©1994 British Editorial Society of Bone and Joint Surgery 0301-620X194/1698 $2.00 82 ThE JOURNAL OF BONE AND JOINT SURGERY OVER-THE-TOP OR TUNNEL RECONSTRUCTION OF THE ANTERIOR CRUCIATE LIGAMENT? A PROSPECTIVE RANDOMISED STUDY OF 54 PATIENTS HAKAN JONSSON, LARS G. ELMQVIST, JOHAN KARRHOLM, YELVERTON TEGNER From the University Hospital in Northern Sweden, Ume#{226}, Sweden We report a review of 54 patients with chronic anterior cruciate ligament rupture treated by patellar-quadriceps tendon graft augmented with polypropylene braid (Ken- nedy-LAD). The femoral placement of the graft was randomised to either a modified over-the-top (OTT) or a tunnel position obtained by an isometric drill guide (ISO). At the two-year follow-up both procedures had resulted in improvement of subjective knee function and activity level. Stereoradiographic measurements showed reduction of anteroposterior laxity to near normal in about one-third of the patients, but muscle strength and objective functional performance showed little or no changes. The OTT group had better subjective knee function. We were unable to confirm the theoretical advantages of the use of the drill guide, partly because it provided a tibial tunnel which was too anterior. J BoneJoint Surg (Br] 1994; 76-B:82-7. Received 26 April 1993; Accepted 23 June 1993 The aim of surgical reconstruction of the anterior cruciate ligament is to restore normal anteroposterior stability without restricting the range of motion. Many factors influence the result, but the position of an intra-articular graft is one of the most important. An isometric placing, with a constant distance between the attachment sites throughout the range of motion, has been considered crucial (Odensten and Gillquist 1985; Graf 1987). If isometry is not achieved, the high tensile forces at extreme positions will elongate or H. Jonsson, MD, PhD L. G. Elmqvist, MD, PhD J. KSrrholm, MD, Associate Professor Department of Orthopaedics, University Hospital in Northern Sweden, 5- 401 85 UmeA, Sweden. Y. Tegner, MD, Associate Professor The Ermeline Clinic, P0 Box 195, S-97106 LuleAna, Sweden. Correspondence should be sent to Dr H. Jonsson. rupture the graft, or restrict joint motion (Arms et al 1984; Graf 1987). Variation in the site of attachment has been shown to be more important at the femoral end than at the tibial end (Hoogland and Hillen 1984; Bradley et al 1988; Hefzy, Grood and Noyes 1989), but improper tibial placement may result in graft impingement at the notch (Penner et al 1988; Howell, Clark and Farley 1991; Yaru, Daniel and Penner 1992; Howell and Clark 1992). There are two basic methods of femoral attachment: the isometric tunnel (ISO) technique first introduced by Hey Groves in 1917 (reprinted 1980), and placement over the lateral femoral condyle, the over-the-top (OTT) technique of McIntosh (1974). OTI’ placement is not isometric: the graft will slacken in flexion and become more tense in extension with the danger of rupture. For this reason some authors disapprove of its use (Hoogland and Hillen 1984; Odensten and Gillquist 1985; Penner et al 1988; Schutzer, Christen and Jakob 1989), but others have reported successful clinical results (Roth et al 1985; Daniel et al 1988; Howe et al 1991). None of these reports has included a controlled clinical study. Our aim was to evaluate the stability and the position of the graft, using clinical and muscle function tests and detailed radiological assessments. PATIENTS AND METhODS From February 1988 to May 1990 a total of 54 patients with chronic anterior cruciate ligament injuries and knee instabil- ity had repair operations. We included only patients with unilateral anterior cruciate ligament rupture who had had no previous intra-articular graft surgery. Preoperatively, the patients were randomised by tossing a coin to either OTT or ISO. Four patients were excluded later because of changes in their postoperative rehabilitation, leaving 26 O1T and 24 ISO cases for analysis. Twenty-nine of the patients had had previous meniscal surgery or ligament suture but these were evenly distributed between the groups (Table I). The ligament reconstructions were performed by three surgeons (LGE, YT, Hi) at two hospitals, but the techniques had been standardised by previous discussion and joint performance of the first few operations. Surgical technique. All reconstructions were made with a graft taken from the central strip of the quadriceps and