Drs. Reichenbach and Leunig contributed equally to this work.
Association between cam-type deformities and magnetic resonance imaging–detected structural hip damage: A cross-sectional study in young men
Version of Record online: 29 NOV 2011
Copyright © 2011 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 63, Issue 12, pages 4023–4030, December 2011
How to Cite
Reichenbach, S., Leunig, M., Werlen, S., Nüesch, E., Pfirrmann, C. W., Bonel, H., Odermatt, A., Hofstetter, W., Ganz, R. and Jüni, P. (2011), Association between cam-type deformities and magnetic resonance imaging–detected structural hip damage: A cross-sectional study in young men. Arthritis & Rheumatism, 63: 4023–4030. doi: 10.1002/art.30589
- Issue online: 29 NOV 2011
- Version of Record online: 29 NOV 2011
- Accepted manuscript online: 8 SEP 2011 01:19PM EST
- Manuscript Accepted: 3 AUG 2011
- Manuscript Received: 3 DEC 2010
- Swiss National Science Foundation (National Research Program 53 on Musculoskeletal Health). Grant Number: 405340-104778
- Deutsche Arthrose-Hilfe e.V.
- PROSPER fellowship from the Swiss National Science Foundation
Femoroacetabular impingement may be a risk factor for hip osteoarthritis in men. An underlying hip deformity of the cam type is common in asymptomatic men with nondysplastic hips. This study was undertaken to examine whether hip deformities of the cam type are associated with signs of hip abnormality, including labral lesions and articular cartilage damage, detectable on magnetic resonance imaging (MRI).
In this cross-sectional, population-based study in asymptomatic young men, 1,080 subjects underwent clinical examination and completed a self-report questionnaire. Of these subjects, 244 asymptomatic men with a mean age of 19.9 years underwent MRI. All MRIs were read for cam-type deformities, labral lesions, cartilage thickness, and impingement pits. The relationship between cam-type deformities and signs of joint damage were examined using logistic regression models adjusted for age and body mass index. Odds ratios (ORs) and 95% confidence intervals (95% CIs) were determined.
Sixty-seven definite cam-type deformities were detected. These deformities were associated with labral lesions (adjusted OR 2.77 [95% CI 1.31, 5.87]), impingement pits (adjusted OR 2.9 [95% CI 1.43, 5.93]), and labral deformities (adjusted OR 2.45 [95% CI 1.06, 5.66]). The adjusted mean difference in combined anterosuperior femoral and acetabular cartilage thickness was −0.19 mm (95% CI −0.41, 0.02) lower in those with cam-type deformities compared to those without.
Our findings indicate that the presence of a cam-type deformity is associated with MRI-detected hip damage in asymptomatic young men.
Osteoarthritis (OA) of the hip is one of the major causes of pain and disability (1, 2), accounting for more than 200,000 hip replacements annually in the US (3). The etiology of OA is multifactorial (4). Current classifications include “idiopathic” OA and “secondary” OA in individuals with clearly visible deformities such as hip dysplasia (5). More than four decades ago, however, Murray suggested that most cases of idiopathic OA were the result of frequently undetected deformities and were therefore secondary OA (6). These deformities were later suggested to cause femoroacetabular impingement and signs of early hip OA (7). Two different types of impingement were distinguished, cam and pincer. (An illustration is available online at www.ispm.ch/journal-downloads.)
Cam impingement is often seen in young male athletes referred to orthopedic surgeons because of groin pain (8), and internal rotation in flexion is usually found to be diminished (9). It is caused by a nonspherical extension of the femoral head or a decreased offset at the anterolateral transition of femoral head to neck (8, 10, 11), which were referred to as cam-type deformities (7, 12). The increased radius of the femoral head entering the acetabulum may result in shear forces at the acetabular cartilage, especially during flexion and internal rotation, and may lead to an abrasion of the acetabular cartilage and to cartilage avulsion at both the labrum and the subchondral bone (13). Conversely, pincer impingement occurs more frequently in women and results from increased acetabular depth with overcoverage of the femoral head, while the femoral head–neck configuration may be normal (14–16).
Cam-type deformities are common in asymptomatic young men (12), but their clinical relevance is unclear. We therefore examined whether cam-type deformities, a nonspherical extension of the femoral head and a decreased anterior head–neck offset, are associated with early signs of hip damage, including labral lesions and decrease in articular cartilage, visible on magnetic resonance imaging (MRI).
SUBJECTS AND METHODS
The Sumiswald Cohort is a population-based inception cohort of consecutive young men undergoing conscription for the Swiss army at a single recruiting center in Sumiswald, Switzerland (12). In Switzerland, ∼97.5% of men of Swiss nationality are required by the army to attend a 3-day recruitment session in specialized centers, regardless of their health status. Consecutive males were asked to participate in the baseline examination for this study, complete a self-administered questionnaire, and undergo clinical examination. We screened all individuals for hip pain using a modified version of the question used in the first National Health and Nutrition Examination Survey (17): During the past 3 months, have you had pain in or around either of your hips? Individuals who reported hip pain of ≥3 on a Likert scale ranging from 1 (no pain) to 5 (extreme pain) were excluded. Additional exclusion criteria were previous surgery in either hip joint, metabolic or inflammatory rheumatic disease or a history of hemophilia, age younger than 18 years, and an inability to give written informed consent. Self-report questionnaires included the pain, stiffness, and function subscales of the Western Ontario and McMaster Universities OA Index (WOMAC) version 3.1 to quantify symptoms within the previous 48 hours (18), as well as the EuroQol 5-domain questionnaire, which includes 5 dimensions and a visual analog scale pertaining to health-related quality of life (19). We measured internal rotation using a recently developed examination chair (20), which enabled us to accurately quantify internal rotation of the hip in a sitting position with the hips and knees flexed 90° and the lower legs hanging unsupported (20). The study was approved by the Research Ethics Committee of the Canton of Bern. All participants gave written informed consent prior to any data collection.
We used a central computer-generated randomization schedule for random selection of participants for invitation for MRI. Stratification was performed according to the extent of internal rotation, with oversampling in the strata with the lowest (<30°) and highest (≥40°) internal rotation. Only 1 hip per participant was examined by MRI. In individuals whose hips had different ranges of motion, the hip with the lower degree of internal rotation was selected. When hips had similar internal rotation (within 1°), the hip for MRI was randomly selected using a concealed central computer-generated randomization schedule (12). All MRI studies were performed with a 1.5T high-field system (Magnetom Avanto; Siemens) using a flexible surface coil and high spatial resolution protocol with patients in the supine position with a neutral position of the hip joint. Radial proton density–weighted sequences were acquired, with all slices oriented parallel to the femoral neck axis, which was used as the axis of rotation. Sequences were performed using a sagittal oblique localizer, which was marked on the proton density–weighted coronal sequence and which ran parallel to the sagittal oblique course of the acetabulum (21).
Pulse sequence parameters of the turbo spin-echo sequence were as follows: repetition time (TR) 2,000 msec, echo time (TE) 15 msec, a field of view (FOV) of 260 × 260 mm, a matrix of 266 × 512, and a slice thickness of 4 mm with a resulting voxel size of 0.98 × 0.51 × 4 mm. The acquisition time for a complete set of 16 slices lasted 4 minutes 43 seconds. In addition, we used a transverse T1-weighted sequence (FOV 200 × 200 mm, slice thickness 4 mm, TR 650 msec, TE 20 msec); transverse fast low-angle shot sequence (FOV 120 × 120 mm, section thickness 2 mm, TR 650 msec, TE 20 msec, flip angle 90°); sagittal true fast imaging with steady-state precession (FISP) 3-dimensional (3-D) sequence (FOV 130 × 130 mm, section thickness 1.5 mm, TR 8.87 msec, TE 3.23 msec, flip angle 28°); sagittal inversion recovery sequence (FOV 180 × 180 mm, section thickness 3 mm, TR 4,800 msec, TE 32 msec, time to inversion 160 msec); and a coronal trueFISP 3-D sequence (FOV 180 × 180 mm, section thickness 1.5 mm, TR 8.16 msec, TE 2.89 msec, flip angle 28°). For ethical reasons, neither intraarticular nor intravenous contrast was injected.
To determine the presence of cam-type deformities, we graded the maximal offset at the head–neck junction on the radial sequences using a semiquantitative scoring system. This system consisted of grades ranging from 0 to 3, where 0 = normal, no evidence of a nonspherical femoral shape (cam deformity) on any sequence; 1 = possible deformity with cortical irregularity and a possible mild decrease in the anterior head–neck offset; 2 = definite deformity with an established decrease in the anterior head–neck offset (cam deformity of <10 mm); and 3 = severe deformity with a large decrease in the anterior head–neck offset (cam deformity of >10 mm) (12). The mean ± SD alpha angle was 44.8 ± 8.4° for grade 0, 48.4 ± 10.1° for grade 1, 57.7 ± 12.7° for grade 2, and 76.4 ± 9.7° for grade 3 deformities (P for trend = 0.001) (12). Grades 2 and 3 were prespecified to indicate a definite cam-type deformity. We used a clock face system to record the localization of cam-type deformities and signs of joint damage on MRI, such as labral disorders on radial sequences, with 12 o'clock denoting a superior location, 3 o'clock an anterior, 6 o'clock an inferior, and 9 o'clock a posterior location (12).
Figure 1 shows all disorders that were scored. The normal labrum has a pointed, triangular shape with sharp margins. Deformed labra (Figure 1A) were defined as those with any shape other than triangular (i.e., oval, round, or irregular) at 12 o'clock (21, 22). The prespecified primary outcome was the presence or absence of a labral lesion at any location from 1 o'clock to 12 o'clock, defined as a linear band of high signal intensity detected in the labrum. We distinguished 2 types of labral lesions: labral avulsion if detected at the basis, i.e., at the transition between labrum and acetabular cartilage (Figure 1B), and intralabral signal alterations if detected in the body of the labrum (Figure 1C). In a labral avulsion, the linear signal change reaches the surface, whereas in an intralabral signal alteration the linear signal intensity remains within the labrum and does not reach the surface (22, 23). Labral ganglia (Figure 1D) were recorded at any location from 1 o'clock to 12 o'clock (21). Impingement pits, formerly called herniation pits (Figure 1E), are well delineated, round to oval fibrocystic changes in the femoral neck, with increased or decreased signal intensity (24–26). These alterations are deemed to be a consequence of repetitive microcontusions of the femoral neck against the acetabular rim. We recorded the presence and locations of such cysts. We measured the combined thickness of the femoral and acetabular cartilage in the anterosuperior location in millimeters using the sagittal trueFISP sequence, where the section through the center of the femoral head was selected (Figure 1F). Separate measurement of the femoral and acetabular cartilage was not possible due to the lack of intraarticular contrast.
All MRIs were read by an experienced radiologist (SW). A random sample of 30 MRIs were also read by 2 other experienced radiologists (CP and HB). Weighted kappa values for intrarater agreement (27) were 1.00 (95% confidence interval [95% CI] 0.49, 1.00) for both labral lesions and impingement pits and 0.59 (95% CI 0.09, 1.00) for labral ganglia, whereas the intraclass correlation coefficient for cartilage thickness was 0.67 (95% CI 0.47, 0.86). Weighted kappa values for interrater agreements were 0.51 (95% CI 0.16, 0.86) for labral lesions, 0.63 (95% CI 0.27, 0.99) for impingement pits, and 0.74 (95% CI 0.25, 1.00) for labral ganglia, whereas the intraclass correlation coefficient for cartilage thickness was 0.74 (95% CI 0.59, 0.91).
The study was set up as an inception cohort study. Assuming 80% followup, an incidence of hip pain of 5% in those without deformity, and a 25% frequency of cam-type deformity, a sample size of 240 patients to be included at baseline would yield >80% power to detect a 16% difference in the incidence of hip pain between those with and those without cam-type deformity, and >90% power to detect a 20% difference. Associations between the presence of definite cam-type deformities and signs of joint damage (any labral lesions as primary outcome, labral avulsions, intralabral signal alterations, labral ganglia, labral deformities, and impingement pits) were determined using crude univariable and multivariable logistic regression models adjusted for age and body mass index (BMI). Sensitivity analyses were restricted to cam-type deformities and signs of joint damage in the anterosuperior location between 1 o'clock and 3 o'clock. An odds ratio (OR) of >1 indicates that hips with cam-type deformities are more likely to have signs of joint damage than those without.
Differences in cartilage thickness between hips with and without cam-type deformities were determined using crude univariable and multivariable linear regression models adjusted for age and BMI. A negative difference in cartilage thickness indicates that cartilage thickness is smaller in hips with cam-type deformities. As a graphical display of the distribution of cartilage thickness, we plotted cumulative frequencies of cartilage thickness separately for hips with and without cam-type deformities.
Baseline characteristics of the participants according to the presence or absence of cam-type deformities were compared using one-way analysis of variance for continuous data and chi-square tests for categorical data. WOMAC scores were standardized to range from 0 to 10, with higher values indicating more severe symptoms. The dimensions of the EuroQol were mapped onto a single health status index based on the European value set and standardized to range from 0 to 10 (28), with higher values indicating better health-related quality of life. P values are 2-sided. Statistical analyses were performed using Stata version 11 software (StataCorp).
The flow of participants through the study has been described previously (12) and is available online at www.ispm.ch/journal-downloads. Of 1,080 eligible individuals, 430 asymptomatic participants were invited for MRI examination and 244 attended (57%). Sixty-seven participants showed definite cam-type deformities (adjusted prevalence of 24%) (12). Table 1 presents a comparison of the participants with and those without cam-type deformity; those with the deformity had a larger BMI and decreased internal rotation.
|Definite cam-type deformity on MRI|
|Yes (n = 67)||No (n = 177)||P†|
|Age, years||20.0 ± 0.7||19.9 ± 0.7||0.35|
|Height, cm||178.8 ± 7.6||178.1 ± 6.9||0.54|
|Weight, kg||77.4 ± 13.7||71.9 ± 11.6||0.002|
|BMI, kg/cm2||24.3 ± 4.2||22.6 ± 3.4||0.002|
|Internal rotation, no. (%) of patients||<0.001|
|<30°||40 (60)||43 (24)|
|≥30° and <40°||17 (25)||64 (36)|
|≥40°||10 (15)||70 (40)|
|Overall||0.2 ± 0.5||0.2 ± 0.4||0.66|
|Pain||0.2 ± 0.8||0.2 ± 0.6||0.66|
|Stiffness||0.4 ± 1.1||0.7 ± 1.7||0.18|
|Function||0.2 ± 0.5||0.1 ± 0.4||0.30|
|EuroQol health state index‡||9.4 ± 1.2||9.3 ± 1.4||0.58|
|EuroQol VAS‡||8.3 ± 1.2||8.5 ± 1.2||0.21|
Table 2 presents crude and adjusted ORs for the association between cam-type deformity and signs of joint damage on MRI. The primary outcome of labral lesions was found in 57 of 67 participants with a cam-type deformity (85%) and 118 of 177 participants without a cam-type deformity (67%), yielding a crude OR of 2.85 (95% CI 1.36, 5.98). The association remained unchanged when adjusting for age and BMI (OR 2.77 [95% CI 1.31, 5.87]). Labral avulsions were observed in 76% of those with a cam-type deformity and 58% of those without, with a crude OR of 2.34 (95% CI 1.24, 4.43) and an adjusted OR of 2.24 (95% CI 1.17, 4.28). Intralabral signal alterations were found in 48% of subjects with a cam-type deformity versus 31% of those without, with a crude OR of 2.08 (95% CI 1.17, 3.71) and an adjusted OR of 2.12 (95% CI 1.17, 3.83). Labrum deformities were found in 27 of 244 MRIs (11%) and were more frequent in those with a cam-type deformity than in those without (18% versus 8%, adjusted OR 2.45 [95% CI 1.06–5.66]). The labrum was oval in 7 participants, round in 6, and of an irregular shape in 14. Impingement pits were found in 30% of the subjects with a cam-type deformity versus 12% of the subjects without (adjusted OR 2.91 [95% CI 1.43, 5.93]), whereas labral ganglia were detected in 31% of the subjects with cam-type deformity versus 25% of those without (adjusted OR 1.26 [95% CI 0.67–2.38]).
|Definite cam-type deformity||Crude OR (95% CI)||Adjusted OR (95% CI)|
|Yes (n = 67)||No (n = 177)|
|Labral lesions||57 (85)||118 (67)||2.85 (1.36, 5.98)||2.77 (1.31, 5.87)|
|Intralabral signal alterations||32 (48)||54 (31)||2.08 (1.17, 3.71)||2.12 (1.17, 3.83)|
|Labral avulsions||51 (76)||102 (58)||2.34 (1.24, 4.43)||2.24 (1.17, 4.28)|
|Labrum deformity||12 (18)||15 (8)||2.36 (1.04, 5.34)||2.45 (1.06, 5.66)|
|Impingement pits||20 (30)||21 (12)||3.16 (1.58, 6.33)||2.91 (1.43, 5.93)|
|Labral ganglion||21 (31)||44 (25)||1.38 (0.74, 2.56)||1.26 (0.67, 2.38)|
Figure 2 shows the localization of definite cam-type deformities and signs of hip damage. Of 67 definite cam-type deformities, 61 were located in the anterosuperior position between 1 o'clock and 3 o'clock (91%). Similarly, most labral abnormalities were located in this position. One hundred thirty-five of 175 labral lesions (77%), 112 of 134 labral avulsions (84%), 57 of 86 intralabral signal alterations (66%), 30 of 41 impingement pits (73%), and 44 of 65 labral ganglia (68%) were located between 1 o'clock and 3 o'clock. The results of a sensitivity analysis restricted to cam-type deformities and signs of joint damage in the anterosuperior location between 1 o'clock and 3 o'clock are available online at www.ispm.ch/journal-downloads. Estimated associations were similar to those observed in the main analyses, even though 95% CIs crossed the line of no difference at 1 for cartilage avulsion and labrum deformity.
The mean ± SD cartilage thickness of anterosuperior femoral and acetabular cartilage combined was 3.96 ± 0.74 mm in those with a cam-type deformity and 4.21 ± 0.77 mm in those without, with a crude difference of −0.24 mm (95% CI −0.46, −0.03 mm). This difference was slightly attenuated after adjustment for age and BMI (adjusted difference −0.19 mm [95% CI −0.41, 0.02]). Figure 3 shows crude cumulative frequency curves of cartilage thickness in those with and those without a cam-type deformity, with a shift of the cumulative frequency curve to the left by ∼0.20 mm in those with a deformity.
In our population-based cross-sectional study of 244 asymptomatic young Swiss men, we found that definite cam-type deformities were associated with labral damage, impingement pits, and decreased anterosuperior cartilage thickness. These relationships persisted after adjusting for age and BMI as potential confounding factors. Most of the cam-type deformities and most of the signs of damage on MRI were located in the anterosuperior position. As reported previously (12), the frequency of cam-type deformities was high (adjusted prevalence 24%), as was the frequency of signs of joint damage.
To our knowledge, this is the first population-based MRI study to examine the role of cam-type deformities of the hip as a potential biomechanical risk factor for joint damage. It is currently believed that the signs of joint damage used in the present study are potential intermediate outcomes in the sequence from normal to OA hips (29). The cross-sectional nature of our study, which was designed to form an inception cohort of asymptomatic young men, prevents us from determining at this point whether this assumption is true and whether the cam-type deformities and their association with joint damage will be associated with an increased risk of developing symptomatic OA of the hip with clinically relevant pain and disability. Only longer term followup in this and other longitudinal studies will clarify the clinical relevance of our findings. Considering that cam-type deformities are predominantly seen in symptomatic men in tertiary care settings, our study was restricted to men only (8). Additional population-based studies are therefore required in women.
Due to ethical considerations, we were unable to perform MRI arthrography because of the invasiveness of the intervention. The high-resolution protocol of the 1.5T MRI device used in our study was sophisticated, and anatomic structures could be adequately evaluated even in the absence of intraarticular contrast. As a result of a lack of contrast, we were unable to separately measure the thickness of femoral and acetabular cartilage, however, and reported only the overall thickness of the cartilage in the anterosuperior position.
We found that a large number of subjects (67%) exhibited labral abnormalities, even in the group without cam-type deformities. We cannot exclude some extent of overreading as an explanation of these findings. However, frequencies of labral signal alterations have previously been reported in asymptomatic individuals (22, 30, 31). Abe et al (22) found abnormal MRI signal intensities in 40 of 71 of volunteers (56%) with an age range of 13 to 65 years. Cotten et al (30) studied 52 hips in 46 asymptomatic volunteers ages 15–85 years. They found intralabral regions of intermediate or high signal intensity in 58% of the cases. Lecouvet et al (31) evaluated high-resolution T1-weighted spin-echo coronal MRIs of 1 hip in each of 200 asymptomatic individuals ages 15–82 years. They reported signal alterations in 44%. All 3 of the studies described above included both men and women, but no information was provided regarding sex differences. In our own experience, labral abnormalities are considerably less frequent in women than in men. Therefore, our results in men are likely to be compatible with those previously published for men and women combined (22, 30, 31). All 3 studies reported on variation in labral shape. Triangular shapes were described in 66% (31) to 88% (30) of participants, whereas other forms were reported in the remaining 12–34%. Lecouvet et al (31) found that a triangular shape was less prevalent in older people compared to younger people, and it is unclear whether deviations from a triangular shape are mere physiologic variations rather than pathologic changes.
Two previously published studies of conventional anteroposterior radiographs of the pelvis determined the association of pistol grip deformities with OA. In a population-based cross-sectional study in Copenhagen, Gosvig et al (32) reported an adjusted risk ratio of 2.2 (95% CI 1.7, 2.8) for the association between pistol grip deformities and hip OA. Doherty et al (33) published the results of a case–control study including cases with symptomatic radiographic hip OA and asymptomatic controls without radiographic OA. Based on conventional anteroposterior radiographs of the pelvis (33), they found an adjusted OR for the association of cam-type deformities with OA of 6.95 (95% CI 4.6, 10.4).
Although pistol grip deformities as observed on conventional radiography largely correspond to the cam-type deformities observed on MRI in our study, both of the previous studies were subject to the same two limitations. First, they used conventional anteroposterior radiographs of the pelvis to determine whether an individual showed a cam-type deformity, which underestimates the prevalence (12). This view predominantly depicts cam-type deformities at the superior position (34, 35), which corresponds to 11 o'clock and 1 o'clock in our MRI study; the majority of deformities that were found in the present study would therefore not have been observed in those previous studies (12, 36). This misclassification is likely to result in an underestimation of the association between cam-type deformities and OA. The second limitation was that, as pointed out by Doherty et al (33), both studies predominantly included older individuals in whom it is difficult to distinguish between genuine cam-type deformities already present at a young age and secondary osseous alterations developing later in life with the progression of OA. This may result in an overestimate of the true associations.
The biomechanical mechanism that could causally explain our findings is that the increased radius of the femoral head associated with cam-type deformity results in shear forces at the acetabular cartilage when the anterosuperior region of the femoral head enters the acetabulum during flexion and internal rotation. This can lead to chronic damage of the joint cartilage through abrasion and to accompanying alterations of labrum and bone, which may occur earlier in the process than initially assumed (13, 37). The high frequency of MRI signs in participants without anterolateral cam-type deformities, however, suggests that the etiology of these intermediate signs is multifactorial and that other factors, such as far-reaching ranges of motion or extensive shear forces at the hip, may contribute to the disease process. Longitudinal studies are needed to determine whether cam-type deformity is a risk factor for symptomatic hip OA in accordance with currently established classification criteria (38).
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Reichenbach had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Reichenbach, Leunig, Odermatt, Hofstetter, Ganz, Jüni.
Acquisition of data. Reichenbach, Werlen, Pfirrmann, Bonel, Jüni.
Analysis and interpretation of data. Reichenbach, Leunig, Nüesch, Hofstetter, Ganz, Jüni.
- 1Murray CJ, Lopez AD, editors. The global burden of disease: a comprehensive assessment of mortality and disability from diseases, injuries and risk factors in 1990 and projected to 2002. Geneva: World Health Organization; 1996.
- 5Definition and classification of osteoarthritis. In: Brandt KD, Lohmander LS, editors. Osteoarthritis. 2nd ed. Oxford: Oxford University Press; 2003. p. 1–8., .
- 9Correlation between internal rotation and bony anatomy in the hip. Clin Orthop Relat Res 2007; 460: 152–8., , , .
- 34Comparison of six radiographic projections to assess femoral head/neck asphericity. Clin Orthop Relat Res 2006; 445: 181–5., , , , .