Decreased intracellular water is associated with sarcopenic obesity in chronic haemodialysis patients

A multicentre, cross-sectional study in 20 dialysis units of tertiary general hospitals in Guizhou Province was conducted from June 1, 2021, to August 30, 2021. All adult haemodialysis patients, who received 4-h bicarbonate-based dialysis treatment thrice weekly for at least 3 months or longer, were invited to participate in our study. Exclusion criteria were as follows: (i) currently unstable patients, defined as patients who were acutely ill or hospitalized at the time of the assessment; (ii) patients with any physical deformities or those with contraindications such as implanted or external electronic devices, metallic implants that hindered the measurement of body composition; (iii) inability to complete questionnaire due to language incompatibility, visual or auditory disability; (iv) patients who did not cooperate with performing muscle performance tests; (v) participants who did not have routine blood tests in the past three months preceding the study; (vi) individuals with active cancer. The flow chart of patient screening is shown in Fig. 1. Ultimately, 3354 patients were included in our analysis.

Sarcopenic obesity is defined as the coexistence of sarcopenia and obesity. The diagnosis of sarcopenia was based on the revised Asian Working Group’s definition of sarcopenia (AWGS 2019) [40]. Patients with low muscle mass (defined as appendicular skeletal muscle index (ASMI) < 7 kg/m² in men and < 5.7 kg/m² in women) together with low muscle strength (defined as HGS < 26 kg in men and < 18 kg in women) were diagnosed with sarcopenia. The ASMI was calculated as appendicular skeletal muscle mass (ASM) in weight (kg)/height (m²), and the ASM was calculated according to an equation developed in Asian chronic HD patients recently [40]. Obesity was defined as a body fat percentage (BFP) ≥ 25% in men and ≥ 35% in women according to WHO recommendation, and these BFP cutoff values are also commonly used in CKD patients [41‐45].

Measurements of body composition were performed using a portable whole-body bioimpedance spectroscopy device, Body Composition Monitor (BCM, Fresenius Medical Care, Bad Homburg, Germany). The BCM measures body composition by analyzing the electrical responses at 50 different frequencies from 5 to 1000 kHz. The accuracy of BCM has been validated against the gold-standard methods [46, 47], and accumulating evidence suggests that body composition, as determined by the BCM, is a key predictor of survival in chronic HD patients [48, 49]. The measurement was carried out approximately 30 min before the HD session by a well-trained renal physician and a dialysis nurse, with four conventional electrodes being placed on the patient lying in supine position: two on the hand and two on the foot contralateral to the vascular access. In order to reduce operator variability, according to the manufacturer’s instructions, all tests were performed by the same operators. intracellular water, extracellular water, total body water and body fat mass were retrieved from the BCM software. BFP was calculated as fat mass divided by body weight.

Body weight was measured with an electronic scale while participants were wearing lightweight clothing. Standing height was assessed using a stadiometer with participants barefoot and standing erect. BMI was calculated as body weight in kilogram divided by height in meter squared. Hand grip strength (HGS) was measured in the dominant or non-fistula hand, using CAMRY® dynamometer with a precision of 0.1 kg. The patients sitting with their arms bent at an angle of 90º on a horizontal level held the tool with the fingers around it. Three measures were taken with 30 s of rest between each test and the maximum score was adopted for the study.

All the patients' individual information including sociodemographic data, comorbid conditions, drug usage, dialysis duration, dialysis mode, blood pressure, and biochemistry indexes was collected from the electronic medical record system, dialysis run sheets and a face-to-face interview with a predesigned questionnaire. Patients with a high school level were defined as individuals who had received more than nine years of schooling education. We collected the above data from each center within a strict quality-control framework, and further checked them to ensure the accuracy of our database. In this study, Stroke included hemorrhagic and ischemic stroke with a clinical history, confirmed by brain imaging test such as computed tomography or magnetic resonance. Cardiovascular disease (CVD) was defined as coronary artery disease (angina, myocardial infarction, percutaneous coronary intervention or coronary bypass grafting), congestive heart failure requiring hospitalization, arrhythmia (atrial fibrillation or other arrhythmia) and peripheral vascular disease (limb claudication needing percutaneous angioplasty and/or bypass grafting as treatment).

Continuous data with a skewed distribution were expressed as the medians and interquartile ranges. Categorical data were expressed as numbers and percentages. There has not been a uniform standard of ICW to define abnormal cellular hydration status as it might change with sex, race, detection methods and so on in HD patients. Therefore, this study used the sex-specific ICW median as surrogate marker for distinguishing between "High” or “Low” ICW level. We split patients into two groups according to sex-specific ICW median levels, resulting in High and Low groups with almost equal group size. To compare the patient characteristics of the two groups, Mann–Whitney U tests or Chi-square tests was used. Multivariable logistic regression analysis was performed to assess odds ratios (ORs) and 95% confidence intervals (CIs) for the association of ICW with risk of sarcopenic obesity, with ICW as continuous variables, or the high ICW groups as the reference. Variables that were statistically significant by univariate analysis were adjusted in the multivariate logistic regression models. To evaluate the robustness of the primary results, the subgroup analysis was performed to explore the potential effect modification by age (< 65 years; and ≥ 65 years), BMI (< 25 kg/m²; and ≥ 25 kg/m²), history of CVD (yes or no), and history of diabetes (yes or no). The potential interactions were evaluated across subgroups. The receiver operating characteristic (ROC) curve analysis was conducted to evaluate the performance and cutoff value of ICW. MedCalc software (version 19.0.4) was used to perform ROC curve analysis. The statistical packages R (http://​www.​r-project.​org; version4.0.1) was used to perform subgroup analysis. Other statistical analyses were performed using SPSS software (version 26.0; SPSS Inc., Chicago, IL, USA). A two-sided p < 0.05 was considered statistically significant.

A total of 3354 participants were included in our final analysis. Clinical characteristics of the enrolled patients are shown in Table 1 for all subjects and the low vs high ICW group. Of the 3354 participants recruited for this study, 2022 (60.3%) were men; the median age of the participants was 57 years (interquartile range: 46.8–68). A total of 752 (22.4%) participants were diagnosed as having sarcopenic obesity according to the criteria used in this study. The median ICW for men was 19.6 kg, and 15.3 kg for women. The prevalence of sarcopenic obesity in patients with low ICW was significantly higher than that in high ICW group (41.3% vs 3.0%). There were statistically reduced trends toward education level, albumin, SBP, DBP, BMI, ASMI and HGS with decreased ICW. Conversely, age, CRP, BFP, and dialysis vintage increased with decreasing ICW levels. Participants were more likely to have a history of diabetes, stroke and CVD with decreasing ICW levels. While usage of L-carnitine, serum lipids level, hemoglobin, platelet, and white blood cell were similar across two groups (p > 0.05). In order to remove medium and large molecular toxins, ensuring the dialysis adequacy of patients, most patients in this study also received two other dialysis modes regularly on the basis of haemodialysis (HD), namely hemodiafiltration (HDF) and hemoperfusion (HP). However, there was also no significant difference in the percentage of patients receiving HDF and HP between the two groups.

In univariate logistic regression analysis, these parameters including ICW, dialysis vintage, age, sex, usage of L- carnitine, history of diabetes, history of CVD, history of stroke, hemoglobin, albumin, SBP, DBP, triglyceride, BMI and CRP were statistically correlated with sarcopenic obesity (p < 0.05). With the high ICW group as the reference or ICW as continuous variables, multivariate-adjusted ORs and 95% CIs for sarcopenic obesity were summarized in Table 2. When regarded as continuous variables, higher ICW was a protective factor for the occurrence of sarcopenic obesity both in male and female (OR, 0.57 and 0.48 respectively). Individuals in low ICW group was 21.9-fold for men and 20.2-fold for women as likely to have sarcopenic obesity compared to those in the high ICW group after adjusting for other above potential confounders.

To further test the robustness of association between ICW and risk of sarcopenic obesity in different subgroups, subgroup analyses of multivariable logistic regression were performed (ICW as continuous variables) in various subgroups (Fig. 2). Overall, after adjusting for age, sex, dialysis vintage, history of stroke, history of diabetes, history of CVD, usage of L-carnitine, albumin, SBP, DBP, hemoglobin, triglyceride, BMI and CRP except for the stratified variable, the association between ICW and sarcopenic obesity stably persisted across subgroups stratified by age, diabetes, BMI, and CVD. The interaction analysis showed that age, BMI and diabetes had interactive roles in the association between ICW and sarcopenic obesity (all p for interaction < 0.05).

In male patients, the area under the ROC curve (AUC) was 0.868, with a sensitivity of 89% and a specificity of 71%, and the cutoff point was 19.1 kg (Fig. 3). In female, the AUC was 0.866, with a sensitivity of 87% and a specificity of 71%, and the cutoff point was 14.5 kg (Fig. 3). Therefore, ICW with different cutoffs had similar recognition value for sarcopenic obesity in male and female.