Total knee replacement (TKR) instability is a poorly defined entity. The diagnosis is based on symptoms and clinical tests of knee laxity. A subjective feeling of instability, such as giving away is not always present in the unstable knee; however, the patient will rather have non-specific symptoms like pain and swelling. On the other hand, symptoms of instability may be present without excessive laxity. The evaluation of abnormal laxity is also uncertain due to individual variations of normal laxity and the subjectivity of manual laxity testing. Different instrumented methods have been developed in order to reduce subjective bias. Stress radiography is one such test which has been utilised to objectively measure medial and lateral joint opening during varus/valgus stress and anteroposterior translation when subjected to sagittal stress. The limits of acceptable laxity measured by stress radiography is however not established. The aim of this study is to determine thresholds of acceptable laxity as measured by stress radiography.

TKR patients will be invited to participate when the decision of primary knee replacement is made. These patients will be scheduled for stress radiography and clinical evaluation at 12 months. Additional patients that are referred to the outpatient clinic due to painful TKR will also be asked to participate. Patients will be categorized as either "satisfied" (control) or "not satisfied" (case). A minimum of 20:80 and a maximum of 50:50 case control ratio is accepted. This ratio will be monitored during inclusion.

Patients are included after informed consent. The primary outcome is to determine diagnostic accuracy of knee laxity measurement by stress radiography in predicting dissatisfaction after total knee arthroplasty. The primary dichotomous outcome variable is patient satisfaction. Patient satisfaction is defined as an Oxford Knee Score (OKS) above the Patient Acceptable Symptom Threshold (PASS) of 30. Patient satisfaction status will be tested against three predictor variables: Coronal knee laxity during varus and valgus stress, and sagittal laxity defined as total anteroposterior translation. Knee laxity will be measured in both the TKR knee and the contralateral healthy knee. The side to side difference in laxity will be used as the primary predictor variable in order to mitigate individual variability. The absolute opening angle will also be analysed separately in order to include those with contralateral knees with substantial Knee laxity measured by stress radiography Coronal joint laxity is measured as the gapping between the femoral and tibial joint line during varus and valgus stress on an anteroposterior radiograph. The joint convergence angle (JCA) is used to quantify the amount of gapping.The knees will be subjected to 150 newton (N) varus and valgus stress in 30 degrees of flexion. A 30-degree knee bolster is used to ensure a consistent flexion angle. The limb is placed in a Telos frame to ensure consistent force application during stress testing.

Sagittal laxity is quantified as the total anteroposterior (AP) translation of the tibia relative to the femur during anterior and posterior stress on a lateral radiograph. In the prosthetic knee, the distance from the most posterior point on the femoral condyles to the posterior margin of the tibial plateau will serve as a reference. The difference of this distance in anterior and posterior stressed position will reflect the total AP translation. The Blumensaat-Anterior Tibia (BAT) method will be used as reference distance in the native knee. The test will be performed with 90-degree flexion angle of the knee, again using the Telos frame to perform the anterior and posterior drawer test with a force of 150N. To ensure consistent rotational position of the femur, the femoral condyles must be superimposed on the lateral stress image.

Reliability of the acquisition and interpretation of images will be evaluated by three parameters: Precision under repeatability conditions, inter- and intra-observer reliability.

Secondary outcomes A) Agreement between the clinical examination of knee laxity and laxity measured by stress radiographs The examiner will evaluate the laxity as "normal" or "excessive". The examiner is blinded for both patient satisfaction and stress radiograph results. Coronal laxity is measured at 0 and 30 degrees of flexion. Anteroposterior laxity is evaluated by anterior and posterior drawer test, as well as the Clunk test. Correlation between the clinical examination and the dichotomized results of normal or excessive in stress radiography will be explored.

B) Agreement of clinical examination of laxity. One dedicated physiotherapist and one orthopaedic surgeon evaluate knee laxity according to the aforementioned clinical examination protocol. For each test, the examiners will rate the laxity as excessive or normal. The agreement between examiners will be explored.

C) Exploration of knee instability symptom score

Patients in the "not-satisfied group" is selected, and divided in to groups of "high" and "normal" laxity. High laxity measured by stress radiography will be defined as laxity above the thresholds calculated by the ROC analysis. The study group will explore low-score items of the Oxford Knee Score that are more frequently encountered in the high-laxity versus the normal laxity group as well as the response (yes/no) to anchor questions which are assumed to be associated with instability:

• Does the knee give away/buckle?
• Do you trust your knee during activity?
• Does the knee mainly become painful/swollen during or after activity?
• Do you have difficulties walking down hill or down stairs? Statistics The sample size is chosen to be 100. Assuming a sampling distribution of 80:20 stable and unstable knees and a moderate discrimination signal between stable and unstable knees, the estimated laxity reference interval (values below the threshold) will cover 0.89-0.98 of all stable knees with a confidence of 90%. The proportion of unstable knees above this threshold will be 0.66 (0.46,0.82). If we assume a high discrimination signal, the expected proportion of unstable knees above the limit will be 0.92 (0.75,0.98)[7].

We will explore diagnostic thresholds of each laxity by receiver operating characteristic curve analysis (ROC). Total sample and gender specific thresholds will be explored.

Reliability of radiographic measurements within and between raters (intra- and interobserver reliability), as well as test-retest reliability is evaluated by calculation of intraclass correlation coefficient. Agreement is likewise evaluated by constructing Bland-Altman plots and calculating the systematic error and 95% limits of agreement (LOA). The sample size for this purpose is 20 patients, in which double examinations will be performed. The patient will get up between examinations, and the whole set up repeated before acquisition of the second image.

Agreement between radiographic laxity and clinical examination, as well as between clinical examiners will be evaluated by calculation of the kappa value, as well as crude descriptive values of absolute agreement.

Data handling All patient data will be de-identified and stored at a secure database (E-reg). The key code used for de-identifying the data will be stored at a separate secure database. The data will be collected at 12-months follow up. Results from patient reported outcome measures, clinical knee examination and results from stress radiography will be stored in E-reg. The data will not be shared by external researchers or institutions, but handled only by the project leader and project coordinator.

Ethical considerations The patients are given standard care and follow-up as given in total knee arthroplasty treatment at Diakonhjemmet hospital. Study participants will additionally be subjected to an additional plain radiographic examination. The radiation dose (0,006mSv) is equivalent to less than 1 day of background radiation in Oslo. The patients are not subjected to additional risks through participation in the project.

Finances and publications An application for research funding has been submitted to the local research fund at Diakonhjemmet hospital. Any funds will be utilized for stress radiographic examination and to cover salary for part time research. Results from the study will be published in an open access journal.

Inclusion Criteria:

• Patients with cruciate retaining (CR) TKR with ≥12 months follow up

Exclusion Criteria:

• TKR implants with higher constrained articulations (posterior stabilized (PS), condylar constrained (CC) and hinged implants)
• Suspected infection
• Range of motion (ROM) limitations that infers with protocol of stress testing; combined flexion contracture >30 degrees and flexion deficit < 80 degrees
• Suspected instability in contralateral knees

All patients that are operated with primary knee replacement.