1. Introduction

Cadmium (Cd) is a ubiquitous toxic metal that occurs naturally but is continuously released into the environment through industrial activities, agricultural practices, and consumer products (Järup & Åkesson, 2009; Charkiewicz, Omeljaniuk, Nowak, Garley, & Nikliński, 2023; Godt et al., 2006). As a non-essential element in the human body, Cd accumulates primarily in the liver and kidneys following absorption, characterized by an exceptionally long biological half-life due to its minimal excretion rate (Menke, Muntner, Silbergeld, Platz, & Guallar, 2009). The kidney acts as the critical target organ for Cd toxicity. Even chronic exposure to low concentrations can lead to progressive nephrotoxicity, including proximal tubular dysfunction, making Cd a high-priority substance for public health management by the European Food Safety Authority (EFSA). (Järup, Berglund, Elinder, Nordberg, & Vahter, 1998; EFSA, 2009; Swaddiwudhipong et al., 2010; Staessen et al., 1999). Accordingly, the EFSA has emphasized the necessity of managing dietary Cd exposure at the population level and has established intake limits, such as the Tolerable Weekly Intake (TWI), to mitigate human health risks (EFSA, 2009; Joint FAO/WHO, 2007).

The primary pathways of Cd exposure include active and passive smoking, the consumption of contaminated foods—particularly soil-based products like grains and vegetables—and contact with environmental media affected by industrial emissions. Individual susceptibility also plays a crucial role; for instance, Cd absorption significantly increases in vulnerable groups, such as individuals with iron deficiency (EFSA, 2009). To monitor internal exposure, blood and urinary cadmium (U-Cd) levels are widely utilized. Specifically, U-Cd is recognized as a reliable biomarker reflecting long-term body burden and renal accumulation (Satarug, Garrett, Sens, & Sens, 2011). To ensure accurate interpretation across varying physiological conditions, U-Cd levels are typically standardized using creatinine correction or adjusted statistically by including creatinine as a covariate (Nordberg, Nogawa, Nordberg, & Friberg, 2018). In this context, urinary cotinine, a primary metabolite of nicotine, serves as a reliable objective biomarker to quantify smoking exposure.

Nationwide biomonitoring studies that characterize the distribution of U-Cd and identify high-exposure groups provide a fundamental basis for establishing environmental health policies (Téllez-Plaza et al., 2012; Åkerström et al., 2013). In South Korea, previous studies using the 1st to 3rd cycles of the Korea National Environmental Health Survey (KoNEHS) have established baseline exposure levels and identified general demographic determinants (Son, Lee, Paek, & Lee, 2009; Joo et al., 2019; Lee et al., 2012; Burm et al., 2016). However, while these earlier reports primarily focused on mean-centered comparisons and broad distribution patterns, they often lacked a granular analysis of the high-exposure tail (e.g., P90 or P95), which represents the most vulnerable segment of the population requiring targeted intervention. Analyzing the 4th Cycle of KoNEHS (2018–2020) offers critical new insights that differentiate it from previous cycles. First, the 4th cycle data reflects the most recent shifts in the socio-environmental landscape, including changes in dietary habits and environmental regulations that may have altered Cd exposure pathways since the 3rd cycle (Kim, Kim, An, Sung, & Sim, 2017; Han et al., 2013). Second, while previous findings suggested a persistent gender gap in Cd accumulation, there is a lack of evidence on whether this vulnerability is intensifying or shifting among specific subgroups of women under current exposure scenarios. This study moves beyond simple trend reporting by employing a weighted quantile-based profiling approach to systematically characterize the “high-exposure group”. By comparing the P90/P95 clusters with the general population, we aim to identify specific lifestyle and environmental determinants that remain hidden in mean-centered analyses (Han et al., 2013; Chaumont et al., 2013). This transition from general distribution reporting to targeted high-risk profiling provides a more direct basis for risk communication and the development of exposure-reduction strategies tailored to the Korean context, where unique dietary patterns (e.g., rice and grain consumption) and occupational heterogeneity persist (Ruiz-Hernandez et al., 2017).

2. Materials and Methods

2.1. Study Design and Data Source

This study is a cross-sectional analysis using raw data from the 4th Korea National Environmental Health Survey (KoNEHS), conducted between 2018 and 2020. KoNEHS is a nationwide biomonitoring program designed to represent the environmental exposure levels and health status of the South Korean population. Of the 5,937 adults who participated in KoNEHS Cycle 4, participants with missing values for key variables—urinary cadmium (U-Cd), creatinine, and cotinine—were excluded. Consequently, a total of 4,239 participants were included in the final analysis. This selection ensures that the associations between cadmium accumulation, renal function, and smoking exposure could be evaluated using a complete dataset for all primary indicators. The study population included both smokers and non-smokers. Smoking status was not used as an inclusion or exclusion criterion. Instead, urinary cotinine concentrations were measured for all participants and utilized as a continuous variable in the analysis to objectively account for varying degrees of active and passive smoking exposure.

2.2. Measurement of U-Cd

U-Cd was measured according to standardized analysis procedures with rigorous internal and external quality control. To account for variations in urine dilution, cadmium concentrations were standardized as µg/g creatinine. Since the distribution of U-Cd was right-skewed, natural log-transformation was applied for continuous variable analysis. Weighted percentiles were calculated to reflect the national population distribution in all descriptive statistics.

2.3. Covariates and Assessment of Smoking Exposure

The primary covariates included age, gender, and smoking exposure. Age was included as a critical factor reflecting the long-term bioaccumulation of cadmium, while gender was used to account for physiological and environmental exposure differences. Notably, smoking exposure was objectively evaluated using urinary cotinine concentration rather than self-reported questionnaire data. The rationale for using urinary cotinine is its reliability as a biomarker that captures both active smoking and significant secondhand smoke exposure while minimizing recall bias inherent in surveys. Cotinine levels were log-transformed for statistical analysis to address their skewed distribution, allowing for a quantitative adjustment of smoking’s independent effect on U-Cd levels.

2.4. Exposure Statistical Analysis

All statistical analyses were performed using Python-based packages, accounting for the complex sampling design of KoNEHS (stratification, clustering, and weighting). Weights were applied to produce estimates representative of the Korean adult population, and cluster-robust standard errors were utilized. The high-exposure group was defined as individuals with U-Cd levels at or above the 90th percentile (P90) of the weighted distribution. While the 95th percentile (P95) is a conventional reference for the upper limit of the general population, the P90 threshold was adopted to ensure a sufficient sample size within the high-exposure subgroup for robust statistical comparisons. The P95 criterion was additionally reviewed as a sensitivity analysis. Differences in age, gender, and cotinine levels between the high-exposure group (≥ P90) and the non-high-exposure group (< P90) were evaluated. For the female-specific subgroup analysis, weighted least squares (WLS) regression was employed, with log-transformed U-Cd as the dependent variable and age and log-transformed cotinine as independent variables. Statistical significance was defined as a two-sided p-value < 0.05.

2.5. Ethical Considerations

This is a secondary analytical study using public national statistical data and was conducted with the approval of the National Institute of Environmental Sciences (NIER-2020-01-016). Use of KoNEHS raw data was carried out in accordance with the data usage guidelines of the Korea Centers for Disease Control and Prevention. All procedures were applied in accordance with the 2013 Declaration of Helsinki of the World Medical Association.

3. Results

3.1. Study Population Characteristics

The final analysis included 4,239 adults from the 4th KoNEHS who had complete data for U-Cd, creatinine, and cotinine. By applying sampling weights, the study subjects were confirmed to be representative of the South Korean adult population in terms of age and gender distribution. The weighted mean age of the total study population was 47.5 years, and 50.2% were female. The overall weighted geometric mean of urinary cotinine was 6.02 ng/mL.

3.2. Distribution of U-Cd Levels

Creatinine-corrected U-Cd levels showed a right-skewed distribution. The weighted median value was 0.64 µg/g creatinine, with the 25th and 75th percentiles at 0.32 µg/g and 1.09 µg/g creatinine, respectively. To define the upper exposure groups, quantile analysis was performed: the 90th percentile (P90) was 1.51 µg/g creatinine, and the 95th percentile (P95) was 1.97 µg/g creatinine (Table 1).

3.3. Characteristics of the High-Exposure Group

The high-exposure group (≥ P90) accounted for 10.0% of the weighted population. This group was significantly older than the non-high-exposure group (< P90), with a mean age of 61.29 years compared to 45.97 years—a difference of over 15 years. A striking disparity was observed in gender distribution: while women made up 46.4% of the non-high-exposure group, they constituted 84.2% of the high-exposure group (Table 2).

3.4. Smoking Biomarker Levels According to Exposure Status

As a result of comparing the U-Cd reflecting smoking exposure, the weighted geometric mean of cotinine in the high exposure group was 3.36, which was lower than 6.42 in the non-exposed group. These results suggest that the urine cadmium high exposure group is not the group with higher smoking exposure.

3.5. Gender-Specific Analyses in Women

Among the 2,350 women analyzed, 16.8% were classified in the high-exposure group (≥ P90). Similar to the overall population, the mean age was significantly higher in the female high-exposure group (61.67 years) compared to the non-high-exposure group (45.81 years). However, the weighted geometric mean of urinary cotinine showed no substantial difference between the high-exposure (2.14) and non-high-exposure (2.26) groups among women (Table 3).

3.6. Weighted Regression Analysis in Women

Weighted least squares regression was performed for women using log-transformed creatinine-corrected U-Cd as the dependent variable. Age showed a significant positive association with U-Cd (β = 0.042, p < 0.001), indicating steady accumulation over time. In contrast, urinary cotinine levels did not show a significant association with U-Cd in women (p = 0.615). The regression model explained 35.6% of the variance (R² = 0.356) (Table 4).

4. Discussion

This study characterized the distribution of U-Cd and the profiles of high-exposure groups among Korean adults using the 4th KoNEHS data. Our findings confirm that Cd accumulation is primarily driven by age rather than short-term behavioral factors like smoking, a trend that is particularly pronounced among women. By using objective biomarkers and focusing on the upper 10% of the exposure distribution, this research provides a more granular understanding of environmental Cd exposure in the Korean context.

The statistically significant positive association between age and U-Cd especially in the female-only regression model, aligns with the biological properties of Cd. As a metal with a biological half-life of several decades, Cd accumulates in the renal cortex over long periods (Satarug, Ruangyuttikarn, Nishijo, & Ruiz, 2018). Unlike blood cadmium, which may reflect recent exposure, U-Cd serves as a reliable indicator of lifetime body burden and chronic kidney accumulation (Åkesson et al., 2002). Our results reconfirm international evidence from studies such as NHANES, which also show that U-Cd increases with age even after adjusting for urinary cotinine levels as an objective biomarker of smoking exposure (Richter, Bishop, Wang, & Swahn, 2009).

A key contribution of this study is the objective assessment of smoking exposure through urinary cotinine. While smoking is a known source of Cd, many previous studies relied on self-reported questionnaires, which are prone to recall bias and exposure misclassification (Richter, Faroon, & Pappas, 2017). By using cotinine, we found that the high-exposure group actually had lower smoking biomarker levels than the general population. This paradox suggests that for the Korean population—particularly women—the primary drivers of high Cd exposure are likely dietary or environmental rather than tobacco-related.

The significant overrepresentation of women in the high-exposure group (84.2%) highlights a specific demographic vulnerability. Physiological factors, such as higher intestinal absorption of Cd in individuals with low iron stores—a condition more common in women—may contribute to this long-term accumulation (Wiseman et al., 2017; Milki et al., 2022). Furthermore, Korea’s unique dietary structure, centered on rice and vegetables which are known to absorb Cd from the soil, may disproportionately affect women and the elderly through chronic low-dose ingestion (Kim et al., 2019; Cirovic & Satarug, 2024).

By shifting the analytical focus from mean values to the upper percentiles (P90 and P95), this study identifies the population segments most in need of health monitoring. The finding that 10% of Korean adults, predominantly elderly women, are in the high-exposure category underscores the need for targeted risk communication and dietary intervention strategies. This approach moves beyond general trend reporting and provides an epidemiological basis for managing Cd as a chronic environmental health risk rather than solely an occupational hazard. Despite the strengths of using a nationally representative dataset and objective biomarkers, this cross-sectional study cannot establish direct causality. Future longitudinal research is needed to investigate specific dietary intake patterns and local environmental factors. Nonetheless, this study provides a critical update on the cadmium exposure landscape in South Korea, offering a practical foundation for public health policy and risk management.

5. Conclusion

This study confirmed that U-Cd in Korean adults, especially women, increased with age independently of smoking. This suggests that urine cadmium is an index that reflects long-term cumulative environmental exposure, not a short-term behavioral factor, and can contribute to reconsidering the direction of cadmium exposure management in the general population.