Short-term creatine use improves strength, lowers fatigue, and speeds recovery in trained athletes

A priori power analysis using G*Power (v. 3.1.5.1) determined that a minimum of nine participants was necessary. The calculation targeted our primary endpoint for total load lifted (kg) in back squat (BS) across sets, using a repeated-measures ANOVA (condition × intensity × time). We assumed a large effect size (Cohen’s f = 0.50; ≈ η_p² ≈ 0.20 based on prior creatine performance data [37]), α = 0.05, power (1–β) = 0.95, and a within-subject correlation of 0.50, which indicated n = 9.

A total of eleven recreationally resistance-trained males were recruited; however, due to incomplete data, one participant was excluded, and a total of ten participants were analyzed (age: 21.3 ± 1.9 y [range: 19–24]; BMI: 21.42 ± 2.36 kg·m⁻² [18.6–25.1]; 1RM BS: 97 ± 14.18 kg [80–120]; 1RM bench press (BP): 56.36 ± 4.52 kg [50–60]). Training experience was ~3 months on average (median 12 weeks, range 6–24 weeks). All participants had engaged in structured resistance training (≥3 times per week for ≥6 weeks) before the study and were familiar with BP and BS exercises. All were CR-supplementation naïve at baseline. Exclusion criteria included the use of medications/supplements affecting muscle biology (e.g. corticosteroids, CrM, anabolic/nutritional substances) during/one month pre-study; pre-existing kidney/liver conditions, low blood pressure, or injuries hindering exercise (confirmed via interviews). Non-compliant participants or those failing to complete all study stages were excluded from the analyses. Participants maintained regular diets and avoided strenuous activity/NSAIDs to limit muscle turnover confounders [38]. Participants abstained from strenuous exercise for 48 h before each testing visit and maintained usual activity otherwise. Before participation, individuals were informed of the study's risks, benefits, and objectives, and provided written consent. The study followed the Declaration of Helsinki and received ethical approval from the local Research Ethics Committee of High Institute of Sport and Physical Education of El Kef, El Kef, Tunisia on 12 October 2024 (ISSEPK-0033/2024). The study was registered at the Pan African Clinical Trials Registry database on 15 May 2025 (PACTR202505827886996). To eliminate confounding variables from hormonal fluctuations, this study's sample was limited to males. Female sex hormones can impact HRV, strength, CrM response, and exercise-induced muscle damage indicators [39]. Additionally, the logistical constraints of this pilot study precluded the inclusion of additional sessions necessary to control for the menstrual cycle.

The study followed a double-blind, crossover, randomised controlled design. The experimental procedure included one familiarisation session and four testing sessions, each separated by seven days. This interval was selected pragmatically to minimise acute carryover and allow sufficient recovery between sessions while maintaining ecological validity within a short-term crossover design. Although full normalisation of intramuscular CR typically requires four to six weeks following cessation [4,24], several recent crossover studies in applied sport contexts have adopted similar seven-day washouts [40,41] as an acceptable compromise to preserve training continuity and participant adherence [10]. The seven-day spacing also supported optimal participant recovery [42]. All experimental tests were conducted in the afternoon to reduce diurnal variability and to standardise testing time for neuromuscular performance and HRV across. Afternoon testing was chosen based on evidence indicating superior strength performance, lower perceived exertion, and more stable physiological responses during these hours [43,44]. Testing sessions were carried out at a controlled temperature of 24 °C (±1 °C) and at a consistent time of day (±0.5 hours) for each participant to minimise the impact of circadian rhythm variations [43,45].

During the first visit, participants’ anthropometric measurements were recorded using bioimpedance analysis (Tanita MC-780MA; Tanita Corporation, Japan). Additionally, they participated in a familiarisation session for the lifting protocol, emphasising maximum explosive velocity during both the BS and BP. A standardised warm-up was performed before determining their 1RM, following the ASEP guidelines for accurately assessing muscular strength and power [46]. The 1RM testing protocol was applied to both BS and BP within the same session. Participants first performed five repetitions at 50% of their estimated 1RM, followed by three repetitions at 70%, with 3-minute rest intervals between sets. After an additional 3-minute rest, participants attempted to reach their 1RM within five trials, each separated by a 3-minute rest period. Movement velocity was continuously tracked to confirm that the 1RM load corresponded to the expected speed for a true 1RM in each exercise [47,48]. The final 1RM values were then used to determine intensity percentages for the subsequent testing sessions. After familiarisation, participants returned to the laboratory four more times for each supplementation protocol (CrM and placebo (PLA)), performing an incremental resistance training test to assess the acute ergogenic effects of supplementation. In total, each participant completed four test sessions over a two-week period (Figure 1).

In this study, CrM (GymBeam GmbH, Berlin, Germany) was mixed with placebo juice, ensuring identical colour, texture, and appearance for blinding. Two independent individuals handled randomisation and kit preparation, which included the assigned supplement, instructions, measuring spoons, and a water bottle. Researchers remained blinded until data collection ended. Blinding efficacy was assessed at the final visit, before debriefing. Participants guessed the supplement received in each phase and rated their confidence.

CrM was administered at 0.3 g·kg⁻¹·d⁻¹ for three days, a well-tolerated dosage for rapidly increasing muscle CR stores. On day one, the full dose was taken two hours before testing; on days two and three, it was divided into three sub-doses (0.1 g·kg⁻¹·d⁻¹). Supplements were provided in a shaker bottle with volume gradations. Participants refrained from food or drink (except water) two hours before testing and avoided other supplements on non-testing days. Adherence was tracked via a compliance log. Participants were instructed to: (i) avoid CR-rich foods, stimulants, gum, sweets, and alcohol for three days before testing; (ii) stay adequately hydrated; (iii) avoid strenuous exercise during the study; and (iv) sleep at least eight hours per night.

During each visit, participants completed an 8-minute supervised warm-up on a treadmill before performing BS and then BP on a Smith machine, with a 3-minute rest between exercises. Prior to the incremental strength test, they warmed up with 12 reps at 10% of 1RM and 6 reps at 30% of 1RM (with 1-minute rests) [49], followed by a 2-minute rest. The test consisted of three sets at 60%, 70%, and 80% of 1RM, performing repetitions to failure with 3-minute rests between sets [50]. Both the BS and BP were executed with a full range of motion—knee flexion and extension for the BS and elbow flexion and extension for the BP [51]. Participants were instructed to perform the concentric phase of each repetition at maximum velocity to optimise muscle strength development [48]. The entire testing session, including the warm-up, lasted approximately 35 minutes.

Statistical analyses were conducted using the R programming language [60]. Descriptive statistics were presented as Mean ± standard deviation (SD). The visualization was conducted with “ggplot2” package [61]. We tested whether the proportion of correct treatment guesses differed from chance (50%) using a chi-square goodness-of-fit test with Yates’ continuity correction (df = 1). The normality of the data was checked using the Shapiro-Wilk test with the “rstatix” package [62]. To assess the statistical effects of supplementation condition, time, and intensity, a three-way repeated measures analysis of variance (ANOVA) with Greenhouse-Geisser correction was performed. When significant main or interaction effects were found, a post-hoc pairwise comparison with Bonferroni adjustment, was performed. The delta change (∆pre-post) was calculated as follows: ∆pre–post (%) = ((score at post-session − score at pre-session)/score at pre-session) × 100. To assess the difference between supplementation condition and time effects of delta changes for HRV indices, a two-way ANOVA was performed, followed by a post-hoc pairwise comparison with the Bonferroni adjustment. All ANOVA models were conducted with the “afex” package [63] and the pairwise comparisons with “emmeans” package [64]. The effect size statistic was calculated as partial eta-squared (η²p) to assess the magnitude of the effects as small (0.01), moderate (0.06), and large (0.14) [65]. Standardised effect size (Cohen’s d) analysis was used to interpret the magnitude of differences between means and classified according to Hopkins [66] as: trivial (d ≤ 0.20), small (0.20 < d ≤ 0.60), moderate (0.60 < d ≤ 1.20), large (1.20 < d ≤ 2.0), very large (2.0 < d ≤ 4.0), and extremely large (d > 4.0). Significance was accepted for all analyses at the a priori level of p < 0.05.

No significant differences in energy or macronutrient intake (Table 1) were found between conditions (p > 0.05).

Blinding assessment indicated guess accuracy near chance. Overall, 55% of participants correctly guessed whether they had received CrM or PLA, which did not differ from chance (χ²(1) = 0.20, p = 0.65), supporting effective blinding.

During the BP, there were significant main effects of condition and intensity (Figure 2, Table 2). Significant differences in reached repetition between all intensities across conditions and sessions were observed (p < 0.01, d = 1.52−4.2). Repetitions were significantly higher in the CrM compared to PLA condition at 60% (p = 0.0034, d = 1.25) and 70% 1RM (p = 0.041, d = 0.75) during the 1^st session, and at 70% (p = 0.008, d = 1.6) and 80% 1RM (p = 0.002, d = 1.37) during the 2^nd session.

Concerning repetitions during the BS (Figure 2), there were significant main effects of condition and intensity, as well as a significant condition × intensity interaction (Table 2). Moreover, there were significant differences between all intensities within both conditions at all session (p < 0.05, d = 1.1−3.12), except between 70 and 80% 1RM for the CrM condition during the 2^nd session (p = 0.159). Additionally, the CrM condition led to significantly greater repetitions compared to PLA for 60% and 80% 1 RM at the 1^st (p = 0.036 and 0.016, d = 0.79 and 0.9, respectively) and only for 60% at the 2^nd session (p = 0.013, d = 0.97).

The analysis revealed significant main effects of condition, session, and intensity (Figure 2, Table 2). Post-hoc comparisons revealed significant differences between all intensities across conditions and sessions (p < 0.05, d = 1.15−2.15), except for 70% vs 80% for CrMduring the 1^st session (p = 0.2228). Velocity during both sessions was significantly higher for the CrM condition as compared to PLA at 60% (1^st: p = 0.007, d = 1.11; 2^nd: p = 0.034, d = 0.78), 70% (1^st: p = 0.024, d = 0.86; 2^nd: p < 0.001, d = 2.95), and 80% 1RM (1^st session: p < 0.001, d = 2.91; 2^nd: p < 0.001, d = 3.98).

During the BS (Figure 2), there were significant main effects of condition, session, and intensity, as well as a significant session × intensity interaction (Table 2). Post-hoc comparisons revealed a significant decrease for the CrM condition from 60% to 70% (p = 0.002, d = 1.6) during the 1^st session and from 60% to 80% 1RM during both sessions (p < 0.001 and 0.046, d = 3.61 and 0.95, respectively for the 1^st and 2^nd sessions). Additionally, the PLA condition’s velocity decreased significantly from 60% to 80% 1RM and from 70% to 80% 1RM during the 1^st session (p < 0.001 and 0.011, d = 2.25 and 1.27, respectively), and from 60% to 80% 1RM during the 2^nd session (p = 0.001, d = 1.79). In the PLA condition, velocity during the 1^st session was significantly higher than the 2^nd session at 60% (p = 0.033, d = 0.75) and 70% 1RM (p = 0.015, d = 0.66), In contrast, the CrM condition showed a significantly higher velocity during the 1^st session compared to 2^nd session at 60% (p < 0.001, d = 1.42), 70% (p = 0.006, d = 1.09), and 80% 1RM (p = 0.002, d = 1.46). During both sessions, the CrM condition was significantly greater than PLA at 60% (1^st: p = 0.001, d = 2.83; 2^nd: p = 0.035, d = 1.57), 70% (1^st: p = 0.002, d = 2.66; 2^nd: p = 0.018, d = 1.94), and 80% 1RM (1^st: p < 0.001, d = 3.84; 2^nd: p < 0.001, d = 4.09).

The analysis revealed significant main effects of condition and intensity (Figure 2, Table 2). The power significantly increased across all intensities for both conditions during both sessions (p < 0.05, d = 1.16−4.71), except from 60% to 70% 1RM for the PLA condition during the 1^st session (p = 0.094).

Regarding power during the BS (Figure 2), there was only a significant main effect of intensity (Table 2), where power significantly increased across all intensities for both conditions during both sessions (p < 0.005, d = 1.98−4.97).

Analysis of peak HR during the BP revealed a significant main effect of intensity (Figure 2, Table 2). Peak HR showed significant increases across all intensities for both conditions during both sessions (p < 0.05, d = 1.1−4.28), but not from 60% to 70% 1RM for the CrM condition during the 2^nd session (p = 0.317). Additionally, peak HR during the 1^st session was significantly lower compared to the 2^nd session for the PLA condition at 60% RM (p = 0.03, d = 0.82).

During the BS, there was significant main effect of intensity, as well as a significant condition × intensity interaction (Figure 2, Table 2). Peak HR significantly increased across all intensities for both conditions during both sessions (p < 0.05, d = 1.32−3.57), but not for the PLA condition from 60% to 70% 1RM during the 1^st session (p = 0.105) or from 70% to 80% 1RM during the 2^nd session (p = 0.64). During the 1^st session, the CrM condition had significantly lower HR at 60% (p = 0.017, d = 1.05) and higher HR at 80% 1RM (p = 0.047, d = 0.82) compared to PLA.

Figure 3 represents the ∆Pre–Post changes (mean ± SD) for HRV indices and blood lactate for CrM and PLA conditions.

Figure 4 represents the mean (±SD) changes in post-session recovery for CMJ, SJ, and upper and lower limb DOMS over time for CrM and PLA.

The ergogenic effects of CrM supplementation manifest distinctly across acute (single-session) and short-term (multi-session) periods. Acutely, CrM supplementation immediately enhanced strength performance, as evidenced by increased repetitions, velocity, and power during the first session. For BP, CrMelevated repetitions at 60% 1RM (d = 1.25) and 70% 1RM (d = 0.75), with peak velocity improvements across all intensities (60% 1RM: d = 1.11; 70% 1RM: d = 0.86; 80% 1RM: d = 2.91). Similarly, in BS, acute effects were observed at 60% (d = 0.79) and 80% 1RM (d = 0.9) for repetitions and at 60% (d = 2.83) and 70% 1RM (d = 2.66) for velocity. Peak power during BP rose acutely (d = 1.16–4.71), with maximal gains at 80% 1RM (d = 4.71). These findings align with CR’s rapid elevation of intramuscular PCr stores, which accelerates ATP resynthesis during near-maximal efforts [4]. Acute benefits may also stem from CR’s osmotic properties, which increase muscle water content and sarcoplasmic calcium handling, facilitating cross-bridge cycling [18,67].

Acute supplementation rapidly increases intramuscular CR content, as demonstrated by Harris, Söderlund [5], who reported that 20 g/day of CR for five days increases total CR storage by roughly 50%, with 20−40% stored as PCr. This biochemical adaptation translates directly to improved performance outcomes, where five days of CR loading (20 g/day) enhanced peak torque production during repeated maximal muscle contractions, particularly in later exercise bouts (bouts 2–4), compared to placebo [7]. Similarly, six days of CR supplementation (5 g/day) mitigated the decline in work output during ten consecutive six-second cycling sprints, reinforcing CR’s role in sustaining high-intensity performance [24]. These findings are in line with Birch, Noble [68], who reported significant increases in peak power (+4.5%) and total work during the first two of three 30-second cycling bouts following five-day CR loading (20 g/day). Notably, performance benefits diminished in the third bout, likely due to ATP depletion.

Mechanistically, acute CR supplementation preserves ATP availability, where five days of CR loading (20 g/day) increased intramuscular PCr by 15%, reducing ATP loss by 30% during repeated 30-second cycling sprints while facilitating a 4% increase in total work output [69]. This highlights CR’s effectiveness in buffering energy demands during near-maximal efforts, reported single 20 g CR doses to improved cycling sprint power due to rapid PCr saturation [70].

When acute ingestion was followed by short-term supplementation (up to three days of CrM intake), the ergogenic effects were further amplified in the subsequent training session, particularly at higher intensities (70–80% 1RM). The second session demonstrated greater BP repetitions at 80% 1RM (d = 1.37) and increased BS peak velocity (d = 4.09), suggesting cumulative performance benefits from repeated CrM dosing. This aligns with Candow, Forbes [71], who observed that five days of CrM loading (20 g/day) increased intramuscular CR retention by ~15%, subsequently enhancing proton buffering during repeated sprint efforts. The intensity-dependent power gains in BP (d = 4.71 at 80% 1RM) and BS (d = 4.97 at 80% 1RM) underscore CR’s critical role in maintaining ATP availability [24]. Interestingly, HR did not significantly decrease during the second session, contrasting with acute HRV improvements, indicating that residual fatigue might obscure cardiovascular benefits during repeated bouts [72].

Overall, acute performance enhancements may result from CrM’s rapid PCr saturation, which buffers ATP depletion during high-intensity efforts [6]. Short-term use could also upregulate the mammalian target of rapamycin (mTOR) signalling and glycogen storage [73], while CR’s antioxidant properties may help mitigate oxidative damage [74]. Additionally, emerging evidence suggests that CrM may play a role in satellite cell activation, potentially facilitating muscle repair [67]. Short-term supplementation protocols (up to six weeks) further amplify these benefits by sustaining intramuscular CR retention. Van Loon, Oosterlaar [75] demonstrated that five days of CrM loading (20 g/day), followed by six weeks of maintenance dosing (2 g/day), elevated muscle CR concentrations and sustained peak power improvements during repeated 12-second cycling sprints, despite partial regression of CR stores toward baseline. Similarly, long-term of CrM supplementation increased fat-free mass, total bench press volume, and cycling work output, highlighting CR’s potential for long-term adaptations [76]. Even shorter CrM protocols yield substantial benefits, while a loading dose of 20 g/day increased peak cycling power by 4% during six consecutive ten-second sprints [77]. Additionally, while another study demonstrated that acute CR supplementation enhanced anaerobic running capacity during a three-minute high-intensity bout [78].

Autonomic effects of acute CrM supplementation were observed in HRV improvements during the first session, with significant increases in RMSSD (d = 2.99) and HF (d = 2.76), reflecting enhanced parasympathetic reactivation [79]. This aligns with recent research reporting that CrM facilitates faster HRV recovery post-resistance exercise, due to its ability to mitigate oxidative stress [70]. Another recent study demonstrated that CR reduces lipid peroxidation and stabilises autonomic function [72]. However, HRV benefits diminished in the second session, reinforcing the hypothesis that accumulated fatigue may overshadow CR’s autonomic advantages during repeated high-intensity efforts [36].

Neuromuscular recovery was superior with CrM supplementation, as sustained CMJ performance at 24 h (d = 1.10) and 48 h (d = 1.26) reflected preserved explosive strength. CrM attenuates post-exercise jump height declines by preserving ATP availability [80]. However, the lack of SJ improvements contrasts with CMJ results, potentially due to SJ’s reliance on concentric strength, which is less sensitive to CR’s effects on stretch-shortening efficiency [81]. Mechanistically, CR enhances titin phosphorylation, improving eccentric force absorption during CMJ [82].

CrM accelerates the restoration of PCr stores, which is critical for rapid ATP regeneration during repeated bouts of exercise. Supplementation of 20 g/day of CrM for 5 days elevates intramuscular total CR by 50% (20–40% as PCr), establishing a biochemical basis for its recovery benefits [5]. This PCr replenishment was directly linked to performance outcomes [18], where CrM supplementation (20 g/day for 5 days) increased peak torque production during repeated maximal contractions, particularly in later exercise bouts (bouts 2–4), compared to placebo. These findings align with Casey, Constantin-Teodosiu [69], who showed that 20 g/day CrM for 5 days reduced ATP loss by 30% during repeated 30-s cycling sprints, despite a 4% increase in total work output. By preserving ATP availability, CrM mitigates the cumulative fatigue associated with high-intensity efforts, a mechanism corroborated by Cooke, Rybalka [80], who linked CrM to sustained CMJ performance at 24 h (d = 1.10) and 48 h (d = 1.26) post-exercise.

Recent molecular insights reveal CR’s role in eccentric force absorption. CR enhances titin phosphorylation, a structural protein critical for passive muscle stiffness, thereby improving energy dissipation during stretch-shortening cycles (e.g. CMJ) [82]. This explains why CR sustains explosive strength recovery but shows limited effects on concentric-dominant tasks, which rely less on eccentric efficiency [81].

Perceptually, CrM supplementation reduced DOMS in upper (d = 1.15) and lower limbs (d = 1.04), consistent with its anti-inflammatory and cell-volumizing properties. CrM lowers creatine kinase (CK) post-exercise by stabilising sarcolemma integrity [83]. Furthermore, CrM demonstrated the capacity to minimise DOMS, particularly in workouts with high eccentric loads (e.g. BS), however benefits are more obvious in the lower limbs due to glycolytic demand [24,84]. CR attenuates exercise-induced muscle damage by stabilising sarcolemma integrity, where Santos, Bassit [83] found that CrM supplementation lowers post-exercise CK levels by 15–20%, reflecting reduced sarcolemma permeability. This aligns with Fukuda, Smith [78], who reported that acute CrM supplementation (20 g/day for 5 days) enhanced anaerobic running capacity during a 3-minute high-intensity bout, likely by minimising membrane disruption and subsequent inflammatory cascades. Doma, Ramachandran [85] confirmed CR’s efficacy in reducing DOMS particularly in exercises with high eccentric loads. These anti-inflammatory effects are partly mediated by CR’s osmotic properties, which increase muscle water content and reduce oxidative stress by scavenging reactive oxygen species [74].

Our trial was designed to evaluate acute and short-term effects; conclusions about long-term adaptations fall outside the study’s scope. Given our study findings, several limitations should be acknowledged. Extensive randomised controlled trials and consensus statements already document long-term safety and efficacy of CrM; our results therefore complement, rather than duplicate, this literature by focusing on acute/short-term responses in resistance-trained males [4,22,86–88]. Although the study was a priori powered for our within-subject primary endpoint, the homogeneous sample (young, recreationally resistance-trained males) limits external validity. Future work should test broader demographics and potential moderators (training status, baseline CR stores, sex). Third, because testing sessions were separated by seven days, complete normalisation of intramuscular creatine cannot be fully ruled out. Although this duration likely minimised acute carryover, as supported by comparable baseline HRV, performance, and lactate measures across sessions, full return to baseline typically requires approximately four to six weeks after supplementation cessation [4,24]. Nevertheless, the present shorter interval aligns with recent applied crossover investigations that also employed seven-day washout periods [40,41], which have demonstrated this duration to be sufficient to limit acute residual effects without disrupting ongoing training. Our decision thus represents a deliberate balance between methodological control and ecological validity, reflecting real-world athletic practice where extended washout periods are often impractical during continuous training or pre-competitive phases. Fourth, only males were enroled to avoid menstrual-cycle–related variability in HRV and performance; cycle-phase control would have required additional sessions beyond this pilot. Future studies are encouraged to include females with appropriate cycle-phase controls. Additionally, while strength performance and HRV were assessed, biochemical markers of muscle damage and oxidative stress were not. Future studies should integrate these measures for a more comprehensive analysis. Lastly, we minimised dietary confounding by instructing participants to avoid creatine-rich foods and all supplements for 72 h before each visit and throughout testing; however, we did not collect detailed diet logs or biomarkers of intake. Future trials will use standardised menus or dietary logs and, where feasible, biomarkers to verify adherence.