SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Objective

Medial knee osteoarthritis (OA) is characterized by pain and associated with abnormal knee moments during walking. The relationship between knee OA pain and gait changes remains to be clarified, and a better understanding of this link could advance the treatment and prevention of disease progression. This study investigated changes in knee moments during walking following experimental knee pain in healthy volunteers, and whether these changes replicated the joint moments observed in medial knee OA patients.

Methods

In a crossover study, 34 healthy subjects were tested on 3 different days; gait analyses were conducted before, during, and after pain induced by hypertonic saline injections (0.75 ml) into the infrapatellar fat pad. Isotonic saline and sham injections were used as control conditions. Peak moments in frontal and sagittal planes were analyzed. The results were compared with data from 161 medial knee OA patients. The patients were divided into less severe OA and severe OA categories, which was based on radiographic disease severity of the medial compartment.

Results

Experimental knee pain led to reduced peak moments in the frontal and sagittal planes in the healthy subjects, which were similar to the patterns observed in less severe OA patients while walking at the same speed.

Conclusion

In healthy subjects, pain was associated with reductions in knee joint moments during walking in a manner similar to less severe knee OA patients. The experimental model may be used to study mechanically-driven knee OA progression and preventive measures against abnormal joint loading in knee OA.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Knee osteoarthritis (OA) is a major cause of pain and physical disability (1). Clinically, knee OA is defined by the presence of both radiographic joint degeneration and pain (2). Pain is traditionally attributed to tissue damage (3), but there is a disparity between joint degeneration in knee OA, as assessed radiographically by the Kellgren/Lawrence (K/L) scale (4) and pain (5, 6). A complex of different mechanisms seems to be implicated in knee OA, and their individual importance for pain and the degenerative processes of the disease remains to be fully elucidated (1).

Medial tibiofemoral OA is the most common, and it has become apparent that medial knee OA is, at least partly, mechanically driven (7–9) and caused by aberrations in knee biomechanics within the context of systemic susceptibility (7, 10, 11). The fulcrum of the biomechanic factors is joint loading, which in fact is associated with the pathogenesis of medial knee OA (9, 11–15). The knee joint loading during walking is of particular interest because walking is the most common method of human locomotion and causes repetitive joint loading.

The most frequently reported difference in knee moments during walking between medial knee OA patients and healthy subjects is in the external knee adduction moment during the stance phase. The external adduction moment is a measure of the knee varus torque and is determinative of medial knee joint loads (16). The maximum adduction moment is a strong predictor of the presence (17), severity (18), and rate of progression (12) of medial knee OA.

The general observation is that peak adduction moments are increased in medial knee OA patients compared with asymptomatic subjects (14, 17, 19–22), yet radiographic disease severity influences the difference between patients and asymptomatic subjects. Less severe medial knee OA patients (K/L grade ≤2) walk with first peak adduction moment similar to healthy subjects, and lower second peak adduction moment than healthy subjects (14, 18). In contrast, severe medial knee OA patients (K/L grade >2) walk with greater first and second peak adduction moments than both less severe patients and healthy subjects (14, 18). The gait differences between K/L classes may be explained by the positive correlation between frontal plane alignment and knee adduction moments (21, 23) because (mal)alignment is correlated to K/L scores (21). In the sagittal plane, the peak internal sagittal plane moments are reduced among knee OA patients compared with healthy subjects (17, 22, 24–26), which can be interpreted as a load-reducing gait pattern, e.g., to avoid pain.

Pain could possibly explain the differences in joint moments between patients with medial knee OA and healthy subjects. Pain relief has been shown to increase the adduction moments (27–29) and sagittal plane moments (25, 28, 29) in knee OA patients, while others have reported the opposite (30). However, knee joint dynamics may be influenced by other factors such as walking speed (31), muscle weakness (32), and joint effusion (33), which may confound the findings. By consequence, it is difficult to assess the isolated effects of pain on knee joint moments in a patient population. The relationship between knee OA pain and gait changes remains to be clarified and a better understanding of this relationship could advance treatment and prevention of disease progression.

Experimental pain in healthy subjects is advantageous with respect to the analysis of isolated effects of pain on knee joint moments. Recently, experimental pain was induced in the infrapatellar fat pad as a model of anterior knee pain leading to changes in quadriceps muscle coordination (34, 35). The infrapatellar fat pad has nociceptive innervation (36) and is a source of pain in knee OA (37). The structure is intraarticular yet extrasynovial, reducing the risk of intrasynovial infection upon injections. It is not known whether the gait abnormalities observed in medial knee OA are replicated during experimental pain in the infrapatellar fat pad.

We hypothesized that experimental knee pain would lead to reduced knee joint moments during walking, replicating those of less severe medial knee OA patients, i.e., reduced internal sagittal plane extensor and flexor moments and the second peak external adduction moment. Accordingly, this study had 2 aims: to investigate changes in knee joint moments during walking with experimental knee pain induced in the infrapatellar fat pad of healthy volunteers, and to investigate whether these changes replicated the joint moments observed in less severe medial knee OA patients.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Subjects.

Thirty-six healthy subjects (18 men and 18 women) volunteered and gave informed consent. The subjects were recruited via an advertisement on the Internet. The study design and methods were approved by the local ethics committee.

Baseline data from 192 knee OA patients included in an ongoing dietary intervention study, the CAROT trial (Influence of Weight Loss or Exercise on Cartilage in Obese Knee Osteoarthritis Patients: a Randomized Controlled Trial), were used in the current study. Patients with no radiographic evidence of medial knee OA or predominantly lateral OA were excluded. According to the radiographic K/L score in the medial compartment of their most affected knee (based on subjective reporting of symptoms), the patients were divided into less severe (K/L grade ≤2) and severe (K/L grade >2) categories, which is similar to previous studies (14, 18).

Study design.

The study was divided into 2 parts: experimental and comparative. The experimental part included only the healthy subjects and was designed as a randomized crossover trial, with each healthy volunteer tested on 3 days separated by at least 1 week.

On each test day, the subjects performed 3 series of gait trials during 3 conditions: baseline, during experimental pain, and 20 minutes after pain induction. There was a 5-minute pause between baseline and trials during the experimental condition. The 20-minute pause was to ensure that the subjects recovered from the pain. Therefore, measurements were performed at time points 0 minutes, 5 minutes, and 25 minutes at each test day, with the second series of trials performed during the experimental conditions.

The experimental pain model consisted of 3 types of injections given to each healthy subject: a hypertonic saline injection inducing pain, an isotonic saline injection as nonpainful control, and a sham injection for further control. In order to investigate the best concentration to use in future studies, the subjects were misinformed that they would receive 3 injections of saline with different concentrations and that they might experience pain or discomfort following all injections. The order of the injections was allocated using a sex-stratified randomization, with 3 men and 3 women receiving the same order of injection types. At the termination of the study, the true nature of the study was revealed to all participating subjects. For the comparative part of the study, the knee OA patients underwent a series of gait analyses and did not receive any injections.

Injections.

The subjects were blindfolded to prevent disclosure of the sham procedure and were told that the saline solutions could be identified by color, therefore making the blindfold necessary. The injections were bolus injections of 0.75 ml sterile saline into the infrapatellar fat pad of the subjects' right knees. The injections were ultrasound guided to ensure correct placement of the bolus in the middle of the fat pad, which is directly behind the patellar tendon. Injections were directed at 45° with a medial approach in a posterior-lateral direction using a 0.4-mm needle (27 gauge) mounted on a 1-ml syringe. Either a sterile hypertonic saline (5.8%) solution or an isotonic saline (0.9%) solution was injected. In trials with sham injections, no injections were given, but the ultrasound scan was performed and the needle was inserted into the skin of the subjects and moved in a similar manner as during the actual injections.

Pain intensity scoring.

The healthy subjects rated the pain intensities following injections after every gait trial using a 100-mm visual analog scale (VAS) where 0 = no pain and 100 = worst imaginable pain.

Questionnaire.

As part of the baseline measurements in the CAROT trial, all knee OA patients filled in a Knee Injury and Osteoarthritis Outcome Score (KOOS). The KOOS is a patient-administered questionnaire with 5 separate subscales assessing pain, symptoms, activities of daily living, sport and recreation function, and knee-related quality of life. Each item is scored 0–4, and items are summed yielding a total KOOS score and separate subscale scores. The scores are transformed to a 0–100 scale, where 100 represents the best result (i.e., more healthy). For the purpose of the present study, the KOOS pain subscale was extracted.

Gait analyses.

All gait analyses were performed in the same laboratory. Kinematic data were acquired using a 3-dimensional (3-D) motion analysis system (Vicon MX, Oxford, UK) with 6 cameras (MX-F20, Vicon) operating at 100 Hz. Two force platforms (AMTI OR 6-5-1000, Watertown, MA) embedded in the laboratory floor captured ground reaction forces at 1,500 Hz synchronized with the kinematic data.

The 3-D orientation of 7 body segments (pelvis, thighs, legs, and feet) was obtained by tracking marker trajectories, with markers placed bilaterally on the anterior and posterior iliac spines, lateral aspect of the thighs, lateral femoral epicondyles, lateral aspects of the legs, lateral malleoli, calcanea, and second metatarsal heads.

Anthropometric parameters required for estimating the joint center locations were first measured. Subsequently, the markers were placed and a capture of a static anatomic landmark calibration trial was done. After the static calibration trial, the healthy subjects and patients practiced their self-selected walking speeds until they could repeatedly walk comfortably within ±0.1 km/hour. The walking speed was measured by photocells, and a display provided the subjects with visual feedback of the walking speed.

Once walking speed and starting points were determined, the healthy subjects performed 3 series of walking trials according to the study design, and the knee OA patients performed 1 series of walking trials. Each series consisted of 5 acceptable trials defined as being within ±0.1 km/hour of the predetermined self-selected walking speed with successful force platform hits without observable force platform targeting. The same self-selected walking speed (within ±0.1 km/hour) was used for all trials in each of the 3 sequences, and on each of the 3 test days, in order to separate the pain-induced adaptive reductions in the joint moment from changes due to changed walking speed.

Statistical analysis.

Analyses of the healthy subjects were performed on the subjects' right knees. The knee OA patients had their affected knee analyzed; if both knees were affected, the most affected knee was chosen for analysis based on the patients' subjective reports. The analyses focused on the stance phase of the gait cycle. Knee joint kinetics and kinematics were calculated using a common commercially available biomechanic model (Plug-In-Gait, Vicon). Knee joint flexion angles at heel strike, maximum knee flexion angle during early stance, and minimum flexion in late stance were extracted. Peak values of the internal sagittal (net extensor and flexor) and external frontal plane knee joint moments were extracted (Figure 1). Moments were normalized to body mass to allow for interindividual comparisons (Nm/kg × 100). All values were used in the statistical analyses, i.e., no within subject averaging.

thumbnail image

Figure 1. Top, external frontal plane moment during stance phase of walking for the healthy subjects at baseline and during experimental knee pain, less severe knee osteoarthritis (OA), and severe knee OA patients. Positive values indicate net external adduction and negative values indicate net external abduction moment. A and B indicate the first and second peak external adduction moments, respectively, extracted for statistical analysis. Bottom, internal sagittal plane moment during the stance phase of walking for the healthy subjects at baseline and during experimental knee pain, less severe knee OA, and severe knee OA patients. Positive values indicate net internal extensor moment (quadriceps moment), and negative values indicate net internal flexor moment (hamstring and gastrocnemius). C and D indicate the peak internal extensor and flexor moments, respectively, extracted for statistical analysis.

Download figure to PowerPoint

In the experimental part of the study, a longitudinal data model was applied to analyze the multiple repeated-measures on the same subject in the crossover design, using the mixed procedure of the SAS system (SAS Institute, Cary, NC), since it enables multiple covariance structures through the random and repeated statements (38). The analysis focused on the fixed effects analysis of time and injection (with 3 levels for each: 0 minutes, 5 minutes, and 25 minutes for time and hypertonic, isotonic, and sham for injections) analyzing whether there was a time × injection interaction and applying each factor as main factors. The estimates were adjusted for walking speed, sex, body mass index (BMI), age, and the baseline estimates. Any significant time × injection interactions were broken down post hoc by exploring the pairwise differences between injection types at 0, 5, and 25 minutes, respectively.

Data from the knee OA patients were analyzed using the same statistical analysis without the repeated statement, focusing on the fixed effect of K/L grouping (2 levels: K/L grades ≤2 and K/L grades >2) and adjusting for walking speed, sex, BMI, and age.

To compare the effects of experimental pain in healthy subjects with the knee OA groups, the estimates from the second series in the experimental study (i.e., during pain/control/sham) were compared with the knee OA group estimates using 2-sample t-tests. Statistical significance was set at P values less than 0.05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Of the 36 included healthy subjects, 1 male subject withdrew consent during the data collection. The remaining 35 subjects completed the study. Another male subject was excluded during data analysis due to corrupted data files.

Of the 192 knee OA patients included in the CAROT trial, gait analyses were completed on 178 patients. Of these, 161 patients had radiographic evidence of predominantly medial knee OA from standard weight-bearing radiographs. Based on the medial compartment K/L scores, there were 92 patients classified as less severe patients (K/L grade ≤2) and 69 patients classified as severe patients (K/L grade >2). Characteristics of the healthy subjects and the knee OA patients are shown in Table 1.

Table 1. Characteristics of healthy subjects and knee OA patients*
CharacteristicMean ± SD (range)P vs. less severe patientsP vs. severe patients
  • *

    OA = osteoarthritis; BMI = body mass index; K/L = Kellgren/Lawrence (range 0–4); KOOS = Knee Injury and Osteoarthritis Outcome Score (range 0–100).

  • Based on 2-sample t-tests.

Healthy subjects (n = 34)   
 Female/male, no.18/16
 Age, years25.6 ± 5.5 (19.0–46.0)< 0.0001< 0.0001
 Height, meters1.75 ± 0.11 (1.54–1.98)< 0.0001< 0.001
 Body mass, kg68.9 ± 13 (48.7–101.4)< 0.0001< 0.0001
 BMI, kg/m222.4 ± 2.3 (18.7–29.8)< 0.0001< 0.0001
 Self-selected walking speed, km/hour4.9 ± 3.8 (4.6–5.4)0.360.12
 Medial compartment K/L score
 KOOS pain score
Less severe knee OA patients (n = 92)   
 Female/male, no.80/12
 Age, years62.1 ± 5.9 (52.0–75.0)0.01
 Height, meters1.66 ± 0.07 (1.49–1.84)0.27
 Body mass, kg98.5 ± 12.2 (73.6–137.2)0.01
 BMI, kg/m236.3 ± 4.0 (30.1–48.7)0.02
 Self-selected walking speed, km/hour4.0 ± 0.7 (2.2–5.4)0.11
 Medial compartment K/L score1.7 ± 0.5 (1.0–2.0)< 0.0001
 KOOS pain score57.9 ± 15.8 (13.9–88.9)0.09
Severe knee OA patients (n = 69)   
 Female/male, no.56/14
 Age, years63.9 ± 7.1 (50.0–78.0)
 Height, meters1.65 ± 0.09 (1.60–1.91)
 Body mass, kg104.7 ± 15.6 (77.2–144.2)
 BMI, kg/m237.9 ± 4.5 (31.3–51.6)
 Self-selected walking speed, km/hour4.0 ± 0.7 (2.0–5.8)
 Medial compartment K/L score3.4 ± 0.5 (3.0–4.0)
 KOOS pain score53.0 ± 15.2 (11.1–100.0)

Experimental knee pain.

There was significant time × injection interaction in the VAS pain intensity scores (P < 0.001). Following hypertonic saline injections, the mean pain intensity was 25.8 mm (95% confidence interval [95% CI] 24.9–26.8 mm, P < 0.0001). Following isotonic saline injections, the mean pain intensity was 0.7 mm (95% CI 0.0–1.7 mm, P = 0.14), and following sham injections the mean pain intensity was 0.3 mm (95% CI 0.0–1.3 mm, P < 0.50).

KOOS pain subscale.

The less severe patients reported an average KOOS pain score of 57.9 (95% CI 50.0–57.3), which was significantly higher (i.e., less painful) than the severe patients who on average reported 53.0 (95% CI 49.4–56.7).

Gait analyses (experimental pain study).

There were significant time × injection interactions in both the first and second peak adduction moments (P = 0.0097 and P = 0.0007, respectively) (Figure 2). During pain, the first peak adduction moments were significantly lower compared with both control (P < 0.0001) and sham conditions (P = 0.0002). During pain, the second peak adduction moment was significantly reduced compared with the control (P < 0.0001) and sham conditions (P < 0.0001).

thumbnail image

Figure 2. Mean (± SEM) of the first (top) and second (bottom) peak external adduction moments during walking (see Figure 1 for definitions) before (0 minutes), during (5 minutes), and after (20 minutes) sham injections, isotonic saline injections, and hypertonic saline injections. a = pain condition (hypertonic saline) significantly different from sham (P < 0.0001) and control (isotonic saline injections) (P < 0.0001). b = control injections of isotonic saline were significantly different from sham (P < 0.0001) and hypertonic saline injections (second peak adduction moment only) (P < 0.0001).

Download figure to PowerPoint

There were significant time × injection interactions in the peak extensor and flexor moments (P < 0.0001 and P = 0.011, respectively) (Figure 3). During pain, the peak extensor moment was significantly reduced compared with both control (P < 0.0001) and sham conditions (P < 0.0001). Also the peak flexor moment was significantly reduced during pain compared with both control (P < 0.0001) and sham (P < 0.0001).

thumbnail image

Figure 3. Mean (± SEM) of the sagittal plane peak internal extensor (top) and flexor (bottom) moments during walking (see Figure 1 for definitions), before (0 minutes), during (5 minutes), and after (20 minutes) sham injections, isotonic saline injections and hypertonic saline injections. The flexor moment is per definition negative and therefore the y-axis scale (bottom) is reversed for visual purposes. a = pain condition (hypertonic saline) significantly different from sham (P < 0.0001) and control (isotonic saline injections) (P < 0.0001). b = pain condition (hypertonic saline) significantly different from control condition (isotonic saline) (P < 0.0001).

Download figure to PowerPoint

There was no significant time × injection interaction in heel strike knee joint angle (P = 0.20), but there was a significant time × injection interaction in maximum and minimum knee flexion angles (P < 0.0001 and P = 0.016, respectively). During pain the maximum flexion angle was reduced by 1.5° (SEM 0.2°) compared with both control and sham conditions (P < 0.0001 for both), and the minimum flexion angle was increased by 0.7° (SEM 0.14°) compared with control and sham conditions (P < 0.0001 for both).

Gait analyses (comparative study).

Compared with the severe knee OA patients, the less severe patients walked with significantly lower first and second peak adduction moments (P < 0.0001 for both) (Figures 1 and 4). During all experimental conditions, the healthy subjects walked with greater first peak adduction moments compared with the less severe patients, although the peak moment was reduced during pain (P < 0.007), and lower second peak adduction moments than the severe patients (P < 0.0001) (Figures 1 and 4). While the healthy subjects' second peak adduction moments during control and sham conditions were higher than among less severe patients (P < 0.042), the healthy subjects walked with second peak adduction moments similar to the less severe patients during pain (P = 0.58) (Figure 4).

thumbnail image

Figure 4. Mean (± SEM) of the first (top) and second (bottom) peak external adduction moments during walking during the experimental conditions in the healthy subjects (time = 5 minutes) for effects of sham, control, and pain compared with estimates from less severe (Kellgren/Lawrence [K/L] ≤2) and severe (K/L >2) knee osteoarthritis (OA) patients. * = significantly different from less severe knee OA patients (P > 0.04). ** = significantly different from severe knee OA patients (P < 0.0001). ¤ = significantly different from sham and control (P > 0.0001).

Download figure to PowerPoint

There were no differences in peak extensor moments between knee OA groups (P = 0.58). The healthy subjects walked with higher peak extensor moments than both knee OA groups during all experimental conditions (P < 0.000), although the peak moment decreased during pain (Figure 5). The severe knee OA patients walked with significantly lower peak flexor moments compared with the less severe patients (P < 0.0001) and the healthy subjects during all experimental conditions (P < 0.004). During none of the experimental conditions did the difference between healthy subjects and less severe patients reach statistical significance (P > 0.09), although the peak flexor moments decreased during pain.

thumbnail image

Figure 5. Mean (± SEM) of the sagittal plane peak internal extensor (top) and flexor (bottom) moments during walking during the experimental conditions in the healthy subjects (time: 5 minutes) for effects of sham, control, and pain compared with estimates from less severe (Kellgren/Lawrence [K/L] ≤2) and severe (K/L >2) knee osteoarthritis (OA) patients. Note reversed y-axis scale (bottom) for visual purposes. * = significantly different from less severe knee OA patients (P < 0.0001). ** = significantly different from severe knee OA patients (P < 0.004). ¤ = significantly different from sham and control (P > 0.0001).

Download figure to PowerPoint

There were no differences between OA severity groups in any of the knee joint angles (P > 0.20), yet the patients walked with 2.5° (SEM 0.8°) more flexed knees at heel strike and 4.2° (SEM 0.9°) more minimal late stance flexion (P < 0.0001) compared with the healthy subjects. During pain in the healthy subjects, the early stance flexion angle was reduced, but not to a level that was different from either of the patient groups (P > 0.34). The late stance flexion angle was increased, but not to a level comparable with either OA groups (P > 0.0013).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

This study shows that experimental pain induced in the infrapatellar fat pad significantly lowered the frontal and sagittal plane knee joint moments in a manner similar to those observed in less severe medial knee OA patients, while walking at the same speed. Therefore, the a priori hypotheses of this study were supported.

The gait adaptation to pain reported here is similar to that reported in other studies of medial knee OA (14, 17, 18, 22, 24, 25). In general, these studies suggest an overall unloading of the medial compartment and support the notion that pain is a protective mechanism. However, the fact that the injection in the infrapatellar fat pad produced a gait adaptation in healthy subjects that was similar to patients with medial knee OA suggests that the unloading stimulus does not necessarily come from mechanical pressure on the medial compartment cartilage, since there are no pain receptors in articular cartilage. Therefore, the common features of the gait adaptations related to the infrapatellar fat pad pain suggest that inflammation or pain receptors in this region could influence the adaptations seen in medial compartment knee OA patients.

The study design with both control and sham injections in healthy volunteers strengthen these observations since placebo effects were controlled. It is useful to examine potential adaptive mechanisms in this population. There were differences in the knee joint kinematics between the patients and baseline measurements of the healthy subjects, but the experimental pain only led to subtle changes in the knee joint angles, which cannot account for the resembling joint moments caused by pain. The changes cannot be explained by painful muscle contractions, direct stimulation of muscle afferents, or motor neurons since no intramuscular injections were done. Moreover, the results are not explained by changes in walking speed because this was kept constant throughout the experimental study. The reductions in external knee adduction moments may be brought about by increased trunk motions and foot progression angles (39–42). Unfortunately, we did not include measurements of trunk lean or foot progression angles in this study. Furthermore, the changes may be caused by changes in muscle coordination or motor neuron recruitment, which have been shown following experimental infrapatellar fat pad pain (35, 43).

It has been shown that peak adduction moments differ according to radiographic disease severity (14, 18); severe patients walk with greater first and second peak adduction moments than less severe patients. The present study corroborates this and adds to this knowledge that less severe patients walk with lower peak adduction moments compared with healthy subjects, and severe patients walk with greater peak adduction moments than both less severe patients and healthy subjects. The differences between the healthy subjects and the less severe patients were generally small. Yet, as little as a 1% increase in the adduction moment has been shown to increase the risk of radiographic disease progression by 6.46 times (12), wherefore pain modulations may have profound consequences. Although the severe patients on average had more pain measured by the KOOS, pain is the predominating feature of less severe knee OA, whereas structural changes are more pronounced in severe knee OA. The relatively larger adduction moments in severe patients compared with less severe patients could be caused by differences in mechanical axis alignment, which is correlated to both radiographic disease severity and adduction moments (21, 23). This suggests that pain may be of less importance for disease progression in severe patients compared with mechanical factors, such as alignment. This is supported by the fact that one severe patient reported no pain on the KOOS (Table 1).

While these data do not warrant conclusions about pain relief, it may be speculated that pain relief may be detrimental to the knee joint by increasing joint loads. In fact, the peak adduction moments and joint loads do increase with pain relief (25, 27–29). Moreover, higher adduction moments have been shown to increase the risk of knee OA progression in a study where the patients were allowed to take pain relieving medication (12). Therefore, pain relief may lead to a more rapid disease progression.

There are limitations to this study. Experimental pain is acute and transient and does not necessarily reproduce knee OA pain, but for obvious reasons it is not possible to induce chronic pain experimentally. Nevertheless, the infrapatellar fat pad has nociceptive innervation (36), and is a source of pain in knee OA (37). The healthy subjects were significantly younger than the knee OA patients, and the patients were obese (BMI >30) in contrast to the healthy volunteers, but there are no relationships between either age or BMI and the peak knee moments used in the present study (44, 45). Asymptomatic elderly people may have radiographic or other changes that may confound any findings related to the experimental pain. Therefore, studying subjects who are as “healthy as possible” is advantageous when the isolated effects of pain are to be assessed, and the present study design allows for such assessment. While the knee OA pain quality may not be replicated, the knee joint moments were.

The reduced moments caused by the experimental pain support the notion of pain as a protective mechanism. A future challenge lies in identifying factor(s) inducing the increase in adduction moment among less severe patients. Pain relief might be such a factor, and pain treatments may be sought that target the pain and, at the same time, prevent increased adduction moment. The present data suggest that less severe patients will benefit the most from such treatments in terms of decelerated disease progression. At present, multidisciplinary efforts including pain relief (46) and noninvasive interventions such as weight loss (47), devices targeting knee mechanics (48), and gait modifications such as reduced walking speed, trunk sway, or toe-out walking (18, 40, 49) seem to be the more promising, although these modalities have not been tested longitudinally against disease progression.

In conclusion, this study shows that knee pain may be induced in healthy volunteers to significantly reduce the adduction and sagittal plane knee joint moments during walking in a manner similar to patterns observed in patients with less severe knee OA. The experimental model may be used to study knee OA pathomechanics and possible preventive measures against abnormal joint loading in knee OA. It is suggested that pain management regimes be tested on the basis of their influence on knee OA progression.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Henriksen had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Henriksen, Graven-Nielsen, Bliddal.

Acquisition of data. Henriksen, Aaboe.

Analysis and interpretation of data. Henriksen, Graven-Nielsen, Andriacchi, Bliddal.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

The authors thank Søren Torp-Pedersen, ultrasound specialist, and Mikael Boesen, radiologist, from the imaging section of The Parker Institute for their help during the planning of this study.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES