Stronger Through the Transition: Evaluating the Effect of Creatine and Resistance Training on Body Composition, Strength, and Sleep in Perimenopause

1. Introduction

Perimenopause is a critical transitional timeframe in the female lifespan, characterized by fluctuations in ovarian hormones, an accelerated loss of lean soft tissue (LST) [1,2], increased adiposity [1,3], and a marked decline in muscle size and quality [4,5]. Secondary to the accelerated reduction of LST, women in perimenopause experience a marked decline in strength [6] and increased symptom burden [7,8], including reduction in sleep quality [9]. The physiological alterations associated with the menopause transition make it a critical window for targeted lifestyle interventions aimed at supporting body re-composition, the process of decreasing fat mass, while increasing LST. Despite the demonstrated need for exercise interventions to support this population, only 9% of exercise physiology research includes midlife women [10], and no studies to date have specifically examined the combination of progressive resistance training (RT) with creatine monohydrate (CM) in perimenopause.

Resistance training is widely-accepted as the most effective way to increase LST [11,12]; however, there is limited research on the direct effects of RT in perimenopause. Existing studies have predominantly used light loads or resistance bands to apply resistance, with only a handful of available of studies in perimenopause [13,14,15,16]. In a study of pre-, peri- and post-menopausal women (N= 70), a low-impact RT program designed to stimulate muscle hypertrophy resulted in modest improvements in LST (~0.2 kg) in 12-weeks, amongst perimenopausal women [15]. The recent American College of Sports Medicine 2026 Position Statement [11] recommends that RT in any form, including use of low-load resistance bands, can support muscle function, hypertrophy, and may be more approachable for novice lifters [11]. The position statement also highlights that progressive RT, increased stimulus across the training period, is considered essential to continue training adaptations, which can be achieved from increasing training load, volume, and frequency [11]. RT with heavier loads allows for greater stimulus with fewer sets, which can enhance time efficiency while still promoting the level of time under tension needed for optimal growth. High intensity resistance training (HIRT) is a progressive RT method that combines higher load, moderate volume, and shorter rest intervals, creating a potent stimulus for both muscle size and strength [17,18,19]. Previous studies implementing HIRT have demonstrated a significant positive impact on LST (+0.15 kg) in young resistance trained women and strength (+62.9 kg) [18] in untrained men and women (Age: 33.0+/- 10.0 years; +0.78 kg LST; +89.9 kg leg press 1RM) [19], in as little as one hour per week.

Creatine monohydrate (CM) has consistently been shown to augment the effects of RT through the increase of intramuscular phosphocreatine stores and rapid adenosine triphosphate re-synthesis, allowing for higher sustained training output and ultimately improving training outcomes, such as greater LST [20,21,22]. In a meta-analysis exploring the effect of CM on LST in postmenopausal women, RT combined with CM resulted in a 0.37 kg increase in LST and an increase in leg press strength maximal strength (+7.5 kg) when pooled across studies [23]. CM supplementation has also demonstrated additive benefit to a 10-week RT program in pre-menopausal healthy, sedentary females (Age: 19-22 years), resulting in an increase in 2.6 kg of LST compared to 1.6 kg in PL [24]. CM has been demonstrated to improve maximal upper body strength in older women (Age: 64.9 ± 5.0 years) by approximately 5.1% compared to PL after 12 weeks of RT (p<0.05) [25]. While CM and progressive RT have not been previously explored in perimenopause, the combination may provide a potent stimulus to rescue adverse changes in body composition and strength.

Sleep disturbances represent one of the most common perimenopausal symptoms, with sleep playing an important role in body composition and exercise adaptations [8,26,27]. CM has demonstrated potential in improving reported insomnia [28], and sleep duration on nights following RT in young women [21]. CM supplementation has specifically demonstrated improvements in cognition during times of sleep deprivation [29], which may align with symptoms experienced during the menopause transition. In a 12-week study examining the effect 5g of CM per day during RT in pre-menopausal women, sleep duration was improved on training versus non-training days (CM: 7.41 hrs; PL: 6.61 hrs, p = 0.013) [21].

Women during the menopause transition often experience adverse changes in body composition, strength, and symptomology, highlighting the need for lifestyle interventions, such as progressive RT and CM supplementation, to support this life phase. The primary purpose of this study was to explore the combined effects of six-weeks HIRT with or without CM on whole-body and regional body composition, muscle size, and muscle quality in perimenopausal women. The secondary purpose of this study was to explore the combined effects of six-weeks HIRT with or without CM on strength, training volume, and sleep characteristics in this population. perimenopause.

2. Methodology

2.1. Participants

A total of 129 women initially qualified via screening with an online questionnaire, of which fifty-five were enrolled in this double-blind, randomized controlled trial (Figure 1; NCT07027800). Participants were randomly assigned via block randomization in a 4:4:3 approach, to either CM supplementation + six-weeks of HIRT (CM; n = 20), non-caloric placebo + six-weeks of HIRT (PL; n = 20), or a smaller planned control group that maintained normal activity with no intervention (CON; n = 15) (Figure 2). A total of four participants dropped due to dissatisfaction being randomized to CON (n = 2) and time commitment associated with the HIRT protocol (n = 2), resulting in a total of 51 participants used for analysis. Participants were included if they were assigned female sex at birth, between 38-60 years of age, and were transitioning through perimenopause, as determined from the National American Menopause Society perimenopause questionnaire section six, including at least two months of menstrual cycle irregularity within the last year and/or experiencing at least two or more perimenopause-related symptoms (Table 1). Female sex hormones (estradiol, progesterone, and follicle stimulating hormone) concentrations were measured via blood serum to confirm perimenopausal status. Participants were included regardless of hormonal contraceptive (HC) or menopausal hormone therapy (MHT) use (HC users: n =21; MHT users: n = 15). Participants demonstrated substantial evidence of menopausal transition, with 100% self-identifying as perimenopausal, 100% experiencing 5 or more symptoms, 55% reporting recent changes in menstrual frequency, 39% reporting changes in bleeding duration, 33% reporting intermenstrual spotting, 35% reporting irregular cycles, and 14% reporting amenorrhea intervals ≥60 days. Women in this study had a body mass index (BMI) between 18.8-34.9 kg/m² and were recreationally active, participating in at least 30 minutes of moderate-intensity physical activity per week, but no more than two days of progressive RT per week. Participants were excluded if they sustained a recent musculoskeletal injury within the previous three months that would limit participation, had a self-identified or clinically diagnosed eating disorder, currently pregnant, or were unwilling to comply with the study protocol. All participants provided written informed consent, approved by the University Institutional Review Board, prior to participation.

2.2. Experimental Design

Participants arrived to all testing visits following a 24-hour abstention from vigorous exercise, 12-hour abstention from caffeine and alcohol, and 8-hour fast from all caloric foods and beverages. Euhydration was confirmed (urine specific gravity, 1.000-1.028) prior to all body composition testing. Body composition, muscle size, and quality were measured from a whole-body dual-energy x-ray absorptiometry (DXA) scan, peripheral computed tomography (pQCT), and ultrasound (US). A WHOOP (WHOOP, Inc., Boston, MA, USA) wearable wristband was provided to all participants and calibrated for seven days to assess sleep and physical activity throughout the duration of the study. Lower and upper body strength was assessed from a one repetition maximum (1RM) of a leg press and bench press, respectively. Participants completed a six-repetition maximum of auxiliary exercises (lunges, shoulder press, bicep curl, tricep extension) to determine the starting training loads. Following baseline testing, participants were randomized; participants in the CM and PL groups completed a 5-day loading phase (4 × 5g Creapure® + 2g non-caloric Crystal Light®) of CM or a non-caloric PL (non-caloric Crystal Light®, 4 × 3.5g), each mixed with 4-6 oz of water, followed by a maintenance phase consisting of 3 × 1.5g CM gummies (Create Wellness, Inc., New York, NY, USA) or 3 PL gummies (Welch’s® Fruit Snacks, Waltham, MA, USA). The maintenance phase was combined with six-weeks of twice-weekly, supervised HIRT. Participants in the training groups were provided with a urinary hormone monitor (Inito, Inc., Bengaluru, India) to use daily, at-home for 30-days during the six-weeks of training. Normal dietary and caffeine intake were maintained prior to HIRT sessions. Baseline assessments were repeated following the 6-week training intervention (CM, PL) and/or maintenance of normal activity (CON).

2.3. Supplementation

Participants were randomized to CM (Creapure® creatine monohydrate AlzChem, Trostberg GmbH, Germany; 5g CM + 2g non-caloric Crystal Light), or a non-caloric PL (3.5g Crystal Light®, Chicago, IL). Participants were instructed to mix their respective supplement with 4-6 ounces of water and consume 4 times per day for 5 days (20g/day). Following the loading phase, a daily maintenance dose was followed for six-weeks. The CM (Create Wellness; 4.5 g/day, 33 kcal; 8 g carbohydrate; 5 g sugar) and PL (Welch’s® Fruit Snacks; 30 kcal; 7 g carbohydrate; 4.3 g sugar) groups both consumed 3 gummies per day. Powder and gummy supplements were provided to participants in pre-measured doses; loading doses were packaged in the same opaque cups, and the maintenance doses were packaged in opaque sachet. Participants were asked to maintain a daily compliance log and return individual dosing packets to the laboratory at the end of the respective loading and maintenance phases. Due to evidence suggesting that consistency of CM supplementation is more important than timing [30], specific timing recommendations were not provided. The research team and participants were blinded to treatments.

2.4. Body Composition

Body composition was assessed using a whole-body DXA (Horizon DXA System, Hologic Inc., Marlborough, MA, USA) scan to determine change in whole body and regional LST and fat mass. Participants were asked to wear light-weight athletic clothing and remove all metal and heavy plastic which could interfere with the DXA scans. Participants laid supine in the middle of the DXA table. Scans were automatically analyzed by the DXA software (APEX Version 5.6.1.1) and confirmed by the DXA technician. Lower extremity custom regions of interest, as previously described in detail [31], were identified to examine contralateral differences between glute, thigh, and lower leg muscle. Test-retest reliability for LST taken from this laboratory with a similar population resulted in an intraclass correlation coefficient (ICC)= 0.998 and standard error of measure (SEM)= 0.78 kg and regional thigh LST ICC= 0.980 and SEM= 0.155kg.

2.5. Muscle Characteristics

2.5.1. Peripheral Quantitative Computed Tomography

Two-dimensional images of the dominant thigh were obtained using a pQCT scanner (pQCT; XCT-3000; Stratec Medizintechnik GmbH, Germany) to determine muscle area (MuA; cm²) and muscle density (MD; mg/cm³). To determine scan position, femur length was manually measured from the greater trochanter to femur lateral epicondyle, and midpoint marked on participant’s skin with a non-permanent marker. Participants laid on a table with their dominant leg extended through the gantry, which were manually positioned above the midpoint mark. The dominant leg was supported by a custom-built support device between the pQCT gantry opening and the table, with the foot secured on the opposite side of the gantry with a Velcro strap placed over the metatarsals. Values were calculated by the pQCT software (Version 6.0). Images were analyzed with via ImageJ software. ImageJ technique reliability based on images from the scanner used in the present study: MuA ICC_2,1 = 0.997, SEM = 2.0544 cm²; MuD ICC_2,1 = 0.761, SEM = 1.50 mg/cm³ [32].

2.5.2. B-mode Ultrasound

A brightness (B)-mode Ultrasound (US; GE LOGIQ-e, Software version R8.0.7, GE Healthcare, Wisconsin, USA) was used to capture a panoramic scan of the thigh to estimate muscle cross sectional area (mCSA; cm²) and echo intensity (EI; arbitrary units [a.u.]) of the vastus lateralis. The same technician performed each scan while the participant laid supine with her right leg extended and relaxed on the examination table for approximately 5 minutes. Measurements were made by applying the device probe directly against the skin at the mid-thigh. During the measurement, the probe slid across the skin surface. All images were exported and analyzed using image-J software. Test-retest reliability for measurements from our laboratory are as follows: mCSA ICC= 0.99, SEM= 0.43 cm² and EI ICC= 0.99, SEM= 0.41 a.u..

2.6. Urinary Hormones

Urine hormone analysis was completed using an at home fertility monitor (Inito Fertility Monitor; Inito Inc., Bengaluru, Karnataka, India). Urinary hormones were measured from all participants in the laboratory at baseline, pre-testing, and post-testing. Participants in the CM and PL groups were provided with an at home phone-mounted device, paired with a smartphone app, along with urine analysis test strips to use for a 30-day period during the HIRT training period and accounted for in training volume analysis. Daily hormones were evaluated at home from the first urine void of the morning. Participants were asked to dip one hormone test strip into the collected urine per day, at a similar time of day, and then insert it into the monitor for analysis. The test provided estimates of hormone concentrations for estrogen (estrone-3-glucuronide; E3G), follicle-stimulating hormone (FSH), luteinizing hormone (LH), and progesterone (pregnanediol-3-glucuronide; PdG). Values were collected automatically and downloaded by the research team. 30-day under the curve was calculated from the trapezoidal method (Equation 1).

Equation 1: AUC = ${\displaystyle \frac{\mathsf{\Delta }x}{2}}{\displaystyle \sum _{i=1}^n}[f\left(x_{i+1}\right)+(x_i\left)\right]$

2.7. Maximal Strength Testing

To assess strength and determine starting weight for HIRT, participants completed a 1RM for leg press and bench press in accordance with the National Strength and Conditioning Association (NSCA) recommendations [33]. Each participant performed a warm-up set of eight to ten repetitions, with a weight that is approximately 50% of the anticipated 1RM. Participants then rested for two minutes, after which a set of four to six repetitions was completed with a load of 80% of the predicted 1RM. After another two-minute rest period, the weight was increased to an estimated 1RM load and participants attempted a single repetition with the weight. After the completion of each successful 1RM attempt, the weight was increased until failure was reached, with two to three minutes of rest between 1RM attempts. Multiple repetition maximum tests, specifically a six repetition-maximum (6RM), were completed with dumbbells for lunges, shoulder press, bicep curl, and triceps extension [34]. The research staff determined what weight participants should begin with for each exercise, based on the prior training of the individual, aiming for three to ten successful repetitions [35]. Participants were allowed two minutes of rest between each accessory exercise. Data obtained from the strength testing assessments were used to determine the load for the HIRT sessions.

2.8. High Intensity Resistance Training

Participants completed two training sessions per week, in person under one-on-one supervision from laboratory staff. Sessions occurred with a minimum of 24 hours in between sessions. Prior to each session, a five-minute warm-up was completed on a treadmill at a self-selected pace. At each session, exercises were performed in the following order: leg press, bench press, lunges, shoulder press, biceps curl, and triceps extension. Three sets of each exercise were completed for six to ten repetitions, with 30 seconds rest between sets and two minutes of rest in between each exercise. To progressively overload, weight was increased once a participant successfully completed ten repetitions for two sets. Load was then increased for lower body exercises by 10% and 5% for upper body. Total volume was calculated by multiplying reps by weight (lbs) for each exercise.

2.9. Sleep Measures

Sleep was assessed continuously throughout the 6-week intervention using the WHOOP strap 5.0 (WHOOP, Inc., Boston, MA), a wrist-worn wearable device that combines accelerometry and photoplethysmography to detect sleep periods of at least one hour in duration. Within each detected sleep period, each 30-second interval was classified into one of four stages: wake, light, slow wave sleep (SWS), and rapid eye movement (REM) sleep. The following sleep outcomes were examined: sleep duration (total sleep time in minutes), sleep efficiency (percentage of time spent asleep relative to total time in bed), SWS percentage (%), REM sleep percentage (%), sleep disturbances (count), normalized sleep disturbances (disturbances per hour of sleep), and sleep consistency (%). To evaluate the acute effects of HIRT on sleep, and whether these effects differed between CM and PL, a within-person comparison of sleep on post-training nights versus non-training nights was performed. For each participant, nights were classified as post-training (the night following a supervised HIRT session) or non-training (all other nights during the intervention period). Sleep consistency is a WHOOP-derived metric adapted from the Sleep Regularity Index, which evaluates sleep onset and offset times over a rolling window, comparing the most recent 24 hours to the preceding four days, with comparisons of days further apart being assigned progressively lower weights [36]. The metric is reported on a 100-point scale, where 0 represents complete misalignment and 100 represents perfect alignment between the most recent and preceding sleep schedules. Wearable-derived sleep detection and staging algorithms from the WHOOP device have been validated against polysomnography, demonstrating 86-89% agreement for two-stage sleep classification [37,38].

2.10. Statistical Analyses

Descriptive statistics were presented as mean ± standard deviation, with a one-way analysis of variance used to evaluate baseline values of participant height, weight, LST, and sex hormones. Baseline sleep characteristics were compared between CM and PL using independent samples t-tests. Pairwise Fisher’s exact tests were conducted for selected symptoms due to small cell counts. A mixed factorial analysis of covariance (ANCOVA) was used to determine between group (CM vs PL vs CON) differences for changes (week 6 – week 0) in LST, fat mass, MuA, MuD, mCSA, EI, and 1RM co-varied for baseline urinary E3G:P concentrations and baseline characteristics (Table 1). 30-day area under the curve was determined for E3G:P concentration and used as a covariate for differences in training volume between CM and PL. When significant interactions were observed, Bonferroni post-hoc pairwise comparisons were used. The sleep outcomes were examined using linear mixed-effects models (LMM) fit with restricted maximum likelihood via the lmerTest package in R. All models included fixed effects for post-training status (post-training vs. non-training), treatment group (CM vs. PL), their interaction, age, body mass index (BMI), and the participant's baseline mean for the respective outcome, with a random intercept for participant. When a significant interaction was detected, estimated marginal means (EMMs) were computed using the emmeans package, and the simple effect of post-training status was evaluated within each group with Tukey adjustment. Statistical significance was set a priori at α = 0.05. Sample size was determined a priori with muscle size, quality, and strength as primary outcomes, producing an average effect size of 0.48 with a sample size of N = 42 with 3 unequal groups to achieve a power of 0.8, nonsphericity correction of 1, and a significance level of 0.05. Power was calculated using G Power Version 3.1.9.2, with an analysis of variance (ANOVA) F-test. All statistical analyses were completed using IBM SPSS Statistics (Version 29.0.2.0, Armonk, NY: IBM Corp) and R (version 4.4.2).

3. Results

There were no significant differences in baseline characteristics between groups for age (p = 0.299), height (p = 0.328), weight (p = 0.789), or BMI (p = 0.402; Table 1). Baseline body composition was similar across groups, with no significant differences in LST (p = 0.939) or body fat percentage (BF%; p = 0.436) between groups. There were also no significant differences observed between groups in baseline urinary E3G:PdG (p = 0.124) or 30-days of serial measurements used to determine area under the curve (AUC) between CM and PL (p = 0.124).

Menopause symptoms were assessed to determine inclusion criteria and baseline status across groups based on a Fisher’s exact test (Table 2). Focused analyses using Fisher’s exact tests for representative symptoms demonstrated no consistent differences in symptom prevalence across groups. For cognitive symptoms, nearly all participants reported memory concerns (CM: 89%, PL: 85%, CON: 92%) and difficulty concentrating (CM: 89%, PL: 85%, CON: 85%), with identical or near-identical prevalence and non-significant Fisher’s exact comparisons (all p = 0.999). Similarly, fatigue was highly prevalent across groups (CM: 94%, PL: 90%, CON: 69%), with modest numerical differences that did not reach statistical significance (all p > 0.10). Vasomotor symptoms such as hot flashes were reported by 55–61% of participants across groups, again with no significant differences (all p ≥ 0.48). Joint pain showed the largest numerical variation (CM: 56%, PL: 70%, CON: 31%); however, this pattern was not consistent across pairwise comparisons, with only one isolated comparison reaching nominal significance (PL vs CON, p = 0.038), likely reflecting small sample size and multiple testing rather than a true group effect. Overall, these findings indicate highly comparable symptom profiles across groups at baseline, supporting baseline equivalence.

3.1. Body Composition

3.1.1. Whole-Body Lean Soft Tissue

There was no significant main effect for the change in whole-body LST (p = 0.066; ŋ² = 0.182) and no significant treatment effects (p = 0.093; ŋ² = 0.105). CM had a greater increase in LST compared to PL [mean difference (MD) ± standard error (SE) 0.380 ± 0.431 kg; p = 0.999] and compared to CON (1.052 ± 0.474 kg; p = 0.096). PL demonstrated a greater change in LST compared to CON (0.672 ± 0.442 kg; p = 0.407) when co-varied for urinary E3G:P concentration and baseline LST values (Figure 3A).

3.1.2. Regional Lean Soft Tissue

There were no significant main effects for the change in any regional outcomes of LST (p = 0.285-0.958; ŋ² = 0.015-0.106) or between treatments (p = 0.308-0.885; ŋ² = 0.006-0.052; Table 3). CM demonstrated larger increases of LST in the glute regions compared to PL and CON (p = 0.858-0.999). PL demonstrated greater increases in LST in the thigh region compared to both CM and CON (p = 0.976 – 0.999). CM had similar changes in right calf LST compared to CON (p = 0.999), but larger increases compared to PL and large increases for left calf compared to PL and CON (p = 0.564 – 0.999). All analyses were adjusted for urinary E3G:PdG and corresponding baseline variables.

3.1.3. Whole-Body Fat Mass

There was no significant main effect for the change in whole-body fat mass (p = 0.072; ŋ² = 0.178) and no significant between treatment effects (p = 0.477; ŋ² = 0.034). CM demonstrated a larger decrease in fat mass compared to PL (MD ± SE: -0.216 ± 0.358 kg; p = 0.999) and compared to CON (-0.486 ± 0.397 kg; p = 0.681). PL also had a greater loss in fat mass compared to CON (-0.270 ± 0.361 kg; p = 0.999) when co-varied for urinary E3G:P concentration and baseline fat mass values (Figure 3B).

3.2. Muscle Characteristics

3.2.1. pQCT: Muscle Area

There was a significant main effect for the change in muscle area (p = 0.011; ŋ² = 0.251) and a significant difference between treatments (p = 0.025; ŋ² = 0.154). Post hoc comparisons demonstrated CM resulted in the largest change in muscle area compared to PL (MD: 2.00 ± 1.99 cm²; p = 0.966) and to CON (MD: 6.11 ± 2.19 cm²; p = 0.023). PL also had a greater change in muscle area compared to CON (MD: 4.12 ± 2.08 cm²; p = 0.162) when adjusted for baseline urinary E3G:P concentration and muscle area values (Figure 4A).

3.2.2. pQCT: Muscle Density

There was a significant main effect for the change in muscle density (p < 0.001; ŋ² = 0.394), but no significant between treatment effects (p = 0.269; ŋ² = 0.058) when adjusted for baseline urinary E3G:P concentration and muscle density values (Figure 4B).

3.2.3. Ultrasound: Muscle Cross Sectional Area

There was a significant main effect for the change in ultrasound derived mCSA (p = 0.018; ŋ² = 0.242) and a significant between treatment effect (p = 0.013; ŋ² = 0.187). Post hoc comparisons demonstrated a significant increase in mCSA for CM when compared to CON (3.222 ± 1.051 cm²; p = 0.011), and a greater, non-significant increase in mCSA compared to PL (1.205 ± 1.071 cm²; p = 0.800). PL resulted in a non-significant increase in mCSA compared to CON (2.017 ± 1.049 cm²; p = 0.184) when co-varied for urinary E3G:P concentration and baseline mCSA (Figure 4C).

3.3. Ultrasound: Echo Intensity

There was no significant main effect for the change in ultrasound-derived EI (p = 0.677; ŋ² = 0.056) and no main between treatment effects (p = 0.397; ŋ² = 0.046). There was no significant change in EI for either group (CM: Δ 2.257 ± 2.866 a.u.; PL: Δ 2.755 ± 2.683 a.u.; CON: Δ 7.235 ± 3.09 a.u.) when co-varied for urinary E3G:P concentration and baseline EI values (Figure 4D).

3.4. Maximal Strength

3.4.1. Leg Press

There was a significant main effect for the change in leg press 1RM (p = 0.002; ŋ² = 0.341) and a significant between treatment effect (p < 0.001; ŋ² = 0.339). Post hoc comparisons demonstrated a significant increase in leg press 1RM for CM, when compared to CON (139.6 ± 34.1 lbs; p < 0.001), and no difference compared to PL (13.4 ± 30.9 lbs; p = 0.999). PL also resulted in a significant increase in leg press 1RM compared to CON (126.2 ± 31.5 lbs; p = 0.001) when co-varied for urinary E3G:P concentration and baseline leg press 1RM (Figure 4E).

3.4.2. Bench Press

There was a significant main effect for the change in bench press 1RM (p = 0.007; ŋ² = 0.277) and a significant between treatment effect (p = 0.015; ŋ² = 0.180). Post hoc comparisons demonstrated a significant increase in bench press 1RM for CM when compared to CON (9.0 ± 3.4 lbs; p = 0.031), and no difference compared to PL (0.9 ± 3.0 lbs; p = 0.999). PL also resulted in a significant increase in bench press 1RM compared to CON (8.1 ± 3.0 lbs; p = 0.032) when co-varied for urinary E3G:P concentration and baseline bench press 1RM (Figure 4F).

3.5. Training Volume and Rating of Perceived Exertion

There was no significant main effect for training volume (p = 0.343; ŋ² = 0.074) and no between treatment effects (p = 0.207; ŋ² = 0.056). CM resulted in an increase in training volume compared to PL (17945.55 ± 13902.16 lbs; p = 0.207) when co-varied for 30-day AUC urinary E3G:P concentration. Weekly training volume was consistently higher in the CM group compared to PL across six-weeks, but differences did not reach statistical significance (p = 0.931 - 0.244; Figure 5). There was no significant main effect for average rating of perceived over time (p = 0.387; ŋ² = 0.066) and no main effect of treatment (p = 0.948; ŋ² < 0.001). The average session RPE over six-weeks of HIRT was similar between CM (17.4 ± 0.3 a.u.) and PL (17.4 ± 0.3 a.u.).

3.6. Sleep Outcomes

3.6.1. Sleep Consistency and Sleep Disturbances

Significant main effects of training status were observed for sleep consistency, sleep disturbances, and normalized sleep disturbances. Sleep consistency was higher on training nights compared to non-training nights across both groups (+3.76%; p < 0.001). Both CM and PL exhibited higher consistency following training sessions (CM: 78.0 ± 6.5% vs. 73.9 ± 6.3%; PL: 74.9 ± 7.8% vs. 72.2 ± 8.8%). Sleep disturbances were also significantly higher on training nights (+0.97; p < 0.001), as were normalized disturbances (+0.13 disturbances/hr; p < 0.001). Raw disturbances increased modestly following training in both groups (CM: 8.7 ± 3.4 vs. 7.8 ± 3.1; PL: 8.5 ± 2.1 vs. 8.1 ± 2.0).

3.6.2. Sleep Duration and Efficiency

No significant main effects of training status were observed for sleep duration (+3.92 min; p = 0.107), sleep efficiency (+0.03%; p = 0.355), SWS% (-0.37%; p = 0.099), or REM% (+0.21%; p = 0.522). No significant training × treatment interactions (p > 0.119) or main effects of treatment group (p > 0.210) were observed for any outcomes.

4. Discussion

Perimenopause is marked by compounding adverse physiological changes, including an accelerated loss of LST, muscle quality, and strength [1,2,4,6]. While these deleterious impacts make perimenopause an opportune time for targeted interventions, we are aware of no studies that have examined the combined impact of CM supplementation and progressive resistance, which has been effectively used to improve muscle in younger females and males [39]. Both training groups (CM and PL) in the present study observed clinically meaningful improvements in body re-composition, with CM supplementation augmenting effects of RT with a 27% greater increase in LST and 28% decrease in fat mass, in comparison to PL. Muscle size increased by 0.9 cm² with CM compared to PL, demonstrating a potential added benefit of CM to enhance RT adaptations. Across both RT groups, significant increases were observed in maximal strength for the 1RM leg press (~67%) and bench press (~83%) when compared to CON. Notably, there were zero sustained injuries from the training intervention or adverse side effects from CM supplementation, further supporting the safety and efficacy of progressive RT+CM in perimenopausal women. No differences were observed in training volume or sleep outcomes between groups. Collectively, these findings demonstrate that HIRT is an effective and time-efficient training strategy. The addition of CM appears to enhance adaptations to further improve LST and strength during the menopause transition.

Perimenopause is commonly associated with adverse changes in body composition, including an accelerated loss of LST and increased central adiposity [1,2,3]. The present study utilized a 6-week combined CM supplementation and HIRT [17,18] protocol leading to meaningful changes in LST and fat mass, despite no dietary changes. CM supplementation resulted in an additive increase of 1.02 kg of LST compared to CON (p = 0.099), while PL had a 0.67 kg greater increase compared to CON, both exceeding our SEM (0.78 kg) from DXA. Consistent with whole-body changes, an increase of 0.28 kg of LST in the thigh region across groups, also surpassing SEM (0.16 kg) [31]. A recent meta-analysis evaluating the impact of CM on LST in post-menopausal women (age: 57-70 yrs) reported similar results; CM when combined with RT resulted in an average 0.32 kg greater LST compared to RT alone [23]. In the present study, fat mass also resulted in a decrease in the CM group with an average 0.81 kg loss compared to 0.41 kg in PL and 0.13 kg in CON groups. This trend is consistent with previous literature in young males [40,41,42], which has reported a greater average fat loss of ~0.70-1.40 kg CM with RT, compared to PL+RT, without dietary alterations. To our knowledge, this is the first study to demonstrate a similar response in perimenopausal women. Changes in body re-composition are multifactorial, and the favorable changes observed with the use of CM supplementation are likely attributable to a combination of greater sustained training volume [43], muscle protein synthesis [44], and recovery [20,44,45,46], all indirectly supporting greater muscle hypertrophy.

When combined with RT, CM has been shown to enhance site-specific hypertrophy, including thigh and lower leg muscle cross-sectional area and thickness [46], with some data also showing enhancements in muscle quality [22]. In the present study, CM supplementation resulted in improvements in muscle size compared to CON for both pQCT MuA (MD: 6.11 cm²) and US mCSA (MD: 3.22 cm²). In contrast, PL pQCT MuA did not differ (MD: 2.00 cm²), nor did US mCSA (MD: 2.017 cm²), compared to CON. CM supplementation has consistently resulted in increased muscular hypertrophy when combined with RT [22,23,39,46,47]. The increase in muscle hypertrophy is largely indirect, as a result of greater ATP resynthesis [46,47], increase in intracellular fluid [48], and thus enhanced training volume and greater time under tension. Previous work has been conducted predominately in young adults, with one study of resistance-trained females and males (Age: 18-38 years) examining six-weeks of RT combined with CM supplementation and found a 0.46 cm increase in ultrasound-derived knee extensor muscle thickness compared to 0.24 cm increase in the PL group [49]. Notably, in the present study muscle density significantly increased by Δ 1.46 g/cm³ in the PL group compared to 0.826 Δ g/cm³ in the CM group, suggesting an improvement in muscle quality as a result of HIRT, independent of CM supplementation. These findings align with previous literature in postmenopausal females demonstrating that one year of RT combined with CM improved muscle density by approximately 0.83 g/cm³ measured with pQCT compared to PL (p = 0.016), despite no differences in muscle area [22]. The discrepancy between findings are likely attributable to differences in training program duration (six weeks versus one year). Additionally, CM has been shown to result in short-term shifts in fluid dynamics, resulting in an increase in transient swelling with CM and high-intensity exercise [48,50]. The increase in intracellular fluid content reported as a result of CM supplementation [48] has been attributed to greater cell swelling and increased muscle hypertrophy [45,51]. Accordingly, the lack of changes in muscle density observed in the present study, may reflect alterations in fluid dynamics, with increases in fluid within the muscle established with CM supplementation impacting imaging quality resulting in lower density values. Future studies should account for intra- and extracellular fluid alterations and cellular fluid dynamics [48].

Strength is a critical component of function and linked to healthy aging [52], yet minimal studies utilize high load RT programs in perimenopausal women [14,16,53]. Maximal lower and upper body strength was significantly improved in both treatment groups in the present study. CM and PL saw similar changes in adjusted 1RM strength for a leg press (CM: Δ 203.1 ± 23.4; PL: Δ 189.8 ± 19.7 lbs) and bench press (CM: Δ 10.2 ± 2.3; PL: Δ 9.3 ± 1.9 lbs), all significantly greater compared to CON. These changes in strength are similar to previous strength outcomes in post-menopausal women, with CM improving leg press 1RM compared to PL by an average ~16.5 lbs [23]. No significant differences were observed in training volume or RPE between treatments in the present study, consistent with previous work examining a 12-month protocol [43]. Collectively, these findings highlight that high-load RT programs are beneficial at improving strength in perimenopausal women, with CM supplementation adding a slight additive benefit to 1RM strength.

Perimenopause is often accompanied by high symptom burden, with adverse changes in sleep being commonly reported, including in the present study’s sample (Table 2). While CM has demonstrated potential for enhancing sleep, this has largely been observed in populations experiencing baseline sleep deprivation [54]. In a study of peri- and post-menopausal women, creatine hydrochloride + creatine ethyl ester supplementation improved symptoms related to insomnia when evaluated on a visual analog scale (Δ-1.3 cm; p = 0.57) [28]. In the present study, CM did not augment any sleep outcomes on training versus non-training nights throughout the HIRT program, but did marginally support sleep consistency on training nights (CM: 78.0 ± 6.5% vs.; PLA: 74.9 ± 7.8%) compared to PL. Notably, regardless of treatment group, HIRT did result in improved sleep consistency on training nights (+3.76%, p < 0.001) increased sleep disturbances and normalized sleep disturbances (+0.97, p < 0.001). Participants in the present sample were already meeting recommended sleep durations, averaging 6.96- 7.45 hours per night, suggesting a potential ceiling effect that would limit the ability for CM supplementation to further improve sleep. While exercise training may also generally enhance sleep outcomes, the research on the impact of HIRT on sleep in midlife has not previously been explored. In post-menopausal women (Age: 66.4-71.8 years), 12-weeks of progressive RT improved subjective sleep scores measured via the Pittsburgh Sleep Quality Index, with the largest benefit in individuals with baseline poor sleep [55]. Similarly, a study of pre-menopausal women (Age: 18-24 yrs) consuming 5g of CM/day, and completing two days of RT per week, reported an increase in total sleep duration (CM: 7.41 hrs; PL: 6.61 hrs) on training days in the CM group only (t(30.92) = 2.65, p = 0.013), measured via Oura rings [21].

Although this is the first study to evaluate CM and progressive RT in perimenopause, it is not without limitations. First, women were intentionally included with a range of RT experience to broaden generalizability. However, this likely impacted the morphological changes seen in LST and mCSA. A longer duration training protocol (≥ 8 weeks) would likely have resulted in a greater magnitude increase in LST, aligning with more pronounced morphological changes. Lastly, dietary intake was not modified; general guidance was given to all women regarding nutrition and nutrient timing following enrollment, but no specific changes were made which may have indirectly impacted recovery from training, body composition, and potentially sleep [56,57]. Despite these limitations, changes in body composition and muscle size were observed beyond measurement error from six-weeks of progressive RT, with and without CM, in perimenopausal women with a wide-range of RT experience.

5. Conclusion

In the present randomized, double-blind, placebo-controlled trial, as little as six-weeks of HIRT elicited clinically meaningfully morphologic changes to muscle in perimenopausal women. Although statistically significant differences between CM and PL in whole-body LST were not detected, the magnitude of change exceeded measurement error, suggesting a greater combined effect from CM + HIRT on LST. Perimenopause represents a critical timeframe for intervention in the female lifespan, yet previous work has primarily used low-loads and high-repetitions [14,15,53,58]. While it is often assumed that lifting heavy weights may inherently lead to injury, both CM and PL groups demonstrated meaningful increases in upper- and lower-body strength following a six-week intervention without injury, highlighting efficacy and safety of participating in heavy progressive RT in perimenopause. Progressive RT programs should be the primary focus for women in perimenopause with body re-composition and strength goals. Combining CM supplementation did not meaningfully enhance muscle quality outcomes, but it may provide slight additive benefit for increasing LST, strength, and muscle size, while simultaneously reducing fat mass. Future research should continue to explore the effects of longer-term resistance plus aerobic training for cardiometabolic health outcomes, as well as explore more the added benefits of dietary alterations. Lastly, continuing to expand data in this population to understand genetic influence, implications of baseline risk stratification, and the addition of menopause hormone therapy should be considered.