A comparison of two different navigated hip replacement techniques on leg length discrepancy

Abstract

We hypothesised that a femoral array placed into bone or an external (pinless) reference marker made no difference to leg length discrepancy in patients undergoing navigated total hip arthroplasty. Consecutive patients undergoing navigated total hip arthroplasty. 162 patients. No statistical difference between preoperative leg length discrepancy (p = 0.524). Mean intraoperative change was 3.7 mm and 4.6 mm (p = 0.262). The mean change in leg length measure post operatively was 4.2 mm and 4.1 mm (p = 0.656). No significant difference in leg length discrepancy between a pinless reference markers and a femoral array placed into the bone.

1. Introduction

Total hip replacement is one of the most successful orthopaedic operations with high patient satisfaction and low revision rates.1, 2, 3 However, there is a constant drive to improve outcomes for patients and reduce revision rates. There is a wealth of literature associating poor component orientation with an increased risk of dislocation and increased wear in hip arthroplasty.4, 5, 6, 7 Increased wear rates can lead to earlier revision.8, 9, 10, 11, 12, 13, 14, 15, 16

The aim to improve outcomes in hip arthroplasty led to the development of computer navigated hip replacements to optimise component positioning and alignment. The effect of these controls is that leg length can be more accurately controlled and adjusted according to the patient. Multiple randomised controlled studies have shown that computer navigation improves component position and reduces leg length discrepancy in hip arthroplasty.17, 18, 19, 20 A meta-analysis has shown computer navigation improved component position (reduced the number of outliers) and reduced dislocation rates.20, 21, 22, 23, 24, 25, 26

Navigation requires the placement of arrays to determine component alignment and position. These arrays were previously placed on pins that were drilled into bone for accurate positioning. One of these arrays was placed in the distal femur. These pins have the theoretical risk of infection, fracture and post-operative discharge.27, 28, 29, 30 This led to the development of a pinless reference system which uses external reference markers that are placed on the skin. This prevents the need for temporary fixation into bone through the skin. This has been demonstrated to have good assessment of offset and leg lengths in cadaveric studies.²⁷ To our knowledge there has been no “in vivo” study to ensure this pinless system had no adverse effect on the accuracy of navigation on leg length. We hypothesised that a femoral array placed into bone or an external reference marker made no difference to leg length difference in patients undergoing total hip arthroplasty. The aim of the study was to demonstrate that the leg length measured using radiographs was the same as that measured by the navigation system

2. Method

All patients were retrospectively selected from the senior author’s data base. 100 consecutive patients were selected from each group − femoral pin and pin-less femur. The author initially performed procedures with femoral pins and the last 100 patients the author performed prior to changing his technique were included. The first 100 consecutive patients undergoing the procedure using pin-less femur array were included for this group. Patient demographics were recorded. All results were tabulated using Microsoft excel (Redmond, Washington: Microsoft, 2007).

Radiographs were viewed using the OrthoView (Martinsville, USA) templating system. Preoperative LLD was determined as the difference in distance (in millimeters) from the trans-ischial line to the centre of the lesser trochanter of each leg.³¹ Both groups were measured independently by two different members of the research team. Inter and intra-observer correlations were calculated.

The first available post-operative antero-posterior pelvic radiograph was then used to determine the LLD. Owing to a lack of recorded magnification on these radiographs, the software’s marker calibration tool was used to correct for magnification error by using the recorded size of the implanted femoral head. LLD was then calculated in the same manner as before.

The absolute difference between the intraoperative recorded change in leg length and the postoperative radiographically measured leg length was then used to give a “measurement error” between the two differing navigation techniques. The direction of length change was not considered (ie leg lengthened or shortened) as it was the relative accuracy of the navigation systems that was being investigated.

2.1. Exclusions

30 (23 pinless femur, 7 femoral pin) radiographs of patients were deemed of inadequate quality to allow accurate measurement and were excluded. Reasons for inadequacy included; pelvic obliquity, limb abduction or adduction, and/or immeasurable lesser trochanter due to rotation of the femur. 8 patients were excluded due to predicted operative complexity owing to conditions such as developmental dysplasia of the hip or significant avascular necrosis.

A power study was performed based on previously published data.³² This confirmed that the sample size obtained allowed detection of 5 mm difference in LLD with power >0.8 (f(α,β) = 7.85).

2.2. Statistics

SPSS (IBM version 20) was used. Data distribution was assessed using a quantile–quantile (Q-Q) plot and found to be normally distributed. Analysis of variance (ANOVA) was used to compare means between groups. Intra-observer error was calculated using intra class correlation (Cronbach’s alpha).

3. Results

All patients in the pinless group were operated on from August 2012–August 2013 when the senior author’s practice changed. All of the control (femoral pin) group were operated on between June 2009 and January 2012 (Table 1).

Table 1.

Results from radiograph measurements.

	N	Range	Minimum	Maximum	Mean	Std. Deviation
Pre op Pinless (mm short)	75	12.00	0.00	12.00	3.9870	3.18092
Pre op Control (mm short)	87	14.00	0.00	14.00	4.7931	3.48464
Pinless Intraop change	75	11.00	0.00	11.00	3.7143	2.45409
Control intraop change	87	14.00	0.00	14.00	4.6782	3.47581
Orthoview pinless (Post op)	75	13.00	0.00	13.00	4.2857	2.95518
Orthoview control (Post op)	87	12.00	0.00	12.00	4.1136	3.00357
Error pinless	75	10.00	0.00	10.00	2.5974	2.40208
Errorcontrol	87	12.00	0.00	12.00	4.0000	2.93337

Following exclusions, a total of 162 patients (75–pinless and 87 control) were included. The indication for surgery in all patients was hip pain secondary to osteoarthritis. All operated legs were radiographically short preoperatively (see Table 1) in comparison to the contralateral side. All results were normally distributed (Q–Q plot). There was no statistical difference between either group’s preoperative leg length discrepancy (p = 0.524).

The mean intraoperative change in the pin and pinless group leg length was 3.7 mm (0–11) and 4.6 mm (0–14) respectively. There was no statistical difference in intraoperative leg length change between groups (p = 0.262).

The mean change in leg length measure post operatively on OrthoView (Martinsville, USA) was 4.2 mm (0–13) (Pinless group) and 4.1 mm (0–12) (Pin group). There was no statistical difference in leg length discrepancy between groups (p = 0.656).

The mean error recorded for the pinless group was 2.59 mm (0–10) compared to 4.00 mm (0–12). Whilst the standard deviation for the pinless group was marginally narrower than for the pin group (2.40 mm vs 2.93 mm) there was no significant difference found between the two groups (p = 0.664).

There were no statistical outliers found in either cohort. Box and whisker diagrams were generated for the data sets using the explore function of SPSS.

3.1. Intraclass correlation (ICC)

To determine intraclass correlation (ICC) 40 (20 pin, 20 pinless) postoperative radiographs were randomly selected from the existing pool of patients and re-measured by the opposite observer. ICC was then calculated (two- way mixed, absolute agreement). Cronbach’s alpha was found to be 0.959 (p < 0.01) showing good interobserver error.

4. Discussion

A common complication with total hip replacements is unacceptable leg length discrepancy. Leg length discrepancy can lead to gait asymmetry, abnormal joint reaction force across the joint and an increased risk of revision.³³ Leg length discrepancy can also be a cause for litigation.³⁴

In the general population up to a third of post THR patients can have a leg length discrepancy of up to 20 mm yet remain asymptomatic.33, 35 The amount of leg length discrepancy following a total hip replacement that leads to symptoms is debated. Generally a leg length discrepancy of less than 1 cm is widely accepted.³⁵ Symptoms from leg length discrepancy may present in one of three ways. These include neurological manifestations, postoperative instability and most commonly a combination of gait disturbance, periarticular muscle spasm, fatigue and groin pain.³⁵

For all these reasons any mechanisms that can reduce unintentional post operative leg length discrepancy should be explored to improve patient outcomes.

There are numerous different techniques that surgeons use to attempt to prevent leg length discrepancy and improve accuracy of the prosthesis alignment. These include pre-operative templating and a wide range of intra-operative techniques, such as measurements from a fixed point with a suture or the insertion of a Steinmann pine.³⁵ None of these have consistently demonstrated a reduction in leg length discrepancy.³⁵

Navigated hip replacements use intra-operative references and measurements to improve the accuracy of leg length discrepancy. Measurements are recorded during the operation, using arrays on the femur and pelvis. The femoral array was initially drilled into the distal to femur to have a constant femoral marker for determining leg length. A pinned femoral array has potential complications such as infection and fracture.27, 28, 29, 30 Due to this a pinless femoral array was developed. The accuracy of this was demonstrated by Renkawitz et al on a cadaveric study measuring leg length discrepancy,²⁷ and on a previous study demonstrated good accuracy in measuring leg length and offset.³⁶

In this study we found no statistical difference in intraoperative leg length change between groups or in leg length discrepancy between the groups. This therefore confirms that in navigated total hip replacements there is no significant difference in leg length discrepancy between a pinless femoral array and an array that required pin placement in the femur.

One known benefit of navigated total hip replacements is the reduction in potential outliers via more accurate control of leg length. Our data would support this view given that no statistical outliers were found.

The limitations for this study are that it is a single surgeons operative results and further study would benefit from analysis data from multiple surgeons. A further weakness is that pinned femoral array was used first. It could therefore be argued that the senior authors surgical accuracy may have improved in this time as he had performed more procedures. Whilst there was no randomisation in this study, the fact that there was no time period overlap between the two groups will have prevented any bias in patient selection. There were 30 patients that were excluded due to inadequate post-operative radiographs due to malrotation and alignment of the image. The main limitation of this study is the inaccuracy of using plain radiographs to assess leg length difference. Heaver et al demonstrated that the technique used in this study was the most accurate method of measuring leg length from plain radiographs and yet had an error of 6 mm increasing to 12.5 mm when there is malpositioning.³¹ Therefore the error within our measurement technique may exceed the error from the navigation system. Whilst pre and post operative CT measurements may be the most accurate modality the radiation exposure and financial restraints of our health system limit their availability.

5. Conclusion

Our results suggest that there was no significant difference in leg length discrepancy between a pinless femoral array and an array that required osseous pin placement. We recommend the use of a pinless femoral array system in navigated total hip replacement that avoids associated issues with pinned arrays without compromising outcome.

Conflict of interest

None.

Mr ED Davis works as a Consultant for Brainlab.

No acknowledgements or other conflicts of interest.