Effects of range of motion on resistance training adaptations: A systematic review and meta-analysis

2.1 Registration of systematic review protocol

This systematic review and meta-analysis was conducted according to the Cochrane Handbook for Systematic Reviews of Interventions³⁶ and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.³⁷ The review protocol was preregistered in PROSPERO (CRD42020160976).

2.2 Eligibility criteria

The PICOS (population, intervention, comparators, outcomes, study design) criteria for the eligibility of studies³⁸ were used to determine the inclusion and exclusion criteria.

2.3 Identification and selection of studies

English language articles were included at the screening level. To ensure a relatively complete census of relevant literature, we performed a backward-forward search, reviewing the references and citations of studies included.⁴⁴ Moreover, a second-level backward reference search was done by pulling the references of the references.⁴⁵ Records retrieved from the database search were imported to Mendeley (v1.19.6, Elsevier, UK) and processed in Microsoft Excel 2016 (Microsoft Corporation, USA) by AHB. After the removal of duplicates, two investigators (AHB and AMC) independently screened the titles and abstracts. References not eliminated were subjected to a second-stage screening of the full text. To ensure a quality appraisal of the review process, we assessed the agreement between the two researchers using an inter-rater reliability test.⁴⁶ Discrepancies at any stage were resolved by discussion with a third investigator (JGP).

2.4 Data extraction

Two reviewers (TV and JCI) independently collected the data of all included studies using standardized forms in Microsoft Excel, including author/s, year, sample characteristics (age, sex, training experience), intervention groups (full and partial ROMs trained), configuration of the resistance training program (exercise/s, duration, frequency, relative intensity, contraction type, movement velocity, and rest intervals), dependent variables of interest (neuromuscular, functional, and structural outcomes), and assessment tests. For quantitative analyses (meta-analyses), we collected the group size and mean differences of the aforementioned outcomes with a 95% confidence interval (CI) or standard deviations (SD) for both intervention groups. Disagreements were adjudicated by JGP.

2.5 Dealing with missing data

Corresponding authors were contacted to provide missing data of relevant variables. Otherwise, data were obtained from figures when possible using WebPlotDigitizer.⁴⁷ Studies with missing mean values were excluded from the meta-analysis but discussed in the review. Missing SD were calculated or estimated from relevant statistics provided (eg, from CI, standard errors, p values) or imputed from an appropriate pretest.⁴⁸

2.6 Risk-of-bias and quality of evidence assessments

Two reviewers (JCI and AHB) used the Cochrane Collaboration Risk-of-Bias Tool (RoB 2)⁴⁹ and the GRADE (Grading of Recommendations, Assessment, Development, and Evaluations)⁵⁰ to judge the level of bias and quality of evidence. The GRADE quality rating was downgraded one level for each of the following limitations: the 95% CI includes both appreciable benefit and harm (imprecision); high variability and heterogeneity across studies (inconsistency); and the presence of high risk of bias. Disagreements were resolved by discussion with JGP.

2.7 Statistical analysis

The effect sizes (ESs) were calculated as the standardized mean differences between the full and partial ROM groups. The sample size and mean ES across all studies were used to calculate the variance around each ES. Meta-analyses were performed using robust variance estimation (RVE) with small-sample corrections.^{51, 52} RVE is a form of random-effects meta-regression for multilevel data structures, which allows for multiple effect sizes from the same study to be included in a meta-analysis, even when information on the covariance of these effect sizes is unavailable. Instead, RVE estimates the variance of meta-regression coefficient estimates using the observed residuals. It does not require distributional assumptions and does not make any requirements on the weights.^{51, 52} A study was used as the clustering variable to account for correlated effects within studies. Observations were weighted by the inverse of the sampling variance. A sensitivity analysis, using alternative correlational values to calculate the standard error, revealed that the choice of correlational value did not impact the overall results of the meta-analysis. Between-study heterogeneity was evaluated using the I² index. Values of I² more than 25%, 50%, and 75% were selected to reflect low, moderate, and high heterogeneity, respectively.³⁶ All analyses were performed using packages robumeta (version 2.0) and metafor (version 2.4-0) in R version 3.5.2 (The R Foundation for Statistical Computing, Vienna, Austria).

3.1 Study characteristics

Details from the 16 RCTs (n = 551 participants) included in the final analysis are presented in Table 1. Resistance training interventions were conducted on male-only samples in ten studies,^{54, 56-59, 61-63, 66, 67} female-only in two studies,^{55, 65} and mixed-samples in four studies.^{53, 60, 64, 68} Nine studies investigated lower-limb exercises,^{53, 54, 56, 61, 63, 64, 66-68} with five of them training the squat,^{53, 56, 63, 66, 67} two a combination of lower-limb exercises,^{61, 64} one the leg press exercise,⁶⁸ and one the knee extension exercise.⁵⁴ Seven studies performed upper-limb exercises,^{55, 57-60, 62, 65} with three studies training the bench press^{57, 59, 65} and the other the elbow flexion,⁶² elbow extension,⁵⁸ arm push-and-pull⁵⁵ and lumbar extension⁶⁰ exercises. The main strength outcome analyzed was the one-repetition maximum (1RM), ^{53, 56, 57, 59, 61-63, 65-67} followed by isometric tests,^{54, 56, 58, 60, 64, 67, 68} isokinetic evaluations,^{54, 55, 68} and barbell velocity assessments.^{59, 63} Functional tests included vertical jumps,^{53, 56, 61, 63, 67} sprints,^{61, 63} Wingate anaerobic test⁶³ and flexibility assessment.⁶⁰ Four and two studies assessed changes in muscle hypertrophy using the cross-sectional area^{54, 56, 58, 64} and muscle volume,^{54, 66} respectively. Five studies included variables of muscle architecture measured by means of ultrasound scans, including the pennation angle,^{54, 56, 64, 68} fascicle length,^{54, 64, 68} and muscle thickness.^{56, 62, 68} Results for the GRADE certainty of the evidence of particular outcomes are presented in Table 2.

3.2 Quality of studies and risk of bias

A summary of the risk-of-bias assessment is shown in Figure 2. No study was considered as a low risk of bias in all categories. The greatest biases were found in the randomization process, measurement of the outcomes, and selection of the reported results. No study provided a trial preregistration. Three studies showed a high risk of bias in the selection of the reported results since they only presented one dependent variable.^{55, 57, 65}

3.3 Muscle strength

Meta-analysis showed that exercise training at full ROM produced a significantly greater effect on the muscle strength with moderate effect size (ES [95% CI] = 0.56 [0.20 to 0.91], p  =  0.004, I²  =  77.6%, studies: n  =  16, Figure 3). Specifically, separate variables analysis (Table 2) revealed that the full ROM training produced significantly greater improvements than partial ROM in 1RM Full ROM lower-limb strength (ES = 1.53, n  =  6, p = 0.001) and nonsignificant but notably higher enhancements in 1RM Full ROM upper-limb strength (ES = 0.69, n  =  4, p = 0.078) and isometric lower-limb strength (ES = 0.74, n  =  5, p = 0.194).

3.4 Functional performance

Meta-analysis showed that training at full ROM produced a greater but not-statistically significant effect on the functional performance (ES [95% CI] = 0.44 [−0.32 to 1.20], p  =  0.186, I²  =  63.1%, studies: n  =  5, Figure 4). Likewise, separate variables analysis (Table 2) revealed greater but not-statistically significant improvements in jump capability after full ROM training (ES = 0.55, n  =  4, p = 0.164) (Table 2). No conclusive evidence was found for the sprint time and Wingate test.

3.5 Muscle hypertrophy

Meta-analysis showed that exercise training at full ROM produced significantly greater muscle hypertrophy on lower-limb muscles, compared to partial ROM training (ES [95% CI] = 0.88 [0.19 to 1.57], p = 0.027, I²  =  80.3%, studies: n  =  4, Figure 5).

3.6 Muscle architecture

Meta-analysis showed no large differences in muscle architecture (ES [95% CI] = 0.28 [−0.26 to 0.82], p = 0.226, I²  =  74.6%, studies: n  =  5, Figure 6). No conclusive evidence was found when variables of muscle architecture were analyzed separately, although fascicle length tended to be favored by the full ROM training (ES = 0.87, n  =  3, p = 0.327).

4.1 Muscle strength

Full ROM repetitions during resistance training were found more effective than partials to enhance muscle strength, particularly lower-limb 1RM Full ROM (ES = 1.53, p = 0.001). The results seemed to be homogeneous in a broad variety of exercises (squat, knee extension, bench press, elbow flexion, arm push-and-pull, and lumbar extension).

Traditionally, resistance training at partial ROM has been suggested as a good strategy to reduce neural inhibition and improve the coordination of primary and stabilizing muscles.^{69, 70} However, this meta-analysis has not found any longitudinal intervention that supports these superior neural benefits in favor of the partial ROM. Moreover, partial ROM resistance training has been believed to produce greater strength adaptations, since it allows us to lift a higher absolute weight, as a result of evading the critical region of the movement (ie, the sticking region).^{25, 71} However, this was not supported by the current meta-analysis, with most of the studies reporting greater neuromuscular adaptations after a full ROM training, both in the upper ^{59, 60, 62, 65} and lower limbs,^{56, 63, 64, 67} even using lower absolute loads (ie, kg) (Table 1, Figure 3). The sticking region would be caused by an interaction between the muscle force-length relationship and the external torque.³¹ On this matter, the sticking region would be the zone at which the maximal amount of contractile material is involved, due to two main reasons: i) the optimal (or close to optimal) muscle length (ie, not excessive stretched or contracted position of the sarcomeres),³⁰ and ii) the minimal velocity (ie, the number of cross-bridges attached increases as the shortening velocity decreases).^{72, 73} Therefore, the fact that partial repetitions systematically avoid this zone of maximal active tension could be the reason behind the lower effectiveness of partial ROM in enhancing strength.^{59, 63} However, future research is needed to understand the kinematics and physiological mechanisms that underlie these findings. On the contrary, the lack of studies executing the partial ROM training at long muscle lengths limited the current research to examine whether there are differences between full and partial ROMs according to the muscle length trained by the latter. Specifically, except for one group that trained by using partial repetitions executed at a long muscle length,⁶⁸ and three studies that executed partial repetitions at an intermediate region of the force-length relationship,^{55, 60, 62} the rest of the investigations trained the partial ROM at short muscle lengths. Nevertheless, since the muscle length trained would be closely related to the moment arm (eg, smaller moment arms at more flexed knee angles, and so at longer muscle lengths),⁷⁴ it would be of great practical value that future investigations compare the full and partial ROM interventions including partial repetitions executed at short and long muscle lengths.

The present review found some controversy about the specificity training principle, which states that responses to training will be adapted in a similar manner to that employed during training. In this regard, some studies found that each training group obtained the greatest 1RM improvements at the specific ROM at which they trained (eg, the partial squat group achieved more 1RM enhancements in the partial squat test than in the full squat test).^{56, 61, 66, 67} Conversely, other investigations showed that, although each training group maximized the strength gains at the specific ROM they trained, the full ROM group obtained the greatest neuromuscular improvements even in the partial tests.^{59, 63} It should be taken into account that the specificity principle could be related to the learning effect of participants, after regular practice. For example, a participant who trained during weeks at a given ROM is expected to obtain greater post-intervention performance in this specific ROM as a consequence of the familiarization with the execution of the exercise.^{75, 76} An interesting approach to reduce the impact of the learning effect when interpreting the main results is the inclusion of complementary neuromuscular tests, not related to the specific resistance training performed during the intervention, for instance, maximal isometric contractions at specific angles.^{56, 64, 67} Thus, taking into account these complementary evaluations, our results continue to support the greater efficacy of the full ROM training to enhance strength gains (Table 2).

4.2 Functional performance

The choice of the optimal ROM to improve sports performance has been under discussion for decades.^{28, 77-79} According to our review, most of the research suggests full ROM resistance training as preferable to increase jump ability^{53, 56, 63, 67} (Table 2), with only one study supporting the superior effectiveness of partial ROM⁶¹ (Figure 4). However, although effect sizes favored the full ROM, the meta-analysis was not significant. Interestingly, except for Rhea et al.,⁶¹ studies reporting specific strength adaptations at the ROM trained (specificity principle) showed higher effectiveness of the full ROM training to increase jump height.^{56, 67} On the contrary, the two studies examining the sprint performance showed conflicting results^{61, 63} ; therefore, we cannot present a clear conclusion about the optimal ROM to maximize this functional capability. Furthermore, only one study examined sports abilities different from jumping or sprinting, by means of the Wingate anaerobic test, with positive results favoring the full ROM.⁶³ Future research should confirm these results.

4.3 Muscle hypertrophy

The present study found superior effectiveness of the full ROM training to produce lower-limb muscle growth (Figure 5). Our results are in line with a previous systematic review³⁵ suggesting a potential greater effect of full ROM resistance training on muscle hypertrophy, especially in the lower limbs.^{54, 56, 64, 66} It is worth noting that, except for Goto et al.⁵⁸ (muscle size measured at a single point of the muscle length), the rest of the investigations used as an indicator of muscle hypertrophy either the muscle volume^{54, 66} or CSA measurements acquired at different lengths of the target muscle (eg, proximal-medial-distal).^{54, 56, 64} Although the assessment of the muscle volume via MRI would be the gold-standard technique,⁸⁰ measuring the CSA at different points would allow researchers to identify regional changes which would be dependent on the exercise trained (eg, leg press and knee extension would maximize hypertrophy in the middle⁸¹ and distal sites³⁹ of the muscle, respectively). Therefore, the results found by the current study regarding the superior effectiveness of the full ROM in generating muscle hypertrophy would be reinforced once the evaluation techniques used by the individual studies have been considered. Nevertheless, two limitations related to the muscle hypertrophy analysis should be noted. Firstly, the scarcity of scientific evidence examining the upper-limb hypertrophy through sensitive methods (ie, muscle volume or multiple CSA measurements) limits us to provide a clear conclusion about the influence of the trained ROM on muscle growth of the upper-limb muscles. Secondly, the limited duration of the training programs designed by the studies included (mean duration = 10.4 weeks; ranging from 6 to 16 weeks) would have influenced the hypertrophy values detected by the current review. Specifically, although significant increases in muscle size have been observed after only a few weeks of training (~3 weeks),³⁹ muscle growth has been proved to be influenced by the duration of the training program in a linear fashion (ie, the longer the duration, the more muscle hypertrophy).⁸² Hence, future investigations comparing the full and partial ROMs in terms of muscle hypertrophy are encouraged to implement training programs of longer duration.

4.4 Muscle architecture

Our results revealed large disparities in the effectiveness of full or partial ROM training to modify the muscle thickness, pennation angle, and fascicle length.^{54, 56, 64} However, two of the three studies analyzing the fascicle length found superior adaptations after full ROM repetitions^{54, 64} (Table 2, Figure 6). On this matter, muscles adapt their structure by adding or removing sarcomeres as a function of different training parameters, including the range at which they are stimulated.^83-85 This may account for the higher enhancements of fascicle length after full ROM training, as a response to stimulate the muscles at lengths that exceed those required by the daily activities,^{54, 64} particularly during the eccentric phase of the movement.^{81, 86} Consequently, the changes in fascicle length would modify the muscle function due to its influence on force-length and force-velocity relationships.³⁴ Thus, a reduction in the number of sarcomeres in series would vary the joint angle where optimal force is produced during the activity (ie, altering the force-length relationship) and reduce the shortening velocity (ie, altering the force-velocity relationship).^87-89 Furthermore, having short fascicles has been related to the rise of microscope muscle damage after repetitive eccentric actions.³⁴ Therefore, athletes’ risk of injury could be reduced by training at full ROM.

This study is not exempt from limitations. Firstly, we had to estimate results from studies that only reported them graphically or lacked some specific statistic (eg, SD). Secondly, most of the meta-analyses indicated moderate to high levels of heterogeneity. This fact could be explained mainly by the different variables included in each quantitative analysis (ie, clinical diversity), as well as the different methodologies (eg, programming, volume, intensity, exercise, duration) used by each study (ie, methodological diversity). Thirdly, the scarce and contradictory results about some effects, both functional (sprint, cycling) and structural (upper-limb hypertrophy and muscle architecture), limit the present study to provide a clear conclusion about these specific adaptations.