This perception is supported by epidemiological data in endurance athletes and limited data from intervention studies using moderate exercise in previously untrained individuals.

The gap

Consensus across the literature

Supporting evidence — 4 representative gaps

• A systematic review and meta-analysis of the effects of exercise training on dysfunction in acute, subacute, and chronic stroke patients (2026) · doi

  This study had several limitations. First, only 16 RCTs and 22 observational study were included, leading to inevitable bias. Second, there are few studies that compare the effects of exercise training across all three stroke phases—acute, subacute, and chronic. Many studies only examine two of these phases. Future meta-analyses should aim for larger sample sizes, better study designs, and multi- center RCTs. Third, although 38 studies were included in this system- atic review, the number of studies available for each meta-analysis was substantially smaller. Specifically, the analyses for MAS, ARAT, and BBS were each based on a limited number of studies, which may have resulted in insufficient statistical power and compromised the robust- ness of the findings. Therefore, these results should be interpreted with caution. Future large-scale studies are needed to validate these pre- liminary findings. Fourth, although subgroup analyses were con- ducted for exercise type, intervention dose, and stroke type, other important variables—such as stroke severity and type of control group intervention—could not be examined due to insufficient data in the included studies. Finally, pooled results should be interpreted cau- tiously due to variations in patient profiles, case numbers, and ethnic backgrounds across studies.

  Keywords: included stroke analyses type rcts exercise across phases future meta number insufficient interpreted intervention several
• Comparative efficacy of different training modes on physical fitness in soccer (football) players: a systematic review and network meta-analysis incorporating randomized controlled trials (2026) · doi

  This study represents the first application of NMA to systematically evaluate the effects of HIIT, SSG, LICT, and HSG on physical fitness indicators in soccer play- ers, encompassing aerobic capacity, anaerobic capacity, and SSE. The NMA methodology enabled not only direct comparisons between different training modalities and control conditions but also quantified relative effects through indirect comparisons. By incorporating SUCRA values to rank optimal training strategies, this approach overcomes the limitation of conventional meta-anal- yses that are restricted to direct comparisons, thereby enhancing the precision and practical relevance of the conclusions. Nevertheless, despite adhering strictly to the PRISMA guidelines, this study inevitably has several limitations. First, despite comprehensive literature retrieval, the number of available primary studies remained limited for certain modalities (e.g., LICT and HSG). The result- ing small sample sizes contribute to statistical impreci- sion, warranting caution in interpreting effect estimates. Furthermore, due to the still limited sample size of the research, some intervention methods cannot be further subdivided (including HIIT, which cannot be differenti- ated in terms of intensity and form). Second, variations in training intervention details (e.g., intensity, frequency) introduced structural heterogeneity within nodes. While pooling was necessary to maintain network connectivity, it implies that findings represent ‘average’ effects. Third, although we conducted sensitivity analyses, the inclusion of studies with high ROB introduces potential quality concerns. Finally, some network links relied on indirect evidence, which inherently carries lower certainty than direct head-to-head comparisons. Li et al. BMC Sports Science, Medicine and Rehabilitation (2026) 18:227 Declarations Ethics approval and consent to participate Not applicable. This study is a systematic review and network meta-analysis based on previously published literature and does not involve the recruitment of human participants or animal subjects. Therefore, the study was conducted in accordance with the ethical standards of the Declaration of Helsinki. As no original or individual-level data were collected, formal approval from an Institutional Review Board (IRB) or an Ethics Committee was not required for this research. The systematic review protocol was initially submitted and published on the Open Science Framework on November 21, 2025, under the registration number DOI h t t p s : / / d o i . o r g / 1 0 . 1 7 6 0 5 / O S F . I O / D 5 B N A.

  Keywords: comparisons effects direct training network review first hiit lict capacity modalities indirect meta despite literature
• Validity and agreement of the 30–15 Intermittent Fitness Test for estimating aerobic and performance-related parameters: a systematic review and meta-analysis (2026) · doi

  While the present meta-analysis supports the 30–15IFT as a valid and practical field-based tool for estimating VO₂max, HRmax, and MRV in athletes, several meth- odological limitations of the included studies warrant attention. According to QUADAS-2 and QUADAS-C assessments, only Paravlic et al. (2022) demonstrated low risk of bias in participant selection, whereas most other studies showed unclear or moderate risk in key domains such as index test and reference standard application. PEDro scores ranged from 3 to 5, reflecting moderate methodological quality, with common limitations includ- ing lack of blinding and incomplete reporting of alloca- tion procedures. Future research should prioritize improving study qual- ity by ensuring rigorous participant selection, blinding of assessors, and standardized administration of both index and reference tests. Larger and more diverse samples, combined with direct physiological measurements (e.g., gas analyzer VO₂max) alongside predictive equations Ghazzagh et al. BMC Sports Science, Medicine and Rehabilitation (2026) 18:228 Page 18 of 20 tailored to intermittent running protocols, could reduce heterogeneity and improve the precision of findings. The findings of the present meta-analysis provide a unique opportunity to refine VO₂max prediction models based on the 30–15IFT. Specifically, the observed pooled bias (1.88 ml·kg⁻¹·min⁻¹) and 95% limits of agreement (− 5.28 to + 9.04 ml·kg⁻¹·min⁻¹) can inform the develop- ment of more precise, population- and protocol-specific predictive equations. For example, regression-based models can incorporate the pooled bias as a group-level correction term, such that predicted VO₂max values are adjusted upward by the mean bias to account for systematic underestimation. Additionally, the limits of agreement can be used to quantify individual-level vari- ability, either by providing prediction intervals for single measurements or by incorporating this variance as a ran- dom effect in mixed-effects models. This approach allows future predictive equations to systematically correct for both average bias and individual variability, thereby enhancing the accuracy and applicability of VO₂max esti- mations derived from the 30–15IFT. In practice, incorporating bias as a fixed effect and the limits of agreement as a random effect or as uncertainty intervals would allow researchers and practitioners to generate VO₂max estimates that are both more accu- rate and individualized. For instance, consider a semi- professional soccer player performing the 30–15IFT. Using a standard predictive approach, the test might esti- mate their VO₂max at 52 ml·kg⁻¹·min⁻¹. Our meta-anal- ysis indicates that the 30–15IFT tends to overestimate VO₂max by 1.88 ml·kg⁻¹·min⁻¹ compared with treadmill- based ITRT. By incorporating this pooled bias as a fixed effect in a regression or mixed-effects model, the esti- mate can be adjusted downward to 50.12 ml·kg⁻¹·min⁻¹. Additionally, applying the 95% limits of agreement (− 5.28 to + 9.04 ml·kg⁻¹·min⁻¹) as a random effect or uncertainty interval allows for individualized prediction ranges—for example, 44.84 to 59.16 ml·kg⁻¹·min⁻¹ for this player. Incorporating further athlete-specific covariates, such as intermittent running experience or fitness level, would refine predictions even more, enhancing the prac- tical utility of the 30–15IFT for individualized training prescription. Moreover, combining these adjustments with athlete- specific factors—such as fitness level, sport-specific adaptations, and intermittent running proficiency—could further enhance predictive accuracy. Such refined model- ing approaches would strengthen the ecological validity of the 30–15IFT and improve its utility for athlete moni- toring and individualized training prescription.

  Keywords: bias predictive effect based limits agreement specific level incorporating individualized meta equations intermittent running prediction
• Chronic exercise training effects on immune function (2000) · doi

  This perception is supported by epidemiological data in endurance athletes and limited data from intervention studies using moderate exercise in previously untrained individuals.

  Keywords: perception supported epidemiological endurance athletes limited intervention using moderate exercise previously untrained individuals