English-language searches of the electronic databases and search engines Medline (PubMed), Scopus, Web of Science, Cochrane Central Register of Controlled Trials (CENTRAL), and Google Scholar were conducted from inception to February 15, 2023, by two independent researchers (T.A. and R.V.). The reference lists of included studies were searched to locate any further relevant articles not found with the initial search.

Articles were retrieved from electronic databases and search engines combining the following terms: (handgrip strength OR grip strength OR physical function OR sarcopenia) AND (resistance training OR strength training OR home-based exercise OR power training OR elastic band) AND (healthy adults OR elderly OR older people OR community-dwelling). No filters were applied to the searched databases to prevent omitting irrelevant articles.

Initially, all files were extracted from databases in either RIS format (Scopus, Web of Science, CENTRAL, and Google Scholar) or nbib format (Medline). The files were then uploaded into Rayyan software, where the titles and abstracts of identified articles were checked for relevance. Subsequently, the reviewers independently reviewed the full text of potentially eligible papers. Any disagreements between the reviewers on inclusion were resolved by a consensus between both researchers (T.A. and R.B.V.). After that, all files selected for inclusion were retrieved from Rayyan software and uploaded into Mendeley software, which was used as a reference management tool to write the first draft of this manuscript.

The PICOS (population, intervention, comparison, outcome, and study design) framework [25] was used to guide this systematic review. Population: Healthy individuals (≥18 years) with and without sarcopenia (low handgrip strength, slow walking speed, and low muscle mass). Intervention (exposure): Different types of resistance training interventions with any session duration (e.g., 30 minutes, 45 minutes), and any weekly frequency (e.g., number of days per week). Comparison: Non-intervention control group. A group of individuals who were not exposed to any exercise or active intervention. Outcome: Changes in handgrip strength. Study design: Any randomized or non-randomized clinical trials comparing different types of resistance exercise intervention versus a non-intervention control group on handgrip strength. Studies enrolling individuals with obesity and/or chronic diseases (e.g., heart disease, diabetes, cancer, chronic lung disease, stroke, Alzheimer, chronic kidney disease) were excluded from this review.

Randomized clinical trials were included in the review if they met the following selection criteria: 1) a research question on the effects of a resistance training intervention, 2) adults or older adult participants without chronic disease (e.g., heart disease, diabetes, cancer, chronic lung disease, stroke, Alzheimer, chronic kidney disease), 3) compared the resistance training intervention with a non-intervention control group, 4) reported at least one outcome related to handgrip strength, and 5) written in English language. Studies were excluded based on the following file types: abstracts, study protocols, conference papers, books, book sections, theses, opinion articles, observational studies, letters to editor, and reviews. Furthermore, studies that used combined interventions (e.g., resistance training plus any other type of intervention [drug, nutritional supplement…]) were excluded from this systematic review. To address our main purpose, studies applying only handgrip strength training were excluded from this review. Comparison groups and study types were not included in the search strategy but were used as inclusion criteria.

The following study characteristics were extracted: authors, publication year, study design, participants’ characteristics (sample size, age, sex, and health status), changes in handgrip strength, device used to test handgrip strength, and characteristics of the exercise intervention program (type and intensity of exercise program, exercise frequency, and duration of intervention program). These data were extracted manually and independently by two researchers (T.A. and R.V.), with disagreements resolved by consensus between both researchers. All data were typed into an excel spreadsheet file and later manually transferred to a word file. When the data reported in the articles were insufficient, additional information was requested from the corresponding authors.

Two authors (R.V. and S.D.) independently assessed the risk of bias in randomized and non-randomized included studies using version 2 of the Cochrane risk-of-bias tool for randomized trials (RoB 2) [26] and the Risk of Bias In Non-randomized Studies of Interventions (ROBINS-I) [27] , respectively. RoB 2 assess randomized trials in the following aspects: 1) bias arising from the randomization process, 2) bias due to deviations from the intended interventions, 3) bias due to missing outcome data, 4) bias in the measurement of the outcome, and 5) bias in the selection of the reported results. The overall risk of bias was expressed as “low risk of bias” if all domains were rated as low risk, “some concerns” if some concern was raised in at least one domain but not rated as high risk in any other, or “high risk of bias” if at least one domain was rated as high risk or has several domains with some concerns [26] . ROBINS-I assess non-randomized trials in the following aspects: a) bias due to confounding, b) bias in selection of participants into the study, c) bias in classification of interventions, d) bias due to deviations from intended interventions, e) bias due to missing data, f) bias in measurement of outcomes, g) bias in selection of the reported result [27] . Traffic light and weighted summary risk-of-bias plots for randomized and non-randomized included studies were produced by the online Risk of bias (robvis) tool (https://mcguinlu.shinyapps.io/robvis/). Any discrepancies were resolved through discussion between both researchers (R.V. and S.D.).

The available data did not allow us to conduct the pre-planned meta-analyses. Thus, only descriptive statistics were performed to summarize data, including the main participants’ characteristics, interventions characteristics, handgrip measurements, and main results reported by the included studies.

Twenty studies were included in this systematic review [28] - [47] . Figure 1 presents the flow of papers through the study selection process. The included studies were published from 1995 [42] up to 2021 [34] , in which six are randomized controlled trials [28] [34] [36] [40] [45] [47] , ten are randomized trials [30] [31] [32] [33] [35] [37] [39] [42] [43] [46] , one is cluster randomized controlled trial [29] , and three are non-randomized trials [38] [41] [44] (Table 1).

Participants’ characteristics are summarised in Table 1. Most of the included studies (95%, n = 19) were conducted with older adults [28] - [42] [44] [45] [46] [47] , while only one study was conducted with young adults [43] . Almost half (45%, n = 9) of the included studies clearly stated that were conducted with healthy individuals [30] [31] [32] [33] [36] [37] [39] [42] [43] , the remaining studies were conducted with older adults without experience in resistance training [38] [40] [45] , older women with cognitive impairment [35] , prefrail and frail older adults [44] , sedentary older men [34] , community-dwelling older adults receiving home care [29] , community-dwelling and independent older adults [47] , sarcopenic and recreationally active older adults [46] , postmenopausal women [28] , and older inner-city African American women [41] .

The number of participants in each study varied from 22 [41] to 419 [37] . Eight studies examined exclusively women [28] [33] [35] [36] [38] [39] [41] [42] , one exclusively men [34] , whilst 11 studies assessed men and women [29] [30]

[31] [32] [37] [40] [43] [44] [45] [46] [47] .

Resistance training programmes are summarised in Table 2. Most of the studies (70%, n = 14) applied one resistance training intervention [29] [31] [32] [33] [35] [38] [39] [41] - [47] , five studies (25%) applied two different resistance training interventions [28] [30] [34] [36] [37] , and the remaining study applied four different resistance training interventions [40] . Resistance training protocols were composed of heavy or moderate intensity or slow eccentric/concentric resistance exercises with rubber bands, elastic band, water canes and/or own body weight [29] [31] [37] [45] [46] , whole-body resistance exercises [39] [44] , home-based resistance exercises [42] [43] , functional-task exercises [28] [36] , suspension resistance exercises [33] [34] , chair-based elastic resistance exercises [35] [41] , traditional moderate/high-intensity resistance exercises [32] [47] ,

RT: resistance training. CON: control group. EBFT: element-based functional training. TSBFT: task-specific-based functional training. FT: function training. MJ: multi-joint resistance training. MJ+SJ: multi-plus single-joint resistance training. SE: supplemented and trained. SN: supplemented and non-trained. NE: non-supplemented and trained. NN: non-supplemented and non-trained. HVLL1: high-velocity, low-load once-weekly. LVHL1: low-velocity, high-load once-weekly. HVLL2: high-velocity, low-load twice-weekly. LVHL2: low-velocity, high-load twice-weekly. F: female. M: male. FM: female and male, m: meters. ^aData presented as mean (standard deviation) or median [interquartile range] or amplitude (minimum – maximum). ^bTotal sample size was 41 individuals (27 females and 14 males), but only females were enrolled in the interventions (resistance training [n = 15] or control group [12] ). ^cStandard error was converted to standard deviation.

high-speed power exercises [38] , high-velocity low-load and low-velocity high-load resistance exercise,40 and low volume multi-joint resistance exercises or a combination of multi- and single-joint resistance exercises [30] .

RT: resistance training. CON: control group. EBFT: element-based functional training. TSBFT: task-specific-based functional training. FT: functional-task exercise. ST: suspension training. TT: traditional training. MJ: multi-joint resistance training. MJ+SJ: multi- plus single-joint resistance training. SE: supplemented and trained. SN: supplemented and non-trained. NE: non-supplemented and trained. NN: non-supplemented and non-trained. HVLL1: high-velocity, low-load once-weekly. LVHL1: low-velocity, high-load once-weekly. HVLL2: high-velocity, low-load twice-weekly. LVHL2: low-velocity, high-load twice-weekly. RPE: rating of perceived exertion. 1RM: one-maximum repetition. AT: aerobic training. HR_reserve: reserve heart rate. min: minute. sec: seconds. ^aTotal sample size was 41 individuals (27 females and 14 males), but only females were enrolled in the interventions (resistance training [n = 15] or control group [12] ). ^bThe final testing, one year later, included 20 exercisers and 18 control subjects.

Intervention duration ranged from four weeks [41] to one year [31] [32] [37] [39] , with 12 weeks being the most common (35%, n = 7) [33] [34] [36] [38] [42] [45] [46] . More than half of the resistance training protocols (60%, n = 12) were performed thrice a week [28] [30] [34] [36] [37] [38] [39] [41] [42] [43] [46] [47] . Five protocols were performed two times per week [29] [31] [32] [33] [44] , one protocol was performed one to two times per week [40] , one protocol was performed two to three times per week [35] , and one protocol was performed five times per week [45] . Session duration ranged from 15 [45] to 60 minutes [31] [32] [33] [34] [36] [38] [39] , with 60 minutes being the most common (35%, n = 7), followed by 50 minutes (15%, n = 3) [28] [41] [47] . Six studies (30%) did not clearly report the session duration [30] [37] [40] [43] [44] [46] .

The number of sets per exercise ranged from one [35] to 10 [31] [32] , with three sets (50%, n = 10) being the most common [34] [36] - [43] [46] (Table 2). Most of the studies (75%, n = 15) adopted a range of eight to 15 repetitions per set [28] - [36] [39] [41] [43] [44] [45] [47] .

The intensity of effort for resistance training protocols was mostly prescribed and monitored by the percentage of one-repetition maximum (35%, n = 7) [37] [38] [39] [40] [44] [46] [47] , and rating of perceived exertion (20%, n = 4) [28] [34] [35] [36] . The remaining studies used participants’ body weight, rubber bands, rice bags, or dumbbells [33] [41] [42] [43] [45] or encouraged the participants to perform the repetitions until fatigue/momentary failure [29] [30] [31] [32] .

Half of the studies (50%, n = 10) did not clearly report the rest interval between sets and/or exercises [28] [29] [31] [32] [37] [39] [41] [42] [43] [45] . Five studies (25%) applied a one-minute rest interval between sets [30] [33] [34] [40] [46] , three studies (15%) applied two minutes [36] [38] [47] , and one study applied 45 seconds [35] . Three studies clearly reported a rest interval between exercises of three minutes [38] [40] and only one study reported two minutes [44] .

Sixteen studies (80%) provide supervision for all training sessions [28] - [38] [40] [41] [44] [46] [47] , one study for the first three months of one-year intervention period [39] , one study for one of three weekly sessions [42] , and one study for only clinic session, but not home sessions [45] . The remaining study [43] did not clearly report the information about supervision.

Settings of the handgrip strength measurements are summarised in Table 3. Eighteen studies (90%) used electronic, digital, or mechanical hand dynamometers, while the remaining two studies [28] [42] did not clearly report what instrument was used to measure handgrip strength. Half of the included studies (50%, n = 10) did not clearly report which position (e.g., standing or sitting) and elbow angle were adopted for handgrip strength measurement [29] [31] [32] [36] [37] [39] [40] [41] [42] [47] , seven studies (35%) adopted a sitting position with a 90˚ elbow flexion position [28] [30] [33] [34] [38] [43] [44] , and three studies (15%) adopted a standing position [35] [45] [46] , in which two of these three studies asked for participants to keep their upper limbs along the side of the body [35] [46] , and one study did not report the arm and/or elbow position [45] .

Most of the studies (55%, n = 11) measured both left and right participants’

RT: resistance training. CON: control group. EBFT: element-based functional training. TSBFT: task-specific-based functional training. FT: function training. ST: suspension training. TT: traditional training. MJ: multi-joint resistance training. MJ+SJ: multi- plus single-joint resistance training. SE: supplemented and trained. SN: supplemented and non-trained. NE: non-supplemented and trained. NN: non-supplemented and non-trained. HVLL1: high-velocity, low-load once-weekly. LVHL1: low-velocity, high-load once-weekly. HVLL2: high-velocity, low-load twice-weekly. LVHL2: low-velocity, high-load twice-weekly. Note: only resistance training and control groups were included in this table? authors did not clearly reported the between-group statistics. ↑: increased. ↓: decreased. ↔: not changed/different.

handgrip strength [28] [30] [31] [35] [36] [38] [40] [42] [43] [44] [46] , five studies (25%) measured only dominant participants’ handgrip strength [32] [33] [34] [39] [47] , one study (5%) measured only right participants’ handgrip strength [45] , one study (5%) measured participants’ preferred arm [29] , and the remaining two studies (10%) did not clearly report which hand (e.g., left, right or both and/or dominant, non-dominant or both) was used to measure handgrip strength [37] [41] .

Twelve studies (60%) reported no significant difference in handgrip strength change between the resistance training group and control group following an intervention study [28] [29] [30] [31] [32] [36] [37] [39] [44] [45] [46] [47] . Two studies (10%) included results for multi-training groups and found increased handgrip strength compared to controls, but only in one training group [34] [40] . Two studies (10%) measured the handgrip strength of the right and left or dominant and non-dominant hands and reported a training effect on one hand but not on the other [38] [43] . Two studies (10%) showed increased handgrip strength in the resistance training group compared with the control group [35] [42] . Finally, two studies (10%) did not clearly report differences in intervention effects [33] [41] .

Overall, the randomized and non-randomized clinical trials presented moderate (“some concerns”) risk of bias (Figure 2(a) and Figure 2(b), respectively). Among the randomized trials in the risk of bias assessment, only three studies (17.6%) reported that the allocation sequence was concealed until participants were enrolled and assigned to interventions [33] [34] [47] . Only four studies (23.5%) used blind assessors [28] [36] [37] [47] . The remaining studies (n = 13) did not blind the assessors, or this information was unclear. Only two studies analyzed the data in accordance with a pre-specified plan [37] [47] . Among the three non-randomized studies included in the risk of bias assessment, none of them used blind assessors [38] [41] [44] . All the non-randomized studies presented a low risk of bias in the classification of interventions due to deviations from intended interventions. Due to the characteristics of the intervention studies, none of the randomized and non-randomized studies could blind participants and personnel (trainers). Supplementary Material shows traffic light risk-of-bias plots for randomized and non-randomized included studies.

Handgrip exercise training may improve handgrip strength in middle-aged and older adults [48] [49] , but this systematic review did not include studies involving such exercise programs. However, when resistance exercise is offered using resistance training machines, study participants sit on a chair. The participants’ hands often grip a bar to maintain body position during the exercise. Even when training with a rubber band, participants may hold onto one end of the band during exercise. This type of exercise makes determining exercise intensity or contraction time difficult but indicates an indirect handgrip exercise. In this systematic review, twelve of the 20 selected studies found no difference in handgrip strength changes between the resistance training and control groups. Most of those studies employed moderate- to high-intensity resistance exercises using resistance training machines and rubber/elastic bands [28] [29] [30] [31] [32] [36] [37] [39] [44] [46] [47] . On the other hand, two studies that reported a significant increase in handgrip strength in the resistance training group compared to the control group involved training programs using their body weight and rubber/elastic bands [35] [42] . These results did not explain the difference in the impact of resistance training on handgrip strength due to differences in exercise modes. Furthermore, there were no differences in other training variables, such as the volume of exercise (number of repetitions and sets) and intervention period, depending on whether they affected handgrip strength. Participants in the two studies [35] [42] that observed a significant increase in handgrip strength with resistance training were older adults with a mean age of about 80. Of the two, in the study where resistance training had the most change in handgrip strength, an increase of approximately 3 kg was observed in the training group [35] . While Labott and colleagues [19] recently concluded in a meta-analysis that different types of exercise training were capable of increasing handgrip strength compared to different control groups (e.g., other exercise interventions or non-exercise control groups), the observed effect size was small. Of the studies included in the analysis, Labott and colleagues [19] observed that only four of the 24 included studies found statistically significant increases in handgrip strength relative to the control group; however, only one of these four studies in fact compared resistance training intervention to a non-exercise control group. Thus, had we been able to perform a meta-analysis, it is possible that pooling all studies together would demonstrate a statistically significant effect of resistance training on handgrip strength relative to the control group, but the effect size would be expected to be small.

When a single study includes two or more training groups, and there is a difference in handgrip strength change between the groups, knowing the factors behind this difference is meaningful from the perspective of handgrip strength improvement strategies. Our selected studies included two [34] or four [40] training groups that found increasing handgrip strength compared to controls in only one training group within each study. Campa and colleagues [34] compared the impact of suspension and traditional resistance training on handgrip strength and found that only suspension training produced increasing handgrip strength. The elastic bands employed in the traditional training program used different tube sizes specific to the given exercise. The suspension training was carried out using gripping straps attached to the tip of the elastic tube, which helped to grip firmly. A predicted factor for the difference in impact on handgrip strength could be attributed to the need for repeated firmer grip during the suspension exercise. Richardson and colleagues [40] observed the impact on handgrip strength when resistance training was performed in eight whole-body exercises (four in the upper body and four in the lower body) at high load (80% 1RM)-low velocity or low load (30% 1RM)-high velocity. In addition to each load-velocity condition, four training conditions differing in frequency (once a week vs. twice a week) were compared. As a result, handgrip strength increased only under the training program with high load-low velocity twice a week. The reasons for these results are unclear, but some possibilities exist. When performing high-load, low-velocity exercises using training machines, the time required to grip the movable bar during upper-body exercise is more extended than under other conditions. For lower-body movements, the time needed to hold the bar to stabilize the body is also longer than other conditions. Training load, volume, and frequency in resistance training using machines may impact the grasping movements of the machine’s bar, which may train handgrip strength indirectly. However, this issue has yet to be investigated.