Synergic association of diabetes mellitus and chronic kidney disease with muscle loss and cachexia: results of a 16-year longitudinal follow-up of a community-based prospective cohort study

Baseline characteristics

The study flow and baseline characteristics of the subjects are shown in Figure 1, Table 1 and Supplementary Table 1. The mean age of the subjects was 51.2 ± 8.6 years, and 48.1% were male. The average estimated glomerular filtration rate (eGFR) was 91.2 ± 14.1 mL/min/1.73 m². Systolic blood pressure (SBP) and BMI were significantly higher in subjects with DM than in those without DM. In addition, subjects with DM were more likely to have cardiovascular comorbidities, such as hypertension, coronary artery disease, congestive heart failure, myocardial infarction, and peripheral vascular disease. In addition, they had worse metabolic profiles, as evidenced by higher homeostasis model assessment of insulin resistance (HOMA-IR), higher levels of total cholesterol and triglycerides, and lower high-density lipoprotein cholesterol (HDL-C) levels. However, kidney function did not differ between subjects with and those without DM.

Figure 1. Flow diagram of study cohort.

Association of DM and CKD with incident muscle depletion

During 73,059 person-years of follow-up (median, 13.7 years; interquartile ranges (IQR), 12.5–14.8 years), muscle depletion occurred in 460 (7.4%) subjects, with a corresponding incidence rate of 6.3 per 1000 person-years. Table 2 details the event rates of muscle depletion, cachexia, and all-cause mortality among the four groups. The adjusted cumulative incidence of muscle depletion was significantly higher in subjects with DM alone and in those with CKD alone than in those without the two conditions (Figure 2A); the incidence in subjects with both DM and CKD was the highest. In the cause-specific hazards regression model after sequential adjustment for confounding factors, the risk of muscle depletion development was significantly higher in subjects with DM alone than in those without DM and CKD (hazard ratio [HR], 1.37; 95% confidence interval [CI], 1.04–1.80). However, CKD alone was not associated with muscle depletion. The HR for muscle depletion was notably higher in subjects with the two conditions (HR, 3.38; 95% CI, 1.30–8.75) (Table 3).

Figure 2. Cumulative incidence function for development of (A) Incident muscle depletion, (B) Cachexia, and (C) all-cause mortality according to diabetes mellitus and chronic kidney disease.

We further examined this association by stratified age groups. Not surprisingly, the incidence rate of muscle loss was higher in older participants (Supplementary Table 2). In addition, the HRs for muscle loss were notably higher in participants aged ≥60 years, particularly when they had both DM and CKD (Supplementary Table 3).

Association of DM and CKD with the development of cachexia and all-cause mortality

The secondary outcome analysis showed results consistent with those of the primary outcome analysis (Table 2, Supplementary Table 4, and Figure 2B). A total of 210 (3.4%) cachexia events were noted during the follow-up. Subjects with DM alone were associated with a 1.55-fold (95% CI, 1.04–2.30) higher risk of cachexia compared to those without DM and CKD, and the risk of cachexia was markedly higher in subjects with the two conditions (HR, 6.07; 95% CI, 1.50–24.6). No significant association between CKD and the risk of cachexia was found.

Furthermore, we analyzed the association of DM and CKD with the risk of all-cause mortality (Table 2, Supplementary Table 4, and Figure 2C). During the follow-up, 203 (3.2%) deaths from any cause occurred. The risk of all-cause mortality was significantly higher in subjects with DM alone than in those without DM and CKD (HR, 1.98; 95% CI, 1.43–2.75). The HR in subjects with both DM and CKD was 2.62 (95% CI, 1.11–6.19).

Longitudinal changes in fat-free mass, fat mass, and body weight

To substantiate our findings further, we compared the slopes of changes in fat-free mass, fat mass, lean soft mass, and body weight among the four groups (Figure 3, Supplementary Figure 1, and Supplementary Table 5). The decline in fat-free mass, FFMI, lean soft mass, and lean soft mass index was greater in subjects with DM alone than in those without DM and CKD. Similar findings were observed in subjects with CKD alone, although the magnitude of the decline difference was less than that in subjects with DM alone. Moreover, the steepest annual decline rates were observed in subjects with both conditions. Consistent with these findings, concomitant declines in body weight and BMI were noted across the four groups. Nonetheless, fat mass declined in subjects with DM and CKD and increased in other groups.

Figure 3. Changes in (A) fat-free mass, (B) fat-free mass index, (C) lean soft mass, and (D) lean soft mass index over time according to 4 groups by diabetes mellitus and chronic kidney disease.

Risk of muscle depletion and cachexia with all-cause mortality

To examine the prognostic significance of muscle loss, we also analyzed whether individuals with muscle depletion have a higher risk of mortality. All-cause mortality occurred in 23 (5.0%) subjects with muscle depletion and in 180 (3.1%) subjects without muscle depletion (Supplementary Table 6 and Supplementary Figure 2A). In the multivariable Cox regression model after adjustment for confounding factors, muscle depletion was associated with a 6.42-fold higher risk of death (95% CI, 3.87–10.6). This association was greater in individuals with cachexia (HR, 10.9; 95% CI, 5.87–20.1) (Supplementary Table 7 and Supplementary Figure 2B).

Glycemic control during follow-up

Lastly, we examined glycemic control among the four groups. As expected, time-averaged glycated hemoglobin (HbA1c), fasting blood glucose (FBG), and 2-h postprandial blood glucose (PBG) levels were significantly higher in subjects with DM than in those without DM. In addition, among the subjects with diabetes, glycemic control was similar between those without CKD and those with CKD (Supplementary Figure 3). These findings suggest that glycemic control is unlikely to contribute to the higher muscle depletion risk in individuals with DM and CKD.

Ethical guidelines statement

All subjects voluntarily participated in this study and provided informed consent. The study protocol was approved by the Ethics Committee of KoGES at the Korean National Institute of Health. This study was performed in accordance with the Declaration of Helsinki and approved by the institutional review board of the National Health Insurance Medical Center (2021-02-018).

Study subjects

We used data from the KoGES, which is a prospective community-based cohort study. The details of the rationale, design, methods, and protocol summary were previously described [34]. Briefly, the study cohort consisted of participants recruited from the national health examinee registry, 10,030 subjects aged 40–69 years with a homogeneous ethnic background who were residents of Ansan (urban area) and Ansung (rural area), which are located near the capital city of Seoul, South Korea. For baseline recruitment, eligible participants were asked to volunteer through on-site invitations, telephone calls, mailed letters, community leader-mediated conferences, or media campaigns. The subjects underwent anthropometric examinations and laboratory tests and completed health-related lifestyle questionnaires in 2001 (baseline), and the tests were repeated biennially until 2016. The subject retention rate was 70.6% at the end of the seventh follow-up phase (16 years of follow-up). Key exclusion criteria were as follows: 1) subjects who did not undergo BIA at baseline (n = 2191); 2) those with muscle depletion at baseline (n = 813), and those with a BMI <20 kg/m² (n = 107). We excluded subjects with other missing data and those who were lost to follow-up after the baseline visit. Thus, a total of 6247 subjects were included in the final analysis (Figure 1).

Demographic, anthropometric, and laboratory data

Demographic and socioeconomic data, including age, sex, education and income levels, smoking status, alcohol intake, and medical histories, were obtained at study entry. Anthropometric parameters, such as height and weight, were measured by trained researchers according to a specified protocol: height was measured to the nearest 0.1 cm using a stadiometer with the subjects standing barefoot, weight was measured with the subjects wearing light clothes and no shoes, and BMI was calculated by dividing weight (kg) by height squared (m²). Moreover, based on surveys with a semiquantitative questionnaire, physical activity level was calculated as the daily estimated metabolic equivalents of task (MET) and was divided into quartiles. One MET is approximately 3.5 mL of oxygen consumed per kilogram of body weight per minute at rest. To obtain MET, subjects reported hours spent on sleep and five types of activities that were categorized according to intensity of activity: sedentary, very light, light, moderate, and heavy. Total MET hours per day were calculated by multiplying the reported hours spent by the MET that were determined based on each activity type [35]. In addition, education level was classified into three groups according to education duration: ≤6 years, 6–9 years, and >9 years. Income level was divided into three groups based on the average per-person monthly income: <$850, $850–$1699, and ≥$1700 per month.

Blood samples were obtained after an 8-h fast, transported to Seoul Clinical Laboratories (Seoul, Republic of Korea) within 24 h of sampling, and analyzed for the following biochemical parameters: serum creatinine, blood urea nitrogen, glucose, HbA1c, calcium, albumin, total cholesterol, triglyceride, HDL-C, and C-reactive protein (CRP). Serum creatinine was measured using the Jaffe method during the follow-up period; thus, we converted the non-isotope dilution-mass spectrometry (IDMS) creatinine to IDMS creatinine using the equation that was previously suggested [36, 37]. Subsequently, eGFR was calculated using the CKD Epidemiology Collaboration equation [38]. To evaluate insulin resistance, HOMA-IR was calculated using fasting glucose and insulin levels [39]. Urine samples were collected in the morning after the first voiding, and dipstick test was performed using URISCAN Pro II (YD Diagnostics Corp., Seoul, Korea). Urine albumin was quantified as absent, trace, 1+, 2+, or 3+ based on a color scale. Albuminuria was defined as albumin level ≥1+ based on the dipstick test.

Assessment of body composition

Body composition was assessed using multifrequency BIA (InBody 3.0, Biospace, Seoul, Korea). BIA was performed at baseline and biennially during the entire study period. Compared with conventional BIA-based methods that rely on formulae to calculate the estimated mass of each body component, multifrequency BIA assumes that the human body consists of five interconnecting cylinders and performs impedance measurements directly on these compartments. Impedances were measured at four specific frequencies (5, 50, 250, and 500 kHz) in five segments (both arms, trunk, and both legs) using a tetrapolar 8-point tactile electrode system.

Definition of diabetes mellitus and chronic kidney disease

DM was defined as DM diagnosis before study enrollment or receiving antidiabetic medications or insulin. Newly diagnosed DM was defined based on the American Diabetes Association criteria: FBG level ≥126 mg/dL (7.0 mmol/L), PBG ≥200 mg/dL (11.1 mmol/L) after 75-g oral glucose tolerance test, or HbA1c ≥6.5% (48 mmol/mol) [40]. Serum concentrations of HbA1c, FBG, and PBG were measured in all subjects at each follow-up visit.

CKD was defined as eGFR <60 mL/min/1.73 m² or presence of albuminuria at baseline examination according to the Kidney Disease Improving Global Outcomes guidelines [41].

Exposure and outcome

The main exposures of interest were DM and CKD. Patients were classified into four groups according to the presence or absence of the two diseases as follows: 1) without DM and CKD, 2) DM without CKD, 3) CKD without DM, and 4) with both DM and CKD. The primary endpoint was de novo development of muscle depletion, which was defined as a decline in FFMI based on BIA below the 10th percentile of the total KoGES population sample (male <16.9 kg/m² and female <15.2 kg/m²).

The secondary outcomes included the development of incident cachexia, all-cause mortality, and the slopes of changes in fat-free mass, fat mass, and body weight during the follow-up. Cachexia is an extreme type of muscle loss; thus, it was defined as muscle depletion with BMI <20 kg/m² or a decrease in body weight ≥10% for 24 months [14]. Primary outcome events were determined based on two or more event measurements, and the first of which was designated as the study endpoint.

Statistical analysis

Data were analyzed using STATA version 15 software (StataCorp, College Station, TX, USA). Continuous variables were expressed as means ± standard deviations; categorical variables, as absolute numbers with percentages. All data were tested for normality before the statistical analysis. The Kolmogorov–Smirnov test was performed to determine the normality of the parameter distribution. Intergroup comparisons were conducted using analysis of variance for normally distributed continuous variables and the χ² test or Fisher’s exact test for categorical variables. Data that did not show a normal distribution were presented as medians with IQR and compared using the Mann–Whitney U test or Kruskal–Wallis test. Cumulative incidence curves for individual muscle depletion and cachexia outcomes were derived using cumulative incidence functions considering competing risks. The Gray test was employed to determine statistically significant differences among groups.

Moreover, survival time was defined as the time interval between the baseline and the onset of the outcome or the last follow-up. Multivariable cause-specific hazards regression models were constructed to determine the association of DM and CKD with the risk of incident muscle depletion and cachexia. Variables that presented statistical significance in the univariable analysis were included in the multivariable models. Model 1 was adjusted for baseline age, sex, and BMI. Model 2 included demographic factors (education, income, alcohol consumption, smoking, and MET) and the presence of comorbidities, such as cardiovascular disease (myocardial infarction, congestive heart failure, unstable angina, peripheral artery disease, and cerebrovascular disease), previous cancer, and chronic obstructive pulmonary disease. Model 3 was further adjusted for SBP, eGFR, albuminuria, HDL-cholesterol, serum calcium, serum albumin, and CRP. Deaths before the occurrence of the primary outcome were treated as a competing risk and censored. Patients who were lost to follow-up were censored at the date of the last examination. We also compared the slopes of changes in fat-free mass, fat mass, lean soft mass, body weight, BMI, and adjusted with height index among the four groups using a generalized linear mixed model that incorporated random slopes and random coefficients. P values <0.05 were considered statistically significant.