Comparing the Interobserver Reliability of 4 Methods Used to Measure Knee Laxity on Coronal Plane Stress Radiograph

Abstract

Background: Varus and valgus knee stress radiographs provide valuable information in the pre- and postoperative evaluation of joint laxity in patients with multiligament knee injuries (MLKIs). Purpose: To review the literature for described techniques of quantifying laxity on coronal stress radiographs of the knee and identify the most reliable method. Study Design: Cohort study (diagnosis); Level of evidence, 3. Methods: A thorough literature search using the MEDLINE and Embase databases identified 4 studies with distinct methods for objectively measuring laxity on varus and valgus stress radiographs: Heesterbeek et al (2008), Jacobsen (1976), LaPrade et al (2004), and Sawant et al (2004). To compare these methods, 200 coronal plane stress radiographs from 50 patients with MLKIs were retrospectively reviewed from an MLKI database at a single institution. The amount of varus and valgus laxity on each radiograph was measured independently by 4 reviewers using each method. Intraclass correlation coefficients (ICCs) with 95% CIs were calculated to assess the interobserver reliability of each method overall and the varus and valgus measurements individually. Results: For all 4 methods, the overall interobserver reliability was considered at least moderate. The method by Heesterbeek et al proved to have the highest interrater reliability in all domains—overall (ICC, 0.87 [95% CI, 0.85-0.90]), valgus (ICC, 0.83 [95% CI, 0.78-0.88]), and varus (ICC, 0.87 [95% CI, 0.83-0.90])—demonstrating good to excellent reliability both overall and in varus measurements and showing good reliability in valgus measurements. The method by Sawant et al demonstrated good reliability in valgus measurements. All other measures demonstrated moderate reliability. Conclusion: Available methods for measuring knee joint laxity on varus and valgus knee stress radiographs in patients with MLKIs demonstrated moderate to good interobserver reliability. The method described by Heesterbeek et al proved to have the highest reliability overall as well as in measurements on varus and valgus views individually.