Effect of hip abductors and lateral rotators’ muscle strengthening on pain and functional outcome in adult patients with patellofemoral pain: A systematic review and meta-analysis

1.1Relationship between hip muscles (abductors and lateral rotators) strength and PFP

Hip muscles (abductors and lateral rotators) are essential for knee and pelvic stabilization during ambulation [5]. The hip abductors and lateral rotators act synergistically to eccentrically control the hip adduction and internal rotation movements [15, 16, 17]. The diminished strength of hip abductors and lateral rotator muscles may result in poor neuromuscular control during activities that require loading on the patellofemoral joint [5, 8, 15]. The weak hip abductors may cause excessive femoral adduction, thereby augmenting lateral forces (Knee Valgum) acting on the patella [16]. In contrast, weak hip lateral rotators result in unrestricted internal rotation of the femur that augments contact pressure between the lateral facet of the patella and lateral femoral condyle [16]. Hence, weak hip muscles (mainly abductors and lateral rotators) are an important aetiological factor for PFP [5, 17, 18, 19].

Many studies compared the effectiveness of hip muscles strength in patients with PFP to matched healthy controls [19, 20, 21, 22]. Ireland et al. reported eccentric muscle strength reduction of 26% in hip abductors and 36% in hip lateral rotators among females with PFP, while Souza and Powers found a reduction of 14% in hip abductors and 17% in hip lateral rotators eccentric muscle strength compared to healthy matched controls [19, 21]. Nevertheless, Piva et al. found no significant muscle strength differences for hip abductors and lateral rotators in patients with PFP compared to healthy age/gender-matched controls; however, Baldon et al. reported significantly reduced strength for eccentric hip abductors, but not for hip lateral rotators among females with PFP to healthy matched controls [20, 22].

The weak hip lateral rotators cause unrestricted internal rotation of the femur about the tibia, enhancing misalignment at the knee joint that in turn leads to a biomechanical imbalance between the hip extensors and lateral rotators that overload the retinaculum and subchondral bone and subsequently potentiate patellofemoral pain and knee dysfunction [19]. Nevertheless, Earl et al. argued that strong hip muscles (abductors and lateral rotators) reverse these effects over the knee joint [3]. Moreover, rotational malalignment and patellar instability are well documented, and weak hip lateral rotator muscles are identified as important contributors [23, 24, 25]. It is important to consider the biomechanical assistance provided by the hip lateral rotator group muscles to maintain the normal alignment of the patella [26].

Ireland et al. and Souza and Powers noted more weakness in hip lateral rotators than hip abductors in patients with PFP [21, 19]. Ferber et al. found that in patients with PFP, the three weeks of isolated hip abductors strengthening reduced patellofemoral pain and increased gait-related knee-joint stability [9]. Two recent randomized controlled trials found that isolated strengthening of hip abductors and lateral rotators effectively relieves pain and improvise knee function in females [4, 27]. The available evidence for PFP considered exercises to strengthen the hip muscles that reduce pain and enhance long-term knee function [3, 4, 6, 27, 28].

1.2Outcome measures of pain, knee function and health status in PFP

The available studies used self-reported Kujala Anterior Knee Pain Scale (AKPS), Visual Analogue Pain (VAS) scale, 11-point Numerical Pain Rating Scale (NPRS) and Pain Severity Scale (PSS) as an outcome measure to document patellofemoral pain in patients with PFP receiving therapeutic interventions [14, 29, 30].

The knee functions for patellofemoral pain were assessed using the Lower Extremity Functional Scale (LEFS), Tegner Activity Scale (TAS), Lysholm Knee Scoring Scale (LKSC)/Tegner Lysholm Knee Scoring Scale (TLKSS), Knee Outcome Survey-Activities of Daily Living Scale (KOS-ADL) and Functional Index Questionnaire (FIQ) [30, 31, 32, 33, 34].

Although the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) is tailored to examine the functional status of osteoarthritis, it is also used to measure health status for patients with patellofemoral pain (because patients with osteoarthritis often present with anterior knee pain, which is similar to patellofemoral pain) [29].

Since systematic reviews evaluating the effect of hip abductors and lateral rotator strengthening for patellofemoral pain, knee function and quality of life in patients with PFP are extremely limited, primarily this study systematically reviews and meta-analyzes the best evidence on the therapeutic value of strengthening hip abductors and lateral rotators muscles for the treatment of PFP. The presumptive hypothesis is that strengthening hip muscles stabilizes the patellofemoral joint, relieves pain and enhances knee functions.

2.1Justification of the systematic review approach

Systematic reviews and meta-analyses are important methodologies for the qualitative and quantitative synthesis of published evidence. Shreds of evidence presented in systematic reviews are key for continuous quality and safety improvements in evidence-based clinical practice and, therefore, useful for clinicians and healthcare policymakers. The present review study used Centre for Reviews and Dissemination (CRD) guidance for undertaking reviews in health interventions to assess the value of hip muscle strengthening as therapeutic interventions in patellofemoral pain and knee function in patients with PFP [35]. Additionally, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement, recommended in CRD’s guidance, is used in literature searches to minimize article selection and reporting bias [36]. This study is exempt from Institutional Review Board approval as it is a literature review. The review has been registered with the Open Science Framework with reference doi: 10.17605/OSF.IO/CWZ8V.

2.2Electronic bibliographic database searches

The controlled clinical trials (randomized and non-randomized), controlled comparative studies and cohort studies ((prospective and retrospective) published in the last 25 years (January 1994 to September 2019) in English language journals were performed across five electronic databases [Medical Literature Analysis and Retrieval System Online (Medline); Excerpta Medica Database (EMBASE); Clinical Index of Nursing and Allied Health Literature (CINAHL); Physiotherapy Evidence Database (PEDro) and The Cochrane Central Register of Controlled Trials (CENTRAL)]. Appropriate combinations of Medical Subject Headings (MeSH) or CINAHL headings with keywords (Table 1) using Boolean Operators (AND, OR and NOT) along with PICOS (target population, intervention, comparator, outcomes and study design) were used [37].

Additional electronic searches are done in the Meta Register of Controlled Trials (mRCT) via the Current Controlled Trials (CCT) database to locate ongoing RCTs with potentially relevant data useful for the present systematic review. The potentially relevant clinical controlled trials and cohort studies (otherwise not indexed in any of the five electronic bibliographic databases and mRCT), electronic searches were supplemented by searching unpublished papers from the OpenGrey (formerly SIGLE) database. The literature searches were additionally supplemented with manual bibliographic searches of relevant systematic reviews, editorials and thesis reports published by the digital libraries of the University of Manchester, University of Central Lancashire and Australian Digital Thesis programmes, including ProQuest. Authors of potentially relevant unpublished reports were contacted by e-mails seeking clarification of their respective studies with the possibility of inclusion in the present review.

2.3Study selection

The study selection was performed using thePRISMA flowchart (Fig. 1), where returned hits for each electronic bibliographic database were screened by two independent reviewers initially based on title and abstracts. The number of potentially relevant articles was noted, citations were imported into the Endnote citation manager (EndNote X7), and full-text articles were retrieved for further eligibility screening by the two independent reviewers. Studies were included based on the following criteria:

Figure 1.

PRISMA flowchart for articles search strategy, screening and eligibility evaluation.

2.4Inclusion criteria

2.5Exclusion criteria

A third senior reviewer was contacted to reach a consensus on any disagreement among the two reviewers regarding the inclusion or exclusion of an article.

2.6Critical appraisal of methodological quality

The McMaster Critical Review Form for Quantitative Studies was applied to examine the methodological quality of all selected studies for study’s objectives, literature survey, study design, sample population, intervention, outcome measures, results, significance, limitations, and conclusions (Table 2) [14]. Knowing that biases are the main threats to RCTs’ internal and external validity, quality appraisal of RCTs was performed using The Cochrane Collaboration’s ‘Risk of bias’ tool tailored specifically for RCTs [38]. The risk of patient selection bias was examined for the selected RCTs for the sufficiency of random sequence generation and concealment allocation to interventional and control groups. This helped to determine the comparability of the study groups at baseline. The risk of performance bias was evaluated based on measures (e.g. single blinding or double-blinding) employed to ensure study participants and personnel are blinded to interventions and outcomes. The risk of detection bias was assessed to know if the assessors were adequately blinded to patient group allocation. The risk of attrition bias and incorporation bias were examined based on the dropout rate and pattern of participants, handling incomplete outcome data, and the indications of intention-to-treat (ITT) analysis. Finally, the risk of reporting bias is evaluated based on the possibility of selective outcome reporting. The reproducibility of exercise therapies prescribed confounding/modifying effects of co-interventions and the levels of supervision and patient compliance to the prescribed physiotherapy during the trial were also evaluated across the RCTs studies.

2.7Data extraction and qualitative synthesis

Data on effect measures were extracted for baseline patellofemoral pain levels, hip exercise interventions, including the comparator treatment, quantitative assessment of patient outcomes for patellofemoral pain and functions, follow-up duration and post-intervention practices during the follow-up periods. Statistical results (mean differences from baseline and effect measures P value at 95% confidence interval) were taken from the evidence tables for interventional studies (separately for controlled clinical trials and cohort studies).

2.8Quantitative synthesis (meta-analysis)

Using MedCalc software version 14.10.2 (MedCalc Software Ltd., Ostend Belgium), data from RCTs that provided the mean difference of pain or knee function between the intervention and the comparator groups were pooled by random or fixed-effect models to obtain standardized mean differences. Separate forest plots were generated for pain and knee function outcomes.

3.1Controlled clinical trials

A total of 383 participants from the 10 CCTs received either hip-strengthening exercises (N= 74) or quadriceps strengthening exercises (N= 157) or hip/quadriceps strengthening exercises (N= 108) or no exercise (N= 44) (Table 2). All CCTs involved true randomization of participants except one, where participants were allocated to their respective groups alternately in a consecutive manner [44].

3.2Intervention protocol

In all CCTs, the hip muscles strengthening protocol focused on hip abductors and lateral rotators. The hip exercise protocol included hip abduction against an elastic band while standing, or with weights in side-lying position coupled with lateral hip rotation against an elastic band while seated and hip extension; quadriceps strengthening involved closed kinetic chain exercise or seated knee extension, leg press, squatting and stretching of hamstrings and quadriceps; and, hip-quadriceps strengthening involved the combination of hip-quadriceps protocol. The duration of intervention ranged from 3 to 8 weeks, while the frequency of therapy sessions ranged from 2 to 4 per week (Table 3).

3.3Outcome measures

All CCTs examined both pain and functional outcomes except one, which assessed only pain [45]. The pain was commonly evaluated using 10-cm VAS by all CCTs except two, which used the 11-point NPRS [4, 28]. The pain was evaluated during ascending and descending stairs [4, 28, 45], squatting, usual pain [45],

Figure 2.

Hip versus quadriceps strengthening on PFP .

Figure 3.

Comparative effect of hip versus quadriceps strengthening on knee function.

Figure 4.

Hip-quad strengthening results in significant pain improvements compared to the standard quadriceps strengthening alone.

Figure 5.

Hip-quad strengthening resulted in a greater functional improvement than the standard quadriceps strengthening alone.

and worst knee pain in the previous week [6, 27, 43, 44, 45, 46]. Functional outcomes were assessed using LEFS [4, 6, 28, 42], AKPS [4, 6, 28, 43], PFJES [41], TLKSS [46] and WOMAC [27, 44].

3.4Follow-up duration

Post-intervention measures were immediately carried out in all studies, at the end of the intervention period. However, the post- interventional follow-up period ranged from one to twelve months (Table 4).

3.5Critical appraisal

Methodological quality assessment of the 10 CCTs based on the Cochrane Collaboration’s ‘Risk of bias’ tool tailored for RCTs is detailed in Tables 5 and 6 [38].

4.1Quality of the summarised evidence

The methodological quality of the fourteen studies except five, i.e., [6, 9, 41, 47, 48] included in the present review is excellent because it fulfilled 14 of the 16-item McMaster critical review criteria. The common methodological issue observed in most of the selected studies was the lack of sample size justification (sample size not determined or not achieved) [6, 27, 41, 43, 44, 47]. All studies with sample power inadequacy issues achieved results with statistical significance, suggesting that the measured pain and functional outcomes reflect the comparative effect of the interventions. However, subject contamination in Dolak et al. was evident because hip and quadriceps groups were combined to receive functional strengthening exercises (as co-interventions for the last four weeks of the intervention) [6]. Such subject contamination might have caused patient bias for their pain and functional outcomes, especially if they know the intervention of their cohorts in the opposite arm of the study [49].

This risk of bias is a critical methodological issue in RCTs and warranted supplementary quality appraisal of all RCTs on the Cochrane Collaboration’s Risk of bias tool [38] to highlight methodological flaws (indicative of ‘Risk of bias’ threatening interval consistency) (Table 6). All RCTs except two recruited participants with a confirmed diagnosis of PFP [27, 44]. However, these studies were included because they enrolled patients presented with anterior keen pain based on symptoms matching the inclusion criteria of the remaining RCTs, which recruited patients with a confirmed diagnosis of PFP. Here, 383 participants from all RCTs presented with anterior knee pain associated with prolonged sitting, climbing stairs and descending stairs in the absence of signs/symptoms of meniscal or other intra-articular pathological conditions or history of other knee pathologies, surgeries and injuries. These are classical symptoms for the diagnosis of PFP [14, 40]. However, these symptoms may indicate knee osteoarthritis, but it may not be so likely because patients enrolled in RCTs were not older than 50 years of age and therefore not likely to present with ageing associated PFP [50].

Four studies included a mixed population of both adults and adolescents aged 17 to 50 years [6, 42, 43, 45]. Since adolescents are physically active and at risk of PFP, hence, the inclusion of this age group [50]. To minimize the possibility of recruiting participants with underlying knee pathologies, i.e. knee osteoarthritis, no studies recruited patients with PFP who were older than 50 [50]. The four studies examined only female participants; therefore, the outcome may only be generalized for the female patients with PFP, but not for the males [4, 6, 28, 42]. The three studies [43, 45, 46] included both male and female participants (proportion of females was higher than males), indicative of females being at a greater risk of PFP than males [6]. This may be attributed to females’ lower hip muscle mass compared to males [51]; therefore, females exhibit lower hip muscle strength than males [51, 52].

The symptom duration is a direct measure of severity of PFP that has a significant influence on therapeutic outcome [53]. Therefore, patients with an early diagnosis of PFP are likely to respond well to therapy compared to those with late diagnoses [18]. Thus, symptom duration is a key confounding variable that must be adjusted via the subject’s stratification. In this systematic review, the mean duration of symptoms of participants with PFP in eight studies ranged from 17 to 21 months. However, six studies [27, 41, 42, 43, 44, 45] did not report the mean duration (months) of PFP symptoms. None of the studies performed the subject’s stratification for the PFP severity and symptom duration. This might have positively skewed pain and functional outcomes in patients with a shorter mean duration of symptoms [18]. Additionally, the subject’s characteristics were barely explained in three studies [42, 44, 45] and not detailed in one study [41]. These findings undermine the quality of the summarised shreds of evidence.

Supervised therapeutic exercises enhance participants’ compliance because unsupervised participants may refrain from pain-provoking exercises [18, 54]. Two previous RCTs reported that supervised exercises for PFP result in less pain and better knee function at short- and long-term follow-up than usual care [18, 54]. In the present systematic review, all studies involved exercises administered in physiotherapy facility/rehabilitation setting under supervision by qualified physiotherapists, except two [6, 45], where two-thirds of exercise sessions were self-administered in patient’s homes (unsupervised). At the same time, one-third had rehabilitation under supervision in a facility. It had an important bearing on patient compliance to intervention and the outcome. Even then, results were significant in these two studies, suggesting that partial supervision too can yield clinically significant results.

4.2Isolated hip musculature strengthening

All fourteen studies demonstrated that isolated strengthening exercises of hip abductors and lateral rotators for two to four times per week up to three to eight weeks duration effectively relieve pain and improve knee function compared to quadriceps strengthening and non-exercise interventions. Kooiker et al. reported variations in quadriceps, hip and hip-quadriceps strengthening protocols in selected studies and opined for the unavailability of standardized protocols for PFP [40]. The common hip exercise protocol included hip abduction against an elastic band while standing and with weights in a side-lying position coupled with lateral hip rotation against an elastic band while seated and hip extension (3 sets of 10 repetitions). Conversely, quadriceps strengthening in all studies generally involved weight-bearing and non-weight-bearing exercises such as closed kinetic chain exercises, seated knee extension, leg press, squatting and stretching of hamstrings and quadriceps (3 sets of 10 repetitions).

The hip protocol generally improved pain and knee function after three to eight weeks of training, with long-term effects observed as late as twelve months post-intervention [28]. Four studies evaluated the comparative therapeutic value of quadriceps versus hip muscle strengthening in treating PFP  [6, 41, 42, 44]. One study by Khayambashi et al. reported superiority of hip muscles strengthening strategy over the quadriceps strengthening for both pain and functional improvement in PFP [44]. The remaining three studies argued that isolated hip and quadriceps strengthening strategies have comparable therapeutic value for the PFP [6, 41, 42]. However, a meta-analysis of the effect measures (pain and function) as measured on VAS and LEFS or WOMAC revealed that hip strengthening significantly favours the standard quadriceps strengthening (p< 0.001) in PFP treatment [6, 42, 44]. Moreover, these findings are strengthened by a study by Na et al., who reviewed the studies related to isolated hip muscle strengthening vs knee strengthening protocol and suggested that isolated hip strengthening is more beneficial in reducing the pain earlier [55]. Thus, these results may aid in providing additional data to fill the existing knowledge gap in the guidelines pertaining to hip targeted exercise therapy in improving the functional performance of patients with PFP [56].

4.3Additional therapeutic effect of hip muscles strengthening

Although the proximal strengthening exercises involving quadriceps and hip muscles are commonly effective in treating PFP, Kooiker et al., Peters and Tyson argued that a combination of hip-quadriceps strategy could add to the therapeutic outcome for patients with PFP [6, 14, 40, 41, 42]. The present systematic review included five RCTs to examine the additional therapeutic outcome of hip-quadriceps strengthening exercises over the standard quadriceps [4, 28, 43, 45, 46]. Except for one, all studies; supported that the hip-quadriceps strategy was superior to the standard quadriceps [46].

The findings of these five RCTs have both internal and external validity and are, therefore, acceptable. Furthermore, meta-analyzed data of these five studies strongly indicated that quadriceps coupled with hip muscle strengthening has significant additional therapeutic benefits over the conventional quadriceps or hip exercises in the treatment of PFP (p< 0.001). Therefore, a hip-quadriceps strategy should be adopted in clinical practices for pain relief and optimal functional improvements in patients with PFP.

4.4Limitations

The summarised evidence supported by meta-analyses indicates that strengthening hip muscles is effective in treating PFP for pain and knee function of physically active male/female adolescents and adults. However, a few important limitations must be noted;

4.5Implications for routine physiotherapy practice

The evidence from the present review has important implications in routine clinical practice for the patients with PFP: