Urinary Cadmium Threshold to Prevent Kidney Disease Development

The frequently observed association between kidney toxicity and long-term cadmium (Cd) exposure has long been dismissed and deemed not to be of clinical relevance. However, Cd exposure has now been associated with increased risk of developing chronic kidney disease (CKD). We investigated the link that may exist between kidney Cd toxicity markers and clinical kidney function measure such as estimated glomerular filtration rates (eGFR). We analyzed data from 193 men to 202 women, aged 16−87 years [mean age 48.8 years], who lived in a low- and high-Cd exposure areas in Thailand. The mean (range) urinary Cd level was 5.93 (0.05–57) μg/g creatinine. The mean (range) for estimated GFR was 86.9 (19.6−137.8) mL/min/1.73 m2. Kidney pathology reflected by urinary β2-microglobulin (β2-MG) levels ≥ 300 μg/g creatinine showed an association with 5.32-fold increase in prevalence odds of CKD (p = 0.001), while urinary Cd levels showed an association with a 2.98-fold greater odds of CKD prevalence (p = 0.037). In non-smoking women, Cd in the highest urinary Cd quartile was associated with 18.3 mL/min/1.73 m2 lower eGFR value, compared to the lowest quartile (p < 0.001). Evidence for Cd-induced kidney pathology could thus be linked to GFR reduction, and CKD development in Cd-exposed people. These findings may help prioritize efforts to reassess Cd exposure and its impact on population health, given the rising prevalence of CKD globally.


Introduction
Exposure to the heavy metal cadmium (Cd) is inevitable for most people as this metal is present in foodstuffs, cigarette smoke and polluted air [1][2][3][4]. By total diet studies, staple foods such as rice, potatoes, and wheat constitute 40-60% of total dietary Cd intake in the average consumer in various populations [4]. In addition, offal, spinach, shellfish, crustacean and mollusks constitute dietary Cd sources [4]. Cd oxide (CdO) in cigarette smoke and polluted air has relatively high bioavailability. Consequently, most smokers show elevated Cd levels in their blood, urine, Numbers are arithmetic mean ± standard deviation (SD). eGFR is determined with CKD−EPI equation, and eGFR < 60 mL/min/1.73 m 2 is defined as CKD [28]. a Both current and ex-smokers are grouped together because of a known long half-life of Cd in the body. b Hypertension was defined as systolic blood pressure ≥ 140 mmHg, or diastolic blood pressure ≥ 90 mmHg, physician diagnosis, or prescription of anti-hypertensive medications. c Tubular Cd toxicity threshold, established by the European Food Safety Agency [29]. d Tubular Cd toxicity threshold, established by the FAO/WHO [30]. e Severe and irreversible tubular dysfunction [19,20].

Ascertainment of Long-Term Cadmium Exposure Levels
Assessment of long-term Cd exposure or body burden was based on creatinine-adjusted urinary Cd concentrations. Urinary Cd is a suitable exposure marker to assess kidney effects since the majority of Cd in urine is ultrafilterable, but not reabsorbed by kidney tubules [2]. The plasma Cd concentration reflects Cd influx into blood circulation from external sources (diet and air) and internal reservoirs (liver). Accordingly, urinary Cd excretion rate is proportional to plasma Cd concentrations, glomerular filtration rates and tubular sequestration rates [2]. For the Bangkok group [26], the urinary Cd concentrations were determined with the inductively-coupled plasma/mass spectrometry, calibrated with multi-element standards (EM Science, EM Industries Inc., Newark, NJ, USA). Quality assurance and control were conducted with simultaneous analysis of samples of the reference urine Lyphochek ® (Bio-Rad, Sydney, Australia), which contained low-and high-range Cd levels. The coefficient of variation of 2.5% was obtained for Cd in the reference urine. Cd concentrations of urine samples reported below the limit of detection (LOD) of 0.05 µg/L were assigned as the LOD divided by the square root of 2. For the Mae Sot group [13], urinary Cd concentrations were determined with an atomic absorption spectrophotometer (Shimadzu Model AA-6300, Kyoto, Japan). Urine standard reference material No. 2670 (The National Institute of Standards, Washington, DC, USA) was used for quality assurance and control purposes.
Assessment of tubular dysfunction was based on a reduction in tubular reabsorption activity, reflected by an increase in urinary excretion rate of β2-MG [12][13][14][15]. Due to a small molecular weight, β2-MG is filtered, reabsorbed by tubules, and approximately 0.3% of filtered β2-MG is excreted in urine [2]. Assessment of tubular integrity was based on urinary excretion of the enzyme NAG [12][13][14][15] and urinary NAG excretion is considered to be proportional to nephron numbers as this enzyme originates mostly from tubular epithelial cells which is released upon cell injury [2]. For the Bangkok group, the urinary β2-MG assay was based on the latex immunoagglutination method (LX test, Eiken 2MGII; Eiken and Shionogi Co., Tokyo, Japan), and the urinary NAG assay was based on an enzymatic reaction and colorimetry. The urinary protein assay was based on turbidimetry (Roche/Hitachi 717, Boehringer Mannheim and Roche Diagnostics, Roche Diagnostics GmbH Mannheim, Germany). The urinary and serum creatinine assay was based on the Jaffe's reaction.
For Mae Sot group, the urinary β2-MG assay was based on an enzyme immunoassay (GLAZYME β2 microglobulin-EIA test kit, Sanyo Chemical Industries, Ltd., Kyoto, Japan), while the urinary NAG assayed was based on colorimetry (NAG test kit, Shionogi Pharmaceuticals, Sapporo, Japan). The urinary protein assay was based on the Kingsbury-Clark method, while the urinary and serum creatinine assay was based on the Jaffe's reaction.

Statistical Analysis
The SPSS statistical package 17.0 (SPSS Inc., Chicago, IL, USA) was used to analyze data. We used the Mann-Whitney U-test to compare two groups of subjects. The distribution of the variables was examined for skewness and those showing right skewing were subjected to logarithmic transformation before analysis, where required. One sample Kolmogorov-Smirnov test was used to detect a departure from normal distribution of variables. We used the logistic regression analysis to estimate Prevalence Odds Ratio (POR) for CKD, attributable to Cd exposure and kidney tubular pathologies. The univariate analysis was used to estimate effect size of Cd exposure levels with adjustment for covariates and urinary Cd quartiles × smoking × gender interactions. In addition, we used a multilinear regression analysis to evaluate the strength of associations between eGFR and its predictors in subjects stratified by gender, smoking status and Cd exposure levels. p values ≤ 0.05 for a two-tailed test was considered to indicate statistical significance.

Characteristics of Study Subjects
Of 395 study subjects, 202 were women and 193 were men. The mean age of women was 4 years older than the mean age of men of 47.4 years (p = 0.024). Smoking was more prevalent in men than women (66.8% vs. 24.3%) (p < 0.001). The mean (SD) values for eGFR were 86.9 (24.2) mL/min/1.78 m 2 (range: 19.6-137.8). The CKD prevalence was 13% in men and 12.4% in women (p = 0.863), while hypertension prevalence was 24.2% in men and 19.7% in women (p = 0.240).
The mean urinary creatinine concentrations in men was higher than women (p < 0.001). The mean urinary Cd concentrations in men (7.48 µg/L) and women (5.87 µg/L) did not differ (p = 0.930). The mean (SD) urinary Cd was 5.93 (7.69) µg/g creatinine (range: 0.05-57.57). The mean urinary Cd tended to be higher in women than men, when data were adjusted for urine dilution by creatinine excretion (6.41 vs. 5.43 µg/g creatinine, p = 0.061). The prevalence of urinary Cd levels above 5.24 µg/g creatinine was 40.3%, while more than half (55.9%) of the subjects had urinary Cd levels, exceeding 1 µg/g creatinine. The prevalence of severe and irreversible tubular dysfunction (urinary β2-MG levels ≥ 1000 µg/g creatinine) was 17.1% in men and 14. 2% in women (p = 0.104). Urinary β2-MG, NAG and protein levels in men and women did not differ.
In all subjects, creatinine-adjusted urinary Cd levels showed a strong correlation with age (Spearman rank's correction coefficient (r) = 0.644, p = < 0.001), and this association between age and urinary Cd levels persisted after stratification by smoking status (r = 0.627, p < 0.001 for non-smokers, r = 0.540, p < 0.001 for smokers). There was an inverse correlation between urinary Cd levels and BMI (r = −0.214, p < 0.005) in all subjects. After controlling for age, the association of urinary Cd levels and BMI persisted in smokers only (r = −0.166, p = 0.027), while there was a tendency for an association in non-smokers (r = −0.119, p = 0.081).
The prevalence rates of various diseases reported by participants differed in men and women (Likelihood Chi-square 15.5, p = 0.03). Osteoporosis was more prevalent in women than men (5.6% vs. 0.5%, p = 0.004). Kidney disease diagnosis tended to be higher in men than women (4.7% vs. 1.5%, p = 0.083).

Discussion
Herein, we have observed for the first time an association of a 5.32-fold rise in CKD prevalence odds and urinary β2-MG levels ≥ 300 μg/g creatinine in Thai subjects with chronic environmental exposure to Cd. This independent association between elevated levels of urinary β2-MG and a marked increase in odds of CKD prevalence suggests a vital role played by kidney tubular cells in the pathogenesis and/or progression of CKD. Indeed, a tubular-glomerular connection is increasingly recognized [31] as is the evidence for β2-MG as marker of a range of kidney disease [32][33][34]. Our

Discussion
Herein, we have observed for the first time an association of a 5.32-fold rise in CKD prevalence odds and urinary β2-MG levels ≥ 300 μg/g creatinine in Thai subjects with chronic environmental exposure to Cd. This independent association between elevated levels of urinary β2-MG and a marked increase in odds of CKD prevalence suggests a vital role played by kidney tubular cells in the pathogenesis and/or progression of CKD. Indeed, a tubular-glomerular connection is increasingly recognized [31] as is the evidence for β2-MG as marker of a range of kidney disease [32][33][34]. Our

Discussion
Herein, we have observed for the first time an association of a 5.32-fold rise in CKD prevalence odds and urinary β2-MG levels ≥ 300 µg/g creatinine in Thai subjects with chronic environmental exposure to Cd. This independent association between elevated levels of urinary β2-MG and a marked increase in odds of CKD prevalence suggests a vital role played by kidney tubular cells in the pathogenesis and/or progression of CKD. Indeed, a tubular-glomerular connection is increasingly recognized [31] as is the evidence for β2-MG as marker of a range of kidney disease [32][33][34]. Our finding concurs with experimental data and clinical outcomes that suggest urinary β2-MG is a predictor of GFR reduction [32][33][34].
Supporting tubular-glomerular connection are data from a prospective cohort study in Japan showing that a sign of tubular impairment (urine β2-MG levels ≥ 300 µg/g creatinine) was associated with a 79% (95% CI: 1.07, 2.99) increase in the likelihood of having eGFR fall at high rates, i.e., 10 mL/min/1.73 m 2 over 5-year observation period [35]. In another cross-sectional study, a milder tubular impairment (urine β2-MG levels ≥ 145 µg/g creatinine) was associated with an increase in the prevalence odds for hypertension in Japanese subjects [36]. Results of these Japanese studies underscored clinical values of urine β2-MG measurement, but Cd exposure levels experienced by Japanese subjects in these two studies were not measured. Thus, it is unknown if these observed outcomes (rapid GFR reduction and hypertension development) in subjects with high urine β2-MG levels could be linked to Cd or other environmental factors.
Urinary Cd levels > 1 µg/L (>0.5 µg/g creatinine) were associated with a 48% increase in the risk of CKD development (95% CI: 1.01, 2.17) in adult participants in the U.S. NHANES 1999-2006 cycle [21]. Consistent with the U.S. study is our finding of an association between elevated Cd body burden, assessed by urinary Cd levels, and an increase in odds of CKD prevalence (2.98 fold). Multilinear regression data indicated also that lower eGFR values were associated with higher urinary Cd levels. Further, in an effect-size analysis, a dose-response between eGFR reduction and urinary Cd quartiles was evident in non-smoking women. This may implicate dietary Cd intake in the pathogenesis of CKD. Likewise, in a Chinese population study, cumulative Cd intake estimate was associated with a 4-fold increase in CKD prevalence (95% CI: 2.91, 5.63) [25].
An association of lower eGFR and higher blood Cd levels was noted in Korean population [37] and the representative of the U.S. population (the U.S. NHANES, 2007-2012) [38]. In a Korean study, blood Cd levels in the highest tertile were associated with 1.85 mL/min/1.73 m 2 lower GFR values (95% CI: −3.55, −0.16), compared with the lowest tertile [37]. However, it is noteworthy the majority of Cd in blood is in red blood cells, which are not filtered (not present in glomerular filtrate) [2]. Consequently, it is impossible to attribute blood Cd to eGFR reduction and to Cd toxicity in the kidney in the absence of data on kidney pathology. A 2.91-fold increase in CKD risk (95% CI: 1.76, 4.81) was associated with blood Cd levels > 0.6 µg/L in the U.S. NHANES 1999-2006 adult participants [23]. Blood Cd levels > 0.53 µg/L were associated with an approximately two-fold increase in risk of CKD development (95% CI: 1.09, 4.50) among adult participants in the NHANES 2011-2012 [22]. In the Korean population study, elevated blood Cd levels, but not blood Pb or blood Hg, were associated with CKD, especially in those with hypertension [24]. Currently, urinary Cd threshold limit for CKD is lacking. However, there are several urinary Cd threshold limits that have been derived for kidney tubular toxicity using benchmark dose method [2,39]. In one study, urinary Cd 0.57-1.84 µg/g creatinine was identified as threshold levels for urinary β2-MG levels ≥ 1065 µg/g creatinine [14]. As discussed above, it was evident that GFR and CKD both were substantially associated with elevated urine β2-MG and that association of GFR and β2-MG was minimal (or absent) in subjects with urinary Cd in quartile 1 (urinary Cd 0.05-0.50 µg/g creatinine). In the absence of threshold limit for CKD and a continuously rising CKD prevalence worldwide, it is argued that urinary Cd of 0.50 µg/g creatinine might be useful. Urinary Cd of 0.50 µg/g creatinine is 2-fold and 10-fold lower than the threshold level for kidney Cd toxicity, established by the European Food Safety Agency [29] and the WHO/FAO [30], respectively. Urinary Cd levels < 1 µg/g creatinine have been found to be associated with kidney pathologies in many previous studies [3,4]. In a study of Swedish women, 53-64 years of age, urinary Cd of 0.67 µg/g and 0.8 µg/g creatinine were found to be associated with markers of tubular impairment and glomerular dysfunction, respectively [40]. Urinary Cd of 0.74 µg/g creatinine was associated with albuminuria in the Torres Strait (Australia) women who had diabetes [41].

Conclusions
For the first time, we have demonstrated that a clinical kidney function measure such as estimated glomerular filtration rates could be linked to both Cd exposure and tubular toxicity in Cd-dose and toxicity severity dependent manner. In addition, we have shown that a urinary Cd level as low as 0.50 µg/g creatinine might be used as a warning sign of excessive Cd intake, Cd toxic burden, kidney pathologies and kidney function deterioration. Urinary Cd of 0.50 µg/g creatinine is 10-fold lower than current threshold for kidney toxicity established by the FAO/WHO of 5.24 µg/g creatinine. This established urinary Cd threshold level does not afford health protection. Consequently, there is an urgent need to reassess Cd toxic burden and urinary Cd toxicity threshold limit that should prevent human population from excessive Cd exposure, and CKD development.

Strengths and Limitations
The strengths of this study include the samples of men and women with homogeneous exposure sources (i.e., none were occupationally exposed) together with a wide Cd-exposure range (urinary Cd 0.05-58 µg/g creatinine) and a wide eGFR range (19.6-137.8 mL/min/1.73 m 2 ) suitable for dose-response relationship analysis. The high CKD prevalence of 12.7% in villages with varying degrees contamination allowed recruitment of sufficient numbers of subjects with low GFR and CKD. The community-based recruitment strategy minimized bias toward certain subpopulation groups, frequently encountered in health center-based studies.
The limitations of this study were its small sample size and its cross-sectional design, which limited an assessment of temporal relationships between variables or causal inference of Cd exposure. A wide age range was another limitation as GFR falls with increasing age due to loss of nephrons [2]. GFR could also fall due to tubular pathologies induced by Cd and other environmental nephrotoxicants. Most subjects with high-Cd exposure were rice farmers, co-exposure to other nephrotoxicants in pesticides might also be a possible confounder. Heavy smoking, and presence of disease notably hypertension and diabetes were likely confounders. GFR may fall because of kidney damage due to smoking. This was evident in Figure 2, where an additional effect of smoking on eGFR was suggested by the increasing β slope in urinary Cd quartiles 3 and 4, relative to quartiles 1 and 2, given the higher prevalence of smokers in urinary quartile 3 (35.4%) and quartile 4 (32%), compared to quartile 2 (20.2%) and quartile 1 (12.4%).
GFR may also fall because of kidney damage due to hypertension, and because of nephron loss, urinary excretion of NAG in heavy smokers, hypertensive and diabetic subjects could be lower than expected. This was evident in Figure 3, where there was a marked drop in the β slope in quartile 4, compared with quartile 3. Such a drop in β slope could be interpreted to be resulted from loss of tubular cells, leading to lower urinary NAG excretion levels than expected in quartile 4.
Urinary Cd concentrations were determined by two methods. For low-Cd exposure group, a high sensitive and high specificity method, known as inductively-coupled plasma mass spectrometry, was used. A less sensitive, but sufficiently high specificity assay with atomic absorption spectrophotometer was used for high-Cd exposure group. However, data from quality control and assurance conduced with standard urine specimens suggest that variation due to different methods was relatively small.
Author Contributions: S.S., W.R. and M.N. designed study protocols. S.S. and W.R. obtained ethical institutional clearances for research on human subjects and supervised biologic specimen collection in Thailand. S.S., W.R. and M.N. supervised biologic specimen analysis in Australia and Japan. S.S., W.R., M.N., and P.R. analyzed and interpreted data. S.S., P.R. wrote and revised the manuscript.