Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Both patients and physicians often use the terms arthritis, arthrosis, and osteoarthritis interchangeably. Arthritis is, however, a nonspecific term that denotes inflammation of a joint, whereas arthrosis is defined as a degenerative affliction of a joint [1]. In contrast, osteoarthritis is a distinct pathologic process: “arthritis characterized by erosion of articular cartilage, either primary or secondary to trauma or other conditions, which becomes soft, frayed, and thinned with eburnation of subchondral bone and outgrowths of marginal osteophytes; pain and loss of function results” [1]. In this vein, the terms degenerative joint arthritis, degenerative joint disease, and osteoarthrosis are true synonyms for osteoarthritis.

Although there is debate about the factors that initiate osteoarthritis, the pathologic process is characterized by progressive loss of the articular surface (Fig. 1) [2]. Initially there is cartilage fissuring, chondrocyte clustering, and some attempt at repair [2]. In this early state when the cartilage damage is confined to the articular surface and there is no associated subchondral reaction, this could also be considered “arthritis” as there is no “osteo” component. As the degenerative process progresses, the subchondral bone remodels and appears sclerotic on radiographs. Among osteoarthritis researchers, there are competing theories about the cause of these subchondral bone changes and whether they occur in response to the cartilage damage or if they occur in response to increased load even before the cartilage has been macroscopically damaged. Nonetheless, as the joint degenerates further, the process is consistent. The synovium and capsule thicken, marginal osteophytes form, and the subchondral bone may develop cysts.

Fig. 1
figure 1

Normal aged cartilage is biomechanically different from normal young cartilage. Over time, the surface (STZ: superficial tangential zone) becomes rougher and there are fewer chondrocytes. Degenerative aged and/or arthritic cartilage has deep surface clefts and fibrillation with chondrocyte clustering. In this diagram, the degree of gray shading indicates biomechanical integrity and loss of glycosaminoglycan (GAG). The table shows the location of changes in the STZ, middle zone (MZ), and deep zone (DZ). (•) indicates the full presence of changes, (ο) the complete absence, with the arrow indicating a change between cartilage states (Reprinted from Osteoarthr Cartilage, Vol. 17, Temple-Wong MM, Bae WC, Chen MQ, Bugbee WD, Amiel D, Coutts RD, Lotz M, Sah RL. Biomechanical, structural, and biochemical indices of degenerative and osteoarthritic deterioration of adult human articular cartilage of the femoral condyle, 1469–1476, Copyright (2009), with permission from Elsevier)

Full size image

Although osteoarthritis is not an inflammatory process in the same sense as the rheumatologic diseases that cause joint destruction, inflammation is clearly part of what causes radiographic osteoarthritis to become painful. When osteoarthritis becomes symptomatic, patients complain of joint pain, decreased motion, effusions, and crepitation and, in more advanced cases, may notice deformity due to ongoing bony destruction. Thus, to gather all of these concepts into a broad definition, osteoarthritis should be defined as a degenerative process where continued cartilage breakdown results from mechanical overload, causing secondary bony and synovial changes and characteristic clinical and radiographic findings.

In the hip, there are many new ideas about anatomic and biomechanical factors that may ultimately cause osteoarthritis (OA), and the basic science in this area is evolving rapidly. When evaluating a potential risk factor or cause of a disease, the Bradford-Hill criteria are helpful for determining if an association between a risk factor and a disease is actually a cause-and-effect relationship [3]. These criteria consist of the following:

  • Strength of association: This refers to the relationship between the possible cause and effect. If there is a stronger relative risk of developing a disease for a patient with a particular risk factor, the risk factor is more likely to be a causal factor. Occasionally, however, the observed association is slight, and the risk factor is nonetheless proven to be a cause of a disease.

  • Consistency: This means that the same association is observed repeatedly, in studies that occur in different populations, with different study designs, and by different observers.

  • Specificity: This describes how precisely a potential risk factor can predict that the disease will occur. It is important, however, to keep in mind that diseases may have more than one cause and that one-to-one relationships between a risk factor and a disease are infrequent.

  • Temporality: This means that the proposed risk factor for the disease always precedes the disease.

  • Dose–response effect: This means that the frequency of the disease increases with the dose or level of exposure. In orthopedics the dose or level exposure can also be the magnitude of a deformity.

  • Biological plausibility: This means that, with what is currently known about biology or biomechanics, the proposed risk factor could reasonably cause the disease. Nonetheless, it is important to keep in mind that sometimes the basic science also needs to advance to elucidate the relationship between the cause and effect.

  • Coherence: This means that the proposed association should not contradict current knowledge about the natural history and biology of the disease.

  • Experimental evidence: This means that an experiment validates the cause-and-effect relationship in the expected manner. For example, modifying a risk factor decreases the frequency of a disease, or addressing the proposed cause of the disease brings about a cure.

  • Analogy: This is the process of thinking about a proposed risk factor by comparing other similar and known cause-and-effect relationships for a particular disease. Reasoning by analogy can help to ascertain a cause-and-effect relationship when the observed association is slighter but similar to a known effect.

Hill himself cautioned, however, that these criteria are not necessary or sufficient for making a causal judgment and should be used more as guidelines for considering whether an observed risk factor truly causes a disease [3]. He also reminded the reader that “the ‘cause’ of illness may be immediate and direct, it may be remote and indirect underlying the observed association.” Thus, returning to the question of the etiology of hip OA, it is more likely that the “cause” is multifactorial and different for different individuals. Furthermore, the Bradford-Hill criteria provide a useful framework for evaluating current hypotheses and evidence about the etiology of hip OA.

Although global prevalence of radiographic hip OA varies considerably, it is a common condition in the United States and Europe. The lifetime risk of developing symptomatic hip OA has been estimated to be as high as 25 % after adjusting for race, body mass index, sex, and prior injury [4]. However, not everyone who has radiographic evidence of hip OA becomes symptomatic. In one recent study, only 20 % of people with radiographic hip OA eventually became symptomatic enough to require total hip arthroplasty [5]. The natural history of asymptomatic radiographic hip OA is, however, difficult to elucidate because it requires long-term prospective cohort studies of large populations. Furthermore, the number of patients who progress to arthroplasty is likely to increase because many middle-aged and elderly patients expect to remain active indefinitely and would rather undergo arthroplasty than modify their activities. Age is one of the known risk factors for developing hip OA, and the incidence of hip OA increases with age. Not only does cartilage accumulate damage over time, but older mesenchymal stem cells also have less repair capacity and a decreased ability to protect cartilage from biomechanical stress [6]. Thus, as the expected human lifespan increases, the amount of hip OA and rates of hip arthroplasty are also projected to increase [7].

Other risk factors for hip OA include physical activity like long-term frequent lifting and standing [8] as well as intense or impact sports in young adulthood [9, 10]. There is an association between higher body mass index (BMI) and hip OA, although this association is much weaker than the association between BMI and knee OA [11]. Sex also appears to be a risk factor, with women having higher rates of hip OA than men [4]. Finally, family history and known congenital deformities have also been categorized as risk factors for hip OA.

Historically, hip OA was categorized as primary or idiopathic and secondary, meaning that the hip became arthritic as a result of a prior traumatic injury, pediatric deformity, or following infection. Primary or idiopathic hip OA was attributed to having “bad genes” – essentially that the patient had inherited weak cartilage. However, evidence is accumulating that primary hip OA is actually secondary to a subtle mechanical problem like femoroacetabular impingement (FAI) or mild dysplasia. The concept that hip OA is a mechanical process was proposed by a number of authors and summarized nicely by Ganz in 2008:

Most, if not all, hip OA is secondary, often secondary to subtle but definite and commonly overlooked, ignored, or not recognized dysplasia or pistol grip deformities (FAI). [12]

The Genetics of Hip Osteoarthritis

While there is clearly an inheritance pattern to hip OA, the nature of the genetic contribution is not entirely known. Have family members with a history of hip OA all inherited a bone structure like dysplasia or FAI that causes chondral damage and subsequent OA, or have they simply inherited less rigorous cartilage that is more likely to be damaged in the setting of subtle FAI or dysplasia? Or, as seems likely, is it some combination of the two factors?

Twin studies done in Caucasian females found a genetic contribution of about 60 % for both center-edge angle (as a measure of acetabular depth) and radiographic hip OA [13]. The magnitude of the genetic contribution is not the same for other joints, meaning that the etiology of OA is likely specific to mechanical factors and anatomy at that joint. This also implicates morphology rather than poor-quality cartilage as the bigger risk factor for hip OA. Other studies have shown that femoral head shape is heritable in families with a history of arthroplasty for “idiopathic” OA. However, in one of these studies, patients with a positive family history were more symptomatic than patients with the same degree of FAI morphology but no family history of hip OA. This suggests that bony morphology may not be entirely responsible for symptom development [14]. Genes have been identified that are associated with both cartilage thickness and hip shape [15, 16]. These genes are expressed in developing limb buds and in developing cartilage as well as being expressed in response to increased biomechanical loads [15, 16]. Thus, the genes associated with hip OA could affect either the hip shape or the cartilage microstructure. Finally, genetic variability influences the association between hip OA and bony morphology, meaning that certain genotypes appear to make the cartilage more vulnerable to mechanical overloading from subtle FAI or dysplasia [15].

Acetabular Dysplasia

Acetabular dysplasia is defined as a shallow or small acetabulum that inadequately covers the femoral head. Moderate to severe acetabular dysplasia has long been recognized as a risk factor for the early development of hip OA [17]. The risk of developing hip OA due to mild or borderline dysplasia is less clear, however, and may be influenced by external factors like soft tissue laxity, femoral version, and sport or dance activities.

Although dysplasia has historically been thought of in the context of infantile hip subluxation or dislocation, there is growing recognition that adolescent- or adult-onset dysplasia may represent a developmental process distinct from infantile dysplasia [18]. Furthermore, very few younger adults undergoing hip arthroplasty for arthritis secondary to dysplasia are identified as neonates [19]. The demographics of the infant and adolescent dysplasia populations are different, with adolescent-onset dysplasia patients having more bilateral hip involvement, a stronger family history, and a higher proportion of male patients [18]. Infantile dysplasia may represent a “packaging problem,” meaning that mechanical factors play a greater role in the shape of the neonatal acetabulum and containment of the femoral head. The risk factors for infantile dysplasia – breech positioning, left-sided laterality, and first-born females – implicate the intrauterine environment as a mechanical factor influencing acetabular development. Furthermore, the historical prevalence of dysplasia was substantially higher in populations that had a tradition of infant swaddling with the legs in extension. When this connection was recognized and parents were instructed not to swaddle their children, the incidence of dysplasia decreased [20].

The prevalence of acetabular dysplasia varies widely [20]. It is somewhat difficult to compare the prevalence of dysplasia between countries or regions because some studies have evaluated adults whereas for some populations the data are only available for infants. In addition, some studies have defined dysplasia as a center-edge angle of <25° whereas others have used a center-edge angle of <20°. Nonetheless, it is well recognized that the prevalence of dysplasia is higher in Asia and is the most common cause of hip OA in Japan [21].

A family history of dysplasia is a known risk factor for dysplasia and is consistent across all studied populations. Dysplasia is even more prevalent in areas where consanguinity (e.g., marriage between first cousins) is common [20]. Twin studies have revealed that the heritability of dysplasia is likely polygenic, with a higher incidence of dysplasia in monozygotic twins as compared to dizygotic twins. These findings have led investigators to propose that the genetic mechanism involves inheritance of excessive soft tissue laxity as well as acetabular shape [20].

Biomechanics of Dysplasia

In normal hips, the peak cartilage contact pressure when standing is located near the acetabular dome. The peak contact site varies between the lateral edge and the superior dome of the acetabulum, becoming more medial if the acetabulum is deeper and more lateral if the acetabulum is shallow [22, 23]. There is a direct relationship between the degree of acetabular coverage (as measured using the center-edge angle) and the contact area of the acetabular surface. As the contact surface area decreases, the peak contact pressure increases – meaning that a small center-edge angle is a marker for higher contact pressure [22]. This translates to increased force on the acetabular rim, particularly in stance, and causes characteristic chondrolabral pathology, including labral tears, ganglia, and, in some cases, acetabular rim fractures [24]. The tissue loss predictably occurs at the superior and anterosuperior regions of the acetabulum [24], which corresponds to the area of the highest load [22]. Acetabular version also influences contact pressures. Highly anteverted dysplastic hips have higher anterior contact stresses as a result of minimal anterior femoral head coverage [25]. In contrast, patients with dysplasia and retroversion have impingement-type contact stresses at the anterior edge of the acetabulum. Correcting the version and coverage with an acetabular reorientation osteotomy has been shown to decrease contact pressure by up to 50 % [23].

Natural History of Dysplasia

Radiographic dysplasia, variably defined as a center-edge angle of <20° or <25°, is clearly associated with an increased risk of hip OA [17]. The risk of developing hip OA is also clearly related to the grade of dysplasia, indicating that hips with worse biomechanics and higher contact pressures have a higher likelihood of sustaining joint damage and ultimately becoming arthritic (Fig. 2) [17]. If hip pain in a young person (<40) is considered to be a precursor of hip OA, it is notable that 25–35 % of young patients with hip pain have dysplasia [26]. Version may also play a role in the natural history of dysplasia. Patients who have retroversion and dysplasia experience an earlier onset of hip pain as compared to patients with normal anteversion [27].

Fig. 2
figure 2

The natural history of a patient with severe left anterior dysplasia (outline). At the time of presentation, the hip had already subluxed laterally, but the joint space was relatively preserved (a). Thirteen years later, there was advanced and severe cartilage degeneration on the femur, with a subchondral cyst and sclerosis (arrow) in the femoral head, as well as on the acetabulum (b). She underwent a total hip arthroplasty 1 year later (c)

Full size image

If the loading biomechanics of a dysplastic hip are changed as a result of a femoral or acetabular osteotomy, the natural history of that hip appears to improve. The results are better for periacetabular or rotational acetabular osteotomy than for femoral osteotomy however. The long-term (20-year) survival rate of the native hip after a periacetabular osteotomy is about 60 % [28]. Even with the improvement in hip biomechanics, most patients have some progression of osteoarthritis and, on average, advance one radiographic Tönnis grade after 10 years [28]. Because dysplasia is largely a problem related to static loading across the hip, one might expect that weight loss in an overweight patient with dysplasia could improve hip pain and natural history because it decreases the overall static load. Although weight loss is known to improve pain and function in patients with knee OA, this has not been studied for patients with dysplasia. The potential for an osteotomy to improve hip function does have some limits. The success rates of osteotomy are poor for patients older than 35 with Tönnis grade 2 or more radiographic hip OA [28]. Thus, there appears to be a “tipping point” of cartilage damage, after which an osteotomy is unlikely to improve the natural history of a dysplastic hip.

Femoroacetabular Impingement (FAI)

Broadly speaking, FAI is defined as the abnormal contact between the femur and the acetabulum during hip motion that occurs as a result of a subtle deformity at the femoral head-neck junction or acetabular rim and that causes progressive chondrolabral damage [29]. Although Reinhold Ganz generally receives credit for describing FAI [29], the earliest description of impingement appears to have been in 1899 in the French literature, with the author noting “empreinte iliaque,” or an impression on the head-neck junction produced by the ilium at the area of the anterior-inferior iliac spine with the hip in flexion [30]. Subsequent authors correctly described impingement in the context of coxa vara, severe protrusio deformities, and slipped capital femoral epiphysis [31–33]. Depending on the site of the deformity, these authors also recommended femoral neck osteoplasty and/or acetabular rim trimming to restore range of motion and provide pain relief. In contrast to all of the previous authors however, Ganz substantiated his ideas about FAI with observations of chondrolabral damage at the site of the impinging lesions and with the results of treatment [29, 34], both of which were made possible with the development of the technique for a safe surgical dislocation of the hip. The description of FAI also coincided with technical improvements in hip arthroscopy that resulted in an increase in hip arthroscopy for labral tears. As a result, arthroscopists began to recognize and describe early chondrolabral damage, which ultimately helped to substantiate the association between hip pain, impingement anatomy, and eventual hip OA [35].

FAI is broadly grouped into cam and pincer types of impingement, which have different mechanisms of damage and different prognoses for the cartilage. Considering the mechanical type of impingement injury is a useful way to think about FAI because it allows for the realization that different anatomic abnormalities can cause the same type of impingement. Although cam impingement can result from many distinct anatomic abnormalities, it ultimately causes an inclusion type of injury where a bony deformity at the head-neck junction enters the joint with hip flexion. Most commonly, the abnormality is a “cam deformity” which occurs as a result of an extension of the physis onto the femoral neck [36] causing either decreased head-neck offset or a prominence at the head-neck junction. However, the femoral head deformities that occur after Legg-Calve-Perthes’ disease and mild or moderate slipped capital femoral epiphysis also cause cam-type impingement and can be considered extreme examples of cam impingement [33, 37]. In cam impingement, the deformity at the head-neck junction causes shear stress and delamination of the acetabular cartilage with separation at the chondrolabral junction [29, 34]. This type of impingement has a worse prognosis for the cartilage and can cause end-stage arthrosis in a relatively young (40–50-year-old) adult. Although the cartilage over the non-spherical portion of the femoral head is abnormal, with histologic changes like cell clustering and surface fibrillation that are consistent with early arthritis [38], the macroscopic chondral damage occurs initially on the acetabulum. The weight-bearing cartilage on the femoral head remains relatively preserved until the acetabular chondral defect advances to the point that the femoral head migrates into the defect. At this time, the chondral damage becomes radiographically apparent, with visible joint space narrowing on x-rays. Pincer impingement, in contrast, causes an impaction type of injury with hip flexion, with the acetabular rim contacting the femoral head, neck, or metaphysis. Global acetabular overcoverage and focal acetabular overcoverage from acetabular retroversion are the two more classic causes of pincer impingement. However, a prominent anterior-inferior iliac spine can also cause rim impingement [39], as can acetabular protrusio [40] and a severe SCFE deformity [33]. The rim impaction causes a crushing injury to the labrum and a linear wear pattern of cartilage damage and, over time, causes rim ossification [29, 34, 41]. In addition, a “pincer groove” is often visible on the femoral neck. Although pincer impingement may not cause chondral damage as rapidly as cam impingement, the crushing injury to the labrum appears to be quite painful for the patient. A smaller number of patients with pincer impingement have femoral levering on the acetabular rim, causing contrecoup injury to the cartilage in the posterior acetabulum [29, 34]. Patients with true acetabular protrusio also develop medial cartilage thinning [40], which may be a result of increased medial contact pressure. While the distinction between cam and pincer FAI helps to explain the nature of the observed cartilage injuries, in practicality most patients with FAI have mixed cam and pincer morphotypes [29].

Biomechanics of FAI

Impingement can be observed directly during a surgical hip dislocation. Nonetheless, these observations have also been validated with finite element analysis of the cartilage contact forces during hip flexion. When the deformity at the head-neck junction is increased (by increasing the alpha angle), the non-spherical portion of the head intrudes into the acetabulum, causing increased cartilage stress on the anterior acetabulum at the site of the cam deformity. In a similar manner, increasing the amount of acetabular coverage (by increasing the center-edge angle) causes higher contact stresses at the acetabular rim and contact with the femoral neck.

Natural History of FAI

There is clearly some heritability for impingement-type anatomy, although the genetic influence may not be as strong for FAI as it is for dysplasia. Interestingly, cam morphology seems to be more heritable than pincer morphology. One sibling study observed a relative risk of 2.8 for inheriting cam-type anatomy and a relative risk of 2.0 for inheriting pincer morphology [14].

Although the evidence that FAI ultimately causes hip OA seems convincing, it is indirect (Fig. 3). Labral tears and FAI morphology are frequent in asymptomatic volunteers [42, 43]. What remains unknown about these populations is whether the subjects are asymptomatic because they are in an early stage of the disease process and have minimal chondrolabral damage or if not all FAI ultimately progresses to become symptomatic. All of the currently available natural history studies are level III or IV prognostic studies based on pelvic radiographs [44, 45]. The rates of radiographic progression for patients with FAI are quite variable, ranging from 18 % to 73 % over 10 years [44, 45]. However, FAI morphology was found in nearly all (96–99 %) hip arthroplasty patients <55 years old who were previously diagnosed with primary or idiopathic hip OA [45].

Fig. 3
figure 3

This patient presented with bilateral FAI from cam deformities and acetabular retroversion on the right. At the initial presentation, there was acetabular subchondral sclerosis, but no joint space narrowing (a). Fifteen years later he had complete joint space destruction in both hips (b). Twenty years after his initial presentation, there are extensive cystic changes and femoral head collapse in both hips (c)

Full size image

Studies of hip OA in athletes provide evidence that OA may result from a combination of FAI and abnormal loading or motion requirements. Compared to the general population, both male and female athletes have higher rates of hip OA. Contact sports and higher exposure to sports increase the risk of hip OA [9, 10]. A few studies have looked at the prevalence of hip OA in former professional dancers. Here the effect is less clear; one study showed an increased incidence of hip OA in former dancers [46], whereas a later study found no difference in rates of hip OA between dancers and the general population [47]. One potential reason for this may be that the range of motion and the amount of trained soft tissue laxity required for dancers weed out patients with impingement morphology before they reach elite or professional levels.

In addition to higher rates of hip OA, high-level athletes also have a higher prevalence of FAI morphology compared to the general population. This was actually first observed in the 1970s but was described as a “tilt deformity” and attributed to a mild subclinical SCFE [48]. More recently, cam deformities were found in 78 % of US collegiate football players, and a radiographic crossover sign was observed in 61 % of these same athletes [49]. Both professional and adolescent soccer players had high rates of FAI-type anatomy, with 72 % of the male professional players and 50 % of the females having some radiographic finding consistent with FAI [50]. Among asymptomatic professional and collegiate hockey players, 39 % had an elevated alpha angle but 77 % had hip and groin abnormalities on MRI [51]. Finally, a study of elite-level basketball players found that 89 % had an elevated alpha angle [52]. Cam deformities appear to occur from an extension of the femoral physis onto the femoral neck. Thus, one cause of high rates of FAI and cam deformities specifically in athletes may be the frequent high-intensity sporting activity itself. High-intensity sports have been shown to affect the proximal humeral physis and glenoid version in the young thrower as well as the distal radial physis in the gymnast. Similarly, repetitive rotational stress across the hip and proximal femoral physis as it is closing may cause the high rates of cam deformity seen in these athletes.

Summary and Conclusions

End-stage hip OA is caused by a combination of a structural deformity (either dysplasia or FAI), wear caused by the motion or activity required from an individual’s hip, the inherent “robustness” of the individual’s cartilage, and the amount of inflammation that the individual experiences. It is clear that not all radiographic hip OA is equally symptomatic and that many, but not all, people with FAI or dysplasia ultimately develop hip OA. The understanding of the structural factors that contribute to hip OA is advancing rapidly. It also appears that identification and treatment of FAI and dysplasia appears to help symptoms that result from early chondrolabral damage. There is good evidence that changing the biomechanics of the dysplastic joint with an acetabular osteotomy changes the natural history of the disease. If this occurs before the cartilage damage has advanced, acetabular reorientation may prevent end-stage OA or at least considerably delay an eventual arthroplasty. For FAI, there is evidence that correcting a cam or pincer deformity improves the symptoms from early OA. Although it seems likely, it is not yet known if surgical treatment can change the natural history of FAI and prevent progression of hip OA. One caveat, however, is that causing further chondral damage with surgery or incomplete treatment of these structural factors, e.g., not recognizing dysplasia in a patient with a cam deformity, might not be helpful and may incite the cascade of OA. Thus, the correct diagnosis and meticulous care of the cartilage are important when treating these patients in an attempt to prevent or delay the onset of hip OA.