This systematic review was conducted in accordance with the criteria established by the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA). The developed protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) under registration number CRD42022369217.

Relevant studies were identified by searching PubMed, Scopus, and Embase databases from the last 15 years from December 2007 to December 2022. The descriptor terms were associated with the Boolean operators as follows: (“strength training” OR “resistance training” OR “weight training”) AND (obes* OR “high fat diet” OR HFD) AND (rat OR mice OR mouse) AND (heart OR cardiomyocyte OR cardiac OR “left ventricle”).

For the studies eligibility, the PICOS strategy was applied as displayed in Table 1. The evaluation of the eligibility criteria was performed blindly by 2 independent researchers (AMOP and SFFC). Disagreements between researchers were discussed with a third researcher (AJN) and resolved in consensus.

Table 1

Inclusion criteria	Exclusion criteria
Population	Population
Rodents	Humans
	In silico studies
	Ex vivo studies
Intervention	Intervention
Resistance training with total duration ≥ 4 weeks	No intervention with resistance training
	Other type of exercise intervention
Comparators	Comparators
Exercised group compared with non-exercised (sedentary), both treated with HFD	Sedentary or exercised group not treated with HFD
Outcomes	Outcomes
Cardiac structure, function, oxidative stress, inflammation, mitochondrial biogenesis, and markers of tissue remodeling	No determination of outcomes with respect to cardiac tissue
Publication parameters	Publication parameters
Original study	Non-original study
Published between December 2007 and December 2022	Letters
English language	Abstracts

HFD: high-fat diet.

Data and information of interest were extracted by AMOP, SFFC, and LBL. The main information obtained were on animal characteristics (lineage, strain, sex, and age); period and composition of HFD treatment (% fat); RT protocol (model, intensity, series, rest period, weekly frequency, and total duration of intervention); and anthropometric, metabolic, and cardiac outcomes.

The risk of bias was examined conforming to the guidelines recommended in the risk of bias tool for animal studies SYRCLE (Systematic Review Centre for Laboratory Animal Experimentation).²⁶ The questions were answered with “yes” (low risk of bias), “no” (high risk of bias), or “unclear” (unclear risk of bias), according to each of the following 10 item tools: random sequence generation, baseline characteristics, allocation concealment, random housing, blinding of caregivers/investigators, random outcome assessment, blinding of outcome assessment, incomplete outcome data, selective reporting, and other bias. The evaluation was performed blindly by 2 independent researchers (AMOP and LBL). Review Manager software, version 5.4 was used to conduct this analysis and to produce the risk of bias figures.

The quality of studies was evaluated using the Collaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Studies (CAMARADES) checklist containing 10 items.²⁷ The articles were scored with one point for reporting the required information and zero points when information was missing, with a total score of maximum 10 points, with 1 to 3 points regarded as low quality, 4 to 7 points regarded as medium quality, and 8 to 10 points regarded as high quality. The evaluation was performed independently by 2 authors (AMOP and LBL).

The search resulted in 70 articles (PubMed n = 11; Scopus n = 34; Embase n = 25) (Figure 1). After excluding duplicate articles (n = 33), 37 articles were selected for the title and abstract reading. After that, 28 articles were excluded for not meeting the inclusion criteria: non-original articles (n = 14); studies without animal models (n = 1); studies without RT intervention (n = 11); studies without evaluation of the heart (n = 1); no treatment with HFD (n = 1).

Figure 1

Subsequently, 9 articles were selected for full text reading and eligibility assessment. Then, 4 articles were excluded: sedentary group not treated with HFD (n = 1); non-treatment with HFD (n = 2); and non-intervention with RT (n = 1). After exclusion, 5 articles remained and were included in the systematic review.

The results from bias analysis are shown in Figure 2. For selection bias, the use of random sequence generation for reducing selection bias was not reported in any of the reviewed articles; therefore, all articles were graded as high risk of bias. Body mass (BM), age, and sex were defined as the baseline characteristics, and the majority of the articles (n = 4; 80%) were graded as low risk. There was an unclear risk of bias for allocation concealment in 4 studies, which were thus graded as unclear on this item.

Figure 2

Regarding performance bias, because none of the articles reported whether random housing was used or whether the participants were blinded (blinding of caregivers/investigators), all articles received high risk. For detection bias, none of the studies reported whether there was a random selection of animals (random outcome assessment) or whether evaluators were blinded (blinding of outcome assessment); thus, we graded all articles as high risk. Additionally, 60% of articles received unclear risk, and 40% received high risk for incomplete outcome data (attrition bias). In relation to selective reporting (reporting bias) and other bias, all articles had low risk of bias.

According to the CAMARADES evaluation (Table 2), all studies presented medium methodological quality, with scores ranging from 6 to 7 points. Based on this, 100% of the studies were published in peer-reviewed journals, reported control of temperature, included statement of compliance with regulatory requirements, and used appropriate animal models.

Table 2

Study	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	Score
Lino et al. 202028	☑	☑	☑			☑	☑		☑		6
Effting et al. 201931	☑	☑	☑			☑	☑		☑	☑	7
Kim et al.33	☑	☑	☑			☑	☑		☑		6
Leite et al.29	☑	☑	☑			☑	☑		☑	☑	7
Melo et al.30	☑	☑				☑	☑		☑	☑	6

(1) Publication in peer-reviewed journal; (2) statement of control of temperature; (3) randomization of treatment or control; (4) allocation concealment; (5) blinded assessment of outcome; (6) avoidance of anesthetics with marked intrinsic properties; (7) appropriate animal model; (8) sample size calculation; (9) statement of compliance with regulatory requirements; (10) statement regarding possible conflict of interest.

Most studies (80%) did not describe the use of any anesthetic; only one study used anesthetic, but not to assess outcomes of interest; therefore, all studies scored 1 point. Two studies did not include a statement of potential conflict of interests, and only one did not present randomly allocated groups. None of the studies reported allocation concealment or blind evaluation of outcome, and none described sample size calculation.

Regarding the characteristics of the animals (Table 3), most studies (n = 4; 80%) used Wistar^28-30 and Sprague-Dawley rats,³⁰while only one article used Swiss mice.³²All studies were performed with male sex, and the initial age of the animals varied between 21 and 90 days.^28,29,31,32 Only those in the study by Kim et al.³³were older than the others (51 weeks).

Table 3

Study	Animal characteristics	HFD treatment	Resistance training protocol	Main effects
Lino et al. 202028	Rats (Wistar) Male 90 days of age	Feeding with HFD (20%) for 11 weeks (3 weeks before RT and 8 weeks during RT)	Model: Ladder climbing MLCT: First climbing with 75% of BM and subsequent addition of 30 g until failure Sessions: 4 climbings with 50%, 75%, 90%, and 100% of the maximum carrying capacity. If the rat reached 100% of the carrying load, an additional 30 g were added for a new maximum carrying capacity with a fifth extra climbing. Rest period between sets: 120 s Frequency: Two days interlarded by rest periods of 72 h (3 to 4 times/week) Total duration: 8 weeks	Anthropometric parameters and fat mass: ↓ BM, ↓ adiposity index Metabolic and biochemical parameters: ↓ Glycaemia, ↓ total cholesterol ↓ TGL, ↓ HDL, ↔ LDL, ↓ VLDL, ↓ rCastelli’s ratio I and II, ↓ TGL/HDL ratio Biometric properties: ↔ HW and LV weight Oxidative stress: Activity: ↔ Total-SOD, ↑ Mn-SOD, ↑ CAT, ↓ GSH, ↔ GPx, ↔ lipid peroxidation Tissue remodeling markers: Activity: ↑ Pro MMP-2, ↑ intermediate MMP-2, ↔ active MMP-2
Effting et al. 201931	• Mice (Swiss) • Male • 40 days of age	• Feeding with HFD (59%) for 26 weeks (18 weeks before RT and 8 weeks during RT)	Model: Ladder climbing MLCT: Not performed Sessions: 5 to 10 climbings per session (volume progression) with 20% to 75% of BM (load progression) Rest period between sets: 120 s Frequency: 48-hour intervals between sessions (3 to 4 times/week) Total duration: 8 weeks	Anthropometric parameters: ↓ BM Metabolic and biochemical parameters: ↓ Fasting glucose, ↑ glucose decay rate in ITT Oxidative stress: Activity: ↔ SOD, ↓ CAT Levels: ↓ DCFH, ↓ MDA, ↔ GSH Inflammatory marker levels: ↓ TNF-α
Kim et al.33	Rats (Sprague-Dawley) Male 51 days of age	Feeding with HFD (50%) for 18 weeks (6 weeks before RT and 12 weeks during RT)	Model: Ladder climbing MLCT: Not performed Sessions: 1 to 8 climbings with 70% to 100% of BM (load progression during the session). If a rat was able to climb the ladder with these loads, additional weights were placed in the cylinder in 30 g increments for each subsequent climbing. Rest period between sets: 120 s Frequency: 3 times/week Total duration: 12 weeks	Anthropometric parameters and fat mass: ↔ BM, ↔ intraperitoneal fat Biometric properties: ↔ HW, ↔ HW/BM Mitochondrial biogenesis: Protein expression: ↑ Cyto-C, ↑ SUD, ↑ PGC1-α, pAMPK/tAMPK. Endoplasmic reticulum stress markers: Protein expression: ↓ p-PERK/PERK, ↔ CHOP, ↔ GRP78
Leite et al.29	Rats (Wistar) Male 21 days of age	Feeding with HFD (30%) for 24 weeks (12 weeks before RT and 12 weeks during RT)	Model: Ladder climbing MLCT: First climbing with 75% of BM, with subsequent additional 30 g loads until failure Sessions: Four climbings with 50%, 75%, 90%, and 100% of the maximum carrying capacity. If a rat reached 100% of its carrying load, an additional 30-g load was added until a new carrying load was determined. Rest period between sets: 120 s Frequency: 3 times/week (48-hour intervals) Total duration: 12 weeks	Anthropometric parameters and fat mass: ↓ BM, ↓ fat (%), ↑ fat-free mass Biometric and hemodynamic properties: ↓ SBP, ↓ SDB, ↓ MBP, ↔ HW, ↑ HW/BM, ↔ LV Tissue remodeling markers: Activity: ↑ Pro-MMP-2, ↑ intermediate MMP-2, ↔ active MMP-2
Melo et al.30	Rats (Wistar) Male 37 days of age	Feeding with HFD (49.2%) for 26 weeks (16 weeks before RT and 10 weeks during RT)	Model: Ladder climbing MLCT: First climbing with 50% of BM, with subsequent additional 30-g loads until to failure Sessions: Climbings with 50%, 75%, 90%, and 100% of the maximum carrying capacity. If the animal reached 100% of the carrying load, an additional 30-g load was added until failure. Rest period between sets: 60 s Frequency: 3 times/week (48-hour intervals) Total duration: 10 weeks	Anthropometric parameters and fat mass: ↔ BM, ↔ body fat, ↔ adiposity index, ↔ rWAT, ↓ eWAT, ↓ visceral WAT Hemodynamic parameters: ↔ SBP, ↔ DBP Metabolic and biochemical parameters: ↔ Glucose, ↔ total cholesterol, ↔ HDL Biometric properties: ↔ Heart, ↔ HW/TL, ↔ LV, ↔ LV/TL Histological properties: ↔ CSA, ↔ collagen Cardiomyocyte contractile function: ↔ FS, ↔ time to 50% of contraction and relaxation, ↔ maximum rate of contraction and relaxation Calcium handling proteins: Protein expression: ↔ Serca2a, ↔ PLB, ↔ pPLBser16, ↔ SERCA2A/PLB, ↔ pPLBser16/PLB

↑: ncrease; ↓: decrease; ↔: unchanged; AMPK: AMP-activated protein kinase; BM: body mass; CAT: catalase; CHOP: proapoptotic C/EBP homologous protein; CSA: cross-sectional area; Cyto-C: cytochrome C; DBP: diastolic blood pressure; DCFH: dichlorodihydrofluorescein; eWAT: epididymal white adipose tissue; FS: fractional shortening; GPx: glutathione peroxidase; GRP78: glucose-regulated protein 78; GSH: glutathione oxidase; HDL: high-density lipoprotein; HFD: high-fat diet; HW: heart weight; ITT: insulin tolerance test; LDL: low-density lipoprotein; LV: left ventricle; MBP: mean blood pressure; MLCT: maximal load carrying test; MDA: malondialdehyde; MMP-2: metalloproteinase-2; Mn-SOD: mitochondrial superoxide dismutase; pAMPK: phospho-AMPK; PERK: PKR-like endoplasmic reticulum kinase; PGC1-α: peroxisome proliferator-activated receptor-γ coactivator 1-α; PLB: phospholamban; pPLBser16: PLB serine16 phosphorylation; RT: resistance training; rWAT: retroperitoneal white adipose tissue; SBP: systolic blood pressure; SERCA2a: sarcoplasmic reticulum Ca²⁺ ATPase; SUD: succinate dehydrogenase; tAMPK: total AMPK; TGL: triglycerides; TL: tibial length; TNF-α: tumor necrosis factor-alpha; VLDL: very-low density lipoprotein; WAT: white adipose tissue.

The fat percentage of HFD ranged between 20% and 59% (Table 3). The total duration of the HFD feeding protocol ranged between 11 and 26 weeks. In all studies, the animals received HFD for a period before the start of RT, which varied between 3 and 18 weeks.^28,29,31-33 The HFD was subsequently maintained during the RT period until the end of the experiment, which lasted 8,^28,32 10,³³and 12 weeks.^29,33

Table 3 shows the characteristics of the RT protocols used in the selected studies. Regarding exercise mode, climbing a vertical ladder was used in all studies. In 3 studies (60%), prior to the beginning of RT, the animals were submitted to maximal load carrying test (MLCT) to determine the load used in the RT sessions. In these studies, MLCT consisted of climbing the stairs with a load equivalent to 50%³²and 75%^28,29of BM, and in subsequent climbs, 30-g loads were added until the point at which the animal reached muscle fatigue. In the other studies (n = 2), BM was used as the basis for determining the RT load; therefore, MLCT was not used.

In 3 studies, the RT session consisted of 4 climbs with 50%, 75%, 90%, and 100% of the MLCT; if a rat reached 100% of its MLCT, an additional 30-g load was added until subsequent climb (without fixed volume). In studies by Leite et al.²⁹and de Souza Lino et al.,²⁸30-g loads were added for each climb until a new carrying load was determined as the RT load. On the other hand, Melo et al.³²readjusted the RT load with a new MLCT; however, the animals started the first climb with the equivalent of 50% of the carrying load from the prior RT session.

In 2 studies, the animals were not submitted to the MLCT; therefore, they performed the stair climbing with weight equivalent to their BM. In the study by Effting et al.,³¹ the animals climbed the ladder with a load equivalent to 20% of BM in weeks 1 and 2, 50% from the third to the sixth week, and 75% in the seventh and eighth. The number of series varied during the RT protocol as follows: in weeks 1 and 3, the animals performed 5 series; in weeks 2, 4, and 7, they performed 7 series; and in weeks 5, 7, and 8, they performed 10 series. In the study by Kim et al.,³³ in the first week the animals climbed with 30% to 50% of their BM, and the weights and the number of repetitions were incrementally increased, but not specified. From the second week until the completion of the exercise program, set 1 was conducted with weights of 70% of BM, sets 2 and 3 with weights of 80%, sets 4 and 5 with weights of 90%, and sets 6 to 8 with weights of 100%. If a rat was able to climb the ladder with these loads, additional weights were placed in the cylinder in 30-g increments for each subsequent climb.

In only one study, the rest period between sets was 60 seconds,³³ while in the others 120 seconds were used.^28,29,31,33 In 3 studies, the frequency of RT was 3 times a week.^29,32,33In the study Effting et al.,³¹ RT sessions were separated by 48 hours; therefore, the weekly frequency ranged from 3 to 4 times, totaling 28 sessions. In de Souza Lino et al.,²⁸the rats performed the RT protocol for 2 days interlarded by rest periods of 72 hours (weekly frequency from 3 to 4). The total duration of the RT intervention ranged between 8 and 12 weeks.

Regarding anthropometric parameters and fat mass, RT reduced the final BM in 3 studies,^28,29,31 which followed a reduction in body fat percentage and increase in fat-free mass,²⁹ although no changes in the adiposity index were observed.²⁸ Furthermore, in another study, RT did not change the BM and adiposity index, but reduced epididymal and visceral fat.³² In the study by Kim et al.,³³RT did not reduce the final BM and intraperitoneal fat (sum of epididymal, mesenteric, and retroperitoneal fats).

Additionally, RT improved insulin resistance,³³ reduced blood glucose,^{28, 33} and counteracted the increase in total cholesterol, triglycerides, HDL, and VLDL induced by HFD feeding.²⁸ Additionally, it reduced the cardiometabolic markers Castelli’s ratio I (total cholesterol/HDL), II (LDL/HDL), and the triglycerides/HDL ratio.²⁸ In another study, HFD feeding and RT did not affect fasting glucose, lipid markers (e.g., total cholesterol and HDL), and systolic and diastolic blood pressure.³²On the other hand, in the study by Leite et al.,²⁹ RT neutralized the increase in systolic, diastolic, and mean blood pressure in response to HFD feeding.

Furthermore, RT reduced levels of inflammatory markers (e.g., TNF-α) and cardiac oxidative stress, denoted by lower levels of DCFH and MDA, despite reduced catalase activity and unchanged SOD activity.³¹In another study,²⁸ RT increased antioxidant enzymatic activity (e.g., Mn-SOD and CAT) and reduced that of GSH and increased the activity of tissue remodeling markers (pro- and intermediate MMP-2). The active MMP-2 isoform did not change after the RT intervention.²⁸ The results of the activities of the MMP-2 isoforms after RT intervention were similar to those shown by Leite et al.²⁹

In this sense, RT intervention increased the expression of proteins related to mitochondrial biogenesis, such as Cyto-C, SUD, and PGC-1α, and reduced the expression of endoplasmic reticulum (ER) stress markers p-PERK/PERK, although it did not change the expression of CHOP and GRP78.³³In the study by Melo et al.,³²RT did not modify the collagen content and cross-sectional area of the LV. In addition to not improving the contractile function of cardiomyocytes (e.g., fractional shortening, maximum rate of contraction and relaxation, and time to 50% of contraction and relaxation), RT did not modify the expression of proteins related to Ca² handling, such as Serca2a, PLB, and pPLBser16, and the respective SERCA2A/PLB and pPLBser16/PLB ratios. Most of the studies reported unchanged biometric properties (i.e., heart weight, LV mass, and their ratios to BM and tibial length) following RT intervention.^28,31,32Only Leite et al.,²⁹ showed unchanged heart and LV mass, but increases in heart rate were normalized for BM.