Phase angle reference values derived from bioelectrical impedance analysis in chinese multi-ethnic adults

Abstract

Objectives

The study aimed to develop age-, sex-, and ethnicity-specific reference values for Phase Angle (PhA) in Chinese adults aged 20–80 years.

Materials and methods

A total of 13256 healthy participants (4793 men and 8463 women) aged 20–80 years from Han, Korean, and Yao ethnic groups in the China National Health Survey were included. PhA was measured using bioelectrical impedance analysis (BIA) at 50 kHz. Generalized Additive Models for Location, Scale, and Shape were used to construct age-, sex-, and ethnicity-specific 5^th, 10^th, 25^th, 50^th, 75^th, 90^th, and 95^th percentile values for PhA.

Results

The study population comprised 9616 Han individuals, 2136 Korean individuals, and 1504 Yao individuals. In men, PhA values peaked between 30–39 years for the Han and Korean populations and 40–49 years for the Yao population, with a steep decline after 50 years. The peak was observed in women at 40–49 years across all ethnic groups, followed by a gradual decline. Men consistently had higher PhA values than women across all ethnicities and age groups. The PhA reference values (5^th to 95^th percentiles) were 5.33-7.16° for Han men, 4.62-6.34° for Han women, 5.53-7.65° for Korean men, 5.04-6.79° for Korean women, 5.15–6.94° for Yao men, and 4.49-6.12° for Yao women.

Conclusion

This study provided age-, sex-, and ethnicity-specific reference values for PhA in diverse Chinese populations aged 20 to 80 years. The reference values for PhA derived from BIA hold the potential for enhancing the diagnosis and prognosis of various health conditions, further highlighting its utility in clinical settings.

Introduction

Bioelectrical impedance analysis (BIA) is a widely used method for body composition assessment because it is inexpensive, portable, time-efficient, and non-invasive [1]. The technique is based on the electrical properties of biological tissues, primarily resistance (R) and reactance (Xc), which are influenced by tissue composition, fluid distribution, and cell membrane properties. In conventional BIA applications, these electrical measurements are incorporated into population-specific predictive equations to estimate body composition parameters such as muscle mass (MM), fat mass (FM), and total body water [2]. However, the accuracy of such estimates strongly depends on the similarity between the target population and the population in which the equations were developed, and even under optimal conditions, estimation errors are unavoidable [3]. Moreover, the existence of numerous proprietary equations limits the comparability of body composition results across studies and devices [4].

Raw bioelectrical parameters, such as Phase Angle (PhA), are increasingly recognized for their clinical significance. The PhA, calculated as the arctangent value of the ratio of Xc to R measured from BIA, is independent of predictive equations for evaluating body composition, making it a valuable indicator of cellular health and nutritional status [5]. R is primarily influenced by total body water and conductive tissue mass, whereas reactance is related to the capacitive properties of cell membranes and tissue interfaces [6,7]. Recent studies have shown that individuals with acute and chronic diseases have lower PhA values than healthy subjects. PhA has emerged as a valuable prognostic marker for many conditions, indicating poorer outcomes in patients with cancer, HIV, cirrhosis, and those undergoing hemodialysis [8–10]. Therefore, lower PhA may predict worse health outcomes, including higher economic costs and increased mortality [11]. Moreover, studies have reported lower PhA was significantly associated with poor muscle function, particularly low muscle quality or strength [12].

For these reasons, the PhA has been used as a health status tool and an important predictor of disease severity and survival in various medical conditions. Hence, the reference value of PhA can serve as a valuable reference for providing a more precise assessment of health and nutritional status, and for helping to tailor medical and nutritional interventions more effectively. Several studies have established the reference value of PhA in Brazil [13], Italy [4,14], and the United States [15]. However, the PhA values have been shown to vary among different populations and ethnic groups [13–15]. These reference values may not apply to other populations or be suitable for general clinical settings so it is necessary to provide reference values of PhA specific to the Chinese population.

China is a multi-ethnic country, with 56 officially recognized ethnic groups, each possessing unique genetic backgrounds, dietary habits, and lifestyle practices that could influence body composition. There is limited data on the PhA reference values for different ethnic groups within China, especially PhA percentile curves. To address research gaps, using large-scale population-based data from the 2023 China National Health Survey (CNHS), we aimed to establish the PhA reference values from BIA among Chinese multi-ethnic adults aged 20–80 years, including those of Han, Korean, and Yao ethnicity.

Materials and methods

Data sources and population

The CNHS is an ongoing, nationally representative, cross-sectional survey in China. This study was part of the CNHS and was conducted in Guangdong (Han and Yao), Jilin (Han and Chinese Korean), and Jiangsu (Han) provinces from April to November 2023. Ethnicity data were extracted from participants’ official identification documents. In accordance with the recommendations of Flanagin et al. race and ethnicity are social constructs without biological meaning [16]. Detailed information about the study design has been reported previously [17]. Briefly, a multi-stage stratified cluster sampling method was used to enroll participants aged 20 years or above from selected communities and villages. We excluded participants < 20 years and >80 years (n = 132), with missing data on body composition (n = 313), with self-reported diseases that could significantly influence body composition including cardio-cerebrovascular diseases, respiratory diseases (e.g. COPD), endocrine and metabolic diseases (e.g. thyroid dysfunction and diabetes mellitus), musculoskeletal disorders (e.g. arthritis and fracture), kidney diseases (e.g. renal insufficiency and renal failure), liver diseases (e.g. liver cirrhosis and hypohepatia), or cancer (n = 5360). To ensure the accuracy of body composition measurements, participants with BMI < mean − 3 SD or > mean + 3 SD were also excluded (n = 48). The extreme outliers of PhA were handled using IQR method (±3 SD criteria) (n = 27). Finally, 13256 individuals (9616 Han, 2136 Korean, and 1504 Yao) were included for data analysis.

The study has been carried out following the Declaration of Helsinki. The Ethical Review Committee of the Institute of Basic Medical Sciences Chinese Academy of Medical Sciences, Beijing, China, has approved this study under the protocol (No. 2022177 and No. 2022134). Written informed consent was obtained from all participants.

Bioelectrical impedance and phase angle assessment

A segmental multifrequency body composition analyzer (MC780MA, Tanita Inc, Japan) was used by trained staff to measure bioelectrical impedance and obtain whole and segmental body composition data. Participants fasted for at least 8 h and avoided alcohol and caffeine for 24 h prior to measurement. Measurements were conducted barefoot with clean, dry skin in a temperature-controlled indoor environment (20–30 °C). Participants were instructed to place their bare feet on the designated markings on the analyzer platform, ensuring full contact, and measurements were obtained after adequate standing stabilization. Simultaneously, they were asked to grip the two metal handles with their hands. Female participants were measured outside the menstrual period when feasible. Individuals who were not in a fasted state, those with pacemakers or other metal implants, and those who consumed alcohol or caffeine before the measurement were excluded from the study to ensure the accuracy and reliability of the results. The device uses three different frequency currents (5, 50, and 250 kHz): low-frequency current to measure extracellular fluid and high-frequency current to measure intracellular fluid.

PhA was calculated using the Xc and R values measured at a frequency of 50 kHz [18]. The formula used to calculate PhA is PhA (°) = arc tangent (Xc/R) × (180°/π).

FM and MM were estimated using the manufacturer’s proprietary prediction equations embedded in the BIA device software (Tanita BC-780MA). In this study, MM refers to the device-derived estimate of total MM.

Statistical analysis

Continuous data was described as mean ± standard deviation (SD). One-way analysis of variance was used to compare differences among the three ethnic groups. When a statistically significant difference was found among the three groups, Tukey’s post-hoc test was performed. Generalized Additive Models for Location, Scale, and Shape (GAMLSS) was used to construct the sex-specific smoothed percentile curves for PhA in Han, Korean, and Yao people. The GAMLSS, an extension of the LMS (lambda-mu-sigma) method for fitting growth curves, is a flexible semi-parametric model that does not require the dependent variable distribution to belong to the exponential family and allows for large skewness and kurtosis [19].

We employed a two-step process to identify the optimal GAMLSS model: 1) Determining the distributions that best fit the data. Given that PhA values are positive continuous variables, we evaluated the fit of 22 continuous GAMLSS family distributions and selected the top three distributions for further exploration. The list of distributions assessed was presented in Supplementary Table 1; 2) Assessing the smoothing techniques that best fit the data for the previously selected distributions. Three smoothing techniques were assessed: polynomial splines (pb), cubic splines (cs), and fractional polynomials (fp). The best-fitting model was determined using the minimum Generalized Akaike Information Criterion (GAIC). We generated sex-age-ethnicity stratified 5^th, 10^th, 25^th, 50^th, 75^th, 90^th, and 95^th percentile values of PhA.

All data analyses were performed using SAS 9.4 software (SAS Institute Inc, Cary, NC, USA), with two-sided significance considered as p < 0.05. Percentile curves of PhA were plotted with R software (version 4.2.2).

Results

Participants characteristics

A total of 13256 subjects (4793 men and 8463 women) aged 20–80 years (mean age: 50.0 ± 12.5 years) were included in the study, including 9616 Han people (3679 men and 5937 women; mean age: 49.8 ± 12.2 years), 2136 Korean people (592 men and 1544 women; mean age: 49.7 ± 12.2 years), and 1504 Yao people (522 men and 982 women; mean age: 51.8 ± 14.1 years). Characteristics of the study population by sex and ethnicity are summarized in Table 1. Overall, men were older and had a higher BMI (24.8 ± 3.4 vs. 24.0 ± 3.4 kg/m²), lower FM (15.3 ± 6.2 vs. 19.2 ± 6.4 kg), and higher MM (51.9 ± 5.9 vs. 36.7 ± 3.2 kg) compared to women. Figure 1 presents PhA values by age group, sex, and ethnicity. Across all age groups, except for women aged 60–80 years, PhA values were highest among Korean individuals and lowest among Yao individuals in both men and women.

Table 1.

Characteristics of the study population by sex and ethnicity.

	Total (n = 13256)	Han (n = 9616)	Korean (n = 2136)	Yao (n = 1504)	P value
Men
Number	4793	3679	592	522
Age (years)	50.8 ± 12.8	50.6 ± 12.7	50.8 ± 11.5	52.1 ± 14.4	a c
Height (cm)	167.7 ± 6.8	168.8 ± 6.4	161.3 ± 5.9	167.5 ± 6.5	a b c d
Weight (kg)	70.0 ± 11.5	70.9 ± 11.4	65.0 ± 10.3	69.2 ± 11.4	a b c d
FM (kg)	15.3 ± 6.2	15.6 ± 6.3	13.2 ± 5.5	15.1 ± 6.2	a b d
MM (kg)	51.9 ± 5.9	52.4 ± 5.9	49.1 ± 5.4	51.2 ± 6.0	a b c d
BMI (kg/m2)	24.8 ± 3.4	24.8 ± 3.4	25.0 ± 3.4	24.6 ± 3.3	a b d
R (Ω)	552.6 ± 60.6	554.9 ± 61.0	537.6 ± 57.5	553.5 ± 57.5	a b d
Xc (Ω)	60.3 ± 7.7	60.3 ± 7.7	61.8 ± 8.1	58.7 ± 7.2	a b c d
Women
Number	8463	5937	1544	982
Age (years)	49.5 ± 12.3*	49.2 ± 11.9	49.2 ± 12.4	51.5 ± 13.9	a c d
Height (cm)	155.6 ± 6.3*	157.1 ± 5.8	150.1 ± 5.5	155.1 ± 5.9	a b c d
Weight (kg)	58.1 ± 8.9*	58.8 ± 8.9	56.3 ± 8.7	56.7 ± 8.5	a b c
FM (kg)	19.2 ± 6.4*	19.3 ± 6.4	18.9 ± 6.3	18.6 ± 6.3	a b c
MM (kg)	36.7 ± 3.2*	37.2 ± 3.2	35.3 ± 3.0	36.0 ± 3.0	a b c d
BMI (kg/m2)	24.0 ± 3.4*	23.8 ± 3.3	25.0 ± 3.5	23.6 ± 3.4	a b d
R (Ω)	647.0 ± 74.7	652.3 ± 75.8	623.0 ± 68.0	652.9 ± 70.3	a b d
Xc (Ω)	61.9 ± 7.8	61.7 ± 7.8	64.0 ± 7.5	60.0 ± 7.6	a b c d

Data were presented as mean ± SD or number.

BMI: body mass index; FM: fat mass; MM: muscle mass.

a, p < 0.05 for the difference among Han, Korean, and Yao.

b, p < 0.05 for the difference between Han and Korean.

c, p < 0.05 for the difference between Han and Yao.

d, p < 0.05 for the difference between Korean and Yao.

^*p < 0.001 vs. men participants.

Figure 1.

PhA values in men and women by age and ethnic groups.

Model results

Supplementary Tables 2–4 showed the best-fitting distributions of the GAMLSS model for PhA among Han, Korean, and Yao people. The best-fitting smoothing technique for all groups was pb for women and cs for men, except for Korean men, where pb was used. The best-fitting combinations of distributions and smoothing techniques are shown in Supplementary Table 5 and Table 2.

Table 2.

The best-fitting combinations of distributions and smoothing techniques.

	Men		Women
	Distribution	Smoothing technique	Distribution	Smoothing technique
Han	BCTo	cs	BCTo	pb
Korean	BCT	pb	exGAUS	pb
Yao	BCCG	cs	GA	pb

pb: polynomial splines; cs: cubic splines.

Normative values

The sex-, age-, and ethnicity-specific 5^th, 10^th, 25^th, 50^th, 75^th, 90^th, and 95^th percentiles of PhA are presented in Figure 2 and Table 3. Due to the large sample size of the Han population, percentile values for PhA in 10-year age groups were also calculated (Table 4). Across all ethnic groups and sexes, the PhA curves followed a similar pattern, characterized by an initial increase to a peak, followed by a subsequent decline. Specifically, in men, PhA values peak at 30–39 years among the Han and Chinese Korean ethnic groups, while in the Yao ethnic group, they peak at ages 40–49, with a sharp decline occurring after age 50. In women, peak values are observed in the 40–49 age group, followed by a gradual decline. Additionally, across all age groups and ethnicities, PhA values are consistently higher in men than in women.

Figure 2.

Reference curves of PhA for men and women in Chinese multi-ethnic adults. (A) Han populations; (B) Korean populations; (C) Yao populations.

Table 3.

Reference values of phase angle by age and sex.

Age (years)	Men										Women
	N	Mean	SD	P5	P10	P25	P50	P75	P90	P95	N	Mean	SD	P5	P10	P25	P50	P75	P90	P95
Han
20–39	777	6.47	0.54	5.53	5.75	6.10	6.45	6.81	7.17	7.41	1307	5.44	0.54	4.63	4.81	5.10	5.41	5.75	6.10	6.34
40–59	1939	6.35	0.55	5.43	5.66	6.00	6.34	6.69	7.04	7.29	3463	5.51	0.53	4.69	4.87	5.16	5.48	5.83	6.18	6.42
60–80	963	5.83	0.58	4.99	5.20	5.51	5.83	6.15	6.48	6.70	1167	5.24	0.53	4.46	4.63	4.90	5.21	5.54	5.87	6.10
Total	3679	6.24	0.61	5.33	5.56	5.89	6.23	6.58	6.92	7.16	5937	5.44	0.54	4.62	4.80	5.09	5.42	5.76	6.10	6.34
Korean
20–39	85	6.87	0.55	5.75	6.03	6.44	6.86	7.28	7.69	7.98	360	5.92	0.55	5.06	5.25	5.56	5.91	6.27	6.59	6.79
40–59	384	6.71	0.62	5.62	5.89	6.30	6.71	7.11	7.52	7.80	861	5.97	0.51	5.12	5.31	5.62	5.97	6.32	6.65	6.85
60–80	123	6.08	0.77	5.11	5.35	5.72	6.09	6.46	6.83	7.08	323	5.73	0.53	4.87	5.06	5.37	5.72	6.08	6.40	6.60
Total	592	6.60	0.70	5.53	5.80	6.21	6.61	7.00	7.39	7.65	1544	5.91	0.53	5.04	5.23	5.55	5.91	6.26	6.59	6.79
Yao
20–39	129	6.34	0.46	5.28	5.53	5.92	6.32	6.69	6.99	7.16	225	5.41	0.53	4.61	4.77	5.06	5.40	5.74	6.07	6.27
40–59	199	6.23	0.60	5.24	5.49	5.88	6.28	6.64	6.94	7.11	428	5.37	0.49	4.57	4.74	5.02	5.35	5.70	6.02	6.22
60–80	194	5.79	0.57	4.88	5.12	5.48	5.85	6.19	6.47	6.63	329	5.06	0.47	4.31	4.47	4.74	5.05	5.37	5.68	5.87
Total	522	6.09	0.60	5.15	5.38	5.74	6.12	6.47	6.77	6.94	982	5.28	0.52	4.49	4.65	4.94	5.26	5.60	5.92	6.12

Table 4.

Reference values of phase angle for han people by age (10-year age intervals) and sex.

Age (years)	Men										Women
	N	Mean	SD	P5	P10	P25	P50	P75	P90	P95	N	Mean	SD	P5	P10	P25	P50	P75	P90	P95
20–29	226	6.45	0.59	5.54	5.75	6.09	6.44	6.78	7.14	7.38	340	5.25	0.53	4.48	4.65	4.92	5.22	5.55	5.88	6.11
30–39	551	6.48	0.52	5.56	5.78	6.11	6.46	6.81	7.16	7.41	967	5.51	0.53	4.70	4.88	5.16	5.48	5.82	6.17	6.41
40–49	819	6.47	0.55	5.55	5.77	6.11	6.45	6.80	7.15	7.40	1502	5.58	0.53	4.76	4.94	5.23	5.55	5.89	6.24	6.49
50–59	1120	6.27	0.54	5.39	5.60	5.93	6.26	6.60	6.95	7.18	1961	5.46	0.52	4.66	4.83	5.12	5.43	5.77	6.11	6.35
60–69	752	5.91	0.55	5.08	5.29	5.60	5.91	6.23	6.55	6.78	925	5.28	0.52	4.51	4.68	4.95	5.26	5.58	5.92	6.15
70–80	211	5.53	0.59	4.75	4.94	5.23	5.53	5.83	6.13	6.34	242	5.07	0.56	4.31	4.47	4.73	5.02	5.33	5.65	5.87

In the Han ethnicity, the PhA reference values (5^th to 95^th percentile) were 5.33-7.16° for men and 4.62-6.34° for women. Similarly, PhA reference values ranged from 5.53-7.65° in Korean men, 5.04-6.79° in Korean women, 5.15–6.94 in Yao men, and 4.49-6.12° in Yao women.

Discussion

In this large-scale population-based study of 13256 multi-ethnic Chinese adults aged 20–80 years, we developed sex-specific smoothed percentile curves for PhA in Han, Korean, and Yao ethnic groups. To the best of our knowledge, this is the first study to present age-, sex-, and ethnic-specific reference values for PhA across diverse Chinese populations. The reference values established in this study can serve as a valuable resource for assessing nutritional status, identifying potential nutritional risks, and providing a normative basis for cross-cultural, cross-ethnic, and cross-age group comparisons across diverse populations.

The reference values provided in this study are descriptive in nature and are not intended to serve as diagnostic cut-offs for disease. Population-specific references are primarily useful for contextualizing individual measurements relative to a comparable population under similar measurement conditions, particularly for physiological indicators such as phase angle that are strongly influenced by body geometry and tissue composition.

Numerous studies have consistently reported sex differences in bioelectrical impedance parameters and age-related changes in PhA [13,20,21]. Consistent with these findings, our results demonstrated that men had significantly higher PhA values compared to women. These sex differences in PhA were observed across all ethnic and age groups, highlighting the need to establish separate normative values for males and females. PhA indicates cell membrane integrity and overall cell quality, and the observed sex differences may be attributed to the greater cell mass typically found in males [7,20]. This difference is also likely related to the generally higher MM and lower body fat in men than women, as PhA is strongly associated with MM [12]. Since muscle tissue has higher electrical conductivity, men typically exhibit higher PhA values than women.

The findings showed that PhA decreased with increasing age. This decline is strongly associated with the physiological aging process, which involves a gradual reduction in cell membrane integrity and cellular quality. As individuals age, MM and the number of healthy cells progressively decrease, while cellular membrane function deteriorates, leading to changes in bioelectrical impedance parameters and a reduction in PhA [22,23]. This trend not only reflects natural tissue loss and declining cell function but also potentially signals a deterioration in cellular health. Consequently, PhA is regarded as a potential biomarker for assessing age-related cellular aging and overall health [6]. Moreover, the rate of PhA decline may vary depending on an individual’s health status, lifestyle, and disease conditions, underscoring the importance of establishing age-specific normative values for PhA in adult populations to provide more accurate interpretations in clinical evaluations. For example, applying the same PhA criteria across adults aged 20–80 years to assess nutritional status may lead to misinterpretation, as age-related physiological changes in PhA could be incorrectly attributed to nutritional abnormalities.

Previous studies have revealed racial differences in PhA values [24,25]. Our study involved three ethnic populations in China: Han, Korean, and Yao. We initially analyzed ethnicity as a factor influencing PhA among Chinese individuals, finding significant differences in PhA values across these ethnic populations within each sex. In detail, PhA values were highest among Korean individuals and lowest among Yao individuals. The differences may be attributed to factors such as genetics, diet, and physical activity. Variations in dietary habits and nutritional intake, such as differences in protein and fat consumption, may influence body composition and, consequently, PhA [26]. Disparities in lifestyle and physical activity levels, such as the frequency of manual labor and cultural exercise practices, may also contribute to the observed variations in PhA [27]. PhA is recognized as a complex bioelectrical parameter reflecting multiple physiological domains, including cellular integrity, membrane function, hydration status, inflammation, and nutritional state [28,29]. For instance, exercise-induced increases in MM are associated with an expansion of total body water, with a preferential increase in intracellular water related to glycogen storage and osmotic shifts, alongside concurrent [30]. This reduces the R and thereby leads to an increase of the PhA. However, the specific mechanism through which ethnic diversity influences PhA in Chinese adults require further study. The findings suggested the significance of establishing ethnicity-specific reference values for PhA.

Sex, age, and ethnicity significantly affected PhA in the study population. These findings highlight the need for population-specific PhA reference values, which provide a normative basis for describing nutritional status and facilitating cross-population comparisons, offering valuable context for clinical interpretation. Given the substantial variations in PhA values across sex, age, and ethnic groups, a unified standard may be insufficient to accurately identify health risks in diverse populations. Hence, a generalized criterion could limit the effectiveness of disease prevention and management strategies.

There are few studies on the reference values of PhA in healthy adults. One study involving 4367 Italian subjects aged 18 to 65 years (2137 males and 2230 females) revealed the PhA mean values of 6.9° for males and 6.1° for females [4], which are significantly higher than the mean PhA reported in our study (Han: 6.24° for males and 5.44° for females; Korean: 6.60° for males and 5.91° for females; Yao: 6.09° for males and 5.28° for females). One possible explanation for the findings could be the younger age of the population included in the former study. Technical factors related to BIA measurement should be considered in addition to biological determinants. PhA is influenced by measurement posture, with standing assessments generally yielding lower values than supine measurements, likely due to posture-related fluid redistribution [31]. In addition, PhA varies across BIA devices because R and Xc depend on electrode configuration, operating frequency, and device-specific signal processing; therefore, raw BIA variables and derived PhA values are not directly interchangeable, highlighting the need for device- and protocol-specific reference values [4]. The lower PhA values observed in our study highlight the importance of establishing population-specific reference values to account for these differences, providing a normative framework for contextualizing individual measurements across diverse populations.

In this study, we explored all distribution families suitable for continuous data within the GAMLSS framework and selected the optimal combination of distribution and smoothing techniques. This comprehensive approach allowed us to model PhA distribution accurately across various ethnic groups and demographic characteristics. The flexibility of GAMLSS methodology enabled a more tailored derivation of normative values by selecting the best-fitting distribution and applying appropriate smoothing methods, ensuring that the generated normative values reflect the actual trends and variations within the population. The adoption of GAMLSS by the WHO for the construction of growth reference curves further highlights the robustness and applicability of this method in generating standardized health metrics [32]. Consequently, the reference values derived in this study are of significant clinical relevance, providing a valuable tool for health assessments, particularly in diverse ethnic populations.

This study has several strengths. First, the application of GAMLSS for estimating normative values enabled the construction of more precise reference intervals, providing more refined estimations than traditional statistical methods. Secondly, the study investigated variations in PhA across different ethnic groups, thereby reflecting the ethnic diversity within China and enhancing the cultural relevance and applicability of the results. Some limitations should also be acknowledged. First, the sample size for Chinese Korean male participants aged 20–40 was relatively small, however, we acknowledge that stratification by ethnicity resulted in unequal subgroup sizes, which may limit the precision of percentile estimates and the generalizability of findings within certain ethnic groups. The Clinical and Laboratory Standards Institute recommends using at least 120 samples to calculate reference intervals for healthy populations to ensure sufficient sample size for reliable statistical estimation [33]. The smaller sample size in this specific age group may affect the generalizability of the reference intervals for Chinese Korean males in this age range. Second, the present study focused on only three ethnic groups. Further research is needed to evaluate whether ethnic disparities exist among additional ethnic groups and to establish appropriate normative values for other ethnic minority populations.

In conclusion, this study established age-, sex-, and ethnicity-specific reference values for PhA in diverse Chinese populations aged 20 to 80 years. Significant differences in PhA were observed across age groups, sexes, and ethnicities. These reference values provide a crucial baseline for both clinical practice and research, offering valuable insights into the health status of Chinese adults. PhA derived from BIA holds the potential for enhancing the assessment of various health conditions, further highlighting its utility in clinical settings. Moreover, this study offers methodological guidance for the establishment of reference values for other physiological indicators, contributing to future research and the development of standardized health assessment criteria.

Funding Statement

This work was supported by Research on the Basic Resources of Science and Technology in the Ministry of Science and Technology (2022FY100800), CAMS Innovation Fund for Medical Sciences (2021-I2M-1-023), and State Key Laboratory Special Fund (2060204).

Ethics approval and consent to participate

The study has been carried out in accordance with the Declaration of Helsinki. The Ethical Review Committee of the Institute of Basic Medical Sciences Chinese Academy of Medical Sciences and the National Center for Health Statistics Research Ethics Review Board, have approved this study under the protocol. Written informed consent was obtained from all participants.

Disclosure statement

The authors declare no conflicts of interest.

Availability of data and materials

The datasets of CNHS analyzed during the present study are available from the corresponding author on reasonable request.