Intraoperative Radiographs for Placing Acetabular Components in Hip Resurfacing Arthroplasty

Abstract

Background

Various clinical and biomechanical studies suggest certain acetabular positions may be associated with higher wear and failure rates in modern metal-on-metal hip resurfacing arthroplasties. However, there are no widely available, reliable, and cost-effective surgical techniques that ensure surgeons are able to place an acetabular component within the safe range of inclination angles after hip resurfacing surgeries.

Questions/purposes

We investigated the accuracy of intraoperative radiographs to determine the acetabular inclination angle in resurfacing arthroplasty procedures.

Patients and Methods

The study group included the first 100 resurfacing arthroplasties performed after we started routinely checking the intraoperative acetabular inclination angles. The acetabular component was repositioned if the intraoperative acetabular inclination angle was out of the target range of 30° to 50°. The control group included the previous 100 resurfacing arthroplasties performed without the benefit of intraoperative radiographs. A posterior minimally invasive surgical approach was used in both groups. Demographics and diagnoses were similar in both groups.

Results

The average (± SD) difference between the intraoperative and 6-week radiographs was 2.7° ± 2.5°. The acetabular inclination angles at 6-week followup were within the targeted range more frequently in the study group than in the control group (outliers: 4% versus 29%).

Conclusions

These data suggest a single intraoperative radiograph is a quick, reliable, and cost-effective method for ensuring the acetabular inclination angle is within the targeted range.

Level of Evidence

Level III, therapeutic study. See Guidelines for Authors for a complete description of levels of evidence.

Introduction

The use of hip resurfacing arthroplasty (HRA) has been resurgent with the development of modern designs of metal-on-metal bearings since the late 1990s [2, 4, 7, 8, 11, 15, 21, 28]. Recent studies reported revision rates of 1% to 5% at an average of 5 to 7 years in young and active patients after HRA [2, 15, 21, 28]. Even though femoral neck fracture and femoral loosening were formerly reported as the most common failure modes in these studies [2, 4, 21], revision related to the acetabular component is being recognized more frequently in recent reports [8, 11]. Previously, higher dislocation rates and increased wear rates were well-documented in standard THA with improper acetabular inclination angles [12, 20, 22, 31]. Although large metal bearings are resistant to dislocation even with relatively high angles of inclination, several clinical studies have implicated high cup inclination angles in wear-related failures after HRA [4, 13, 33]. One biomechanical analysis suggested higher wear rates after two million cycles when acetabular component inclination angles were greater than 55° in 39-mm metal-on-metal bearing hip implants [31]. Also, while several studies [3, 8] suggest the upper limit for ideal inclination angle for large metal bearings is approximately 55°, we found no data suggesting a lower limit.

Many methods, including visual inspection [32], use of specialized jigs [23], intraoperative fluoroscopic guidance [21], image-free navigation [16, 24], and CT-based computer navigation [10, 17, 27], reportedly assist surgeons in achieving the safe target during the placement of acetabular components. However, all of these techniques have disadvantages, such as a lack of accuracy, high cost, a requirement for special training, and extra preparation and operating room (OR) time. There are no widely available, reliable, and cost-effective surgical techniques that ensure surgeons are able to place an acetabular component within the safe range of inclination angles after HRAs.

We therefore (1) asked whether a single intraoperative AP pelvis radiograph would ensure the inclination angle of an acetabular component was in a target range of 30° to 50° during a metal-on-metal HRA; (2) if so, asked whether the intraoperative assessment of inclination was accurate when compared to the standard postoperative supine AP pelvis radiographs and whether this method was time-consuming compared to our routine surgical protocol; and (3) developed guidelines to prevent outliers with acetabular inclination angles out of the target range of 30° to 50° after HRA.

Patients and Methods

The senior author (TPG) had performed more than 1100 metal-on-metal HRAs and had already focused on placing acetabular components with acetabular inclination angles within the range of 30° to 50° before this study. He began routinely using plane radiographs to check the inclination angles of the acetabular components during HRA procedures in November 2007. The study group included the first 100 HRAs performed with the aid of an intraoperative radiograph between November 2007 and March 2008. To determine whether this technique led to an improvement in component positions, the study group was compared to a control group consisting of the last 100 HRAs performed between July 2007 and November 2007, just before the protocol of an intraoperative radiographic check was started. No patients were excluded from this study for any reason. For the study group, all inclination angles of acetabular components on the intraoperative (intraoperative acetabular inclination angle [IOIA]), immediate recovery room (recovery room acetabular inclination angle [RRIA]), and 6-week followup (6-week followup acetabular inclination angle [SFIA]) radiographs were prospectively measured by the same author (LW). The radiographic records (RRIA and SFIA) of the control group were retrospectively reviewed. All of the selected cases had good-quality AP pelvis radiographs in the recovery room and at their 6-week followup for analysis of the acetabular inclination angle. We defined a possible cup shift when a measured change in component position between SFIA and RRIA was greater than the average difference plus three SDs of variability. There were no differences between the two groups with respect to demographic and clinical parameters (Table 1).

Table 1.

Demographic and intraoperative data for both groups

Variable	Study group	Control group	p Value
Number of hips	100	100
Age (years)*	51 ± 7	52 ± 9	0.70†
Gender			0.44‡
Women	29	32
Men	71	68
Diagnosis			0.20‡
Osteoarthritis	73	69
Dysplasia	18	13
Posttrauma	4	5
Avascular necrosis	3	9
Other	2	4
Component size (mm)*
Acetabular	56 ± 4	56 ± 4	0.81§
Femoral	49 ± 6	50 ± 4	0.72§

* Values are expressed as mean ± SD; all other values are expressed as number of hips; ^†the data are normally distributed; the standard t test was used here; ^‡the chi square test was used here; ^§the data are not normally distributed; a Wilcoxon-Mann-Whitney test was used here.

All HRA procedures were performed (not supervised) by the senior author (TPG) with the same first assistant (LW) through a posterior minimally invasive approach. The Recap^® femoral resurfacing component (Biomet, Inc, Warsaw, IN) was combined with the Magnum™ acetabular component (Biomet, Inc) in all cases. Patients were placed in the lateral decubitus position using a standard OSI positioning system (OSI-Orthopaedic Systems Inc, Union City, CA) consisting of a set of four pads clamped to the OR table to hold the patient firmly in position. The radiology staff was called when acetabular reaming had begun. In both groups, the acetabular bone was underreamed by 1 mm. A 1 mm smaller smooth trial component was placed to assess the appropriateness of reaming. Then the component was impacted. No alignment guides were used. If the component position was not perfect by visual criteria, an edge tamp was used and the cup was impacted again with a secondary impactor. If this method was unsuccessful, the cup was removed and reimpacted with the primary impactor. In no cases did we change to another component or use supplementary fixation. In the study group, as soon as the acetabular component was impacted, a portable crosstable AP pelvis radiograph was obtained. A sterile drape was clamped to the front of the patient, and the portable radiograph unit was moved into place. The anesthetist tilted the OR table to align the anterior plane of the pelvis visually perpendicular to the floor according to the surgeon’s best estimate. The xray beam was angled approximately 10° caudal. The xray plate was placed on a stand posterior to the patient, oriented perpendicular to the xray beam, and covered with a sterile sheet. While the radiograph was being developed, the supplemental sterile drape was removed, the femoral component was implanted, and closure was started to avoid unnecessary delay. The film was immediately developed and hung on the OR view box. The inclination angle of the acetabular component was measured by drawing a line across the inferior pubic rami and another line across the opening face of the acetabular component [6] (Fig. 1). The acetabular inclination angles were measured on the intraoperative radiograph by an assistant (LW) and immediately reviewed by the senior author (TPG). The component position was not changed if the inclination angle was in the target range of 30° to 50°. If the component required adjustment, the process of repositioning occurred again. A new radiograph was then ordered in the recovery room. The extra OR time required to obtain this intraoperative radiograph was measured in 20 consecutive cases.

Fig. 1.

Measurement of the acetabular inclination angle (α) is shown.

Twenty HRAs were randomly selected and their 6-week followup radiographs were measured three times each by two observers (FL, LW) to evaluate the intraobserver and interobserver variability of this technique. One author (FL) did not have any experience measuring acetabular component positions on radiographs before this study. The average absolute interobserver variability (± SD) was 2.4° ± 3.0° (range, 0.4°–5.5°), and the average absolute intraobserver variability (± SD) was 0.9° ± 0.3° (range, 0.4°–1.6°). The average correlation coefficient value was 0.89 (95% confidence interval, 0.74–0.95) for intraobserver variability and 0.99 (95% confidence interval, 0.98–0.99) for interobserver variability (correlation coefficients range from 0 to 1, with 0 representing no agreement and 1 representing perfect agreement [1]).

We evaluated the normal distribution of numeric parameters with the use of Shapiro-Wilk W tests. If the tested parameters were normally distributed, we determined the differences between the study and control groups with the use of standard t tests; otherwise, we evaluated the differences with the use of nonparametric Wilcoxon-Mann-Whitney tests. We evaluated the differences of categorical parameters between the two groups with the use of chi square tests. To assess the differences among IOIA, RRIA, and SFIA in each group, we calculated the mean and SD for each of them. A univariable linear regression was performed based on a standard least-squares technique with a 95% confidence interval to determine the target range of IOIA, which ensures SFIA falls into the safe range of 30° to 50°. Cases where the acetabular component was determined to be possibly shifted were excluded from the data used to build the model. All statistical analyses were performed with the use of OrthoTrack (Midlands Orthopaedics, PA, Columbia, SC), except the regression analysis, which was generated with the use of JMP (SAS Institute, Inc, Cary, NC). Institutional Review Board approval was obtained for this study.

Results

There were more cases (p < 0.001) with an acetabular inclination angle within the target range in the study group than in the control group, both on the immediate recovery room and on the postoperative 6-week followup radiographs (Table 2). In the study group, there was one case with a RRIA of more than 50° (51°) and four cases with a SFIA of less than 30° (range, 20°–26°). In the control group, there were 16 cases with a RRIA of more than 50° (range, 51°–57°) and one case with a RRIA of less than 30° (29°); at the 6-week followup, there were 29 outliers with a SFIA of more than 50° (range, 51°–59°).

Table 2.

Intraoperative, recovery room, and 6-week followup acetabular inclination angles between the study and control groups

Variable	Study group			Control group
	Average ± SD	Range	Missing rate	Average ± SD	Range	Missing rate
IOIA	39° ± 5°	30°–50°	0%	NA	NA	NA
RRIA	39° ± 4°	30°–51°	1%	46° ± 5°	29°–56°	17%
SFIA	39° ± 5°	20°–50°	4%	47° ± 5°	51°–59°	29%

IOIA = intraoperative acetabular inclination angle; RRIA = recovery room acetabular inclination angle; SFIA = 6-week followup acetabular inclination angle; NA = not applicable.

For the study group, the average absolute difference (± SD) was 2.5° ± 2.0° (range, 0°–10°) between IOIA and RRIA, 2.7° ± 2.5° (range, 0°–14°) between IOIA and SFIA, and 2.1° ± 1.8° (range, 0°–10°) between RRIA and SFIA. For the control group, the average absolute difference (± SD) between RRIA and SFIA was 2.1° ± 1.7° (range, 0°–13°). The average additional intraoperative radiograph time (± SD) during which the operation was delayed was 116 ± 28 seconds per case.

In the study group, the acetabular component appeared shifted between the immediate recovery room and the 6-week radiographs in two cases (Table 3). Both of these acetabular components were now out of the target range, with SFIAs of less than 30°. These two cases were excluded from the data to generate the regression model. To achieve a SFIA in the target range of 30° to 50° with a 95% confidence interval, the surgeon had to keep the IOIA in the range of 32° to 46° according to the regression model (R² = 0.52) (Fig. 2). In the control group, the acetabular component was suspected of having shifted in one case. It shifted to a more inclined position outside of the target range (Table 3). There was no difference (p = 0.56) in the incidence of acetabular component shifting between the two groups. All three patients with shifted acetabular components were doing well with a stable component at their minimal 2-year followup visit. There were no acetabular revisions or dislocations in either group. One femoral component was revised in each group due to femoral neck fracture. The acetabular component was left in place in both cases. The patients were doing well at the latest followup. Because the acetabular components were not revised, they were not excluded from this study.

Table 3.

Possible asymptomatic acetabular component shifts

Case	IOIA	RRIA	SFIA	Latest Harris hip score*
Study group
1	38°	33°	24°	100
2	45°	36°	20°	82
Control group
1	NA	51°	59°	100

* Minimal 2-year followup; IOIA = intraoperative acetabular inclination angle; RRIA = recovery room acetabular inclination angle; SFIA = 6-week followup acetabular inclination angle; NA = not applicable.

Fig. 2.

A graph shows the regression model fitting of the acetabular inclination angles intraoperatively versus at 6-week followup (R² = 0.52). Black dots = original data points; Lines a and a′ = 95% confidence limits of an individual predicted value; Lines b and b′ = 95% confidence curves of predicted values.

Discussion

It appears likely hip arthroplasty based on large metal-on-metal bearings will have increased metal ion levels and higher rates of wear-related failure with higher acetabular inclination angles [20, 21, 31]. Previously, a few studies were directed at proper femoral component stem orientation in HRA [19, 26, 27]. However, it now appears proper placement of the acetabular component may be even more critical to the success of this procedure [8, 9, 11, 21, 30]. We investigated the value of using a single intraoperative portable crosstable AP radiograph to improve the accuracy of placing the acetabular component within a target range of inclination.

Our study is associated with some limitations. First, the data analysis and the regression model in this study were based on cases from only one surgeon who had performed a large number of HRAs; more variability would likely occur when this technique is used by less experienced surgeons. Second, we could not determine whether cup repositioning led to asymptomatic cup shifts because we were not aware of the problem of possible cup shifts during this study and we did not record when and how often cups were repositioned in both groups. It is possible there were more cups repositioned based on unacceptable radiographic criteria in the study group. Uncertainty of cup position without a radiograph could have led to more cup changes in the control group as well. We found no difference in the incidence of possible cup shifts between the two groups. More accurate standardized radiographic and accurate intraoperative logs of details of cup insertion and repositioning should be required in the future studies to answer these questions. Third, we could not determine an acetabular anteversion angle. We are unaware of any rigorously established ideal target criteria for acetabular anteversion position in HRA. Until further data become available, we believe it is reasonable to keep the acetabular inclination angle of 30° to 50°, with the anteversion determined intraoperatively by surgeons to maximize the ROM and to avoid impingement in both flexion and extension.

A “safe” zone of acetabular component position was suggested for small-bearing metal-on-polyethylene THA in a retrospective study [20]. Although this study had some weaknesses, the safe range of 30° to 50° suggested in that study has been generally accepted as a gold standard by surgeons [14, 16, 25], which was also targeted in our study. An AP pelvis radiograph remains a universal tool for assessing acetabular inclination angles after hip arthroplasties in the United States. Generally, the inaccuracy in radiographic measurements of inclination angles is introduced when patients are lying on their side on an OR table. In this study, we found our overall frequency of ending up in this desired range at 6-week followup was improved after we began checking the intraoperative radiograph (4% versus 29%). Thus, our data suggest, despite the disadvantages of using plane radiographs, the intraoperative radiograph can still provide enough information to ensure the inclination angle of an acetabular component was in a target range of 30° to 50° during a metal-on-metal HRA.

Normally, it is difficult to place an acetabular component within the target range in one attempt for every case without any assistance. Numerous studies have documented surgical navigation holds the promise for providing most accurate intraoperative guidance for acetabular component position [10, 16, 29, 34]. However, the equipment is not available to most surgeons. Two German studies performed cost-saving analyses and reported the incremental cost of using a navigation system in joint arthroplasties ranged from 300 Euro to 442 Euro per case [5, 18]. Also, surgical navigation is usually more time-consuming. A study reported an additional 10 to 15 minutes of OR time was needed to install the image-guided navigation system [10]. Comparatively, our technique required less than 2 minutes of additional OR time in our study. On the other hand, when acetabular inclination is evaluated, errors may be added up due to data calibration and data acquisition by introducing navigation system (Table 4). Furthermore, measurements using current navigation systems are based on the anterior pelvic plane and cannot be directly compared with measurements made on conventional supine or standing radiographs. These differences require further complicated translations that may result in additional possible errors [22]. Fluoroscopy has also been suggested in metal-on-metal HRA [21]. The major disadvantage is the inability to actually perform a measurement of an inclination angle and a surgeon has to make a visual estimate of the angle because the whole pelvis cannot be captured on one film. Intraoperative radiographic evaluation does not allow guided placement; it only allows checking a final position. To avoid repositioning an acetabular component, another option would be to implant a trial component, obtain a radiograph, and then implant an actual component. This might avoid changing the actual cup position. It is unknown whether reorienting an acetabular component influences the durability of the implant. There may be a trade-off between accepting a suboptimal position and accepting some risk of cup shift or failure of ingrowth due to the process of repositioning the component. Further studies will be required to answer these questions.

Table 4.

Comparison of accuracy among different studies

Study	Method	Absolute deviation
Kalteis et al. [16] (2005)	Image-free navigation system	2.3° ± 2.8°
Kalteis et al. [16] (2005)	Free hand	6.2° ± 7.0°
DiGioia et al. [10] (1998)	Image-guided navigation system	3.7° ± 2.6°
Ybinger et al. [34] (2007)	Imageless navigation system	3.5° ± 4.4°
Gross et al.	Intraoperative radiograph	2.7° ± 2.5°

Values are expressed as mean ± SD.

We suspect surgical navigation technology will eventually emerge as the most accurate and reproducible method of placing acetabular components with three-dimensional accuracy. However, several obstacles remain. First, it is not clear they are currently more accurate than cheaper and quicker methods. Second, substantial additional costs for these advanced technologies may not be supported in the current climate of cost containment. Therefore, we currently suggest the use of a single portable intraoperative radiograph as an inexpensive, quick, and widely available method to improve the frequency of placing the acetabular component within the desired range of inclination angles in metal-on-metal HRA. Also, we believe this technique could be applicable to any other metal-on-metal hip arthroplasty.