Hip morphology and its relationship with hip strength, mobility and lower limb biomechanics: a systematic review in adults

–This systematic review (PROSPERO registration n.43640) aimed to summarise and determine the quality of evidence relating hip bone morphology to (i) hip strength, (ii) mobility and (iii) lower limb biomechanics during functional activities. A standardized search on MEDLINE/PubMed, Web of Science, ScienceDirect and Scopus resulted in 17 papers that met inclusion criteria: i) original investigations with a minimal sample of n=10, ii) studies on humans and iii) presence of at least one quantitative hip morphological parameter and one hip functional (i.e. strength and mobility) and/or one lower limb biomechanical parameter. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies tool with adaptations. Sixteen out of the 17 included studies showed high risk of bias. We observed that primary evidence pointed to the influence of hip morphology on hip mobility in the transverse plane. Specifically, positive correlations between femoral anteversion angle and range of internal hip rotation in physical examination were observed. Regarding biomechanical parameters, no clear evidence of association between hip morphology, and kinematic and kinetic parameters were found. Our results point to a field that is currently under explored and future studies with low risk of bias addressing these relationships are required.


INTRODUCTION
Over the last two decades, great attention has been directed to the hip in the assessment and investigation of lower limb injuries and disorders [1][2][3] . There has been an increasing body of literature suggesting that poor biomechanical performance within the hip joint, such as hip weakness and limited range of motion, is linked to the development of lower limb injuries, such as knee joint pain and osteoarthritis [2][3][4][5] , and to traumatic injuries, such as knee and ankle sprains [1][2][3][4][5][6] .
During weight bearing activities, increased hip internal rotation and adduction was shown to be a risk factor 1 and associated factor 7 for lower limb disorders. These findings have directed the focus of prevention and rehabilitation programs towards achieving control of hip internal rotation during functional gestures 8,9 . However, despite this fairly common practice, our understanding towards non-modifiable and modifiable factors related to abnormal hip biomechanics is incipient.
Among parameters that are not modifiable -not without surgery, the femur and hip bone morphology are likely strong determinants of the mechanical environment and behaviour of hip joints. Hip morphology varies depending on age, gender, ethnicity, activity level and developmental stage 10,11 . It is known that increased femoral anteversion influences lower extremity alignment during standing and also changes hip muscles' moment arms and hip range of motion 12,13 . Modifiable factors are those related to muscle strength, joint mobility and motor control 2 .
In some studies, deviations in hip morphology have been associated to the development of lower limb disorders 14,15 . However, this relationship is often non-specific, with many individuals presenting radiographic alterations and not evolving to lower limb injury or pain [11][12][13][14][15][16] . While, for example, cam morphology seems to be associated with hip osteoarthritis, odds ratio varies between 2.2 and 20.6 14,16 , with most people with cam morphology not developing hip osteoarthritis 11 . In the case of other morphological findings, such as pincer morphology, femoral and acetabular orientation, associations to injury and/or pain during functional daily activities seem to be even harder to draw from the literature 11,17,18 .
Given the complex relationship between hip morphology and injury, investigations focusing on intermediate parameters may help on the identification of subgroups of higher risk of becoming symptomatic and/or developing lower limb injuries. Characteristics that are worth considering may include hip muscle strength, hip mobility, and lower limb biomechanical behaviour during functional activities such as gait, squatting, etc. Investigations on that matter may help to clarify the level of association between changes in hip morphology and the development of musculoskeletal disorders and to identify modifiable intermediate risk factors to prevent, stop, or slow down disease progression and/or to avoid overtreatment.
This systematic review was designed to summarise and investigate the quality of current available evidence relating hip bone morphology to hip strength, mobility and lower limb biomechanics in humans.

METHODS
We performed this systematic review according to the recommendations contained in Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 19,20 . The protocol was registered within the International prospective register of systematic reviews (PROPESRO) under the number 43640.
After a careful preliminary assessment of the literature related to hip morphology and clinical and biomechanical parameters, the three following sets of key words were elaborated: i) morphology set, ii) imaging findings set and iii) functional parameters set. For keywords within a set, the command OR was used in order to represent a given construct, while, between sets, the command "AND" was used to limit the search to the association between constructs (Box 1). Box 1. Search strategy in MEDLINE (via Pubmed).

1#
hip OR femur OR femoral OR "iliac bone" OR acetabulum OR acetabular OR pelvis OR pelvic 2# morphology OR morphological OR morphologic OR anatomy OR anatomical OR "caput-collum-diaphyseal angle" OR "neck-shaft angle" OR "coxa valga" OR "coxa vara" OR "femoral version" OR "femoral anteversion" OR "femoral neck anteversion" OR "femoral retroversion" OR "femoral neck retroversion" OR "femoral shape" OR "acetabular version" OR "acetabular anteversion" OR "acetabular retroversion" OR "center edge angle" OR "angle of Wiberg" OR "Wiberg angle" OR "Wiberg's angle" OR "angle of Lequesne" OR "vertical-center anterior angle" OR "cross-over sign" OR "alpha angle" OR "anterior offset" OR "X-ray" OR radiograph OR radiographic OR radiography OR radiographical OR "computed tomography" OR CT OR "magnetic resonance" OR MRI OR "Lyon's protocol" OR "tibial tubercle-trochlear groove distance" OR "tibial tuberosity-trochlear groove" OR TT-TG Prior to the official search, these key words were used to verify in the Cochrane and PROSPERO database whether a similar systematic review had been published or registered. No relevant results were found. For the official search, four databases were screened from the date they were concepted to July 2016: MEDLINE/PubMed, Web of Science, ScienceDirect e Scopus. Last search was performed in July 2016. According to each database, filters were used ( Figure 1). Inclusion criteria were: i) original investigations with a minimal sample of 10 participants per group comparison, ii) studies on humans, iii) presence of quantitative hip morphological parameters and quantitative hip functional parameters and/or lower limb biomechanical parameters.
Articles that were not available in full format or articles that contained morphological findings that solely represent the progression of joint diseases were excluded from the analysis. No language restrictions were imposed. Three independent reviewers conducted the selection process. First, Reviewer #1 excluded duplicates, editorials, case studies, incomplete articles and non-original investigations. Second, Reviewers #1 and #2 analysed all titles and excluded unrelated titles. After reading all selected abstracts, articles that could potentially fit to the inclusion/exclusion criteria were evaluated in their full format and a final independent decision was made by reviewers. The list of references cited in each article was screened for potential inclusion in the review. Third, in case of disagreement between reviewers #1 and #2 that persisted after discussion, reviewer #3 was consulted.
Bias and quality analysis of the selected papers was performed using a set of questions that were adapted from the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool 21 . QUADAS contains 14 questions that are answered with "yes", "no" or "unclear". The instrument used in this systematic review was composed of 12 items (in contrast to 14 in the original QUADAS). Maximum score was 12 and meant that the paper met all requirements listed for a good quality. If the information was "unclear", the item was scored with 0 points, similarly to answer "no". As a cut point, articles that scored nine points or more, were considered of high quality, while articles that scored below nine were considered of poor methodological quality 22 .
For the analysis of associations between morphology and functional parameters the following interpretation of correlation scores was adopted: r values between 0 and 0.20 were considered indicative of a very poor correlation; r values between 0.21 and 0.40, a poor correlation; between 0.41 e 0.60, moderate; 0.61 and 0.80, high and r values greater than 0.80 were considered indicative of an excellent correlation 23 .

RESULTS
Up to the date this review was submitted (June, 2019), no systematic review addressing the relationship between hip morphological parameters and functional parameters of the lower limbs had been registered or published.
Muscle strength was assessed based on torque capacity or force against resistance. All contractions were isometric and the peak force achieved was the parameter estimated 34,37,39 .
Concerning the quality assessment of studies, raters independently agreed in 61.8% of the items (126) scored in the risk of bias scale, while, for the remaining 38.2%, agreement was met after discussion ( Table 1). The study with the lowest score 40 received 4/12 points and the one with the highest score 37 received 10/12 points. A representative sample of the population of interest was present in only seven studies 28,30,36,[38][39][40] , with most studies based on samples that were not randomly selected. In addition, in most studies (9 out of 17), the selection criteria were not clear (Item 2) 25 Goniometer, full-body inertial-based motion-capture system suit (inertialbased motion-capture system) and CT.
Passive  To compare the foot rollover process during running between male semiprofessional soccer players with increased alpha angles and agematched amateur soccer players.
14 male semiprofessional soccer players and 14 male amateur soccer players.
MRI, a piezoelectric force platform, an accelerometer, an electrogoniometer, regular running shoe and the same shoe with inserted valgus wedges. Alpha angle of the right hip (in all cases, the right leg was the kicking leg). In a biomechanical laboratory setting, each participant of both groups ran in two shoe conditions across a piezoelectric force platform. Running speed was controlled. Simultaneously, in-shoe pressure distribution (on seven anatomical locations of the foot: medial and lateral heel; lateral midfoot; first, third and fifth metatarsal heads; hallux), tibial acceleration and rearfoot motion measurements of the right foot were performed.

Alpha angle ( o ).
Loading rate (bw/s), peak tibial acceleration (g), median power frequency (Hz), peak vertical force (bw), peak horizontal force (bw), horizontal impulse (bw.s), maximum rearfoot motion (°), peak pressure lateral heel (kPa), peak pressure medial heel (kPa), peak pressure lateral midfoot (kPa), peak pressure metatarsal head V (kPa), peak pressure metatarsal head III (kPa), peak pressure metatarsal head I (kPa), peak pressure hallux (kPa). Three-dimensional motion analysis was performed using a computer-aided video motion analysis system. Ground reaction forces were obtained using force plates. Strength testing was performed using a dynamometer. Femoral shape was quantified using MRI. Kinematic evaluation during running at a fixed velocity. The kinematic variable of interest was average hip internal rotation angle during the first 50% of the stance phase of running.
Hip strength was performed in 4 different positions: standing pelvic drop (isometric, isokinetic, and isotonic endurance), seated hip external rotation (only isometric endurance), prone hip extension (isometric, isokinetic, and isotonic endurance) side-lying hip abduction (only isometric endurance). Femoral inclination and femoral anteversion hip.  In 15 out of 17 studies, the imaging exam used was considered appropriate for the morphological parameter of interest (Item 3). However, in 1 study 39 the measurement of the neck-shaft angle was performed using densitometry; in another study 40 , the alpha angle was measured using an anteroposterior incidence. Both imaging procedures may be considered inappropriate for the measurement of interest. In relation to the time between exams and the possibility of change in the parameters of interest, both reviewers agreed that the time was often irrelevant given the time required to modify a morphological measurement (Item 4). However, there were studies that involved participants with degenerative alterations, in which the time between exams could potentially influence results but was not mentioned 34 . In three studies 31,32,38 , the exam of interest was not applied to the entire sample and the choice of which participants would or would not take the exam was not random (Item 5).
In relation to the biomechanical measurements, nine studies 24,26,30,31,[35][36][37][38][39] described the procedures with enough detail to allow for replication (Item 6). The most commonly missing information in studies seemed to be the description of instruments 25,27,29 and patient positioning 29,33,39,40 used in the assessment of hip range of motion. With regards to the imaging exam procedure (Item 7), five studies did not seem to provide enough detail to allow for replication. Only one study clearly stated that measurements were blinded between the morphological and biomechanical tests (items 8 and 9) 25 .
Considering the sample participation, one study did not clearly describe what happened to all participants that were initially included in the study (item 10) 34 . Five studies 27,33,34,36,37 presented a clear descriptive analysis of data, with mean/median and dispersion statistics (Item 11). With regards to the inferential statistics (Item 12), seven studies 25,27,29,[32][33][34]37 introduced a correlation analysis.
Significant correlations between morphological and biomechanical parameters were found in some studies 25,27,29,30,35,36,37 , while in some others descriptive statistics allowed for the exploration of possible associations between morphology and the parameters of interest in this review 24,25,29,30,35,36,38,40. The significance, context and interpretation of these relationships are specified in the discussion section.

DISCUSSION
Based on the parameters evaluated, 16 out of the 17 studies performed poorly in the risk of bias and quality assessment with only one study reaching a score equal or above 9 in the scale and being considered of high quality 37 . We speculate the high risk of bias present in the literature regarding the relationship between morphological and biomechanical parameters is likely related to the difficulty in inspecting hip morphology in healthy control individuals and to the cost and ethical issues related to imaging tests. The results of our analysis emphasize the need of interpreting available evidence with caution. Significant correlations between morphological and biomechanical parameters were found in some studies 25,29,30,36 . Patients with small femoral offset were associated to a greater pelvic drop in the Trendelenburg test (r=0.416; p=0.0137) 26 . In a study with female runners 39 , a moderate correlation between neck-shaft angle and the strength of hip abductor muscles was found (r=-0.47; p=0.02) but not between neck-shaft angle and gait kinematic parameters.
Significant correlations were also observed in studies that focused in patients that had undergone hip video arthroscopy. In a study with 123 patients 27 , the femoral anteversion angle was found to be related to the range of internal rotation in physical examination (r=0.36; p<0.001). Ejnisman et al. 35 evaluated 188 patients and also found a similar correlation (r=0.231; p=0.002) accompanied with a negative correlation between femoral anteversion and external rotation (r=-0.208; p=0.027). It is important to note however that despite reaching statistical significance, all observed correlations are poor, indicating that other parameters may contribute largely. Finally, Souza and Powers 37 investigated biomechanical and morphological factors related to hip internal rotation angle at the first 50% of stance phase during running in female runners with patellofemoral pain syndrome. No significant relationship was found between hip internal rotation and femoral version or neck-shaft angle (p=0.11 and p=0.10; respectively).
Some studies presented descriptive analysis that allow for the exploration of possible associations between morphology and the parameters of interest in this review. Chadayammuri et al. 25 35 grouped patients in I) femoral version <5°; II) femoral version between 5 and 15°; and III) femoral version > 15°. Group I presented significantly greater range of movement towards external rotation (45±14°, 38±12° and 36±13° for group I II and III, respectively). No significant differences between groups were observed for internal rotation. Ferro et al, 29 using the same grouping criteria, also In general, there seems to be a significant relationship between femoral and acetabular version angles and hip mobility in the transverse plane. Most of the evidence available points to a small effect on range of motion with little being known about possible effects of morphology on movement biomechanics. Although there is a theoretical expectation of hip morphology to affect strength at a certain degree due to changes in moment arms, only one study was found supporting experimental evidence towards this hypothesis 39 .
Some limitations should be considered in this systematic review. Although no restrictions in language were imposed to the search, the fact that only English keywords were entered may have restricted access to some non-English papers. With regards to the instrument used to access quality and risk of bias, although the original QUADAS scale from which the instrument was created has been extensively used, the adaptations made for the purpose of this review have not been validated. Finally, we used a cut point of 70% of affirmative responses to define a study of high quality 22 , while there has been criticism on the use of scores to classify studies on quality 21 .

CONCLUSION
This study assessed the quality and summarized the evidence available on the role of hip morphology in determining hip strength, mobility and the biomechanics of lower limb movement during functional and sport activities. Literature in the area is limited and the studies available are mostly accompanied by a high risk of bias. Primary evidence seems to be focused on the influence of hip morphology on the hip mobility in the transverse plane, primarily on changes in range of motion. With regards to biomechanical parameters that could provide insights into the association between hip morphology and kinematic and kinetic characteristics of movement, we were not able to draw a clear conclusion based on the available findings. Our results point to a field that is currently under explored and future studies addressing these relationships will be beneficial to our understanding of modifiable and non-modifiable parameters related to lower limb disorders.

Funding
This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior -Brasil (CAPES) -Finance Code 001.

Ethical approval
This research is in accordance with the standards set by the Declaration of Helsinki.