"Effect of Posterior Tibial Slope on Postoperative Flexion and Functional Outcome in Total Knee Arthroplasty. A Double Blinded Randomised Controlled Trial"

We present a prospective trial examining the effect of posterior tibial slope at total knee arthroplasty (TKA) on the range of movement and functional outcome. Current literature shows little difference clinically in TKA with increasing posterior tibial slope. Previous studies have been retrospective or involved small numbers and may represent poor ligament balancing or inaccurate alignment. In a prospective, double-blinded, randomised controlled trial, 250 patients undergoing primary Profix TKA, were randomised to receive either a 0 or 4 degree posterior tibial cut. Range of movement (ROM) was measured pre-operatively, at 3 months and 1 year by a single clinical physiotherapist. SF-12 and WOMAC scores were calculated at the same visit. Both patient and physiotherapist were blinded to the angle of tibial slope. Mean one year post operative ROM was greater by 2 degrees (p=0.470) in those with a 4 degree tibial slope. Post operatively both groups had significant improvement in functional outcome scores. A 0.2 (p= 0.892) and 0.51 (p= 0.707) greater improvement in SF12 physical score and mental scores respectively was found in the 0 degree group at one year. There was also a 1.09 (p=0.718) greater improvement in WOMAC score with a 0 degree slope. In conclusion increased posterior tibial slope gives a marginally better but non-significant post operative ROM and makes no significant difference to functional outcome.

operative range of movement being undeniably the most consistent predictor. The range of flexion ultimately achieved however is likely to be multifactorial and surgical factors including maintenance of posterior tibial slope, posterior condylar offset and alteration of the joint line, although contentious, may play a significant role [6,7].
Most modern knee arthroplasty systems allow posterior tibial slope options, typically zero, three, four some up to seven degrees.
There is little consensus on the optimum tibial slope with surgeons using their 'preferred' cutting block to obtain the desired slope or attempt to recreate the preoperative physiological posterior tibial slope.
A posterior tibial slope may improve femoral roll-back, especially laterally, providing more 'normal' knee kinematics, leading to increased flexion. In a skeletal model, a posterior tibial slope reduces the height of the posterior tibial wall and hence delays contact between the posterior tibia and femur during flexion and theoretically leading to an increased range of movement. In vivo, however, the soft tissues may affect the flexion gained due to tissue impingement or even stiffness. An excessive posterior slope may lead to flexion instability, posterior wear, loss of full extension and early loosening of the tibial component [8]. The benefit of altering the tibial cutting angle therefore remains unclear. Numerous cadaveric studies have suggested that recreating the normal 5˚ to 10˚ tibial slope increases the flexion gap and improves post-operative range of movement [9,10]. However, many clinical trials have failed to replicate these results [11][12][13][14]. The studies comparing tibial slope angle with post-operative movement have been retrospective case series involving different implants with multiple variants. Because the flexion gap and therefore range of movement are influenced by many factors in addition to the tibial slope, including posterior cruciate ligament tension, femoral component size (posterior condylar offset) and quadriceps tightness, it is difficult to analyse the effect of the slope using retrospective case series. We therefore undertook a prospective double-blind randomised controlled trial comparing the posterior tibial slope angle with post-operative range of movement and functional outcome based on the null hypothesis that the tibial slope makes no difference to final movement at one year and that there is no difference in functional outcome between one group with a tibial slope cut at 0˚ and another with a slope cut at 4˚.

Ethics
Ethical approval was obtained before study start (Research Ethics Committee reference 05/Q1502/80). A National Health Service grant funded a research fellow (CAJ) to run the trial.

Statistical Analysis
Pre-trial calculation showed that detecting a mean 7.5˚ difference in flexion between the two groups with 80% power required a sample size of 125 patients per arm. Statistical analysis was carried out using Stats Direct (www.statsdirect.co.uk). The Mann-Whitney U test was used to test the null hypothesis, with p < 0.05 as significant and all results reported with two-sided p values.

Patient Recruitment
The for health-related physical and mental wellbeing. In the WOMAC scoring system a lower score indicates a better outcome; in the SF-12 scoring system a higher score indicates a better outcome.

Surgical Techniques
Patient randomisation to the 0˚ or 4˚ tibial slope group was relayed to the surgeon in the anesthetic room from a numbered sealed envelope by a pre-determined randomisation allocation.
The operations were done through a standard trivector approach and the posterior cruciate ligament was retained as standard. The tibial resection was made as per Smith and Nephew operative technique recommendations using either the 0˚ or 4˚ cutting block.
An intramedullary tibial alignment rod was inserted and securely fastened to the allocated tibial cutting block with a set screw and two supplementary pins were placed in the tibia. All cuts were made using an implant specific 'Profix' saw blade designed to minimize toggle and increase accuracy of the cuts. The standard poly-ethylene inserts used had no 'inbuilt' posterior slope, hence the slope cut on the tibia produced the final posterior slope. A 'Conforming Plus' insert with anterior lip was used if the surgeon felt the posterior cruciate ligament was non-functional (n = 7 patients). All components were cemented using CMWII gentamicinenhanced cement (DePuy CMW, UK). All operations were carried out by the two senior authors or by trainees under their direct supervision as first assistant. All patients adhered to the standard hospital post-operative protocol. Cefuroxime was given on induction with subcutaneous dalteparin was started on the day of surgery and continued for ten days, along with thromboembolism-deterrent stockings. Patients were mobilised on the day or day after surgery. They were discharged home once independently mobile with at least 80 degrees of flexion and had standardised outpatient physiotherapy follow up.

Outcome Measures
At 3 and 12 months post-operatively, the same specialist physiotherapist re-measured knee flexion using a goniometer and assigning WOMAC and SF-12 scores. Both the patient and physiotherapist remained blinded to the slope of the tibial insert.
Patients were informed of its value after the database was frozen after the one-year review. Two authors (JRF and CJ), also blinded to the randomisation groups, independently reviewed the standard lateral postoperative radiographs to measure the prosthetic posterior tibial slope. Radiographers were instructed to obtain as true a lateral image as possible. The tibial slope was measured digitally as the angle between the medial tibial plateau and the proximal tibial anatomical axis (a line connecting the mid point of the tibia at 5 and 15cm below the joint line) on a lateral knee radiograph. Both authors measured each patient twice on different occasions giving a total of four measurements for each patient and a mean of these values was calculated. Larger tibias have a larger antero-posterior distance, hence there will be a larger posterior drop in height with a 4˚ posterior tibial slope than with smaller tibias and thus potentially greater flexion achieved. Further analysis was undertaken to assess whether tibial size was a factor. Patients were also divided into two groups (small and large) by tibial implant size. The Profix arthroplasty comes in six tibial sizes, numbered 1 to 6, of which size 1 is not of standard availability and rarely used.
It was not used in any patient in this cohort. We compared range of movement in those with a 'small' tibial implants, sizes 2, 3 and 4, against those with a 'large' implants, sizes 5 and 6.

Results
Two hundred and fifty six patients were recruited to the trial; six withdrew before surgery, leaving 250 who were randomised to a 0˚ or 4˚ tibial slope. Nine patients were subsequently lost to follow-up, leaving 241 patients with outcome measures. Baseline characteristics, pre-operative movement and quality of life measures were similar in both groups (Table 1), with a slight female predominance. No parameter was statistically significant.

Primary Outcomes
There was no significant difference in range of movement or quality of life measures at any point between the 0˚ and 4˚ groups ( Table 2). The range of movement decreased in both groups at three months before improving back to pre-operative levels at one year.
Patient reported outcomes improved significantly post-operatively.
WOMAC scores also improved significantly with similar scores in both groups postoperatively. SF-12 scores likewise improved in both groups, with no significant difference between the two groups.
There was no difference in the post-operative scores of any measure when comparing the 'small' and 'large' tibial sizes. The WOMAC scores were significantly higher pre-operatively for the 4 degree group with a greater improvement at 12 months but no significant difference in scores at 12 months compared to the 0 degree group.

Secondary Outcomes
On measurement of the lateral x-rays, the mean tibial slope in

Discussion
This randomised controlled double-blinded study shows used [19]. Although there would have been an element of rotational inaccuracy in some films, the same form of x-ray was used in the entire series. Accuracy of the method was suggested by the fact that the 0˚ group averaged 0.2˚ on x-ray, confirming that cutting a 0˚ slope indeed gives that angle. However, cutting a 4˚ slope gave an average of 2.2˚ radiographically. Both surgeons obtained similar results with a mean tibial slope of 2.4 and 2.1 radiographically indicating it was not a single surgeon technical error that led to a lower tibial slope than expected with the cutting jig. The differences between intra-operative and radiographically measured tibial slope may be explained by subtle inaccuracies in placing the jig on the anterior tibia; any internal or external rotation would lead to a 4˚ slope but not in the true sagittal plane, resulting in a reduced slope.
Inaccuracies in radiographical measurement of slope will also occur with variation depending on what technique is used. According to the cadaveric work by Bellemans even with an increase of 2˚ tibial slope we would still expect to see an additional 3.5˚ to 4˚ of flexion.
This was not mirrored in our results. Despite possible inaccuracies in radiographic measurement the slope obtained, when using this surgical technique no difference in postoperative range of movement and patient reported outcomes are seen. Ismalilidis published work to suggest that differences in the posterior tibial slope should not contribute to a rotational malalignment if the anatomical tibial axis was used to align an arthroplasty [20].
Cutting the tibial slope parallel to the physiological preoperative slope rather than to a pre-determined angle has been suggested to result in better soft tissue balancing and more 'normal' knee kinematics. As this study was designed to look for a difference in post-operative range of movement and outcomes scores using a widely accepted technique used by many surgeons of pre-operatively planned tibial slope of 4˚ or 0˚, the better softtissue balancing in a 0˚ knee may explain why we see no difference in flexion or patient-related outcome between the two groups.
Comparison of small and large tibias also failed to yield any difference in movement. A posterior tibial slope in a larger tibia will cause a bigger drop in posterior tibial height than in a small tibia, hence delaying contact between posterior tibia and femur during deep flexion. Theoretically this should lead to increased flexion. However, we failed to reveal a significant difference in flexion between subgroups with similar body mass index. This may be explained by Nagamine's work who showed that in 208 patients undergoing total knee replacement, the proximal tibial condyle did not have a posterior slope in 86.5% of patients [21]. There showed that the posterior slope occurs in the proximal metaphysis of the tibia and that the tibial condyle rotates posteriorly. Hence the posterior tibial slope is mainly created by the poser rotation of the tibial condyle.
Our results should be extrapolated to other knee replacement designs with caution. Posterior stabilised designs show different kinematics from cruciate-retaining implants, while subtle differences in design in both cruciate-retaining and cruciate-sparing implants may also affect femoro-tibial component congruity, influencing kinematics with a suggested optimal posterior tibial slope ranging from 0˚ to 10˚ [22,23]. However, the relatively deep dish design of the Profix prosthesis used in this study is probably more prone to impingement than less conforming knees. More modern designs may negate these issues, in particular if they are intended to achieve more normal kinematics. Hence, as in all trials, each implant should be assessed individually. It should also be noted that this study is limited to a single site and patient population. The vast majority of our patients are White British; other tibial slope values are seen in different populations and ethnic groups [24].
Given that the influence of the posterior tibial slope can only be demonstrated in a cadaveric setting, in vivo studies have failed to demonstrate a correlation and therefore the clinical significance of the tibial slope within a narrow therapeutic window may be negligible.