PURPOSE: In a mobile-bearing unicompartmental knee arthroplasty (UKA), stability is of utmost importance to promote knee function and to prevent dislocation of the insert. Gap balancing can be guided by the use of spacers or a tensioner. The goal of this study is to compare laxity of a tension-guided implantation technique versus a spacer-guided technique for medial UKA with a mobile bearing. Also clinical function was compared between the groups. METHODS: The tension-guided UKA system (BalanSys™, Mathys Ltd, Bettlach, Switzerland) was compared with a retrospective group with a spacer-guided system (Oxford, Biomet Ltd, Bridgend, UK). A total of 30 tension-guided medial UKAs were implanted and compared with 35 spacer-guided medial prostheses. In both groups, valgus laxity was measured at least 4 months postoperatively in extension and 70° flexion using stress radiographs. Knee Society Scores (KSS) were obtained at the 6-month follow-up. RESULTS: Valgus laxity in flexion was significantly higher in the tension-guided group compared with the spacer-guided group: 3.9° (SD 1.8°) versus 2.4° (SD 1.2°), respectively, P < 0.001). In extension, valgus laxity was significantly different: 1.8° (SD 1.0°) in the tension-guided group compared with 2.7° (SD 0.9°) in the spacer-guided group (P < 0.001). There was no significant difference between the KSS for the two groups (n.s.). CONCLUSIONS: The tensor-guided system resulted in significantly more valgus laxity in flexion compared with the spacer-guided system. However, in extension, the situation was reversed: the tension-guided system resulted in less valgus laxity than the spacer-guided system. Clinically, there were no differences between the groups. The valgus laxity found with the spacer-guided system better approximates the valgus laxity values of the healthy elderly.