Targeted Treatment Protocol in Patellofemoral Pain: Does Treatment Designed According to Subgroups Improve Clinical Outcomes in Patients Unresponsive to Multimodal Treatment?

Abstract

Background:

Targeted intervention for subgroups is a promising approach for the management of patellofemoral pain.

Hypothesis:

Treatment designed according to subgroups will improve clinical outcomes in patients unresponsive to multimodal treatment.

Study Design:

Prospective crossover intervention.

Level of Evidence:

Level 3.

Methods:

Patients with patellofemoral pain (PFP; n = 61; mean age, 27 ± 9 years) were enrolled. Patients with PFP received standard multimodal treatment 3 times a week for 6 weeks. Patients not responding to multimodal treatment were then classified into 1 of 3 subgroups (strong, weak and tight, and weak and pronated foot) using 6 simple clinical tests. They were subsequently administered 6 further weeks of targeted intervention, designed according to subgroup characteristics. Visual analog scale (VAS), perception of recovery scale (PRS), 5-Level European Quality 5 Dimensions (EQ-5D-5L), and self-reported version of the Leeds Assessment of Neuropathic Symptoms and Signs scale (S-LANSS) were used to assess pain, knee function, and quality of life before and after the interventions.

Results:

In total, 34% (n = 21) of patients demonstrated recovery after multimodal treatment. However, over 70% (n = 29/40) of nonresponders demonstrated recovery after targeted treatment. The VAS, PRS, S-LANSS, and EQ-5D-5L scores improved significantly after targeted intervention compared with after multimodal treatment (P < 0.001). The VAS score at rest was significantly lower in the “weak and pronated foot” and the “weak and tight” subgroups (P = 0.011 and P = 0.008, respectively). Posttreatment pain intensity on activity was significantly lower in the “strong” subgroup (P = 0.006).

Conclusion:

Targeted treatment designed according to subgroup characteristics improves clinical outcomes in patients unresponsive to multimodal treatment.

Clinical Relevance:

Targeted intervention could be easily implemented after 6 simple clinical assessment tests to subgroup patients into 1 of 3 subgroups (strong, weak and tight, and weak and pronated foot). Targeted interventions applied according to the characteristics of these subgroups have more beneficial treatment effects than a current multimodal treatment program.

Patellofemoral pain (PFP) is a chronic musculoskeletal problem that causes persistent anterior knee pain.^{3,5,13,14,24,25,32,37,45} Despite its widespread use in clinics, it is difficult to suggest that the current multimodal treatment approach leads to successful outcomes in the majority of patients with PFP, as studies have shown that only 46% of patients’ knees were pain-free at discharge.² This indicates that over half of the patients with PFP do not respond to treatment and may continue their lives with chronic anterior knee pain.

Identification of the factors leading to these low treatment success rates has consistently been a priority of previous International Patellofemoral Pain Research Retreats.^9,11,34,52 The most important factor that has emerged affecting the success of treatment is the variety of musculoskeletal and biomechanical differences between patients. The current multimodal treatment, therefore, may not affect the heterogeneous PFP patient population with the same efficiency.^26,28,43 Clinically subgrouping patients with PFP and delivering targeted treatments has been strongly recommended for future investigations of PFP treatment from the International Patellofemoral Pain Research Retreats.^11,34,52 An overview of PFP subgroups and the methods used to derive these subgroups has been published previously.^39,41 They exhibit different anthropometric and biomechanical characteristics and do not form a homogeneous group. There are 3 subgroups in the PFP population: strong, weak and tight, and weak and pronated foot.⁴⁰ The purpose of this study was to assess the clinical outcomes of targeted treatments designed according to the characteristics of these 3 subgroups of PFP patients.⁴⁰ The hypotheses were that the assessment and subgroup classification would be clinically feasible and that targeted treatments designed according to the characteristics of the 3 subgroups of PFP patients would show clinical benefits over and above a multimodal intervention.

Methods

Study Design

A prospective crossover intervention study design was used (Figure 1).

Figure 1.

Study flowchart.

Participants

Patients aged between 18 and 40 years attending a physical therapy outpatient clinic at a university hospital with a clinical diagnosis of PFP were approached for eligibility in this study. Eligibility criteria were based on previously defined PFP criteria.^6,40,47 Patients were excluded if they had any of the following: previous knee surgery, clinical evidence of ligamentous instability and/or internal derangement, a history of patellar subluxation or dislocation, joint effusion, true knee joint locking and/or giving way, bursitis, patellar or iliotibial tract tendinopathy, Osgood-Schlatter disease, Sinding-Larsen-Johansson syndrome, muscle tears or symptomatic knee plicae, another serious comorbidity that would preclude or affect compliance with the assessment, or were pregnant.

Subgroup Classification

Quadriceps and hip abductor muscle strength,^30,50 patellar glide test,^44,54 quadriceps length,⁵³ gastrocnemius length,⁵³ and foot posture index³⁶ assessments were performed to classify all consenting patients into 1 of 3 subgroups (strong, weak and tight, and weak and pronated foot) using the algorithm derived from the work by Selfe et al.⁴⁰

Intervention

Multimodal Treatment

The multimodal treatment program was designed based on the usual exercises and modalities used in local clinics.^19,20,31,49 All patients received a standard, supervised, 60-minute multimodal treatment 3 times per week for 6 weeks. Table 1 shows the details of the multimodal rehabilitation program.

Table 1.

Multimodal treatment program

Modality	Application Type
Thermotherapy	Cold packs, 20 min
Transcutaneous electrical neural stimulation	Conventional mode, 20 min 50-100 Hz, 20-60 pulse/s
Therapeutic ultrasound	1 W/cm2, 5 min around knee joint
Hamstring/tensor fascia lata/iliotibial band stretching	30 s/5 rep
Isometric quadriceps strengthening	10 rep × 3 set
Isometric hip adductor strengthening	10 rep × 3 set
OKC knee extension exercise	3 sets of patients’ 8-10 RM, in painless ROM
OKC hip adductor exercise	Side-lying/3 sets of patients’ 8-10 RM
Home-based exercise program a

OKC, open kinetic chain; rep, repetition; RM, repetition maximum; ROM, range of motion.

^aHome-based exercise program included the same applications except transcutaneous electrical neural stimulation, neuromuscular electrical stimulation, and therapeutic ultrasound.

Targeted Treatment

Patients who did not respond to multimodal treatment were assigned to 1 of the 3 treatment groups: strong, weak and tight, and weak and pronated foot. They then followed a further 6-week targeted intervention program administered for 45 minutes, 3 times per week. The targeted treatment program was designed according to the key deficits identified in each patient by the subgrouping clinical assessment tests. The patients in the “strong” subgroup had no muscle strength deficit; therefore, the intervention program for this subgroup was targeted at improving neuromuscular control and coordination ability using proprioceptive exercises, such as progressive balance exercises, and knee braces,^46,47 which have been shown to offer improvements in movement control in patients with PFP,⁴² reductions in patellofemoral reaction forces,⁴⁴ and reduction of pain at 6 and 12 months during a PFP rehabilitation program.⁴⁸ In the “weak and tight” subgroup, the exercise program consisted of closed kinetic chain (CKC) muscle strengthening and stretching and weight management advice, as a larger body mass index was identified as a potentially relevant clinical feature in this subgroup.⁴⁰ In the “weak and pronated foot” subgroup, muscle weakness and abnormal foot alignment were identified as the key factors. Therefore, the intervention program included CKC strengthening exercises and foot orthoses.^4,23 Table 2 shows the details of each of the specific targeted intervention programs.

Table 2.

Targeted treatment program

Modality	Application Type
Strong Subgroup
Progressive balance/proprioception exercises	Standing on 1 leg on wobble board 3 sets of 1 min exercise each leg 1-3 sets per session depending on pain Progression a : Eyes closed, bouncing ball against wall, bouncing ball against wall on an unstable surface
Patellar bracing b	Patient was asked to put on knee brace during ADLs
Activity modification	Activity reduction to fit within envelope of function locally determined and negotiated with individual patient
Weak and tight subgroup
CKC strengthening exercises	Plié/lunge/single-limb squat Pain-free ROM 10 reps per set/1-3 sets depending on pain
Gastrocnemius and quadriceps stretching exercises	30 seconds static stretch × 3 reps × 1 per day
Weight management strategies	Locally determined and negotiated with individual patient
Weak and pronated foot subgroup
CKC strengthening exercises	Plié/lunge/single-limb squat Pain-free ROM 10 reps per set/1-3 sets depending on pain
Foot orthoses	Custom-made insole supporting medial longitudinal arch of foot c
Activity modification	Improve activity levels locally determined and negotiated with individual patient

ADLs, activities of daily living; CKC, closed kinetic chain; rep, repetition; ROM, range of motion.

^aProgression timing in balance exercise was decided by clinician based on patient pain-free achievement.

^bOff-the-shelf knee support with patellar pad was used (Orthocare material: 5-mm neoprene/styrene butadiene rubber/nylon jersey/pk). Brace size was selected by clinician according to patient comfort and patellar coherence (S/M/L/XL sizes were used).

^cCustom-made insoles were tailored individually based on static and dynamic examination of load distribution on foot, using CAT-CAM free step V.1.3.30.

Outcome Measures

Pain during activity measured using a visual analog scale (VAS) was the primary outcome measure of this study.¹⁸ Activity was specified by patients.

The perception of recovery scale (PRS) was measured using a 7-point Likert-type scale ranging from “completely recovered” to “worse than ever.” Patients were classified as “recovered” if they rated themselves as “completely recovered” or “strongly recovered.” Patients rating themselves in 1 of the other 5 categories from “slightly recovered” to “worse than ever” were categorized as “not recovered.”³⁵

The 5-Level European Quality 5 Dimensions (EQ-5D-5L) was used as a self-reported generic measure of health and quality of life. Patients rated their overall health on the day of the interview on a 0 to 100 hash-marked, vertical VAS (EQ-5D-5L-VAS). A higher EQ-5D-5L-VAS score indicates better health status.²¹

Neuropathic pain was measured using the Self-Administered Leeds Assessment of Neuropathic Symptoms and Signs (S-LANSS) questionnaire. The S-LANSS comprises a 5-item questionnaire regarding pain symptoms and 2 items for clinical signs involving self-administered sensory tests for the presence of allodynia and decreased sensation to pinprick. This was used to discriminate the small number of patients who may have neuropathic knee pain from those with nociceptive pain.³⁸ The possible scores range from 0 to 24, with a score of 12 or greater considered to be suggestive of neuropathic pain.²⁷ Finally, a single leg hop test was used to determine functional performance.¹ Distance was measured from toe to heel and the mean score of 3 repetitions was recorded.

Data Analysis

A sample size calculation was performed based on the minimal detectable change on the pain VAS. Data from a previous study indicate that the mean VAS score in patients with PFP was 4.3 ± 1 cm,⁸ with 30% of the maximum score of the VAS-pain considered to be the detectable change; the sample size for each treatment subgroup to achieve a 90% power at the 0.05 level of significance was determined to be 8 patients. Data were not normally distributed when analyzed with the Kolmogorov-Smirnov test. Consequently, nonparametric tests were indicated. Therefore, the Wilcoxon signed-rank test was used to compare pre- and posttreatment outcomes with an alpha value of 0.05. In addition, the mean of rank scores, SEs, and Z scores were reported, along with descriptive statistics to describe the general features of the patients. All statistical analysis was conducted using SPSS (Version 21.0; IBM).

Results

Of the 128 patients who were screened, 95 were included in the present study. Of these, 61 patients completed the multimodal treatment (Figure 1, Table 3). Twenty-one patients (34%) demonstrated recovery after multimodal treatment (phase 1) and were discharged. Forty patients (64%) unresponsive to multimodal treatment were administered 6 further weeks of targeted intervention, designed according to subgroup characteristics (phase 2). Twenty-nine (72.5%) patients demonstrated recovery after targeted intervention (phase 2), and 11 (27.5%) patients did not respond to either of the treatment approaches (Table 4).

Table 3.

Demographic data of patients who participated in the study

Patients (n = 61)	Mean	SD
Age, y	27	9
Height, cm	170	8
Weight, kg	65	13
Time since symptoms started, mo	24	28
Body mass index, kg/m2	22.5	3

Table 4.

Perception of recovery after treatments^a

	Phase 1, Multimodal Treatment (n = 61)				Phase 2, Targeted Treatment (n = 40)
Perception of Recovery Scale	Overall, % (n)	Weak and Tight, % (n)	Weak and Pronated, % (n)	Strong, % (n)	Overall, % (n)	Weak and Tight, % (n)	Weak and Pronated, % (n)	Strong, % (n)
Fully improved	11 (7)	16 (4)	—	9 (2)	7.5 (3)	8 (1)	—	11 (2)
Great improvement	23 (14)	36 (9)	29 (4)	9 (2)	65 (26)	92 (11)	80 (8)	39 (7)
Some improvement	48 (29)	36 (9)	57 (8)	55 (12)	17.5 (7)	—	20 (2)	28 (5)
No change	16 (10)	12 (3)	14 (2)	18 (4)	10 (4)	—	—	22 (4)
A little worse	2 (1)	—	—	9 (2)	0 (0)	—	—	—

^aDashes indicate no patient.

Pain intensity (VAS) at rest and during activity and PRS were significantly improved after targeted intervention (P < 0.001) (Table 5). S-LANSS, EQ-5D-5L, and EQ5D-5L-VAS scores were significantly improved after targeted intervention compared with pretargeted treatment scores (P = 0.001, P < 0.001, and P = 0.02, respectively) (Table 5).

Table 5.

Outcome measure differences with targeted treatment

	Before Targeted Treatment		After Targeted Treatment
Outcome Measures (n = 40)	Median	Min-Max	Median	Min-Max	Z	P
Perception of recovery	3	3-5	2	1-4	−5.034	<0.001*
VAS activity, cm	4.4	0.1-8.8	1.8	0-7.5	−4.075	<0.001*
VAS rest, cm	1.7	0-7.4	0.5	0-7.0	−3.599	<0.001*
S-LANSS	5	0-16	0	0-24	−3.449	0.001*
EQ-5D-5L	7	5-10	6	5-11	−3.704	<0.001*
EQ-5D-5L-VAS	80	30-95	85	50-100	−2.322	0.020*
Quadriceps muscle strength, N·m/kg	1,1	0.5-2.1	1,2	0.6-2.3	−3.644	<0.001*
Hip abductor muscle strength, N·m/kg	1,3	0.7 - 2,6	1,3	0.6-1.9	−1.456	0.145
Patellar mobility test, mm	12	7-25	11	2-18	−2.062	0.039*
Foot posture index	6	0-11	6	0-12	−0.372	0.710
Quadriceps length, deg	142.7	115-156	145.2	128-155	−2.150	0.032
Gastrocnemius length, deg	19.6	8-40	20.5	12.3-40	−1.358	0.174
Jump, cm	90.2	30-180	91	38-179	−1.472	0.141

EQ-5D-5L: 5-Level European Quality 5 Dimensions; Max, maximum; Min, minimum; S-LANSS, Self-Report of the Leeds Assessment of Neuropathic Symptoms and Signs; VAS, visual analog scale.

^*Statistically significant (P < 0.05).

Within the 3 subgroups, the findings showed that the PRS score was significantly improved after targeted treatment compared with pretargeted treatment levels in the “strong,” “weak and tight,” and “weak and pronated foot” subgroups (P = 0.005, P = 0.001, and P = 0.004, respectively).

VAS pain intensity at rest was also significantly lower after targeted intervention in the “weak and pronated foot” and “weak and tight” subgroups (P = 0.011 and P = 0.008, respectively); however within the “strong” subgroup, no change was seen between pre- and posttreatment (P = 0.245) (Table 6). However, pain intensity during activity was significantly lower after treatment in the “strong” (P = 0.006), “weak and pronated foot,” and “weak and tight” subgroups, although the latter 2 reductions were not statistically significant (P = 0.059 and P = 0.06, respectively) (Table 6).

Table 6.

Differences in subgroups before and after targeted treatment (n = 40)

		Before Treatment		After Treatment
		Median	Min-Max	Median	Min-Max	Z	P
VAS in activity	Weak and pronated group (n = 10)	5.3	0.5-8.8	2.7	0.2-6.6	−1.886	0.059
	Weak and tight group (n = 12)	3.7	0.4-7.7	3	0-6.5	−1.883	0.060
	Strong group (n = 18)	5.0	0.1-8.2	2.0	0-7.5	−2.741	0.006*
VAS at rest	Weak and pronated group (n = 10)	3.9	0-7.1	0.8	0-3.4	−2.547	0.011*
	Weak and tight group (n = 12)	1.0	0-3.5	0.68	0-1.6	−2.667	0.008*
	Strong group (n = 18)	1.8	0-7.4	0.7	0-7	−1.161	0.245
PRS	Weak and pronated group(n = 10)	3	3-4	2	2-3	−2.887	0.004*
	Weak and tight group (n = 12)	3	3-4	2	1-2	−3.213	0.001*
	Strong group (n = 18)	3	3-5	2.5	1-4	−2.830	0.005*

Max, maximum; Min, minimum; PRS, perception of recovery scale; VAS, visual analog scale.

^*Statistically significant (P < 0.05).

Other measures, including quadriceps length test, S-LANSS, EQ-5D-5L, and EQ-5D-VAS, were significantly improved in the “weak and tight” subgroup. S-LANSS, EQ-5D-5L, and patellar mobility were significantly improved in the “weak and pronated foot” subgroup. In the “strong” group, only gastrocnemius length was significantly different between pre- and posttargeted treatment (P = 0.03). Results for outcome measures are shown in Table 7.

Table 7.

Outcome measures in subgroups before and after targeted treatment

	Weak and Tight Subgroup (n = 12)				Weak and Pronated Subgroup (n = 10)				Strong Subgroup (n = 18)
	Before Median (Min-Max)	After Median (Min-Max)	Z	P	Before Median (Min-Max)	After Median (Min-Max)	Z	P	Before Median (Min-Max)	After Median (Min-Max)	Z	P
S-LANSS	5 (0-11)	0 (0-6)	−2.716	0.007*	6 (0-11)	0 (0-10)	−2.410	0.016*	5 (0-169)	1.5 (0-24)	−0.947	0.344
EQ-5D-5L	7.5 (5-10)	6 (5-9)	−2.556	0.011*	9 (6-9)	6 (5-11)	−2.203	0.028*	6 (5-10)	6 (5-10)	−1.613	0.107
EQ-5D-5L-VAS	80 (50- 90)	90 (50-95)	−2.034	0.042*	80 (50- 90)	80 (50-100)	−1.027	0.305	82.5 (30-95)	82.5 (55-100)	−1.444	0.149
Quadriceps muscle strength, N·m/kg	0.84 (0.5-1.3)	1.05 (0.6-1.4)	−3.061	0.002*	1.06 (0.6-2.1)	1.3 (0.7-1.6)	−1.887	0.059	1.2 (0.9-1.6)	1.2 (0.9-2.2)	−0.893	0.372
Hip abductor muscle strength, N·m/kg	0.9 (0.7-1.4)	1.1 (0.6-1.6)	−1.844	0.065	1.1 (0.7-1.6)	1.2 (0.9-1.6)	−0.593	0.553	1.4 (0.9-2.6)	1.5 (1-1.9)	−0.259	0.796
Patellar mobility test, mm	10 (7- 15)	10 (8-15)	−0.103	0.918	15 (11- 22)	12 (2-18)	−2.325	0.020*	12 (8-25)	11 (7-17)	−0.803	0,422
Foot posture index	5 (0-9)	5.5 (2-10)	−1.725	0.084	7.5 (4-11)	7.5 (2-12)	−0.679	0.497	5 (0-11)	6 (0-12)	−0.178	0.859
Quadriceps length, deg	137 (115-149)	140 (128-152)	−2.134	0.033*	140 (118-152)	146 (130-155)	−1.481	0.139	147 (117-155)	148 (128-155)	−0.071	0.943
Gastrocnemius length, deg	18.2 (10-26)	17.4 (12.6-27)	−1.295	0.195	21.3 (10-40)	17.3 (12.6-34)	−1.244	0.214	19.6 (8-27)	21.5 (12.3-40)	−2.120	0.034*
Jump test, cm	79.1 (30-115)	81 (38-115)	−1.718	0.286	85.4 (40-149)	84.2 (65-154)	−1.718	0.086	104.5 (49.3-180.6)	107.2 (57.3-179.3)	−0.305	0.760

EQ-5D-5L, 5-Level European Quality 5 Dimensions; Max, maximum; Min, minimum; S-LANSS, Self-Report of the Leeds Assessment of Neuropathic Symptoms and Signs; VAS, visual analog scale.

^*Statistically significant (P < 0.05).

Discussion

The results of this study suggest that the targeted subgroups and the algorithm used to classify PFP patients as “strong,” “weak and tight,” or “weak and pronated foot”⁴⁰ are valid and able to be implemented clinically. The findings from this study were in agreement with previous work¹² that reported differential response patterns in outcomes at 12 months in their subgroups. This suggests that targeted interventions based on subgroups provide an important development in the treatment strategy for patients with PFP.^34,52

The “strong” subgroup demonstrated a poor response to multimodal treatment, but a significant improvement was observed after targeted treatment. This finding is consistent with Yosmaoğlu et al,⁵⁵ Greuel et al,¹⁷ and Gallina et al,¹⁶ who reported results confirming that motor control of the quadriceps is problematic in some patients with PFP. One explanation for this is improved neuromuscular control in patients classified as “strong.” Since these patients already demonstrated relatively high quadriceps muscle torque, targeted intervention was delivered focusing on progressive development of motor control on unstable surfaces instead of conventional muscle strength exercises. Given that quadriceps strength did not change as a result of the targeted intervention, these progressive balance exercises and patellar bracing improved motor control and stability.⁴² In addition, bracing may reduce patellofemoral forces during activities of daily living and sporting tasks⁴⁴ and improvements within rehabilitation protocols.⁴⁸ This was reflected in the improvement in the other pain-related parameters. However, since the average pretreatment VAS pain level at rest in this subgroup was already low, a decrease from 1.8 to 0.7 has minimal clinical relevance.

Clinically, the “weak and tight” subgroup appeared to be the most responsive group to treatment overall, with a relatively even split of 52% responding to multimodal treatment, and all of the remaining patients responding to targeted intervention. This finding was not surprising as multimodal treatment routinely includes strengthening and stretching exercises. However, closer analysis of the outcomes in the “weak and tight” subgroup suggest that although patients’ perception of recovery improved, the VAS activity pain intensity was not significantly decreased after targeted treatment in this subgroup. Considering muscle weakness is the main issue in this subgroup, the probable cause of this unexpected finding is persistent inability to compensate patellofemoral loads, especially during relatively high-level activities of daily life such as ascending/descending stairs even after the targeted treatment. Targeted intervention consisting of functional strengthening may still be insufficient for high-level activities of daily living, which demand considerable muscular activity,⁴⁹ although it caused approximately a 30% increase in muscle torque and a significant improvement in perception of recovery in this subgroup.

Findings from the “weak and pronated foot” subgroup suggest that targeted treatment, including foot orthoses and pain-free strengthening exercises, was also successful in terms of perception of recovery and VAS pain on rest, although the same improvement was not observed in VAS pain during activity. One explanation for this could be the indirect corrective effect of the foot orthoses^15,22 on the knee, as the patients showed no improvement in strength after targeted treatment. Moreover, optimum correction is very difficult to determine during the intervention of foot orthoses. Special single physical therapy interventions or combined interventions for patellar taping, mobilization, or manual therapy may have beneficial effects on pain-related functional symptoms in PFP.^10,29,33 However, the therapeutic effects of these applications remain limited because PFP patients exhibit a wide variety of structural features and biopsychosocial differences. The biomechanical and anthropometric characteristics of patients were not similar. Foot pronation, for example, was noticeably high in some patients, while some had neutral foot alignment. Similarly, quadriceps muscle strength, which is a predisposing factor or the most common symptom in previous studies,^7,54 has been high in some patients, with the remainder having considerable muscle weakness. Therefore, specific applications such as foot orthoses, knee braces, tape, and even exercises may not be required by every patient.

The functional hop test is often used in clinics to measure functional capability.⁵¹ Considering that there was no increase in quadriceps muscle strength in the “weak and pronated foot” and “strong” subgroups, an improvement in hop test scores was not expected.

Because of the methodological design of this study, patients received 6 weeks of multimodal treatment before 6 weeks of targeted treatment with no intervening washout period. This is a study limitation since the cumulative effects of the previous treatment (multimodal) were ignored. Therefore, the observed difference in some parameters could be the result of regression to the mean.

Conclusion

PFP patients are not a homogeneous group and have biomechanical and structural differences. Both the targeted interventions for patellofemoral pain syndrome (TIPPs) assessment and the subgroup classification algorithm are clinically feasible.