A review of the sports medicine literature reveals a clear underrepresentation of female athletes in research. In the current era of precision medicine, increasing attention has been directed toward the regulatory roles of estrogen and progesterone in athletic performance and health optimization. Regular fluctuations in estrogen and progesterone across menstrual cycle phases (i.e., early follicular, late follicular, and mid-luteal phases) may influence strength performance, hydration status, body composition, and energy balance. However, few studies have examined these outcomes using hormonal confirmation of menstrual cycle phases. Monophasic oral contraceptive use also represents a highly relevant hormonal condition among female athletes, as exogenous hormones suppress endogenous ovarian fluctuations and create distinct hormonal profiles across active pill-consumption and withdrawal phases. Nevertheless, the influence of oral contraceptive phases on strength-related outcomes, hydration markers, body water regulation, body composition, and energy balance remains insufficiently characterized, particularly in comparison with naturally menstruating athletes. In response to these gaps, this longitudinal observational study primarily aims to examine variations in strength-related outcomes across three distinct menstrual cycle phases (early follicular, late follicular, and mid-luteal) in eumenorrheic athletes. Secondary objectives include: (i) conducting within- and between-group comparisons of hydration status, energy balance, and strength outcomes (maximal, endurance, and explosive torque) in naturally menstruating athletes and oral contraceptive users; (ii) testing the reliability of methods used to assess body water, energy expenditure, and body composition across the menstrual cycle; (iii) exploring associations between energy availability, resting energy expenditure, and sex hormone concentrations across menstrual cycle phases; and (iv) testing, validating, and proposing methodological recommendations for the use of bioelectrical impedance analysis in tracking fluid-related changes across hormonal phases. To achieve these goals, the study will use a longitudinal observational design involving 40 female athletes, including 24 naturally menstruating athletes and 16 oral contraceptive users. Naturally menstruating athletes will be assessed during the early follicular, late follicular, and mid-luteal phases of the menstrual cycle, while oral contraceptive users will be assessed across pill-consumption and withdrawal phases. Measurements will be conducted across the three menstrual cycle phases and across oral contraceptive use phases, and will include: i) maximal voluntary isometric strength assessed using handgrip dynamometry, bench press, and leg press; ii) serum estrogen and progesterone; iii) body water and its compartments, and water turnover by dilution techniques; iv) hydration status by plasma osmolality, sodium, and vasopressin; v) energy balance by doubly labeled water and body composition changes; vi) resting energy expenditure by indirect calorimetry.
BACKGROUND: Sports medicine has revealed a clear underrepresentation of female athletes in research, with findings from male participants often being used as a proxy for females. In the new era of precision medicine, the focus has shifted to the regulatory role of sex hormones in enhancing performance and optimizing athletic health. Sex hormone levels are relatively stable on a daily basis in men but fluctuate throughout the menstrual cycle (MC) in women. Menstrual status can range from amenorrhoea to natural menstruation (i.e., MC lasting 21-35 days), with the latter being classified as anovulatory, luteal phase-deficient, or ovulatory. The MC is divided into two main phases: the follicular phase (lasting from menstruation to ovulation) and the luteal phase (lasting from ovulation to the onset of menstruation). Regular fluctuations in sex hormones occur during an ovulatory MC, with three hormonal phases based on estrogen and progesterone fluctuations: early follicular (low levels of estrogen and progesterone), late follicular (high levels of estrogen and low levels of progesterone), and mid-luteal (high levels of estrogen and progesterone). MC duration and hormone profiles show large individual variability due to lifestyle factors or oral contraceptive (OC) use. The most commonly used are combined OCs, which lead to ovarian suppression through negative feedback from exogenous estrogen and progesterone. Estrogen may contribute to increases in muscle mass by improving intrinsic muscle quality, enabling muscle fibres to generate force, and increasing collagen levels in connective tissues, whereas progesterone is associated with protein catabolism, potentially reducing muscle strength. Adequate levels of strength are critical for improving performance, reducing injury risk, and enhancing sport-specific skills (e.g., managing body mass, running, and jumping). Maximal strength, endurance strength (fatigue-resistance-related motor tasks), and power are crucial in sports, but the variations associated with the MC are not clearly understood. These misconceptions are partly related to hormone fluctuations throughout the MC. Some authors suggest that strength performance varies across MC phases and improves when estrogen levels are higher (e.g., during the follicular phase), while others consider these variations inconclusive. Additionally, variations in fluid and energy balance due to hormonal fluctuations throughout the MC are related to muscle power and strength, but their role has yet to be explored and requires further investigation. On the other hand, low energy availability, defined as less than 30 kcal/kg fat-free mass/day in women, adversely affects the secretion of sex hormones (e.g., estrogen and progesterone) and causes functional hypothalamic amenorrhea, which decreases sarcoplasmic and myofibrillar protein synthesis in athletes, thereby reducing strength. These hormonal fluctuations and their impact on energy availability also extend to fluid balance. Fluid balance is affected because both estrogen and progesterone can influence thermoregulatory and fluid-regulatory systems, likely causing shifts in water compartments. In women, estrogen and progesterone can affect thirst, fluid intake, and sodium regulation, thereby influencing hydration status throughout the MC. Estrogen affects the threshold for vasopressin release, whereas progesterone increases aldosterone and vasopressin levels, collectively increasing fluid retention and potentially expanding extracellular water, especially during the mid-luteal phase (higher estrogen and progesterone levels). To date, no studies have investigated hydration status, including water compartment changes, across MC phases or in OC users among athletes. Hydration status affects strength and athletic health, but how fluid shifts occur throughout MC-related hormonal fluctuations remains poorly understood. Previous research has observed that intracellular water gains, as assessed by dilution techniques, improved strength and jump height over the course of an athletic season; however, neither MC phases nor OC use were considered. Although dilution techniques are considered the reference method for assessing water pools, the cost and time involved limit their use in research, clinical, and field settings. Therefore, new methods that assess hydration status safely, accurately, reliably, and feasibly are needed. Bioelectrical Impedance Analysis (BIA) has been used as a simple and feasible method for assessing water pools. However, concerns remain regarding its applicability across MC phases. In summary, evidence is needed regarding variations in strength-related outcomes under different hormonal environments throughout the MC. Additionally, variations in hydration status and water compartments throughout the MC warrant further investigation. Understanding the differences in strength- and hydration-related outcomes between naturally menstruating athletes and those using OCs also requires clarification. Concerns regarding the applicability of BIA during the MC persist, and further clarification is needed to distinguish physiological from bioelectrical characteristics as markers of fluid-related shifts. Considering the gaps mentioned above, the primary aim of this study is to analyse variations in strength-related outcomes (i.e., endurance, maximal, and explosive strength) resulting from hormonal fluctuations across three distinct and identified MC phases (i.e., early follicular, late follicular, and mid-luteal) in naturally menstruating athletes. Secondary aims include: (1) comparing hydration status, energy balance measures (energy expenditure and body composition), and strength-related outcomes (maximal, endurance, and explosive) within and between naturally menstruating athletes and those using OCs; (2) testing the reliability of methods used to assess fluid changes, energy expenditure, and body composition throughout the MC; (3) exploring whether energy availability and resting energy expenditure are associated with sex hormone production across MC phases in naturally menstruating athletes; and (4) testing, validating, and proposing recommendations for the use of bioelectrical impedance in tracking fluid changes throughout the MC. STUDY DESIGN: This study will use a longitudinal observational design involving two groups of female athletes: naturally MC athletes and OC users. Participants will be assessed across three distinct phases of their menstrual cycle or oral contraceptive cycle. Testing procedures will be standardized across all sessions to minimize between-session variability. A block randomization approach will be used to determine whether participants will begin testing during the follicular or luteal phase, while OC users will begin assessments either during the withdrawal phase or the active pill phase. For naturally menstruating athletes, day 1 of a new menstrual cycle will be defined as the onset of menstruation. For OC users, day 1 will correspond to the first day of active pill consumption within the 21-day active pill cycle. The identified menstrual cycle phases for naturally menstruating participants will include the early follicular, late follicular, and mid-luteal phases. For OC users, assessments will be conducted during week 1, weeks 2-3, and week 4 of the contraceptive cycle. All assessments will take place at the Exercise and Health Laboratory, Faculty of Human Kinetics, University of Lisbon. SUBJECTS: Female athletes (N=40) will participate, divided into two groups: natural MC (N=24) and OC users (N=16). MEASURES: Measurements will be conducted across three menstrual cycle phases (i.e., early follicular, late follicular, and mid-luteal) and across the oral contraceptive cycle during week 1, weeks 2-3, and the withdrawal week (week 4). Assessments will include: (i) maximal voluntary isometric strength assessed using handgrip dynamometry, bench press, and leg press; (ii) serum estrogen and progesterone concentrations; (iii) body water compartments and water turnover assessed using dilution techniques; (iv) hydration status assessed using plasma osmolality, sodium concentration, and vasopressin; (v) energy balance assessed using doubly labeled water, and body composition changes assessed using a 4-compartment model; and (vi) resting energy expenditure assessed by indirect calorimetry. SAMPLE POWER ANALYSIS: Sample size was calculated based on observed effect-sizes of 0.7 and 0.9, respectively for within (MC phases) and between-group differences (natural MC vs. OC). To detect differences using a type I error of 5% and a power of 80% (G\*Power v.3.1.9.2), within two MC phases using an effect size of 0.7, resulting in 15 participants while for detecting differences between groups, 16 athletes would be required. Sixteen athletes per group are needed but 8 additional athletes will be enrolled in the natural MC as \~50% of the exercising women had ovulatory cycles, totalling 40 athletes (24 for the natural MC and 16 under OC). STATISTICAL ANALYSIS: Statistical analysis will be performed using IBM SPSS statistics version 28.0 (IBM, USA) and RStudio (Version 1.4.1717, RStudio Team, MA). To perform the block randomization of the testing session to start in either the follicular or luteal phases or to start in the placebo pill or OC pill phase, a computer algorithm written in RStudio (Version 1.4.1717, RStudio Team, Boston, MA) will be employed with randomly selected block sizes. To assess the variation in primary outcomes due to the MC phases (i.e., early follicular, late follicular, and mid-luteal), linear mixed-effects models including the menstrual phases and potential confounding factors will be used. The covariance matrix for repeated measures within subjects over time will be modeled as Unstructured or, if necessary, Compound Symmetry. Model residual distributions will be examined graphically to identify which specific phases differ from each other. To compare natural MC and OC users for the secondary outcomes, mixed-design analysis of variance (ANOVA), with factors (phases) and group (natural MC and OC users), will be performed, and Student's unpaired or paired t-tests will be used post hoc to investigate any significant model effects or interactions. Multiple regression analysis will be employed to test the association between energy availability and sex hormones in natural MC athletes, adjusting for potential confounding factors. To test the validity of alternative techniques for assessing energy expenditure, body composition, and water compartments, Student unpaired or paired t-tests will be performed. Also, multiple regression analysis will be used to investigate the association between alternative and reference methods whereas the agreement between methods will be assessed by using the Bland-Altman approach (also used for testing the reliability of the methods), including the analysis of the correlation between the mean and the difference of the methods. Statistical significance will be set at p\<0.05 (2-tailed).
Study Type
OBSERVATIONAL
Enrollment
40
Faculdade de Motricidade Humana
Cruz Quebrada, Oeiras, Portugal
RECRUITINGMaximal Strength
Upper- and lower-body maximal strength will be assessed during the three evaluation periods using maximal voluntary isometric contractions performed on a bench press machine and a horizontal leg press, respectively. For each exercise, participants will perform three 5-second maximal voluntary isometric contractions, with 1-minute rest intervals between trials. All contractions will be executed isometrically, without joint movement. Maximal strength will be defined as the highest peak force value obtained during the maximal voluntary isometric contraction trials.
Time frame: 4 weeks
Explosive Strength
Explosive strength of the upper and lower limbs will be assessed by analysing the rate of force development derived from the force-time curve obtained during maximal voluntary isometric contractions performed on a bench press machine (upper limbs) and a horizontal leg press (lower limbs). The rate of force development will be calculated as the slope of the force-time curve (Δforce/Δtime) and expressed in newtons per second (N/s). Peak rate of force development will be identified as the maximum slope of the force-time curve using a 20-millisecond sliding window. Sequential rate of force development will be computed over consecutive 50-millisecond intervals from contraction onset (0 ms) up to 250 ms. Values at 50, 100, 150, 200, and 250 ms will be reported in both absolute terms and normalised to maximal voluntary isometric contraction force, as previously described.
Time frame: 4 weeks
Endurance Strength
Muscular endurance will be assessed for both upper and lower limbs using a bench press machine and a horizontal leg press, respectively. Participants will perform a sustained isometric contraction at 40% of their maximal voluntary isometric contraction force until task failure. Task failure will be defined as a force decline exceeding 10% below the target level (40% of maximal voluntary isometric contraction) sustained for more than 10 seconds. Standardised verbal encouragement and real-time visual feedback will be provided throughout the test.
Time frame: 4 weeks
Handgrip strength
Handgrip strength will be measured as the maximal voluntary isometric contraction of the hand and forearm muscles using a portable hand dynamometer (JAMAR Plus+, Patterson Medical, USA). Participants will be assessed in a standing position with the arm in a neutral position (midway between supination and pronation) and the elbow fully extended alongside the body. Prior to testing, the dynamometer handle will be adjusted to each participant's hand size. Measurements will be performed alternately on both hands until three valid attempts are completed for each hand. For each attempt, the participant will exert maximal grip force for 5 seconds, followed by a 60-second rest interval.
Time frame: 4 weeks
Menstrual cycle determination
For both groups (naturally menstruating athletes and oral contraceptive users), cycle phase identification will initially be based on calendar tracking of menstrual cycle phases, and oral contraceptive cycle weeks. To confirm the different cycle phases, serum estrogen and progesterone concentrations will be measured throughout the study. Blood samples will be collected via venipuncture, and serum samples will be stored at -80°C until analysis by electrochemiluminescence immunoassay (Cobas e411; Roche Diagnostics, Switzerland). Additionally, in the naturally menstruating group only, urinary ovulation tests (ClearBlue®, Switzerland) will be used to detect the luteinizing hormone (LH) surge, beginning on day 10 of the menstrual cycle (or four days before the predicted ovulation date).
Time frame: 4 weeks
Total Body Water
Total body water will be measured by deuterium dilution using a Hydra stable isotope ratio mass spectrometer (PDZ Europa, Crewe, UK). Following a 12-hour overnight fast, a baseline urine sample will be collected. Each participant will then ingest an oral dose of 0.1 g/kg body mass of 99.9% deuterium oxide (²H₂O; Sigma-Aldrich, St. Louis, USA). Post-dose urine samples will be collected at the 4- and 5-hour time points during an equilibration period in which no food or fluid intake will be permitted. Urine and diluted dose samples will be prepared for ²H/¹H isotope ratio analysis.
Time frame: 4 weeks
Extracellular Water
Extracellular water will be determined by sodium bromide dilution. Following collection of a baseline saliva sample, each participant will ingest 0.030 g/kg body mass of 99.0% sodium bromide (NaBr; Sigma-Aldrich, St. Louis, USA) diluted in 50 mL of distilled deionised water. After a 3- to 4-hour equilibration period, during which no food or fluid intake will be permitted, a post-dose saliva sample will be collected. All saliva samples will be collected using Salivette collection devices (Sarstedt, Nümbrecht, Germany), then centrifuged and stored at -20°C until analysis.
Time frame: 4 weeks
Intracellular Water
Intracellular water will be calculated as the difference between total body water and extracellular water (ICW = TBW - ECW).
Time frame: 4 weeks
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.
Water Balance/Turnover
Water turnover (rH₂O) and water intake will be assessed using the doubly labelled water method. Total body water will be derived from the dilution spaces of deuterium (²H) and oxygen-18 (¹⁸O) measured in urine samples, with correction factors applied for non-aqueous hydrogen exchange (1.041 for ²H and 1.007 for ¹⁸O). Water turnover will be calculated by multiplying the ²H dilution space by its elimination rate constant. Assuming steady-state conditions, water influx will be considered equivalent to water efflux. Water intake will then be estimated by subtracting metabolic water production and respiratory and transcutaneous water gains from total water influx.
Time frame: 4 weeks
Hydration status: Plasma osmolality
Plasma osmolality will be measured using a freezing-point osmometer (OSMO1, Advanced Instruments, Norwood, USA) and expressed in mOsm/kg.
Time frame: 4 weeks
Hydration status: Urine osmolality
Urine osmolality (mOsm/kg) will be assessed using an osmometer (Model OSMO1, Advanced Instruments, Canada). Urine osmolality (Uosm) is defined as the number of osmotically active particles per kilogram of water in urine. This parameter reflects the kidney's ability to regulate water balance through concentration and dilution mechanisms over a 24-hour period. Participants will be provided with a sterile container for urine collection and storage.
Time frame: 4 weeks
Hydration status: Saliva osmolality
Saliva osmolality will be measured using a freezing-point depression osmometer (OSMO1, Advanced Instruments, Norwood, USA) and expressed in mOsm/kg. Saliva samples will be collected using Salivette (Sarstedt, Nümbrecht, Germany). Participants will sit quietly for 2 minutes, allowing saliva to accumulate passively in the mouth, then place the swab under the tongue and gently roll it until fully saturated. Participants will be instructed to avoid touching the swab with their fingers throughout the procedure.
Time frame: 4 weeks
Hydration status: Urine Specific Gravity
Urine-specific gravity will be measured in a fasted baseline urine sample using a digital refractometer (PAL-10S, Atago, Tokyo, Japan).
Time frame: 4 weeks
Fat mass
Fat mass, a molecular-level component, will be calculated using a reference four-compartment model: \[FM (kg) = 2.748 × BV - 0.699 × TBW + 1.129 × Mo - 2.051 × BM\] where BV is body volume (L) obtained by air displacement plethysmography (described below), TBW is total body water (kg) assessed by dilution techniques (described above), Mo is bone mineral mass (kg) obtained by dual-energy X-ray absorptiometry (described below), and BM is body mass (kg). In this model, soft tissue minerals (Ms), a small molecular-level component, are estimated as Ms = 0.00129 × TBW.
Time frame: 4 weeks
Body Volume
Body volume will be measured by air displacement plethysmography (BOD POD, COSMED, Concord, USA). Participants will wear a swimsuit and a swimming cap to minimise air trapped in clothing and hair. Body mass will be measured to the nearest 0.1 kg using the electronic scale integrated with the plethysmograph system. Body volume will be computed from the initial measurement corrected for thoracic gas volume, which will be measured directly in all participants, and for body surface area artefact.
Time frame: 4 weeks
Bone Mineral Content
Bone mineral content (BMC) will be estimated using dual-energy X-ray absorptiometry (DXA; MA, USA). The technique is based on the differential attenuation of X-ray beams at two energy levels (70 and 140 kV), synchronized with the line frequency, for each pixel of the scanned image. All analyses will be performed by a trained laboratory technician in accordance with the manufacturer's operator manual, using the standard analysis protocol. Since BMC represents ash bone mass, BMC values will be converted to bone mineral mass (Mo) by multiplying them by a factor of 1.0436.
Time frame: 4 weeks
Fat Free Mass
After the determination of fat mass (FM), fat-free mass (FFM) is calculated by subtracting FM from total body mass.
Time frame: 4 weeks
3DO Application
In addition to the four-compartment model described above, body composition will also be assessed using the 3DO application (Prism Labs, USA), a three-dimensional body reconstruction system based on a mobile platform that enables the estimation of body fat percentage, fat mass, fat-free mass, body volume, and multiple body circumferences.
Time frame: 4 weeks
Bioelectrical Impedance Analysis: Resistance
Whole-body and segmental bioelectrical impedance will be assessed using the AKERN BIA 101/BIVA PRO, a phase-sensitive, single-frequency bioelectrical impedance analysis (BIA) device used to measure resistance (R) values.
Time frame: 4 weeks
Bioelectrical Impedance Analysis: Reactance
Whole-body and segmental bioelectrical impedance will be assessed using the AKERN BIA 101/BIVA PRO, a phase-sensitive, single-frequency bioelectrical impedance analysis (BIA) device used to measure reactance (Xc).
Time frame: 4 weeks
Bioelectrical Impedance Analysis: Phase Angle
Whole-body and segmental bioelectrical impedance measurements will be performed using the AKERN BIA 101/BIVA PRO, a single-frequency, phase-sensitive bioelectrical impedance analysis (BIA) device that provides phase angle (PhA) values.
Time frame: 4 weeks
Resting Energy Expenditure
Resting energy expenditure (REE) will be assessed by indirect calorimetry under standardized conditions. A breath-by-breath metabolic cart (QUARK RMR, Cosmed, Italy), equipped with a mask system, will be used to continuously measure oxygen consumption (V̇O₂) and carbon dioxide production (V̇CO₂). The total testing duration will be 45 minutes, including an initial 15-minute rest period followed by 30 minutes of gas exchange measurement. The first and last 5 minutes of data collection will be excluded from the analysis. REE will be determined from the mean V̇O₂ and V̇CO₂ obtained during a 5-minute steady-state interval occurring between minutes 5 and 25 of the measurement period. Steady state will be defined by a coefficient of variation ≤10% for both V̇O₂ and V̇CO₂ and a respiratory exchange ratio (RER) between 0.70 and 1.00. The Weir equation will be applied to calculate REE.
Time frame: 4 weeks
Total Energy Expenditure
Total energy expenditure (TEE) will be assessed using the doubly labeled water (DLW) method. At baseline, participants will provide a 10 mL urine sample after an overnight fast. Following baseline urine collection, participants will ingest a dose of water containing 1.8 g/kg of predicted total body water (TBW) enriched with 10% H₂¹⁸O (Taiyo Nippon Sanso Corporation, Japan) and 0.16 g/kg of predicted TBW enriched with 99.9% ²H₂O (Sigma-Aldrich, USA), diluted in 50 mL of tap water. Predicted TBW will be estimated as 61% of body mass. Additional urine samples will be collected at 4 and 5 hours post-dose, and again after 7 days (first morning urine sample and one hour later). Urine samples will be analyzed in triplicate using isotope ratio mass spectrometry. Carbon dioxide production will be calculated using the two-point method, and TEE will be determined using the modified Weir equation.
Time frame: 4 weeks
Energy intake
Energy intake (EI) will be estimated using the intake-balance method, which accounts for changes in body energy stores, specifically fat mass (FM) and fat-free mass (FFM), as well as total energy expenditure (TEE). Energy intake will be calculated using the following equation: Energy Intake (kcal/day) = TEE + \[1020 \* (ΔFFM / Δtime) + 9500 \* (ΔFM / Δtime)\]
Time frame: 4 weeks
Energy availability
Energy availability (EA) is determined by subtracting exercise energy expenditure (EEE) from total energy intake (EI) and normalizing this value to fat-free mass (FFM). Energy availability will be calculated using the following equation: \[Energy availability (kcal/kg of FFM) = (Energy Intake - Exercise Energy Expenditure) / FFM\]; Exercise energy expenditure will be measured via accelerometry. A resting energy expenditure ratio (REE ratio) below 0.90 will be considered indicative of energy deficiency and low energy availability.
Time frame: 4 weeks
Dietary intake
To estimate dietary intake, three non-consecutive food records (including one weekend day) will be completed under the supervision of a trained nutritionist. Prior to completing the records, participants will receive written instructions, including specific guidelines, portion size images, and examples of common recording errors. Dietary data will be analyzed using Food Processor Plus® software (ESHA Research, USA).
Time frame: 4 weeks
Physical activity
Physical activity will be assessed using an ActiGraph GT9X accelerometer worn on the non-dominant wrist for a minimum of five consecutive days during each menstrual cycle phase or oral contraceptive pill cycle week. Data will be processed using ActiLife and GGIR software. Outcome measures will include time spent at different physical activity intensity levels and mean daily physical activity.
Time frame: 4 weeks
Physical Pain
Physical pain will be assessed using a Visual Analogue Scale (0-8), in which 0 represents the lowest level of pain and 8 represents the highest level.
Time frame: 4 weeks
Motivation to perform strength testing
Motivation to perform the strength test will be assessed using a Visual Analogue Scale (0-8), in which 0 represents the lowest level of motivation and 8 represents the highest level.
Time frame: 4 weeks
Perception of performance
Perceived performance assessed using a Visual Analogue Scale (0-8), where 0 represents the lowest and 8 the highest level of perceived performance.
Time frame: 4 weeks
Affective responses
Affective responses (i.e., pleasure and arousal levels) were assessed using the Affect Grid, a single-item, two-dimensional measure with anchors ranging from 1 to 9 on the pleasure to displeasure and arousal to sleepiness dimensions. The Affect Grid was designed as a rapid method for assessing affect along these two orthogonal dimensions. Participants were instructed to place a single cross (X) in the square on the grid that best represented their current affective state.
Time frame: 4 weeks