A correction factor is needed for the intraoperative measurement of cup inclination in total hip arthroplasty

Abstract

BACKGROUND:

The cup inclination in total hip arthroplasty is key to minimising complications. Stereometric effects (parallax) in two-dimensional projections can alter angle measurements. Even for different approaches, fluoroscopy causes different results in the measurement of inclination. A previous study has introduced a corrective factor for intraoperative radiographic cup inclination measurements compared to the postoperative standing radiographs.

OBJECTIVE:

The aim of this study was to find out whether, first, the correction factor is reproducible and second, whether the correction factor is independent of the surgical approach and C-arm model used.

METHODS:

A series of 377 cases of primary total hip arthroplasty was reviewed. We compared the cup inclination angle in the intraoperative and postoperative radiographic images. Based on this, it was possible to specify a standard of correction factor in defined ranges.

RESULTS:

The mean cup inclination in intraoperative images was 37.47 ${}^{\circ }$ and the mean angle in postoperative images was 41.42, resulting in a mean difference of 3.95 ${}^{\circ }$ with a strong correlation ( $r=$ 0.706).

CONCLUSION:

An added correction factor of 4 ${}^{\circ }$ should be respected in intraoperative inclination measurements to adapt for parallax. An increased correction factor of 6 ${}^{\circ }$ in particularly low-positioned cups ( $<$ 31 ${}^{\circ }$ ) and a decreased correction factor of 2 ${}^{\circ }$ in particularly steeply positioned cups ( $>$ 43 ${}^{\circ }$ ) is recommended.

1. Introduction

The inclination in primary total hip arthroplasty (THA) is key to minimising complications and guaranteeing an optimal fitting of the arthroplasty into the natural pelvis. In general, cup inclination is oriented from the trajectory of both tear-drop figures of the pelvis. The targeted cup positioning is within the so called Lewinnek’s safe zone with an inclination of 40 ${}^{\circ }$ $\pm$ 10 ${}^{\circ }$ and an anteversion of 15 ${}^{\circ }$ $\pm$ 10 ${}^{\circ }$  [1]. Malpositioning, which means an suboptimal orientation of the acetabular component outside of the described safe zone, leads to notable complications such as impingement, early polyethylene wear or dislocation [2, 3, 4].

The use of fluoroscopy is controversial in primary THA, as fluoroscopy increases the surgical time and costs [5]. However, intraoperative X-ray analysis gives surgeons the possibility to correct implant orientation to desired angles. Thus, using intraoperative fluoroscopy can significantly improve cup positioning and inclination as previous publications have shown.

Nevertheless, stereometric effects (parallax) in two-dimensional projections can alter angle measurements [6]. Even for different surgical approaches, fluoroscopy causes different results in the measurement of inclination [7]. Although exact measurements or calculation of angles are not very common in daily practice, the above-mentioned considerations must be taken into account, when the angles of intraoperative fluoroscopy pictures are estimated by the surgeon and his team. A previous study has introduced a corrective factor for intraoperative radiographic cup inclination measurements (RI) compared to the postoperative standing radiographs [8]. In the mentioned study a correction factor of 5 ${}^{\circ }$ was determined for the direct anterior approach (DAA) when using the 12-inch GE 9900 elite C-arm (GE Healthcare, Chicago, IL) for intraoperative fluoroscopy. The aim of this study was to find out whether, first, the correction factor is reproducible in a new data set, obtained in a different surgical center and, second, whether the correction factor is independent of the surgical approach and C-arm model used.

2. Material and methods

We retrospectively reviewed cases of primary total hip arthroplasty performed in the period of 2013 to 2017 in the university hospital Cologne. Only cases of primary osteoarthritis were eligible, any revision surgery or different cup fixation (e.g. screw) techniques were excluded. Of the 516 eligible patients, 113 were excluded due to alternative cup fixation. Further 34 patients were excluded due to inappropriate imaging, leaving 377 patients with complete images and data. All patients included in the study underwent THA with an uncemented acetabular component (Pinnacle, DePuy Synthes, Warsaw, IN, USA) and with either a cemented or uncemented stem. On all patients the modified lateral approach to the hip was performed. The study was approved by the institutional ethics committee (EN 21-1038).

2.1 Radiographic evaluation

A Philips BV Endura (Philips Medical Systems DMC GmbH, Hamburg, Germany) Image Intensifier was used for intraoperative fluoroscopy. The patient was in a supine position and the detector of the C-arm was placed as close to the patient as possible to reduce radiation. Images were taken centred over the symphysis with symmetrical obturator foramina, thus both tear-drop figures were visible. All patients underwent standardised preoperative and postoperative weight-bearing anterior to posterior (AP) pelvic radiographs with the x-beam centred on the symphysis (source-to-film distance (SID) of 100 cm). In the underlying study, inclination is defined as radiographic inclination (RI). Radiographic inclination (RI) is defined as the angle $\rho$ between the long axis of the projected ellipse of the cup opening and a trajectory through both tear-drop figures. $\rho$ was measured in both the intraoperative images and postoperative a.p. pelvis radiographs, see Figs 3 and 4.

We performed double-observer measurements to determine in RI of the intraoperative fluoroscopy images (IRI) and of the postoperative standing a.p. radiographs (PRI). Measurements were performed using the institutional picture archiving and communication system (IMPAX EE; Agfa Healthcare N.V., Mortsel, Belgium). To check the internal consistency of the measurements, a reliability analysis was performed. The intraclass correlation coefficient (ICC) is determined as an indicator of reliability.

2.2 Statistical analysis

Descriptive statistics were performed to summarise the means, medians, and ranges for all variables. Paired t test (for normally distributed values) or the Wilcoxon-rank-sum test (for non-normally distributed values) was performed to identify significance. Statistical analysis was performed for a 95% confidence interval. Results with $p<$ 0.05 were considered as statistically significant results, whereas results with $p<$ 0.01 were considered as highly significant. Bivariant correlation analyses (e.g. angles of inclination in relation to the age, surgical side and sex of the patients) were performed using Pearson’s correlation ( $r$ ). A linear regression equation was used to calculate the angular deviation as a function of the intraoperative angle. Based on this, it was possible to specify a standard of correction factor in defined ranges. The power analysis showed that with a Pearson correlation of $r=$ 0.5 (effect size), a statistical power ( $1-\beta$ ) of 95%, and a significance level of $\alpha =$ 0.05, one would need a sample size of $n=$ 46 for a significant result. All statistical analyses were performed using SPSS Statistics software version 27.0.0.0 (SPSS Inc., Chicago, IL, USA).

3. Results

The study included 369 patients with a modified lateral surgical approach, 182 (49.32%) were male and 187 (50.68%) were female. At the time of surgery, the patients had an overall mean age of 64.25 years (range 18–90). Male patients had a mean age of 61.51 years (range 22–85) and female patients had a mean age of 66.97 years (range 18–90). This study analyzed 188 left-sided THAs and 181 right-sided THAs. The mean angle of IRI was 37.47 ${}^{\circ }$ (range 21–56 ${}^{\circ }$ ) and the mean angle of PRI was 41.42 ${}^{\circ }$ (range 25.8–62.2 ${}^{\circ }$ ), hence the mean difference of radiographic inclination was 3.95 ${}^{\circ }$ (DRI) with a strong correlation ( $r=$ 0.706). For the difference of DRI our intraclass correlation coeffivient (ICC) was calculated as 0.89, which can be interpreted as very good.

The inclination angles IRI and PRI and the difference between IRI and PRI (DRI) were analyzed using $t$ -tests for independent variables for side and gender related differences. Only IRI and DRI showed statistically significant side related differences ( $p<$ 0.05) and only PRI showed a statistically significant gender related difference of mean ( $p<$ 0.05). Table 1 summarises the side and gender related differences of the mean values of IRI, PRI and DRI.

Table 1

Side and gender related differences of the mean values of DRI

	Sex (men/women)	Side (right/left)
Mean DRI	3.62 ${}^{\circ }$ /4.27 ${}^{\circ }$	3.47 ${}^{\circ }$ /4.41 ${}^{\circ }$
Standard deviation (SD)	4.19 ${}^{\circ }$ /4.96 ${}^{\circ }$	4.63 ${}^{\circ }$ /4.54 ${}^{\circ }$
Significance ( $p=$ )	0.172	0.048
$N=$	182/187	181/188

Table 2

Values of intraoperative and postoperative inclination angles

	$N=$	Range	Minimum	Maximum	Mean	SD
Intraoperative radiographic inclination (IRI)	369	35.0 ${}^{\circ }$	21.0 ${}^{\circ }$	56.0 ${}^{\circ }$	37.466 ${}^{\circ }$	6.137 ${}^{\circ }$
Postoperative radiographic inclination (PRI)	369	36.4 ${}^{\circ }$	25.8 ${}^{\circ }$	62.2 ${}^{\circ }$	41.416 ${}^{\circ }$	5.854 ${}^{\circ }$
Difference of PRI and IRI (DRI)	369	37.8 ${}^{\circ }$	$-$ 13.3 ${}^{\circ }$	24.5 ${}^{\circ }$	3.950 ${}^{\circ }$	4.603 ${}^{\circ }$

Furthermore, patient age was examined for any relationship with IRI, PRI and DRI. Pearson’s correlation showed that IRI ( $r=$ $-$ 0.106) and PRI ( $r=$ $-$ 0.130) have a statistically significant but weak negative correlation with patient age ( $p<$ 0.05). There was also a very weak negative correlation between DRI ( $r=$ $-$ 0,025) and patient age which was not statistically significant ( $p=$ 0.633).

In another step, we added a correction factor (CF) to IRI and tested against PRI. We added 2 ${}^{\circ }$ ( $p⩽$ 0.000), 4 ${}^{\circ }$ ( $p=$ 0.835) and 6 ${}^{\circ }$ ( $p⩽$ 0.000), see Fig. 1. In a further subgroup analysis, the study cohort was divided into three subgroups (Subgroup 1 with IRI $<$ 31 ${}^{\circ }$ ; subgroup 2 with IRI $=$ 31–42 ${}^{\circ }$ ; subgroup 3 with IRI $>$ 42 ${}^{\circ }$ ) and tested again against PRI for each correction factor. No statistically significant differences were found between IRI and PRI in subgroup 1 with a CF of 6 ${}^{\circ }$ ( $p=$ 0.269), in subgroup 2 with a CF of 4 ${}^{\circ }$ ( $p=$ 0.974) and in subgroup 3 with a CF of 2 ( $p=$ 0.217).

Figure 1.

Correction of the intraoperative angle correlating to the postoperative inclination angle.

4. Discussion

Through the preliminary study by Rueckl et al., the idea of a correction factor (CF) in cup positioning during implantation of total hip arthroplasty was introduced [8]. We analyzed whether the correction factor is reproducible and, second, whether the correction factor is independent of the surgical approach and C-arm model used. The group found a deviation of 5 ${}^{\circ }$ between the intraoperative inclination measured by C-arm and the postoperative inclination measured by conventional radiographs. To determine a suitable correction factor (CF) in our data set, three scenarios of correction were investigated. We added 2 ${}^{\circ }$ , 4 ${}^{\circ }$ and 6 ${}^{\circ }$ to the intraoperative radiological inclination (IRI) of the overall study population and tested each of the three variables against the postoperative radiological inclination (PRI) using $t$ -tests for dependent variables. Only a correction of 4 ${}^{\circ }$ showed no statistically significant deviation when testing the corrected IRI against PRI ( $p=$ 0.835), which was interpreted as a successful correction (see Fig. 1). Thus, the correction factor determined by us deviates by 1 ${}^{\circ }$ from the correction factor calculated by Rueckl et al. We interpreted the deviation of our correction factor of just 1 ${}^{\circ }$ as a validation of the idea of a correction factor per se. It should be noted that the study results of Rueckl et al. originate from an endoprosthetic center with a very high level of standardization and data from operations were examined that were all performed by the same individual high-volume surgeon. In contrast, our data were collected at a German university hospital, where the procedures cannot be designed down to the last detail for endoprosthetic care alone and the operations were performed by several surgeons. This inevitably leads to a higher variance of surgical results and thus different data was drawn. Consequently, a different correction factor was calculated.

sDue to the distinct range of IRI among our study population (21 ${}^{\circ }$ to 56 ${}^{\circ }$ ), the question arose whether a general correction factor for the entirety of the intraoperative angles can be applied, or whether an adjustment of the correction factor should be made for the particularly high or particularly low inclination angles. Therefore, we divided the study population into three subgroups depending on the IRI. Subgroup 1 consisted of IRI $⩽$ 31 ${}^{\circ }$ , subgroup 2 consisted of IRI 32–42 ${}^{\circ }$ , and subgroup 3 included IRI $⩾$ 43 ${}^{\circ }$ . Each subgroup was tested separately using $t$ -tests for dependent variables for any significant differences between the three corrected IRI and PRI. No statistically significant differences were found between IRI and PRI in subgroup 1 with a CF of 6 ${}^{\circ }$ , in subgroup 2 with a CF of 4 ${}^{\circ }$ and in subgroup 3 with a CF of 2 ${}^{\circ }$ ( $p=$ 0.269, $p=$ 0.974 and $p=$ 0.217 respectively), which again, was interpreted as a successful correction of the IRI. This was further supported by the fact that conclusions about the extent of angular correction were possible on the basis of the linear regression equation. The larger the intraoperative angle, the smaller the difference between the inclination angles (see Fig. 2).

Figure 2.

Difference in angular inclination depending on the intraoperative angle.

It is known that different C-arm models produce different parallax effects. The correction factor for the 12-inch GE 9900 elite C-arm (GE Healthcare, Chicago, IL) was calculated as 5 ${}^{\circ }$ , while we calculated a correction factor of 4 ${}^{\circ }$ when using the Philips BV Endura (Philips Medical Systems DMC GmbH, Hamburg, Germany). The difference in the calculated correction factors for the two different C-arm models appears irrelevant when transitioning the results into clinical practice. In addition, it can be assumed that the influence on the difference in the data due to the device-specific parallax effects is likely to be significantly less than the influence of the operator-dependent variance in the postoperative results.

Figure 3.

Intraoperative measurement of cup inclination angle.

Figure 4.

Postoperative measurement of cup inclination angle.

In the study by Rueckl et al., the direct anterior approach (DAA) was used as the only surgical approach. In contrast, the modified lateral approach was used as the only surgical approach in our study. Since an almost identical factor was found in both studies, we interpreted this as being independent of the surgical approach.

In synopsis, the consideration of a correction factor of about 4 ${}^{\circ }$ to 5 ${}^{\circ }$ is recommended. Especially in clinics with a certain variability of cup positioning, depending on different surgeons, an individual correction factor is recommended, with an increased correction factor of 6 ${}^{\circ }$ in particularly low-positioned cups and a decreased correction factor of 2 ${}^{\circ }$ in particularly steeply positioned cups.

5. Conclusions

We were able to successfully reproduce the correction factor as described in the preliminary study by Rueckl et al. and thus substantiate the relevance of an intraoperative correction factor for optimal cup positioning during implantation of a total hip arthroplasty.

While it was calculated for the direct anterior approach before, we have shown that the correction factor is also valid for the lateral approach to the hip and independent of the C-arm model used. From our calculations we can recommend adding a correction factor of 4 ${}^{\circ }$ on intraoperative radiographic cup inclination. When faced with a very low tilted cup of $<$ 31 ${}^{\circ }$ inclination or a very high tilted cup of >42 ${}^{\circ }$ inclination we recommend using a 6 ${}^{\circ }$ respectively 2 ${}^{\circ }$ correction factor instead to get the best possible post operative result.

5.1 Limitations

The parallax effects may differ from various C-arm models. The given correction factor is calculated for the Philips BV Endura. However, the actual effects and their significance of different C-arm models on parallax effects remains unclear. Furthermore, we calculated the correction factor based on the data that we collected in a university hospital as described. The resulting variance in cup positioning influences the calculation of the correction factor. Other clinics with a different setting will presumably produce a different variance in cup inclination and thus possibly receive a slightly different correction angle.

We did not calculate our findings with correlation to the body mass index of the patients. Body mass alters the distance (in mm/cm) of C-Arms to the individual hip cup center. This may also affect the calculation of the correction factor.

We did not perform a subgroup analysis for different cup sizes. From our findings it remains unclear, whether the cup size itself influences the angulation or the calculated correction factor.

Funding

The authors received no financial support for the preparation, research, authorship, and/or publication of this manuscript.

Ethical approval

A positive vote was given by the medical ethics committee of the University of Cologne (EN 21-1038).

Informed consent

The authors declare that there is no information (names, initials, hospital identification numbers, or photographs) in the submitted manuscript that can be used to identify patients.