Is there a biomechanical link between patellofemoral pain and osteoarthritis? A narrative review
Authors: Narelle Wyndow1, Natalie Collins1,2, Bill Vicenzino1, Kylie Tucker3, Kay Crossley1,4
1. Division of Physiotherapy, School of Health and Rehabilitation Sciences, The University of Queensland, St Lucia, Queensland, Australia.
2. Department of Mechanical Engineering, Melbourne School of Engineering, The University of Melbourne, Melbourne, Victoria, Australia.
3. School of Biomedical Sciences, The University of Queensland, St Lucia, Queensland, Australia.
4. School of Allied Health, College of Science, Health and Engineering, La Trobe University, Bundoora, Victoria, Australia.
Corresponding Author:
Professor Kay Crossley
Level E, Office Of Allied Health, HS3|05
Latrobe University
Plenty Road & Kingsbury Drive, Melbourne VIC 3086
Ph: 03 9479 3902
email:
Factors / Study Type / PFP Findings / PF OAFindings
Local
Sulcus Angle / Systematic Review / Lankhorst[1] /
  • Larger sulcus angle (WMD 1.61 95% CI 0.44 to 2.77)

Prospective / Stefanik[2] /
  • High sulcus angles increases odds of cartilage damage and bone marrow lesions (OR: 1.5, 95% CI: 1.1–2.1; p . 0.01; OR: 1.6, 95% CI: 1.1–2.3; p . 0.02)

Hunter [3] /
  • Midrange sulcus angles (2nd and 3rd quartiles) demonstrate trend to increased odds of medial joint space narrowing progression (J-shaped trend p = 0.07)
  • Greatest or smallest sulcus angles (1st or 4th quartiles) demonstrate trend to lateral joint space narrowing progression (P for trend = 0.04).

Kalichman[4] /
  • Increasing sulcus angleincreases risk of medial joint space narrowing (p for linear trend = 0.01)
  • Increasing sulcus angle increases risk of lateral patellar osteophytosis (p for linear trend = 0.08)
  • Increasing sulcus angle increases risk for medial patellar osteophytosis (p for linear trend = 0.05)

Cross-sectional / Tuna [5]
/
  • Larger sulcus angles in those with chondromalacia patella on MRI (147.10° ±7.29° vs. 134.61°±6.47°, p=<0.01)
/ Wu [6] /
  • Greater sulcus angle in those with patella malalignment and symptomatic knee OA vs those without patella malalignment and symptomatic knee OA (144.8 ± 7.37 vs 138.5 ±4.68, p = 0.001)

Harbaugh[7] /
  • Larger sulcus angle in the lateral vs. non-lateral symptomatic maltrackers (7.1%; p=0.009)
  • Sulcus angle not different between all maltrackers and asymptomatic individuals

Local
Trochlea Morphology / Cross-sectional / Tuna [5] /
  • Reduced trochlea depth3.14±0.79 mm vs. 4.81± 0.96 mm (p=<0.01)
/ Stefanik[2] /
  • Low lateral trochlea inclination angle increases odds of cartilage damage & bone marrow lesions vs those with high lateral trochlea inclination in lateral PFJ (OR: 2.6, 95% CI: 1.9–3.7; p < 0.0001; and OR: 2.3, 95% CI: 1.5–3.3; p < 0.0001)
  • Low lateral trochlear inclination increases odds of cartilagedamage vs those with high lateral trochlear inclination (OR: 1.6, 95% CI: 1.2–2.1; p = 0.003)
  • Low trochlear angles increases odds of cartilage damage in the lateral PFJvs those with high trochlear angles (OR: 2.0, 95% CI: 1.2–3.5; p < 0.0001)

Carrillon[8] /
  • Smaller lateral trochlea inclincation angle on MRI (6.17° ±, 4.97 vs. 16.93°, ± 4.76; P <0.001)

Keser [9] / Smaller lateral trochlea inclination angle on MRI (17.32° vs. 21.5°; p=<0.05) / Kalichman[10] /
  • Increasing trochlear angle associated with both WOMAC (p= 0.06) pain and WOMAC function subscale (p = 0.04).

Harbaugh[7] /
  • Smaller articular cartilage depth in maltrackersvs. asymptomatic controls (25.7%; p<0.001)
  • Smaller sulcus groove length in maltrackersvs. asymptomatic controls (8.0%; p=.031)
  • Greater trochlear depth in the lateral vs. non-lateral maltrackers (22.5%; p=0.015)
  • Greater lateral trochlea inclination angle in non-lateral maltrackers vs. both maltrackers and asymptomatic controls (20.0%; p=0.008 and 11.3% ;p=0.016 respectively)
  • Greater trochlear depth in the lateral vs. non-lateral maltrackers (22.5%; p=0.015)

Local
Patella Alta / Cross-sectional / Aglietti [11]
/
  • Higher patella height ratio on x-ray (Insall-Salvati method. 1.08 vs. 1.04, p=<0.01)
/ Kalichman[4] /
  • Increasing patella length ratio increases risk of lateral joint space narrowing (p for linear trend = 0.01)
Increasing patella length ratio associated with increasing lateral patellar osteophytosis (p for linear trend = 0.01)
Stefanik[2] /
  • Patella alta not associated with quadriceps weakness, lateral PFJ cartilage damage orbone marrow lesions (p for interaction =0.80)

Local
Patella displacement & tilt / Systematic Review / Lankhorst[1] /
  • Larger patellar tilt angle (WMD 4.34 95% CI 1.16 to 7.52)

Prospective / Hunter [3] /
  • Medial patella displacement predisposed to medial joint space narrowing progression (odds for each quartile 1, 1.2, 1.2, 2.2, p for trend=0.03)
  • Medial patella displacement protects from lateral joint space narrowing progression (odds for each quartile 1, 0.7, 0.6, 0.4, p for trend=0.0004)
  • Increasing patella tilt protected from medial joint space narrowing progression (odds for each quartile 1, 0.8, 0.5, 0.2, p<0.0001) and trended to increasing pain severity (p=0.09)

Cross-sectional / Tuna [5]
/
  • Lateral patellar tilt angle lower in those with CMP on MRI (10.08°± 4.34° vs. 16.97°± 3.55°; p=<0.01)
/ Kalichman[10] /
  • Increasing lateral patellar tilt angle and decreasing bisect offset (increasing lateral subluxation) show trends for increasing WOMAC pain (p=0.07, p=0.16)

Salsich[12] /
  • Higher bisectoffsetindex at 0° (0.69 ± 0.13 vs.. 0.64 ± 0.09, p = 0.04)
  • Higher patellar tilt angle at 0° (patellar tilt angle: 12.5 ± 7.6◦ vs.. 9.2 ± 5.8°, p = 0.04).
/ Kalichman[4] /
  • Low lateral patella tilt angle values (-25 to 13°) associated with greatest lateral joint space narrowing (p for linear trend < 0.0001).
  • Low lateral patella tilt angleassociated with and lateral patellar osteophytosis(p for linear trend < 0.0001).
  • Lateral patella displacement associated with increased lateral joint space narrowing (p for linear trend < 0.0001).
  • Lateral patella displacement associated with lateral patellar osteophytosis (p for linear trend < 0.0001)
  • Lateral patella displacement negatively associated with medial joint space narrowing (p for linear trend < 0.0026).

Souza [13]
/
  • Greater lateral patella tilt at 0°, 15° and 30° knee flexion on MRI (p=0.03)

Draper [14]
/
  • Greater bisect offset between knee flexion angles 0° and 50° on MRI (average 10%, p=0.03)
  • Greater lateral patella tilt between knee flexion angles 0° and 20° on MRI (average twice as large (10°vs. 4°), p=0.04)
  • Greater lateral patella displacement at 0°, 15°, 30° and 45° knee flexion on MRI (p=0.011)
/ Wu [6] /
  • Greater congruence angle in those with symptomatic knee OA and malaligned patella vs those without malaligned patella and symptomatic knee OA and healthy controls (mean ± SD: 14.66 ± 20.34, –4.61 ± 9.11, –4.80 ± 17.77, p = 0.001)
  • Greater lateral PFJ angle in those with symptomatic knee OA and malaligned patella vs those without malaligned patella and symptomatic knee OA and healthy controls (1.41 ± 6.81, 9.03 ± 4.04, 11.20 ±5.91, p 0.001)
  • Greater percentage of lateral patellar displacement in those with symptomatic knee OA and malaligned patella vs those without malaligned patella and symptomatic knee OA and healthy controls (91.3± 9.99, 103.7 ± 9.32, 101.5 ± 7.65, p 0.001)

Local
Q-angle, tibiofemoral alignment angle / Systematic Review / Lankhorst[1]
/
  • Larger Q-angle (WMD 2.08 95% CI 0.64 to 3.63)

Prospective / Myers [15] /
  • Knee abduction moments >15Nm on landing from a drop vertical jump a risk factor for PFP (6.8% risk)
/ Cahue[16] /
  • Varus malaligned knees have increased odds of radiographic medial PF OA progression (adjusted OR 1.85, 95% CI: 1.00–3.44)
  • Valgus malaligned knees have increased odds of radiographic lateral PF OA progression (adjusted OR 1.64, 95% CI 1.01– 2.66)

Witrvouw[17] /
  • No difference in Q-angle (11.45 vs. 13.01; p =0.394)

Cross-sectional / Nakagawa [18] /
  • Increased knee abduction at 15°, 30°, 45° and 60° during stepping (p = 0.013–0.001)
/ Gross [19] /
  • Varus malaligned knees (<−2°) higher prevalence of medial than lateral PFJ cartilage damage on MRI
  • Neutral aligned knees (-2° to 2°) have higher prevalence of medial than lateral PFJ cartilage damage on MRI
  • Valgus malaligned knees (>2°) lateral PFJ not greater than medial PFJ cartilage damage on MRI

Aglietti[11] /
  • Larger Q-angle on x-ray (20° vs. 15°, p=<0.001)
  • Larger congruence angle (-2° vs. -8°, p=<0.001)
/ Elahi[20] /
  • Valgus malaligned knees had greater lateralthan medial radiographic PF OA (57.3% vs 23.8%, p=0.0066)
  • Varus malalignment more likely in the medial PF OA
  • Valgus malaligned knees more likely to have isolated PF OA than isolated TF OA (p=0.0002), as were knees with mixed PF/TF OA (p =0.0006)

Sheehan [21] /
  • Q-angle associated with medial not lateral patella maltracking (non-weight bearing MRI)
/ Peat [22] /
  • Greater intermalleolar gap > 0 cm (valgus malalignment) more indicative of moderate to severe PF OA

Salsich[23]
/
  • Higher tibiofemoral rotation at 0°(6.4 ± 5.9◦ vs. 4.0 ± 4.6◦ , p = 0.07)

Local
Quadriceps strength/force/torque / Systematic Review / Lankhorst[1]
/
  • Greater peak force in VL, vastusintermedius and semitendinosusduring walking
  • Lower knee extension peak torque at 60° (Nm) compared to controls (WMD: −37.47 95% CI −71.75 to −3.20)

Lankhorst [24] /
  • Lower knee extension peak torques a risk factor for PFP:
    (a) standardized relative to body weight at 60°/s, –0.24 Nm (95% CI: –0.39, –0.09)
    (b) standardized relative to body weight at 240°/s, –0.11 Nm (95% CI: –0.17, –0.05)
    (c) standardized relative to body mass index at 60°/s, –0.84 Nm (95% CI: –1.23, –0.44)
    (d) standardized relative to body mass index at 240°/s, –0.32 Nm (95% CI: –0.52, –0.12)
    (e) nonstandardized in a concentric mode at 60°/s, –17.54 Nm (95% CI: –25.53, –9.54)
    (f) nonstandardized in a concentric mode at 240°/s, –7.72 Nm (95% CI: –12.67, –2.77).

Prospective / Amin [25] /
  • Greater quadriceps isokinetic knee extension strength protective against cartilage loss at lateral PFJ (highest vs lowest tertile of strength, OR 0.4, 95% CI 0.2, 0.9)

Cross-sectional / Callaghan [26] /
  • Lower peak torque difference PFP vs. controls (p = 0.002)
  • Lower peak torque between affected vs. unaffected limb (18.4%, 95% CI 13 to 23.8)
/ Fok[27] /
  • Lower knee extension moment during stair ascent in isolated PFOA or combined PF/TF OA vs controls (29%, –1.56% body weight x height, 95% CI –2.38, –0.74; p= 0.001; 44%, –2.38% body weight x height, 95% CI –3.27, –1.49, p 0.001)
  • Lower quadriceps forces during stair ascent in isolated PF OA and combined PF/TF OA vs controls (26%, –0.65 body weight, 95% CI –0.99, –0.31, p =0.001; 35%, –0.86 body weight, 95% CI –1.16, –0.53, p 0.001)
  • Lower knee extension moment during stair descent in isolated PF OA and combined PF/TF OA vs controls (51%, –2.78% body weight x height, 95% CI –4.30, –1.27, p 0.001; 38%, –2.04% body weight x height, 95% CI –3.68, –0.41, p 0.016)
  • Lower quadriceps forces during stair descent in isolated PF OA vs controls (42%, –0.94 body weight, 95% CI –1.52, –0.37, p = 0.002)
  • Lower quadriceps forces during stair descent in combined PF/TF OA vs controls (35%, –0.77 body weight, 95% CI –1.40, –0.16, p = 0.015)

Rathleff[28] /
  • 12-16 yr olds have no difference in isometric strength (81.5 vs. 81.8% body weight, p=0.97)

Rathleff[29] /
  • 15-19yr olds have lower normalized maximum isometric quad torque (2.27 vs. 2.80 N-m-kg-1 )

Yosmaoglu[30]
/
  • Lower peak isokinetic torque at 60° & 180°/s (60°: 58.1 ± 24.1 vs. 197.8 ± 9.8 Nm, p=<0.001; 180°: 40.2 ± 18.2 vs. 139 ± 11.5, p=<0.001)
  • Lower muscle endurance (concentric 357.5 ± 20.3 vs. 502 ± 72.1 (Nm); eccentric 285.6 ± 26.3 vs. 378.1 ± 57.6 (Nm) \0.001*

Local
Quadriceps strength/force/torque (cont.) / Cross-sectional
(Cont.) / Stefanik[2] /
  • Lowest concentric isokinetic quadriceps strength tertile have higher prevalence of lateral PFJ cartilage damage, medial PFJ cartilage damage, and lateral PF bone marrow lesions vs. knees in the highest strength tertile (10.2% lower, 95% CI 3-18; 9.1%, 95% CI 2-16; 7.1%, 95% CI 1, 13)

Baker[31] /
  • Reduced isometric quadriceps strength in isolated PF and combined PF/TF OA in women (OR 0.6, 95% CI 0.4–0.9; OR 0.4, 95% CI 0.3–0.6)
  • Similar trends in men (p =0.12)

Peat[22] /
  • Markedly reduced isometric knee extensor strength a clinical indicator of moderate to severe PF OA

Farrokhi[32] /
  • Lower quadriceps strength in those with severe PF OA and TF OA vs. no PFOA and mild PF OA groups (p < 0.001).

Farrokhi[33] / Lower knee extension strength associated with moderate/severe PF OA vs no PF OA (1.4±0.5 Nm/BW; 1.8±0.5 Nm/BW, p =0.03)
Hoglund[34] /
  • Decreased peak body mass-normalized isometric quadriceps force vs. controls (p = .009)

Within-subject / Kaya [35] /
  • Decreased torque in affected vs. unaffected limb at 60°/s (29%, p=<0.05).
  • No difference at 180°/s

Local
Quadriceps CSA / Cross-sectional / Callaghan [26] /
  • No significant difference in cross-sectional area in affected vs. unaffected limb(3.38%, 95% CI: 1.3 to 5.45, non-significant)
  • No significant difference between groups (1.31%, 95% CI: 0.06 to 2.55)
/ Hart [36] /
  • Smaller normalized VM, VL and rectus femoris volumes than controls (mean diff 0.90 cm3kg_1, α= 0.011; 1.50 cm3kg_1, α= 0.012; 0.71 cm3kg_1, α= 0.009)
  • No difference in VM/VL ratios observed

Within-subject / Kaya [35] /
  • Decreased total volume (55 cm3)and cross-sectional area (2.1 cm2) in affected vs. unaffected limb (p=<0.01)

Local
Vasti timing / Systematic Review / Chester [37] /
  • Delayed VMO relative to VL onset during stair ascent (pooled mean difference 17.7 ms, 95% CI: 3.8 ms to 31.6 ms). Not significant when Boling [38] excluded (pooled mean difference 12.03 ms, 95% CI. -0.17 ms to 24.23 ms).
  • Delayed VMO onset relative to VL during stair descent (pooled mean difference 30.25 ms, 95% CI: 16.68 ms to 43.81 ms). Not significant when Boling et al [38]excluded (pooled mean difference 21.33 ms, 95% CI. 16.47 ms to 26.19 ms)
  • Trends for delayed onset of VMO relative to VL during voluntary tasks such as lateral step downs, isometric/isokinetic knee extension tasks and reflex activity

Prospective / Van Tiggelen[39] /
  • Delayed onset of VMO relative to VL a contributing factor to PFP development (p=0.023)
  • Delayed VMO onset relative to VL in all participants with PFP after military training (P= <0.001)

Witvrouw[17] /
  • Shortened VMO reflex response time a risk factor for PFP development (p=0.005)

Cross-sectional / Wu[6] /
  • Lower VMO/VL ratio across all velocities in those with symptomatic knee OA +malaligned patella vs. those without malaligned patella + symptomatic knee OA and healthy controls (80°/sec p = 0.001; 120°/sec p = 0.37; 240° /sec p = 0.031)

Hinman[40] /
  • No differences VMO/VL onset timing were found between symptomatic knee OA vs. controls in either the concentric or eccentric phase of the stair-stepping task (F1,72= 0.45, p < 0.05)

Local
Hamstring timing, strength, length / Systematic Review / Lankhorst[1] /
  • Shorter hamstrings (79.1° ± 11.5°, 88.6°± 10.5°, 95% CI: –15.1; –4.1, t value: –3.6, p=<0.001)

Cross-sectional / Patil[41]
/
  • Earlier lateral hamstring activation relative to medial hamstring during a seated isometric task (8.2ms vs. 62 ms difference in onset timing)

Yosmaoglu[30] /
  • Lower peak isokinetic strength at 60 & 180 °/s (60°: 33.4 ± 15.8 vs. 103.5 ± 18.2, p=<0.001; 180°: 28.5 ± 12.3 vs. 105.3 ± 14.7, p=<0.001)

Local
Other / Systematic Review / Lankhorst[1]
/
  • Greater co-contraction of quadriceps and hamstrings at heel strike (p=.025)

Cross-sectional / Yosmaoglu[30] /
  • Higher tracking-trajectory error (concentric 1.8 ± 0.4 vs. 1.5 ± 0.5 cm; eccentric 1.7 ± 0.3 vs. 1.2 ± 0.3 cm; p=<0.001)
/ Farrokhi[32] /
  • Reduced loading-response knee flexion excursions and increased peak single-leg stance external knee flexion moments in severe PF OA (p = 0.002, p < 0.05)

Fok[27] /
  • Lower PFJ reaction forces in isolated PF and combined PF/TF OA during stair descent vs controls (47%, –0.80 body weight, 95% CI –1.24, –0.37, p = 0.001; 35%, –0.60 body weight, 95% CI –1.08, –0.13, p = 0.013)
  • Lower PFJ reaction forces in isolated PF OA and combined PF/TF OA during stair ascent vs controls (25%, –0.54 body weight, 95% CI –0.81, –0.28, p 0.001; 30%, –0.65 body weight, 95% CI –0.94, –0.37, p 0.001)
  • Higher PFJ reaction forces during stair descent associated with higher scores on KOOS sport/recreation and ADL subscales (r = 0.483, p = 0.007; r = 0.368, p = 0.045)

Study Type / PFP Findings / PF OA Findings
Proximal
Femoral rotation/adduction / Prospective / Boling [42] /
  • Increased femoral internal rotation angle during a jump landing a risk factor for PFP development (Unadjusted RR 1.30, 95% CI 0.60-2.82, confidence limits ratio 4.68, p = 0.50. Domain specific models RR 1.99, 95% CI 0.72 – 5.50, confidence limits ratio 7.66, p = 0.19)

Cross-sectional / Nakagawa [18]
/
  • Increased hip adduction at the angles evaluated except for 15° and 30° of knee flexion during stair descent (p = 0.021–0.001)
/ Hoglund[34] /
  • No difference in peak hip adduction angle during sit to stand (p = 0.61)

Barton [43] /
  • Reduced peak hip internal rotation during walking (7.08° vs. 11.88°, p = 0.024)
/ Pohl [44] /
  • No differences between PFOA vs. control groups for contra-lateral pelvic drop, hip adduction and hip internal rotation angles during walking

Souza [13] /
  • Greater medial femoral rotation at 0°, 15° and 45° of knee flexion on MRI (p = 0.037)

Souza [45] /
  • Greater peakhip internal rotation (7.6° ± 7.0° versus 1.2° ± 3.8°;p=<0.05). Females only

Powers [46] /
  • Less femoral internal rotation during walking (2.10° external rotation vs. 1.60° of internal rotation, p=0.03)
  • Later peak femur rotation (17.0% vs. 13.4% of the gait cycle, p = 0.05)

Within-subject / Barton [47] /
  • Greater hip adduction range of motion associated with greater rearfoot eversion range of motion (33% of variance, r = 0.573, P = 0.002)

Proximal
Gluteus activation / Systematic Review / Barton [48] /
  • Delayed and shorter activation of gluteus medius during stair negotiation and running (moderate-strong evidence, limited evidence)
  • Increased gluteus maximus activity during stair ascent (limited evidence)

Cross-sectional / Nakagawa [49] /
  • Less gluteus medius activation at 60° kneeflexion for both stair descent and ascent (p = 0.015, 0.005)

Souza [45] /
  • Increased activation of gluteus maximus during step-down and running tasks (mean± SD, 44.1%±30.6% vs. 23.1% ± 11.7% and 15.2%±8.8% vs. 9.3% ±4.8% MVIC respectively)

Proximal
Hip strength/forces / Systematic Review / Rathleff[50] /
  • Lower isometric hip strength not a riskfactor for PFP (Moderate to strong evidence, SMD: −0.24 to 0.06, I2=0–75%, p=0.02–0.93)
  • Lower isometric hip adduction and flexion strength not a risk factor for PFP (Moderate evidence, SMD: −0.06 to 0.19, I2=0–58%, p=0.12–0.41)

Prins[51]
(data not pooled) /
  • Reduced external hip rotation, abduction and extension strength compared to healthy controls (strong evidence)
  • Reduced hip flexion and internal rotation strength compared to healthy controls (moderate evidence)
  • Decreased external rotation and abduction strength compared to unaffected side (moderate evidence)

Lankhorst[1] /
  • Less hip abductor strength (percentagebody weight, %BW) (WMD −3.30; 95% CI −5.60 to −1.00)
  • Lesship external rotation strength (%BW) (WMD −1.43; 95%CI −2.71 to −0.16)

Cross-sectional / Rathleff[28] /
  • In 12-16yr olds there were no difference in isometric hip abduction, adduction, external or internal rotation strength
/ Pohl [44] /
  • Lower maximum isometric hip abduction strength vs. controls (28.1%BW vs. 37.4%BW, p = 0.01, ES = 0.97)
  • No difference in maximum isometric hip external rotation strength (p= 0.42)

Souza [45] /
  • Less peak isometric hip abduction torque (mean ± SD, 1.39 ± 0.41 vs. 1.62 ± 0.26 Nm/kg of body mass; p = .02; t value,–2.07; df, 39)
  • Less hip extension torque (mean ±SD, 1.98 ±0.50 vs. 2.35 ± 0.38 Nm/kg of body mass; p = .005; tvalue, –2.69; df, 39)
/ Crossley[52] /
  • Lower gluteus medius & gluteus minimus muscle forces vs. controls during walking (mean diff 0.15, 95% CI: 0.01 - 0.29 BW; mean diff: 0.03, 95% CI 0.01 - 0.06 BW)
  • No differences for Vasti or gluteus maximus muscle force (p=0.35; p=0.76)

Proximal
Hip strength/forces (cont.) / Cross-sectional
(Cont.) / Willson [53] /
  • Reduced external hip rotation strength (14%, p = 0.03)
/ Fok[27] /
  • Lower hip abductor forces during stair descent in isolated PF OA vs controls (27%, –0.47 body weight, 95% CI –0.81, –0.14, p = 0.006)
  • Lower hip abductor forces during stair descent in combined PF/TF OA vs controls (35%, –0.63 body weight, 95% CI –0.99, –0.28, p= 0.001)
  • No difference for gluteus maximus or hip abductor muscle forces (p=0.49; p=0.33)
  • Greater hip flexion in isolated PFOA and combined PF/TF OA during stair ascent vs controls (mean difference 4.6° [95% CI 0.7, 8.5, p = 0.023; 7.1°, 95% CI 2.8, 11.3, p=0.002)

Hoglund[34] /
  • Decreased peak body mass-normalized isometric force of the hip abductors and hip extensors vs. controls (ρ = .021; ρ = .021)
  • Decreased peak body mass-normalized isometric hip external rotator force trends (p = .121)

Proximal
Pelvic Kinematics / Cross-sectional / Nakagawa [49] /
  • Greater contra-lateral pelvic drop for all angles evaluated, except for 15° and 30° of knee flexion during stair descent (p = 0.034–0.001)

Willson[54] /
  • Greater contralateral pelvic drop at the end of single leg jump exertion protocol (PFP group = 1.1° less contralateral pelvic elevation, control group = 1.4° greater elevation, p = .003)

Proximal
Trunk strength & kinematics / Cross-sectional / Cowan [55]
/
  • Lower trunk side flexion strength (p=0.03)
/ Fok[27] /
  • Greater anterior pelvic tilt in isolated PF OA and combined PF/TF OA during stair ascent vs controls (mean difference 4.1°, 95% CI 1.6, 6.7, p= 0.002; 6.9°, 95% CI 4.1, 9.7, p 0.001)

Nakagawa [49] /
  • Greater ipsilateral trunk lean for all angles evaluated, except for 15° and 30° of knee flexion during stair descent (p<=0.001)

Willson[54] /
  • Reduced average lateral trunk flexion strength (24%, p = .06)

Study Type / PFP Findings / PF OA Findings
Distal
Rearfoot / Cross-sectional / Aliberti[56] /
  • Larger contact area over medial (1.8cm2) and central rearfoot (3.60cm2) at initial contact
  • Lower peak pressure under medial forefoot at propulsion

Barton [47] /
  • Greater rearfoot eversion associated with greater internal tibial rotation & hip adduction during walking (r=0.573, p=0.002)

Rodrigues
[57] /
  • Relatively more internal tibial rotation than rearfoot eversion at 34-38% stance phase of running vs. controls who had relatively more rearfoot eversion than tibial internal rotation (205° - 195°, Ev/TIR: 2.1-3.7 vs. 241° - 246°, Ev/TIR 0.6-0.4)

Rodrigues [58] /
  • Subtalar joint functions closer to end (4.21° eversion buffer vs. 7.25° buffer, p=0.03, ES=0.77)
  • No difference in “traditional” measures of foot pronation

Silva [59] /
  • Rearfoot eversion rangeof motion during stair ascent has moderate diagnostic accuracy to predict PFP from health controls (area under the curve = 0.72)

Duffy [60] /
  • Less pronation change in first 10% of stance (mean difference 25%, -1.30°, SD = -2.27 - -0.33, F = 7.52, p = 0.007
  • Less calcaneus-tibia touchdown angle (MD 2.80 95% CI 0.46 to 5.14)

Noehren[61] /
  • No difference in peak rearfoot eversion during running (p = 0.27)
  • Increased internal tibial rotation during running (−10.0° ± 5.3) −6.5 ± 3.0, p = 0.03, ES: 0.83)

Powers [46]
/
  • No difference in magnitude of peak foot pronation (walking) (8.9°, vs. 8.3°; p=0.29)
  • No difference in timing of peak foot pronation (19.8% gait cycle vs.19.8% gait cycle; p=0.49)
  • Slower walking speed (71.6 vs. 82.9 m/min; p=0.0002)
  • Decreased stride length (1.3 vs. 1.4 m; p=0.008)
  • Decreased cadence (113.9 vs. 122.1 steps/min; p=0.0009)

Distal
Rearfoot (Cont.) / Cross-sectional
(Cont.) / Callaghan [62]
/
  • Delayed timing of peak rearfoot eversion during walking (mean 0.211s ± 0.032 vs. 0.018s ±0.047, p = <0.05)

Levinger[63] /
  • Delayed timing of peak rearfoot eversion during walking (46 ± 6.5% vs. 39 ± 7%, p = 0.02)
  • No difference in peak rearfoot eversion angle during walking (p = 0.79)

Carvalho e Silva [59] /
  • No difference in rearfoot eversion range (p = 0.3)

Neal [64] /
  • Increased navicular drop a risk factor for PFP (Very limited evidence. SMD: 0.33, 95% CI: 0.02 to 0.65)
  • No association between pronated foot posture (defined by FPI and navicular drop) and increased risk of PFP (Limited evidence. RR: 1.22, 95% CI: 0.73 to 2.02)

Distal
Midfoot / Systematic Review / Lankhorst[1] /
  • No association between height index and PFP foundafter pooling (WMD 0.01; 95% CI −0.01 to 0.03)

Cross-sectional / Barton [65] /
  • Lower longitudinal arch angle in relaxed stance
    (Diff: 6.9°, 95% CI: 1.3 to 12.6, ES: 0.90, p= 0.019)
  • Greater longitudinal arch angle differencerelative to subtalar joint neutral (Diff: 3.0°, 95% CI: 0.9 to 5.1, ES: 0.90, p = 0.007)
  • Greater foot posture index (Diff: 2.4, 95% CI: 0.5 to 4.4. p = 0.15)
  • Greater normalized navicular drop (Diff: 1.6% of foot length 95% CI: 0.6 to 2.7, ES: 1.02, p = 0.003)
  • Greater normalized dorsal arch height difference[65] (Diff: 0.7% of foot length, 95%CI: 0.2 to 1.2, ES: 1.02, p = 0.005)
  • Greater normalized navicular drift (Diff: 1.6% of foot length, 95% CI: 0.6 to 2.7, ES: 0.92, 0.005)
  • Greater calcaneal angle difference relative to subtalar joint neutral (Diff: 2.6°, 95% CI: 0.8 to 4.3, ES: 0.75, p = 0.006)

Noehen[61] /
  • No difference in peak rearfoot eversion (P=0.27)
  • No difference in forefoot dorsiflexion (P=0.66), or forefoot abduction (P=0.16)

McPoil[66] /
  • Larger than normal differences in dorsal arch height between non-weight bearing and weight-bearing (odds ratio: 4.04, 95% CI: 1.45 to 11.32, p = 0.0002)
  • Greater foot mobility magnitude (odds ratio: 1.00, 95% CI: 0.61 to 3.99, p = 0.001)

Molgaard[67] /
  • Greater navicular drop (4.2 mm, 95% CI: 3.2 to 5.3 mmvs. 2.9 mm, 95%CI: 2.5 to 3.3 mm)
  • Greater navicular drift (2.6 mm, 95% CI: 1.6 to 3.7 mm] vs. 1.4 mm, 95% CI: 0.9 to 2.0 mm)

Carvalho e Silva [59] /
  • Greater rearfoot eversion range of motion during stair climbing (F:1,52 = 8.34, p = 0.006)

Distal
Whole foot pronation / Cross- sectional / Powers [46] /
  • No difference in peak magnitude or peak timing of foot pronation during walking (p = 0.29, p = 0.49 respectively)

Distal
Shank & Ankle / Systematic
Review / Lankhorst[1]
/
  • Greater peak force in gastrocnemius during walking & running (p=<0.001, p=<0.002)
  • Reduced gastrocnemius length (7.4° vs. 17.6°, p=<0.001)
  • Reduced soleus length (14.8° vs. 21.7°, p=0.001)

Cross-sectional / Collins [68] /
  • Reduced ankle joint dorsiflexion range of motion associated with greater peak hip adduction (-0.56, p=0.04) & hip internal rotation (-0.67, p=0.009) during a SLSq
  • No association with midfoot width mobility or arch height and hip angles
/ Hoglund[34] /
  • Greater peak tibial abduction angle during sit to stand (median diff = 4.8°, ρ = .040; mean diff means = 5.2°)
  • Greater tibial abduction angle at the end of sit to stand vs. control subjects (median diff = 4.7°, ρ = .021; mean diffs = 5.8°)
Increased peak tibial abduction angle moderately correlated with decreased hip abductor peak body mass-normalized isometric force (r = -.568, ρ = .027)
Leitch [69] /
  • Greater rearfoot eversion (peak & angle at toe-off)
  • Decreased rearfoot dorsiflexion (peak & excursion) (non-sig)
  • Higher knee internal rotation (peak & excursion) (non-sig)
  • Higher forefoot dorsiflexion (peak) (non-sig)
/ Fok[27] /
  • No difference in plantar flexor forces between isolated PF OA, combined PF/TF OA and controls when ascending or descending stairs (p = 0.69; p = 0.45)

Powers [46] /
  • No difference in peak magnitude or peak timing of tibial rotation during walking (p = 0.25, p = 0.26 respectively)

Molgaard[67] /
  • Higher passive ankle joint dorsiflexion in 16-18yr olds (22.28°, 95% CI: 188° to 268° vs. 17.78°, 95% CI: 158° to 208°)

Abbreviations: