Patellar Kinematics, Part II:The Influence of the Depth of theTrochlear Groove in Subjects Withand Without Patellofemoral Pain

Background and Purpose. A shallow intercondylar groove has beenimplicated as being contributory to abnormal patellar alignment. Thepurpose of this study was to assess the influence of the depth of theintercondylar groove on patellar kinematics. Subjects. Twenty-threewomen (mean age526.8 years, SD58.5, range514–46) with a diagno-sis of patellofemoral pain and 12 women (mean age529.1 years,SD55.0, range524–38) without patellofemoral pain participated.Only female subjects were studied because of potential biomechanicaldifferences between sexes. Methods. Patellar kinematics were assessedduring resisted knee extension using kinematic magnetic resonanceimaging. Measurements of medial and lateral patellar displacementand tilt were correlated with the depth of the trochlear groove (sulcusangle) at 45, 36, 27, 18, 9, and 0 degrees of knee flexion usingregression analysis. Results. The depth of the trochlear groove wasfound to be correlated with patellar kinematics, with increased shal-lowness being predictive of lateral patellar tilt at 27, 18, 9, and 0degrees of flexion and of lateral patellar displacement at 9 and 0degrees of flexion (r 5.51–.76). Conclusions and Discussion. Theresults of this study indicate that bony structure is an importantdeterminant of patellar kinematics at end-range knee extension(0°–30°). [Powers CM. Patellar kinematics, part II: the influence of thedepth of the trochlear groove in subjects with and without patello-femoral pain. Phys Ther. 2000;80:965–973.]

Key Words: Magnetic resonance imaging, Patellar kinematics, Patellofemoral joint.

Physical Therapy . Volume 80 . Number 10 . October 2000 965

Rese

arch

Repo

rt

Christopher M Powers

v

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

IIIIII

IIIIII

IIIIII

IIIIII

IIIIII

IIIIII

IIIIII

IIIIII

IIIIII

IIIIII

IIIIII

IIIIII

IIIIII

III

Patellar malalignment is thought to be amongthe etiological factors contributing to patello-femoral pain (PFP).1 The cause of PFP appearsto be multifaceted, with components being

defined by 2 distinct categories: structural and dynamic.Structural considerations include abnormal bony config-uration1–6 or tightness of noncontractile elements.7–9

Dynamic components have been hypothesized as involv-ing unequal activity of the different heads of the quad-riceps femoris muscle10,11; however, evidence to supportthis premise has not been consistent.12,13

Brattstrom2 reported that dysplasia of the femoral troch-lea is the most important etiological factor in recurrentpatellar subluxation. Because the lateral femoral condyleis larger and projects farther anteriorly than the medialcondyle, the trochlear groove is thought to provide bonystability resisting laterally directed forces.7 Althoughsome authors2,14 have reported that the decreased depthof the intercondylar sulcus is a primary cause of lateral-ization of the patella, other authors15–18 have hypothe-sized that abnormal patellar kinematics are the result ofthe patella resting above the trochlear groove. Recentwork by Farahmand and colleagues,19,20 however, sug-gests that stability of the patella is more a function of theincreased tension of the patellar tendon and quadricepstendon as the knee flexes, and not necessarily a functionof the depth of the trochlear groove.

Although bony abnormalities have been implicated asbeing contributory to abnormal patellar alignment, therelationship of these factors to patellar tracking patternshas not been established. With the advent of kinematicmagnetic resonance imaging (KMRI) and cine phasecontrast imaging techniques,21 quantification of patellarmovement throughout an arc of resisted knee extensionis possible.22–24 These diagnostic techniques have a dis-tinct advantage over imaging procedures used withoutallowing for knee movement because contributions ofthe extensor mechanism to patellofemoral joint kine-matics can be assessed.25

The purposes of this investigation were to comparepatellar tracking patterns between subjects with PFP andsubjects without PFP and to assess the influence of the

depth of the intercondylar groove on patellar kinemat-ics. I hypothesized that subjects with PFP would exhibitgreater amounts of lateral patellar displacement andlateral patellar tilt compared with subjects without PFPand that the magnitude of lateral patellar displacementand lateral patellar tilt would be associated with thedepth of the trochlear groove. For results and discussionconcerning the influence of vastus muscle activity inpatellar kinematics, the reader is referred to the articleby Powers titled “Patellar Kinematics, Part I: The Influ-ence of Vastus Muscle Activity in Subjects With andWithout Patellofemoral Pain” in this issue.

Method

SubjectsTwenty-three women with a diagnosis of PFP and 12women without PFP participated in this study. Onlyfemale subjects were studied because of potential biome-chanical differences between sexes. Both groups weresimilar in age, height, and weight (Tab. 1). Age, height,and weight were found to be normally distributed withineach group and when data from both groups werecombined. No attempt was made to match each subjectspecifically for age, height, and weight, as there is noevidence in the literature to suggest that individuals ofdifferent ages, heights, and weights will demonstratedifferences in patellar kinematics.

The subjects with PFP were patients of the SouthernCalifornia Orthopaedic Institute who were deemed to beappropriate candidates by the treating physician. Priorto participation, all subjects with PFP were screened torule out ligamentous instability, internal derangement,and patellar tendinitis. Each subject’s pain originatedfrom the patellofemoral joint, and only patients withhistories relating to nontraumatic events were accepted.In addition, pain had to be readily reproducible with atleast 2 of the following activities: stair ascent or descent,squatting, kneeling, prolonged sitting, or isometricquadriceps femoris muscle contraction.1,19 Subjects wereexcluded from the study if they reported previous kneesurgery or a history compatible with acute traumaticpatellar dislocation.

CM Powers, PT, PhD, is Director, Musculoskeletal Biomechanics Research Laboratory, and Assistant Professor, Department of Biokinesiology andPhysical Therapy, University of Southern California, 1540 E Alcazar St, CHP-155, Los Angeles, CA 90033 (USA) (powers@hsc.usc.edu).

Dr Powers provided concept/research design, writing, data collection and analysis, subjects, project management, and fund procurement.

This study was approved for human subjects by the Los Amigos Research and Education Institute Inc of Rancho Los Amigos Medical Center(Downey, Calif).

This study was partially funded by a grant from the Foundation for Physical Therapy.

This article was submitted December 28, 1999, and was accepted May 29, 2000.

966 . Powers Physical Therapy . Volume 80 . Number 10 . October 2000

Individuals comprising the comparison group wererecruited by word of mouth and were either employeesof Rancho Los Amigos Medical Center (Downey, Calif)or students from the University of Southern California.Subjects had to have no history or diagnosis of kneepathology or trauma and they had to be free of kneepain at the time of the study. In addition, these subjectsdid not report pain with any of the activities listedearlier. The kinematic data from the comparison groupwere previously described in an article discussing the useof magnetic resonance imaging (MRI) for assessingpatellar tracking.23

InstrumentationKinematic magnetic resonance imaging of the patello-femoral joint was assessed with the transmit and receivequadrature body coil of a 1.5T magnetic resonancesystem* using a pulse sequence that allowed fast imagingtimes with the best possible temporal resolution (fast-spoiled gradient recall acquisition in the steady state).Axial-plane imaging was performed using the followingparameters: time to repeat56.5 milliseconds, time toecho52.1 milliseconds, number of excitations51.0,matrix size5256 3 128, field of view538 cm, flipangle530 degrees, and a 7-mm section thickness with aninterslice spacing of 0.5 mm.23 Acquisition time was 6seconds to obtain 6 images (ie, 1 image per second).

All imaging was performed using a specially constructed,nonferromagnetic positioning device† that permittedbilateral knee extension against resistance (in the proneposition) from 45 degrees of flexion to full extension(see Fig. 1 in the companion article by Powers in thisissue). The device was designed to allow uninhibitedmovement of the patellofemoral joint and normal rota-tion of the lower extremities. I believe that these designfeatures are important because patellar tracking may beinfluenced by tibial rotation.26

Resistance was accomplished through a pulley systemwith a constant 30.5-cm lever arm. The design of the

device was such that the application of the force wasalways perpendicular to the tibia to ensure a constant(isotonic) torque throughout the entire range ofmotion.23 Weights constructed of nonmagnetic, 316Lseries stainless steel‡ supplied the resistive force for thismaneuver. These plates were placed on a movablecarriage that was attached to the pulley apparatus (seeFig. 1 in the companion article by Powers in this issue).

ProcedurePrior to testing, all procedures were explained to eachsubject and written informed consent was obtained. Allimaging was performed at Tower Imaging Center in westLos Angeles, Calif. Subjects were placed prone on thepositioning device in a position designed to allow fornatural lower-extremity rotation. After this position wasachieved, Velcro straps§ were used to secure the subjects’thigh and tibia to the positioning device. Resistance onthe device was then set at 15% of body weight.

After familiarization with the knee extension apparatus,subjects were instructed to practice extending theirknees at a rate of approximately 9°/s. This rate ensured6 evenly spaced images throughout the 45-degree arc ofmotion (including the 45° position) and permittedimaging at 45, 36, 27, 18, 9, and 0 degrees of kneeflexion. Approximation of this rate was made by theprincipal investigator (CMP) with the use of a stopwatch.

Once the subject, in the opinion of the principal inves-tigator, was able to reproduce the desired rate of motionin a smooth and even manner, imaging commenced.Subjects were instructed to initiate extension upon ver-bal command and continue until full extension hadbeen reached. Imaging was done at 3 different imageplanes to assess the entire excursion of the patella inrelation to the trochlear groove (ie, 3 slices wereobtained for each angle of knee flexion). These proce-dures were repeated if I thought the rate of kneeextension was too fast or too slow, or not performed in asmooth manner. In addition, the procedure was

* General Electric Medical Systems, 3200 N Grandview Ave, Waukesha, WI 54601.† Captain Plastic, PO Box 27493, Seattle, WA 98125.

‡ Esco Corp, 6415 E Corvette St, Los Angeles, CA 90242.§ Velcro USA Inc, PO Box 5218, 406 Brown Ave, Manchester, NH 03108.

Table 1.Subject Characteristics

Subjects With Patellofemoral Pain(n523)

Subjects Without Patellofemoral Pain(n512)

PaX SD Range X SD Range

Age (y) 26.8 8.5 14–46 29.1 5.0 24–38 .38Height (cm) 165.6 7.2 151.3–177.1 168.4 8.0 153.6–183.5 .29Weight (kg) 62.2 9.1 42.0–82.7 61.2 8.0 48.7–74.1 .76

a Probability values based on independent t tests.

Physical Therapy . Volume 80 . Number 10 . October 2000 Powers . 967

IIIIII

IIIIII

IIIIII

IIIIII

IIII

repeated if 6 adequate images were not obtained. Anadequate image was one in which the medial and lateralborders of the midsection of the patella, the trochleargroove, and the posterior femoral condyles were welldefined. Visualization of these landmarks was necessaryfor subsequent analysis.

Data ManagementPrior to analysis, all images were screened by the princi-pal investigator to ascertain the midsection of the patella(maximum patellar width) at each angle of knee flexion.Once the midsection of the patella was determined,measurements for these images were obtained. Onlyimages containing a midpatella slice were analyzed.

To examine patellofemoral joint relationships at thevarious degrees of knee flexion, measures that wereindependent of the shape of the patella and the anteriorfemoral condyles were used.23 This was done in an effortto avoid measurement variability resulting from thecontinually changing contour of these structures whenviewed at different angles of knee flexion and to allowassessment of patellar orientation when the intercondy-lar groove was not well visualized. All measurementswere made with a computer-assisted program andincluded assessment of medial and lateral patellar dis-placement, medial and lateral patellar tilt, and the sulcusangle.

Medial and lateral patellar displacement were deter-mined by the “bisect offset” measurement as describedby Stanford et al27 and modified by Brossmann et al.22

The bisect offset was measured by drawing a line con-necting the posterior femoral condyles and then project-ing a perpendicular line anteriorly through the deepestpoint (apex) of the trochlear groove. This line inter-sected with the patellar width line, which connected thewidest points of the patella (see Fig. 2 in the companionarticle by Powers in this issue).23 The perpendicular linewas projected anteriorly from the bisection of the poste-rior condylar line to obtain data when the trochleargroove was flattened (see Fig. 2 in the companion articleby Powers in this issue). All bisect offset data representedthe extent of the patella lying lateral to the projectedperpendicular line and were expressed as a percentageof total patellar width.

Medial and lateral patellar tilt were measured using amodification of the technique described by Sasaki andYagi.28 The patellar tilt angle was the angle formed bythe lines joining the maximum width of the patella andthe line joining the posterior femoral condyles (seeFig. 3 in the companion article by Powers in this issue).All tilt measurements were reported in degrees.

The sulcus angle was described by Brattstrom2 as theangle formed by the highest points of the medial andlateral femoral condyles and the lowest point of theintercondylar sulcus (Fig. 1).23 To obtain data when thetrochlear groove lacked discernable depth, the center ofthe sulcus angle was defined by a perpendicular line thatwas drawn anteriorly from the bisection of the posteriorcondylar line (Fig. 1). The estimation of the center ofthe sulcus angle was based on the evaluation of normalimages that showed that the deepest portion of theintercondylar groove typically overlies the midpoint ofthe posterior condyle interval. All sulcus angles werereported in degrees.

The day-to-day reliability for obtaining the KMRI datausing the procedures and measurements described wasdetermined in a previous study to have intraclass corre-lation coefficients ranging from .66 to .82).23 Based onrepeated testing, intraobserver measurement error(standard error of measurement) was determined to be3.4% for the bisect offset measurement, 2.9 degrees forpatellar tilt, and 2.0 degrees for the sulcus angle.Although anatomical landmarks were identified manu-ally, all lines used for angle and displacement measure-ments were drawn by the computer software. Quantifi-cation of all angles and distances was performed by thissame program. This procedure assisted in minimizingmeasurement error.

Figure 1.Method used to measure the sulcus angle. This angle was defined bylines joining the highest points of the medial and lateral condyles andthe lowest point of the intercondylar sulcus (AB and CB) (left). In order toobtain data when the trochlear groove lacked discernible depth, thecenter of the sulcus angle was defined by a perpendicular line that wasprojected anteriorly from the bisection of the posterior condylar line(right). All sulcus angle measurements were reported in degrees.Reprinted by permission of Lippincott Williams & Wilkins from PowersCM, Shellock FG, Beering TV, et al. Effect of bracing on patellarkinematics in patients with patellofemoral joint pain. Med Sci SportsExerc. 1999;31:1714–1720.

968 . Powers Physical Therapy . Volume 80 . Number 10 . October 2000

Data AnalysisAll statistical procedures were performed with BMDPstatistical software.\ Prior to analysis, descriptive statisticswere calculated for all variables, and normality of distri-bution was assessed using the Wilk-Shapiro test. Basedon the analysis of distribution, all data were analyzedusing parametric tests. Significance levels were set atP,.05.

To determine whether patellar indexes varied betweengroups or angles of knee flexion, a 2 3 6 (group 3angle) analysis of variance for repeated measures on onevariable (angle) was performed. This analysis was per-formed for each kinematic variable. A regression analysiswas performed to determine whether the sulcus angle(independent variable) was predictive of patellar tilt orpatellar displacement (dependent variables). This anal-ysis was repeated for both dependent variables at eachangle of knee flexion. To control for differencesbetween the 2 groups of subjects, the grouping variablewas included in all regression equations.

Results

Patellar KinematicsA difference was found in patellar tilt between the 2groups. Compared with the comparison group, thesubjects with PFP demonstrated a greater degree oflateral patellar tilt when the data were averaged across allangles of knee flexion (10.7° versus 5.5°, P,.02) (Fig. 2).The largest difference between the 2 groups was 7degrees (11.7° in the subjects with PFP versus 4.7° in the

subjects without PFP), which occurred at 27 degrees ofknee flexion.

In contrast, there was no difference in bisect offsetbetween the 2 groups (no group effect or interaction)(Fig. 3). When the data were averaged across all kneeflexion angles, the average bisect offset measurement forthe subjects with PFP was 57.9% of the patella lateral tomidline, as compared with 53.8% of the patella lateral tomidline in the subjects without PFP.

Similarly, there was no difference in the sulcus anglebetween the subjects with PFP and the subjects without PFP(no group effect or interaction) (Fig. 4). When averagedacross all angles of knee flexion, the mean sulcus angle was149.4 degrees for the subjects with PFP, as compared with144.6 degrees for the subjects without PFP.

Relationship Between Sulcus Angle and PatellarKinematicsThe Pearson correlation coefficients obtained whenassessing the relationship between the sulcus angle andpatellar displacement at the various knee flexion anglesranged from .15 to .74 (Tab. 2). Similarly, the correla-tion coefficients obtained when assessing the relation-ship between the sulcus angle and patellar tilt at thevarious knee flexion angles ranged from .26 to .76(Tab. 2).

The sulcus angle was a predictor of patellar displace-ment at 9 degrees of knee flexion (r 5.46, R25.21);however, it was a stronger predictor of patellar displace-ment at 0 degrees (r 5.74, R25.55; Fig. 5). In general, asthe sulcus angle increased (ie, became more shallow),the amount of lateral patellar displacement alsoincreased.\ SPSS Inc, 444 N Michigan Ave, Chicago, IL 60611.

Figure 2.Comparison of patellar tilt between the subjects with patellofemoral pain(PFP) and the subjects without PFP from 45 to 0 degrees of knee flexion.Positive values indicate lateral tilt. Lateral patellar tilt was greater for thesubjects with PFP than for the subjects without PFP (P,.05). Error barsindicate one standard deviation. Data for subjects without PFP previ-ously reported by Powers et al.23

Figure 3.Comparison of patellar displacement (bisect offset) between the subjectswith patellofemoral pain (PFP) and the subjects without PFP from 45 to 0degrees of knee flexion. Error bars indicate one standard deviation.Data for subjects with PFP previously reported by Powers et al.23

Physical Therapy . Volume 80 . Number 10 . October 2000 Powers . 969

IIIIII

IIIIII

IIIIII

IIIIII

IIII

The sulcus angle also was a predictor of patellar tilt at27 degrees (r 5.51, R25.26), 18 degrees (r 5.54,R25.29), 9 degrees (r 5.63, R25.40), and 0 degrees ofknee flexion (r 5.76, R25.58; Fig. 6). As with patellardisplacement, an increase in the sulcus angle resulted ingreater amounts of lateral patellar tilt.

DiscussionThe sulcus angle, as measured in this study, was repre-sentative of the depth of the femoral trochlea at themidsection of the patella. In general, there was a trendtoward a more shallow groove in the subjects with PFPwhen the data were averaged across all knee flexionangles. It is evident from these data, however, thatalthough the 2 groups had similar sulcus angles at 45, 36,and 27 degrees of flexion, a substantial increase (loss ofdepth) was observed in the subjects with PFP as the kneeextended beyond 27 degrees. This increase in the sulcusangle is similar to the increases reported by Schutzeret al29 and Kujala et al30 and suggests that bony stabilityat the end-range of extension may be compromised inpeople with PFP.

The sulcus angle was found to be a predictor of lateralpatellar tilt at 27, 18, 9, and, 0 degrees, as well as apredictor of lateral patellar displacement at 9 and 0degrees. This finding underscores the importance of thebony anatomy in contributing to patellar stability andcould theoretically explain the clinical manifestation oflateral patellar subluxation during terminal knee exten-sion. The association between bony anatomy and patel-lar stability was evident in the PFP data, where it wasobserved that the point at which the sulcus angle beganto deviate from the data obtained for the comparisongroup (approximately 27°) was at the same point atwhich the lateral displacement became more pro-nounced (Figs. 3 and 4). The finding that more than halfof the variability in patellar tilt and displacement could beexplained by the sulcus angle at 0 degrees supports theargument of Brattstrom2 that a shallow femoral sulcus is apredisposing factor with regard to abnormal patellar kine-matics at terminal knee extension.

During knee extension, the sulcus angle of the subjectswithout PFP increased an average of 10 degrees, indicat-

Figure 4.Comparison of sulcus angle between the subjects with patellofemoralpain (PFP) and the subjects without PFP from 45 to 0 degrees of kneeflexion. Error bars indicate one standard deviation. Data for subjectswith PFP previously reported by Powers et al.23

Figure 5.Relationship between the sulcus angle (in degrees) and bisect offset(percentage of the patella width lateral to midline) for the subjects withpatellofemoral pain (PFP) and the subjects without PFP at 0 degrees ofknee flexion (r 5.74; F519.3; df52,33; P,.05).

Figure 6.Relationship between the sulcus angle (in degrees) and patellar tilt (indegrees) for the subjects with patellofemoral pain (PFP) and the subjectswithout PFP at 0 degrees of knee flexion (r 5.76; F520.6; df52,33;P,.05). Positive values of patellar tilt indicate lateral tilting.

Table 2.Pearson Correlation Coefficients for Sulcus Angle and KinematicVariables

DependentVariable

Knee Flexion Angle (°)

45 36 27 18 9 0

Patellar displacement .15 .23 .16 .35 .46a .74a

Patellar tilt .26 .34 .51a .54a .63a .76a

a Significant at P,.05.

970 . Powers Physical Therapy . Volume 80 . Number 10 . October 2000

ing that the patella was moving to a more shallow portionof the femoral trochlea. Because the patella migratessuperiorly as the knee extends,31,32 this observation, in myopinion, suggests that the bony stability afforded by thecranial portion of the trochlear groove is less than thatprovided by the caudal portion. This hypothesis is sup-ported by the findings of Malghem and Maldague,14 whoreported that the depth of the proximal trochlear groove(as determined by lateral radiographs) was less than thedepth of the middle portion in subjects who were pain-free.

In contrast, the finding of an increasing sulcus anglewith knee extension in my investigation appears tocontradict the data of Farahmand and colleagues,20 whoreported that the geometry of the trochlear groove (asencountered by the sliding patella during knee flexion)

changed very little. Their findings, how-ever, were based on their analysis ofcadaver specimens under low-level,static loading conditions. I contend it islikely that the conditions used in myinvestigation (active quadriceps femorismuscle contraction/shortening) pulledthe patella farther superiorly in thetrochlear groove, thereby accounting forthe differences in the sulcus angles.

Although not significant, the averageincrease (flattening) of the sulcus angleduring extension in the subjects withPFP (19°) was almost twice that of thesubjects without PFP (10°). Althoughthis increase in the sulcus angle is indic-ative of compromised patellar stability,the etiological factor underlying thisfinding is not entirely evident. Forexample, there are 2 possible explana-tions for the increase in the sulcusangle: (1) dysplasia of the cranialportion of the femoral trochlea and(2) patella alta (excessive superiormigration of the patella with respect tothe trochlear groove). Although bothof these alternatives are possible, it isdifficult to separate the effects of eachwith regard to patellar tracking. Hvidand colleagues33 reported data thatsuggest that both findings are typicallyfound in conjunction with each other.Without knowing the vertical positionof the patella within the femoral troch-lea, however, it would be difficult toascertain whether an increased sulcusangle was the result of dysplasia or ofpatella alta, or a combination of both.

This determination would require further radiologicalevaluation, using lateral-view techniques that have beendescribed for assessing trochlear dysplasia14,34 and patellaalta35–37 or serial axial views to determine the exact positionof the patella within the trochlear groove.38

Despite the fact that the KMRI data collected in thisstudy were limited for assessing the exact vertical posi-tion of the patella, I contend that some qualitativeinformation was gained. For example, in 22% of thesubjects with PFP, it appeared that the patella wassuperior to the femoral trochlea, which would be sug-gestive of patella alta. As shown in Figure 7, the patella ofpatient 3 is situated on the shaft of the femur, well abovethe level of the femoral condyles. In contrast, patient 2demonstrates a relatively shallow trochlear groove,

Figure 7.Axial-plane images obtained from a subject without patellofemoral pain (PFP) and 3 subjectswith PFP (patients 1–3). The subject without PFP and patient 1 demonstrate a centered patellawithin the trochlear groove. Patient 2 demonstrates a moderate degree of lateral displacement(lateral border of patella lateral to the anterior femoral condyle) and lateral tilting as well as arelatively shallow trochlear groove. In patient 3, the patella is positioned well above thetrochlear groove, and there is extreme lateral displacement and lateral tilting of the patella.

Physical Therapy . Volume 80 . Number 10 . October 2000 Powers . 971

IIIIII

IIIIII

IIIIII

IIIIII

IIII

although the posterior femoral condyles are still visible,suggesting that this image section was not above the levelof the femoral trochlea. Therefore, an argument couldbe made that the diminished sulcus depth in this subjectwas more likely the result of trochlear dysplasia.

The bisect offset data obtained for both groups indi-cated that the patella was lateral to the midline through-out the range of motion. On the average, the subjectswith PFP demonstrated greater patellar lateralization atall angles of flexion. This finding, however, was notstatistically significant. The normal kinematic pattern forpatellar displacement was characterized by slight medialdisplacement from 45 to 18 degrees of knee flexion,followed by subtle lateral displacement as the kneeextended from 18 to 0 degrees (Fig. 3). This pattern ofmovement is consistent with that previously described asa frontal-plane “C” curve.39 Although, the average patel-lar displacement pattern of the subjects with PFP wassimilar to that of the subjects without PFP from 45 to 27degrees of flexion, there was a reversal to a progressivelymore lateral alignment as the knee continued to extend.The largest difference between groups was evident at 0degrees (62% versus 54% of the patella lateral to themidline), which coincides with the contention of Fulk-erson and Hungerford1 that patellar subluxation typi-cally occurs during terminal knee extension.

The bisect offset data of the subjects with PFP demon-strated large variability at 18, 9, and 0 degrees of flexion.At these angles, the standard deviations were approxi-mately 2 to 3 times those of the subjects without PFP,indicating that these subjects exhibited a wide range ofhorizontal patellar displacement (Fig. 3). At 0 degrees,for example, 22% of the subjects with PFP had a bisectoffset value greater than 2 standard deviations of thecomparison group, whereas 61% had a bisect offset valuewithin 1 standard deviation of the control group. Thesefindings support the work of Shellock et al,40 whoreported that only 26% of their subjects demonstratedlateral subluxation of the patella. Although the data ofShellock and colleagues40 were based on qualitative MRIassessment, the results of these previous studies, as well asthe data of my investigation, indicate that excessive lateraldisplacement of the patella is not a universal finding in thispopulation. The role of abnormal patellar kinematics as aprimary cause of PFP, in my view, may be questioned.

The patellar tilt data showed that the patella was laterallytilted throughout the range of motion in both groups,with the subjects with PFP demonstrating greater mag-nitudes compared with the subjects without PFP whenthe data were averaged across all knee flexion angles.These results suggest that excessive lateral tilt may be amore frequent radiological finding in PFP comparedwith lateral displacement or subluxation. A larger sam-

ple size (including male subjects), however, would benecessary to confirm this observation.

The subjects without PFP demonstrated an overall pat-tern of decreasing lateral tilt as the knee extended, which isconsistent with findings obtained with cadaver speci-mens41,42 and cine phase contrast imaging techniques.21

The average tilt values for the subjects, with PFP, however,remained fairly consistent across all knee flexion angles.This finding is in contrast to the data of Brossmann andcolleagues,22 which showed an overall tendency towardprogressive lateral tilt as the knee extended. This pattern ofmovement was evident in only 27% of the subjects with PFPin my investigation, which suggests that this should not beconsidered the dominant motion pattern. This discrepancycould have been the result of the difference in subjects inthe 2 studies, as well as the different measurement tech-niques used to determine patellar tilt.

The results of my study may have clinical implications forthe treatment of people with patellar malalignment. Forexample, if patellar tracking is primarily dictated by bonystructure, then treatment procedures that address onlysoft-tissue components (such quadriceps femoris musclestrengthening or a lateral retinacular release) may havelimited success. Likewise, the long-term success of aprocedure such as a distal realignment may depend onwhether the patella can be relocated within the bonyconfines of the trochlea.

A limitation of my study was the fact that a relativelysmall comparison group was used to provide comparisondata. Although differences were found with respect topatellar tilt, a larger sample size might have increasedthe ability to find group differences in the bisect offsetand sulcus angle measurements. Additional study in thisarea should consider larger sample sizes, particularlygiven the large variability among individuals with PFP. Apost hoc power analysis revealed that approximately 80and 110 subjects would be required to find group effects(10% differences) for the sulcus angle and bisect offset,respectively.

As a result of the limitations imposed by the size of theMRI bore, the loading condition used in this study (non–weight bearing) was not consistent with the loading condi-tion that would be evident with weight-bearing activities.Therefore, care should be taken in interpreting the resultsof this study until differences in patellar kinematics can beestablished between various loading conditions.

ConclusionsThe results of this study indicate that the sulcus angle isa predictor of both lateral patellar tilt and lateral patellardisplacement during terminal knee extension. This findingsuggests that bony structure is an important determinate of

972 . Powers Physical Therapy . Volume 80 . Number 10 . October 2000

patellar kinematics during this particular activity in youngwomen. Further research should be directed toward iden-tifying additional factors that can improve the predictabilityof patellar kinematics as well investigating the influence oflower-extremity function on patellar alignment.

References1 Fulkerson JP, Hungerford DS. Disorders of the Patellofemoral Joint. 2nded. Baltimore, Md: Williams & Wilkins; 1990.

2 Brattstrom H. Shape of the intercondylar groove normally and inrecurrent dislocation of the patella. Acta Orthop Scand. 1964;68:85–138.

3 Hvid I, Andersen LI, Schmidt H. Chondromalacia patellae: therelation to abnormal patellofemoral joint mechanics. Acta OrthopScand. 1981;52:661–666.

4 Paulos L, Rusche K, Johnson C, Noyes FR. Patellar malalignment: atreatment rationale. Phys Ther. 1980;60:1624–1632.

5 Vainionpaa S, Laasonen E, Patiala H, et al. Acute dislocation of thepatella: clinical, radiographic and operative findings in 64 consecutivecases. Acta Orthop Scand. 1986;57:331–333.

6 Wiberg G. Roentgenographic and anatomic studies on the femoro-patellar joint. Acta Orthop Scand. 1941;12:319–410.

7 Fox TA. Dysplasia of the quadriceps mechanism: hypoplasia of thevastus medialis muscle as related to the hypermobile patella syndrome.Surg Clin North Am. 1975;55:199–226.

8 Jeffreys TE. Recurrent dislocation of the patella due to abnormalattachment of the ilio-tibial tract. J Bone Joint Surg Br. 1963;45:740–743.

9 Puniello MS. Iliotibial band tightness and medial patellar glide inpatients with patellofemoral dysfunction. J Orthop Sports Phys Ther.1993;17:144–148.

10 Mariani PP, Caruso I. An electromyographic investigation of sub-luxation of the patella. J Bone Joint Surg Br. 1979;61:169–171.

11 Souza DR, Gross MT. Comparison of vastus medialis obliquus:vastuslateralis muscle integrated electromyographic ratios between healthysubjects and patients with patellofemoral pain. Phys Ther. 1991;71:310–316.

12 Boucher JP, King MA, Levebvre R, Pepin A. Quadriceps femorismuscle activity in patellofemoral pain syndrome. Am J Sports Med.1992;20:527–532.

13 Powers CM, Landel R, Perry J. Timing and intensity of vastus muscleactivity during functional activities in subjects with and without patel-lofemoral pain. Phys Ther. 1996;76:946–955.

14 Malghem J, Maldague B. Depth insufficiency of the proximaltrochlear groove on lateral radiographs of the knee: relation to patellardislocation. Radiology. 1989;170:507–510.

15 Geenen E, Molenaers G, Martens M. Patella alta in patellofemoralinstability. Acta Orthop Belg. 1989;55:387–393.

16 Insall J, Goldberg V, Salvati E. Recurrent dislocation and thehigh-riding patella. Clin Orthop. 1972;88:67–69.

17 Insall J, Falvo KA, Wise DW. Chondromalacia patellae: a prospectivestudy. J Bone Joint Surg Am. 1976;58:1–8.

18 Moller BN, Krebs B, Jurik AG. Patellar height and patellofemoralcongruence. Arch Orthop Trauma Surg. 1986;104:380–381.

19 Farahmand F, Senavongse W, Amis AA. Quantitative study ofquadriceps muscles and trochlear groove geometry related to instabil-ity of the patellofemoral joint. J Orthop Res. 1998;16:136–143.

20 Farahmand F, Tahmasbi MN, Amis AA. Lateral force-displacementbehavior of the human patella and its variation with knee flexion: abiomechanical study in vitro. J Biomech. 1998;31:1147–1152.

21 Sheehan FT, Zajac FE, Drace JE. Using cine phase contrast mag-netic resonance imaging to non-invasively study in vivo knee dynamics.J Biomech. 1998;31:21–26.

22 Brossmann J, Muhle C, Schroder C, et al. Patellar tracking patternsduring active and passive knee extension: evaluation with motion-triggered cine MR imaging. Radiology. 1993;187:205–212.

23 Powers CM, Shellock FG, Pfaff M. Quantification of patellar track-ing using kinematic magnetic resonance imaging. J Magn Reson Imag-ing. 1998;8:724–732.

24 Shellock FG, Mink JH, Deutsch A, Pressman BD. Kinematic mag-netic resonance imaging of the joints: techniques and clinical applica-tions. J Magn Reson Imaging. 1991;7:104–135.

25 Shellock FG, Mink JH, Deutsch AL, Foo TK. Kinematic MR imagingof the patellofemoral joint: comparison of passive positioning andactive movement techniques. Radiology. 1992;184:574–577.

26 van Kampen A, Huiskes R. The three-dimensional tracking patternof the human patella. J Orthop Res. 1990;8:372–382.

27 Stanford W, Phelan J, Kathol MH. Patellofemoral joint motion: evalu-ation by ultrafast computed tomography. Skeletal Radiol. 1988;17:487–492.

28 Sasaki T, Yagi T. Subluxation of the patella: investigation bycomputerized tomography. Int Orthop. 1986;10:115–120.

29 Schutzer SF, Ramsby GR, Fulkerson JP. The evaluation of patello-femoral pain using computerized tomography: a preliminary study.Clin Orthop. 1986;204:286–293.

30 Kujala UM, Osterman K, Kormano M, et al. Patellofemoral relation-ships in recurrent patellar dislocation. J Bone Joint Surg Br. 1989;71:788–792.

31 Goodfellow J, Hungerford DS, Woods C. Patello-femoral jointmechanics and pathology, 2: chondromalacia patellae. J Bone Joint SurgBr. 1976;58:291–299.

32 Seedhom BB, Takeda T, Tsubuku M, Wright V. Mechanical factorsand patellofemoral osteoarthritis. Ann Rheum Dis. 1979;38:307–316.

33 Hvid I, Andersen LI, Schmidt H. Patellar height and femoraltrochlear development. Acta Orthop Scand. 1983;54:91–93.

34 Grelsamer RP, Tedder JL. The lateral trochlear sign: femoraltrochlear dysplasia as seen on a lateral view roentgenograph. ClinOrthop. 1992;281:159–162.

35 Blackburne JS, Peel TE. A new method of measuring patellarheight. J Bone Joint Surg Br. 1977;59:241–242.

36 de Carvalho A, Andersen AH, Topp S, Jurik AG. A method forassessing the height of the patella. Int Orthop. 1985;9:195–197.

37 Insall J, Salvati E. Patella position in the normal knee joint.Radiology. 1971;101:101–104.

38 Shellock FG, Kim S, Mink JH, et al. “Functional” patella altadetermined with axial-plane imaging of the patellofemoral joint:association with abnormal patellar alignment and tracking. [abstract].J Magn Reson Imaging. 1992;2:93.

39 Hungerford DS, Barry M. Biomechanics of the patellofemoral joint.Clin Orthop. 1979;144:9–15.

40 Shellock FG, Mink JH, Deutsch AL, Fox JM. Patellar trackingabnormalities: clinical experience with kinematic MR imaging in 130patients. Radiology. 1989;172:799–804.

41 Nagamine R, Otani T, White SE, et al. Patellar tracking measure-ments in the normal knee. J Orthop Res. 1995;13:115–122.

42 Reider B, Marshall JL, Ring B. Patellar tracking. Clin Orthop.1981;157:143–148.

Physical Therapy . Volume 80 . Number 10 . October 2000 Powers . 973

IIIIII

IIIIII

IIIIII

IIIIII

IIII