Seventy patients with chronic anterior instability underwent anterior cruciate ligament reconstruction with a Dacron prosthesis pretensioned to 60 N. Of these patients, 49% (34%) had combined medial instability, 32% (22) had failed previous anterior cruciate ligament surgery, and 37% (26) had previous meniscectomy. At reconstruction, 12 patients had their medial instability treated; 22 did not. Follow-up intervals were 3, 6, and 12 months and then each year to 5 years. The 5-year followup included 69 patients; the other 1 had the ligament removed because of a synovial fistula at 8 months. Results were 23% prosthesis ruptures, 3% poor, 17% fair, 16% good, and 39% excellent. The 2-year results showed the same distribution, but a lower rupture rate, which was affected by placement of the tibial tunnel within the anterior one-third of the tibia (9 times increase) and coexisting nonrepaired medial instability (5 times increase). Those patients with perfect placement of the ligament who also had good medial stability and no previous ligament surgery had no rupture at 5 years. The stability that was gained at surgery was gradually lost (-11.2% per year). At 5 years, the uninjured knee also had lost 41% of the preoperative stability; the mean laxity difference was within +/- 2 mm. The mean improvement in subjective knee function (Lysholm score 74.5 to 91.9) was maintained during the followup. The mean preoperative activity level improved significantly, but did not reach the preinjury level. These results show that the Dacron prosthesis will not give acceptable results in salvage cases where other instabilities are left untreated.