We randomised 129 knees which were to be replaced using a standard posterior-cruciate-ligament (PCL)-retaining cemented total knee replacement into two groups. In one the PCL was retained in the normal way and in the other it was resected. They were well matched, with a predominance of women, and a mean age of 67 years.

There was no statistically significant difference in the Hospital for Special Surgery scores at a mean of 57 months (56 to 60) between the two groups although 21 patients (24 knees) were lost to follow-up. Relief from pain, correction of deformity, range of movement, stability and strength were comparable in both. Radiological assessment showed femoral rollback in approximately 20% of knees with a slightly higher incidence in the PCL-resected group. There was no significant loosening detected in either group at review at two years.

At five years, one knee in the PCL-retained group had been revised because of infection and one patient in each group was awaiting revision for loosening.

Our findings have shown no significant difference in the five-year results for a PCL-retaining total knee replacement if the PCL is excised or preserved. This suggests two important points. First, the PCL is not functional in most patients with a total knee replacement even when retained. Secondly, patients with an excised PCL show a good result with a PCL-retaining implant, thereby questioning the need for a posterior stabilised design in such a situation.

We prospectively randomised 100 patients undergoing cemented total knee replacement to receive either a single deep closed-suction drain or no drain.

The total blood loss was significantly greater in those with a drain (568 ml versus 119 ml, p < 0.01; 95% CI 360 to 520) although those without lost more blood into the dressings (55 ml versus 119 ml, p < 0.01; 95% CI −70 to 10). There was no statistical difference in the postoperative swelling or pain score, or in the incidence of pyrexia, ecchymosis, time at which flexion was regained or the need for manipulation, or in the incidence of infection at a minimum of five years after surgery in the two groups.

We have been unable to provide evidence to support the use of a closed-suction drain in cemented knee arthroplasty. It merely interferes with mobilisation and complicates nursing. Reinfusion drains may, however, prove to be beneficial.

Total knee arthroplasty (TKR) using a medial capsular approach gives worse results in arthritic knees with valgus deformity than in those in varus, usually because of swelling, poor wound healing and stiffness, instability, recurrent valgus deformity and poor patellar tracking.

A technique for replacement TKR of valgus knees using a lateral capsular approach was described several years ago, but was not routinely adopted because of the difficulties with and complexity of the procedure which included deliberate elevation of the tibial tubercle. In order to avoid this we have modified and simplified the procedure. Our preliminary results suggest that this lateral approach is safe and may give a better outcome than that through the medial capsule for the replacement of valgus knees.