*Article* **The NOAEL Equivalent of Environmental Cadmium Exposure Associated with GFR Reduction and Chronic Kidney Disease**

**Soisungwan Satarug 1,\*, Aleksandra Buha Ðordevi´c ¯ 2 , Supabhorn Yimthiang <sup>3</sup> , David A. Vesey 1,4 and Glenda C. Gobe 1,5,6**


**Abstract:** Cadmium (Cd) is a highly toxic metal pollutant present in virtually all food types. Health guidance values were established to safeguard against excessive dietary Cd exposure. The derivation of such health guidance figures has been shifted from the no-observed-adverse-effect level (NOAEL) to the lower 95% confidence bound of the benchmark dose (BMD), termed BMDL. Here, we used the PROAST software to calculate the BMDL figures for Cd excretion (ECd) associated with a reduction in the estimated glomerular filtration rate (eGFR), and an increased prevalence of chronic kidney disease (CKD), defined as eGFR <sup>≤</sup> 60 mL/min/1.73 m<sup>2</sup> . Data were from 1189 Thai subjects (493 males and 696 females) mean age of 43.2 years. The overall percentages of smokers, hypertension and CKD were 33.6%, 29.4% and 6.2%, respectively. The overall mean ECd normalized to the excretion of creatinine (Ecr) as ECd/Ecr was 0.64 µg/g creatinine. ECd/Ecr, age and body mass index (BMI) were independently associated with increased prevalence odds ratios (POR) for CKD. BMI figures <sup>≥</sup>24 kg/m<sup>2</sup> were associated with an increase in POR for CKD by 2.81-fold (*<sup>p</sup>* = 0.028). ECd/Ecr values of 0.38–2.49 µg/g creatinine were associated with an increase in POR for CKD risk by 6.2-fold (*p* = 0.001). The NOAEL equivalent figures of ECd/Ecr based on eGFR reduction in males, females and all subjects were 0.839, 0.849 and 0.828 µg/g creatinine, respectively. The BMDL/BMDU values of ECd/Ecr associated with a 10% increase in CKD prevalence were 2.77/5.06 µg/g creatinine. These data indicate that Cd-induced eGFR reduction occurs at relatively low body burdens and that the population health risk associated with ECd/Ecr of 2.77–5.06 µg/g creatinine was not negligible.

**Keywords:** benchmark dose; BMDL; BMDU; cadmium; creatinine clearance; chronic kidney disease; eGFR; NOAEL; urine cadmium

## **1. Introduction**

Environmental exposure to cadmium (Cd) is inevitable for most people because the metal is present in almost all food types [1–3]. The realization in the 1940s that the condition referred to as "itai-itai" disease was due to the consumption of rice heavily contaminated with Cd brought into focus the real threat to health posed by this metal [4,5]. Itai-itai disease is the most severe form of human Cd poisoning, characterized by severe damage to the kidneys and bones, resulting in multiple bone fractures due to osteoporosis and osteomalacia [4,5]. The pathologic symptoms of the itai-itai disease have been replicated in Cd-treated cynomolgus monkeys [6].

**Citation:** Satarug, S.; Ðordevi´c, A.B.; ¯ Yimthiang, S.; Vesey, D.A.; Gobe, G.C. The NOAEL Equivalent of Environmental Cadmium Exposure Associated with GFR Reduction and Chronic Kidney Disease. *Toxics* **2022**, *10*, 614. https://doi.org/10.3390/ toxics10100614

Academic Editors: Virgínia Cruz Fernandes and Diogo Pestana

Received: 22 September 2022 Accepted: 14 October 2022 Published: 15 October 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

To safeguard against excessive dietary Cd exposure, health guidance such as a tolerable intake level of Cd was established [7]. The Joint FAO/WHO Expert Committee on Food Additives and Contaminants (JECFA) considered the kidney to be the critical target of Cd toxicity [8]. By definition, the provisional tolerable weekly intake (PTWI) for a chemical with no known biological function is an estimate of the amount that can be ingested weekly over a lifetime without appreciable health risk. Subsequently, the PTWI for Cd was amended to a tolerable monthly intake (TMI) of 25 µg per kg body weight per month, equivalent to 0.83 µg per kg body weight per day [8]. This tolerable intake level for Cd was derived from a risk assessment model that assumed an increase in excretion of β2 microglobulin (β2M) (Eβ2M) above 300 µg/g creatinine as the point of departure (POD) [8]. However, we have shown that such an increase in Eβ2M reflected tubular dysfunction and nephron loss, evident from a reduction in estimated glomerular filtration rate (eGFR) to 60 mL/min/1.73 m<sup>2</sup> or below [9,10]. In effect, a tolerable intake level of Cd derived from the Eβ2M-based POD is not sufficiently low to be without an impact on human health.

Current evidence suggests that sufficient tubular injury disables glomerular filtration and leads to nephron atrophy and a decrease in GFR [11–13]. Accordingly, we argue that a reduction in eGFR due to Cd nephropathy could serve as the POD from which health guidance values should be derived. Owing to some shortcomings of the no-observed-adverseeffect level (NOAEL), the benchmark dose (BMD) has been used as the POD [7,14–16]. The BMD is a dose level, derived from an estimated dose–response curve, associated with a specified change in response, termed benchmark response (BMR) which can be set at 1%, 5%, or 10% as required [14–16].

The present study had two major aims. The first aim was to characterize a reduction in eGFR and risk factors of chronic kidney disease (CKD) in a sufficiently large group of people with a wide range of environmental Cd exposure. The risk factors considered included age, body mass index (BMI), smoking, hypertension, and Cd exposure measured as excretion of Cd (ECd). The second aim was to compute the lower 95% confidence bound of BMD (BMDL) and the BMD upper confidence limit (BMDU) of ECd associated with eGFR reduction and an increase in the prevalence of CKD.

#### **2. Materials and Methods**

#### *2.1. Participants*

To represent a large group of subjects with a wide range of environmental Cd exposure levels suitable for the dose–response analysis and health risk calculation, we assembled archived data from 1189 persons who participated in large population-based studies undertaken in a Cd contamination area in the Mae Sot District, Tak Province (*n* = 537), and low exposure locations in Bangkok and Nakhon–Si–Thammarat Province (*n* = 652). The Institutional Ethical Committees of Chulalongkorn University, Chiang Mai University and the Mae Sot Hospital approved the study protocol for the Mae Sot and Bangkok groups. The Office of the Human Research Ethics Committee of Walailak University in Thailand approved the study protocol for the Nakhon Si Thammarat group [17,18].

All participants gave informed consent prior to participation. They had lived at their current addresses for at least 30 years. Exclusion criteria were pregnancy, breastfeeding, a history of metalwork, and a hospital record or physician's diagnosis of advanced chronic disease. Because occupational exposure was an exclusion criterion, we presumed that all participants had acquired Cd from the environment. Diabetes was defined as fasting plasma glucose levels ≥ 126 mg/dL or a physician's prescription of anti-diabetic medications. Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, a physician's diagnosis, or prescription of anti-hypertensive medications.

### *2.2. Collection and Analysis of Biological Specimens*

Simultaneous blood and urine sampling are required to normalize ECd, to Ccr. Accordingly, second-morning urine samples were collected after an overnight fast, and whole blood samples were obtained within 3 hours after the urine sampling. Aliquots of urine, whole blood and plasma were stored at −20 ◦C or −80 ◦C for later analysis. The assay for urine and plasma concentrations of creatinine ([cr]<sup>u</sup> and [cr]p) was based on the Jaffe reaction.

For the Bangkok group, urine concentration of Cd ([Cd]u) was determined by inductivelycoupled plasma mass spectrometry (ICP/MS, Agilent 7500, Agilent Technologies, Santa Clara, CA, USA). Multi-element standards (EM Science, EM Industries, Inc., Newark, NJ, USA) were used to calibrate the Cd analyses. Quality assurance and control were conducted with simultaneous analyses of samples of the reference urine Lyphochek® (Bio-Rad, Gladesville, New South Wales, Australia), which contained low- and high-range Cd levels. A coefficient of variation value of 2.5% was obtained for Cd in the reference urine. The low limit of detection (LOD) of urine Cd was 0.05 µg/L. The urine samples containing Cd below the LOD were assigned as the LOD divided by the square root of 2 [19].

For the Nakhon–Si–Thammarat group, [Cd]<sup>u</sup> was determined with the GBC System 5000 Graphite Furnace Atomic Absorption Spectrophotometer (AAS) (GBC Scientific Equipment, Hampshire, IL, USA). Instrumental metal analysis was calibrated with multi-element standards (Merck KGaA, Darmstadt, Germany). Reference urine metal control levels 1, 2, and 3 (Lyphocheck, Bio-Rad, Hercules, CA, USA) were used for quality control, analytical accuracy, and precision assurance. The analytical accuracy of metal detection was checked by an external quality assessment every 3 years. The LOD of urine Cd was 0.1 µg/L. When [Cd]<sup>u</sup> was below its detection limit, the Cd concentration assigned was the detection limit divided by the square root of 2 [19].

For the Mae Sot group, [Cd]<sup>u</sup> was determined with AAS (Shimadzu Model AA-6300, Kyoto, Japan). Urine standard reference material No. 2670 (National Institute of Standards, Washington, DC, USA) was used for quality assurance and control purposes. The LOD of Cd quantitation, defined as 3 times the standard deviation of blank measurements was 0.06 µg/L. None of the urine samples from this group contained [Cd]<sup>u</sup> below the detection limit.

#### *2.3. Estimated Glomerular Filtration Rates (eGFR)*

The GFR is the product of nephron number and mean single nephron GFR, and in theory, the GFR is indicative of nephron function [20–22]. In practice, the GFR is estimated from established chronic kidney disease-epidemiology collaboration (CKD-EPI) equations and is reported as eGFR [21].

Male eGFR = 141 <sup>×</sup> [plasma creatinine/0.9]<sup>Y</sup> <sup>×</sup> 0.993age, where Y = <sup>−</sup>0.411 if [cr]<sup>p</sup> <sup>≤</sup> 0.9 mg/dL, Y = <sup>−</sup>1.209 if [cr]<sup>p</sup> > 0.9 mg/dL. Female eGFR = 144 <sup>×</sup> [plasma creatinine/0.7]<sup>Y</sup> <sup>×</sup> 0.993age, where Y = <sup>−</sup>0.329 if [cr]<sup>p</sup> <sup>≤</sup> 0.7 mg/dL, Y = <sup>−</sup>1.209 if [cr]<sup>p</sup> > 0.7 mg/dL. For dichotomous comparisons, CKD was defined as eGFR <sup>≤</sup> 60 mL/min/1.73 m<sup>2</sup> . CKD stages 1, 2, 3a, 3b, 4, and 5 corresponded to eGFR of 90–119, 60–89, 45–59, 30–44, 15–29, and <15 mL/min/1.73 m<sup>2</sup> , respectively.

## *2.4. Normalization of ECd to Ecr and Ccr*

E<sup>x</sup> was normalized to Ecr as [x]u/[cr]u, where x = Cd; [x]<sup>u</sup> = urine concentration of x (mass/volume); and [cr]<sup>u</sup> = urine creatinine concentration (mg/dL). The ratio [x]u/[cr]<sup>u</sup> was expressed in µg/g of creatinine.

E<sup>x</sup> was normalized to Ccr as Ex/Ccr = [x]u[cr]p/[cr]u, where x = Cd; [x]<sup>u</sup> = urine concentration of x (mass/volume); [cr]<sup>p</sup> = plasma creatinine concentration (mg/dL); and [cr]<sup>u</sup> = urine creatinine concentration (mg/dL). Ex/Ccr was expressed as the excretion of x per volume of filtrate [23].

#### *2.5. Benchmark Dose Computation and Benchmark Dose–Response (BMR) Setting*

We used the web-based PROAST software version 70.1 (https://proastweb.rivm. nl accessed on 13 October 2022) to compute the BMD figures for ECd/Ecr and ECd/Ccr associated with glomerular dysfunction. A specific effect size termed the benchmark response (BMR) was set at 5% for a continuous eGFR reduction endpoint and at 10%

for a quantal endpoint where eGFR <sup>≤</sup> 60 mL/min/1.73 m<sup>2</sup> . For a continuous endpoint, BMD values were computed from fitting datasets to four dose–response models, including inverse exponential, natural logarithmic, exponential, and Hill models. For a quantal endpoint, BMD values were calculated from fitting datasets to seven dose–response models that included two-stage, logarithmic logistic, Weibull, logarithmic probability, gamma, exponential and Hill models. The BMD 95% confidence intervals of ECd/Ecr and ECd/Ccr were from model averaging using bootstrap with 200 repeats.

The BMDL and BMDU corresponded to the lower bound and upper bound of the 95% confidence interval (CI) of BMD. The wider the BMDL-BMDU difference, the higher the statistical uncertainty in the dataset [23–26]. BMDL/BMDU figures of ECd for the glomerular endpoint were calculated for males, females and all subjects.

## *2.6. Statistical Analysis*

Data were analyzed with IBM SPSS Statistics 21 (IBM Inc., New York, NY, USA). The one-sample Kolmogorov–Smirnov test was used to identify departures of continuous variables from a normal distribution, and a logarithmic transformation was applied to variables that showed rightward skewing before they were subjected to parametric statistical analysis. The Mann–Whitney U-test was used to compare mean differences between the two groups. The Chi-square test was used to determine differences in percentage and prevalence data. The multivariable logistic regression analysis was used to determine the Prevalence Odds Ratio (POR) for CKD in relation to six independent variables; age, BMI, gender, smoking, hypertension and Cd exposure measures as ECd. We employed two models in each logistic regression analysis: model 1 incorporated log2(ECd/Ecr) or three ECd/Ecr groups; model 2 incorporated log2(ECd/Ccr) or three ECd/Ccr groups. All other independent variables in models 1 and 2 were identical. For all tests, *p*-values ≤ 0.05 for two-tailed tests were assumed to indicate statistical significance.

#### **3. Results**

#### *3.1. Characterization of Cadmium Exposure by Sex and Smoking*

Table 1 provides demographic data of participants (493 males and 696 females) stratified by sex and smoking status.

The overall mean age of participants was 43.2 years, and the overall percentages of current smokers plus those who had stopped smoking for less than 10 years, hypertension and low eGFR were 33.6%, 29.4% and 6.2%, respectively. The overall mean [Cd]<sup>u</sup> and mean ECd/Ecr were 0.94 µg/L and 0.64 µg/g creatinine, while the overall mean ECd/Ccr × 100 was 1.02 µg/L filtrate.

Smoking was higher among males (57.4%) than females (16.4%). In both sexes, % of smokers and non-smokers with hypertension did not differ. However, % of low eGFR among smokers was 3.7- and 3.8-fold higher than non-smokers in female and male groups, respectively. For the female group only, the mean BMI was 6 % lower in smokers than non-smokers (*p* = 0.004).

For the male group, the mean [Cd]<sup>u</sup> in smokers was 5.4-fold higher than nonsmokers (1.73 vs. 0.32 µg/L, *p* < 0.001). Mean ECd/Ecr and mean ECd/Ccr in smokers were 2.9- and 4.1-fold higher than in nonsmokers, respectively.

For the female group, the mean [Cd]<sup>u</sup> in smokers was 6.4-fold higher than nonsmokers (4.84 vs. 0.75 µg/L, *p* < 0.001). Mean ECd/Ecr and mean ECd/Ccr in smokers were 3.2-and 6-fold higher than in nonsmokers, respectively.


**Table 1.** Characteristics of participants stratified by sex and smoking status.

*n*, number of subjects; BMI, body mass index; eGFR, estimated glomerular filtration rate; Ex, excretion of x; cr, creatinine; Ccr, clearance of creatinine. <sup>a</sup> eGFR determined with Chronic Kidney Disease Epidemiology Collaboration (CKD–EPI) equations [20]; <sup>b</sup> eGFR of 90–119, 60–89, 45–59, 30–44, 15–29, and <15 mL/min/1.73 m<sup>2</sup> corresponded to CKD stages 1, 2, 3a, 3b, 4, and 5, respectively. <sup>c</sup> Ex/Ecr = [x]u/[cr]u; <sup>d</sup> Ex/Ccr = [x]u[cr]p/[cr]u, where x = Cd [23]. Data for age, eGFR and BMI are arithmetic means ± standard deviation (SD). Data for all other continuous variables are geometric means ± SD. Data for BMI are from 951 subjects; data for hypertension are from 917 subjects; data for all other variables are from 1189 subjects. For each test, *p* ≤ 0.05 identifies statistical significance, determined by Chi-Square test and Mann–Whitney U test for % differences and mean differences, respectively. Compared with non-smoking males \* *p* = 0.029–0.042, \*\* *p* = 0.001–0.006, \*\*\* *p* ≤ 0.001. Compared with non-smoking females, # *<sup>p</sup>* = 0.004, ## *<sup>p</sup>* = 0.001, ### *<sup>p</sup>* <sup>≤</sup> 0.001.

#### *3.2. Characterization of CKD Risk factors*

Table 2 provides the results of a logistic regression analysis where ECd/Ecr and ECd/Ccr were continuous variables, while age and BMI were categorical variables.

An independent effect on the POR for CKD was observed for ECd/Ecr, BMI and age (Table 2). Sex, smoking and hypertension were not associated with the POR for CKD. Doubling of ECd/Ecr was associated with an increase in POR for CKD by 1.47-fold (*<sup>p</sup>* < 0.001). BMI figures <sup>≥</sup> 24 kg/m<sup>2</sup> were associated with 2.81-fold increase in POR for CKD (*p* = 0.028). Compared with those aged 16–45 years, the POR values for CKD were 14-, 28- and 141-fold higher in those aged 46–55, 56–65, and 66–87 years, respectively.

In an equivalent analysis of the Ccr-normalized datasets, ECd/Ccr, BMI and age were independently associated with increased POR for CKD. Sex, smoking and hypertension were not associated with the POR for CKD. Doubling of ECd/Ccr was associated with an increase in POR for CKD by 1.96-fold (*<sup>p</sup>* < 0.001). BMI figures <sup>≥</sup> 24 kg/m<sup>2</sup> were associated with a 3.12-fold increase in POR for CKD (*p* = 0.022). Compared with those aged 16–45 years, the POR values for CKD were 10-, 35- and 199-fold higher in those aged 46–55, 56–65, and 66–87 years, respectively.



POR, Prevalence Odds Ratio; S.E., standard error of mean; CI, confidence interval. <sup>a</sup> CKD was defined as estimated glomerular filtration rate (eGFR) <sup>≤</sup> 60 mL/min/1.73 m<sup>2</sup> . Coding; female = 1, male = 2, hypertensive = 1, normotensive = 2, smoker = 1, non-smoker = 2. Data were generated from logistic regression analyses relating POR for CKD to six independent variables, listed in the first column. For all tests, *p*-values < 0.05 indicate statistical significance. Log2[(ECd/Ecr) <sup>×</sup> <sup>10</sup><sup>3</sup> ] was incorporated into model 1; log2[(ECd/Ccr) <sup>×</sup> <sup>10</sup><sup>5</sup> ] was incorporated into model 2. Other independent variables in models 1 and 2 were identical. β coefficients indicate an effect size of each independent variable on POR for CKD.

#### *3.3. Cadmium Excretion in Relation to the Risk of CKD*

Table 3 provides the results of a logistic regression analysis where age and BMI were continuous variables, while ECd/Ecr was a categorical variable in model 1, and ECd/Ccr was categorical in model 2.


**Table 3.** Dose–response relationship between cadmium excretion and the risk of chronic kidney disease.

POR, Prevalence Odds Ratio; S.E., standard error of mean; CI, confidence interval. <sup>a</sup> CKD was defined as estimated glomerular filtration rate (eGFR) <sup>≤</sup> 60 mL/min/1.73 m<sup>2</sup> . Coding; female = 1, male = 2, hypertensive = 1, normotensive = 2, smoker = 1, non-smoker = 2. Data were generated from logistic regression analyses relating POR for CKD to six independent variables listed in the first column. For all tests, *p*-values < 0.05 indicate statistical significance. Three ECd/Ecr categories were incorporated into model 1; three ECd/Ccr × 100 categories were incorporated into model 2. Other independent variables in models 1 and 2 were identical. β coefficients indicate an effect size of each independent variable on POR for CKD.

Age and BMI were independently associated with increased POR for CKD in both models 1 and 2. Compared with ECd/Ecr ≤ 0.37 µg/g creatinine (model 1), the POR for CKD was increased by 6.2- and 10.6-fold in those with ECd/Ecr values of 0.38–2.49 and ≥2.5 µg/g creatinine, respectively. Compared with ECd/Ccr ≤ 9.9 ng/L filtrates (model 2), the POR for CKD was increased by 4.4- and 20.8-fold in those with ECd/Ccr values of 10–49.9 and ≥50 ng/L filtrate, respectively.
