To examine the effect of high-speed power training (HSPT) on muscle performance, mobility-based function, and pain in older adults with knee osteoarthritis.
To examine the effect of high-speed power training (HSPT) on muscle performance, mobility-based function, and pain in older adults with knee osteoarthritis.
Thirty-three participants (mean ± SD age 67.6 ± 6.8 years) were randomized to HSPT (n = 12), slow-speed strength training (SSST; n = 10), or control (CON; n = 11) for a 12-week intervention. HSPT performed 3 sets of 12–14 repetitions at 40% of the 1-repetition maximum (1RM) “as fast as possible,” SSST performed 3 sets of 8–10 repetitions at 80% of the 1RM slowly, and CON performed stretching and warm-up exercises. Outcome measures included leg press (LP) 1RM and LP peak power (PP) from 40–90% of the 1RM and the corresponding PP velocity (PPV) and PP force; 400-meter walk, Berg Balance Scale, and timed chair rise; and self-reported function and pain using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC). Analysis of variance models were used to compare changes from baseline to 12 weeks. Statistical significance was accepted at P < 0.05.
LP PP improved in both HSPT and SSST compared to CON (P = 0.04). LP PPV improved only in HSPT (P = 0.01). There were also improvements in timed chair rise (P = 0.002), WOMAC function (P = 0.004), and WOMAC pain (P = 0.02) across all of the groups.
HSPT was effective at improving function and pain, but no more so than either SSST or CON. Because HSPT improved multiple muscle performance measures (strength, power, and speed), it is a more effective resistance training protocol than SSST and may increase safety in this population, especially when high-speed movements are required during daily tasks.
There is no cure for osteoarthritis (OA) (1); therefore, treatment modalities to improve quality of life for older adults with this disease have focused on improving measures of muscle strength, function, and reducing pain. While a comprehensive program to treat knee OA may include a variety of exercises (aerobic, resistance training, flexibility), a principal component of rehabilitation to reduce the symptoms of the disease has been resistance training (RT) with a strengthening component (2–7). Current recommendations for RT in healthy adults encourage strength-enhancing contractions at moderate to high resistances performed slowly (50–80% of maximal strength) (8). Surprisingly, RT protocols prescribed for patients with knee OA are often identical to those performed by healthy adults, despite considerable differences in levels of impairment and functional ability.
Difficulty arises when attempting to choose the most effective RT protocol due to inconsistent findings and controversy in the literature regarding strength-enhancing RT. For example, strengthening of the quadriceps in patients with knee OA has been shown to improve measures of both pain and function in some (2, 6), but not all, studies (4, 9). Moreover, some evidence suggests that greater muscle strength surrounding malaligned joints or those with increased laxity increased the odds for progression of OA (10). The effect of muscle strength on OA may also differ according to joint status or whether the joint is considered healthy (incident arthritis) (11, 12) or arthritic (progressive) (13). While RT with a strengthening component is currently the preferred recommendation for patients with knee OA to improve measures of function and reduce pain, “muscle-enhancing interventions” (10) that emphasize speed of movement may be preferable, and possibly safer, alternatives.
We believe it is timely to question the traditional assumptions of improved strength translating to better function in patients with knee OA. In the older adult population, muscle power (force × velocity) has been identified as an important muscle performance characteristic (14–17), and a recent meta-analysis has demonstrated that RT programs using high-speed movements (i.e., power training) were more effective than traditional slow-speed RT at improving muscle power and function (18). A critical component of muscle power development is the speed at which force is applied. Our laboratory has shown that high-speed power training (HSPT) at low external resistances (40% of the 1-repetition maximum [1RM]) significantly improved muscle performance (peak power [PP] and PP velocity) (19) and high-speed function (20) compared to traditional slow-speed RT at high external resistances (80% of the 1RM) in older adults.
In addition, high-speed contractions may improve the metabolic and regenerative processes within the joints of individuals with knee OA. One study showed that high-speed contractions (180°/second) improved oxygen partial pressure compared to slow-speed contractions (60°/second) in arthritic knees (21). Because increased oxygen partial pressure improves blood flow and diffusion within the joint (21), a speculative mechanism can be proposed as to why HSPT might improve measures of muscle performance, function, and pain in patients with knee OA. To our knowledge, no studies have examined a regimen of HSPT in older adults with knee OA.
The purpose of this study was to examine the effect of HSPT on muscle performance, mobility-based functional tasks, and pain in older men and women with knee OA. We hypothesized that HSPT would improve power and the velocity component of power and improve measures of function and pain compared to slow-speed RT.
No studies in the literature have examined high-speed power training and its impact on functional measures and pain in older adults with knee osteoarthritis (OA).
Twelve weeks of either high-speed power training or slow-speed strength training both improve strength similarly in older adults with knee OA, but muscle power demonstrated greater improvement with high-speed training.
Only high-speed power training improved muscle speed, which could have significant implications for safety in this population when high-speed movements are required for activities of daily living.
Despite improvements in muscle performance with high-speed power training, these improvements did not necessarily translate into improved function. High-speed power training was effective at improving some measures of function and pain, but no more than that of slow-speed strength training or an active control group.
Eligible participants were ages 55 years or older with physician-diagnosed evidence of knee OA. Study eligibility required meeting criteria of the American College of Rheumatology (ACR) clinical classification of knee OA (22), which consisted of knee pain and inclusion of 3 of the following 6 criteria: age >50 years, crepitus on active motion, less than 30 minutes of stiffness upon waking in the morning, bony tenderness, bony enlargement, and no palpable warmth of synovium. A qualifying level of pain or functional deficit as noted on the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain or WOMAC function scale (pain: 1 response of at least “moderate” or 2 responses of “minimal”; function: 2 responses of at least “moderate” or 4 responses of “mild”) was also required for inclusion in the study. Exclusion criteria consisted of history of heart disease, severe visual impairment, presence of neurologic disease, pulmonary disease requiring the use of oxygen, uncontrolled hypertension, hip fracture or lower extremity joint replacement in the past 6 months, and current participation in structured exercise. A study physician determined medical eligibility for all of the participants.
The study was conducted as a randomized, single-blind, controlled, intervention trial. A random-numbers generator was used to assign participants to the experimental conditions. Concealment of the treatments was facilitated by maintaining blinding of the lead investigator (SPS), who was responsible for all of the evaluations at baseline and 12 weeks. The research assistants had knowledge of the participant group assignments, and each participant agreed to randomization at the outset of the study.
Fifty-seven individuals were contacted to participate in the study. Once eligibility was determined, 45 individuals were randomized to 1 of 3 groups: HSPT, slow-speed strength training (SSST), and control (CON) (Figure 1). Twelve participants dropped out during baseline before training had begun. Four participants withdrew during the intervention, 1 in HSPT, 2 in SSST, and 1 in CON, 3 of which were determined to be study related (HSPT: n = 1 and SSST: n = 2). Using an intent-to-treat design, the data from 33 participants (HSPT: n = 12, SSST: n = 10, and CON: n = 11) were analyzed in this study. Twenty of these 33 participants were part of a previous study on muscle performance (20). The muscle performance, function, and pain data reported in the present study, however, are unique to these 33 participants and have not been reported elsewhere. The study was approved by the University of Missouri Institutional Review Board and written consent was obtained from all of the participants.
The study evaluated the effects of 12 weeks of explosive HSPT on muscle performance, function, and pain compared with traditional SSST. Outcome measures included muscle performance: leg press (LP) 1RM and LP PP, velocity at PP (PPV), and force at PP (PPF) from 40–90% of the 1RM. Mobility-based measures of function consisted of the 400-meter walk (400-m W), the Berg Balance Scale (BBS), and the timed chair rise (TCR). The WOMAC was used to assess self-reported function and pain.
Participants completed 2 weeks of baseline assessments. On visit 1, body mass and height were measured using a platform scale and stadiometer. Cognitive function was assessed by the Mini-Mental State Examination (23), and the Geriatric Depression Scale was administered to assess depression over the previous week (24). Number of falls in the past year and daily medications were assessed via questionnaire. Participants also completed the WOMAC to determine initial study eligibility. On visit 2, a physical therapist evaluated whether the participant met the ACR criteria for clinical classification of knee OA. If the WOMAC criteria and ACR criteria were met, participants were scheduled for a physical with the study physician to determine medical eligibility. To characterize the severity of knee OA, all of the participants underwent an anteroposterior bilateral weight-bearing radiograph of both knees with the participant standing with both toes pointing straight ahead, both knees fully extended, and weight equally distributed on both feet. Both the left and right knees were graded from 0–4 using the Kellgren/Lawrence scale. On visits 3 and 4, muscle performance and functional measures were obtained. The following week, muscle performance and functional measures were repeated to establish reliability. If baseline 1RM measurements deviated by more than 10% in repeated attempts, a third measure was obtained. At the end of baseline testing, participants were randomized to treatment. Following the 12-week RT intervention, posttraining muscle performance, function, and pain measures were obtained.
Participants exercised 3 times per week for 12 weeks using computer-interfaced Keiser a420 pneumatic LP and knee extension (KE) RT equipment. HSPT performed 3 sets of 12–14 repetitions at 40% of the 1RM, whereas SSST performed 3 sets of 8–10 repetitions at 80% of the 1RM. HSPT performed an explosive movement at high speed during the concentric phase of the contraction, paused for 1 second, and performed the eccentric portion of the contraction over 2 seconds. SSST performed each contraction at slow velocity (2 seconds for the concentric phase of the contraction), paused for 1 second, and performed the eccentric portion of the contraction over 2 seconds. Repetition number was higher in the HSPT group to more closely equate work performed between the groups and to remain consistent with RT guidelines (8). CON met 3 times a week for approximately 15–20 minutes of assisted stretching exercises using 12 back, trunk, and lower extremity stretches. Each stretch was initiated and held for 30 seconds by a trained physical therapy student research assistant. Following the stretching protocol, a 5-minute warm-up on a cycle ergometer was performed. HSPT and SSST also particpated in the stretching and cycling protocol.
LP and seated KE 1RM were tested using Keiser pneumatic RT equipment fitted with a420 electronics. The a420 equipment captured measures of PP, PPV, and PPF during the concentric portion of each contraction by sampling the system pressure at 400 Hz and making calculations based on an appropriate algorithm.
The seat of the recumbent LP and seated KE machines were positioned to ensure the hip and knee joints were at 90–100 degrees of flexion. The 1RM was obtained by progressively increasing resistance until the subject was no longer able to push out 1 repetition successfully. The Borg Scale (25) was used to assist in evaluating when 1RM was reached. Peak muscle power was obtained at 40%, 50%, 60%, 70%, 80%, and 90% of the 1RM approximately 30 minutes after 1RM testing (26, 27). Participants were instructed to exert “as fast as possible” at each relative percentage of the 1RM. Three attempts were made at each resistance and the greatest PP output obtained at each resistance was used in the analysis. Therefore, there was a total of 6 PP measures obtained for analysis from 40–90% of the 1RM at both baseline and posttraining. The corresponding PPV and PPF for the PP value at each resistance were also obtained. The 1RM was measured biweekly in HSPT and SSST only and relative training intensity was adjusted accordingly to ensure adequate overload during training. Only LP data are included in this report.
Participants were instructed to walk at a pace they could maintain without overexerting themselves until they had completed the 400-m W or could no longer continue. The 400-m W was administered using a premeasured distance of 400 meters. Heart rate was monitored continually. The 400-m W has strong face validity as a mobility measure and test–retest reliability is high (28).
The BBS was used to measure the participants' ability to maintain balance. The BBS utilized 14 tasks involving sitting to standing, transferring, reaching, bending forward, picking up objects from the floor, single-leg and tandem standing, turning, and stepping that were scored on a 0–4 scale and summed to obtain an aggregate balance score (range 0–56). This test is widely used in the literature to assess balance and has been shown to be valid and reliable in older adults (29, 30).
Volunteers placed their folded arms across their chests and stood up from a sitting position 5 times as quickly as they could. Subjects completed the activity using an armless chair (seat height 43.18 cm). TCR was reported as the number of repetitions per minute calculated from the time required to complete 5 chair stands. The TCR is part of a physical performance battery that has been shown to be significantly related to future disability in older adults (31).
Self-reported function was assessed by the WOMAC function subscale, a 17-item Likert scale questionnaire (where 0 = none, 1 = mild, 2 = moderate, 3 = severe, and 4 = extreme) developed specifically for the study of OA. Self-reported function was represented by summation of the component item scores (range 0–68; higher scores indicate greater functional loss). The reliability and construct validity of the function subscale have been established in OA patients (32).
Pain was assessed by the WOMAC pain subscale, a 5-item Likert scale questionnaire (where 0 = none, 1 = mild, 2 = moderate, 3 = severe, and 4 = extreme) developed specifically for the study of OA. Pain was represented by summation of the component item scores (range 0–20; higher scores indicate greater levels of pain). The reliability and construct validity of the pain subscale have been established in OA patients (32).
Descriptive statistics were run on all of the variables. Associations among variables of age, sex, height, and weight with muscle performance and functional outcomes were evaluated using Pearson's correlation coefficient. When significant associations were found, those variables were used as covariates in all analysis of variance (ANOVA) models. If the variable was not a significant variable in the full ANOVA model, a reduced model was run without the covariate. In all repeated-measures ANOVA models, when equal variances were not observed using Mauchly's test of sphericity, corrections to the F ratio were applied using Greenhouse-Geyser epsilon values. Statistical significance for all tests was accepted at P values less than 0.05. Data are reported as the mean ± SD.
To evaluate baseline differences in subject characteristics, a one-way ANOVA (continuous variables) or chi-square test (categorical variables) was run. To evaluate baseline differences among the groups in 1RM, a one-way ANOVA was performed. To evaluate baseline differences among the groups in muscle performance (PP, PPV, and PPF) from 40–90% of the 1RM, a 3 × 6 (group by condition) repeated-measures ANOVA was run for each muscle performance measure. To evaluate baseline differences among the groups in mobility-based measures of function, WOMAC function, and WOMAC pain, a one-way ANOVA was performed.
To evaluate differences among the groups in 1RM from baseline to posttraining, a 3 × 2 (group by time) repeated-measures ANOVA was run. To evaluate differences among the groups in muscle performance from 40–90% of the 1RM, the change scores in PP, PPV, and PPF with training were calculated (posttraining value minus baseline value) and a 3 × 6 (group by condition) repeated-measures ANOVA was run for each measure. To evaluate differences among the groups in functional measures from baseline to posttraining, a 3 × 2 (group by time) repeated-measures ANOVA was run. If significant group main effects or interactions were found, post hoc testing on the change score using Tukey's honest significant difference test was performed.
There were no differences among the groups in age, sex, height, weight, body mass index, depression, cognitive function, or number of daily medications taken (Table 1). Fifty eligible knees had Kellgren/Lawrence scale grades of 0 (n = 2 [4%]), 1 (n = 22 [44%]), 3 (n = 16 [32%]), and 4 (n = 2 [4%]), and there were no differences in scores among the groups (all P > 0.05) (Table 1). Kellgren/Lawrence scores indicate that the severity of knee OA was mild in this study. There were no differences among the groups in baseline LP 1RM, LP PPV, or LP PPF from 40–90% of the 1RM (all P > 0.05). There was no difference among the groups in baseline 400-m W, BBS, TCR, WOMAC function, or WOMAC pain (all P > 0.05) (Table 2). Therefore, all of the groups were similar in all baseline characteristics at the start of training.
|HSPT (n = 12)||SSST (n = 10)||CON (n = 11)||P|
|Age, years||66.9 ± 4.9||65.9 ± 8.3||68.4 ± 8.1||0.73|
|Height, cm||172.2 ± 10.4||169.0 ± 10.8||166.0 ± 11.4||0.40|
|Weight, kg||84.4 ± 18.8||94.5 ± 23.7||84.7 ± 18.1||0.44|
|BMI, kg/m2||28.4 ± 5.7||33.1 ± 8.9||30.8 ± 6.8||0.31|
|GDS||5.6 ± 2.6||6.5 ± 4.5||5.3 ± 2.8||0.69|
|MMSE||28.8 ± 1.2||28.6 ± 1.2||28.8 ± 0.9||0.87|
|No. of prescribed medications||5.1 ± 5.5||4.4 ± 2.5||5.8 ± 3.6||0.74|
|K/L score (right)||1.5 ± 0.8||1.9 ± 0.9||2.0 ± 0.9||0.55|
|K/L score (left)||1.6 ± 0.5||1.3 ± 1.0||1.5 ± 1.0||0.67|
|HSPT (n = 12)||SSST (n = 10)||CON (n = 11)|
|400-meter self-paced walk, seconds|
|Pre||333 ± 19.7||366 ± 75.7||363 ± 89.7|
|Post||331 ± 31.7||374 ± 71.8||371 ± 65.8|
|Berg Balance Scale (range 0–56)|
|Pre||51.7 ± 3.4||53.1 ± 1.6||52.1 ± 4.1|
|Post||52.9 ± 3.1||52.9 ± 2.6||50.5 ± 4.4|
|Timed chair rise, seconds|
|Pre||17.3 ± 4.4||16.7 ± 3.7||16.8 ± 4.2|
|Post||15.1 ± 2.9||14.4 ± 2.2||15.1 ± 2.6|
|WOMAC function (range 0–68)|
|Pre||41.4 ± 9.7||41.9 ± 9.8||39.5 ± 11.0|
|Post||26.5 ± 6.1||33.5 ± 12.6||34.8 ± 13.9|
|WOMAC pain (range 0–20)|
|Pre||11.5 ± 2.8||12.2 ± 3.4||11.7 ± 2.6|
|Post||9.3 ± 3.2||10.4 ± 2.8||10.2 ± 2.5|
There was a significant improvement in LP 1RM with training (P < 0.001). Mean ± SD HSPT and SSST improved from 1,397 ± 461 N to 1,781 ± 539 N (27%) and from 1,453 ± 543 N to 1,788 ± 683 N (23%), respectively, compared to CON (from 1,270 ± 351 N to 1,350 ± 312 N [6%]). Post hoc tests showed that the changes in 1RM in HSPT and SSST were greater than in CON (both P = 0.02), but were not different from each other (P = 0.99). There was a significant improvement in LP PP, LP PPV, and LP PPF with training (all P ≤ 0.05). Post hoc tests showed that LP PP in HSPT was greater than CON from 40–70% of the 1RM (all P ≤ 0.05), whereas SSST was greater than CON only at 40% and 60% of the 1RM (all P ≤ 0.04) (Figure 2). LP PP in HSPT was greater than SSST at 50% of the 1RM (P ≤ 0.05). In addition, LP PPV in HSPT was greater than CON at 40–50% of the 1RM (all P ≤ 0.02), while SSST was not significantly greater than CON at any time point (all P ≤ 0.05) (Figure 2). Finally, LP PPF in HSPT was greater than CON at 40% and 70–90% of the 1RM (all P ≤ 0.04), while SSST was greater than CON only at 40% and 80–90% of the 1RM (all P ≤ 0.02) (Figure 2). These findings indicate that HSPT exerted a somewhat broader training effect than SSST with regard to power, velocity, and force across a range of external resistances typically encountered in daily activities.
There were no significant group main effects or group by time interaction for 400-m W, BBS, TCR, WOMAC function, or WOMAC pain (all P > 0.05). When collapsing across the groups, there were significant improvements in TCR, WOMAC function, and WOMAC pain (all P < 0.05) (Table 1). Therefore, despite overall improvements in some functional measures and pain, different modes of training did not impact these changes.
Muscle power has previously been examined as a predictor of performance (14, 17, 33) and as an outcome variable in RT and power training studies in older adults (26, 34–38). This is the first study we are aware of that has explored the effects of power training in older adults with knee OA. The major finding from this study was that HSPT and SSST were both successful at improving muscle performance compared to CON in patients with knee OA. HSPT demonstrated greater improvements in power than SSST and also demonstrated improvements in velocity not seen in SSST. Interestingly, the improvements in power in both HSPT and SSST and velocity in HSPT did not have a greater effect on function than those observed in CON.
In this study, we chose to examine muscle performance across a range of external resistances (40–90% of the 1RM) that might be encountered in everyday living as opposed to simply reporting one “peak” measure from this range. This is important because different functional tasks may require power obtained at higher or lower velocities. Moving the body quickly (e.g., crossing a busy intersection) may require a different type of power, with a greater velocity and lower force component compared to moving the body upward (e.g., climbing a flight of stairs), which may require a lower velocity and greater force component. Therefore, it is important to examine how different RT programs impact power and its components (velocity and force) across a wide range of external resistances.
Our findings indicate that older adults with knee OA improved muscle performance measures similarly to community-dwelling older adults without a diagnosis of knee OA (19). In the present study, improvements in muscle power were greater in HSPT compared with SSST. HSPT improved power from 40–70% of the 1RM compared to CON, whereas SSST improved only at 40% and 60% of the 1RM compared to CON. It was also found that HSPT demonstrated improvement in velocity at relative intensities close to the training intensity (40–50% of the 1RM), but this did not occur with slow-speed RT. There were only 2 time points at which velocity did increase, but it is important to note that strength had improved significantly in both groups after 12 weeks of RT and participants were exerting against a resistance relative to the posttraining 1RM, 23–27% higher than at baseline. Our findings suggest that speed is trainable in older adults with knee OA, which could have significant implications for safety when high-speed movements need to be employed. Finally, increases in 1RM strength were similar in HSPT and SSST. Therefore, improvements in power and velocity with HSPT compared to SSST do not come at the expense of other muscle performance measures, making HSPT an appealing option for older adults with knee OA.
Currently there is a paucity of data on the mechanisms responsible for the improvements observed in muscle performance with power training. It is believed that neural factors (i.e., changes in agonist–antagonist voluntary neural activation) may demonstrate greater contributions to muscle performance improvements than muscle hypertrophy with explosive power training compared to heavy RT (39). Also, although increased activation of type II motor units has been observed during acute bouts of explosive resistance loading (40), it is not currently known whether type II fiber type adaptations result from prolonged power training regimens.
An interesting finding from our study was that improvements occurred in some, but not all, functional tasks, and improvements in CON were similar to those in the training groups. Despite large increases in several muscle performance measures in our training groups (strength, power, velocity), there was not a clear positive transfer to function. It may be that greater improvements in muscle power and velocity observed in HSPT simply did not transfer to typical mobility-based functional tasks; therefore, positive transfer of power and speed to function may require novel functional tasks performed at higher speeds. In a previous study, we did find that HSPT demonstrated a transfer of power and velocity to function when the high-speed functional movements (braking speed on a driving simulator) were similar to the nature of the high-speed training (LP exercise) (20). A limitation of the present study was that our choice of functional assessments may not have adequately allowed for the functional impact of power training and that more power and speed-related functional tasks should have been included. We also hypothesized that function and pain would be improved in HSPT compared to SSST and CON because of the potential increase in oxygen partial pressure and diffusion of oxygen in the knee that high-speed movements have shown in patients with knee OA (21). However, any improvement in the metabolic or regenerative processes within the joint exposed to HSPT did not appear to contribute to differences in function among the groups.
Because improvements in TCR and both the WOMAC function and WOMAC pain measures were observed across all of the groups, including CON, it may be that our active control group greatly benefited from the 15-minute program of assisted lower extremity stretching and the 5-minute warm-up on a cycle ergometer. Although many studies employ stretching as part of an intervention that includes either RT or aerobic training, few studies have focused solely on stretching as an intervention program. Recently, 2 studies have employed stretching as a control intervention (41, 42); however, these studies did not show significant improvement as we did in WOMAC function (mean ± SD 1.7 ± 2.6 versus 4.8 ± 3.9) or pain (0.1 ± 1.0 versus 1.5 ± 2.5). Perhaps the assisted stretching program in the present study, the additional day per week of stretching compared to recent studies (3 days versus 2 days) (39, 40), or a combination of stretching and cycling was enough of an “intervention” to improve measures of function and pain in the present study. Because Kellgren/Lawrence scores were not different among the groups, the response of the CON group cannot be explained by differences in the severity of knee OA.
In conclusion, HSPT and SSST improved strength and peak muscle power in older adults with knee OA. Most importantly, high-speed training improved muscle speed, which may have important safety implications when high-speed movements are required. Both high-speed and slow-speed RT improved some measures of function and pain, but no more than an active control group. Because of the significant improvements in power and velocity with high-speed power training, we recommend the inclusion of this type of training in exercise programs for older adults with knee OA.
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 published. Dr. Sayers 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. Sayers.
Acquisition of data. Sayers.
Analysis and interpretation of data. Sayers, Gibson, Cook.