Abstract
nd: We aimed to (i) identify the factors associated with performance in non-elite runners, (ii) present the terms and definitions/attributes used to characterize runners, and (iii) identify how performance has been operationalized.Methods: Our search was conducted using the databases PubMed, Web of Science, Medline Ovid, Cochrane, PsycInfo, Scielo, Scopus, and SportDiscus in October 2023 and updated in February 2026. Original articles that assessed factors associated with performance in non-elite runners competing in distances ranging from 5 km to ultramarathons were included. The findings were sum- marized by race distance. The Joanna Briggs Institute Analytical Cross-Sectional Studies critical appraisal tool was used for quality assessment.Results: A total of4151 studies were identified, and 66
October 2023 and updated in February 2026. Original articles that assessed factors associated with performance in non-elite runners competing in distances ranging from 5 km to ultramarathons were included. The findings were sum- marized by race distance. The Joanna Briggs Institute Analytical Cross-Sectional Studies critical appraisal tool was used for quality assessment.Results: A total of4151 studies were identified, and 66 studies were included in the final selection. “Recreational” and “athletes” were the most used terms, and finish time was the most common indicator of performance. Performance decline was influenced by arm circumference and mid-axillary skinfold thickness, smoking, body mass index, alcohol consumption, and weather character- istics. Training variables, physiological determinants, and social variables were positively related to performance.Conclusions: The field struggles with a lack of clarity regarding the nomenclature and criteria used to categorize runners. The relevance of a predictor differs according to race distance, with physiological aspects becoming less important at higher distances (i.e., marathon and ultramarathon). Keywords:endurance; performance; exercise; non-elite runners; systematic review 1. Introduction Performance in long-distance running is a multifactorial phenomenon, influenced by a plethora of factors related to the subject and the environment [1]. In general, an- thropometric, training, and physiological variables have been considered to be important J. Funct. Morphol. Kinesiol.2026,11, 124
J. Funct. Morphol. Kinesiol.2026,11, 124 2 of 17 predictors of performance for runners at different competitive levels [1]. For example, a negative association between body fat, body mass, body mass index, and performance has been shown, while training and physiological determinants have shown a positive association [2,3] . Additionally, physiological determinants and training experience have been found to influence marathoners [2,4], with increased maximal oxygen uptake, anaero- bic threshold, improved running economy, weekly training volume, and intensity leading to reduced time [2]. Beyond these domains, psychological variables and social support have been inves- tigated. In ultra-trail runners, performance has been associated with mental toughness, resilience, and obsessive passion [5]. While no significant influence has been observed for social support among Brazilian amateur runners competing in 5 km, half-marathon, and marathon events [6], French track-and-field athletes have been shown to be positively influenced by family members [7]. Additionally, environmental factors such as wind, temperature, and humidity [8], as well as race course characteristics, exert an influence on athletes competing in marathon and ultramarathon events [9]. These results indicate the need for a context-sensitive approach. Despite the multitude of studies aimed at addressing the gap regarding factors as- sociated with performance, the field still grapples with important limitations. A primary limitation pertains to the lack of clarity surrounding the concept of performance [10], in- cluding the meaning of performance for non-elite runners, and how performance has been operationalized. A previous study identified that the social representation of performance for amateur athletes includes terms such as “effort”, “improvement”, and “results”, sug- gesting that performance is perceived as a self-driven objective [11]. Moreover, a wide array of terms has been employed to classify runners. “Amateur”, “recreational”, “well-trained”, “health”, “competitive”, and “novice” are often used interchangeably to describe runners, without clear conceptual distinctions or criteria for differentiation [12]. The heterogeneity of concepts, the diverse training backgrounds of various types of runners, the diverse motivations for engagement in training and competition, and differences in anthropometric characteristics limit guidance, synthesis of the literature, and professional intervention, especially for those focused on performance enhancement [12,13]. Similarly, inconsistency in
interchangeably to describe runners, without clear conceptual distinctions or criteria for differentiation [12]. The heterogeneity of concepts, the diverse training backgrounds of various types of runners, the diverse motivations for engagement in training and competition, and differences in anthropometric characteristics limit guidance, synthesis of the literature, and professional intervention, especially for those focused on performance enhancement [12,13]. Similarly, inconsistency in the application of diverse metrics to evaluate performance in running studies hampers a comprehensive understanding of the subject, which could otherwise offer insights into factors that may be modifiable to enhance outcomes such as health and injury prevention, in addition to performance alone. Gaining clarity on these gaps is the first step to designing interventions to improve running performance while respecting the characteristics and needs of runners. Given how widespread running is among the general population [14], and given the increasing popularity of ultramarathons in recent decades [15], we aim to identify the factors associated with performance in non-elite runners competing in distances ranging from 5 km to ultramarathons. We hope to advance a previous narrative review that identified predictive models for runners of long-distance races (i.e., 5 km to marathon) at different competitive levels [1] by analyzing the terms, concepts, and attributes used to characterize runners and identifying how performance has been operationalized in running studies. 2. Materials and Methods 2.1. Protocol and Registration This systematic review was conducted according to the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [16] (Supplementary Material File S1) and the PERSiST guidance [17]. Before initiating the review, the protocol https://doi.org/10.3390/jfmk11010124
J. Funct. Morphol. Kinesiol.2026,11, 124 3 of 17 for the analysis was registered on the PROSPERO International prospective register for systematic reviews website (https://www.crd.york.ac.uk/PROSPERO/ October 2023)) (registration number: CRD42023471483). 2.2. Databases, Search Strategy, and Study Eligibility Criteria A search strategy was developed by the lead author (M.T.) in collaboration with co- authors of this review (T.N.G. and M.S.). The following key databases, relevant to our research question, were searched: PubMed, Web of Science, Medline Ovid, Cochrane, PsycINFO, Scielo, Scopus, and SportDiscus. The search was conducted and completed in October 2023 and updated in February 2026. Table search and the eligibility criteria adopted. We defined non-elite runners as individuals who participate in running without sponsorship and who do not compete in world cham- pionships, the Olympic Games, or continental championships. Studies including mixed samples (i.e., both elite and non-elite athletes), as well as those allowing participants to self-classify as elite, were excluded. Articles were excluded if the full text was unavailable or written in a language other than English, Spanish, or Portuguese. Table 1.Search terms and eligibility criteria following the PICOS strategy.Search Terms Inclusion Criteria Exclusion Criteria P Runners OR “long-distance runners” OR “amateur runners” OR “recreational runners” OR joggers OR “non-professional” NOT “elite-athletes” Women and men; non-elite runners competing in race distances ranging from 5 km to ultramarathons; runners competing in mountain, trail, or sky runs Animals and in vitro studies; children or adolescents; elite athletes; triathletes or biathletes; I “Factors associated” OR correlates OR predict* OR determinants Studies reporting predictors, correlations, or associations Do not focus on running performance or performance predictors C Not applied Not applied Not applied O Performance OR “running pace” OR velocity OR “finish time” Performance (finish time, velocity, running pace) is the main outcome Walking; performance in sprint; running performance in team sports; assessing performance in lab tests S All designs Original study articles published in peer-reviewed journals, adopting both quantitative and/or qualitative approaches Conference abstracts, commentaries, book chapters, or editorials The asterisk (*) was used as a wildcard to truncate word stems in the search strategy, allowing the retrieval of different variations of
Walking; performance in sprint; running performance in team sports; assessing performance in lab tests S All designs Original study articles published in peer-reviewed journals, adopting both quantitative and/or qualitative approaches Conference abstracts, commentaries, book chapters, or editorials The asterisk (*) was used as a wildcard to truncate word stems in the search strategy, allowing the retrieval of different variations of the terms sharing the same root. 2.3. Study Selection The identification and screening steps were carried out by two authors (M.T. and M.S.). All records retrieved by the search query were imported into the Rayyan tool (https://www.rayyan.ai/ duplicates were removed by the lead author (M.T.). Subsequently, the title and abstract were independently screened by the two authors (M.T. and M.S.). Any disagreements were resolved through discussion. Next, the full texts of potentially eligible studies were evaluated independently, and any disagreements were resolved through discussion. https://doi.org/10.3390/jfmk11010124
J. Funct. Morphol. Kinesiol.2026,11, 124 4 of 17 2.4. Data Extraction and Research Reporting Quality Data extraction was performed by the lead author (M.T.), with a quality control cross- check performed by both M.T. and M.F. One week after data extraction, both authors reviewed the extraction of half of the studies. Data extraction included: (1) the first author and year of publication; (2) sample characteristics; (3) term used to classify runners; (4) definition of the term/attributes used to characterize the sample (when reported); (5) performance indicator (e.g., mean speed, finish time, and running pace); (6) variables studied; and (7) main outcomes. The findings were summarized in a general way and by race distance (i.e., 5 km, 10 km, half-marathon, marathon, and ultramarathon). The Joanna Briggs Institute Analytical Cross-Sectional Studies critical appraisal tool [18] was used for quality assessment. Two reviewers (MT and TNG) conducted individual quality assessments, resolving discrepancies through consensus. The quality appraisal checklist comprised eight criteria: whether (1) the sample inclusions were de- fined; (2) the study subjects and settings were described; (3) exposure was measured validly and reliably; (4) objective, standard measurement criteria for the conditions were used; (5) confounding factors were identified; (6) strategies to deal with confounding factors were stated; (7) the outcomes were measured validly and reliably; and (8) appropriate statistical analysis was used. Each criterion was scored as being “met = yes” or “not met = no” or “unclear”, or, in some instances, as “not applicable”. To verify the quality of the study, we summed up the “yes” responses. Results were used to assign an a priori quality rat- ing to each study (0–2 points = very low; 3–4 points = low; 5–6 points = moderate; and 7–8 points = high). 3. Results The search retrieved 4151 studies. After excluding duplicates, 3676 studies remained. A total of 3516 studies were excluded by title/abstract, and 160 were assessed, of which 94 were excluded. Reasons for exclusion included not assessing performance prediction, labo- ratory measures, focusing on uphill/downhill running, using the wrong design/distance or sample, and not reporting beta or r values. Upon reading the
The search retrieved 4151 studies. After excluding duplicates, 3676 studies remained. A total of 3516 studies were excluded by title/abstract, and 160 were assessed, of which 94 were excluded. Reasons for exclusion included not assessing performance prediction, labo- ratory measures, focusing on uphill/downhill running, using the wrong design/distance or sample, and not reporting beta or r values. Upon reading the entire text, 94 studies were excluded, leaving 66 that met the eligibility criteria. Figure screening process. 3.1. Research Reporting Quality Table vast majority of studies (92.1%) were classified as moderate quality, four studies (5.3%) achieved a high-quality rating, and two studies (2.6%) were classified as low quality; no study was rated as very low quality. For question Q1 (“Were the criteria for inclusion in the sample clearly defined?”), approximately 62.1% of the studies were rated as “no”, indicating that clear inclusion criteria were frequently not reported. In contrast, question Q2, which evaluates whether study participants and settings are described in detail, was positively rated for 98.5% of the studies, as most reported participants’ age, sex, training experience, and anthropometric characteristics, although contextual and time frame details were sometimes limited. Regarding questions Q3 and Q4, most of the studies were rated positively, suggesting that exposures and outcomes were generally measured in a valid and reliable manner. For confounding factors (Q5 and Q6), important methodological limitations were observed, with few studies clearly identifying confounding factors (Q5) or reporting strategies to deal with them (Q6), although some applied statistical approaches such as stratification by sex, age, or competitive level. All studies were positively rated for Q7, indicating appropriate validity and reliability of outcome measurements. Similarly, Q8 was rated positively in 100% of the studies, reflecting the use of appropriate statistical https://doi.org/10.3390/jfmk11010124
J. Funct. Morphol. Kinesiol.2026,11, 124 5 of 17 analyses aligned with the study objectives, although some studies could have benefited from more detailed reporting. Figure 1.PRISMA flow diagram of the screening process. Table 2.Quality assessment for included studies. Author, Year Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Total Ueno, et al. (2021) [19] 5 Balducci, et al. (2017) [20] 5 Nikolaidis & Knechtle (2020) [21] 7 Knechtle et al. (2008) [22] 5 Knechtle et al. (2011) [23] 5 Knechtle et al. (2011) [24] 5 Knechtle et al. (2009) [25] 5 Knechtle et al. (2010) [26] 5 https://doi.org/10.3390/jfmk11010124
J. Funct. Morphol. Kinesiol.2026,11, 124 6 of 17 Table 2.Cont. Author, Year Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Total Dellagrana et al. (2015) [27] 5 Gómez-Molina et al. (2017) [4] 6 Thuany et al. (2023) [6] 6 Thuany et al. (2021) [28] 6 Coates et al. (2021) [29] 5 Nikolaidis et al. (2023) [30] 7 Thuany et al. (2021) [31] 6 Clemente-Suarez & Nikolaidis (2017) [32] 5 Daniela et al. (2012) [33] 5 Knechtle et al. (2010) [34] 5 Coquart (2023) [35] 6 Knechtle et al. (2011) [36] 5 Knechtle et al. (2010) [37] 5 Paavolainen et al. (1999) [38] 6 Sinnett et al. (2001) [39] 6 Del Rosso et al. (2021) [40] 5 Martínez-Navarro et al. (2021) [41] 6 Alvero-Cruz et al. (2019) [42] 5 Takeshima & Tanaka (1995) [43] 5 Kilding et al. (2006) [44] 5 Florence & Weir (1997) [45] 5 Adams et al. (2017) [46] 6 Ueno et al. (2019) [47] 5 Scheer et al. (2019) [48] 6 Nummela et al. (2006) [49] 6 Alvero-Cruz et al. (2019) [50] 4 Wiswell et al. (2000) [51] 5 Rossuello et al. (2009) [52] 6 https://doi.org/10.3390/jfmk11010124
J. Funct. Morphol. Kinesiol.2026,11, 124 7 of 17 Table 2.Cont. Author, Year Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Total Fornasiero et al. (2018) [53] 5 Radosavljevi´c et al. (2016) [54] 5 Masters & Ogles (1998) [55] 5 Rubaltelli (2018) [56] 5 Ueno et al. (2018) [57] 5 Knechtle et al. (2014) [58] 6 Ueno et al. (2018) [59] 6 Knechtle et al. (2021) [60] 4 Ramsbottom et al. (1989) [61] 5 Schmid et al. (2012) [62] 6 Marti et al. (1988) [63] 5 Machado et al. (2013) [64] 5 Siqueira et al. (2022) [65] 6 Kubo et al. (2015) [66] 5 Matos et al. (2020) [67] 6 Manzi et al. (2009) [68] 6 Lerebourg et al. (2023) [69] 6 Landman et al. (2012) [70] 6 Bertuzzi et al. (2014) [71] 5 Forsyth et al. (2017) [72] 6 Valentino et al. (2016) [73] 6 Alves et al. (2024) [74] 6 Inamura et al. (2024) [75] 6 Knechtle et al. (2024) [76] 6 de Anda Martín et al. (2024) [77] 6 Gutiérrez et al. (2025) [78] 6 Knechtle et al. (2025) [79] 6 Knechtle et al. (2025) [9] 6 https://doi.org/10.3390/jfmk11010124
J. Funct. Morphol. Kinesiol.2026,11, 124 8 of 17 Table 2.Cont. Author, Year Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Total Turnwald et al. (2025) [80] 6 DeJong Lempke et al. (2026) [81] 8 Legend: Q1 (Were the criteria for inclusion in the sample clearly defined?); Q2 (Were the study subjects and the setting described in detail?); Q3 (Was the exposure measured in a valid and reliable way?); Q4 (Were objective, standard criteria used for measurement of the condition?); Q5 (Were confounding factors identified?); Q6 (Were strategies to deal with confounding factors stated?); Q7 (Were the outcomes measured in a valid and reliable way?); Q8 (Was appropriate statistical analysis used?). Yes ( ); No ( ); Unclear ( ); Not applied ( ). The total is based on the sum of the “yes” responses. The results of this assessment were used to assign an a priori quality rating to each study (0–2 points = very low; 3–4 points = low; 5–6 points = moderate; 7–8 points = high). 3.2. Sample Characteristics and Terms Used A total of 2,347,977 runners were sampled from 1988 [63] to 2026 [81]. Of the total, 1,525,555 were men (~65%), and 822,422 were women (~35%). “Recreational” was the term most used (12 studies), followed by “athletes” (11 studies), “well-trained/moderately trained runners” (ten studies), and “runners” (five studies). Less frequent terms were “competitors/competitive”, “endurance runners/endurance trained”, “master”, “trail/ultra-endurance or ultrarunners”, “road runners”, “healthy runners”, “amateur”, “non-professional”, “joggers”, “finishers”, “starting runners”, and “subjects”, and three other studies used two different terms (“recreational” and “nonprofessional athletes”, “recreational healthy runners”, and “endurance-trained”). Of these studies, only three presented an explanation for the terms used. For example, the term master athletes was used and defined in two studies: Wiswell et al. [51] considered "master athletes" to be runners aged≥40 years who had at least 5 years of training and a running volume of≥15 km/week, and who participated at least once per year in organized running competitions. Similarly, Forsyth et al. [72] considered veteran runners to be those aged 40 years or older who had at least five years of training, reached
Description
This systematic review analyzes factors influencing performance in non-elite runners across various race distances.