ABSTRACT
Background and Purpose
The mechanistic rationale for treating the foot in anterior knee pain (AKP) is well established in the biomechanics literature; however, translating it into a clinically actionable evidence‐supported rehabilitation protocol has proven difficult. This investigation pursued two aims: to characterize the functional and biomechanical differences between patients with AKP with and without pronated foot posture, and to compare the short‐term effectiveness of two foot‐targeted physiotherapy regimens, manual therapy with augmented low‐dye taping (MT + ALDT) and neuromuscular training with ALDT (NMT + ALDT), against routine physiotherapy care.
Methods
A sequential two‐phase design was employed in the outpatient physiotherapy department of Mediclinic Al Noor Hospital, Abu Dhabi, UAE. Phase 1 was a prospective case‐control study (n = 50; pronated foot, n = 30; neutral/supinated, n = 20). Phase 2 was a single‐blind parallel‐group RCT that randomized 31 participants with confirmed pronation 1:1:1 to MT + ALDT (n = 11), NMT + ALDT (n = 10), or routine physiotherapy (n = 10) across 12 sessions in 4 weeks. Four outcomes were assessed at baseline and post‐intervention: NPRS, Kujala AKPS, FPI‐6, and DVI.
Results
Phase 1 revealed a 33.7‐point AKPS gap between pronated and neutral/supinated participants (50.20 vs. 83.90; p < 0.001, Mann–Whitney U) and a significant FPI‐6–DVI association (χ 2 = 4.22, p = 0.04). In Phase 2, ANCOVA confirmed that both foot‐targeted regimens significantly outperformed routine care on all four outcomes (all p < 0.001), with MT + ALDT demonstrating the greatest improvements. Effect sizes were large (η 2 = 0.646–0.958; 95% CIs reported in Results) and are treated as preliminary estimates warranting cautious interpretation given the small sample and single‐center setting.
Discussion
Foot pronation was strongly associated with impaired patellofemoral function, although the cross‐sectional design of Phase 1 precluded causal inference. Both foot‐targeted regimens outperformed routine care, and MT + ALDT demonstrated a stronger clinical profile. The equivalent biomechanical outcomes across both experimental arms were consistent with a contribution from ALDT, but its independent role could not be isolated without a taping‐isolated comparator arm. These preliminary single‐center findings support further investigation of foot posture assessment and distal intervention in AKP pending confirmation in adequately powered multicenter trials.
Trial Registration
ClinicalTrials.gov identifier: NCT05917080
1. Introduction
Anterior knee pain is deceptively common, and in a meaningful proportion of patients, it is deceptively difficult to resolve. The annual prevalence ranges from 22% to 29% in adolescent and young adult populations (Smith et al. 2018), and longitudinal follow‐up data suggest that approximately half of the patients report significant symptoms years after the first clinical encounter (Willy et al. 2019). These numbers raise an honest question about whether current rehabilitation strategies adequately address the full scope of the problem.
However, this is not the case for many patients. Conventional AKP management has remained heavily concentrated on the knee itself, quadriceps strengthening, patellar taping, and hip abductor protocols (Willy et al. 2019). These approaches yield genuine benefits for many, but patellofemoral dysfunction is genuinely multifactorial, arising from the convergence of biomechanical, neuromuscular, and kinematic variables distributed across the entire lower‐limb kinetic chain (Neal et al. 2014). Persistently high recurrence rates in clinical practice suggest that relevant distal contributors are routinely underweighted.
Pronated foot posture is among the most biomechanically plausible contributors. Excessive pronation medial arch collapse, rearfoot eversion, and midfoot hypermobility under load propagate proximally through subtalar motion coupling: internal tibial rotation and femoral adduction amplify dynamic knee valgus and concentrate patellar contact stress laterally on the trochlear surface (Powers et al. 2017; Barton et al. 2011). This cascade is well‐described. However, whether systematic correction modifies patient outcomes in AKP remains to be studied in controlled clinical settings.
Three physiotherapy approaches offer mechanistically coherent entry points for treatment. Talonavicular joint mobilization (TJM) addresses periarticular restriction at the midfoot that sustains compensatory pronation under load (Kim et al. 2022). Neuromuscular training (NMT) targets motor control deficits permitting dynamic arch collapse during functional movement (Hübscher et al. 2010; Myer et al. 2005; Hewett et al. 1999; Praveenkumar et al. 2025). Augmented Low‐Dye taping (ALDT) provides session‐by‐session mechanical arch realignment, establishing a corrected substrate for both exercise and manual therapy (Kumar et al. 2024; Franettovich et al. 2010; Vicenzino et al. 2007; Srinivasan et al. 2024). A direct head‐to‐head comparison of two ALDT‐anchored regimens against routine care has not been formally evaluated. A fully combined TJM + NMT + ALDT arm was not included in this design; the rationale is explained in Section 2.3 and revisited in the limitations section.
This study addressed this gap using a two‐phase sequential design. Phase 1 characterized the functional and biomechanical differences between patients with AKP stratified by foot posture. Phase 2 evaluated, within a single‐blind RCT, whether MT + ALDT or NMT + ALDT produced superior outcomes compared with routine physiotherapy without foot‐specific intervention and whether the two active regimens differed meaningfully from each other. A priori hypotheses were specified as follows: (i) pronated‐foot individuals would show significantly greater functional impairment and dynamic valgus than neutral/supinated counterparts, and (ii) both foot‐targeted regimens would produce statistically and clinically superior improvements over routine care.
2. Methods
2.1. Study Design and Ethical Approval
This investigation used a sequential two‐phase design: a prospective cross‐sectional case‐control study (Phase 1) followed by a prospective, assessor‐blinded, parallel‐group three‐arm RCT (Phase 2). Both phases were conducted at the outpatient physiotherapy department of Mediclinic Al Noor Hospital in Abu Dhabi, UAE.
Ethical approval was granted by the Mediclinic Middle East Clinical Research Ethics Committee (Reference: MCME.CR.310.MNOO.2023). The Phase 2 trial was prospectively registered on ClinicalTrials.gov (NCT05917080) before the first participant was enrolled. All participants provided written informed consent prior to any study procedure, and all aspects of the study conformed to the 2013 revision of the Declaration of Helsinki. Phase 2 reporting followed the CONSORT 2025 Statement for randomized trials (Table S1).
2.1.1. Phase 1: Case‐Control Component
Fifty adults aged 18–55 years with clinically diagnosed AKP were enrolled by convenience sampling and stratified into a case group (n = 30; FPI‐6 ≥ +6) and a control group (n = 20; FPI‐6 < +6). Clinical diagnoses were established by the principal investigator using criteria consistent with prior AKP trials (Dos Anjos Rabelo et al. 2014). Three validated outcome measures were administered: AKPS, FPI‐6, and DVI. The inclusion and exclusion criteria are presented in Table 1.
TABLE 1.
Presents participants fulfilling the inclusion and exclusion criteria.
| Inclusion criteria | Exclusion criteria |
|---|---|
|
|
Abbreviations: AKPS, Anterior Knee Pain Scale; NPRS, Numerical Pain Rating Scale; NSAIDs, non‐steroidal anti‐inflammatory drugs; PFPS, patellofemoral pain syndrome.
2.1.2. Phase 2: RCT Component
Phase 2 required AKP of ≥ 12 weeks’ duration and FPI‐6 ≥ +6 at screening, with enrollment from the same clinical setting as Phase 1.
2.2. Participants, Sample Size, and Randomization
The target Phase 2 sample was 48 participants (16 per arm), based on an a priori power calculation in G*Power 3.1 (one‐way ANOVA; AKPS between‐group difference = one MCID = 13 points; SD ≈ 12; α = 0.05; power = 0.80), yielding a minimum of 42, inflated to 48 for 10%–15% projected attrition. Recruitment commenced on 15 June 2023.
Enrollment fell short of the target. Insurance pre‐authorization delays, scheduling constraints, and cultural factors limiting female participation are well‐recognized access barriers in UAE private‐sector physiotherapy; they confined recruitment to 31 participants (Figure 1). Crucially, no participant withdrew after randomization; all 31 completed the 4‐week program with complete outcome data, making the per‐protocol analysis analytically equivalent to intention‐to‐treat. The power implications are discussed in Section 5.
FIGURE 1.
Flowchart of the randomized controlled trial (RCT) design. OPD, outpatient department; VMO, vastus medialis oblique.
The allocation sequence was computer‐generated using blocked randomization (block size = 3, ensuring 1:1:1 balance) by an independent statistician uninvolved in any other aspect of the study. The sequence was concealed in sequentially numbered, opaque, sealed envelopes held by a designated patient administrator with no treatment or assessment role; envelopes were opened in order only after eligibility and consent were confirmed.
Outcome assessors administering the NPRS, AKPS, FPI‐6, and DVI remained blinded to group allocation and played no treatment role, establishing a single‐blind, assessor‐blinded design. Participants and treating physiotherapists could not be blinded, given the physically distinct nature of taping, joint mobilization, and progressive exercise. Blinding success was not formally evaluated, which is a limitation noted in Section 5.
2.3. Interventions
All groups attended 12 sessions over 4 weeks (three per week; 30–40 min each). A common base program was shared by all participants: intrinsic foot muscle exercises using a short‐foot protocol (2 sets, 5 min), calf and plantar fascia stretching (four 30‐s holds), and progressive VMO‐targeted isometric quadriceps strengthening across three levels. This shared foundation ensures that between‐group differences in outcomes reflect group‐specific components and not variations in baseline exercise exposure.
A brief note on the design rationale: ALDT was paired separately with MT in one experimental arm and with NMT in the other, rather than combined in a single three‐component arm, because administering TJM, a full progressive NMT circuit, the base program, and taping within one 30–40‐min session was judged clinically impractical and potentially dilutive of individual component dose. The paired design allows for the direct evaluation of which adjunct produces a greater benefit alongside ALDT and supports a head‐to‐head comparison between regimens. Whether an integrated TJM + NMT + ALDT arm would add incremental benefits remains a specific priority for future investigations (Section 5).
2.3.1. Experimental‐I (MT + ALDT)
Talonavicular joint mobilization was performed in the prone position using the Maitland Grade III technique, with high‐amplitude oscillations at the end‐range and one‐second mid‐range vibrations until tissue resistance was encountered. The therapist stabilized the calcaneus with the dominant hand while applying a controlled dorsal glide to the navicular; two five‐minute sets per session. ALDT followed the Vicenzino protocol: a spur strip from the first metatarsal head around the posterior heel to the fifth metatarsal head, reinforced with mini‐strips, reverse strips, lock‐off strips, a calcaneal sling, and circumferential anchors to limit rearfoot eversion and support the arch of the foot.
2.3.2. Experimental‐II (NMT + ALDT)
The progressive NMT program targeting dynamic balance, lower‐limb coordination, agility, and proprioception included knee‐over‐toe lunges on stable and unstable surfaces, lateral sidestepping, tandem and semi‐tandem walking, high‐knee marching, and graded TheraBand foam‐pad stability work (green‐to‐silver progression). Advancement to unstable surfaces required sustained single‐limb support for > 10 consecutive seconds without observable postural compensation. The ALDT was applied identically to Experimental‐I.
2.3.3. Control (Routine Physiotherapy)
The control group completed the shared base program only, with no joint mobilization, structured NMT, or taping, reflecting standard outpatient physiotherapy for AKP at the study site.
2.4. Outcome Measures
2.4.1. NPRS
Average pain during daily functional activities over the preceding week was rated on an 11‐point scale (0 = no pain; 10 = worst imaginable). The test‐retest ICC > 0.85, and it has strong criterion validity against the visual analogue scale in musculoskeletal samples (Pagé et al. 2012; Suzuki et al. 2020).
2.4.2. AKPS
The 13‐item Kujala Anterior Knee Pain Scale (0–100; higher = better) has a test‐retest ICC of 0.95 and an MCID of 13 points (Haddad et al. 2021; Da Cunha et al. 2013). Both validated Arabic and English versions were available. The MCID was applied throughout the study to distinguish clinically meaningful changes from statistically incidental changes.
2.4.3. FPI‐6
The Foot Posture Index evaluates six standardized criteria in a relaxed bilateral stance (composite −12 to +12). Scores ≥ +6 define a pronated foot per Redmond et al.’s original validation; inter‐rater ICC 0.79–0.88 in clinical settings (Redmond et al. 2006).
2.4.4. DVI
The mean knee valgus angle across the three single‐leg squat phases (initial contact, peak flexion, propulsive rebound) was captured by a standardized smartphone video and analyzed with 2D software; test‐retest ICC 0.68–0.83. DVI ≥ 10° was considered clinically significant (Scholtes and Salsich 2017; Crossley et al. 2004).
2.5. Statistical Analysis
SPSS v22.0 (IBM Corp., Armonk, NY, USA; α = 0.05, two‐tailed). Shapiro–Wilk testing confirmed non‐normality for Phase 1 AKPS data; the Mann–Whitney U test compared functional scores between foot‐posture groups, and a chi‐square test examined the FPI‐6–DVI association. In Phase 2, one‐way ANOVA verified baseline homogeneity; within‐group pre‐to‐post changes were assessed using paired t‐tests with Cohen's d (mean difference/pooled pre‐intervention SD). Between‐group post‐intervention differences were estimated using ANCOVA with age and BMI as a priori covariates (pre‐specified in the registered protocol on biomechanical grounds), with Bonferroni correction for pairwise comparisons. Partial η 2 = SS_treatment/(SS_treatment + SS_error) per established convention; ANCOVA assumptions (linearity, slope homogeneity, error‐variance homogeneity via Levene's test) were verified prior to interpretation, and no violations were identified.
The effect size estimates in this dataset are unusually large (Cohen's d up to 8.80; partial η 2 up to 0.958), warranting independent verification. Ninety‐five percent confidence intervals (95% CIs) for Cohen's d were computed using the noncentral t‐distribution approach, with SE(d) ≈ √(1/n + d 2/[2(n − 1)]). Ninety‐five percent CIs for partial η 2 were derived from the noncentral F‐distribution method using the MBESS package in R. Wide CIs are expected, given the small per‐arm sample sizes, and are reported to aid in the transparent interpretation of effect size magnitudes. A second analyst was provided direct access to the raw data and SPSS output files and independently recomputed all calculations; the recalculated values confirmed those reported previously. These magnitudes most plausibly reflect large absolute between‐group differences combined with low within‐group variability in a highly selected, motivated clinical sample with high protocol adherence, a pattern consistent with the pronounced clinical contrast between foot‐targeted and routine care in this specific context, but one that almost certainly overestimates what would be observed in larger, more heterogeneous populations. All effect sizes were treated as preliminary. The raw data, syntax, and verification files are available from the corresponding author upon request.
3. Results/Findings
3.1. Phase 1: Case‐Control Findings
Phase 1 demographic and clinical data are shown in Table 2. Mean age was 28.62 ± 4.75 years; mean BMI 27.28 ± 1.83 kg/m2; 88% male. The case and control groups were well‐matched for both demographic variables.
TABLE 2.
Phase 1 demographic and clinical characteristics by group.
| Variable | Case group (pronated) n = 30 | Control group (neutral/supinated) n = 20 | Overall N = 50 | p‐value |
|---|---|---|---|---|
| Age, years | 28.0 ± 5.4 | 29.5 ± 3.53 | 28.62 ± 4.75 | 0.22 |
| Gender (M/F) | 26/4 (86.7%) | 18/2 (90%) | 44/6 (88%) | — |
| BMI, kg/m2 | 27.58 ± 1.85 | 26.84 ± 1.74 | 27.28 ± 1.83 | 0.16 |
| Kujala AKPS | 50.20 ± 5.28 | 83.90 ± 8.36 | 63.68 ± 17.94 | < 0.001 a |
| DVI (moderate/severe) | 17 (56.7%)/13 (43.3%) | 5 (25%)/0 | — | — |
| FPI (pronated/highly pronated) | 8 (26.7%)/22 (73.3%) | — | — | — |
Note: Values are expressed as: Mean ± SD or n (%).
Abbreviations: AKPS, Anterior Knee Pain Scale; DVI, Dynamic Valgus Index; FPI, Foot Posture Index.
Mann–Whitney U test.
The functional gap between the groups was clinically significant. Pronated‐foot participants averaged an AKPS of 50.20 ± 5.28 (moderate‐to‐severe functional impairment) versus 83.90 ± 8.36 in the neutral/supinated group, a 33.7‐point differential exceeding the MCID by more than twofold (Mann–Whitney U = 0.000, p < 0.001; Figure 2A–C). All 30 case group participants demonstrated clinically significant dynamic valgus: moderate in 56.7% and severe in 43.3%; 73.3% were also classified as highly pronated on FPI‐6. A significant chi‐square association was observed between the FPI‐6 category and DVI severity (Figure 3). (χ 2 [1, N = 30] = 4.22, p = 0.04) confirmed that greater pronation correlated with more pronounced dynamic valgus. DVI was not significantly associated with sex (p = 0.43), BMI (p = 0.25), or Kujala severity category (p = 0.15).
FIGURE 2.
Phase 1 case‐control results. (A) Kujala AKPS scores by foot posture group (***p < 0.001, Mann–Whitney U). (B) Distribution of Dynamic Valgus Index categories. (C) Foot Posture Index classification. Error bars represent ± 1 SD.
FIGURE 3.
Scatter plot of FPI‐6 against DVI scores in the Phase 1 case group (n = 30). A significant positive association was identified (χ 2 = 4.22, p = 0.04). Reference lines indicate clinical threshold values for pronation (FPI ≥ +6) and dynamic valgus (DVI ≥ 10°).
3.2. Phase 2: Randomized Controlled Trial Findings
3.2.1. Baseline Group Comparability
One‐way ANOVA confirmed baseline homogeneity across groups for all outcomes: NPRS (F = 0.248, p = 0.782), AKPS (F = 0.100, p = 0.905), FPI‐6 (F = 0.161, p = 0.852), and DVI (F = 0.673, p = 0.518) (Table 3). Randomization produced comparable groups, supporting valid between‐group comparisons.
TABLE 3.
Baseline descriptive statistics and one‐way ANOVA—Phase 2.
| Outcome | Experimental‐I: MT + ALDT (n = 11) | Experimental‐II: NMT + ALDT (n = 10) | Control‐III (n = 10) | F (p) |
|---|---|---|---|---|
| NPRS | 7.18 ± 0.75 | 7.30 ± 0.67 | 7.40 ± 0.70 | 0.248 (0.782) |
| AKPS | 44.73 ± 3.72 | 45.40 ± 3.27 | 45.00 ± 3.30 | 0.100 (0.905) |
| FPI‐6 | 8.45 ± 1.21 | 8.60 ± 1.17 | 8.30 ± 1.16 | 0.161 (0.852) |
| DVI, ° | 14.84 ± 1.12 | 15.27 ± 1.09 | 15.29 ± 0.81 | 0.673 (0.518) |
Note: Values: mean ± SD.
Abbreviations: AKPS, Kujala Anterior Knee Pain Scale; DVI, Dynamic Valgus Index; FPI‐6, Foot Posture Index; MT + ALDT, Manual Therapy + Augmented Low‐Dye Taping; NMT + ALDT, Neuromuscular Training + Augmented Low‐Dye Taping; NPRS, Numerical Pain Rating Scale.
3.2.2. Within‐Group Changes
Within‐group data are presented in Table 4 and Figure 4 (graphical data for effect sizes: Figure S1). Both experimental groups showed significant and clinically meaningful improvements across all four outcomes (all p < 0.001). MT + ALDT achieved NPRS reduction of 5.36 points (7.18 → 1.82; Cohen's d = 5.22, 95% CI [2.86, 7.58]) and an AKPS gain of 45.00 points (44.73 → 89.73; d = 6.94, 95% CI [3.84, 10.04]), exceeding the MCID by more than threefold. NMT + ALDT produced comparable but somewhat smaller gains: NPRS Δ = 3.80 (d = 4.82, 95% CI [2.51, 7.13]); AKPS Δ = 32.70 (d = 6.94, 95% CI [3.68, 10.20]). DVI fell from 14.84° to 11.73° in MT + ALDT (d = 8.53, 95% CI [4.74, 12.32]) and from 15.27° to 12.04° in NMT + ALDT (d = 8.80, 95% CI [4.69, 12.91]). The wide CIs reflect small per‐arm sample sizes and are consistent with the preliminary interpretation of these estimates. The control group showed negligible changes: NPRS Δ0.30 (p = 0.157), FPI‐6 Δ0.40 (p = 0.104), and DVI Δ0.40° (p = 0.076). An AKPS gain of 1.50 points reached statistical significance (p = 0.033) but fell below the 13‐point MCID and carried no clinical weight.
TABLE 4.
Pre‐ and post‐intervention within‐group comparisons (paired t‐tests).
| Outcome | Group | Pre (Mean ± SD) | Post (Mean ± SD) | Mean Δ | t (df) | p | Cohen’s d |
|---|---|---|---|---|---|---|---|
| NPRS | Experimental‐I | 7.18 ± 0.75 | 1.82 ± 0.75 | 5.36 ± 1.03 | 17.323 (10) | < 0.001 | 5.22 |
| Experimental‐II | 7.30 ± 0.67 | 3.50 ± 0.71 | 3.80 ± 0.79 | 15.234 (9) | < 0.001 | 4.82 | |
| Control | 7.10 ± 0.72 | 6.80 ± 0.75 | 0.30 ± 0.50 | 1.560 (9) | 0.157 | 0.40 | |
| AKPS | Experimental‐I | 44.73 ± 3.72 | 89.73 ± 4.41 | 45.00 ± 6.48 | 23.029 (10) | < 0.001 | 6.94 |
| Experimental‐II | 45.40 ± 3.27 | 78.10 ± 6.28 | 32.70 ± 6.63 | 15.587 (9) | < 0.001 | 6.94 | |
| Control | 44.50 ± 3.50 | 46.00 ± 3.80 | 1.50 ± 1.20 | 2.500 (9) | 0.033 | 0.80 | |
| FPI‐6 | Experimental‐I | 8.45 ± 1.21 | 3.00 ± 1.34 | 5.45 ± 2.02 | 8.964 (10) | < 0.001 | 2.70 |
| Experimental‐II | 8.60 ± 1.17 | 3.10 ± 1.37 | 5.50 ± 1.72 | 10.136 (9) | < 0.001 | 3.21 | |
| Control | 8.50 ± 1.19 | 8.10 ± 1.20 | 0.40 ± 0.60 | 1.800 (9) | 0.104 | 0.50 | |
| DVI, ° | Experimental‐I | 14.84 ± 1.12 | 11.73 ± 1.15 | 3.11 ± 0.36 | 28.285 (10) | < 0.001 | 8.53 |
| Experimental‐II | 15.27 ± 1.09 | 12.04 ± 1.10 | 3.23 ± 0.42 | 25.420 (9) | < 0.001 | 8.80 | |
| Control | 15.00 ± 1.10 | 14.60 ± 1.15 | 0.40 ± 0.50 | 2.000 (9) | 0.076 | 0.60 |
Note: Experimental‐I: MT + ALDT; Experimental‐II: NMT + ALDT. Values: mean ± SD. MCID for AKPS: 13 points.
All p‐values for the experimental groups were < 0.001.
FIGURE 4.
Pre‐ versus post‐intervention comparisons across all four outcome measures and three groups. Light bars = pre‐intervention; solid bars = post‐intervention. ***p < 0.001 (paired t‐test). Error bars = ± 1 SD.
3.2.3. Between‐Group ANCOVA and Post‐Hoc Comparisons
The between‐group effect sizes in this section are unusually large and were independently verified, as described in Section 2.5; raw materials are available upon request.
ANCOVA confirmed highly significant between‐group differences across all four outcomes after adjusting for baseline values, age, and BMI (Table 5; Figure S2). The largest effect was observed for DVI (F [2, 25] = 282.027, η 2 = 0.958, 95% CI [0.940, 0.967]), with group membership explaining approximately 96% of the post‐intervention DVI variance after covariate adjustment. Large effects were also confirmed for AKPS (η 2 = 0.838, 95% CI [0.778, 0.884]), NPRS (η 2 = 0.778, 95% CI [0.694, 0.838]), and FPI‐6 (η 2 = 0.646, 95% CI [0.562, 0.724]).
TABLE 5.
ANCOVA‐adjusted post‐intervention means, effect sizes, and Bonferroni post‐hoc comparisons.
| Outcome | Adj. mean Experimental‐I (SE) | Adj. mean Experimental‐II (SE) | Adj. mean control (SE) | F (2, 25) | p | η 2 | Post‐hoc |
|---|---|---|---|---|---|---|---|
| NPRS | 1.853 ± 0.231 | 3.480 ± 0.237 | 4.982 ± 0.234 | 43.702 | < 0.001 | 0.778 | I > II > III* |
| AKPS | 89.912 ± 1.532 | 77.835 ± 1.577 | 64.762 ± 1.553 | 64.823 | < 0.001 | 0.838 | I > II > III* |
| FPI‐6 | 2.853 ± 0.398 | 3.210 ± 0.410 | 6.352 ± 0.405 | 22.762 | < 0.001 | 0.646 | I = II > III* |
| DVI, ° | 12.004 ± 0.087 | 11.899 ± 0.089 | 14.516 ± 0.089 | 282.027 | < 0.001 | 0.958 | I = II > III* |
Abbreviations: η, partial eta squared; Adj. Mean, ANCOVA‐adjusted mean; SE, standard error.
p < 0.001 (Bonferroni). I–II indicates no significant difference between experimental groups.
Bonferroni post‐hoc comparisons confirmed that MT + ALDT outperformed both NMT + ALDT and the control on NPRS and AKPS (all p < 0.001). NMT + ALDT significantly exceeded the control group in all four outcomes (all p < 0.001). Critically, MT + ALDT and NMT + ALDT did not differ in terms of FPI‐6 (mean difference = −0.356, p = 1.000) or DVI (mean difference = 0.106, p = 1.000). Because ALDT was present in both experimental arms and absent from the control, these equivalent biomechanical outcomes are consistent with a contribution from ALDT to structural correction. However, the study design did not include a taping‐only or no‐taping comparator arm; therefore, the independent contribution of ALDT cannot be isolated from these data. These results indicate that adjunctive pairing of MT or NMT did not differentially influence biomechanical endpoints in this sample, with implications for resource allocation discussed in Section 4.2.
4. Discussion
Two findings stand out and require discussion. First, pronated foot posture was associated with a clinically substantial functional deficit in adults with AKP that was larger than expected if the foot was merely a secondary contributor. Second, both foot‐targeted regimens produced substantially greater short‐term improvements than routine physiotherapy. This section discusses the implications of these findings, what they cannot support, and how they fit into the broader literature.
4.1. Phase 1: Association, Not Causation
A 33‐point AKPS differential is not subtle; it approximates the difference between a patient who can descend stairs with manageable discomfort and one who cannot. That magnitude plausibly maps onto the biomechanical cascade described by Buldt et al. (2013), in which subtalar pronation drives internal tibial rotation that, in a closed kinetic chain, amplifies valgus and shifts patellar contact stress laterally (Yang et al. 2022; Wyndow et al. 2016). The significant FPI‐6–DVI association (p = 0.04) within the pronated group corroborates this cascade in a routine outpatient setting rather than a controlled laboratory (Udhaya Kumar et al. 2025).
An important caveat is causality. Phase 1 was cross‐sectional and demonstrated co‐occurrence, not directionality. Foot pronation cannot be concluded to cause patellofemoral dysfunction from these data; unmeasured confounders chronicity, prior treatment history, and fear‐avoidance cognitions may account for part or all of the association. The absence of DVI associations with sex (p = 0.43) or BMI (p = 0.25) is consistent with foot pronation operating as a relatively independent kinematic correlate in this specific cohort, which was predominantly male.
4.2. Why Foot‐Targeted Regimens Outperformed Routine Care
The effect sizes in this dataset are far larger than those typically reported in rehabilitation research, and transparency requires addressing this at the outset. Cohen's d values of 5–9 do not appear in routine physiotherapy RCTs. After independent verification confirmed the arithmetic (Section 2.5), the most plausible explanation is that very large absolute between‐group differences themselves reflecting a substantial clinical contrast between foot‐targeted and standard care in this selected sample were accompanied by low within‐group variability, producing exceptionally large effect estimates. These should be treated as preliminary, context‐specific observations that are almost certain to attenuate in larger and more heterogeneous populations.
None of the proposed mechanisms were measured directly. Rearfoot kinematics, muscle activation patterns, and talonavicular arthrokinematics were not assessed; the following are hypothesis‐driven interpretations grounded in existing biomechanics literature and not direct findings of this study. The ALDT is theorized to limit rearfoot eversion before each loaded repetition, potentially providing a more favorable mechanical substrate within which subsequent exercise or manual therapy may operate; however, this mechanism, although biomechanically plausible, was not directly assessed in this study. Franettovich et al. (2010) demonstrated an approximately 12% reduction in tibialis anterior activation during walking with ALDT, suggesting that taping modifies the neuromuscular demand across the kinetic chain. TJM may additionally restore arthrokinematics freedom at the talonavicular joint, disrupting the restriction proposed to perpetuate pronation during gait (Kim et al. 2022). These are testable mechanistic hypotheses requiring instrumented biomechanical investigation.
The equivalent biomechanical outcomes across both experimental arms on the FPI‐6 and DVI are the findings with the clearest resource‐allocation implications. The choice of MT versus NMT as an adjunct to ALDT did not appear to influence the structural endpoints in this sample. Earlier drafts of this manuscript described ALDT as the ‘primary driver’ of biomechanical correction; that characterization overstates what this design can support and has been removed from the manuscript. The design supports the interchangeability of adjunctive modality MT or NMT in its structural effects, a finding with potential relevance for service planning, pending replication with taping‐isolated comparator arms.
4.3. Contextualizing the Findings
Kim et al. (2022) reported approximately 40% pain reduction and 25% DVI improvement following TJM with NMT in an adolescent cohort with AKP. MT + ALDT in the present study achieved 75% pain reduction, and NMT + ALDT achieved 52% substantially larger pain reduction, likely reflecting ALDT's added mechanical contribution, the VMO‐targeted strengthening protocol, or population‐specific characteristics. None of these explanations can be isolated from the data; each is a hypothesis worth testing (Silva et al. 2019).
The negligible outcomes in the control group align with the position of Willy et al. (2019) that knee‐focused physiotherapy may be insufficient when the primary biomechanical driver of AKP lies distal to the treated joint. Importantly, the control condition was not inert; participants completed a structured strengthening and stretching program across 12 sessions with regular clinical contact. The fact that even this level of structured care failed to produce threshold‐level functional gains strengthens the argument that foot‐targeted components add something specific to the treatment.
5. Limitations
These findings carry several material limitations.
The Phase 2 sample (n = 31) fell substantially short of the planned 48 owing to UAE‐specific access barriers rather than treatment failure. With per‐arm ns of 10–11, the statistical power for smaller between‐group contrasts is limited, and the obtained effect sizes should be regarded as preliminary, upwardly biased estimates. An unmeasured site‐ or population‐specific contributor to the effect magnitude cannot be excluded.
The 88% male composition substantially restricts generalizability to women, in whom sex‐specific differences in the Q‐angle, ligamentous laxity, and lower limb biomechanics are well‐documented (Zumwalt 2023). There is no basis for assuming equivalent responses in predominantly female samples; future trials should set explicit sex‐balance targets.
Dynamic alignment was assessed using 2D smartphone videos rather than 3D motion capture or electromyography. Therefore, the proposed mechanistic explanations in Section 4.2 are untested hypotheses and not the findings of this study. The 4‐week observation window precludes conclusions about the durability, recurrence, or maintenance of biomechanical correction. The success of assessor blinding was not formally evaluated.
The absence of a taping‐only or no‐taping experimental arm prevented the isolation of ALDT's independent contribution. Any inference about taping as the key structural agent is indirect and must be treated as a hypothesis. The absence of a fully combined TJM + NMT + ALDT arm is an equally important design gap; whether integrating all three components would confer incremental benefits remains untested and is identified as a research priority.
Adequately powered multicenter RCTs with sex‐balanced sampling, a taping‐isolated comparator arm, objective kinematic and electromyographic assessments, formal blinding verification, and at least 6 months of follow‐up are needed before these findings can inform practice recommendations.
6. Conclusion
Across both phases of this investigation, pronated foot posture was associated with substantially impaired patellofemoral function and more severe dynamic knee valgus in adults with AKP, although causal inference is precluded by the cross‐sectional Phase 1 design. Both foot‐targeted regimens produced considerably greater short‐term improvements than routine physiotherapy, and MT + ALDT achieved a stronger profile on pain and functional recovery. The biomechanical equivalence of both experimental arms on FPI‐6 and DVI is consistent with a contribution from ALDT to structural correction, but the design cannot isolate that contribution, and a fully combined TJM + NMT + ALDT arm was not evaluated. The findings are preliminary, with a small sample size, single center, 4‐week follow‐up, and predominantly male cohort, and should not inform routine practice change in isolation.
For practitioners: Based on these preliminary findings, assessing foot posture and incorporating foot‐targeted interventions alongside knee‐focused care may be worth considering in AKP presentations with confirmed pronation, pending confirmation in larger trials. ALDT was common to both regimens and outperformed routine care; its independent contribution cannot be quantified from this design, and a taping‐isolated comparator arm should be incorporated in future trials. For researchers, the equivalence of both experimental arms on biomechanical endpoints, the inactivity of the control group, and the unresolved role of ALDT define a productive research agenda for kinetic‐chain rehabilitation trials in AKP.
6.1. Plain Language Summary
Anterior knee pain is one of the most common reasons adults attend physiotherapy, yet many patients find that standard treatment provides only limited or temporary relief. Flat feet, where the arch rolls inward underweight, have long been proposed as a contributing factor that is frequently overlooked in routine assessments.
In the first part of this study, we compared 30 adults with AKP and flat feet to 20 without flat feet and found that those with flat feet had markedly worse knee function and more pronounced inward knee movement during squatting. In the randomized trial, 31 participants with AKP and flat feet were allocated to one of three 4‐week programs: manual mobilization of the foot arch with supportive taping, a progressive balance and neuromuscular exercise program with the same taping, or standard knee physiotherapy without any foot component. Both foot‐targeted groups improved substantially more than the standard care group.
These are early, preliminary results from a small, single‐center study with a 4‐week follow‐up period and a predominantly male sample. They cannot be used to change clinical practice in isolation, and larger and longer trials in more diverse patient populations are needed before firm recommendations can be made regarding their use. The results suggest that assessing foot posture in patients presenting with anterior knee pain may be worthwhile, as some patients may have a foot‐related biomechanical contributor that is amenable to targeted intervention alongside standard knee rehabilitation.
6.2. Implications for Physiotherapy Practice
Foot posture screening in AKP assessment. FPI‐6 and DVI each take less than 5 min at the bedside with no specialist equipment. In this preliminary single‐center investigation, elevated scores on both measures were associated with meaningful between‐group functional differences in Phase 1 and a differential treatment response in Phase 2. The formal integration of these tools into standard AKP pathways cannot be recommended based on this evidence alone; confirmation in larger multicenter trials with diverse populations is required. Clinicians using these tools should interpret the findings as one element of a broader assessment.
Augmented Low‐Dye taping. ALDT was present in both experimental arms, both of which outperformed the control and achieved equivalent biomechanical correction. This pattern is consistent with a clinically meaningful contribution from taping to structural change, but the design included no taping‐only or no‐taping comparator; therefore, ALDT's independent contribution cannot be quantified. Characterizing ALDT as ‘indispensable’ or the ‘primary driver’ of biomechanical correction is not supported by this design; a dedicated taping‐comparator arm is needed to resolve this question.
Manual therapy with ALDT. MT + ALDT produced the best outcomes in this trial. These results are hypothesis‐generating and not confirmatory, with a small sample size, single center, and short follow‐up. Where manual therapy competency is available and foot pronation appears biomechanically relevant, MT + ALDT may be considered by clinicians with relevant manual therapy competency when foot pronation has been confirmed and biomechanical assessment suggests a distal contributor. Adoption as a standard recommendation for this population should await confirmatory evidence from larger, well‐powered multicenter trials.
NMT with ALDT as an alternative. NMT + ALDT produced biomechanical outcomes statistically equivalent to MT + ALDT in this sample, with slightly smaller pain reductions. In settings where specialist manual therapy is unavailable, NMT + ALDT may represent a pragmatic alternative; however, this should not be treated as established guidance pending independent, adequately powered replication.
Author Contributions
Contributed to and critically revised the manuscript. All authors have approved the final version for submission.
Funding
The authors have nothing to report.
Ethics Statement
Approved by the Mediclinic Middle East Clinical Research Ethics Committee (MCME.CR.310.MNOO.2023) on 8 June 2023.
Consent
Written informed consent was obtained from each participant prior to enrollment.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figure S1: Within‐group Cohen's d effect sizes—Pre to post intervention.
Figure S2: ANCOVA‐adjusted post‐intervention means and effect sizes (η 2).
Table S1: CONSORT 2025 checklist of information to include when reporting a randomised trial (Phase 2).
Acknowledgments
The authors gratefully acknowledge Mediclinic Al Noor Hospital, Abu Dhabi, UAE, for providing the clinical infrastructure and institutional support that facilitated the conduct of this study. We sincerely appreciate the cooperation of the physiotherapy department and allied clinical staff in enabling patient access and data collection.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- Barton, C. J. , Levinger P., Crossley K. M., Webster K. E., and Menz H. B.. 2011. “Relationships Between the Foot Posture Index and Foot Kinematics During Gait in Individuals With and Without Patellofemoral Pain Syndrome.” Journal of Foot and Ankle Research 4, no. 1: 1–8. 10.1186/1757-1146-4-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Buldt, A. K. , Murley G. S., Butterworth P., Levinger P., Menz H. B., and Landorf K. B.. 2013. “The Relationship Between Foot Posture and Lower Limb Kinematics During Walking: A Systematic Review.” Gait & Posture 38, no. 3: 363–372. 10.1016/j.gaitpost.2013.01.010. [DOI] [PubMed] [Google Scholar]
- Crossley, K. M. , Bennell K. L., Cowan S. M., and Green S.. 2004. “Analysis of Outcome Measures for Persons With Patellofemoral Pain: Which Are Reliable and Valid?” Archives of Physical Medicine and Rehabilitation 85, no. 5: 815–822. 10.1016/s0003-9993(03)00613-0. [DOI] [PubMed] [Google Scholar]
- Da Cunha, R. A. , Costa L. O. P., Hespanhol L. C., Pires R. S., Kujala U. M., and Lopes A. D.. 2013. “Translation, Cross‐Cultural Adaptation, and Clinimetric Testing of Instruments Used to Assess Patients With Patellofemoral Pain Syndrome in the Brazilian Population.” Journal of Orthopaedic & Sports Physical Therapy 43, no. 5: 332–339. 10.2519/jospt.2013.4228. [DOI] [PubMed] [Google Scholar]
- Dos Anjos Rabelo, N. D. , Lima B., dos Reis A. C., et al. 2014. “Neuromuscular Training and Muscle Strengthening in Patients With Patellofemoral Pain Syndrome: A Protocol of Randomized Controlled Trial.” BMC Musculoskeletal Disorders 15, no. 1: 1–10. 10.1186/1471-2474-15-157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Franettovich, M. , Chapman A. R., Blanch P., and Vicenzino B.. 2010. “Augmented Low‐Dye Tape Alters Foot Mobility and Neuromotor Control of Gait in Individuals With and Without Exercise Related Leg Pain.” Journal of Foot and Ankle Research 3: 1–9. 10.1186/1757-1146-3-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Haddad, B. I. , Hamdan M., Isleem U., et al. 2021. “Validation of the Cultural Adaptation of the Kujala Score in Arabic.” Journal of Orthopaedic Surgery and Research 16, no. 1: 1–5. 10.1186/s13018-021-02489-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hewett, T. E. , Lindenfeld T. N., Riccobene J. V., and Noyes F. R.. 1999. “The Effect of Neuromuscular Training on the Incidence of Knee Injury in Female Athletes.” American Journal of Sports Medicine 27, no. 6: 699–706. 10.1177/03635465990270060301. [DOI] [PubMed] [Google Scholar]
- Hübscher, M. , Zech A., Pfeifer K., Hänsel F., Vogt L., and Banzer W.. 2010. “Neuromuscular Training for Sports Injury Prevention: A Systematic Review.” Medicine & Science in Sports & Exercise 42, no. 3: 413–421. 10.1249/mss.0b013e3181b88d37. [DOI] [PubMed] [Google Scholar]
- Kim, H. J. , Cho J., and Lee S.. 2022. “Talonavicular Joint Mobilization and Foot Core Strengthening in Patellofemoral Pain Syndrome: A Single‐Blind, Three‐Armed Randomized Controlled Trial.” BMC Musculoskeletal Disorders 23, no. 1: 150. 10.1186/s12891-022-05099-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kumar, A. A. U. , Ramalingam V., Moussallem C. D., and Gopal K.. 2024. “Neuromuscular Training, Manual Therapy, and Augmented Low‐Dye Taping for Pronated Foot Correction in Anterior Knee Pain: RCT Study Protocol.” Fizjoterapia Polska 24, no. 5: 68–75. 10.56984/8ZG020CZM0A. [DOI] [Google Scholar]
- Myer, G. D. , Ford K. R., Palumbo O. P., and Hewett T. E.. 2005. “Neuromuscular Training Improves Performance and Lower‐Extremity Biomechanics in Female Athletes.” Journal of Strength & Conditioning Research 19, no. 1: 51–60. 10.1519/13643.1. [DOI] [PubMed] [Google Scholar]
- Neal, B. S. , Griffiths I. B., Dowling G. J., et al. 2014. “Foot Posture as a Risk Factor for Lower Limb Overuse Injury: A Systematic Review and Meta‐Analysis.” Journal of Foot and Ankle Research 7: 1–13. 10.1186/s13047-014-0055-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Pagé, M. G. , Katz J., Stinson J., Isaac L., Martin‐Pichora A. L., and Campbell F.. 2012. “Validation of the Numerical Rating Scale for Pain Intensity and Unpleasantness in Pediatric Acute Postoperative Pain: Sensitivity to Change over Time.” Journal of Pain 13, no. 4: 359–369. 10.1016/j.jpain.2011.12.010. [DOI] [PubMed] [Google Scholar]
- Powers, C. M. , Witvrouw E., Davis I. S., and Crossley K. M.. 2017. “Evidence‐Based Framework for a Pathomechanical Model of Patellofemoral Pain: 2017 Patellofemoral Pain Consensus Statement From the 4th International Patellofemoral Pain Research Retreat, Manchester, UK: Part 3.” British Journal of Sports Medicine 51, no. 24: 1713–1723. 10.1136/bjsports-2017-098717. [DOI] [PubMed] [Google Scholar]
- Praveenkumar, R. , Kamalakannan M., Hariharan J., Rajendra Prasad P. B., Kumar R., and Husna H.. 2025. “Effectiveness of Jump Training Along With Core Neuromuscular Training in Football Players on Improving Their Anterior Knee Pain, Functional Performance, and Knee Valgus Using Dartfish Software.” Fizjoterapia Polska 25, no. 1: 19–26. 10.56984/8ZG007DMFP7. [DOI] [Google Scholar]
- Redmond, A. C. , Crosbie J., and Ouvrier R. A.. 2006. “Development and Validation of a Novel Rating System for Scoring Standing Foot Posture: The Foot Posture Index.” Clinical biomechanics 21, no. 1: 89–98. 10.1016/j.clinbiomech.2005.08.002. [DOI] [PubMed] [Google Scholar]
- Scholtes, S. A. , and Salsich G. B.. 2017. “A Dynamic Valgus Index That Combines Hip and Knee Angles: Assessment of Utility in Females With Patellofemoral Pain.” International Journal of Sports Physical Therapy 12, no. 3: 333–340. PMCID: PMC5455183. 10.26603/ijspt20170333. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Silva, N. C. , de Castro Silva M., Guimarães M. G., de Oliveria Nascimento M. B., and Felicio L. R.. 2019. “Effects of Neuromuscular Training and Strengthening of Trunk and Lower Limbs Muscles in Women With Patellofemoral Pain: A Protocol of Randomized Controlled Clinical Trial, Blinded.” Trials 20: 1–11. 10.1186/s13063-019-3650-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Smith, B. E. , Selfe J., Thacker D., et al. 2018. “Incidence and Prevalence of Patellofemoral Pain: A Systematic Review and Meta‐Analysis.” PLoS One 13, no. 1: e0190892. 10.1371/journal.pone.0190892. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Srinivasan, P. , Jothilingam M., Anbarason J. P. A., Siddiq A. R. A., and Ramesh B.. 2024. “Effect of Corrective Exercise Strategies on Tapping Skill Among University Level Kho Kho Players With Overpronated Foot.” Central European Journal of Sport Sciences and Medicine 48: 37–46. 10.18276/cej.2024.4-04. [DOI] [Google Scholar]
- Suzuki, H. , Aono S., Inoue S., et al. 2020. “Clinically Significant Changes in Pain Along the Pain Intensity Numerical Rating Scale in Patients With Chronic Low Back Pain.” PLoS One 15, no. 3: e0229228. 10.1371/journal.pone.0229228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Udhaya Kumar, A. A. , Ramalingam V., Moussallem C. D., and Gopal K.. 2025. “Biomechanical Impact of Foot Pronation on Anterior Knee Pain—Case Controlled Study.” Fizicna Reabilitacia ta Rekreacijno‐Ozdorovci Tehnologii 10, no. 1. 10.15391/prrht.2025-10(1).05. [DOI] [Google Scholar]
- Vicenzino, B. , McPoil T., and Buckland S.. 2007. “Plantar Foot Pressures After the Augmented Low Dye Taping Technique.” Journal of Athletic Training 42, no. 3: 374–380. PMCID: PMC1978458; PMID: 180599. [PMC free article] [PubMed] [Google Scholar]
- Willy, R. W. , Hoglund L. T., Barton C. J., et al. 2019. “Patellofemoral Pain: Clinical Practice Guidelines Linked to the International Classification of Functioning, Disability and Health From the Academy of Orthopaedic Physical Therapy of the American Physical Therapy Association.” Journal of Orthopaedic & Sports Physical Therapy 49, no. 9: CPG1–95. https://www.jospt.org/doi/10.2519/jospt.2019.0302. [Google Scholar]
- Wyndow, N. , Collins N., Vicenzino B., Tucker K., and Crossley K.. 2016. “Is There a Biomechanical Link Between Patellofemoral Pain and Osteoarthritis? A Narrative Review.” Sports Medicine 46, no. 12: 1797–1808. 10.1007/s40279-016-0545-6. [DOI] [PubMed] [Google Scholar]
- Yang, C. , Best T. M., Liu H., and Yu B.. 2022. “Knee Biomechanical Factors Associated With Patellofemoral Pain in Recreational Runners.” Knee 35: 87–97. 10.1016/j.knee.2022.02.007. [DOI] [PubMed] [Google Scholar]
- Zumwalt, M. 2023. “Prevention and Management of Common Musculoskeletal Injuries in the Adult Female Athlete.” In The Active Female: Health Issues Throughout the Lifespan, 243–258. Springer. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1: Within‐group Cohen's d effect sizes—Pre to post intervention.
Figure S2: ANCOVA‐adjusted post‐intervention means and effect sizes (η 2).
Table S1: CONSORT 2025 checklist of information to include when reporting a randomised trial (Phase 2).
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.