Sensitivity of different measures of frontal plane alignment to medial and lateral joint space narrowing: From the osteoarthritis initiative

Abstract

ObjectivesTo explore cross-sectional relationships between a new radiographic measure of the femorotibial angle (FTA), the hip–knee–ankle angle (HKA), and the goniometry; to quantify sex differences between measures; and to evaluate the sensitivity of these measures to medial and lateral joint space narrowing (JSN). MethodsConcurrent validity was evaluated in 2123 knees from the osteoarthritis (OA) initiative using Pearson correlation coefficients (r) and Bland–Altman plots (offsets). Diagnostic validity with respect to JSN was evaluated using areas under the receiver-operating characteristic curves (AUC) and standardized mean differences (SMD). Analyses were stratified by sex and JSN. ResultsJSN and sex contributed significantly to regression models predicting offsets between measures (p < 0.001), after controlling for age, BMI, and OA severity. There were weak correlations between FTA vs. goniometry (r: 0.16–0.22), and moderate correlations between FTA vs. HKA (r: 0.25–0.53) and HKA vs. goniometry (r: 0.40–0.67). The offset between the new FTA measure and HKA was larger in females than males (p < 0.001). Offsets between radiographic measures and goniometry also varied by sex and JSN (FTA: 2.9°–7.6°; HKA: 0.5°–2.4°). AUC (0.74–0.91) and SMD (0.53–3.80) between JSN strata were largest for FTA, whereas diagnostic validity was moderate for HKA (AUC: 0.69–0.80; SMD: 0.43–2.04) and weakest for handheld goniometry (AUC: 0.56–0.63; SMD: 0.07–0.90). ConclusionsCompared to HKA, goniometry demonstrated poor diagnostic validity with respect to JSN. The new FTA measure, in contrast, represented a reasonable surrogate of radiographic disease severity (JSN). The new FTA measure was skewed in the varus direction when compared to HKA and more so in females. This requires cautious interpretation when measurements are related to previous studies.