Abstract. Bone mineral density (BMD) and clinical status of 40 patients with a chronic, unilateral patellofemoral pain syndrome (PFPS) were determinated.
Calcif Tissue Int (1998) 62:548–553
© 1998 Springer-Verlag New York Inc.
Bone Mineral Density in the Chronic Patellofemoral Pain Syndrome J. Leppa¨la¨,1 P. Kannus,1 A. Natri,2 H. Sieva¨nen,1 M. Ja¨rvinen,3 I. Vuori1 1
Accident and Trauma Research Center, UKK Institute for Health Promotion Research, Tampere, Finland Tampere Research Center of Sports Medicine, UKK Institute, Tampere, Finland 3 Department of Surgery, Tampere University Hospital and Medical School, University of Tampere, Tampere, Finland 2
Received: 30 May 1997 / Accepted: 8 January 1998
Abstract. Bone mineral density (BMD) and clinical status of 40 patients with a chronic, unilateral patellofemoral pain syndrome (PFPS) were determinated. The mean duration of the disease at the time of the follow-up was 7.6 ± 1.8 (SD) years. The BMD was measured at the spine (L2–L4), and the femoral neck, trochanter area of the femur, distal femur, patella, proximal tibia, and calcaneus of both lower extremities using a dual-energy X-ray absorptiometric (DXA) scanner. The mean BMD of the affected limb (compared with the unaffected side) was significantly lower in the distal femur (−3.3%; P 4 0.002), patella (−2.5%; P 4 0.016), and proximal tibia (−1.9%; P 4 0.008). The femoral neck, trochanter area of the femur, and calcaneus showed no significant side-to-side differences, and the spinal BMDs of men and women with the PFPS were comparable with the manufacturer’s age-adjusted reference values for Western European men and women. The relative BMDs of the affected knee showed strongest correlation with the muscle strength of the same knee: the better the muscle strength compared with the healthy knee, the higher the relative BMD (r 4 0.56–0.58 with P < 0.001 in each anatomic site of the knee). In the stepwise regression analysis, low body weight or low body mass index, high level of physical activity, the patient’s good subjective overall assessment of his/her affected knee, and short duration of the symptoms were also independent predictors of the high relative BMD in the affected knee so that along with the muscle strength these variables could account for 51% of the variation seen in the relative BMD of the femur, 61% in the patella, and 54% in the proximal tibia. In conclusion, chronic patellofemoral pain syndrome results in a significantly decreased BMD in the knee region of the affected limb. The spine, proximal femur, and calcaneus are not affected. Recovery of normal muscle strength and knee function seems to be of great importance for good BMD. Key words: Knee disorders — Bone mineral density — Patellofemoral pain — Bone loss — Rehabilitation.
Low bone mineral density (BMD) is found to be a major risk factor for osteoporotic fractures among the elderly [1– 4]. Insufficient physical activity is, in turn, one of the obvious risk factors for low bone density: prolonged immobi-
lization, such as bed rest, results in extensive loss of bone [5–7]. Many diseases and disabilities also cause bone loss, at least partly because of physical inactivity of the patient [8, 9]. Also, musculoskeletal injuries with subsequent immobilization and disuse of the injured limb lead to a rapid loss of bone and it is questionable whether full recovery is possible [10, 11]. Patellofemoral pain syndrome (PFPS) is a characteristic musculoskeletal pain state that is often seen in young adults [12–14]. The classic form of PFPS, chondromalacia patellae, is characterized by macroscopic softening, fissuring, and fragmentation of the undersurface of the patella [12– 14]. Once it has begun, PFPS frequently becomes chronic, and the pain forces the patient to stop physical activities. The etiology and pathogenesis of PFPS are still unknown, but several predisposing factors have been proposed, including acute trauma, overuse, immobilization, overweight, genetic predisposition, malalignment of the knee extensor mechanism, congenital anomalies of the patella, prolonged synovitis, recurrent hemorrhage into a joint, joint infection, and repetitive intraarticular injections of corticosteroids [12–16]. In many cases, however, there are no obvious factors behind the PFPS [14]. The most common symptoms are retropatellar pain and crepitation during various activities, such as squatting, running, jumping, and going up or down stairs. Also, patients may feel patellar pseudolocking, snapping, knee stiffness, and periodical effusions. The literature provides a wide scale of conservative and operative managements for PFPS but so far there is no general consensus on treatment. Nevertheless, quadriceps muscle exercises have been recommended by many authors as an effective way to correct the faulty patellar tracking and reduce the symptoms [13–18]. The objective of our study was to examine, with dual energy X-ray absorptiometry (DXA) the long-term effects of a PFPS on the BMD of the affected limb and to determine whether the results are associated with the patients’ age, weight, height, duration of symptoms, pain assessment, muscle strength, and function of the affected knee. In addition, the effect of gender, limb dominance, general physical activity level, one-leg jumping, duck-walking, and subjective and objective overall assessment of the knee was studied. Materials and Methods Subjects
Correspondence to: Jari Leppa¨la¨, The Bone Research Group, the UKK Institute, Kaupinpuistonkatu 1, FIN-33500 Tampere, Finland
Forty-nine consecutive patients with a characteristic symptomatic PFPS in one limb were initially treated in our clinic by nonopera-
J. Leppa¨la¨ et al.: Patellofemoral Pain Syndrome and BMD
Table 1. Criteria for inclusion into the study of chronic Patellofemoral pain syndrome
1. Age 15–50 years, with closed epiphyseal growth plates 2. No general illnesses or medication use 3. Characteristic history and symptoms of patellofemoral pain syndrome at one extremity: At least 2 months duration of retropatellar pain during physical activity such as jumping, running, squatting, and going up or down stairs Characteristic patellar crepitation during squatting and a positive ‘‘movie sign’’; that is, retropatellar pain when the knee is kept in flexion for a prolonged period, with relief of pain on extension of the knee Possibly one or more of the following symptoms in conjunction with physical activity: patellar snapping, pseudolocking, giving way, stiffness, and periodic knee effusions 4. Characteristic clinical signs of patellofemoral pain syndrome at one extremity: Retropatellar pain and crepitation on patella compression, grinding, and/or apprehension tests Retropatellar pain on one-leg squatting 5. Negative findings on clinical examination of knee ligaments (instability), menisci (tears), bursae and synovial plicae (inflammation), Hoffa’s fat pad (rupture and inflammation), iliotibial band (friction syndrome), and hamstring, quadriceps, and patellar tendons and their insertions (rupture and inflammation) 6. Negative findings on radiographs of the knee (anteroposterior and lateral weight-bearing views, a tunnel view, and a tangential view of the patella, with the knee in 30° of flexion) with regard to changes of osteoarthritis, osteochondritis dissecans, or loose bodies in the patellofemoral and tibiofemoral joints
tive methods; the treatment included intensive quadriceps muscle rehabilitation for 6 weeks [14]. The criteria for initial inclusion in the study (Table 1) were based on previous studies of the PFPS [17–19]. Forty-five of the 49 patients (92%) attended the current 7-year follow-up visit. One patient had died, one had moved to Thailand, and the remaining two could not be traced. Additionally, we had to exclude five patients from the current study because of patellofemoral osteoarthritis (two patients) or bilateral disease (three patients), thus a total of 40 patients (19 men and 21 women) were studied. Their mean age at time of the follow-up was 33 ± 6 (SD) years, ranging from 21 to 58 years. Their general physical activity level was divided into three categories: very active 4 competitive athletes (2 subjects); moderately active 4 regular recreational sporting activity (33 subjects), and sedentary 4 no mentionable physical activity during a normal day (5 subjects). The mean duration of the disease at the time of follow-up was 7.6 ± 1.8 years. The right limb was involved in 26 patients and the left in 14. Each patient was examined by the same physicians (PK and AN) at study entry, after 6 months [14], and after 7 years. All subjects were told of the study procedure, purposes, and any known risks, and all gave their written informed consent. The study was approved by our clinic’s Ethical Committee on Human Research. Study Protocol Subjective Evaluation. Pain during physical activity was assessed using a 100 mm visual analog scale, from no pain 4 0 to extremely intense pain 4 100 [20]. A 3-point scale was used to classify the subjective overall assessment: 1 4 excellent, completely asymptomatic knee and 3 4 fair or poor, no relief of
549
symptoms or even more symptoms than before during physical activities (see the footnote in Table 4). Functional Tests. To evaluate the function of the affected knee, the patient was asked to walk, duckwalk, run, run in place, jump on one leg, do a half-squat, and squat fully (repeatedly) [14, 21, 22]. With information from these functional tests and the above subjective evaluation, the standardized functional scores of Lysholm and Gillquist [21] and Tegner and Lysholm [22] were calculated for the affected knees. In the scoring system of Lysholm and Gillquist, the affected knee is given a score of from 0 to 100 (100 the best); in that of Tegner, the scores range from 0 to 10 (10 the best). Clinical Evaluation. The following parameters were recorded at the clinical follow-up visit: pain during the patella compression test (compression and grinding of the patella against the femur); pain during the patella apprehension test (tightening of the quadriceps muscles with the knee extended while the examiner prevents the patella from moving upwards); and patella crepitation during the compression test [13, 23]. These tests were evaluated using a 4-point scale, where 1 4 no pain or crepitation and 4 4 intense pain or crepitation. The physician’s opinion of the overall assessment was based on a 3-point scale, where 1 4 complete subjective, functional, and clinical recovery and 3 4 overall status not improved or even worse than before (see the footnote in Table 4). Muscle Strength Measurement. Isometric quadriceps muscle strength was measured with the knee in 60° of flexion (fully extended knee 4 0°) using a standardized isometric dynamometer (Digitest Inc., Muurame, Finland) [14, 24]. Three maximal efforts were allowed and the best result was recorded. In each subject, the value for the unaffected side was used as the reference point. BMD Measurement. The areal bone mineral density (BMD in g/ cm2) was measured from the femoral neck, trochanter area of the proximal femur, distal femur, patella, proximal tibia, and calcaneus of both lower limbs, and the lumbar spine (L2–L4) using a Norland XR-26 DXA scanner (Norland Inc., Fort Atkinson, WI): (Fig. 1.). All the measurements were performed by the same technician according to our established procedures [25]. The coefficients of variation for the sites used in this study were 0.7% for the lumbar spine, 0.5% for the femoral neck, 0.5% for the trochanter area, 1.0% for the distal femur, 0.7% for the patella, 0.7% for the proximal tibia, and 0.7% for the calcaneus [26]. Statistical Analysis The data were analyzed by an IBM compatible 486 microcomputer using the 1990 version of the statistical package of the BMDP [27]. The relative BMD of the affected limb was calculated by dividing the affected-to-unaffected side difference by the absolute value of the unaffected side and multiplying the outcome by 100. BMD of the affected and unaffected sides was compared using a matched, paired t-test. The effect of gender, limb dominance, level of physical activity, ability for one-leg jumping and duckwalking, and subjective and objective overall assessments of the knee on the relative BMD in the affected limb was determined by an analysis of variance. The associations among age, weight, height, body mass index, duration of symptoms, pain assessment, muscle strength, the two functional knee scores, and the relative BMD in the affected limb were determined by the regression equations and corresponding Pearson product moment correlation coefficients (r). A multiple stepwise regression analysis (forward stepping) was then performed, starting with the predictor (independent variable) that had the strongest (significant at least in 5% level, P < 0.05) association with the dependent variable (4 sideto-side difference in the BMD of the distal femur, patella, and proximal tibia) in the individual comparisons. The 5% significance level was also used as a criterion to include any additional predictive variable in the model. The results are expressed as the mean ± SD (95% confidence
550
J. Leppa¨la¨ et al.: Patellofemoral Pain Syndrome and BMD
Fig. 1. Anatomic sites of the BMD measurements. The numbers refer to millimeters.
interval) throughout the study. The given significance levels refer to two-tailed tests. An alpha level (type I error) of less than 5% (P < 0.05) was considered significant. Using this level and regarding 5% difference in BMD between the affected and unaffected limb as clinically important findings, the sample size of the study was calculated to give 80% statistical power (beta level or type II error 0.20) for a decision to accept the null hypothesis if no difference could be found. Results Lower Limbs
The absolute (g/cm2) and relative (side-to-side difference in percentage) BMDs are reported in Table 2. For the 40 patients with the PFPS, there were significant differences between the affected and unaffected limb in the knee region (Table 2). The mean BMD of the affected limb was significantly lower in the distal femur (−3.3%; P 4 0.002), patella (−2.5%; P 4 0.016), and proximal tibia (−1.9%; P 4 0.008). In the other sites, the mean BMDs of the affected limb were also lower than those in the unaffected side, but the differences were not significant (femoral neck −1.0%, trochanter area −1.1%, and calcaneus −0.1%, all NS). Neither age, height, weight, body mass index, duration of symptoms, nor pain assessment and functional knee scores had a clear and systematic association with the relative BMD of the affected knee (Table 3). Neither gender, the side of the affection, limb dominance, level of physical activity, nor the results of patella compression and apprehension tests, and the functional tests of one-leg jumping and full-squatting were associated with the outcome. However, muscle strength correlated significantly with the relative BMDs in the distal femur, patella, and proximal tibia (r 4 0.58, 0.56 and 0.57 with P < 0.001 in all sites) (Table 3): the better the muscle strength in the affected lower limb (compared with the other limb), the higher the relative BMD in the same limb. Also subjective overall assessment (P 4 0.004), objective overall assessment (P 4 0.038), and duckwalking test (P 4 0.028) had a significant association with the relative BMDs of the affected patella (Table 4). In the proximal tibia, only subjective overall assessment showed a significant association (P 4 0.023). In the distal femur, these above-noted three variables showed a tendency for association with the relative BMD, but none of the associations was significant (Table 4). According to these results, the forward-stepping regression analysis started by fixing the muscle strength as the predictive variable, with the muscle strength alone (r2) accounting for 34%, 31%, and 33% of the variation seen in the
relative BMD of the distal femur, patella and proximal tibia, respectively. In the distal femur, inclusion of body weight (11%) and physical activity level (6%) in the model significantly improved the overall prediction so that together with muscle strength, these predictors accounted for 51% of the variation seen in relative BMD of the distal femur. In the patella, subjective overall assessment (15%), body mass index (9%), and limb dominance (6%) were additional independent predictors, all accounting for 61% of the variation. In the proximal tibia, body weight (13%) and duration of symptoms (8%) were such additional predictors with all the predictors together accounting for 54% of the variation. Lumbar Spine
The spinal BMD of the 19 men with PFPS was 1.077 ± 0.100 g/cm2 (95% confidence interval 1.029–1.125), and that of 21 women 1.079 ± 0.110 g/cm2 (95% CI 1.026– 1.132). The mean value of the patients was comparable to the manufacturer’s age-adjusted reference values for Western European men (1.120 ± 0.120 g/cm2) and women (1.090 ± 0.120 g/cm2). Discussion
The effects of different types of acute knee injuries on BMD at various skeletal sites have been studied previously [28– 34], but investigations of effects of various chronic knee disorders, such as PFPS, on BMD are scarce. Our study showed that PFPS had created a statistically significant bone loss to the affected knee (distal femur 3.3%, patella 2.5%, and proximal tibia 1.9%). As such, however, this amount of bone loss is not very much, especially if compared with the side-to-side deficits we have seen after acute knee injuries [28–30]. One explanation for the difference could be that patients with PFPS never suffer from complete immobilization of the affected limb, whereas patients with acute injuries often do. In other words, despite pain and discomfort, the patients with PFPS still have to bear the body weight and this may have a considerable bonepreserving effect in the affected limb. Because the original BMDs at various skeletal sites were not determined in the patients before their PFPS, two factors should be considered when interpreting the results of a sideto-side comparison. First, there may be natural BMD differences between the dominant and nondominant limbs. In the upper limbs of healthy men never involved in unilateral activities, the BMDs in the dominant side are on average
J. Leppa¨la¨ et al.: Patellofemoral Pain Syndrome and BMD
551
Table 2. BMD (G/CM2) in the affected and unaffected lower extremities of 40 patients with a chronic PFPSa Anatomic site
Affected
Unaffected
Difference
% Difference
P valueb
Femoral neck
0.964 ± 0.078 (0.925–1.003) 1.028 ± 0.077 (0.984–1.072) 1.337 ± 0.121 (1.276–1.398) 1.131 ± 0.092 (1.085–1.177) 1.009 ± 0.098 (0.960–1.058) 0.675 ± 0.060 (0.645–0.706)
0.974 ± 0.077 (0.935–1.013) 1.039 ± 0.085 (0.997–1.082) 1.387 ± 0.131 (1.321–1.452) 1.165 ± 0.104 (1.113–1.217) 1.030 ± 0.104 (0.978–1.082) 0.676 ± 0.060 (0.646–0.706)
−0.010 ± 0.023 (−0.024–0.003) −0.011 ± 0.021 (−0.026–0.005) −0.025 ± 0.038 (−0.053–−0.014) −0.025 ± 0.043 (−0.047 to −0.004) −0.019 ± 0.027 (−0.033–0.006) −0.001 ± 0.001 (−0.012–0.011)
−1.0 (−2.4–0.3) −1.1 (−2.6–0.5) −3.3 (−5.3 to −1.4) −2.5 (−4.7 to −0.4) −1.9 (−3.3 to −0.6) −0.1 (−1.2–1.1)
0.133
Trochanter area Distal femur Patella Proximal tibia Calcaneus a b
0.166 0.002 0.016 0.008 0.852
Mean ± SD (95% confidence interval) The matched, paired t-test
Table 3. The relationship between the continuous-type predictive variables and relative BMD in the distal femur, patella, and proximal tibia of the affected extremitya Distal femur
Patella
Predictors
r
r
Age Height Weight Body mass indexb Duration of symptoms (years) Pain assessmentc Functional scoresd Lysholm Tegner Muscle strength
0.04 −0.24 −0.34 −0.22 −0.08 −0.08
0.800 0.143 0.034 0.176 0.608 0.616
−0.10 −0.13 −0.21 −0.16 −0.15 −0.36
0.536 0.431 0.201 0.342 0.342 0.023
−0.12 −0.33 −0.37 −0.36 −0.36 −0.20
0.455 0.040 0.020 0.296 0.022 0.208
0.24 0.11 0.58
0.129 0.514
