Assessment of human maximal muscular strength and power ability is routinely performed within strength and conditioning (McMaster et al., 2014; Paul and Nassis, 2015; Suchomel et al., 2016), as well as physical therapy and rehabilitation practice (Dickoff-Hoffman and Davies, 1993; Myer et al., 2006). Strength and power assessment may focus on the capacity of individual muscle groups or the whole kinetic chain (Schrama et al., 2014; Suchomel et al., 2016; Zapparoli and Riberto, 2017; Petrigna et al., 2019). For instance, during the rehabilitation of the anterior cruciate ligament injury, it is common to perform both single-joint knee strength assessment and jump or hop tests for height or distance (a multi-joint task; O’Malley et al., 2018). The body of knowledge on strength and power assessments is continuously growing, with novel measurement devices (Hickey et al., 2018; Whinton et al., 2018; Palmer et al., 2020) and methodological approaches to assessment (Bellumori et al., 2011; García-Ramos et al., 2019; Hughes et al., 2019) being proposed and validated. In terms of single-joint strength assessments, most of the research has been dedicated to the knee joint, which is likely due to the high reliability of these measurements and their implications in rehabilitation (de Araujo Ribeiro Alvares et al., 2015; Muñoz-Bermejo et al., 2019) and implications in the rehabilitation of several prevalent and detrimental lower-limb injuries (Myer et al., 2006; Cvjetkovic et al., 2015; Kaeding et al., 2017; Bourne et al., 2018; Goff et al., 2018; O’Malley et al., 2018). Knee extension and flexion strength measurements have been consistently shown as reliable when conducted under isometric (De Carvalho Froufe Andrade et al., 2013; de Araujo Ribeiro Alvares et al., 2015; Sung et al., 2019) or isokinetic (Sole et al., 2007; De Carvalho Froufe Andrade et al., 2013; de Araujo Ribeiro Alvares et al., 2015; Törpel et al., 2017; Zapparoli and Riberto, 2017; Muñoz-Bermejo et al., 2019) conditions using state-of-the-art dynamometers. Moreover, moderate to high reliability and validity have been reported for knee strength assessments performed by hand-held dynamometers (Mentiplay et al., 2015; Chamorro et al., 2017; Chopp-Hurley et al., 2019; Lesnak et al., 2019; van der Made et al., 2019).

A large body of evidence related to knee strength has facilitated the formulation of guidelines for coaches, physical therapists, and other practitioners within sport science and medicine, orthopedics, rehabilitation, and prevention. For instance, the ratio between hamstring and quadriceps muscles (assessed by the knee flexion and extension torque capacity, respectively) has been shown as one of the risk factors for future lower limb injuries (Devan et al., 2004; Dallinga et al., 2012; Lee et al., 2018) and has been suggested as one of the criteria for safe return to sport (Cvjetkovic et al., 2015; Erickson and Sherry, 2017). Moreover, inter-limb knee strength asymmetries have been linked to history of injury (Schiltz et al., 2009) and reduced sport performance (Bishop et al., 2018). One of the paramount goals of rehabilitation of unilateral knee injuries is to establish the level of strength similar to the uninjured side (Myer et al., 2006; O’Malley et al., 2018). Although some ambiguities exist (Kaplan and Witvrouw, 2019), the guidelines related to inter-limb asymmetries in knee strength as well as flexor to extensor imbalances are useful in prevention and rehabilitation of knee injuries. However, less is known regarding the absolute reference values for knee flexion and extension strength that could be used as a reference in rehabilitation of athletic injuries, prevention or assessing the general functional capacity in older adults or patient populations. In addition to quantifying the ratios between the limbs or opposite muscle groups, it could be useful to have a reference value to which absolute strength levels of an individual could be compared.

Several previous studies have attempted to establish reference values for knee strength in different populations. These studies have encompassed athletes (Zvijac et al., 2014; Risberg et al., 2018; Hannon et al., 2019), children (Holm et al., 2008; McKay et al., 2017), older adults (Pereira et al., 2019), and general adult population (Meldrum et al., 2007; Harbo et al., 2012; Bohannon, 2017; McKay et al., 2017). Therefore, albeit limited, reference values in the literature do exist. However, comparing these studies is difficult, because they reported different units of measurement. Most notably, some studies reported their results as absolute force (Meldrum et al., 2007) or absolute torque values (Harbo et al., 2012; McKay et al., 2017), while others also reported body-mass normalized results (Holm et al., 2008; Zvijac et al., 2014; Risberg et al., 2018; Pereira et al., 2019). Moreover, although the single-joint strength assessments are generally valid and reliable, some of the differences could occur due to measurement errors, use of different devices and different assessment protocols. Finally, several factors pertaining to measurement protocol, such as knee angle during the isometric measurements (Marginson and Eston, 2001; Krishnan and Williams, 2014) or velocity during isokinetic measurements (Perrine and Edgerton, 1978; Grbic et al., 2017), are important to consider in strength assessment. Since these factors are not standardized across studies, the generalization of findings from single studies and translation into practice is further limited. On the other hand, there is a large body of methodological, observational and interventional studies that examined and reported knee strength as one of the outcomes. Examining this body of literature would provide a scoping overview of reference values that could be useful for practitioners during the assessment of their clients’ level of strength, as well as scientists to compare the outcomes from their measurements to typical values obtained across previous investigations. To the best of our knowledge, no attempt has been made to date to review all studies that reported knee flexion or extension strength as an outcome.

The aim of this paper was to review all available studies that reported isometric knee extension and/or flexion strength as assessed during maximal voluntary contraction. Because the body of literature investigating knee strength is very large, we decided to limit our review to isometric contractions in healthy populations, including children, adults, older adults, and both genders. While isokinetic strength is more functionally relevant in some aspects than isometric strength, this decision was based on the fact that most isokinetic dynamometers allow the measurement of isometric strength but not vice versa. Moreover, acceptable to high reliability of hand-held dynamometry to assess isometric knee strength has been reported (Mentiplay et al., 2015; Chamorro et al., 2017; Chopp-Hurley et al., 2019; Lesnak et al., 2019; van der Made et al., 2019). To facilitate the comparison of the studies and study subgroups, we aimed to obtain body-mass normalized torque values. For this purpose, we converted absolute force/torque values into body-mass normalized scores by using appropriate estimates (see section Estimating Body Mass Normalized Torque From Absolute Values for details).

This paper aimed to obtain reference values for knee strength in athletes and general adult and older adult populations. For this purpose, we performed a broad review of studies that assessed and reported isometric knee extension and/or flexion strength during maximal voluntary contraction. To the best of our knowledge, this is the first attempt at obtaining reference knee strength values by performing an all-encompassing review of the literature. Several previous studies have attempted to obtain reference values for specific populations or subgroups by performing measurements on larger samples of participants. Zvijac et al. (2014) examined over 1,200 collegiate male American football players and reported the means for peak isokinetic (60°/s) knee torque that ranged (based on the playing position and limb dominance) from 2.48 to 3.08Nm/kg and from 1.69 to 2.10Nm/kg for extension and flexion, respectively. Furthermore, Whiteley et al. (2012) examined 216 professional male soccer players and reported mean isokinetic (60°/s) knee extension strength at 3.1–3.3Nm/kg and knee flexion strength at 1.7–1.8Nm/kg. Hannon et al. (2019), likewise using isokinetic assessment at 60°/s, assessed elite female handball (n=150) and soccer (n=200) players, and observed a similar normalized peak torque in both player groups (mean extension: 2.3–2.4Nm/kg; mean flexion: 1.3–1.4Nm/kg). Similar to our results, the extension torque was highest in the mid-range knee position, while the flexion torque was relatively constant across the range of motion. In sum, the values obtained in the previous studies roughly agree with our results. Somewhat lower torque values sometimes obtained in the abovementioned studies likely resulted from using isokinetic measurements. According to the force-velocity relationship (Perrine and Edgerton, 1978; Gülch, 1994; Grbic et al., 2017), isokinetic assessment should result in lower values than isometric measurements.

For the non-athletic populations, comparing our results to the available literature is challenging, because previous studies that attempted to obtain reference values did not report body-mass normalized units (Meldrum et al., 2007; Harbo et al., 2012; McKay et al., 2017). Interestingly, our analyses showed only modest differences between athletes and general adult population. Accordingly, Ushiyama et al. (2017) reported the mean isometric knee extension torque of 2.64Nm/kg (range: 1.98–4.61) in 30 (20 males) healthy adult participants. Some studies reported values exceeding the confidence intervals that we calculated. For instance, Dalgaard et al. (2019) reported isometric knee extension (knee angle=70°) in 28 young females at 3.0Nm/kg, and even at 3.3–3.5Nm/kg after a period of resistance training. Unfortunately, to the best of our knowledge, no study examined knee strength in a larger sample (e.g., >100) of adults and reported the outcomes in Nm/kg. Harbo et al. (2012) tested 93 male and 85 female non-athlete adults and reported isometric knee extension peak torque at 246.6±56.3Nm for the males and 166.6±38.2Nm for the females, which is within our calculated confidence intervals for non-normalized values (although close to the upper limit; Supplementary File 1).

The reference values that we calculated for the older adults also seem to mostly agree with individual previous studies. Periera et al. examined the isokinetic (60°/s) knee extension strength in 453 older women, separated by age groups. In the 60–65years old group, they reported the mean knee extension peak torque values at 1.5Nm/kg, which corresponds to our results for mid-range knee angle measurements. Interestingly, they observed only a modest decline in strength throughout the groups, reaching 1.2Nm/kg in the oldest group (80–85years old), with barely noticing any decline until 75years of age. This supports our decision to pool all available data from the older adult studies since most of these included older adults aged between 65 and 75 years. Spinoso et al. (2019) also assessed isometric knee flexion strength (knee angle=60°) in 30 older females and reported the mean value at 0.64±0.17Nm/kg, which was statistically significantly lower than the mean value in the group of 16 younger female participants (1.09±0.23Nm/kg). A similar observation was reported for the isometric knee extension strength (1.65±0.23Nm/kg in the older group; 2.45±0.52Nm/kg in the younger group; Spinoso et al., 2019). These results all fit the knee strength reference values that we calculated. Lower mean values for isometric knee extension strength (1.12–1.15Nm/kg) were reported by Kim et al. (2012) for 55 community-dwelling elderly women (aged 75 and older) with sarcopenia. Lower-limb muscle strength is one of the determinants of risk of falling (Pijnappels et al., 2008; Hicks et al., 2020) and is also associated with general physical functioning of older adults (Liu and Latham, 2009; Brady et al., 2014). Therefore, the values obtained in this review could serve as a reference when trying to improve strength in older adults.

Some limitations of this review need to be acknowledged and discussed. The authors generally followed the PRISMA guidelines, with some exceptions being made that are limiting the rigor of this review. Firstly, although the authors performed the review with the best possible care and scrutiny, the relevant body of literature was large, which makes it possible that some studies were omitted. Nevertheless, we believe that a significant proportion of the relevant studies have been included in the review, and additional studies would have no or small effects on the pooled results. For the same reason, only two literature databases were considered. The risk of bias of individual studies was not assessed. This was mainly due to the fact that a multitude of study designs were eligible for inclusion in the review, which make the comparison of papers very difficult. Moreover, we omitted some of the possible bias sources (e.g., allocation bias) by extraction only baseline data in longitudinal studies. However, other types of bias, such as selection bias, could have still affected our results. Unfortunately, several studies that examined knee strength did not report descriptive data. Normalization of the data to Nm/kg was also not possible in all studies because body mass data were absent. Furthermore, it is not entirely clear whether normalizing the torque values by body mass is always the most appropriate. The first studies that examined the methods used for adjusting strength values to body size suggested that torque values should indeed be divided by body mass, while force values should be divided with 2/3 of the body mass for appropriate between-participant comparisons (Jaric et al., 2002). However, some of the subsequent studies showed that different calculations might be more suitable in different populations. For instance, Abdalla et al. (2020) suggested that the knee extension strength should be divided by 96% of body mass in older males and 70% of body mass in older females when establishing sarcopenia status. For children, it was suggested that 140% of the body mass should be used in the division (Wren and Engsberg, 2007). This does not affect our results, which represent pooled data from multiple studies that reported body mass normalized values as torque divided by 100% of the body mass. Rather, future researchers should be careful when using our values to establish different criteria or cut-offs.

This review provides reference values for isometric single-joint knee flexion and extension strength. The obtained values seem to mostly agree with previous single studies that attempted to obtain reference knee strength values in specific subgroups. These values may serve as a reference for future researchers, as well as practitioners during rehabilitation of musculoskeletal injuries, prevention or general assessment of physical abilities of the older adult or patient populations. Researchers and practitioners should be mindful of factors affecting knee extension and flexion strength (and the ratio of the two), such as sex, age, and knee angle. We have to stress that the values should be viewed cautiously, as some factors influencing strength could not be taken into account.

NS conceptualized the paper. NS and ZK carried out the study selection process. MP and ZK analyzed the data. ZK wrote the first manuscript draft. MP and NS finalized the paper. All authors contributed to the article and approved the submitted version.

The study was supported by the Slovenian Research Agency through the project TELASI-PREVENT (L5-1845; Body asymmetries as a risk factor in musculoskeletal injury development: studying aetiological mechanisms and designing corrective interventions for primary and tertiary preventive care). The funder played no role in conceptualization of the study, data acquisition, article writing nor any other phase of the study.

NS was employed by company S2P, Science to Practice, Ltd.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Authors alone are responsible for conceptualization of the study and data analysis. The funders provided part of the authors’ salaries through national research fund and were not involved in any part of the study.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.