When does the patella dislocate? A systematic review of biomechanical & kinematic studies

Abstract

Background

Patellar dislocations are a significant injury with the potential for long term problems. Little work has been done on establishing the mechanism by which this injury occurs.

Objectives

To determine the mechanism of injury of a patella dislocation based on the available published literature and compare them to already proposed theories.

Methods

A systematic review of the literature was conducted following searches performed on MEDLINE, EMBASE and ProQuest from the earliest year of indexing using the following search terms in any combination: “patella”, “dislocation”, “mechanism of injury”, “anatomy”, “biomechanical” and “risk factor”. A broad inclusion criteria was used that included studies that looked at patellar dislocations and instability with respect to the patellofemoral joint (PFJ) kinematics or altered kinematics of the PFJ. Studies that did not address the kinematics or biomechanics of the PFJ were excluded. Studies were appraised based on their methodology using a combination of the Critical Appraisal Skills Programme tool and the Quality Appraisal for Cadaveric Studies.

Results

113 studies were identified from a search of MEDLINE, EMBASE and ProQuest databases. Following application of our inclusion criteria, a total of 23 studies were included in our review. 18 of these studies were cadaveric biomechanical studies. The remaining studies were anatomical, imaging based, and a computer simulation based study.

Conclusions

These biomechanical and kinematic studies provide some evidence that a dislocation is likely to occur during early knee flexion with external rotation of the tibia and contraction of the quadriceps. There is limited evidence to support other elements of proposed mechanisms of dislocation.

1. Introduction

Patella dislocations account for 3% of all knee injuries¹ and is associated with young and active individuals.² An injury such as this can result in lifelong problems for the individual.³ A great deal of work has been conducted in recent years in establishing the internal and external risk factors that result in this injury. An important aspect of injury prevention requires not only to establish these risk factors, but also to determine a mechanism of injury as suggested by Van Mechlen.³ Through determination of the mechanism of injury, it is possible to develop injury prevention programmes and improve rehabilitation protocols. However, the exact mechanism by which a patella dislocates remains unclear.

The aim of this systematic review is to assimilate the findings of the biomechanical and kinematic studies that further enhance our understanding of the mechanism of injury by which a patella dislocation may occur and compare these to proposed theories to assess their feasibility. This may allow further development of current injury prevention protocols.

2. Method

An electronic database search was conducted on MEDLINE, EMBASE and ProQuest from the earliest year of indexing using the following search terms in any combination: “patella”, “dislocation”, “mechanism of injury”, “anatomy”, “biomechanical” and “risk factor” on 6^th April 2017 and then again on June 15, 2017.

A further manual search of the references of all papers and review articles was performed in order to identify any further relevant studies. This methodology was used to maximize the number of studies that could be identified and therefore included.

2.1. Inclusion criteria

Level I and II studies cannot be performed for mechanism of injury studies due to ethical considerations of causing such an injury. Therefore, given the likely low number of studies performed on this topic, lower levels of evidence were included with a primary focus on studies that provide insight into the behaviour of the patella under various conditions that would assist us in formulating a potential mechanism of injury. Therefore, only studies published in the English language that looked at patellar dislocations and instability with respect to the patellofemoral joint (PFJ) kinematics or altered kinematics of the PFJ were considered. No restriction was placed on when the article was published.

2.2. Exclusion criteria

Studies that do not address the kinematics or biomechanics of the PFJ were excluded. Furthermore, studies that assessed anatomical risk factors were also excluded as they would bias the interpretation and applicability of the results.

2.3. Review & analysis

The abstracts of all studies identified by the search were assessed with a particular focus on aims and study design. Papers identified at this stage were assessed according to the inclusion criteria.

No formal assessment tool for cadaveric or biomechanical studies currently exists. Therefore, studies were appraised based on their methodology using a combination of the Critical Appraisal Skills Programme (CASP) tool⁴ and the Quality Appraisal for Cadaveric Studies (QUACS).⁵ These tools were modified by removing elements that were not felt to be relevant in order to provide a framework for assessment of all studies in this review. Data was extracted using a structured proforma based on this combined appraisal tool.

3. Results

One hundred and thirteen studies were identified in our search. Following review of the abstracts, papers and references a total of 23 papers were included. The main reason for studies being excluded was that they did not address the kinematics of the PFJ with respect to instability or they were assessing the incidence of anatomical and biomechanical risk factors. The PRISMA flow diagram of the literature can be seen in Fig. 1.

Fig. 1.

PRISMA flow diagram of literature search.

The most common type of studies identified were cadaveric biomechanical studies which represented 18 of the 24 studies. There were 2 anatomical studies, 2 imaging based studies and 1 computer simulation based study. The results of the appraisal of these studies can be seen in Table 1.

Table 1.

Critical appraisal table of methodological quality of included studies.

	Objective Stated	Appropriate Study Design	Basic Information About Specimens/Participants	Conditions Of Specimens	Study Protocol Clearly Stated	Exposure Accurately Measured	Outcome Accurately Measured	Results Presented Thoroughly	Stats Appropriate	Limitations/Confounding Factors Discussed	Clinical Implications Discussed	Conclusion In Keeping With Results	Results Fit With Other Studies
Senavongse212005	✓	✓	✗	✗	✓	✓	✓	✓	✓	✓	✓	✓	✓
Philipott152012	✓	✓	✗	✓	✗	✓	✓	✓	✓	✓	✓	✓	✓
Brunet13 2003	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Yamada142007	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Goh26 1995	✓	✓	✗	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Raimondo271998	✓	✓	✓	✗	✗	✗	✗	✓	✓	✓	✓	✓	✓
Balcarek342013	✓	✓	✓	✓	✓	✓	✓	✓	✗	✓	✓	✓	✓
Conlan191993	✓	✗	✓	✗	✓	✓	✓	✓	✓	✓	✓	✓	✓
Desio201998	✓	✓	✗	✗	✓	✓	✓	✓	✓	✓	✓	✓	✓
Hautmaa251998	✓	✓	✗	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Mountney242005	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Burks23 1997	✓	✓	✓	✗	✓	✓	✓	✓	✗	✓	✓	✓	✓
V. Kampen161990	✓	✓	✗	✓	✓	✓	✓	✓	✗	✓	✓	✓	✓
Munro35 2012	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Diederichs312013	✓	✓	✓	✓	✓	✓	✓	✓	✓	✗	✓	✓	✓
Kan30 2009	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Sakai32 1994	✓	✓	✓	✓	✓	✓	✓	✓	✓	✗	✓	✓	✓
Farahmand221998	✓	✗	✗	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Heegard31 1994	✓	✓	✗	✗	✓	✓	✓	✓	✗	✗	✓	✓	✓
Nagamine331995	✓	✓	✗	✗	✓	✓	✗	✓	✓	✗	✓	✓	✓
Hefzy171992	✓	✓	✗	✗	✓	✗	✓	✓	✓	✗	✓	✓	✓
Farahmand282004	✓	✓	✗	✗	✓	✓	✓	✓	✓	✓	✓	✓	✓
Zaffagnini122013	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Farahmand291998	✓	✓	✗	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓

4. Discussion

4.1. Patellar tracking

The theory that the patella will dislocate during the early phases of flexion is strongly supported by a number of the biomechanical studies that look at patellar tracking. A total of 8 studies assessed patella tracking as the knee flexes and its possible relationship to instability. The results of these studies can be seen in Table 2.

Table 2.

Studies demonstrating findings of normal patella kinematics.

Study	Tilt	Shift	Flexion	Rotation
Heegard et al.10	Medial for first 20° of flexion and then lateral up until a maximum of 100°	Decreased medial shift with external tibial rotation	Flexion of patella increased with knee flexion but not as much	Not assessed
Nagamine et al.11	Initial tilt was medial but then became lateral from 45 to 90°	Initial shift medially before moving laterally	Not assessed	Rotated from neutral rotation to a medial rotation as it progressed through knee flexion
Zaffagnini et al.12	All tilt was lateral with passive range of motion until 85° at which point it became medial	Initial medial shift of patella which then begins to move laterally after 20–25°	Not assessed	Not assessed
Brunet et al.13	Medial tilt for first 22° then lateral in eccentric and concentric muscle contraction	Medial shift for first 22° then lateral	Flexion followed the knee joint but at a slower rate	No consistent pattern
Yamada et al.14	Lateral tilt began at 20° of flexion	Direction of shift not commented upon	Increased with knee flexion	Varied up until 20° flexion and then became lateral
Philipott et al.15	Minimal tilt up until 45° at which point it became lateral	Translates laterally with flexion reaching its maximum at 90°	Increased with knee flexion in	Patella rotates medially up until 30° and then rotates laterally
Van Kampen et al.16	Described as ‘wavering’: medial to lateral back to medial again with flexion	Initial medial shift, that lateralises with increasing knee flexion	Flexes with knee flexion	Not significant in first 40° of flexion
Hefzy et al.17	Patella tilts medially for first 20–30° of flexion before becoming lateral	Shifts laterally with knee flexion	Flexes with knee flexion	Did not comment on normal patella rotation

Patella shift has been stated as being the most important aspect of PFJ kinematics when it comes to instability.⁶ Seven of the eight studies commented on shift. All seven reported that following an initial shift medially, the patella then begins to move laterally. However, only 2 studies take this one step further and gave details about the point at which this lateral shift begins: between 20 and 25°.^7,8 Yamada⁹ did not comment on the direction of shift.

There were similar findings for tilt in 6 of 8 studies, but the point at which the movement became lateral was variable. Heegard,¹⁰ Brunet⁸ and Yamada⁹ all reported that this lateral tilt began at 20–22° of flexion. Nagamine¹¹ reported that the lateral tilt did not commence until 45° of flexion, whereas Zaffagnini⁷ found that tilt is lateral up to 85°. Van Kampen⁶ did not comment on when the lateral tilt began.

The findings from these studies tend to suggest that movement of the patella is lateral from approximately 20°. This would be in keeping with the proposed mechanism that dislocation predominantly occurs in the early phase of flexion. These studies looked at movement from extension to flexion except for Brunet⁸ who looked at both concentric and eccentric muscle contractions. They demonstrated that net patella shift, tilt and rotation in both contraction patterns was lateral. This, therefore, supports the theory that the dislocation may happen during an extension movement as proposed by Nikku.¹²

Seven of these eight studies were cadaveric, but only 2 gave information about the age and the condition of the specimens used.^7,8 Four of the cadaveric studies do not report the age of the specimens that they use^6,10,11,13 and the other three have an average age of 50 or above .^7,8,14 It is likely that the studies that do not report the age of their specimens are also likely to be old. However, given that dislocations are generally an injury of the younger population,¹⁵ the generalisability of these results may be brought into question. Furthermore, the presence of any history of previous PF problems or symptoms would be hard to exclude. The studies did attempt to exclude specimens with previous surgery or evidence of degenerative changes. This was done either by visual inspection or the use of x-rays. Finally, the tracking systems used in these studies are well established but they are associated with an element of human error regarding placement of the markers which is not addressed by any of the studies.

One element in these cadaveric studies of which the influence is not clear is with respect to the tensions applied to the quadriceps. All of these studies applied differing amounts of tension. However, true patella tracking is under the influence of the quadriceps throughout its movement and as a result these studies, it could be argued, failed to truly recreate the in vivo biomechanics of the PFJ. Van Kampen⁶ did, however, demonstrate that increasing quadriceps tension did not influence patellar tracking in the experimental setting. Nevertheless, due to this factor, interpretation of the results and application to real life must be done with a degree of caution. The role of the quadriceps is discussed later in this review. A further element that may have influenced the tracking of the patella in these cadaveric studies is dissection and removal of the skin. The patella is a very subcutaneous bone and we are unsure the effect it has upon tracking as it tightens through flexion.

One strength of the cadaveric studies, however, was that they were dynamic studies that used a tracking systems to monitor the movement of the patella. Yamada,⁹ on the other hand, used a simulation model based upon static images with the quadriceps relaxed. Using such simulation models may result in simplification of the knee joint and make assumptions that are not true. This may have been the reason that they were not able to comment upon patella shift in this study. Nevertheless, they used a younger cohort of patients, with the study and control groups having an average age of 21 and 24 respectively. This does improve its generalisability.

Despite the limitations of these studies, the likely conclusion is that lateral tilt and shift of the patella appears to begin at around 20° of flexion. And given that at this point movement is occurring in this plane, it would suggest that the force vector on the patella is becoming lateral at this point. This may well represent the point at which it would become easier for the patella to dislocate in this direction.

4.2. Soft tissue restraints

Soft tissue restraints play an important role in preventing dislocations of the patella. Numerous experiments have shown that medial patellofemoral ligament (MPFL) has a major role to play in preventing a lateral patellar dislocation. Studies have shown that it provides 50–60% of restraint at varying degrees of flexion.^16,17,18 There has also been work done on a number of the other medial structures and their role is now better understood than it has been previously. One study demonstrated that rupturing of the medial retinacular structures resulted in reduced stability throughout knee flexion.¹⁹ Philipott¹⁴ on the other hand showed that the medial patellomeniscal ligament and medial patellotibial ligament also had a role to play in stability.

For a patella to dislocate laterally it needs to overcome these structures as well as the MPFL. In total 8 studies were identified that performed biomechanical testing of this aspect of the knee with 5 of them determining its influence on stability. The finding of these studies can be seen in Table 3.

Table 3.

Findings of studies looking at the function of the MPFL.

Study	Knee Positions	Quadriceps Tension	MPFL Findings
Conlan et al.19	Full extension	No	MPFL provides 53% of medial restraint Restraining force with all structures intact was 225 N/cm
Hautmaa et al.25	Tests performed with knee at 30° (±5°)	Yes - 2-pound weight applied as resting physiological load	MPFL responsible for 50% of medial restraint with a displacement of 13.8 mm with 5lb of force applied
Farahmand et al.22	Patellar displacement (10 mm) was measured at 0, 10, 20, 30, 45, 60 and 90°	Yes – varying forces used	Decrease in force required to displace patella between 0° and 20° which then remains reasonably constant until 60°
Zaffagnini et al.12	Testing performed at 0, 30, 60 and 90° of flexion	Yes - 60 N isolated force through the quadriceps tendon	With a resected MPFL, there is increased laxity at 30–60° of knee flexion
Desio et al.20	Testing performed at 20° of flexion	Yes – 2lb weight attached to centralise patella	MPFL represents 60% of restraint Average force of 182 N required to achieve maximal patella translation
Senvongse et al.21	Testing performed from 0 to 90°	Yes – 175 N	75 N of force required to displace patella at 20° of knee flexion compared with 126 N and 125 N at 0° and 90° respectively
Mountney et al.24	Knee placed in external rotation	No	Mean tensile strength of 208 N required for MPFL to rupture
Burks et al.23	Secured in 20° of flexion	Yes – 2lb weight attached to centralise patella	Mean tensile strength of 209 N required for MPFL to rupture

These studies confirm that the MPFL certainly has an important role to play in preventing a lateral patella dislocation. Studies by Farahmand et al.²⁰ and Zaffagnini et al.⁷ showed that the point at which the patella enters or leaves the trochlea (20–30° of flexion) until mid-flexion (60°) is the point of least resistance.

Burks²¹ and Mountey²² attempted to determine the force required to overcome the MPFL. They found that forces of 209 N and 208 N were required for it to rupture, respectively. However, the patella was not dislocated. Desio¹⁷ reported that 182 N was required to fully translate the patella at 20° of flexion, but it is unclear if this resulted in a rupture of the MPFL or if the patella was fully dislocated.

Many of the limitations discussed with respect to the cadaveric studies in patellar tracking are also applicable here. Again, the investigators did not routinely report the demographics of the specimens used and their conditions. Only Zaffagnini⁷ and Mountney²² reported both of these elements. The average age of the specimens, when reported, was greater than 50. As well as affecting the generalisability of their results, this may have adversely affected the quality of the tissue and, hence, the results. The tissues may also have been altered by the embalming process and it is difficult to account for this in our interpretation of the results.

The role of the quadriceps also poses another problem when it comes to interpretation. As has been previously been mentioned, it does play a role in patella tracking and, therefore, will affect the tension of the soft tissue structures depending upon the direction of the force. As can be seen in Table 3, a variety of tensions were used. If the aim of the study is to assess the influence of the MPFL as part of the knee structure then it is important to adequately tension the quadriceps. If the aim is to look at the MPFL in isolation, then the amount of tension plays less of role and is more important in ensuring that the patella is centered. Based on the aims of these papers, it could be argued that they have tensioned the quadriceps correctly with respect to their aims.

These studies do add further credibility to the fact that a lateral patella dislocation is likely to occur during the early-mid stages of flexion. The MPFL clearly is an important restraint in preventing a lateral dislocation and would need to be overcome. It would appear from these studies that this structure is at its weakest point between 20° and 60°. This does not, however, mean that dislocation laterally cannot occur at other points. The force required may be greater and other elements, such as the trochlea, play a more prominent role at that stage of the joint's movement. Overall, it is difficult to draw conclusions regarding the amount of force required for the patella to dislocate based on these studies as there is a great deal of variation in the methodologies used and the way in which it has been reported. Nevertheless, they will need to be overcome in order for a dislocation to occur.

4.3. The quadriceps muscle

The role of the quadriceps has already been briefly discussed in this review with respect to its influence on patella tracking and stability. Unlike the other soft tissue structures, this forms part of the dynamic element which influences the patella. The role of the vastus medialis oblique (VMO), in particular, has been the source of investigation. Four papers were identified commenting specifically on the role of the VMO. A further four papers looked at the morphology and influence of other elements of the quadriceps. The results of these papers can be seen in Table 4.

Table 4.

The role of quadriceps muscles in the patellofemoral joint.

STUDY	QUADRICEPS/VMO FINDINGS
Philipott et al.15	VMO has no significant effect on patella stability
Senavongse & Amis21	Relaxing the VMO, reduced the lateral restraining force throughout flexion. At 30°, it reduced it by 30%
Goh et al.26	No tension on the VMO meant that it’s translated more laterally throughout all ranges of flexion
Raimondo et al.27	VMO plays an important role in functioning as a dynamic restraint
Farahmand et al.29	Greater variation in physiological cross sectional area of the vastus lateralis than of the vastus medialis
Kan et al.30	Patients with known PF instability were found to have more of an equal VLO:VMO ratio whereas the control group had more of a VMO inclination
Farahmand et al.28	Subluxations that occur at 0–30 of flexion are likely to be due to abnormal medial retinacular structures. Muscle imbalances had an influence throughout flexion
Farahmand et al.22	Increasing quadriceps tension did not result in an increase in the restraining force to translate the patella laterally between 0 and 60° flexion

Other than Philipott's study,¹⁴ these studies demonstrate that different parts of the quadriceps have a key role to play in patellar stability. However, Philipott¹⁴ acknowledge that their study was poorly designed in this regard. It is clear from several studies that the VMO does play an important role in the stability of the patella.^19,23,24 Its lack of function results in greater lateral movement. However, the patella was not actually dislocated in the cadaveric studies by Senavongse¹⁹ or Goh.²³ Sevangonse¹⁹ displaced the patella laterally by 10 mm whereas Goh²³ tracked the movement of the patella as the knee was extended from 90° to 20° of flexion. Therefore, the results of these studies cannot be taken to represent that of a dislocation. Nevertheless, it does give some insight into the role of the VMO and its potential effects on tracking and subluxation symptoms.

Raimondo's²⁴ findings, however, are based on the morphological characteristics of the VMO and not on dynamic studies. They noted that the VMO represented 30% of the physiological cross-sectional area (PCSA) of the vastus medialis complex. This combined with the pennation angle of the VMO led the authors of this study to conclude that it has an important role to play in patellar stability due to the direction of its pull when contracted. However, as a purely anatomical study they did not offer a great deal of information regarding the conditions of the specimens nor their dissection methodology. The QUACS appraisal tool also suggests that the findings should be observed by more than one researcher. This may well have been the case, but is not alluded to in their report.

The work looking at the role of the quadriceps is probably more applicable to real life as this focuses more on their morphology and contraction patterns. Two such studies demonstrated that increasing tension in the quadriceps did not increase restraining force between 0 and 60°.^20,25 They do, however, show that some tension is required. However, again these studies did not dislocate the patella and the measurements were based on a lateral displacement of 5 mm.

One study appears to suggest that there is a muscle imbalance in those with instability.²⁶ Their work indicated that the imbalance is related to the size of the vastus lateralis (VL) rather than the VMO as had been expected. Whilst Farahmand²⁶ gave a full description of the method of dissection and the basic characteristics of their specimens, they did lack data for 5 out of 12 of their specimens. Their findings are backed up, to a degree, by the work an imaging based study using diffusion-tensor magnetic resonance imaging (DT-MRI) which showed that the ratio of vastus lateralis oblique (VLO):VMO in patients with instability is more equal than in the control group.²⁷ This series, however,had a small sample size and may, therefore, limit the generalisability of their findings.

Nevertheless, it is clear from these studies that the quadriceps has an important role to play in the stability of the PFJ. However, not a great deal of information can be gathered regarding their role in an acute dislocation other than to suggest that in an individual with a more lateral favouring quadriceps muscle complex, the patella is likely to track more laterally. This may be extrapolated further to suggest that a strong contraction in such a scenario may play a role in a lateral dislocation.

4.4. Tibial rotation

External tibial rotation occurs as a normal part of knee biomechanics during the last 20° of extension. One element of the mechanisms proposed is that there is external rotation of the tibia at the time of injury. Five studies looked at this element specifically.

All of these studies were cadaveric in nature and, other than Heegard et al.,⁹ concluded that external rotation of the tibia increases the lateral force upon the patella. Heegard⁹ reported that there was a decrease in the amplitude of the medial force at the initial part of flexion. It would also seem that some studies have reported that the only point at which there is a statistically significant difference in the patellar tracking is from 0 to 30° of flexion. A summary of the findings of these studies can be seen in Table 5.

Table 5.

Findings of the effects of tibial rotation.

STUDY	RESULT OF TIBIAL ROTATION
Van Kampen & Huiskes16	External rotation of the tibia results in a lateralising force
Sakai et al.32	External rotation of the tibia when going from sitting to standing, squatting to standing and in the stance phase of walking results in increased lateral patellar shift
Nagamine et al.33	External rotation resulted in lateral shift and tilt of the patella during the early phases of flexion
Hefzy et al.17	External rotation of the tibia causes increased lateral shift which was significant during the first 30° of flexion
Heegard et al.10	External rotation resulted in decreased medial shift & tilt of the patella. Internal rotation of the tibia had no effect on patellar shift and tilt from 0 to 30° of flexion and reported these parameters as independent

Sakai et al.'s study is of particular interest as they conducted a series of experiments replicating real-life movements.²⁸ In a very well designed study, they showed that tibial rotation had the same effect in all scenarios. This adds a degree of credibility to the fact that the dislocation may well occur during extension of the knee as well as during flexion.

The remaining 4 studies all followed a similar methodology and as a result had similar limitations. The limitations of cadaveric studies has been discussed previously, but whilst these studies were well designed, they suffered from their sample sizes. The study with the largest number of samples was Nagamine²⁹ with 11 specimens. The remaining three studies had a total of 10 specimens combined. However, given the results of these studies are so similar it is reasonable to conclude that at the time of injury, the tibia is likely to be externally rotated.

4.5. Femoral rotation

Internal rotation of the femur is considered part of the mechanism by which a lateral patellar dislocation occurs. No biomechanical analysis of this aspect has been conducted. Biomechanically, in real life, the femur is believed to be in a fixed position.⁶ Therefore, the other studies discussed in this systematic review have a rigid femur and as a result do not investigate their contribution to patellar instability. Two papers by Balcarek et al.³⁰ and Diedrichs et al.³¹ do look at this aspect. Both of these studies are imaging based studies and focused on patients with chronic instability.

The findings of both studies need to be interpreted with caution as they are investigating patients with chronic instability who have suffered a dislocation. Furthermore, given that these tests are not dynamic, their applicability to a mechanism of injury model is hard to justify. The results of this do not exclude the possibility that there is some internal rotation at the time of a dislocation but on the basis of these studies, it can only really be considered as an anatomical risk factor.

4.6. Mechanism of inury

The biomechanical and kinematic studies included in this systematic review reveal that a patella dislocation is likely to occur with the tibia externally rotated, the quadriceps contracted and the knee flexed to approximately 20–30°. The knee may be flexing or extending at the time of dislocation. The force required for the dislocation to occur would also have to be sufficient to overcome the soft tissue restraints of the patella, in particular the MPFL. The role of femoral internal rotation is less clear. These findings are in keeping with elements of previously proposed theories:

• •Hughston³² hypothesizes that at the time of injury, the foot is planted and the individual is performing a cutting action in the opposite direction. This results in internal rotation of the femur, external rotation of the tibia and the knee in a valgus position. The quadriceps contracts resulting in a lateral force acting upon the patella. This was thought to occur in the early part of knee flexion.
• •Malteuis et al.³³ reported a dislocation at 70° flexion occurring in an individual with trochlea dysplasia during eccentric quadriceps contraction testing. They felt the tibia was externally rotated with the knee in a valgus position
• •
  Nikku et al.⁶ conducted a series of patient interviews and identified 2 further potential mechanisms:

  • o further extension of the knee from a nearly straight start
  • o further flexion of the knee from an already flexed position
• •Cash and Hughston³⁴ also reported a direct valgus blow to the knee resulting in a patella dislocation

These theories, however, are based predominantly on retrospective interviews and case reports. Hence, this review is the first attempt in the published literature at amalgamating the available biomechanical and kinematic evidence to develop such an injury model.

4.7. Limitations

Despite there being well-designed studies included, many of the limitations of this review revolve around the usage of cadaveric methods. This reflects the fact that it is difficult ethically to conduct trials on injury mechanisms and as a result there are no level I or II trials available. Cadaveric studies do allow us to perform direct injury studies and passively assess the biomechanics of a joint. However, often it is difficult to obtain specimens of the correct age to represent the demographic most affected by that injury and it is not possible to truly replicate the involvement of the muscles surrounding the joint. Furthermore, it is difficult to recreate the types of loads that a joint may undergo during an injury as well the influence of any external factors. These factors can be negated to a certain degree using computer simulation models, but these may be simplified and not account for the complex nature of a joint. These models are often not validated and this, therefore, may call into question the final results and conclusion. This review also only focused on studies that have been published in the English language. This may have meant studies were not identified and included that may have provided further insight.

5. Conclusion

The patella in an anatomically normal knee is not expected to dislocate. In order for it to do so requires a complex mechanism to overcome the numerous restraints that have evolved for it to maintain its position and tracking.

This review represents the first attempt to analyse the numerous papers that have been published and attempt to determine a possible mechanism by which a patella dislocation may occur. The biomechanical and kinematic studies in this review have assessed differing elements in isolation as well as in combination with other factors. This has allowed us to explore the various elements that are believed to be involved in a dislocation and elements of a possible mechanism to be devised. At the time of injury, the tibia is likely to be in an externally rotated position, the quadriceps contracted and the knee flexed. It is likely to have 20–30° of flexion, but a dislocation can happen in deeper flexion with a greater force or if there is an anatomical anomaly. The knee may be flexing or extending at the time of dislocation.

Higher levels of evidence are required for determining the mechanism by which a patella dislocation occurs. With the advent of dynamic imaging, we may gain greater in vivo insight into the patellofemoral joint. Until a time that we are able to conduct such studies, an important element that we have highlighted as a result of our review is that there a number of well-designed studies that have been conducted. The combined results of these as demonstrated in this paper may help guide future research in this domain. However, further work is required to repeat these experiments with larger sample sizes to increase the generalisability of the results. Work should also be undertaken to determine the influence of internal rotation of the femur and the effect of a valgus force to the knee. This should further supplement ongoing work in further understanding the anatomical risk factors in individuals who suffer with recurrent dislocations.

Contributorship

V Dewan: Conception, Design, Collection Of Data, Data analysis & interpretation, Writing of manuscript for publication, final approval of manuscript, MSL Webb: Data analysis and interpretation, writing of manuscript, final approval of manuscript, D Prakash: Data analysis and interpretation, writing of manuscript, final approval of manuscript, A Malik: Data analysis and interpretation, writing of manuscript, final approval of manuscript, S Gella: Data analysis and interpretation, writing of manuscript, final approval of manuscript, C Kipps: Data analysis and interpretation, writing of manuscript, final approval of manuscript.

Funding info

Nil.

Ethical approval

N/A.

Data sharing statement

All relevant data are included in the article.

Declaration of competing interest

None of the authors have any conflicts of interest to disclose.