Abstract
is systematic review and network meta-analysis aimed to compare the effects of dietary protein, creatine, and omega-3 fatty acid supplementation on muscle strength, endurance performance, and recovery outcomes in trained athletes. A comprehensive literature search across MEDLINE, Embase, Cochrane CENTRAL, Web of Science, SPORTDiscus, and Sco- pus identified randomized controlled trials evaluating these supplements in individuals engaged in structured training for a minimum of six months. Network meta-analysis employing a frequentist random-effects model synthesized direct and indirect evidence, with treatment rankings determined using Surface Under the Cumulative Ranking curve probabilities. The analysis incorporated 35 trials enrolling 1211 participants. Creatine sup- plementation demonstrated superior effects for muscle strength (SMD = 0.46, 95% CI: 0.29 to 0.63, SUCRA = 82.4%), protein supplementation proved most effective for endurance performance (SMD = 0.28, 95% CI: 0.08 to 0.48, SUCRA = 85.2%), and omega-3 supplemen- tation yielded the greatest benefits for recovery outcomes (SMD = 0.40, 95% CI: 0.18 to 0.62, SUCRA = 88.7%). Network consistency assessment revealed no significant disagree- ment between direct and indirect evidence across all outcomes. These findings reveal an outcome-specific efficacy pattern supporting targeted supplementation strategies aligned with primary training objectives in athletic populations. Keywords:network meta-analysis; dietary supplements; athletic performance; creatine; omega-3 fatty acids 1. Introduction Trained athletes, who are considered to be individuals following organized exercise programs for at least six months, can be considered a specific group with increased nu-
across all outcomes. These findings reveal an outcome-specific efficacy pattern supporting targeted supplementation strategies aligned with primary training objectives in athletic populations. Keywords:network meta-analysis; dietary supplements; athletic performance; creatine; omega-3 fatty acids 1. Introduction Trained athletes, who are considered to be individuals following organized exercise programs for at least six months, can be considered a specific group with increased nu- tritional demands and specialized physiological responses to nutritional supplements. Among several nutritional supplements currently utilized, protein, creatine, and omega-3 fatty acids are recognized as the most well-examined and popular supplements consumed by athletes [1–3]. Each of these supplements targets the fundamental performance criteria that athletes seek to optimize: muscle strength, characterized by maximal force-generating ability; endurance, reflecting the ability to perform protracted physical exertion; and re- covery, involving the replenishment of physiological abilities after physical stress. The specific mechanisms by which each is believed to work indicate a differential effect for these Nutrients2026,18, 909 https://doi.org/10.3390/nu18060909
Nutrients2026,18, 909 2 of 20 domains of interest, although the specific relative efficacy of protein, creatine, and omega-3 fatty acids in trained athletes has not yet been sufficiently well-characterized within the current literature. The physiological basis for each supplement relates to well-established biochemical pathways that impact athletic performance. Protein supplements increase muscle protein synthesis and provide the necessary amino acids for exercising individuals to repair muscle damage, with the International Society of Sports Nutrition recommending 1.4 to 2.0 g per kilogram of body weight per day for individuals on regular training programs [1]. Creatine monohydrate increases muscle phosphocreatine levels, which in turn accelerates the regeneration of adenosine triphosphate during intensive exercise [2]. Omega-3 fatty acids, particularly eicosapentaenoic and docosahexaenoic acids, affect inflammation and membrane fluidity, which in theory could have a positive effect on heart function during endurance exercise or assist with recovery from muscle-injury exercise [3]. While each of these supplements has proven effective within a specific arena, athletes are often faced with dilemmas regarding what supplement to emphasize, given certain limitations regarding simultaneous usage. Although the current body of evidence is large, it is fraught with important limitations that impair its translational value for athletic populations. Conventional pairwise meta- analyses have integrated findings for direct comparisons between specific supplements and placebo or control groups, proving protein supplementation to be efficacious in augmenting lean body mass increases during resistance training [4] and creatine supplementation to be effective in improving strength and power for varied forms of exercise [2,5]. However, these analyses are inherently unable to inform questions of relative efficacy since they evaluate each supplement in parallel rather than comparing it to other options. There are no head-to-head trials evaluating the relative efficacy of protein, creatine, and omega-3 supplementation since the overwhelming majority of study designs are placebo-controlled rather than active-comparator frameworks. This creates a knowledge gap for athletes and individuals to make informed decisions on the use of supplements for specific performance goals such as gaining maximal strength, increasing endurance capacity, and facilitating rapid recovery from intensive exercise. Network meta-analysis, which provides a method of analysis, is able
since the overwhelming majority of study designs are placebo-controlled rather than active-comparator frameworks. This creates a knowledge gap for athletes and individuals to make informed decisions on the use of supplements for specific performance goals such as gaining maximal strength, increasing endurance capacity, and facilitating rapid recovery from intensive exercise. Network meta-analysis, which provides a method of analysis, is able to overcome these challenges by simultaneously comparing a variety of interventions in a combined analysis of direct and indirect evidence [6]. A network of studies is constructed using trials that share a common comparator, allowing for a comparison of interventions that have not been paired in a randomized trial. The method allows for a probabilistic ranking of treatments based on the Surface Under the Cumulative Ranking Curve, which measures the probability that a treatment is the best option for a given outcome. In the field of pharmacology and medicine, network meta-analysis has revolutionized the process of clinical decision-making by providing an effective means to aggregate heterogeneous data into a coherent framework for comparison, though its applicability to sports nutrition is still limited. It is especially useful within the context of dietary supplements, wherein logistical and commercial realities prevent the conduct of multi-arm trials with active comparators. This systematic review and network meta-analysis fills this evidence gap by deter- mining the effects of dietary protein, creatine, and omega-3 supplementation on muscle strength, endurance performance, and recovery outcomes in trained athletes. This review analyzes the results of 35 randomized controlled trials involving 1211 participants to estab- lish a network through which direct comparisons among the three groups and the placebo groups are possible. It conforms to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [7] and the PRISMA extension for network meta-analysis [6] and uses the Grading of Recommendations Assessment, Development https://doi.org/10.3390/nu18060909
Nutrients2026,18, 909 3 of 20 and Evaluation (GRADE) framework [8] to evaluate the evidence certainty. Through the production of quantitative estimates of the likely effects and probabilistic ranking for each supplement on the three outcome measures—muscle strength, endurance performance, and recovery—this study provides the first comprehensive comparison of three widely stud- ied ergogenic supplements among athletic populations. The results provide implications for athletes, coaches, and sports nutrition professionals seeking to apply evidence-based supplementation in line with specific training goals. 2. Materials and Methods 2.1. Search Strategy This systematic review and network meta-analysis was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [7]. The completed PRISMA checklist is provided as Supple- mentary Material. The literature search carried out by the research team utilized different electronic databases from inception through October 2025, without applying any publica- tion date restrictions, to retrieve randomized controlled trials that examined the impact of protein, creatine, and omega-3 fatty acid supplementation on athletic performance. The search was restricted to articles published in the English language. Although no date limits were imposed, all eligible studies identified were published between 2008 and 2025, reflecting the relatively recent emergence of rigorous randomized controlled trial evidence examining these supplements in trained athletic populations. The search included MED- LINE (via PubMed), Embase, Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science Core Collection, SPORTDiscus, and Scopus. These databases were chosen to include diverse fields such as biomedical literature and sports science that could be relevant to our research question. The search strategy combined Medical Subject Headings (MeSH) search terms with free-text terms, which were further divided into four conceptual fields: supplement in- terventions (protein supplementation, whey protein, casein, amino acids, creatine mono- hydrate, omega-3 fatty acids, fish oil, eicosapentaenoic acid, docosahexaenoic acid,n-3 polyunsaturated fatty acids), athletic populations (athletes, trained individuals, resistance training, endurance training, sports performance), performance outcomes (muscle strength, muscular power, endurance performance, recovery, muscle damage, delayed onset muscle soreness), and randomized controlled trial methodology. Boolean operators (AND/OR) were deliberately used to search the terms in and across the domains
mono- hydrate, omega-3 fatty acids, fish oil, eicosapentaenoic acid, docosahexaenoic acid,n-3 polyunsaturated fatty acids), athletic populations (athletes, trained individuals, resistance training, endurance training, sports performance), performance outcomes (muscle strength, muscular power, endurance performance, recovery, muscle damage, delayed onset muscle soreness), and randomized controlled trial methodology. Boolean operators (AND/OR) were deliberately used to search the terms in and across the domains with the necessary truncation for variant spellings. To maximize comprehensiveness, supplementary search strategies were implemented, which included the evaluation of references from the retrieved studies, forward and back- ward citation tracking of included studies through Google Scholar to identify additional relevant publications citing or cited by the retrieved articles, and grey literature citations. The grey literature citations included dissertation databases, conference proceedings from the International Society of Sports Nutrition and the American College of Sports Medicine, and clinical trial registries (ClinicalTrials.gov). Efforts were made to obtain unpublished data or ongoing trials that fulfill the eligibility criteria from the corresponding authors in the included articles. This review was not prospectively registered in the PROSPERO database; however, the protocol was developed prior to data analysis in accordance with a prespecified protocol. 2.2. Eligibility Criteria The eligibility criteria for study enrollment were defined following the PICOS pa- rameters (Population, Intervention, Comparator, Outcomes, Study design). The study https://doi.org/10.3390/nu18060909
Nutrients2026,18, 909 4 of 20 population included trained athletes and physically active individuals who had engaged in organized exercise for at least six months before study enrollment. This particular criterion was set to enable a homogeneous study population and to distinguish it from untrained athletes. The target participants included sports competitors in different sports disciplines as well as persons who regularly took part in resistance or endurance training. Clinical studies conducted among participants with metabolic disorders or among older participants only (mean age greater than 65 years) with no training backgrounds in sports were excluded. Intervention components included the supplementation of dietary protein in any form (whey protein, casein, soy protein, or amino acid supplements providing at least 1.2 g/kg body weight per day for a minimum of four weeks), creatine supplementation (loading protocol of approximately 20 g/day for 5–7 days followed by maintenance dosing, or chronic supplementation of at least 3 g/day for a minimum of four weeks), and omega-3 fatty acid supplementation (fish oil or purified EPA/DHA supplements providing at least 1 g/day combined EPA and DHA for a minimum of two weeks). The minimum supplemen- tation durations were determined based on established pharmacokinetic evidence: Protein supplementation requires approximately four weeks to produce measurable adaptations in muscle protein synthesis and performance outcomes [1,4], creatine supplementation requires at least four weeks for adequate intramuscular phosphocreatine saturation when using chronic low-dose protocols [2], and omega-3 fatty acid supplementation requires a minimum of two weeks for meaningful incorporation into cell membrane phospho- lipids [3]. Multi-ingredient supplements were excluded unless specific nutrient effects could be isolated through study design. Control conditions included placebo supplementa- tion (iso-caloric or non-caloric), no supplementation, or alternative active comparators. Primary outcomes were categorized into three domains: muscle strength (one- repetition maximum, isokinetic peak torque, maximal voluntary contraction), endurance performance (time to exhaustion, time trial performance, maximal oxygen consumption, running economy), and recovery (creatine kinase, muscle soreness ratings, inflammatory markers, functional recovery assessments). Only randomized controlled trials with parallel or crossover designs were included, with crossover studies requiring adequate washout periods (minimum two weeks for protein, four weeks
strength (one- repetition maximum, isokinetic peak torque, maximal voluntary contraction), endurance performance (time to exhaustion, time trial performance, maximal oxygen consumption, running economy), and recovery (creatine kinase, muscle soreness ratings, inflammatory markers, functional recovery assessments). Only randomized controlled trials with parallel or crossover designs were included, with crossover studies requiring adequate washout periods (minimum two weeks for protein, four weeks for creatine, and eight weeks for omega-3). Studies published in English reporting at least one outcome within these domains were eligible. 2.3. Study Selection and Data Extraction Literature screening took place independently by two members of our team using Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia). The two-stage process comprised title and abstract screening, followed by full-text reviews of potentially eligible publications. Any discrepancies were resolved either by consensus or adjudication by a third senior researcher. On calibrating the process of screening fifty randomly selected records, the Cohen’s kappa value of 0.87 revealed excellent inter-rater reliability. The complete screening process was represented in a PRISMA flow diagram along with reasons for exclusion. The extracted data were entered into an electronic form consisting of study-level variables (first author, publication year, country, study design), participant-level variables (sample size, age, sex distribution, training status, sport discipline), intervention-related variables (supplement type, dosage, timing, duration), comparator information, outcome measures and techniques, and any recorded adverse events. For continuous outcomes, means, standard deviations, and sample sizes were abstracted for each treatment group. When standard deviations were not given, they were calculated from standard errors https://doi.org/10.3390/nu18060909
Nutrients2026,18, 909 5 of 20 or confidence intervals using standard formulas. For studies reporting medians with interquartile ranges or ranges, means and standard deviations were estimated using the methods described by Wan et al. [9], implemented via the ‘estmeansd’ package in R software (version 4.3.2). Authors were solicited when key data were missing or unclear, or when data transformation was not feasible. 2.4. Risk of Bias Assessment Methodological quality was assessed using the Cochrane Risk of Bias (RoB 2.0), which considers five domains: the randomization process, the occurrence of deviations from intended interventions, missing outcome data, the measurement of outcomes, and the selection of the results reported. Both appraisers reviewed them independently, with any inconsistencies settled through discussion. Studies were classified as having an overall low risk, some concerns, or high risk of bias based on domain-level ratings. 2.5. Statistical Analysis Data analysis was performed using R software (version 4.3.2). The netmeta and gemtc packages were utilized for analysis. Standardized mean differences (SMDs) were calculated using Hedges’ g with a small-sample bias correction and reported with 95% confidence intervals as effect sizes. Effect sizes were interpreted as negligible (<0.2), small (0.2–0.5), moderate (0.5–0.8), or large (≥0.8). The frequentist random effects model was chosen for the network meta-analysis to account for expected heterogeneity between the studies. The structure of the network was represented in the network plot, in which the size of the nodes represented the number of participants and the width of the edges represented the number of studies. The transitivity assumption was tested by examining the distribution of potential effect modifiers (age, training status, intervention duration). Heterogeneity was assessed using Cochran’s Q statistic (withp< 0.10 indicating significant heterogeneity), I-squared (I 2 ) statistics, and between-study variance (tau-squared,τ 2).I 2 values were interpreted as low (<25%), moderate (25–75%), or high (>75%) heterogeneity in accordance with the Cochrane Handbook [10]. Node splitting and design-by-treatment interaction tests were used to test for inconsistencies in direct and indirect estimates (p< 0.10 indicating significant inconsistency). The treatment outcomes were presented in league tables, while the intervention rank was defined using surface under the cumulative
values were interpreted as low (<25%), moderate (25–75%), or high (>75%) heterogeneity in accordance with the Cochrane Handbook [10]. Node splitting and design-by-treatment interaction tests were used to test for inconsistencies in direct and indirect estimates (p< 0.10 indicating significant inconsistency). The treatment outcomes were presented in league tables, while the intervention rank was defined using surface under the cumulative ranking curve (SUCRA) probabilities, ranging from 0% (worst) to 100% (best). Subgroup analyses were pre-planned for participant characteristics (sex, age, type of sport) and intervention features (dose, duration). However, sex-stratified subgroup analysis was not feasible due to insufficient reporting of sex-disaggregated outcome data in the included studies, and age-based stratification was precluded by the narrow age range of participants (mean 24.6 years, range 18–35 years). Consequently, subgroup analyses were conducted for three moderators: type of sport, intervention duration, and supplement dosage. Sensitivity analyses examined the robustness of the estimates by excluding studies at high risk of bias and by using different statistical models. Small study effects were investigated using comparison-adjusted funnel plots and Egger’s regression test. 2.6. Certainty of Evidence Assessment Evidence certainty was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach [8] adapted to network meta-analyses, evaluating five domains: risk of bias, inconsistency, indirectness, imprecision, and publi- cation bias. Risk of bias was evaluated based on the proportion of information derived from studies at high risk of bias. Inconsistency was assessed through the magnitude of heterogeneity statistics (I 2 andτ 2). Indirectness reflected the degree to which the evidence https://doi.org/10.3390/nu18060909
Nutrients2026,18, 909 6 of 20 directly addressed the comparison of interest. Imprecision was judged by the width of confidence intervals relative to clinically meaningful thresholds. Publication bias was assessed through funnel plot asymmetry and statistical tests. For comparisons informed primarily by indirect evidence, intransitivity was additionally assessed. Evidence certainty was classified as high, moderate, low, or very low, and the results are presented in summary of findings tables. 3. Results 3.1. Literature Search Results A systematic electronic literature search of six databases, namely MEDLINE, Embase, Cochrane CENTRAL, Web of Science, SPORTDiscus, and Scopus, generated 3859 hits that could meet the inclusion criteria. Once 1703 duplicate citations had been removed, 2156 citations underwent title and abstract screening. In this initial screening process, 1869 citations were removed due to reasons such as no relevant content, non-human studies, and citations that did not meet the inclusion criteria, leading to 287 citations that underwent detailed screening. The result of the full-text screening was the removal of 252 articles that did not fulfill the inclusion criteria. The reasons behind exclusion were properly recorded: 87 studies recruited participants who did not fulfill criteria regarding a minimum of six months of structured exercise training to be considered a trained athlete; 62 studies had improper comparison groups, which included active-controlled trials without any placebo groups; 48 studies had incomplete results regarding strength, endurance, and recovery, which were not sufficient to determine effect sizes; 31 articles were duplicate studies of the same study population; and 24 studies utilized non-randomized controlled trial designs. The complete study selection process is illustrated in Figure. Figure 1.PRISMA 2020 flow diagram of the study selection process. * Records identified from databases and registers. ** Records excluded during title/abstract screening. The results from the final network meta-analysis are based on 35 randomized con- trolled trials with a total of 1211 participants who satisfied all the criteria for the analysis. The distribution across intervention categories demonstrated reasonable balance: 14 trials (40.0%) investigated protein supplementation involving 583 participants, 14 trials (40.0%) https://doi.org/10.3390/nu18060909
network meta-analysis are based on 35 randomized con- trolled trials with a total of 1211 participants who satisfied all the criteria for the analysis. The distribution across intervention categories demonstrated reasonable balance: 14 trials (40.0%) investigated protein supplementation involving 583 participants, 14 trials (40.0%) https://doi.org/10.3390/nu18060909
Description
This study analyzes the effects of various supplements on athletic performance.