Background
Patellofemoral pain syndrome (PFPS) is a common complaint in adolescents and young adults. The symptom most frequently reported is a diffuse peripatellar (around the knee cap) and retropatellar (behind the knee cap) localised pain, typically provoked by ascending or descending stairs, squatting and sitting with flexed knees for prolonged periods of time. Other common symptoms are crepitus and giving-way (Cutbill 1997; Nissen 1998; Powers 1998; Thomee 1999; Zomerdijk 1998).
Several factors have been implicated in the etiology of PFPS. Malalignment of the lower extremity, sometimes due to excessive pronation of the foot, may result in a compensatory internal rotation of the tibia and increased valgus stress (Shelton 1991). The vastus medialis obliquus (VMO) plays a major role in stabilising patellar glide through the femoral groove. Weakness of the VMO relative to other muscle groups of the quadriceps and aberrant firing patterns of the nerves innervating the VMO and vastus lateralis (VL) have been demonstrated in patients with PFPS (Gilleard 1998). This muscle imbalance may cause maltracking of the patella through the femoral groove, resulting in an abnormal distribution of the patellofemoral joint reaction stress (PFJRS) (Grelsamer 1998). Tight anatomical structures (hamstrings, iliotibial band, patellar retinaculum) (Puniello 1993; Witvrouw 2000b) and overactivity (Holmes 1998; Thomee 1999; Witvrouw 2000b) may also increase the PFJRS. Poor congruence angles between the posterior aspect of the patella and the intercondylar sulcus of the femur predispose for subluxation or even dislocation of the patella, causing cartilage damage (McNally 2001). Clinical studies have not however been able to demonstrate biomechanical or alignment differences between patients with PFPS and healthy individuals (Fairbank 1984; Thomee 1999). Thomee 1997 argues that the combination of malalignment and muscle function deficit may increase the risk of overload and thus PFPS. Increased intrapatellar pressure may cause subchondral degeneration which progresses to the surface and ultimately results in chondral lesions ( Arnoldi 1991; Goodfellow 1976). As cartilage is not innervated, subchondral bone may cause the pain. However, many authors (Bourne 1988; Natri 1998; Nissen 1998; Thomee 1999) report a poor correlation between pain and cartilage damage. Peripatellar soft tissues, such as the patellar retinaculum may also play a role.
The uncertainty regarding the etiology of the complaint also extends to the diagnostic criteria and terms. PFPS is sometimes referred to as 'anterior knee pain' (Clark 2000), a term that may also indicate other medical conditions causing pain in the anterior part of the knee (Cutbill 1997; Bourne 1988) and which often refers more to symptoms than a clear diagnosis. Chondromalacia patellae or chondropathy are often used as a synonyms for PFPS. Nevertheless, in literature there is some agreement that chondromalacia or chondropathy are applied to patients with actual patellar cartilage damage and PFPS is a term to be applied only to patients with retropatellar pain in which no cartilage damage is evident (Arroll 1997; Cutbill 1997; Holmes 1998; Juhn 1999; Thomee 1999; Wilk 1998; ). However, retropatellar pain is generally thought of as a self-limiting condition with a good prognosis, especially for patients who are young (Kannus 1994), patients who have unilateral complaints and patients in which crepitation is absent (Natri 1998). This means that patients are usually managed in primary care and are rarely referred to specialist care (Bourne 1988). Therefore reliable diagnostic techniques for determining cartilage damage such as computed tomography (CT), magnetic resonance imaging (MRI) or arthroscopy (Cutbill 1997; Nissen 1998) are seldom applied. In fact a diagnosis based solely on symptoms and physical examination of the knee is not uncommon. Diagnostic tests often applied are listed here.
Palpation of the lateral and medial aspects of the patella can determine sensitivity of the retropatellar surface. "Clarke's test", "compression test" or "axial pressure test" are synonyms for pressing the patella against the femur and asking the patient to contract the quadriceps. The test is positive when pain or crepitations are present. The patellar grind test is similar but requires pressure to the patella in distal direction. Resisted knee extension can also elicit pain with PFPS. The specificity and sensitivity of these tests is debated in literature, but validation studies are absent. Gaffney found that only half of the patients with PFPS were positive on Clarke's test (Gaffney 1992). In the apprehension test a lateral pressure is applied to the patella. Patients with a history of (sub)luxation will react with sudden contraction of the quadriceps muscles. The relevance of determining cartilage damage with more reliable techniques than physical examination is minimal, as Natri found that neither the radiologic nor the MRI changes seen in affected knees showed a clear association with the seven year outcomes for pain and knee function. (Natri 1998). All things considered the distinction between chondromalacia and PFPS seems theoretical rather than practical, so patients with chondromalacia as well as PFPS will be included in this review.
Most researchers advocate conservative treatment of PFPS or chondromalacia (Arroll 1997; Cutbill 1997; Juhn 1999; Thomee 1999), though there is still insufficient clarity about the effectiveness of the conservative treatment methods (Powers 1998; Wilk 1998; Zomerdijk 1998). This review is being undertaken to clarify the effectiveness of quadriceps strengthening exercises, the most promising conservative treatment method for patellofemoral pain syndrome available (McConnell 1986; Natri 1998; Powers 1998; Powers 2000; Thomee 1999;; Witvrouw 2000a).
Quadriceps strengthening exercise therapy encompasses a broad range of possible variations and accompanying terms. To offer the reader some support with the interpretation of these terms, an overview of the possibilities is given here. Exercises involving contact of the foot with a surface are referred to as "closed kinetic chain exercises", as opposed to "open kinetic chain" exercises which are often prescribed because of the limited forces they elicit in the knee joint. Contractions of the quadriceps muscles can either be concentric, eccentric or isotonic. During concentric contractions the muscles shorten (e.g. when raising a straight leg, extending a bent knee or squeezing a pillow between both legs), whereas during eccentric contractions they lengthen in an actively controlled manner (e.g. when slowly lowering a straight leg, descending stairs or squatting down). Isotonic contractions require a constant strain without changes in the length of the muscle (e.g. during wall squats with knees flexed in 90 degrees and the back against the wall). Exercises in which the position of the knee does not change are referred to as static or isometric. Hence, exercises can be described in three dimensions: the presence of reaction forces caused by contact of the foot with a surface (open versus closed kinetic chain), type of muscle activity (concentric, eccentric, isotonic), and knee movement (flexion/extension versus isometric or static). Combinations of above denominations apply to every type of exercise, and the terminology used for exercise programs reflects the emphasis intended by the therapist.
Quadriceps strengthening exercises are usually combined with stretching exercises, to loosen tight structures like hamstrings, the iliotibial band and the patellar retinaculum. Additional tools provided by therapists to facilitate exercise therapy are patellar taping (McConnell 1986) or Coumans bandaging to adjust the patellofemoral congruence angle and thereby relieve pain and facilitate exercising. Therapists may also apply additional technology in treatment programs. Isokinetic exercises (exercises in which the lower leg moves at a predetermined, constant speed) require an isokinetic dynamometer to control the velocity with which the knee goes through a large range of motion. This device can also measure the concentric as well as eccentric force applied by knee extensors (quadriceps) or flexors (hamstrings) at predetermined velocities. The velocity spectrum for these dynamometers ranges from 0 to 360 degrees per second. Electromyographic biofeedback visualises specific muscle contractions and may help the patient target the Vastus Medialis Obliquus (VMO) during exercise. Electrostimulation provides external stimuli for specific muscles resulting in contractions and thus exercise.
Objectives
The objective of this review was to assess the effectiveness of exercise therapy in the treatment of PFPS, by
- comparing exercise therapy with 'placebo' treatment or no treatment/waiting list controls
- comparing different types of exercise therapy
- comparing exercise therapy with other conservative or surgical treatment
using anterior knee pain and knee function as clinically relevant outcome measures. Measurements up to one year follow-up were considered short term outcomes, thereafter long term.
Methods
Criteria for considering studies for this review
Types of studies
Concurrent, randomised or quasi-randomised controlled trials (RCTs) and concurrent controlled trials without randomisation (CCTs) on exercise therapy for patellofemoral pain were considered. Quasi-randomised treatment allocation pertains to which were not strictly random, such as date of birth, alternation etc. Retrospective studies were excluded.
Types of participants
Adolescent and adult patients suffering from patellofemoral pain syndrome (designated by the author as such or as "anterior knee pain syndrome", "patellar dysfunction" "chondromalacia patellae" or "chondropathy"). Studies which specifically focused on other named knee pathologies such as Hoffa's syndrome, Osgood Schlatter syndrome, Sinding-Larsen-Johansson syndrome, iliotibial band friction syndrome, tendinitis, neuromas, intra-articular pathology including osteoarthritis, rheumatoid arthritis, traumatic injuries (such as injured ligaments, meniscal tears, patellar fractures and patellar luxation), plica syndromes, and more rarely occurring pathologies were excluded (Nissen 1998; Thomee 1999).
Types of interventions
Only controlled trials including at least one treatment arm consisting of exercise therapy aimed at strengthening knee extensor musculature, either at home or under supervision of a therapist were included in this review.
Types of outcome measures
The primary outcome was knee pain. Secondary outcomes focus on functional disability level (i.e. decreased knee function in activities of daily living) and subjective perception of recovery. Questionnaires focusing on knee function (such as Functional Index Questionnaire, WOMAC Osteoarthritis Index, and Kujala Patellofemoral Function Scale, Lysholm scale etc.) and the ability to perform tests (squatting, hopping on one leg etc.) were considered measures for functional disability. Adverse effects like knee swelling or substantially increasing pain levels as a direct effect of treatment were taken into consideration as well. As changes in knee function on impairment level alone (i.e. range of motion, muscle strength etc.) do not directly represent changes in the symptoms of patellofemoral pain or the resulting disability, they were not considered clinically relevant outcome measures in this review.
Search methods for identification of studies
We searched the Cochrane Bone, Joint and Muscle Trauma Group and Cochrane Rehabilitation and Related Therapies Field specialised registers, the Cochrane Controlled Trials Register (The Cochrane Library current issue), PEDro - The Physiotherapy Evidence Database (http://ptwww.cchs.usyd.edu.au/pedro), MEDLINE (1966 to December 2001), EMBASE (1988 to December 2001), CINAHL (1982 to December 2001), and reference lists of articles. No language restriction was applied.
In MEDLINE (OVID WEB), the search strategy was combined with all phases of the optimal trial search strategy (Clarke 2003a) and was modified for use in other databases (see Appendix 1).
Data collection and analysis
Selection of studies
Two reviewers (MB, SBZ) independently selected the trials, initially based on title and abstract. From the title, keywords and abstract they assessed whether the study met the inclusion criteria regarding diagnosis, design and intervention. Of the selected references, the full article was retrieved for final assessment. Next, they independently performed a final selection of the trials to be included in the review, using a standardised form. Disagreements were solved in a consensus meeting.
Data extraction and management
Two reviewers (EH, RB) independently extracted the data regarding the interventions, type of outcome measures, follow-up, loss to follow-up, and outcomes, using a standardised form.
Assessment of risk of bias in included studies
In this review, risk of bias is implicitly assessed in terms of methodological quality.
The methodological quality was assessed by two reviewers (BK, JV) independently. They used the criteria list recommended by the Cochrane Bone, Joint and Muscle Trauma Group, combined with the Delphi list (Verhagen 1998) and one additional question adapted from the criteria list for Methodological Quality Assessment (van Tulder 1997) (see Table 1). Disagreements were solved in a consensus meeting.
For each item Cohen's Kappa and the percentage agreement between both reviewers was calculated, after dichotomising the data into optimal and suboptimal scores (i.e. value 1 was converted to 0). Trials presenting an adequate or concealed randomisation procedure and adequate blinding (Cochrane code A), or a maximum score of five or more Delphi items were labelled "high quality" trials.
Analysis
Analysis of pooled study outcomes was only to be implemented if the studies or subgroups of studies were considered clinically homogeneous and if statistical heterogeneity was not demonstrated. If the trial results were heterogeneous, the factors possibly underlying this phenomenon were considered and summarised. A further analysis using a rating system with levels of evidence based on the overall quality, and the outcome of the studies, was used (van Tulder 1997; van Tulder 2001):
- strong evidence - provided by generally consistent findings in multiple high quality RCTs;
- moderate evidence - provided by generally consistent findings in one high quality RCT and one or more lower quality RCTs, or by generally consistent findings in multiple low quality RCTs;
- limited evidence - provided by only one RCT (either high or low quality) or generally consistent findings in CCTs;
- conflicting evidence - inconsistent findings in multiple RCTs and CCTs;
- no evidence - no CCTs or RCTs.
Where possible, the results of each RCT were expressed as Relative Risks (RR) with corresponding 95 per cent confidence intervals for dichotomous data and weighted mean differences and 95 per cent confidence intervals for continuous data. MetaView, the statistical analysis component of RevMan (RevMan 2000), was used to graphically present the comparisons of each study.
Results
Description of studies
See: Characteristics of included studies; Characteristics of excluded studies.
A total of 12 studies were included in this review: three CCTs and nine RCTs. The studies proved to be rather heterogeneous with respect to participant characteristics (including diagnostic criteria), the type, intensity and duration of therapy, follow-up duration, outcome measures and measurement instruments. Methodological quality was also variable. The studies are presented here, classified for similarities in comparisons. Detailed descriptions can be found in the Characteristics of Included Studies table.
Three studies compared exercise therapy with a control group not receiving exercise therapy.
- In the high quality RCT by Clark 2000 half of the participants received eccentric exercise therapy and were encouraged to exercise daily at home, and half of both the exercise and the non-exercise groups received patellar taping therapy. All four groups received education on the background of PFPS, footwear and appropriate sporting activities, pain controlling drugs, stress relaxation techniques, ice compresses and massage, diet and weight advice, prognosis and self help. All participants attended six sessions for each treatment over three months. Follow-up took place after 3 and 12 months.
- The low quality RCT by Timm 1998 compared a group receiving daily exercise using a Protonics
® device for four weeks with a control group that received no therapy and was not contacted between baseline and follow-up at four weeks. A Protonics® device is a special brace designed to provide progressive resistance exercise during activities of daily living, without restraining motion or protecting knee ligaments. - The low quality CCT by McMullen 1990 compared static, open kinetic chain exercise with isokinetic exercise and a waiting list control group. Control group patients were contacted weekly to monitor the condition of the involved knee, and were promised the most effective therapy of the other two groups after the trial. Exercise training took place in 12 sessions over four weeks. All participants were instructed to refrain from excessive, strenuous daily leg activities during the treatment program. Follow-up measurements were done at four weeks.
The remaining studies compared different types of exercise with each other. Descriptive terms used by the authors differ, but closer consideration of the descriptions of the exercises performed in the trials, enables five studies to be classified as closed kinetic chain exercise versus open kinetic chain exercise.
- The high quality RCT by Witvrouw 2000 compared open with closed kinetic chain exercise three times weekly for five weeks, and patients were advised to maintain their muscle strength until follow-up at three months. During the training program patients were not allowed to participate in sports.
- The high quality RCT by Wijnen 1996 compared the McConnell program (Gerrard 1990) including taping and a closed kinetic chain exercise program to be performed twice daily at home, with a Coumans bandage during six weeks combined with standard home exercises (not further defined). Follow up ended at 6 weeks.
- The high quality RCT by Gaffney 1992 compared concentric exercises in a pain free range (straight leg raises and progression to knee extensions from 90 degrees) with progressively increased eccentric exercises (isometric self-resisted quadriceps, squats, step-ups and step-downs with stretching and McConnell taping to enable pain free training). Participants trained daily at home for six weeks. Participants were encouraged to remain at their desired level of activity and as their symptoms abated, further activity was encouraged. Follow up ended at 6 weeks.
- The low quality RCT by Colón 1988 compared isometric open kinetic chain straight leg raises with closed kinetic chain Pogo-stick bounces, twice daily for six to eight weeks. A Pogo stick is a stick with foot holds, which contains springs to enable bouncing. Only participants with a positive patellar compression sign as well as crepitation were included. Stretching was encouraged. Follow up ended at 8 weeks.
- The low quality CCT by Stiene 1996 compared velocity spectrum isokinetic training (open kinetic chain) with closed kinetic chain exercises, both groups received three training sessions per week and were encouraged to apply ice and stretching during an eight week period. Seven patients with patellar dislocation and acute onset of patellofemoral pain were equally distributed over both groups. Follow up ended at one year.
Four studies compared exercise programs that could not be classified as open versus closed kinetic chain exercise. They fit the inclusion criteria for this review, but cannot be compared to any other study and hence are not used in the best evidence synthesis.
- The low quality RCT by Harrison 1999 compared a group receiving home therapy including stretching and one education session with a group receiving the same exercise under supervision and with a group receiving a more extensive physiotherapist directed treatment program including patellar taping. Participants of both physiotherapist supervised groups attended three sessions a week for four weeks and were instructed to perform individualised stretching exercises at home. Follow-up measurements were performed at 1, 3, 6 and 12 months.
- The low quality study by Thomee 1997 compared eccentric and isometric exercise, focusing on female patients only. Both groups perform both open and closed kinetic chain exercises. The 12 week training period started after three familiarisation sessions, and was performed daily for the first two weeks with supervision three times weekly. Thereafter training was performed three times weekly with physical therapist contact once or twice weekly. Follow-up measurements were performed at 3 and 12 months.
- The low quality study by Gobelet 2001 compared three times weekly supervised isometric proprioceptive training including stretching exercises with three times weekly supervised velocity spectrum isokinetic training and with twice daily electrostimulation at home, during four weeks. Follow-up ended at 4 weeks.
- The low quality study by Dursun 2001 compared groups receiving identical exercise programs using both closed and open kinetic chain exercise which differ only in the use of EMG-biofeedback of VMO and VL (vastus lateralis) contractions. Supervision by a therapist drops from five days a week for the first four weeks, to three times weekly thereafter, up to three months, when follow-up ends.
Details on times of measurements, patient characteristics and diagnostic terms are shown in the Characteristics of Included Studies table.
Risk of bias in included studies
Two reviewers (BK, JV) independently determined the methodological quality of the 12 selected studies. Consensus was reached after a meeting between both reviewers. The methodological quality of the selected studies as determined during the consensus meeting is shown in Figure 1. Cohen's Kappa and % agreement were calculated for the initial scores given by each reviewer independently (Figure 1). If the number of Delphi items that received a maximum score exceeded four, the study was labelled as "high quality". If the assigned treatment was adequately concealed prior to allocation (item M-A/D2) the study received Cochrane code A, which was used as an alternative criterion for high quality of a study and is listed in the Characteristics of Included Studies table.
 | Figure 1. Methodological quality scores after consensus meeting |
Effects of interventions
Of the 750 titles and abstracts identified by the systematic search of the literature, two reviewers (SB, MB) selected 16 studies that met the inclusion criteria. The methodological quality assessment of these studies is described in the previous section. The remaining two reviewers (EH, RB) extracted data from the publications. Four studies (Beetsma 1996; Eburne 1996; Kowall 1996; Roush 2000) had to be excluded from the review: Beetsma 1996 and Eburne 1996 due to lack of detail in description of procedures and outcomes; Kowall 1996 because both treatment arms performed the same exercises, and the objective of the study was to evaluate the effectiveness of additional taping. Furthermore, Roush 2000 also included patients with Osgood-Schlatter and plica syndromes. Twelve studies were included in the review, representing 697 patients, with an equal number of males and females, and an age ranging from 11 to 65, with an average of 24. An overview of further patient characteristics and the outcome measures is given in the Characteristics of Included Studies table.
Quantitative meta-analysis of pooled high quality studies was impossible due to the heterogeneity of the interventions used for comparison, heterogeneity of gathered outcome measures and applied instruments and heterogeneity of assessment times. For qualitative analysis we identified two comparisons that were addressed by more than one trial. First of all, the question whether patients receiving exercise therapy improve more than patients on a waiting list or patients receiving conservative treatment without exercise. Second, the question whether weight bearing exercises, more closely resembling activities of daily living (closed kinetic chain) provide better results than non-weight bearing exercises (open kinetic chain). Descriptions of each treatment were closely examined to determine whether the study under investigation could contribute to a best evidence synthesis for either one of these questions. Evidence provided by the studies is summarised in Table 2 and Table 3. Four studies describe unique comparisons not addressing these questions.
Exercise versus no exercise
- In the high quality RCT by Clark 2000 patient groups receiving exercise therapy were pooled and compared to pooled patient groups not receiving exercise therapy. It was shown that functional ability improves equally in both pooled groups. Pain reduction was not significantly different at 3 months. At the 12 month assessment Clark states that the groups receiving exercise therapy experienced significantly greater pain reduction. Clark reports means and SD of changes only for the 3 month assessment, based on individual changes. Our calculations based on means per time-point do not exactly reproduce these figures nor the statistical difference at 12 months. The number of patients discharged from therapy because they were satisfied with the results were significantly greater for the group that exercised. The number needed to treat was 3 (95%CI: 1.6 to 3.3), so three exercising patients yielded one more satisfied patient than expected in the control group.
- The low quality RCT by Timm 1998 showed that resistance during ADL provided by the Protonics® device almost halves the pain-scores compared to the control participants, and drastically improves functional ability after four weeks. Both effects differ significantly from the control group that did not receive exercise therapy.
- McMullen 1990 found in his low quality CCT that static exercise improved function more than isokinetic exercise, though both types provided only minimal improvement compared to the waiting list controls. Pain levels are not reported, though the author states that they remained unchanged for all groups after four weeks.
From the best evidence synthesis it follows that there is limited evidence to support the hypothesis that exercise therapy reduces anterior knee pain in patients with PFPS: one high quality RCT and one low quality RCT claim significant pain reduction, and one CCT with a small number of patients contradicts this. There is conflicting evidence of functional improvement: one high quality RCT and one small CCT do not find improvement whereas one low quality RCT does.
Open kinetic chain versus closed kinetic chain
For categorising the studies, the descriptions of the exercises rather than the terminology in the publications was used.
- The high quality RCT by Witvrouw 2000 showed that both function and pain improve significantly with both types of exercise, though no significant differences between the groups are found.
- The high quality RCT by Wijnen 1996 showed no statistical differences for pain and function. However patient satisfaction with the therapy is significantly greater in the group combining closed kinetic chain exercises with McConnell taping.
- The low quality RCT by Gaffney 1992 reported no significant differences in pain and function outcomes between eccentric closed kinetic chain and concentric open kinetic chain exercises.
- The low quality CCT by Stiene 1996 shows that though muscle strength improves in both groups, the closed kinetic chain exercise results in significantly better function as determined through retro-step up performance. This result is dubious as baseline values differ significantly between groups. The representation of Functional Index Questionnaire results was inadequate for interpretation. Pain was not reported in this study.
- The low quality RCT by Colón 1988 focuses completely on muscle strength, but does not provide statistical analyses to compare groups. He found that almost all patients in both groups report substantial (>50%) pain relief, but pain levels are not reported and differences between groups are not apparent.
The results of both high and low quality RCTs are consistent for both pain and function, so there is strong evidence to support the hypothesis that closed kinetic chain exercises provide equal results to open kinetic chain exercises for either pain reduction or function improvement.
Other comparisons
- The low quality RCT by Harrison 1999 showed improvement in all groups for both pain and function, which is stated to be significant for the Patellar Function Scale. However, these outcomes were not significantly different between home exercise and the supervised exercise groups. Interestingly, our analysis of the presented data revealed that significantly more patients from the physical therapy group rated their clinical change as "significant improvement" compared to the home exercise group, though the author states there is no significant difference.
- In the low quality RCT by Thomee 1997 a significant reduction of pain in all visual analogue scales is reported, both at three months and again at 12 months, though no differences between isometric and eccentric exercise groups were found. No pain levels are reported, only frequencies of patients with pain in three situations. Lysholm knee function scores are not reported. Muscle strength increased significantly in both groups, though no significant differences were found except in a 25 degree range during eccentric contractions.
- The low quality RCT performed by Dursun 2001 did not reveal any differences between the outcomes of the groups exercising with or without EMG-biofeedback.
- The low quality RCT by Gobelet 2001 found significant increases in clinical evaluation of the knee using the Arpège score list for the groups receiving electrostimulation and isometric exercise. Isokinetic exercise did not improve the status. Isokinetic muscle strength improved in the groups receiving electrostimulation and isokinetic training, but in the group receiving isometric training strength did not improve at all isokinetic velocities at which muscle strength was measured.
Discussion
Exercise versus no exercise
Only one of the three trials comparing exercise with no exercise was of high quality. The best evidence synthesis suggests that there is some indication that exercise is effective, but the data are not straightforward.
McMullen 1990 argues the time period of four weeks may be too short, though other authors, such as Timm 1998 have found significant improvement in this period. The intensity of the exercises may be the clue, as Timm's participants received daily therapy for several hours during activities of daily living. However, the Protonics® device will not be universally applied and is therefore of clinically limited relevance. The first follow-up assessment by Clark 2000 was found after three months, at which time point improvement was made in all groups, though the difference between the exercising and non exercising groups only became apparent after one year. It is possible that the 60% drop-out rate after 12 months in Clark's study contributed to this significant difference by introducing attrition bias.
But what explains the difference in effect seen in different control groups? One might argue that the improvement observed in Clark's study reflects the natural course of the affliction. However, the duration of symptoms prior to the study makes this unlikely. Another explanation may lay in the effect that participating in and fulfilling the requirements of a study alters an individual's behaviour, thereby contributing to the improvement. This is the so called Hawthorne effect. It is also possible that education may affect the behaviour of patients more than mere enrolment in a study when the treatment comes down to being placed on a waiting list. The duration of the trials by Timm 1998 and McMullen 1990 may also be too short to establish the Hawthorne effect, because it may take longer than four weeks for behavioural changes to result in clinical improvement. However, the assumption that behavioural changes occur, cannot be established from the reported results.
Although the studies performed by Clark 2000 (N=81) and Timm 1998 (N=100) have the largest number of patients of all included studies, it should be noted that the number of patients in these studies is still modest.
Open kinetic chain exercise versus closed kinetic chain exercise
The concept that closed kinetic chain exercises would be more effective than open kinetic chain exercises because they more closely resemble activities of daily living was not supported by evidence in any of the studies considered in this review. Greater satisfaction with McConnell treatment found by Wijnen 1996 could either be attributed to the closed kinetic chain exercises or to the application of McConnell tapes instead of Coumans bandages. This touches a problem that calls for reservations in the interpretation of this best evidence synthesis. It should be noted that though the common factor in these five studies is the contrast of open versus closed kinetic chain exercise, the differences in all other aspects of the interventions are considerable. The terminology the authors use for their exercise programs reflects the factor the author is most interested in and hence the different accents in each exercise program.
Methodological quality
Quality assessment
Overall the agreement between reviewers on the methodological scoring was reasonable, and consensus was reached without problems. Poor reporting of the studies was partly responsible for the poor agreement between the reviewers for item M-G: Were care programmes, other than the trial options, identical? The sometimes meagre descriptions of the treatment programs made evaluation of comparability harder, but interpretation of reported facts also led to problems: is the mention of differences in permission to use patellar taping, analgesics or infra-red treatment part of the trial options, or does it supplement these options? The duration of the treatments was always identical. The different scores for item M-K can be attributed partly to the fact that the term diagnostic tests raised confusion as to whether the tests are used for screening purposes or for outcome assessment. Furthermore, it is open to interpretation whether assessment of symptoms like pain during certain activities can be viewed as diagnostic tests.
Cut-off point for high quality
The nature of exercise therapy makes it impossible to conceal treatment allocation to the patients or for the treatment providers, which results in a maximum feasible score of 7 out of 9 Delphi items. The cut-off point for the number of Delphi items needed for the qualification "high methodological quality" coincided with the allocation of Cochrane code "A", and the difference between the high quality scores and the low quality scores always amounted to at least 2 Delphi items. Dursun's study is the only study that might be qualified as high quality when a different cut-off point is chosen. However, this study does not answer any of the clinically relevant research questions. Therefore, the cut-off point for classification of high or low quality was deemed justified for use in our planned qualitative analysis and no analysis was performed using an alternative cut-off point.
Methodological shortcomings
Though all studies intend to compare treatments, some authors have failed to provide a statistical analysis between treatment groups. They suffice with stating whether within each group significant changes occur. However, when significant changes occur within each group, the question whether some treatments provide better effects is not answered. Worse, when significant changes occur within one group, but not another, comparison of both groups may not produce statistically significant differences. Especially in studies where blinding of treatment allocators during randomisation was not described (i.e. all low quality studies), and where baseline characteristics and measurements were not equal, the method of reporting within group changes can be very misleading.
Though some authors of low quality trials describe their methods in detail, this detail is sometimes lacking in the reporting of outcome measures. Shortcomings range from failing to report outcomes that are mentioned in the methods section (Thomee 1997 (VAS), McMullen 1990 (Cincinnatti Rating Scale (CRS))), not mentioning the number of patients (Gaffney 1992: VAS and diagnostic tests), to methods of data reduction that prevent insight in the data. For example originally continuous data are converted to ordinal (Harrison 1999 (Functional Index Questionnaire (FIQ))), or even dichotomous data (Thomee 1997 (VAS)), which also hampers insight in variability of the data. McMullen 1990 and Colón 1988 fail to report baseline data. Although McMullen 1990 presents ANCOVA outcomes and post-treatment values giving the reader an opportunity to deduce estimators of baseline values, Colón 1988 only presents the number of patients with at least 50% pain reduction. Furthermore, drop-outs have rarely been reported properly and intention to treat analyses were even more rare.
Timm 1998, McMullen 1990 and Dursun 2001 include only patients with unilateral afflictions which may give a biased representation of the patient population. Witvrouw 2000, Harrison 1999, and Thomee 1997 have taken the approach of including both unilateral and bilateral patients, choosing the most afflicted leg as object of investigation. However, Gaffney 1992 uses both unilaterally and bilaterally afflicted patients, but has reported data that represent knees instead of patients, without giving the number of patients involved.
Outcome measures
Pain is the symptom that prompts the patient's visit to a doctor, and function may be limited as a result. Muscle imbalance and/or weakness may be the underlying problem or a condition for PFPS to evolve, so muscle strengthening is a means to treat PFPS, but it is not a goal in itself in the management of PFPS. However, isokinetic power and torque measurements as quantifiable measure for muscle strength are used as outcome measures by some authors. Natri 1998 showed that restoration of quadriceps strength is important for good recovery of the patient, as determined by the difference between affected and unaffected leg: the smaller the difference in extensor strength, the better the outcome. However, none of the authors in this review chose the difference between legs as outcome parameter, which is understandable, given the fact that some patients have bilateral complaints. Presentation of these results would therefore muddy the overview given here, so we chose to leave them out. Not surprisingly, for all groups receiving exercise therapy, muscle strength increased, and differences found when comparing exercise treatments were usually minimal. Stiene 1996 notes that improving muscle strength did not improve the patient's function and Dursun 2001 found that improved muscle function appeared to have no effect on the clinical and functional status. Gobelet 2001 found that isokinetic training increased muscle strength, though not clinical improvement, whereas isometric training did not increase muscle strength, but improved the clinical outcome. These findings illustrate the difficulty of interpreting the effect of therapy using muscle strength as an outcome measure for knee function. Therefore we chose to determine effectiveness using outcomes more directly related to the wellbeing of the patients involved. Hence, our choice not to include muscle strength as relevant outcome measure in determining the effectiveness of PFPS seems justified.
Compliance and withdrawal
Compliance problems can be viewed as an inescapable element of exercise therapy, so compliance problems in trials can be viewed as truthful representations of medical practice, which is why an intention to treat analysis is imperative. Harrison notes that many drop-outs showed good results, and suggests an underestimation of the effect of treatment is given. Unfortunately, few authors have reported compliance in a satisfactory manner. Colón 1988 reports one participant dropping out because of increased symptoms. Stiene 1996 reports non-compliance and unavailability for final testing as reasons for dropping them from analysis. Gobelet 2001 has withdrawn patients from analysis because of poor compliance, defined less than 70% attendance of training sessions. If no intention to treat analysis is performed, at least a comparison of baseline values of outcome measures of the drop-outs would be useful, to determine the possible bias of results. As most authors have not reported an intention to treat analysis and most studies struggle with high drop-out rates and small population sizes the effect of compliance as a confounder must be deemed significant, though elusive. High drop-out rates are evident in many studies, and make the feasibility of long term assessments problematic.
Power
If one looks at the limited evidence for the effectiveness of exercise therapy, one can see that benefits from exercise therapy seem relatively small, and variances (if reported) are rather large. When comparing different types of exercise therapy it is only logical that differences between treatment groups are even smaller. It is therefore regrettable that patient numbers in the included studies were, in general, rather small, and in some cases alarmingly so. This makes it almost impossible to detect differences between treatment groups (type II error). When reading this review it should be kept in mind that the low power and the other methodological flaws discussed above make it hard to reach any firm conclusions.
Authors' conclusions
Implications for practice There is limited evidence for the effectiveness of exercise therapy for PFPS. |
Implications for research Prior to the study an assessment of the disease burden, the pain levels and the level of function impairment of the expected study population should be made, and patients should be asked how much improvement they expect from exercise therapy for it to be called successful, given the effort it requires. Taking into account the variance of these outcome measures, a power calculation should be made to determine the minimal number of patients required for detection of the desired effect. A factorial design aimed at studying the additional effect of education, taping, or any form of pain relief may be considered to determine the role of various co-interventions commonly applied. The population size required would have to be determined with adequate power analysis. Future researchers should beware of the misleading notion that muscle function represents the clinical status of PFPS, and use pain and function as the primary outcome measures in any trial studying the effectiveness of exercise therapy for PFPS. Questionnaires to assess the status of knee function often include questions about pain. However, separate pain measures are a valuable addition to the assessment of knee status, as can be seen from Clark's study, where pain reduction is significantly greater in the exercise group, whereas function assessments do not show this significant difference. The limited evidence for effectiveness of exercise therapy for PFPS shows that the ethical objections of several authors against using a control group not receiving any therapy are based more on the assumption of effectiveness of exercise therapy than on sound scientific evidence. This observation should be considered by investigators who wish to contribute to the discussion on effectiveness of exercise therapy by performing studies of high methodological quality, which should compare exercise therapy to a control group not receiving exercise therapy. |
Acknowledgements
We thank the following for helpful comments at editorial review: Lesley Gillespie for her extensive help with the search strategy and text editing, Prof. William Gillespie, Prof. Marc Swiontkowski, Prof. Rajan Madhok, Dr. Janet Wale, Dr. Bruce Arroll, Leeann Morton, Peter Herbison, Prof. Tracey Howe, and Kate Rowntree. We would also like to thank Hilda Bastian for her help with the production of the synopsis. We also thank Dr. Arianne Verhagen for helpful advice on the subject of methodological quality scoring and qualitative data analysis.
Data and analyses
- Top of page
- Background
- Objectives
- Methods
- Results
- Discussion
- Authors' conclusions
- Acknowledgements
- Data and analyses
- Appendices
- What's new
- History
- Contributions of authors
- Declarations of interest
- Sources of support
- Notes
- Index terms
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Appendices
Appendix 1. Search strategy for MEDLINE (OVID WEB)
1. Arthralgia/
2. Knee Joint/ or Knee/ or Patella/
3. and/1-2
4. anterior knee pain.tw.
5. ((patell$ or femoropatell$ or femoro-patell$ or retropatell$) adj2 (pain or syndrome or dysfunction)).tw.
6. ((lateral compression or lateral facet or lateral pressure or odd facet) adj syndrome).tw.
7. ((chondromalac$ or chondropath$) adj2 (knee$1 or patell$ or femoropatell$ or femoro-patell$ or retropatell$)).tw.
8. or/4-7
9. or/3,8
10. exp Physical Therapy/
11. (exercis$ or strengthen$ or stretch$).tw.
12. (stabil$ adj3 train$).tw.
13. or/10-12
14. and/9,13
What's new
Last assessed as up-to-date: 16 June 2003.
| |||||||||
History
Protocol first published: Issue 1, 2002
Review first published: Issue 4, 2003
Contributions of authors
Edith Heintjes: data extraction, data analysis and text of the review
Marjolein Berger: primary support, study selection, feedback on clinical aspects
Sita Bierma: study selection and feedback on physical therapy terminology and textual feedback
Roos Bernsen: data extraction and statistical feedback
Jan Verhaar: methodological scoring and textual feedback
Bart Koes: methodological scoring and textual feedback
Declarations of interest
None known.
Sources of support
Internal sources
- Department of General Practice of the Erasmus University Rotterdam, Netherlands.
- Department of Orthopaedics of the University Hospital Rotterdam, Netherlands.
External sources
- No sources of support supplied
Notes
This study will serve as the basis of a thesis for a PhD study in 'Non traumatic knee injuries in adolescents' at The Institute of General Practice at the Erasmus University Rotterdam by E.M. Heintjes (MSc) and will be supervised by Dr. M.Y. Berger (MD) and Dr. S.M.A. Bierma-Zeinstra (PhD).
Index terms
Medical Subject Headings (MeSH)
*Knee Joint; *Patella; Adolescent; Arthralgia [*rehabilitation]; Exercise Therapy [*methods]; Syndrome
MeSH check words
Adult; Humans