A review of plain radiographic methods for measuring femoral version after total hip arthroplasty and validation of modified Budin’s method in Indian population

Abstract

Background

Multiple plain radiographic methods; cross table radiographic method, modified Budin method, projected neck-shaft angle method, and trans lateral decubitus view method using Ogata-Goldsand formula, have been described in literature for measurement of femoral stem version (FSV) after total hip arthroplasty (THA).

Purpose

To review these multiple radiographic methods and also determine validity and accuracy of modified Budin method in Indian population.

Patients and methods

A literature search for different methods described for FSV measurement was performed and these methods were reviewed. In addition, for validation of modified Budin method in Indian population, data was collected for 36 THAs prospectively. A posteroanterior radiograph with patient sitting in 90° hip flexion and 30° abduction was taken 3 weeks and 6 weeks after surgery for calculating the FSV using modified Budin method. At 3 weeks, a CT scan was also done for version measurement. Intra and interobserver reliability, and reproducibility of radiographic FSV measurement, and the correlation between CT scan and radiographic FSV measurement were statistically calculated.

Results

All the reviewed studies demonstrated that their method is comparable to CT method for FSV measurement, except FSV measured on cross table radiographic view. However, there were only one or at most two studies of every method described except for modified Budin technique. The mean FSV in our prospective case series using ‘modified Budin’ method was 11.6° which was comparable to the mean of CT scan version measurement (12.3°). The mean difference was 0.7° which was not statistically significant (p value > 0.05). In addition, there was high intra-class correlation coefficient in radiographic FSV for both intra- and inter-observer reliability.

Conclusion

Multiple methods have been described and validated in literature, however, the ‘modified Budin’ view have proved multiple times to have excellent reliability and validity for easy measurement of FSV.

1. Introduction

Total hip arthroplasty (THA) is a successful procedure; however, its longevity depends mainly on accurate femoral stem and acetabular cup position, among many other factors. Inaccurate femoral stem version (FSV) is a major risk factor for postoperative pain, rotational instability,⁹ early dislocation,^12,17 decreased range of motion,³ peri-prosthetic fracture, bony impingement¹³ and component wear.¹⁶ A practical and reliable method is required for FSV measurement owing to its impact on the functional outcome of THA.¹⁷ However, a wide variation (−19° retroversion to +33° anteversion),¹⁸ measurement in axial plane (which is difficult in plain radiographs), patient’s hip and leg position influence on measurement, complex mathematical formulae, and requirement of special views have been the obstacles for FSV accurate measurement.

CT scan, highly accurate and reliable, has been the reference standard for the precise post-operative radiological measurement of FSV in THA,¹¹ but this is expensive,¹¹ requires additional radiation exposure,¹ artefacts are created by the implants¹⁹ and is not available ubiquitously in developing countries like India. Many approaches have been developed for FSV measurements using plain postoperative radiographs; cross table radiographic method,⁸ modified Budin method on PA radiograph,¹¹ projected neck shaft angle method on AP radiograph,²⁰ trans-lateral decubitus method⁶ and method with the biplanar (A/P and lateral) X ray system from EOS,⁵ but none have proved to be as accurate as CT scan. These descriptions of multiple methods are the proof that assessment of FSV on plain radiographs is challenging. The goal of the current study was to review all the plain radiographic methods in literature and validate the precision and reproducibility of modified Budin method for measurements of FSV in Indian population.

2. Materials and methods

A literature search in PubMed was done through June 2020 with the search words ‘femoral stem anteversion’ OR ‘femoral version’ OR ‘stem torsion’ OR ‘femoral stem version’. Only the clinical studies comparing plain radiographic method with CT scan for FSV after THA were included. Biomechanical studies, studies without comparison with CT scan, and the radiographic method validation study were excluded. In addition, non- English literature, duplicates and letter to editor were not considered for literature review. Finally, a total of 7 studies were included in this systematic review. The mean age, mean radiographic FSV, mean CT FSV, mean difference and standard deviation between these values, Pearson correlation coefficient, and inter- and intraobserver reliability were noted.

We also tried to evaluate the validity of modified Budin method for FSV measurement in our prospective case series of 36 THAs operated from June 2015 to January 2020 by a fellowship trained senior arthroplasty surgeon (AP). Modified Budin view (a postero-anterior radiograph) was taken at 3 weeks post-surgery with x ray beam centred on the hip and parallel to the floor (Fig. 1A). The patient is seated with hip and knee flexed to 90° and hip 30° abducted (Fig. 1B). The line between the centre of the modular head and the centre at the base of the trunnion of the stem was defined as stem-neck axis. Occasionally, the centre of the acetabular component was taken as a surrogate marker of the centre of the head when the entire margin of the modular head was not seen on the radiographs. The line joining the posterior-most portions of the medial and lateral femoral condyles was defined as the posterior intercondylar line of the femur. The angle between the stem-neck axis and the posterior intercondylar line was defined as the radiological stem anteversion (Fig. 2).

Fig. 1a.

Position for radiographic femoral anteversion using modified Budin view as seen from the side-posteroanterior beam, knee in 90°, and distance between source and film is 110 cm.

Fig. 1b.

Position for radiographic femoral anteversion using modified Budin view as seen from above- 30° abduction of hip.

Fig. 2.

Radiograph obtained using modified Budin method. a) centre of head, b) centre of trunnion, c) posterior intercondylar line, and d) femoral anteversion angle.

At the same time CT scan of the hip and knee was done and its three cross sectional images were used to calculate FSV, i) the image showing the largest circular section of the head or acetabulum in which the centre was marked (Fig. 3A), ii) the image showing the largest stem neck width in which the centre of the trunnion was identified (Fig. 3B). The line connecting these two centres was defined as stem neck axis, and iii) in the knee, the image showing the largest section of condyles was identified where posterior most points of each condyle were marked (Fig. 3C). The line that connected these two points was defined as posterior intercondylar line. The angle measured between these two lines, stem neck axis and posterior intercondylar line was defined as CT femoral anteversion (Fig. 3C).

Fig. 3a.

Measurement of femoral version on CT scan. Centre of the modular head (c).

Fig. 3b.

Measurement of femoral version on CT scan. Centre of the base of the trunnion (c′), line (C) was between the two centres, which is the stem-neck axis.

Fig. 3c.

Measurement of femoral version on CT scan. Posterior-most point of each condyle (d and d′), and line (D) between these two points, which is the posterior intercondylar line of the femur. Angle between C and D is femoral version.

Radiographs were repeated again after three weeks to ascertain the effects of positioning and other variables affecting the assessment on x rays, however, CT scan was done only once at 3 weeks as it is associated with high radiation dose and is costly. All radiographs were taken in the same department using a standardized protocol. Two blinded observers made each observation at two different points in time for measurement of inter- and intra-observer reliability. The mean of all the observations of radiographic method was regarded as femoral version. The CT method was regarded as the reference standard.

3. Assessment and statistical analysis

Intra-observer reliability was assessed by one examiner on two occasions three weeks apart on each CT scan and radiograph for the case series of 36 THA. Interobserver reliability was assessed across the two examiners. Intra observer and interobserver reliability were tested using intra class correlation coefficient (ICC) for radiographs as well as CT for femoral version. Reproducibility of modified Budin view and femoral version measurements in them were also analysed using ICC for each. An ICC of 1 means perfect reliability or reproducibility and an ICC of 0 means the opposite. The Pearson’s correlation coefficient (PCC) was used to evaluate correlation between mean radiological and CT measurements. PCC was characterized as poor (0.00–0.20), fair (0.21–0.40), moderate (0.41–0.60), good (0.61–0.80) or excellent (0.81–1.00) [26]. Statistical analyses were conducted using IBM-SPSS for Windows version 22.0 (SPSS Inc., Chicago, Illinois) and statistical significance was set at p < 0.05.

4. Results

In our case series of 36 THA patients, the mean difference between the radiological modified Budin method (11.6°) and CT (12.3°) measurement of femoral anteversion was 0.7° which was statistically insignificant (p value > 0.05). The Pearson correlation coefficient (PCC) between the two measurements was r = 0.987 (p < 0.001, r² = 0.975) which is excellent (Fig. 4). The mean difference between radiological and CT measurement of FSV can be expected to be between 0.4° and 1.9° according to 95% confidence interval with an ICC of 0.914. With regard to the radiological femoral version, excellent intraobserver reliability with an ICC of 0.905 (95% CI: 0.873–0.933) was observed. Likewise there was excellent interobserver reliability, with an ICC of 0.887 (95% CI: 0.844–0.914) between the two observers (Table 1). Similarly, reproducibility of the observations taken by two observers at two different points of time was excellent with ICC of 0.985 (95% CI: 149 0.970–0.993).

Fig. 4.

Correlation between radiographic method and CT scan.

Table 1.

Femoral stem version: Intra- and interobserver reliability of measurement on CT scans and Modified Budin method.

	Intra-observer reliability (ICC, 95% CI)	Inter-observer reliability (ICC, 95% CI)
CT scan	0.924 (0.891–0.958)	0.908 (0.874–0.935)
Modified Budin method	0.905 (0.873–0.933)	0.887 (0.844–0.914)

ICC, intraclass correlation coefficient; CI, confidence interval.

The results and findings of all the studies describing and comparing multiple methods for measurement of FSV are summarised in Table 2. Except FSV measured on cross table radiographic view,⁸ all other studies demonstrated that their method is comparable to CT method. However, there are only one or two studies of every method described except for modified Budin method.^11,20,21

Table 2.

Studies comparing radiographic method with CT scan method for measurement of femoral version.

Author	Sample Size	Age (years)	Assessment	Radiograph Mean (°)	CT scan Mean (°)	Mean difference (°)	SD in MD	Correlation (r)	Radiograph Interobserver reliability	Radiograph Intraobserver reliability ICC	CT Interobserver reliability ICC	CT Intraobserver reliability ICC
Ha et al., 2020	36	59.8	Trans -lateral decubitus view vs CT	23.72	23.91	−0.19β	NR	0.729 (p = 0.0001)	ICC: 0.935	0.934	0.975	0.941
Esposito et al., 2018	20	NR	EOS vs CT	NR	NR	4 β	4	≥0.88 (p < 0.01)	Cronbach’s α = 0.89	NR	NR	NR
Woerner et al., 2016	24	62.6	AP CCD vs 3D-CT Scan	9.4	7.2	2.2 β	6.8	0.78 (p < 0.001)	Mean CCC:≥0.79	≥0.88	NR	NR
Weber et al., 2015	115	62.5	AP CCD vs 3D-CT Scan	9.1	8	1.2 β	6.2	0.88 (p < 0.001)	Mean CCC:0.87 (excellent)	≥0.94	NR	NR
	115	62.5	Modified Budin vs 3D-CT	6.7	7.2	−0.5 β	4.2	0.84 (p < 0.001)	Mean CCC:≥0.86	≥0.94	NR	NR
Kanazawa et al., 2015	103	65.5	Cross table radiograph vs CT	19.4	38.8	19.5 α	NR	0.30 (fair)	NR	NR	NR	NR
Guenoun et al., 2015	28 composite femurs	NA	EOS vs CT	4.06	3.82	0.24 β		0.98	ICC: 0.997	0.998	0.992	0.998
	30 patients	NR	EOS	7.78	NA	NA	NA	NA	ICC: 0.914	0.998	NA	NA
Lee et al., 2013	36	60.8	Modified Budin vs CT	12.35	13.36	1.01 β	NR	0.877 (p = 0.001)	ICC: 0.934	0.944	0.943	0.942
Hermann et al., 1998	40	62	Lateral oblique view vs 3D-CT	16	16	NR	NR	0.91	NR	NR	NR	NR

AP CCD, Method based on anteroposterior (AP) hip radiographs using the projected caput-collum-diaphyseal (CCD) angle; SD, Standard deviation; MD, Mean difference; CCC, Concordance correlation coefficient for interobserver reliability; ICC, Intraclass correlation coefficient; α, Significant difference; β, No significant difference; NR, Not reported; NA, Not Applicable.

The rotation-dependent projected neck–shaft angle (NSA) of the femoral component on the full weight bearing standing AP radiographs was measured (the true NSA was 135° in their study) by Weber et al.²⁰ FSV was calculated with the help of the projected and true NSA, and applying a newly developed mathematical formula.¹⁰ Using a mathematical formula and a low cost ubiquitously available AP radiograph, this method allowed a simple and practicable evaluation of post-operative FSV. However, the authors cautioned about its use when a range of accuracy of ±3° is required. This method was also later validated by another study.²¹

In addition to measurement of cup version, Kanazawa et al. used cross table radiographs (where patient is supine and contralateral hip is flexed) for measuring FSV and compared it with CT measured FSV.⁸ Crosstable FSV was defined as the angle formed between the axis of prosthetic neck and the long axis of body. The authors found wide variation between the two values along with poor correlation between them. This was attributed to the difference in the measurement plane, inconsistency in maintain the position for obtaining the radiographs and errors due to inherent femoral bow and inbuilt prosthetic version.

A low-dose digital stereoradiography, EOS (EOS Imaging, Paris, France) system, was compared with CT scan for FSV evaluation in two studies.^4,5 This system produces a full length, weight bearing images using minimal radiation. It is composed of orthogonally placed two image acquisition systems each containing an x-ray tube and a linear detector in a C-shaped vertically travelling arm. FSV and other parameters can be automatically calculated by specially designed software (SterEOS, EOS Imaging, Paris, France) which is included in the workstation as it allows three-dimensional (3D) modelling of the bone envelope. There was excellent correlation between EOS and CT FSV value which makes the biplanar radiography a reliable low-radiation alternative for measuring femoral version as well as other measurements compared to CT.

A combination of trans lateral view and AP radiograph were used by Ha et al. for measurement of FSV.⁶ For trans lateral view, the patient is placed in a lateral decubitus position on the affected side with hip extended and the ipsilateral knee flexed by 90° and opposite hip at 100° flexion and 30° abduction (maintained by placing a foam block below the opposite leg). First, the stem neck-femoral shaft angle was calculated on both hip AP view and trans-lateral decubitus view. Then, FSV was measured using the Ogata and Goldsand formula.¹⁵ There was excellent correlation between the FSV calculated with this method and CT scan, especially when the radiological anteversion was <30°.

In 1998, a total of three views, an AP view, a lateral oblique view at 45°, and a lateral view of the knee, and Norman’s trigonometric principles were used by Hermann and Egund for calculating FSV.^7,14 The FSV measured with this method was comparable to CT FSV value. Also, there was also an acceptable level of accuracy and reproducibility with this method.

5. Discussion

When determining the causes of mechanical failures in THA, measurement of implant position plays a vital role. CT scan is a gold standard and used as a reference method for measuring various parameters, especially FSV as it is difficult to evaluate stem version in plain radiographs. Besides, no single radiographic method has proved to be as accurate as CT. All plain radiographic methods have their outliers and limitations regarding their applicability. For instance, Weber’s method; i) unable to differentiate between the anteversion and retroversion, ii) an extra radiograph (axial view) is necessary, iii) vulnerable to patient’s position (which can influence the projected NSA), iv) true prosthetic NSA should already be known, v) not applicable for preoperative evaluation of native femoral version, vi) the mathematical model used is only defined for angles exceeding the known true prosthetic NSA, and vii) requirement of a highly standardized radiographic technique for precise patient positioning and projection errors minimization. However, no special view (as in modified Budin method¹¹) is required, AP view is sufficient which saves the patient from an additional radiation exposure.²⁰

Kanazawa’s method using cross table radiograph for FSV measurement also have some limitations.⁸ The patient’s position while obtaining cross table radiographs plays a crucial role in accurate determination of FSV. The hip and knee should be extended and there should be neutral hip rotation for ideal radiograph. These requirements may not be fulfilled if patient has hip and/or knee contractures and when the examiner is not aware of the correct position. If the hip is inadvertantly internally rotated by the examiner, FSV will be factitiously decreased, and vice versa. Crosstable radiographs, thus, were not recommended for clinical use in measuring FSV.

EOS method has proved to be a reliable low dose radiation alternative to CT scan for measuring FSV. However, anatomic landmarks on biplanar radiographs are difficult to identify which requires training and practice.⁵ The acquisition time for EOS radiography is about 10 s and the patient has to stand motionless. This can be difficult for the patients having some postoperative complications and when the examination is performed within a short duration after surgery. Despite these drawbacks, EOS method has the advantage of having utility in preoperative planning as well as postoperative implant positioning assessments. Besides, this technique can calculate other parameters also without any additional films, such as acetabular cup orientation, stem offset, and limb length discrepancy. In addition to measurement of anatomic definitions of component position, this 3D stereoradiography technique has a major advantage of evaluating the functional component position in standing or sitting position which can be helpful in considering the impact of pelvic tilt and femoral torsion on combined anteversion.

Recently described method by Ha et al. to measure FSV requires two different films and accurate technique of radiography along with complex mathematical calculation.⁶ The reference points for measurement of FSV in this technique are similar to CT scan method, thus accurate measurements can be expected. However, this method has to be modified when neck-shaft angle is other than 132°. In addition, it is not applicable in deformed femurs. There can be discrepancy in the measurement as softfoam block was used to support the contralateral leg which might not hold the patient’s leg in the desired position consistently.

Modified Budin method used by Lee et al.¹¹ was their modification of the method used by Budin and Chandler² in 1957 for femoral version measurement in children with cerebral palsy or developmental dysplasia of hip to plan femoral osteotomy. Similar to this study, our prospective case series showed excellent Pearson correlation coefficient (0.987) with no mean difference between radiological measurement and CT measurement. The radiological method also has high intraobserver as well as interobserver reliability with excellent reproducibility, thus different observers can make the measurement with similar precision at multiple times. However, this method has limitation in obese and tall patients because of difficult visualisation of landmarks and in stiff hip or knee patients due to difficult positioning. Moreover, the radiology technician requires expertise and experience for proper position of the patient, and appropriate exposure of the film.

6. Conclusion

There is sparse literature on reliability and accuracy of different methods of femoral version measurement on radiographs. However, despite multiple limitations, these conventional radiographic approaches harbour the possibility of a simple and feasible evaluation of postoperative FSV. Among all of them, the modified Budin method is a better alternative to CT scan because of its excellent validity and reliability as it does not require multiple films, complex mathematical calculation or difficult limb position.

Location of work

This work was performed at Guru Teg Bahadur Hospital, Dilshad Garden, Delhi-110095.

Declaration of competing interest

There is no conflict of interest for any authors.