VALIDITY OF HAND-HELD DYNAMOMETRY IN MEASURING QUADRICEPS STRENGTH AND RATE OF TORQUE DEVELOPMENT

Abstract

Background

A hand-held dynamometer (HHD) offers a reliable and valid method to quantify quadriceps strength in a clinical environment. While measures of peak strength provide functional insights, most daily activities are performed quickly and do not require maximum strength. Rate of torque development (RTD) measures better reflect both the demands of daily activity and athletic movements. The capacity to obtain RTD measures in clinical settings is possible with an HHD, but the validity of RTD measures has not been quantified.

Hypothesis/Purpose

To determine the validity of an HHD to measure quadriceps isometric strength metrics compared to isometric strength measures obtained on an isokinetic dynamometer. It was hypothesized that the HHD would be a valid measure of peak torque and RTD at all time intervals when compared to the isokinetic dynamometer.

Study Design

Descriptive laboratory study.

Methods

Twenty healthy participants (12 male, 8 female) (age = 23.7 ± 2.9 years, height = 174.6 ± 10.1 cm, mass = 76.4 ± 15.9 kg, and Tegner = 6.7 ± 1.2) performed maximum isometric quadriceps contractions on an isokinetic dynamometer and with an HHD. Outcome measures included quadriceps peak torque and RTD at three intervals (0-100, 0-250 ms, and average). Pearson product-moment correlation coefficients and Spearman's rank correlation coefficient were used to determine relationships between devices. Bland-Altman Plots with Limits of Agreement (LOA) calculations were used to quantify systematic bias between measurement techniques.

Results

There was a significant correlation between the isokinetic dynamometer and the HHD for peak torque (p<.001, r = .894) and all RTD measurements (p<.002, r = .807; ρ = .502-.604). Bland-Altman plot LOA indicated the HHD overestimated peak torque values (19.4 ± 53.2 Nm) and underestimated all RTD measurements (-55.2 ± 190.7 Nm/s to -265.2 ± 402.6 Nm/s).

Conclusion

These results show it is possible to obtain valid measures of quadriceps peak torque and late RTD using an HHD. Measures of early RTD and RTDAvg obtained with an HHD were more variable and should be viewed with caution.

Level of Evidence

Diagnostic, Level 3

INTRODUCTION

Insights into functional ability and health status may be estimated based on isolated muscle function measures, such as hand grip strength or quadriceps strength (e.g. isometric, isokinetic).^1–4 Quadriceps strength is a prognostic indicator for chronic diseases including coronary artery disease,^5,6 chronic obstructive pulmonary disease,^7–9 and knee osteoarthritis.^10,11 Since quadriceps strength reflects health status and contributes to function it is an objective measure widely used in rehabilitation settings.

Currently, the gold standard method to quantify quadriceps strength utilizes an isokinetic dynamometer. However, this option lacks clinically applicability due to cost and size. A hand-held dynamometer (HHD) provides a valid and reliable testing alternative.^12,13 Although maximum quadriceps strength is an important measure of function, an activity such as walking requires submaximal (<60%) quadriceps activation¹⁴ and is accompanied with rapid joint loading (∼200 milliseconds [ms]).¹⁵ From an injury perspective, non-contact anterior cruciate ligament (ACL) injuries have been shown to occur within the first 100 ms after initial foot contact with the ground.^16,17 Quantifying the capacity to quickly develop muscle tension better reflects the demands of daily and sport activities.^18–21 Since quadriceps peak torque measured isometrically is typically performed by asking the participant to gradually increase muscle contraction intensity over a period of 1-5 seconds, it is not an accurate measure of an individual's ability to quickly develop muscle tension. Rate of torque development (RTD) is an alternative measure of quadriceps strength that quantifies how quickly muscle tension is developed. RTD is the change in torque relative to time and provides insights into different neuromuscular properties that contribute to muscle force production. Specifically, neurological properties have been shown to correspond with early RTD (<100 ms) and structural properties with later RTD (>150 ms).²² These insights can provide clinicians with information to select interventions to better address underlying causes of impairment.^23,24

Since RTD is calculated from an isometric contraction, an HHD provides a clinically feasible method to determine RTD. However, the validity of RTD measures obtained using an HHD has not been quantified. Therefore, the purpose of this study was to determine the validity of an HHD to measure quadriceps isometric strength metrics compared to isometric strength measures obtained using an isokinetic dynamometer. It was hypothesized that the HHD would be a valid measure of peak torque and RTD at all time intervals when compared to the isokinetic dynamometer.

METHODS

Twenty physically active adults (12 males, 8 females; age = 23.7 ± 2.9 years, height = 174.6 ± 10.1 cm, mass = 76.4 ± 15.9 kg, and Tegner = 6.7 ± 1.2) volunteered for the study. Inclusion criteria was age between 19 and 40 years. Exclusion criteria included a history of traumatic spine or lower extremity injury within the past six months. The Institutional Review Board at Creighton University approved the study (IRB 960835) and informed consent forms, which were compliant with the Declaration of Helsinki. Prior to testing, all participants completed an approved informed consent form, standardized health history form, and a form to quantify physical activity level (Tegner Activity Scale).

First, participants were tested on an isokinetic dynamometer (Biodex System 3; Computer Sports Medicine Inc., Stoughton, MA, USA) (Figure 1). The dynamometer was interfaced with a data acquisition system (MP150; Biopac Systems, Inc., Goleta, CA, USA), sampled at 2000 Hz, and recorded using AcqKnowledge software (version 4.2, Biopac Systems, Inc., Goleta, CA, USA). Torque signals were low-pass filtered at 15 Hz. Measures were obtained on both limbs, with the dominant limb tested first (leg used to jump). Participants completed a standardized warm up consisting of four submaximal isometric contractions, three at 50% effort and one at 75% effort, and two maximal isometric contractions with one minute of rest between each contraction. After the warm up was complete, participants performed six maximum isometric contractions, “as fast and as hard as possible,”²⁵ holding for approximately three seconds, with one minute rest between contractions. Loud verbal encouragement and real-time visual biofeedback (two lines, 100% and 90% of peak torque obtained during warm-up) were provided to ensure maximum effort during each trial.²⁶ If a participant was unable to perform near-maximum capacity (<90% peak torque) or if initial countermovement was identified via visual inspection of torque tracing, the trial was repeated until the six successful trials were completed. Initial countermovement occurs when a participant initially contracts the hamstring muscle, producing a stretch shortening cycle, prior to quadriceps maximum isometric contraction. This countermovement leads to inaccurate measures of quadriceps RTD due to the quadriceps isometric contraction not starting from rest. Once testing was completed on the dominant limb, the nondominant limb was tested using the same methods.

Figure 1.

Participants performed testing on an isokinetic dynamometer with hips flexed to 85 ° and knees flexed to 90 ° The chest and pelvis were secured with straps, and a padded ankle strap was placed 5 cm proximal to the distal aspect of the lateral malleolus. Participants kept arms crossed during testing to isolate the quadriceps muscle.

Participants then transitioned to a treatment table for isometric testing using an HHD (microFET2, Hoggan Scientific, LLC; West Jordan, UT) (Figure 2A). The HHD has a fixed sampling frequency of 100 Hz and the capacity to wirelessly transmit force signal data to a laptop computer via a manufacturer-supplied USB receiver. A modified belt-stabilized configuration ¹² was used to interface the HHD against a treatment table leg (Figure 2B). This testing configuration removes the limitation of tester strength present during a traditional non-belt stabilized testing configuration.²⁷ This configuration also helped maintain the position of the tibia pad and HHD during rest intervals. Participants performed a warm up consisting of three isometric contractions at 50% effort, 75% effort, and 100% effort with one minute of rest between each contraction. After the warm up, participants completed six maximal isometric contractions, holding for approximately 3 seconds, with one minute of rest between each contraction. Methods were consistent with the isokinetic dynamometer trials, but did not include visual biofeedback.

Figure 2.

HHD testing was performed in a seated position with legs hanging from the edge of a wooden treatment table, with the knee in approximately 90 ° of flexion (A). A small wedge bolster was placed at the posterior aspect of the distal thigh to minimize posterior thigh discomfort, and a strap (standard gait belt) was used to stabilize the thighs to the table. Participants kept arms crossed during testing to isolate the quadriceps muscle. A foam pad was placed across the anterior shin, in a location consistent with the isokinetic dynamometer (5 cm proximal to lateral malleolus), with the strap looped through the pad and the HHD secured against the leg of the table (B). A small elastic wrap was positioned between the table leg and triceps surae muscle to minimize belt slack (i.e., compliance).

Data processing was performed using specific algorithms created in MATLAB (The MathWorks, Inc., Natick, MA, USA). Peak torques were extracted for each isometric contraction and the trials with the two highest peak torques were used for data analysis. Since the HHD measures force (N), torque was calculated by multiplying the isometric force (N) by the moment arm (distance between lateral knee joint line to distal aspect of the lateral malleolus minus 5 cm).¹² The onset of contraction was determined by visual inspection of the torque tracing (1.0 Nm y-axis; 100 ms x-axis) and was defined as the last peak or trough before the torque signal deflected away from baseline noise (<1 Nm).^22,28,29 Repetitions with countermovement ( > 2 Nm) were discarded. In the HHD condition, there was an initial tension (approximately 40 N) created by the resting leg, strap, and bolster that was corrected for during the calculation of peak torque and RTD. RTD was calculated for each contraction by identifying the slope of the force-time curve (Δforce/Δtime) at the time intervals of 100 and 250 ms and to the time of peak torque (RTDAvg) from the onset of each contraction.

Statistical Analysis

Descriptive statistics were calculated for all outcome variables from the two contractions with the highest peak torque values. Additionally, all peak torque and RTD measures were normalized to body mass (Nm/kg or Nm/kg*s⁻¹) to allow comparison between studies which have expressed normative values for healthy individuals. The independent variable was testing method (isokinetic dynamometer, HHD), and the outcome variables were quadriceps peak torque (Nm) and RTD at 100ms and 250ms and RTDAvg (Nm/s). Data were examined for normal distribution using Shapiro-Wilk tests. When data were not normally distributed, nonparametric analyses were conducted. Differences between testing methods were determined using paired t-tests (parametric test) or Wilcoxon signed-rank test (nonparametric test) and validity was quantified using Pearson product-moment correlation coefficient (parametric test) or Spearman's rank correlation coefficient (nonparametric test). Bland-Altman Plots with Limits of Agreement (LOA) calculations were used to provide insights into systematic bias between measurement methods.³⁰ Significance was set a priori at p<0.05. Statistical analyses were performed with SPSS Version 25.0 (SPSS Inc., Chicago, IL).

RESULTS

Descriptive statistics and statistical summaries are available in Table 1. HHD data from one participant was excluded due to an equipment malfunction. Data for peak torque and RTD250 were normally distributed and analyzed using parametric tests while RTD100 and RTDAvg were not normally distributed and analyzed using nonparametric tests. There was a significant difference between devices for peak torque (p  =  .03), RTD100 (T < .001), and RTDAvg (T  =  .002), and no significant difference for RTD250 (p  =  .09). There was a significant correlation between the isokinetic dynamometer and the HHD for peak torque (p<.001, r = .894), RTD100 (t  =  .001, ρ = .604), RTD250 (p < .001, r = .807), and RTDAvg (p = .002, ρ = .502). Bland-Altman plot LOA indicated the HHD overestimated peak torque values (19.4 ± 53.2 Nm) and underestimated all RTD measurements (-55.2 ± 190.7 Nm/s to -265.2 ± 402.6 Nm/s) (Figure 3).

Table 1.

Quadriceps Strength Measures Between Devices. Values are mean ± SD. Normalized values are in parentheses.

	Isokinetic Dynamometer	Hand-Held Dynamometer	Significance	Correlation Coefficient	Limits of Agreement for Bland-Altman Plots
Peak Torque	284.1 ± 95.5 Nm (3.57 ± 0.61 Nm/kg)	303.5 ± 118.2 Nm (3.81 ± 0.91 Nm*s−1kg)	p  =  .03*	r = .894	19.4 ± 53.2 Nm
RTD100	789.9 ± 442.6 Nm/s (9.62 ± 4.40 Nm*s−1/kg)	524.6 ± 391.8 Nm/s (6.38 ± 4.08 Nm*s−1kg)	T =  < .001*	r = .527	-265.2 ± 402.6 Nm/s
RTD250	832.8 ± 311.1 Nm/s (10.41 ± 2.43 Nm*s−1kg)	777.6 ± 354.7 Nm/s (9.71 ± 2.94 Nm*s−1kg)	p  =  .09	r = .839	-55.2 ± 190.7 Nm/s
RTDAvg	328.0 ± 122.4 Nm/s (4.17 ± 1.08 Nm*s−1/kg)	269.0 ± 209.7 Nm/s (3.37 ± 2.17 Nm*s−1kg)	T  =  .002*	r = .280	-59.0 ± 208.3 Nm/s

* HHD values were significantly different (P or T < .05) than isokinetic dynamometer.

Figure 3.

Bland-Altman Plots and Limits of Agreement for Isokinetic Dynamometer versus Handheld Dynamometer. Black line indicates average difference, red and yellow lines indicate upper and lower Limits of Agreement.

DISCUSSION

The purpose of this study was to determine the validity of an HHD to measure isometric quadriceps peak torque and RTD compared to an isometric measure of the same on an isokinetic dynamometer. The results of this study support the hypothesis that the HHD provides a valid measure of quadriceps peak torque and RTD. Measures obtained using the HHD were significantly correlated with isometric peak torque measures obtained using the isokinetic dynamometer (r = .894), which is in agreement with previous research.¹² The HHD significantly overestimated torque values (Figure 3A) with an average difference of 19.4 Nm (SD = 53.2; 95% CI -87.0 to 125.8). This overestimation is slightly greater than the 15.1 Nm minimal detectable change of an isometric quadriceps contraction measured with an HHD.¹² Previous studies^12,13 indicate using a similar belt-stabilized HHD configuration tends to result in an underestimation of peak torque (13-36 Nm; 95% CI -50 to 70 Nm), although the confidence intervals have overlap with results from the current study. The possible cause of this overestimation could be due to the manual calculation of torque for the HHD trials. Torque was calculated by taking the product of the force (N) and the distance (m) from the lateral joint line of the knee to the center of the pad placed just proximal to the lateral malleolus. Recording the distance between the lateral joint line of the knee and the center of the pad, with a flexible tape measure, introduces the possibility of measurement error which could have contributed to the overestimation HHD peak torque relative to values obtained on the isokinetic dynamometer.

RTD values obtained with the HHD demonstrated a significant correlation with the isokinetic dynamometer for RTD100 (ρ  =  .604) and RTD250 (r = .807) RTDAvg (ρ  =  .502). The HHD underestimated RTD values (Figure 3B-D) with a difference ranging between 55.2 Nm/s to 265.2 Nm/s. The lower correlation for early RTD values and the underestimation of all RTD values (significantly lower early RTD) may be due to differences in testing set-up, specifically the lack of back support for HHD testing (Figure 2A). Since participants were unable to lean against a rigid surface to stabilize their trunk during testing this may contribute to a longer time to peak torque. This is reflected in the lower RTDAvg values obtained with the HHD despite relatively similar peak torque values. Also, early RTD values have been shown to have lower reliability and higher variability when compared to later time intervals³¹ and potentially due to neural factors such as motor unit recruitment and rate.²⁵ This could be a potential reason for the higher correlation for RTD250 as compared to RTD100. An additional factor contributing to the lower correlation for early RTD values may have been due to slack in the belt and elastic properties of the pad, which would alter the sensitivity of the HHD.²⁵ Although we attempted to minimize this slack (Figure 2B), it is likely there was still a level of compliance with the belt and pad. While customized dynamometers with minimal padding are suggested,²⁵ researchers should attempt to best identify methods which have the capacity to be replicated in clinical environments and minimize participant discomfort. Another contributing factor to the underestimation of RTD values could be due to the presence of pre-tension in the belt, which has been shown to increase early RTD.³² While participants were instructed to start from a relaxed state, it is possible participants were extending their knee to maintain constant tension on the belt. Pre-tension was limited by placing a small elastic wrap between the leg of the table and the triceps surae muscle. This created enough tension to keep the HHD secured against the table, allowing the participant to relax between trials. All trials were visually inspected for pre-tension and countermovement during data processing. Since RTD was lower with the HHD relative to the isokinetic dynamometer, it is likely pre-tension was not a consistent issue, but may have contributed to the variance in values. Future studies may incorporate the use of electromyography for visual biofeedback and verification that trials start from a relaxed state.

The results from this study indicate that an HHD can be used, as a substitute for an isokinetic dynamometer, to obtain measures of quadriceps peak torque and later RTD values but not for early RTD and RTDAvg values. This study showed a significant difference between isokinetic dynamometer and HHD for RTD100 (T<.001) and RTDAvg (T = .002) values. To determine an individual's early RTD values, it may be more appropriate to test them on a more stable or less compliant testing device. The advantage is that the HHD offers a less expensive and time-consuming way to obtain measures in a clinical environment. Quadriceps RTD measures can be used to monitor rehabilitation progress and inform clinical decisions regarding treatment approaches and offer insights beyond peak torque. Deficits in early RTD are due to neuromuscular properties, such as neural drive, while RTD at later stages is due to contractile properties, such as cross-sectional area.²² Training programs that focus on both peak torque and reaching maximum force as quickly as possible have shown to lead to increases in peak torque and RTD in healthy populations^24,33 and individuals with ACL reconstruction.³⁴

A limitation of this study was that the sample population only consisted of young, healthy individuals. Thus, results cannot be extrapolated to individuals beyond this age range or those with pathology such as knee injury (e.g., ACL, osteoarthritis). Additionally, the knee angle for quadriceps strength was only tested at 90 ° of flexion. While quadriceps peak torque values are highest around 60 ° of knee flexion versus other angles (e.g. 30 °, 90 °),³⁵ the selected knee joint angle for testing varies across studies.^36–38 Knee flexion angle of 90 ° was selected because it is easier to reproduce in the clinic and eliminates the need to correct for gravity (limb mass). Future research should study individuals with knee joint or other lower extremity pathologies, quadriceps strength at different knee joint angles, and different muscle groups using similar testing configurations.

CONCLUSIONS

These results show that measures of quadriceps peak torque and late RTD (RTD250) obtained with an HHD are valid compared to an isokinetic dynamometer. Measures of early RTD (RTD100) and RTDAvg should be viewed with caution. Overall the HHD overestimated quadriceps isometric peak torque and underestimated quadriceps RTD measures. The relatively wide confidence intervals for the LOA indicate that these results should be interpreted with caution. These inexact estimations of strength need to be kept in mind while progressing patients through rehabilitation programs. By using strength benchmarks for progression, clinicians could either inaccurately advance or block patients from progressing through rehabilitation after ACL reconstruction.