Genetic influences in the aetiology of femoroacetabular impingement

Identification of proband.

The study had ethical approval. Probands were recruited from three centres which routinely performed surgery for FAI by arthroscopy, anterior arthrotomy, or surgical dislocation. Clinical and radiological criteria were applied to validate the diagnosis of FAI among the probands. The proband had to have a clinical diagnosis of anterior FAI and a positive anterior impingement test¹⁵ documented in the initial clinical correspondence. Then, the patient had to have an anteroposterior (AP) radiograph of the pelvis in the supine position with the x-ray beam centred in the midline and on the point midway between the superior border of the pubic symphysis and a line drawn connecting the anterior superior iliac spines,^16,17 and a cross-table lateral radiograph of the hip taken in 15° of internal rotation.^16,18,19

The following exclusion criteria were applied to ensure that only probands with primary anterior FAI were recruited to the study. Patients had to be between the ages of 18 and 55 years and not have a diagnosis of posterior FAI. They should not have had previous investigation of or treatment to either hip as a child or adolescent, or have had acquired FAI before or after skeletal maturity (for example secondary to SCFE, trauma or infection), as determined from the case notes. Those undergoing a surgical dislocation who were found to have advanced disease and had a resurfacing arthroplasty performed at the initial procedure were excluded, as were patients with definite radiological OA of the hip equivalent to grade 2 of the classification of Kellgren and Lawrence.²⁰ Patients with acetabular dysplasia as diagnosed by a lateral centre-edge angle on the AP radiograph of less than 20°,^16,21 were also excluded as were those of non-Caucasian origin to eliminate confounding because of racial variation. Finally, the obturator-foramen index²² was measured on all the AP pelvic radiographs to ensure that there was no significant pelvic rotation. Only patients with an index between 0.7 and 1.4 were accepted in the study.²³

Probands were identified both prospectively and retrospectively from a database. They were asked to detail siblings resident in the United Kingdom, and their spouse or partner. In addition, to validate the information from the case notes, the probands were asked whether they had been investigated or treated for problems with either hip as a child, adolescent or adult. All the probands had prospectively completed the Oxford hip score²⁴ (OHS) and the non-arthritic hip score pre-operatively. The OHS is validated, patient-based, sensitive to clinically important change and scored from 0 (asymptomatic) to 48 (most severe symptoms).²⁵ The non-arthritic score is validated, patient-based and scored from 0 (asymptomatic) to 100 (most severe symptoms).²⁶

Recruitment and assessment of the siblings and control group.

Siblings and control subjects were invited to participate in the study. In order to increase the number of the latter, all participating siblings were asked to detail their spouses or partners. All participating subjects were then assessed in a dedicated research clinic by a single experienced observer (TCBP).

The clinical assessment included a history, examination and completion of the OHS and non-arthritic hip score for each hip and the UCLA (University of California, Los Angeles) activity score.²⁷ The last is scored from 0 to 10, is validated and incorporates occupation and sporting activity.^27,28 A proforma facilitated standardisation of the clinical assessment. All the subjects were asked whether they had previously had surgery on either hip, undergone investigation of or treatment for pain in the hip as a child or adolescent, had experienced any groin pain or clicking on either side in the last two years and whether either of their parents had undergone total hip replacement or resurfacing arthroplasty for OA. A routine examination of the hips was performed and the presence of a positive anterior impingement sign or irritability of movement of the hip recorded.

All participants had an AP radiograph of the pelvis in the supine position and a cross-table lateral radiograph of each hip taken as described above, using a 15° wedge placed beneath the femoral condyles to standardise rotation. The radiographer repeated the AP radiograph of the pelvis if necessary to ensure that the obturator foramen index was within 0.7 and 1.4.

Radiological measurement.

There are two principal mechanisms of FAI, namely, cam and pincer.^1,2 Cam impingement is characterised by a reduction of offset or abnormal asphericity at the anterolateral femoral head-neck junction. Pincer impingement results from over-coverage of the femoral head by the acetabulum, which may be global (coxa profunda) or focal (cranial retroversion of the acetabulum). Most patients have combined cam and pincer morphology and are described as having ‘mixed FAI’.^1,2 Based on the radiological measurements described below, the presence of cam and pincer deformities in the probands, siblings and the control group was recorded and their morphology classified as being normal, pure cam, pure pincer or mixed.

Proximal femoral morphology.

From the cross-table lateral radiograph, the alpha angle^16,29–31 and anterior offset ratio^16,18,31,32 were measured by saving the digital images as TIFF files, which were then analysed in a custom-designed software program in Matlab version R2007a (The Math-works Inc, Natick, Massachusetts). The accuracy of this program had been tested and confirmed during its development.³¹ Based on a previous study from our institution of 157 normal subjects using the same standardised radiological and measurement technique,³¹ 95% reference intervals³³ for the alpha angle and anterior offset ratio were calculated and a cam deformity was defined as an alpha angle > 62.5° or an anterior offset ratio < 0.135.

Acetabular morphology was determined from the AP radiograph of the pelvis. Global overcoverage is suspected when the hip has coxa profunda or protrusio.² Because of its high prevalence in normal hips, especially in women,³⁴ we defined coxa profunda based on the measurement of the lateral centre-edge angle,^16,21 the acetabular index,^16,35 and the acetabular depth-to-width ratio.^36,37 Because there is limited evidence for the thresholds of abnormality for these measurements in the context of pincer FAI,^16,38,39 we calculated 95% reference intervals³³ from the control subjects who had no evidence of OA and were asymptomatic with a normal examination. There were 23 such men and 23 women. Global overcoverage was thus defined as a lateral centre-edge angle > 38.0° (men and women) or an acetabular index < −3.7° (men) or < −4.9° (women), or an acetabular depth-to-width ratio of > 0.55 (men) or > 0.64 (women) or coxa protrusio.² All the measurements were made from the TIFF files using the Matlab software program.

Focal overcoverage was diagnosed by the presence of a crossover sign.^16,40 Because of the sensitivity of this assessment to pelvic tilt,^16,41 we measured the distance from the sacrococcygeal joint to the pubic symphysis and only recorded the crossover sign if this distance was within 40 mm to 55 mm for women and 25 mm to 40 mm for men.^41,42 The distances were measured using the Picture Archiving and Communication System (PACS, 2004; GE Healthcare UK Ltd, Little Chalfont, United Kingdom). Assessment of global overcoverage was made in all subjects regardless of the sacrococcygeal joint-pubic symphysis distance, since it was considered that referencing the margins of the sourcil¹⁶ rather than the outermost point of the acetabular rim²¹ would protect against error due to variation in pelvic tilt.²³ A pincer deformity was diagnosed if there was either focal or global overcoverage or both.

Presence of osteoarthritis.

Osteophytosis was graded using an atlas⁴³ and the minimum width of the joint-space recorded in each hip. Narrowing of the joint space was defined as a minimum width of < 2.80 mm in men and < 2.87 mm in women.³¹ An overall grade of OA was assigned using the system of Kellgren and Lawrence,²⁰ whereby grade-2 is defined as definite but mild OA, with narrowing of the joint space and osteophytosis.

Observer variability.

All the measurements were made by one observer (TCBP). Intraobserver repeatability was obtained by the re-measurement of 20 sets of radiographs (40 hips) after an interval of four weeks. Interobserver reproducibility was obtained by a second observer (MRW) measuring the same 20 sets of radiographs.

Power analysis.

The sibling relative risk for OA of the hip is approximately 2.^8,10,11 The incidence of morphology of FAI and its clinical features in the general population is unknown. Assuming a similar sibling risk for abnormal morphology, with an estimation of the prevalence of abnormal morphology in the general population of 20%, then the prevalence in the siblings would be 40%. On this basis, to achieve 80% power (alpha 0.05), the estimated sample size in each group was 82 subjects.

Statistical analysis.

The one-sample Kolmogorov-Smirnov test was used to confirm a Gaussian distribution of the measurements for global overcoverage in the normal control subjects. Differences in continuous variables between groups and genders were compared using the unpaired t-tests. The chi-squared test and Fisher’s exact test were used to determine variations in the prevalence of abnormal morphology, clinical features and OA between groups. Intra- and inter-observer reliability of numerical data was assessed by intra-and interclass coefficients and of categorical data by the kappa coefficient. All the statistical calculations were performed using SPSS version 13.0 software (SPSS Inc., Chicago, Illinois). A p-value ≤0.05 was deemed to be significant.

Hip morphology.

Table II summarises the morphological classification of each hip in the sibling and control groups. The prevalence of a cam deformity in the siblings of those probands with a cam deformity was significantly greater than that in the control group (p < 0.001, Table III).

The presence of focal acetabular cover (a crossover sign) was not assessed in 28 female (56 hips) and 19 male (38 hips) control subjects, and in 26 female (52 hips) and 23 male (46 hips) siblings because the sacrococcygeal joint-pubic symphysis distance was outside the acceptable range. Assessment of global overcoverage was still made in these subjects. The prevalence of a pincer deformity in the siblings of those probands with a pincer deformity was significantly higher than that in the control group (Table IV). The prevalence of bilateral FAI morphology was higher among the siblings (42 of 96 siblings versus 13 of 77 control subjects, relative risk 2.6, p = 0.002). Overall, 16 of 77 (20.8%) control subjects had at least one hip with a cam deformity compared with 45 of 96 siblings (46.9%) (relative risk = 2.3, p < 0.001). Of the 77 control subjects, 22 (28.6%) had at least one hip with a pincer deformity compared with 39 of the 96 (40.6%) siblings (relative risk = 1.4, p = 0.111).

Clinical features and morphology of FAI.

The prevalence of clinical features in the siblings with abnormal morphology was significantly higher, particularly in men, and was statistically significant because of the magnitude of the relative risks in spite of the small numbers (p = 0.040 Table V). Table VI shows the pre-operative hip scores in the probands and symptomatic siblings. Six siblings had hip scores which were worse than the means for the probands.

Potentially these findings could be explained by the deformity being more severe in those siblings with FAI morphology compared to the control subjects with FAI morphology. However, the prevalence of a mixed pattern of FAI was similar in these subgroups (21 of 109 siblings hips versus 6 of 46 control hips, p = 0.487, Table II). Furthermore, the mean alpha angle and anterior offset ratios were not significantly different in those siblings and controls with cam deformities (p = 0.578 and p = 0.320 respectively).

Because of the potential confounding effect of the level of activity on the likelihood of an individual with FAI morphology having clinical features, the siblings with the morphology and signs of FAI were compared with the control subjects with FAI morphology without signs for age and level of activity (Table VII).

Osteoarthritis.

In the sibling group, 11 of 96 siblings (5 unilateral, 3 bilateral) had grade-2 OA compared with 0 of 77 control hips (p = 0.002). In no case was OA more advanced than grade 2.¹⁹ Of the 64 proband families 13 had at least one parent who had undergone a hip replacement for OA, compared with six of the 77 control families (p = 0.046).

Patients with cam FAI often present with a distinct ‘bump’ at the femoral head-neck junction. Because there have been no longitudinal observational studies of patients with FAI before presentation, the reasons for the development of these bumps are unknown. Three unrelated individuals with cam morphology in the sibling group had clear erosive lesions at the head-neck junction. It is possible that a reactive phenomenon results in bony deposition at these sites, with a resultant bump which exacerbates the cam deformity. Figure 1 illustrates this in a family of three brothers.

Fig. 1a, Fig. 1b, Fig. 1c

Radiographs of three brothers with different stages of clinical and radiological progression showing a) a pre-operative view in a 27-year-old proband with an alpha angle of 81.3°, b) normal clinical examination in a 21-year-old brother with an alpha angle of 63.1° and an erosive lesion at the head-neck junction and c) positive impingement sign in a 23-year-old brother with an alpha angle of 78.8° and bony deposition at the head-neck junction forming a bump.

Reliability.

Inter- and intraobserver reliability data are presented in Table VIII.