BMI, gender (decreased risk in women), WOMAC Pain

^cAlso significant: gender (decreased risk in women), WOMAC Pain