Dry-land training in developmental competitive swimmers: a systematic review and meta-analysis of comparative effects across three modalities, dose-response patterns, and associations between physical fitness and swimming-specific performance

Abstract

Dry-land training is an important component of competitive swimming and is used to develop the physical fitness (PF) required for sport-specific training. However, the relative effects of different dry-land training modalities, their dose-response patterns, and associations between PF and swimming-specific performance (SSP) remain unclear. Eight databases were searched from inception to 15 February 2026, with additional records identified through websites and citation searching. Controlled intervention studies of three dry-land training modalities (core, plyometric, or resistance training) in youth, adolescent, age-group, or comparable competitive swimmers who trained with the intention to compete and remained within a sport-specific developmental pathway were included. Methodological quality was assessed using the PEDro scale. Hedges’ g was calculated from pre–post changes. Random-effects meta-analyses, sensitivity analyses, continuous dose-response meta-regressions, and comparison-level analyses of associations between PF and SSP effects were performed. Thirty-seven studies involving 829 swimmers were included: 14 core training, 8 plyometric training, and 15 resistance training studies. All three modalities showed positive conventional pooled effects on SSP (g = 0.204, 95% CI 0.082 to 0.326; g = 0.376, 95% CI 0.163 to 0.588; and g = 0.291, 95% CI 0.202 to 0.380, respectively) and PF (g = 0.541, 95% CI 0.325 to 0.756; g = 0.247, 95% CI 0.066 to 0.428; and g = 0.356, 95% CI 0.242 to 0.470, respectively). Heterogeneity was absent to moderate across the six models (I² = 0.0%–47.7%). Statistical support was stronger for CT–SSP, RT–SSP, and RT–PF, whereas CT–PF and both PT outcome domains were more inference-sensitive. Exploratory comparison-level dose-response models indicated modality- and outcome-specific patterns within the observed dose ranges. The comparison-level PF–SSP effect association was unstable for CT (β = 0.021, p = 0.905), positive but conditional for PT (β = 0.579, p = 0.017), and showed the strongest statistical support for RT (β = 1.180, p < 0.001). The three included dry-land training modalities showed positive conventional pooled effects on PF and SSP in developmental competitive swimmers, although the strength and stability of support differed across combinations of training modality and outcome domain. Dose-response findings represented exploratory comparison-level patterns within the observed ranges and should not be treated as fixed prescriptions or definitive thresholds. Training planning may be informed by training modality, target outcome, observed-range dose-response characteristics, and comparison-level associations between PF and SSP effects. PROSPERO CRD420261367300.