This systematic review was designed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The PRISMA statement provides a 27-item checklist intended to guide the reporting of systematic reviews, particularly those involving randomized controlled trials (Page et al., 2021).

To search all relevant studies, three electronic databases were chosen (PubMed, Web of Science, and Scopus). For additional search, the backward search was used (i.e., assessing the reference lists of included articles). The terms such as “Nordic curl” and “Nordic hamstring” were combined with the terms “football” and “volume” to find any title and abstract that is related to our topic. Only publications in the English language were included. Boolean terms “AND” and “OR” were used for the keyword’s combination.

The main inclusion criteria were that studies had to examine the NHE effect of either high or low volume or a comparison of both on muscle architecture adaptations and/or eccentric strength variables of hamstring muscles in amateur, semi-professional, or professional soccer players of both genders. Other criteria were based on the publication date of 2014–2025 and full text availability of the articles. The duration of the included study had to be more than 4 weeks, and the experimental group could not have possessed any lower limb injury at least 6 months before the start of the study. The muscle architecture had to be measured on BF. Mean ± standard deviation (SD) pre- and post-intervention were provided for the measured variables for secondary analysis. Additionally, there had to be an absence of any health disorders that would interfere with the study results. The next criteria were based on assessing the post-PHV players (peak height velocity) to prevent the influence of the maturation status on our results (Drury et al., 2020). Studies were excluded if they collected data solely through injury incidence questionnaires without measuring physiological or performance adaptations. Additionally, isokinetic data collected at angular velocities greater than 120°/s were excluded due to decreased reliability and a reduced percentage of the range of motion maintained at a constant velocity as angular speed increases (Kellis and Baltzopoulos, 1995).

If the studies failed to meet our criteria, they were excluded from this review. It should be mentioned that although women may possess lower peak strength values than men (Medeiros et al., 2021), we decided to include the participants of both genders, as our analyses considered the strength and architecture adaptation differences between pre- and post-intervention rather than peak strength values of athletes (Medeiros et al., 2021).

The methodological quality of the studies was assessed using the Testex scale (Smart et al., 2015), a quality assessment instrument specifically designed for exercise training studies. It assesses methodological quality based on 12 criteria, with a total possible score of 15 points. Reference scores were used to express the quality level of the studies and are presented as follows: <4 points “poor quality,” 4–7 points “moderate quality,” 8–10 points “good quality,” and >11 points “excellent quality” (Davies et al., 2021). Two types of bias assessments were carried out in this review: the Rosenthal fail-safe N method (Rosenthal, 1979) and the Cochrane risk of bias tool for randomized controlled trials (Sterne et al., 2019). The Cochrane tool evaluates RCT’s across several domains, such as the randomization process, deviations from the intended intervention, missing outcome data, etc. Each domain is rated as having a “low risk of bias,” “high risk of bias,” or “some concerns.” The fail-safe N was applied to estimate how many unpublished or missing studies with null results would be required to reduce the observed effect to nonsignificance (p > 0.05).

Means and standard deviations for strength outcomes and muscle architecture measures were independently extracted from the included studies. Strength outcomes included isokinetic assessments such as eccentric peak torque (measured at 30°/s, 60°/s, and 120°/s) and eccentric force. Muscle architecture variables included fascicle length, pennation angle, and muscle thickness.

Effect sizes (ES) were calculated to provide standardized comparisons of mean differences between groups or experimental conditions. Hedges’ g, along with 95% confidence intervals (CI), was used to quantify mean differences between pre- and post-intervention, as it adjusts for sample size disparities. Hedges’ g was determined using the following formula (Hedges and Ingram, 1985): g = M e a n   p o s t − M e a n   p r e SD   pooled

Interpretation of ES was followed by Hopkins thresholds (Hopkins, 2010): trivial (≤0.20), small (0.20–0.59), moderate (0.60–1.19), large (1.20–1.99), and very large (≥2.00). Consistency across studies was assessed using Higgins test for heterogeneity (I²) (Hopkins, 2010), with values interpreted as low (<25%), moderate (25%–75%), or high (≥75%) heterogeneity. Between-study variance estimates for both strength and architecture outcomes were calculated using random-effects models, with results expressed as 95% confidence intervals (The Jamovi project, 2019). Effect sizes and random-effects models were calculated using Jamovi software (version 2.7.6). For clarity and improved visualization, forest plots were created separately based on the extracted effect size data.

The PRISMA flow diagram (Figure 1) illustrates the search and selection stages (Page et al., 2021). Initially, we identified 855 articles from the main databases and another 29 articles, which provided an additional backward search from the reference lists of relevant studies. After the exclusion of duplicates (n = 47), we removed studies that did not research hamstring strength variables or architectural adaptations after NHE intervention in a soccer population. Thirty-nine relevant articles were included in the full-text analyses, and eleven of them reported suitable data for quantitative analysis. The outcomes extracted from the studies were eccentric knee flexor strength or eccentric knee flexor torque. Other outcomes obtained from the studies were fascicle length (FL), pennation angle (PA), and muscle thickness (MT) of BF^lh, SM, and ST muscles, which represent the hamstring muscle architecture. For qualitative analysis Pre-post differences were calculated from mean values and expressed as percentages. These indicators serve to improve the understanding of the collected data (Table 3). For quantitative analysis, data were collected from the interventions reported in the included studies, comparing outcomes before and after the interventions. Information from both interventions or/and control groups was extracted when available; however, having a control group was not part of the inclusion or exclusion criteria. As a result, three of the included studies, de Oliviera et al. (2020), Vianna et al. (2021), and Siddle et al. (2024) did not include control groups.

Quality of assessment was assessed using TESTEX criterion. The mean score of the included studies was 8 out of 15 points, with the highest-scoring studies being randomized trials (Lovell et al., 2018; Medeiros et al., 2020). The scores in 6 of the presented studies were of moderate quality, 3 studies were of good quality, and 2 studies were presented with excellent quality scores (Table 1).

For the risk of bias, two tools were also performed. The Cochrane risk of bias (Table 2) shows an overall “some concerns” for the review, mostly due to unclear randomization methods or incomplete reporting of missing data. The second tool identifies that the results of this meta-analysis are not subject to publication bias (p < 0.001), with 178 “filed away” studies for eccentric strength and 30 studies for muscle architecture needed to prove null effects of NHE interventions on eccentric strength and muscle architecture. These results indicate that the findings for eccentric strength are robust and unlikely to be overturned by unpublished null studies. In contrast, the muscle architecture outcomes appear more fragile, as relatively few unpublished studies could change the statistical significance.

Within-study pre–post differences showing the magnitude of change (Hedge’s g, 95% CI) across all included trials are illustrated in Figures 2–6 below. For eccentric hamstring strength, high-volume interventions (de Oliviera et al., 2020; Suarez-Arrones et al., 2019; Amundsen et al., 2022; Ishøi et al., 2018; Vianna et al., 2021) consistently demonstrated moderate improvements (g = 0.77, p < 0.001, 95% CI 0.49–1.06), with moderate heterogeneity (I² = 51%). Low-volume protocols (Cadu et al., 2022; Amundsen et al., 2022) also produced significant but small effects (g = 0.46, p < 0.05, 95% CI 0.06–0.87) with no heterogeneity (I² = 0%), while control groups (Suarez-Arrones et al., 2019; Ishøi et al., 2018) showed only trivial or negative effects (g = −0.06, p = 0.75, 95% CI −0.40–0.29), with moderate heterogeneity (I² = 51%). In the case of eccentric torque, low volume interventions (Amundsen et al., 2022; Siddle et al., 2024) did not result in significant gains, with only trivial effect (g = 0.04, p = 0.74, 95% CI −0.21–0.29), as well as high-volume interventions (g = 0.05, p = 0.55, 95% CI −0.12–0.23) (Medeiros et al., 2020; Amundsen et al., 2022; Sebelien et al., 2014; Lovell et al., 2018). No heterogeneity (I² = 0%) was presented in low and high volume groups, respectively. Two studies provided a control group (Sebelien et al., 2014; Lovell et al., 2018), resulting in a trivial negative effect (g = −0.07, p = 0.67, 95% CI −0.67–0.51, I² = 0%).

Changes in muscle architecture were more variable across outcomes. For fascicle length, high-volume interventions (Medeiros et al., 2020; Lovell et al., 2018; Vianna et al., 2021; Mendiguchia et al., 2020) elicited significant changes with a small effect size (g = 0.43, p < 0.001, 95% CI 0.20–0.65) and no heterogeneity (I² = 0%), whereas control groups (Lovell et al., 2018; Mendiguchia et al., 2020) showed a small negative effect (g = −0.31, p = 0.25) with low heterogeneity (I² = 23.7%). Low volume intervention (Siddle et al., 2024) resulted in non-significant gains with only trivial ES (g = 0.12, 95% CI −0.37–0.61).

In contrast, adaptations in pennation angle were not statistically significant (g = −0.16, p = 0.16, 95% CI −0.38–0.06, I² = 0%) with trivial to small negative ES for either high-volume groups (Medeiros et al., 2020; Lovell et al., 2018; Mendiguchia et al., 2020) or control groups (g = −0.51, p = 0.39, I² = 78.6%) (Lovell et al., 2018; Mendiguchia et al., 2020). Low volume group (Siddle et al., 2024) again resulted in no significant changes (g = −0.11, −0.61–0.38) after the training intervention. For muscle thickness, high-volume interventions (Medeiros et al., 2020; Lovell et al., 2018; Mendiguchia et al., 2020) demonstrated significant increases with a small effect (g = 0.48, p < 0.001, 95% CI 0.28–0.68), whereas control groups showed a moderate negative effect (g = −0.91, p = 0.39) with high between-study heterogeneity (I² = 88.3%). A trivial negative effect (g = −0.10; 95% CI −0.59–0.39) was found in low volume group (Siddle et al., 2024).

Although the NHE has been widely used by strength and conditioning coaches, no consensus about dosage and its manipulation in training has been reached yet. There was a wide range of volumes and durations of included studies from our literature search (Table 3). In summary, 10 of 11 articles included the measuring of eccentric strength variables of hamstring muscles. The high volume of NHE has been performed in 8 studies (de Oliviera et al., 2020; Amundsen et al., 2022; Suarez-Arrones et al., 2019; Ishøi et al., 2018; Vianna et al., 2021; Lovell et al., 2018; Sebelien et al., 2014; Medeiros et al., 2020), whereas only 3 articles applied a low volume of NHE (Amundsen et al., 2022; Cadu et al., 2022; Siddle et al., 2024). Closer insights into the training program and results of included studies are presented in Table 3.

Performing high volumes of NHE leads to a significant increase in peak eccentric strength after 4 weeks (de Oliviera et al., 2020) and after longer periods (>8 weeks) (+10%–16.5%). (Amundsen et al., 2022; Vianna et al., 2021; Suarez-Arrones et al., 2019; Ishøi et al., 2018; Medeiros et al., 2021). All studies, where peak eccentric strength was investigated, found significant gains after high volumes of NHE. However, in a study where the players had been previously trained using NHE, no significant increase was seen (Suarez-Arrones et al., 2019). In addition, a high volume of NHE leads to significant improvements in eccentric peak torque only in 2 (Lovell et al., 2018; Medeiros et al., 2020) of 4 studies (Lovell et al., 2018; Medeiros et al., 2020; Amundsen et al., 2022; Sebelien et al., 2014).

Low volume of this exercise leads to a significant increase of peak eccentric strength in female soccer players (Amundsen et al., 2022), with no significant differences between low and high volume groups. Significantly higher peak eccentric strength was found after performing a very low volume (1 set of 3 reps) for 21 weeks and with a low compliance rate (47.2%) and with higher changes for the “high compliance subgroup” (Cadu et al., 2022). Two studies in our review examined the effect of low volume on eccentric peak torque in professional female (Amundsen et al., 2022) and academy soccer players (Siddle et al., 2024) with no significant gains.

The variability of results in increasing eccentric torque after NHE can be explained by the poor correlation between peak eccentric strength measured on NordBord and eccentric peak torque measured on isokinetic dynamometers (Amundsen et al., 2022; Wiesinger et al., 2020). This fact can be explained by the low similarity of movement when performing tests on the NordBord device versus the isokinetic machine (Medeiros et al., 2021). Furthermore, any significant correlations between the eccentric peak torque during NHE and peak eccentric knee flexion torque (r = 0.24–0.3, p = 0.26–0.4) were found when measured on a NordBord and isokinetic dynamometer (Nishida et al., 2022). Similarly poor correlation (r = 0.35) was found between eccentric force during NHE and isokinetic eccentric peak torque at 60°/s (Van Dyk et al., 2018). Another issue seems to be the difference in the body positions (prone, sitting) and different movement velocities (30°/s, 60°/s, 180°/s, 270°/s) that were used in the studies to measure this parameter. Significant changes were observed after performing a high volume of NHE in 2 studies (Lovell et al., 2018; Medeiros et al., 2020), but only when using lower angular velocity during testing on isokinetic devices (30°/s, 60°/s). Measuring eccentric peak torque has therefore seemed to be highly influenced by other factors such as the movement performed and the velocity, which must be taken into consideration. On the contrary, emphasizing higher angular velocities during testing with an isokinetic dynamometer seems to have some applicability, as sprinting is the most frequent action during which HSI is occurring. The use of high velocities can better explain specific hamstring work during this task, although it is performed in isolation.

The peak eccentric strength measured during NHE on the NordBord device (Opar et al., 2013) is the most used variable in practice and research. However, there are some misconceptions about which variable is the best for assessing the eccentric hamstring strength. Opar et al. (2015) suggest implementing torque measurement during NHE on NordBord, as there are some differences between athletes with longer or shorter lower leg levers. Despite this, they claim that measuring the force output during NHE still provides useful information for HSI risk. Buchheit et al. (2016), on the other hand, found that measuring the knee flexor strength is largely body mass dependent, but simply dividing absolute strength by athlete body mass is not a sufficient method. They assume to use the provided equation (eccentric strength [N] = 4 x BM [kg] + 26.1) for the estimation of players’ expected strength based on their own body mass (BM) and compare it to their actual peak force value from the test when monitoring players over longer periods when BM changes may occur. This equation was developed to differentiate what effect the athlete’s BM had compared to his true eccentric strength. On the other hand, Opar et al. (2021) did not find any difference between prospectively injured or non-injured soccer players irrespective of the quantification tool used (between-limb asymmetry, relative or absolute strength) in pre-season eccentric strength. Therefore, practitioners can use different quantifications of eccentric strength, based on their goal.

Only 6 of 11 investigated studies provided a compliance rate greater than 80%. Some previous studies (Chesterton et al., 2021) have shown that the low compliance rate in soccer players is a result of DOMS presenting after the NHE. Goode et al. (2015), in their meta-analysis, found that the effect of the exercise is also highly influenced by intervention compliance. Similar results were reported by Rudisill et al. (2023) and Biz et al. (2021), who demonstrated that eccentric hamstring training, particularly the NHE and other injury prevention protocols, can reduce hamstring injury incidence by up to 70%, but poor compliance and limited implementation in team practice remain major barriers to effectiveness.

The groups where a high volume of NHE has been performed increased intensity via increased volume, whereas in low volume groups it is more likely that the increase in intensity was due to increasing the breakpoint angle of the hamstrings during movement because no volume has been increased (Cuthbert et al., 2020). The athlete can last longer and get his torso closer to the ground, which increases the torque due to force being applied over a greater momentum. For improving muscle strength, the performed intensity must be over 85% of 1 repetition maximum (RM) and ∼6 reps (Haff et al., 2016). This mostly applies to more advanced athletes, whereas in novice athletes, lower intensity is also preferred to elicit improvements in strength capabilities. The NHE is supramaximal in nature, and intensity is above 1RM; therefore, it applies true eccentric stimulus on the hamstring muscles. The assumption would be that lower repetitions are enough to stimulate hamstrings at the same level. Severo-Silveira et al. (2021), on the other hand, found that a progressive workload (236 reps over 8 weeks) had better results than a constant workload (138 reps over 8 weeks) from an eccentric strength perspective.

From the results of our studies, it seems that a lower volume of this exercise can be as sufficient as high volume in improving the peak eccentric hamstring strength measured on a NordBord. On the contrary, the effect of both volume training types on eccentric peak torque seems to be inconsistent in findings. The variability of different testing methods on isokinetic dynamometry and small correlations with other methods introduce challenges in comparing eccentric strength outcomes. Practitioners must consider the low validity of NHE when measuring the eccentric strength on an isokinetic device.

In recent decades, the introduction of 2D image ultrasound has allowed the study of muscle architecture (Blazevich et al., 2006). This cost-effective and time-saving noninvasive method has helped to expand the assessment of muscle thickness, pennation angle, and fascicle length, particularly in the BF^lh, which is the most researched muscle regarding the HSI. However, the use of two-dimensional ultrasound in estimating fascicle length has been presented with some methodological limitations. Entire BF^lh fascicles are too large for the field of view, and thus, an estimation of fascicle length is required through equation (Kellis et al., 2009).

Eccentric training appears to elicit greater increases in muscle CSA than concentric or traditional resistance training (Douglas et al., 2017). The mechanism seems to be a high level of mechanical tension per active motor unit (Prilutsky, 2000), stretch-induced strain (Toigo and Boutellier, 2006), and a greater propensity for exercise-induced muscle damage (McHugh, 2003), which may stimulate the hypertrophic signaling response to a greater extent. Increasing distal muscle hypertrophy with eccentric training results in increasing muscle CSA via the addition of sarcomeres in series in contrast to the addition of sarcomeres in parallel with concentric training (Franchi et al., 2014). This fact can explain the increase in fascicle angle after the eccentric (Duclay et al., 2008; Leong et al., 2014) as well as concentric training (Franchi et al., 2017). Simultaneously, stretch-induced strain from eccentric contractions, sensed within the Z-line region of titin, appears to elicit a specific anabolic signaling response (Franchi et al., 2017).

The influence of NHE on muscle architecture in our review was investigated in 5 of 11 articles (Lovell et al., 2018; Vianna et al., 2021; Medeiros et al., 2020; Mendiguchia et al., 2020; Siddle et al., 2024). Four included articles examined the effect of high volume in NHE, and one article (Siddle et al., 2024) examined the effect of the low volume protocol. Previous studies that investigated the effect of low volume on muscle architecture were mostly conducted on recreationally active males but not elite athletes (Presland et al., 2018). A similar exercise program (2–4 sets of 6–12 reps) was used across the studies where high volume has been presented. The difference was only in program duration (6–12 weeks). Only one study used a protocol with a low volume program (144 reps in 8 weeks) (Siddle et al., 2024).

Fascicle length of BF^lh is the most examined variable regarding the muscle architecture assessment. Recent studies examined that possessing BF^lh fascicles <10.56 cm significantly increased the future risk of HSI ∼4 times in elite Australian soccer players (Timmins et al., 2016). On the other hand, an increase in FL of 0.5 cm decreases the risk of hamstring strain by ∼ 74% on average. The influence of a high volume training program on this variable was present in 4 of 5 studies (Lovell et al., 2018; Vianna et al., 2021; Medeiros et al., 2020; Mendiguchia et al., 2020), while only one study used a low volume training program (Siddle et al., 2024). A significant increase by + 6.4% has been found after 8 weeks (456 reps) in elite female players (Vianna et al., 2021). An small increase of FL was found in professional male players with the same program and duration with low (1x per week) and high training frequency (2x per week) (Medeiros et al., 2020). This variable also increased by +7.38% after 6 weeks in the elite soccer players (Mendiguchia et al., 2020) and by +1.58% (small ES) after 12 weeks in amateur soccer players (Lovell et al., 2018) performing NHE before training sessions but not after. Low volume NHE programs elicit significant increases in FL of BF^lh after 6 weeks in recreational athletes with no previous NHE experience (Presland et al., 2018), while no meaningful difference between the low and high volume groups has been found. Contrary to that, no significant change was found after the low volume intervention in elite male academy soccer players (Siddle et al., 2024). One possible explanation is that the protocol has been done on elite young soccer players with previous exposure to NHE; therefore, to elicit significant improvements, the exercise intensity had to be higher (Siddle et al., 2024). Low volume also increased the FL of BF^lh and SM after 6 weeks in elite youth soccer players, but this change can be influenced by the inclusion of the second exercise (bilateral stiff-leg deadlift), which is more hip dominant movement and elicits greater activation of BF^lh, opposite to NHE (knee dominant movement), where more ST muscle activity has been presented (Bourne et al., 2017).

Three of our included studies (Lovell et al., 2018; Mendiguchia et al., 2020; Medeiros et al., 2020) investigated the influence of high volume NHE on the MT. Small and medium increases in MT were presented after 6 and 8 weeks, respectively (Mendiguchia et al., 2020; Medeiros et al., 2020). The same result was presented with the group performing the NHE after, but not before, the soccer session (Lovell et al., 2018), where the compliance rate was presented below 50% in all groups; interestingly, this fact did not affect the training outcome as stated previously with higher volumes of NHE (Chesterton et al., 2021; Chebbi et al., 2022). This is not in line with results from the meta-analysis by Cuthbert et al. (2020), where the authors found no meaningful changes in MT. The authors assume that the problem can be in low training duration (8 weeks = <), though this did not affect the training outcome in the study where the lower training duration (6 weeks) has been implemented (Mendiguchia et al., 2020). On the contrary, MT of SM, ST, and BFlh did not change after the low volume program (Siddle et al., 2024), which is also probably influenced by lower intensity of exercise and application of programme on previously NHE trained players.

PA increased after a high volume intervention program in duration of 8 weeks in “after training group” (small ES) (Lovell et al., 2018) and 6 weeks (ES = 0.41) (Mendiguchia et al., 2020). No significant improvements were found in the angle of pennation after 8 weeks in high volume groups (Medeiros et al., 2020) and “before training group” (Lovell et al., 2018). Similarly, no significant difference in PA was found after low volume intervention with 8 weeks duration (Siddle et al., 2024). High volume resistance training is primarily responsible for increasing PA and CSA (Cuthbert et al., 2020). A decrease in PA as a desired outcome after eccentric resistance training (Gérard et al., 2020) was presented in meta-analysis by Cuthbert et al. (2020) in both low and high volume groups. On the other hand, from our review, we observed only a trivial increase in the angle of pennation after high volume training, while no changes were seen after low volume training application. The increase results in a higher physiological CSA with the addition of myofibrils in parallel. This essentially improved force transmission through the muscle-tendon unit and a higher architectural gear ratio (Douglas et al., 2017; Azizi and Roberts, 2014). The ratio allows the pennate biceps femoris to limit the strain from active fascicles, which results in protection during fast-velocity lengthening actions (Azizi and Roberts, 2014). Mechanical tension and intramuscular metabolic stress determine the hypertrophic signal of the muscle during high-intensity resistance training (Douglas et al., 2017). Although the mechanical tension applies in both types of volumes, low volume of NHE seem to not elicit the intramuscular metabolic stress enough, resulting in no hamstring architecture changes.

Despite the methodological differences in our review (total volumes, duration, playing levels), the high volume of NHE seems to increase the FL of hamstring muscles and MT with trivial but non significant changes observed in PA. Contrary to this, low-volume protocol yields no significant adaptations on these parameters. Intensity of exercise is a significant factor that must be considered when the goal is to further elicit the outcomes in previously trained players. We emphasize that more research must be done on this particular topic because of the lack of literature investigating the effect of low-volume NHE protocols on hamstring architecture in soccer players. Summary of the gaps, limitations and suggestions for the future studies are mentioned in Table 4.

We suggest that coaches may prioritize the high volume of NHE in the pre-season period in order to improve architectural adaptation of the hamstring muscles and, when starting with novice athletes, to increase compliance and decrease DOMS. Novice athletes should first use low-intensity exercise (i.e., band-assisted NHE, decreasing ROM) and gradually increase the intensity to perform exercise in full range of motion. Low volume of exercise can be used within the in-season period, where a high match schedule is presented to decrease compliance with exercise. Progressive overload must be used in experienced players with the aim to gradually increase the hamstring strength and architecture of hamstring muscles. Additionally, coaches should try to find the minimal effective dose of NHE, respecting the desired outcomes of exercise, compliance, and strength level of the players.