Phase Angle and Extracellular Water‐to‐Total Body Water Ratio Measured by Bioelectrical Impedance Analysis Are Associated with Anemia and Hemoglobin Levels in Women with Type 2 Diabetes: A Cross‐Sectional Study Authors: Niloufar Abdollahpour, Najmeh Seifi, Ali Mottaghi Moghaddam Shahri, Habibollah Esmaily, Gordon A Ferns, Majid Ghayour‐Mobarhan Published: 2026 ABSTRACT ABSTRACTBackground and AimsAnemia is common in women with diabetes, increasing disability and mortality risks. Bioelectrical impedance analysis (BIA) measures phase angle (PhA), reflecting cellular integrity and muscle quality, and the extracellular water to total body water ratio (ECW/TBW), indicating fluid balance and inflammation. Both are associated with malnutrition and chronic inflammation, key factors in anemia. This study explored their association with anemia in diabetic women.MethodsThe current study is a cross‐sectional study that was done in Phase II of the Mashhad stroke and heart atherosclerotic disorder (MASHAD) study (2018–2020). We evaluated 1381 women with diabetes, aged 35–65. PhA and ECW/TBW were assessed using BIA (InBody S10 Biospace). To assess the relationships between PhA, ECW/TBW and anemia, multiple logistic regression was performed. Linear regression was used to evaluate the associations of PhA and ECW/TBW with hemoglobin (Hb) and hematocrit (Hct) levels.ResultsIn the fully adjusted logistic models, a higher PhA and lower ECW/TBW were associated with a lower risk of anemia (OR: 0.55; 95% CI: 0.33, 0.94 and OR: 1.86; 95% CI: 1.08, 3.19, respectively). Furthermore, in linear regression analysis, PhA exhibited a positive correlation with Hb (β = 0.20; p < 0.001) and Hct levels (β = 0.34; p = 0.03). ECW/TBW was negatively correlated with Hb (β = −16.96; p < 0.001) and showed a non‐significant negative correlation with Hct (β = −19.39; p = 0.09).ConclusionsPhA is associated with lower odds of anemia and higher Hb and Hct levels in women with diabetes, while a higher ECW/TBW is linked to an increased risk of anemia and lower Hb levels. These findings underscore the novel clinical relevance of PhA and ECW/TBW in identifying anemia risk in diabetic women, highlighting their potential as associative markers for early detection and intervention.Keywords: anemia, bioelectrical impedance analysis, diabetes, extracellular water‐to‐total body water ratio, hematocrit, hemoglobin, phase angle Introduction Diabetes mellitus (DM) is a prevalent metabolic disease and is also a major cause of mortality worldwide [1]. The main feature of this condition is hyperglycemia, which is the product of a deficiency in insulin action, secretion, or both [2]. At present, diabetes affects over 415 million individuals across the globe [3]. This figure is estimated to reach 642 million by 2040 [4]. In Iran, nearly 5 million people have been reported to have diabetes [5]. Hyperglycemia needs to be treated; otherwise, severe micro‐ and macro‐vascular complications are expected to arise [6]. A systematic review and meta‐analysis study conducted in 2024, which included 26,485 people with diabetes, reported an overall anemia prevalence of 35.45% [7]. Notably, a recent study focused on adults with Type 2 diabetes mellitus (T2DM) found that women with diabetes experienced anemia at a higher rate, with a prevalence of 38.5% [8]. The higher prevalence of anemia among individuals can likely be attributed to several factors. Chronic inflammation plays a crucial role by impairing erythropoiesis and reducing red blood cell lifespan through increased cytokine production and bone marrow suppression [9]. Additional factors that may contribute to this condition are deficiencies of iron, folate, and vitamin B12, medication side effects, reduced erythropoietin (EPO) production, absorption disorders, and oxidative stress [10]. In addition, diabetic nephropathy is known as one of the most significant microvascular complications leading to a deficiency of EPO and anemia [11, 12]. This, in turn, highlights the urgent need for the assessment of anemia risk among patients with diabetes, specifically in the early stages of the disease. In recent years, bioimpedance analysis (BIA) has gained prominence for assessing body composition and cellular health on the basis of electrophysiological characteristics of the human organism [13]. BIA provides crucial data on muscle and fat mass, total protein and mineral ratios, and hydration status [14]. In addition, phase angle (PhA), representing the phase difference between electrical voltage and current waveforms, has become a key BIA parameter in clinical studies. PhA is used as a marker of cell mass, as well as the integrity of cells and the functional status of cell membranes [15]. PhA, a predictive indicator, offers valuable insights into various diseases, including sepsis, malignancies, cirrhosis, and kidney diseases [16, 17, 18]. PhA can be used as a medically helpful index for general well‐being and also for predicting disease progression. A fall in this index is related to both indicatory evidence of illness and advancing age, which leaves older individuals with a diminished PhA exposed to an elevated risk of mortality [19]. Through the analysis of cell membrane resistance against frequency currents, BIA is also capable of assessing extracellular water (ECW), intracellular water (ICW), and total body water (TBW). Indicating cellular fluid volume equilibrium, the percentage resulting from the proportion of ECW over TBW presents itself as a winning preclinical marker of disease [20]. Recent studies suggest that PhA is a promising marker for anemia risk in chronic diseases [21, 22]. In diabetes, lower PhA and higher ECW/TBW ratio are linked to reduced hemoglobin (Hb) and hematocrit (Hct) [21]. In hemodialysis patients, low PhA is associated with poor response to erythropoiesis‐stimulating agents and iron supplements [22]. PhA and ECW/TBW serve as noninvasive methods to study and potentially detect the catabolic state and disease development in persons with DM [23, 24]. Moreover, it has been demonstrated that blood conductivity is significantly influenced by the cellular composition of blood [25]. Given that bioelectrical impedance parameters reflect cellular integrity and hydration status, they may provide indirect insights into hematological alterations, including anemia [21]. Considering these factors and previous studies, we were led to the proposal that PhA and ECW/TBW may play pivotal roles in the interplay with anemia [21]. Uncertainties still abound as regards the relationship between these factors in the case of women, especially those affected by diabetes‐related coexisting conditions. In addition, studying anemia in this demographic group is important because of their higher risk from menstrual cycles, pregnancy, and childbirth. Hormonal differences and increased cardiovascular complications make the management of the case more complex [8]. Anemia management in diabetic women could benefit from noninvasive assessment tools, such as PhA and ECW/TBW, which might help in evaluating tissue hydration, nutritional status, and overall cellular health—factors that are closely linked to anemia progression and severity [21]. Anemia significantly impacts women's daily lives due to their multiple roles in society and family, and they may respond in different ways to treatments compared to men [26]. For diabetic women, anemia may further complicate glycemic control, exacerbate fatigue, and increase the likelihood of diabetes‐related complications, ultimately affecting their quality of life and disease prognosis [27]. Given the urgent need to better understand anemia in women with diabetes, this study aims to investigate the associations between PhA, ECW/TBW ratio, and anemia‐related parameters (Hb, Hct). By evaluating these relationships, we seek to clarify their clinical relevance and potential role as non‐invasive markers in this population. Materials and Methods Study Design and Participants This cross‐sectional investigation included 1381 women with type 2 diabetes. They were selected from the second phase of the Mashhad stroke and heart atherosclerotic disorder (MASHAD) study (2018–2020). These women were part of the diabetic subgroup within the larger MASHAD cohort and were selected based on the presence of diabetes. Further inclusion was determined by meeting the study's eligibility criteria, such as having complete data on body composition (especially PhA and ECW/TBW) and laboratory assessments (Hb and Hct). The MASHAD study itself commenced in 2008–2010, and tracked 9704 participants aged 35–65 years. These participants were recruited using a cluster‐randomized method and followed up over a period of 10 years [28]. This study was prepared and presented in line with the STROBE reporting guidelines (Table S1). The following criteria were applied to exclude women diagnosed with diabetes from the study: (1) individuals who did not have complete data about body composition (especially PhA and ECW/TBW) or laboratory assessment (Hb and Hct); (2) pregnant or breastfeeding women, as pregnancy‐related physiological changes can influence body composition and hematological markers; (3) patients suffering from severe renal impairment characterized by a serum creatinine (Cr) level greater than 2 mg/dL, since impaired kidney function affects fluid balance and body composition measurements; (4) individuals with cirrhosis or malnutrition (serum albumin (Alb) level below 3 g/dL), as these conditions alter fluid distribution and body composition parameters; and (5) those on iron supplements, because iron therapy can affect hematological markers (Hb and Hct), potentially confounding the study results. The flow diagram of study participants is presented in Figure 1. Flow diagram of study subject subgroups analyses. All individuals, before being included in this study, signed informed consent, and all other procedures were approved by the Human Research Ethics Committee of Mashhad University of Medical Sciences under the ethics approval number IR.MUMS.MEDICAL.REC.1395.458. Bioelectrical Impedance Analysis Body composition assessment was performed by using InBody S10 (Biospace Co. Ltd., Korea; Model JMW140). Measurements taken after overnight fasting, and when performed strictly according to the manufacturer's instructions. Participants were asked to follow standard pre‐assessment conditions for improved accuracy. These included avoiding over‐ or under‐hydration, avoiding alcohol, caffeine, and smoking, and fasting for at least 4 h. They also had to void urine before the test and refrain from any strenuous exercise for at least 12 h before the measurement [29]. At first, participants were asked to remove any metal objects, like watches, rings, bracelets, earrings, etc. After that, they were instructed to lie relaxed on a couch with electrodes installed on both ends of their limbs. Electrodes were attached to both middle fingers and thumbs of the hands and between the ankle bone and heel of both feet, following the manufacturer's guidelines. BIA used six frequencies (1, 5, 50, 250, 500, and 1000 kHz) to enhance accuracy. Low frequencies (1, 5 kHz) assess ECW, while high frequencies (250, 500, 1000 kHz) penetrate cell membranes to measure ICW and TBW. The 50 kHz frequency is standard due to its balance of penetration and reliability. Bioelectrical impedance analysis (BIA) evaluates various parameters, including ICW, ECW, TBW, Body Fat Mass (BFM), Free Fat Mass (FFM), and Skeletal Muscle Mass (SMM). In addition to these parameters, whole body resistance and reactance were evaluated with an electrical current of 50 kHz, and phase angle in degrees calculated according to the following equation: phase angle (°) = arctangent (Xc/R) ∗ (180/π), where resistance (R) represents the opposition to electrical current flow through bodily fluids, mainly influenced by water content, and reactance (Xc) reflects the cell membranes' ability to store and release electrical energy. Both of these parameters are measured in Ohms. Laboratory Assessment Following the study protocol, venous blood samples were obtained from all participants between 8:00 a.m. and 10:00 a.m. after they had completed at least a 14‐h overnight fast [28]. Blood samples were collected and used to determine blood cell counts, fasting blood glucose (FBG), and serum biochemical parameters, including urea, creatinine, aspartate aminotransferase (AST), and alanine transaminase (ALT), high‐sensitivity C‐reactive protein (hs‐CRP). According to the study protocol, individuals with type 2 diabetes were defined as participants whose FBG was over 126 mg/dL or who were under treatment by hypoglycemic agents. In accordance with WHO recommendations, anemia was defined as an individual Hb level lower than 12 g/dL [8]. In this study, estimated glomerular filtration rate (eGFR) was calculated using the Cockcroft–Gault equation, adjusted for body surface area (BSA), as follows: (multiplied by 0.85 for females) BSA was calculated using the Du Bois formula: This adjustment to 1.73 m2 allows for standardization of kidney function estimates across individuals of different body sizes. Covariates Participants' data about demographic characteristics, medical history, and lifestyle factors like smoking, marital status, etc., were collected by healthcare professionals. Data on menstrual status were obtained from patient self‐reports. To evaluate the activity levels of study participants, a set of questions from the Scottish Heart Health Study (SHHS)/MONICA questionnaire was used. This questionnaire, which has been previously validated, was employed to measure physical activity level (PAL) by assessing individuals' activity during work, non‐work, and rest periods [30, 31]. The energy intake of each participant was estimated by utilizing a previously validated 65‐item food frequency questionnaire [28]. In addition to these lifestyle factors, all subjects completed a questionnaire on cardiovascular disease risk and a resting 12‐lead ECG and complementary test, if required, to evaluate cardiovascular status. A person was classified as having hypertension (HTN) if SBP was 140 mmHg or higher, and/or DBP was 90 mmHg or higher, or if taking anti‐hypertension medication. Statistical Analysis Participants were categorized into tertiles according to whole‐body phase angle values to facilitate clinically interpretable comparisons across low, intermediate, and high PhA groups while maintaining balanced sample sizes within each category. Statistical analyses were performed using SPSS software version 20.0 (SPSS Inc., Chicago, IL, USA). All statistical tests were two‐sided. A priori statistical significance was defined as p < 0.05. The distribution of continuous variables was evaluated using the Kolmogorov–Smirnov test, and homogeneity of variances was assessed using Levene's test. Continuous variables are presented as mean ± standard deviation (SD), whereas categorical variables are expressed as frequency (percentage). Differences across tertiles of phase angle were evaluated using one‐way analysis of variance (ANOVA) for continuous variables, followed by Bonferroni post hoc comparisons for pairwise comparisons. The χ 2 test was used for categorical variables. Multiple logistic regression analyses were conducted to examine the associations of phase angle and ECW/TBW tertiles with anemia status, and findings are reported as odds ratios (ORs) with 95% confidence intervals (95% CIs). Multiple linear regression analyses were additionally performed to evaluate the associations of phase angle and ECW/TBW with hemoglobin and hematocrit concentrations, and results are presented as beta coefficients (β) with corresponding 95% CIs. Four models were designed in this study; the first model was adjusted for age, BMI, and PAL. The second model was adjusted based on the menstrual state as well as other adjustments in model one, and the third model, in addition to previous adjustments, was also adjusted for energy intake, CVD status, and hypertension. In the fourth model, hs‐CRP and eGFR were also adjusted, along with all the variables from the previous models. Exploratory sensitivity analysis was conducted by classifying individuals based on several subgroups, including kidney function (eGFR ≥ 60 vs. eGFR < 60), menstruation status (yes vs. no), age (≥ 60 vs. < 60 years), BMI (≥ 30 vs. < 30 kg/m2), hypertension status (yes vs. no), and hs‐CRP levels (> 3 vs. ≤ 3 mg/L). Results A total of 1381 women with type 2 diabetes, with a mean age of 59.35 ± 7.76 years, were included in the present study. Overall, 9.7% (n = 134) of study participants met the criteria for anemia. The mean BMI, ECW/TBW, and whole body PhA were 30.52 ± 7.35 kg/m2, 0.38 ± 0.01, and 5.26 ± 0.67° respectively. The tertiles of 50 kHz‐whole body PhA were used to classify subjects into three PhA categories. Physical and laboratory characteristics of participants over PhA categories are presented in Table 1. Table 1Clinical characteristics of the study participants by categories of whole body phase angle.Whole body phase angleVariableTotal (n = 1381)T1 (n = 433) ≤ 4.9T2 (n = 520) 5–5.5T3 (n = 428) ≥ 5.6 p value p trendDemographic dataAge (year)59.35 ± 7.7662.89 ± 7.19a 59.55 ± 7.12b 55.52 ± 7.30c < 0.001< 0.001Education, n (%)Diploma or under1192 (93.49%)409 (94.90%)419 (92.29%)364 (93.33%)0.29—University educated83 (6.51%)22 (5.10%)35 (7.71%)26 (6.66%)Marriage, n (%)Married1081 (78.40%)293 (67.80%)424 (81.70%)364 (85.20%)< 0.001—Smoking, n (%)Current‐smoker211 (15.35%)75 (17.36%)73 (14.07%)63 (14.86%)0.03—Non‐smoker1046 (76.07%)313 (72.45%)400 (77.07%)333 (78.54%)Past‐smoker118 (8.58%)44 (10.19%)46 (8.86%)28 (6.60%)Menstrual status, n (%)Yes637 (46.60%)167 (38.93%)238 (46.21%)232 (54.85%)< 0.001—No730 (53.40%)262 (61.07%)277 (53.79%)191 (45.15%)Weight (kg)72.12 ± 12.6169.95 ± 12.69a 72.09 ± 12.78a 74.36 ± 11.94a < 0.001< 0.001BMI (kg/m2)30.52 ± 7.3529.98 ± 8.77a 30.22 ± 5.00a 31.42 ± 8.08a 0.008< 0.001HTN, n (%)665 (48.30%)228 (52.90%)263 (50.80%)174 (40.70%)< 0.001—CVD, n (%)220 (19.10%)98 (27.10%)83 (18.90%)39 (11.20%)< 0.001—Anemia, n (%)134 (9.70%)60 (13.90%)34 (6.50%)40 (9.30%)0.001—PAL1.78 ± 0.341.79 ± 0.31a 1.81 ± 0.37a 1.74 ± 0.33a 0.020.03Energy intake (kcal/day)1857.81 ± 1186.741860.54 ± 1310.40a 1851.69 ± 1200.16a 1861.81 ± 1015.38a 0.990.99Bioelectrical impedance analysisBFM (kg)29.90 ± 9.0529.37 ± 9.08a 29.97 ± 9.32a 30.35 ± 8.67a 0.270.11SMM (kg)22.91 ± 3.4821.73 ± 3.55a 22.80 ± 3.03b 24.24 ± 3.47c < 0.001< 0.001FFM (kg)42.12 ± 5.6040.50 ± 5.50a 41.97 ± 5.16b 43.92 ± 5.68c < 0.001< 0.001ECW/TBW0.38 ± 0.010.39 ± 0.01a 0.38 ± 0.00b 0.37 ± 0.01c < 0.001< 0.00150kHz‐RA phase angle (°)5.15 ± 0.634.58 ± 0.40a 5.15 ± 0.33b 5.73 ± 0.54c < 0.001< 0.00150kHz‐LA phase angle (°)4.87 ± 0.564.37 ± 0.39a 4.87 ± 0.37b 5.38 ± 0.44c < 0.001< 0.00150kHz‐TR phase angle (°)7.16 ± 1.776.49 ± 1.41a 7.20 ± 1.73b 7.77 ± 1.90c < 0.001< 0.00150kHz‐RL phase angle (°)5.29 ± 1.054.37 ± 0.62a 5.28 ± 0.51b 6.26 ± 1.03c < 0.001< 0.00150kHz‐LL phase angle (°)5.17 ± 0.964.36 ± 0.65a 5.17 ± 0.63b 5.98 ± 0.88c < 0.001< 0.00150kHz‐whole body phase angle (°)5.26 ± 0.674.55 ± 0.34a 5.25 ± 0.16b 6.00 ± 0.47c < 0.001< 0.001Laboratory assessmentHb (g/dL)13.65 ± 1.2413.42 ± 1.25a 13.72 ± 1.15ab 13.79 ± 1.30b < 0.001< 0.001Hct (%)40.08 ± 3.2639.60 ± 3.36a 40.27 ± 3.15a 40.32 ± 3.25a 0.0010.001MCV (fL)84.89 ± 5.6485.26 ± 6.01a 84.95 ± 5.33a 84.45 ± 5.61a 0.100.04MCH (pg)28.92 ± 2.3328.90 ± 2.37a 28.95 ± 2.14a 28.90 ± 2.49a 0.920.98MCHC (g/dL)34.06 ± 1.5333.89 ± 1.49a 34.08 ± 1.39ab 34.20 ± 1.72b 0.0120.003RDW (fL)13.17 ± 1.0613.27 ± 1.16a 13.10 ± 0.89a 13.15 ± 1.11a 0.040.08Urea (mg/dL)32.32 ± 9.7434.44 ± 10.88a 32.19 ± 9.32ab 30.35 ± 8.55b < 0.001< 0.001Cr (mg/dL)1.03 ± 0.201.06 ± 0.23a 1.01 ± 0.19a 1.01 ± 0.18a < 0.0010.001AST (U/L)21.49 ± 9.7020.83 ± 9.89a 21.30 ± 9.60a 22.37 ± 9.59a 0.060.02ALT (U/L)19.85 ± 12.8317.46 ± 10.81a 19.99 ± 12.82a 22.09 ± 14.25b < 0.001< 0.001FBG (mg/dL)157.57 ± 63.63166.34 ± 72.23a 156.36 ± 61.52b 150.19 ± 55.47b < 0.001< 0.001eGFR (mL/min/1.73 m2)69.82 ± 21.5368.52 ± 19.75a 68.88 ± 20.80a 72.36 ± 23.97b 0.001< 0.001 Note: Continuous variables are expressed as mean ± SD, while categorical variables are represented as number (percent). ANOVA was utilized for analyzing continuous variables followed by Bonferroni post‐hoc tests. Means with different superscript letters (a,b,c) are significantly different from one another (p < 0.05), whereas means sharing at least one common letter are not significantly different. The χ 2 test was applied for categorical variables. p value < 0.05 were considered as meaningful.Abbreviations: ALT, alanine transaminase; AST, aspartate aminotransferase; BFM, body fat mass; BMI, body mass index; Cr, creatinine; CVD, cardiovascular disease; ECW, extra cellular water; eGFR, estimated glomerular filtration rate; FBG, fasting blood glucose; FFM, fat free mass; Hb, hemoglobin; Hct, hematocrit; HTN, hypertension; LA, left arm; LL, left leg; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; PAL, physical activity level; RA, right arm; RL, right leg; RDW, red cell distribution width; SMM, skeletal muscle mass; TBW, total body water; TR, trunk. Clinical characteristics of the study participants by categories of whole body phase angle. Note: Continuous variables are expressed as mean ± SD, while categorical variables are represented as number (percent). ANOVA was utilized for analyzing continuous variables followed by Bonferroni post‐hoc tests. Means with different superscript letters (a,b,c) are significantly different from one another (p < 0.05), whereas means sharing at least one common letter are not significantly different. The χ 2 test was applied for categorical variables. p value < 0.05 were considered as meaningful. Abbreviations: ALT, alanine transaminase; AST, aspartate aminotransferase; BFM, body fat mass; BMI, body mass index; Cr, creatinine; CVD, cardiovascular disease; ECW, extra cellular water; eGFR, estimated glomerular filtration rate; FBG, fasting blood glucose; FFM, fat free mass; Hb, hemoglobin; Hct, hematocrit; HTN, hypertension; LA, left arm; LL, left leg; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; PAL, physical activity level; RA, right arm; RL, right leg; RDW, red cell distribution width; SMM, skeletal muscle mass; TBW, total body water; TR, trunk. Compared with participants in the lowest PhA tertile, those in the highest tertile were younger individuals (55.52 vs. 62.89, p < 0.001), having higher BMI (31.42 vs. 29.98, p = 0.008), and lower physical activity levels (1.74 vs. 1.79, p = 0.02). There was significant difference in smoking status (p = 0.03), marital status (p < 0.001) and history of having HTN (p < 0.001) and CVD (p < 0.001), across tertiles of PhA. SMM, FFM, whole body phase angle and all its segmental values, Hb, and ALT differed significantly across tertiles (all, p < 0.001), with a significant increasing trend (all, p‐trend < 0.001). There was a statistically significant difference between the groups in terms of ECW/TBW, Hct, urea, Cr, FBG, and eGFR (all, p < 0.05). Specifically, a decreasing trend was observed as participants progressed from the first to the third tertile (all, p‐trend < 0.05). The associations between PhA and ECW/TBW with anemia are shown in Table 2. In multivariable logistic regression analyses, higher whole‐body phase angle values were consistently associated with lower odds of anemia across all adjustment models. In the fully adjusted model, participants in PhA tertile 3, the highest, were 45% less likely to develop anemia in comparison with those placed in tertile 1, the lowest. (OR: 0.55; 95% CI: 0.33, 0.94). Regarding ECW/TBW, no statistically significant association with anemia was observed in the participants positioned in the unadjusted model (OR: 1.44; 95% CI: 0.94, 2.21). However, after adjustment for confounding variables, higher ECW/TBW values demonstrated higher chances of anemia in all multivariable models (all, p < 0.05). A significant trend across tertiles was observed for both phase angle and ECW/TBW in relation to anemia across all models (all, p‐trend < 0.05). Table 2Multiple logistic regression on the association between Phase angle, ECW/TBW, and anemia.VariablePhase angle (°)ECW/TBWT1 (n = 433)T2 (n = 520)T3 (n = 428) p‐trendT1 (n = 464)T2 (n = 458)T3 (n = 461) p‐trendAnemiaCrude10.44 (0.28, 0.68)0.64 (0.42, 0.98)0.0310.99 (0.62, 1.57)1.44 (0.94, 2.21)0.09Model 110.39 (0.25, 0.62)0.51 (0.32, 0.81)0.00311.10 (0.69, 1.77)1.77 (1.10, 2.83)0.03Model 210.40 (0.25, 0.63)0.51 (0.32, 0.82)0.00311.10 (0.69, 1.76)1.75 (1.09, 2.81)0.02Model 310.38 (0.23, 0.63)0.58 (0.35, 0.97)0.0211.07 (0.63, 1.83)1.80 (1.07, 3.05)0.03Model 410.37 (0.22, 0.61)0.55 (0.33, 0.94)0.0211.14 (0.67, 1.96)1.86 (1.08, 3.19)0.02 Note: Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using multiple logistic regression. p value < 0.05 were considered as meaningful.Abbreviations: Crude, unadjusted; ECW, extracellular water; Model1, adjusted for age, BMI, and PAL; Model2, Model1 + menstruation status; Model3, Model2 + energy intake, CVD status, hypertension; Model 4, Model 3 + hs‐CRP and eGFR; TBW, total body water. Multiple logistic regression on the association between Phase angle, ECW/TBW, and anemia. Note: Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using multiple logistic regression. p value < 0.05 were considered as meaningful. Abbreviations: Crude, unadjusted; ECW, extracellular water; Model1, adjusted for age, BMI, and PAL; Model2, Model1 + menstruation status; Model3, Model2 + energy intake, CVD status, hypertension; Model 4, Model 3 + hs‐CRP and eGFR; TBW, total body water. Segmental analyses demonstrated that right arm phase angle was inversely associated with anemia in the fully adjusted model (OR: 0.53; 95% CI: 0.32, 0.87; p = 0.01). In addition, elevated ECW/TBW ratios in the left arm (OR: 1.67; 95% CI: 1.01, 2.75; p = 0.04), trunk (OR: 1.93; 95% CI: 1.12, 2.31; p = 0.02), right leg (OR: 1.82; 95% CI: 1.05, 3.15; p = 0.03) and left leg (OR: 2.09; 95% CI: 1.21, 3.60; p = 0.01) positively associated with anemia risk. While no significant association was observed in other segments (p > 0.05 for all) (Table S2). Multiple linear regression analysis for all subjects showed that phase angle was positively correlated with Hb level in all models (fully adjusted β = 0.20; 95% CI: 0.08, 0.31; p < 0.001). Conversely, ECW/TBW demonstrated a significant inverse relationship with hemoglobin levels across all models (fully adjusted β = −16.96; 95% CI: −25.60, −8.32; p < 0.001) (Table 3). Table 3Associations between phase angle, ECW/TBW, and hemoglobin in multiple linear regression analysis.VariablePhase angle (°)Beta coefficientStandard error95% CI of beta p valueHemoglobinCrude0.180.50(0.08, 0.27)< 0.001Model 10.230.05(0.13, 0.34)< 0.001Model 20.220.05(0.12, 0.33)< 0.001Model 30.210.06(0.1, 0.32)< 0.001Model 40.200.06(0.08, 0.31)< 0.001ECW/TBWCrude−14.493.96(−22.25, −6.73)< 0.001Model 1−17.724.23(−26.02, −9.43)< 0.001Model 2−16.464.20(−24.69, −8.23)< 0.001Model 3−16.954.37(−25.52, −8.37)< 0.001Model 4−16.964.40(−25.60, −8.32)< 0.001 Note: Beta coefficients (β), standard errors, and 95% confidence intervals (CIs) were estimated using linear regression. p value < 0.05 were considered as meaningful.Abbreviations: Crude, unadjusted; ECW, extra cellular water; Model1, adjusted for age, BMI, and PAL; Model2, Model1 + menstruation status; Model3, Model2 + energy intake, CVD status, hypertension. Model4, Model 3 + hs‐CRP and eGFR; TBW: total body water. Associations between phase angle, ECW/TBW, and hemoglobin in multiple linear regression analysis. Note: Beta coefficients (β), standard errors, and 95% confidence intervals (CIs) were estimated using linear regression. p value < 0.05 were considered as meaningful. Abbreviations: Crude, unadjusted; ECW, extra cellular water; Model1, adjusted for age, BMI, and PAL; Model2, Model1 + menstruation status; Model3, Model2 + energy intake, CVD status, hypertension. Model4, Model 3 + hs‐CRP and eGFR; TBW: total body water. In line with other analyses, we conducted multiple linear regression with PhA, ECW/TBW, and the Hct level. PhA exhibited a positive association with hematocrit level in all models (fully adjusted β = 0.34; 95% CI: 0.03, 0.64; p = 0.03). Unexpectedly, the association between ECW/TBW and Hct turned out not to be statistically significant, in neither the crude (β = −12.69; 95% CI: −33.25, 7.88 p = 0.23) nor the adjusted analyses (β = −19.39; 95% CI: −42.28, 3.50; p = 0.09) (Table 4). Table 4Associations between phase angle, ECW/TBW, and hematocrit in multiple linear regression analysis.VariablePhase angle (°)Beta coefficientStandard error95% CI of beta p valueHematocritCrude0.290.13(0.03, 0.55)0.03Model 10.430.14(0.15, 0.71)0.003Model 20.420.14(0.14, 0.70)0.003Model 30.380.15(0.08, 0.69)0.01Model 40.340.15(0.03, 0.64)0.03ECW/TBWCrude−12.6910.48(−33.25, 7.88)0.23Model 1−21.3511.20(−43.32, 0.62)0.06Model 2−18.4211.09(−40.18, 3.35)0.10Model 3−18.3911.69(−41.32, 4.55)0.12Model 4−19.3911.66(−42.28, 3.50)0.09 Note: Beta coefficients (β), standard errors, and 95% confidence intervals (CIs) were estimated using linear regression. p value < 0.05 were considered as meaningful.Abbreviations: Crude, unadjusted; ECW, extracellular water; Model1, adjusted for age, BMI, and PAL; Model2, Model1 + menstruation status; Model3, Model2 + energy intake, CVD status, hypertension; TBW, total body water. Associations between phase angle, ECW/TBW, and hematocrit in multiple linear regression analysis. Note: Beta coefficients (β), standard errors, and 95% confidence intervals (CIs) were estimated using linear regression. p value < 0.05 were considered as meaningful. Abbreviations: Crude, unadjusted; ECW, extracellular water; Model1, adjusted for age, BMI, and PAL; Model2, Model1 + menstruation status; Model3, Model2 + energy intake, CVD status, hypertension; TBW, total body water. In the exploratory sensitivity analysis, the inverse association between PhA and anemia remained significant in participants without hypertension and those with hs‐CRP ≤ 3 mg/L. Regarding ECW/TBW, a significant positive association with anemia was observed among participants without hypertension, those with hs‐CRP ≤ 3 mg/L, as well as individuals with BMI < 30 and eGFR < 60 mL/min/1.73 m2. Detailed results are presented in Figures 2, 3. Subgroup sensitivity analysis of the association between phase angle and anemia across clinical and demographic subgroups. Subgroup sensitivity analysis of the association between ECW/TBW ratio and anemia across clinical and demographic subgroups. Discussion In the present study, we explored the association between PhA and ECW/TBW for levels of Hb, Hct, and anemia in a subgroup of women with diabetes. The results of this study found that PhA was negatively associated with anemia, while its connection with hemoglobin and hematocrit levels was positive. As for ECW/TBW, it was positively associated with anemia and negatively associated with hemoglobin levels. Furthermore, a potential negative link was identified between ECW/TBW and Hct levels. The findings of the present study proved to be in accordance with previous findings [21]. Given that PhA is considered an indirect indicator of cellular health and that the ECW/TBW ratio shows systemic fluid retention, this study identified an association between these BIA parameters and anemia. A number of other studies have come up with a close association as regards the relationship between RBC bioimpedance on the one hand and RBC‐associated conditions on the other. Varlet‐Marie et al., for instance, confirmed an association between whole body impedance and agreeability along with deformability of RBCs at 50 kHz [32]. In a comparable observation, Tran et al. reported that when red blood cells undergo hemolysis, the release of cytoplasmic contents is associated with alterations in the electrical properties of the cells [33]. Furthermore, Ringaitiene et al. arrived at the conclusion that RBCs impedance is linked with their lifespan, a process which affects the membrane stability and deformability of RBCs [34]. Lower PhA values signal a compromised integrity of cell membranes and have been suggested to indicate inefficient metabolic function, which may be linked to impaired erythropoiesis. Poor cellular health may be associated with reduced red blood cell production. This can increase the risk of the development of anemia [35]. Although nutritional status was not directly assessed in this study, PhA has been proposed as a surrogate indicator of cellular function, which may reflect aspects of nutritional intake and metabolic efficiency. Proper cellular function may reflect the availability of essential nutrients like iron, vitamin B12, and folate, all of which play a crucial role in Hb and Hct synthesis [36]. Previous studies have indicated that malnutrition is associated with hypoalbuminemia, a condition known to impair erythropoiesis and contribute to hemodilutional anemia [37]. Additionally, studies on ESAs suggest a link between PhA and erythropoietic efficiency. A 2022 study found that higher PhA correlated with reduced ESA resistance in hemodialysis patients, independent of inflammation. While endogenous EPO levels were not measured, PhA may reflect erythropoietic function by indicating better cellular integrity and ESA responsiveness [38]. The body's hydration status also affects PhA. Lower PhA values may be a sign of poor fluid distribution, which may be associated with variations in blood volume and red blood cell concentration [21]. Gomes et al. highlighted the significance of calculating PhA to assess intracellular and extracellular hydration in the case of patients suffering from cancer [39]. Since there are reports which claim that renal anemia is in concordance with hydration in patients suffering from CKD [40], the likelihood is that PhA can be associated with levels of Hb and Hct, potentially reflecting underlying hydration status. A subsequent proposal for this mechanism could be that PhA is inversely related to hs‐CRP levels in both men and women [41], and long‐standing inflammation plays an important role in the formation of anemia [42]. Taken together, the results showed a positive association between PhA and Hb and Hct levels. This finding suggests that higher PhA may co‐occur with better cellular function, balanced hydration, and lower levels of inflammation, without implying directionality (Figure 4). Proposed mechanisms linking PhA and ECW/TBW with anemia‐related parameters in women with type 2 diabetes. A higher ECW/TBW ratio, resulting from fluid displacement rather than a true deficiency of red blood cells, may contribute to lower measured hemoglobin and hematocrit levels through hemodilution, causing pseudo‐anemia [42]. There are reports disclosing a connection between anemia and fluid retention in CKD patients [43], also the case with those affected with heart failure [44, 45]. Hemodilution anemia is mostly detected in patients who have had heart failure resulting from volume overload. Although we did not include volume‐overload heart failure cases in the present study, the observed counter relationship between ECW/TBW and Hb could probably be accounted for by subclinical edema. Moreover, fluid retention observed in conditions such as patients who have been on continuous ambulatory peritoneal dialysis is also positively connected with response to EPO and diminished erythropoiesis [46]. As for PhA, findings point to a positive relationship between ECW/TBW ratio and hs‐CRP levels in patients undergoing peritoneal dialysis [47] (Figure 4). Mendelian randomization studies suggest a causal link between increased body water and reduced kidney function [48]. While animal models have shown no significant changes in fluid balance with iron deficiency anemia [49], it is important to note that low hemoglobin may affect fluid balance through neurohormonal mechanisms, particularly in conditions like heart failure [50]. This highlights the possibility of reverse causality, whereby anemia itself may influence hydration status and cellular integrity, rather than being solely a consequence of these factors, and future longitudinal studies are needed to investigate the temporal relationship between these variables more comprehensively. The findings of our study are indicative of the idea that PhA and ECW/TBW could be valuable indicators in women with diabetes, potentially leading to more effective management strategies. Higher PhA, reflecting better cellular function, may indicate a lower anemia risk in diabetics, while elevated ECW/TBW suggests hemodilution, potentially masking true anemia. Integrating these parameters into routine assessments can improve anemia diagnosis by distinguishing true from pseudo‐anemia, guiding interventions in nutrition, hydration, and metabolic control. Assessment of PhA and the ECW/TBW ratio opened a more comprehensive view of body composition and fluid balance. Inclusion of 1381 women with diabetes gave our statistical power a stronger boost, thus making the findings more reliable to be later generalized. However, there were limits and confounders to the study as well because of its cross‐sectional nature, which impeded our ability to draw causal inferences. The findings are specific to women with diabetes. This means that caution should be applied in attempts to generalize the results of the study populations other than women with diabetes. What is more, the physiological part of PhA and ECW/TBW in erythropoiesis is likely to be incompatible in the case of patients with T1DM and those experiencing T2DM. There are also potential confounders, namely variability in dietary intake, medication used, and the presence of other comorbid conditions, which may have influenced the results. Additionally, information on iron supplementation was not collected in the present cohort study. As a result, we could not adjust for its potential confounding effect on hemoglobin and hematocrit levels. Reliance on BIA for assessing body composition and fluid distribution may bring measurement errors into play. This is especially so as the hydration status and recent physical activity of the participants could have exerted an undetected influence on BIA accuracy. Further specific examination for the estimation of RBC lifespan and erythropoiesis is needed because of the fact that the parameters estimated by BIA in this study demonstrate the whole‐body cell conditions. Another reason for this need is that RBCs are a very small segment of the whole‐body cell mass, which contributes to PhA and ECW/TBW. Future longitudinal studies are recommended to verify these findings and inspect the mechanisms fundamental to associative observations in this study. Conclusions In conclusion, this study underscores that PhA and ECW/TBW are associated with anemia‐related parameters in women with diabetes and may serve as potential indicators for further investigation. The findings reveal that higher PhA is associated with a lower risk of anemia and better Hb and Hct levels, without implying a causal relationship, suggesting its function as a non‐invasive marker for nutritional and hydration status. Conversely, a higher ECW/TBW ratio is linked to an increased risk of anemia and lower Hb levels, indicating fluid retention's potential association with hematological health. Caution should be exercised with the interpretation of these findings because the study is cross‑sectional and there still exists a possibility for reverse causality. Author Contributions Niloufar Abdollahpour: investigation, writing – original draft. Najmeh Seifi: conceptualization, methodology, formal analysis. Ali Mottaghi Moghaddam Shahri: investigation, writing – original draft. Habibollah Esmaily: methodology, formal analysis. Gordon A. Ferns: writing – review and editing. Majid Ghayour‐Mobarhan: supervision, project administration. Ethics Statement All experiments were performed in accordance with the Declaration of Helsinki and Mashhad University of Medical Sciences ethical guidelines and regulations. The research protocol was approved by the School of Medicine, Mashhad University of Medical Sciences, Biomedical Research Ethics Committee (IR.MUMS.MEDICAL.REC.1395.458). Consent All participants signed a written informed consent before participating in the study. Conflicts of Interest The authors declare no conflicts of interest. Transparency Statement The corresponding author, Majid Ghayour‐Mobarhan, affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained. Supporting information Supporting File 1 Supporting File 2