Does a knee replacement reduce the load on both knees during walking?

In our last article I told you about a group of 20 patients we measured just before they had a knee joint replacement. We were able to confirm the suspicion of many patients, that walking with a sore knee puts additional load on the good knee. We also recorded the muscles on either side of the joint, showing that they were working at the same time, in other words against each other. This phenomenon, known as co-contraction, can make walking feel safer but only at the cost of even higher loads on the joints.

Our volunteers came back to the lab in ORLAU once they had recovered from their knee surgery and we repeated all our tests of their walking. Overall we found that they were walking faster and they were certainly more comfortable with a new knee but what had happened to the loading? The good news is that the loading on the side with the joint replacement had reduced, as had the abnormal co-contraction of the muscles. Unfortunately, however, the high loading did not fully resolve on the good side and co-contraction generally remained high. We were left with an important question. Had our volunteers always had high joint loads or had they learned to walk in a particular way as their knee became painful?

So far I have described the general pattern observed in the whole group of patients. Inevitably, however, some did better than others. We were curious to know if the patients who did well were already different from those who did less well before their operation. Predicting the outcome of surgery is notoriously difficult because there are so many factors involved including the surgery itself, any physiotherapy or exercise habits afterwards and the patient’s personal outlook. Some of these are difficult to measure. We wondered whether there was anything we could detect using our measurements of walking?

To try and answer this question Andy Metcalfe used computer modelling techniques. He used his software to analyse 20 sets of control data, from subjects without arthritis, and also data from our patients. We asked the software to find which features were the best at telling the two groups apart. We then assumed that a patient who did well would move towards normal values for those specific parameters. The process is complicated but we produced a simple triangle graph to illustrate the results. A patient who does well should move from the bottom right corner (labelled {OA} in the blue half of the triangle) towards the bottom left (labelled {Normal} in the yellow half).

The good news is that lots of our patients did move in the right direction, with 8 out of 15 ending up on the ‘Normal’ side of the triangle for their arthritic knee.

Having found 8 patients who did well we could compare them with the 7 who remained in the ‘OA’ half of the plot. The two groups were very similar in terms of their body weight (BMI) and how sore their knees were before the operation. The group who did well were a little younger (on average by 7 years), but the strongest prediction came from their walking. Those who were closer to the bottom right corner (ie those with the worst gait patterns) before their surgery were less likely to cross over into the normal half afterwards. On this basis we were able to predict success in 87% of our patients for their affected knee.

Not all patients are the same but we would like to help everyone to get the best out of their treatment and avoid patients needing further joint replacements wherever possible. From our results it looks like walking well before surgery is important so perhaps that is something we can work on with our physiotherapy. Clearly there is more work to do.

If you would like to read our full paper you can find it here.