Are the Guidelines for Dietary and Workplace Exposure to Cadmium Adequate?

1. Introduction

Exposure to pollution from environmental cadmium (Cd) has been linked to worldwide rising prevalences of ill-health conditions and malignant and non-malignant diseases [1,2,3]. Hypertension [4] and iron deficiency anemia [5] are the poor health conditions, while osteoporosis [6,7,8], type 2 diabetes [9,10] and chronic kidney disease (CKD) [11,12] are consistently found to be associated with environmental Cd exposure.

CKD is a highly prevalent disease, affecting 8-13% of adult population worldwide, reaching epidemic proportion in many parts of the world [1,2]. It is diagnosed when the estimated glomerular filtration rate (eGFR) falls to one-third of a normal range and/or the presence of albuminuria for at least 3 months [1,2]. A fall of eGFR at a high rate (≥ 3 mL/min/1.73 m² per year) has been causally linked to environmental Cd exposure in a prospective cohort study from Switzerland [13]. Using the Bradford-Hill Criteria, Hsu et al. presented evidence that strongly supports causal relationships between Cd exposure and the disease of kidneys and bones in the general population [3].

Cd is a cumulative toxicant because of its excretion rate is miniscule, resulting in the long half-life, varying from 7.4 to 30 years [14,15,16]. It preferentially accumulates within the kidney tubular cells, and is released into the urine as tubular cells die from any cause [17]. For this reason, excretion of Cd is used reliably as a quantitative measure of cumulative exposure or the body burden of the metal [17,18]. Cd can cause damage to its target cells via multiple mechanisms, like oxidative stress plus depleting antioxidant defenses [1,2]. It can induce also functional iron (Fe) and zinc (Zn) deficiencies [5,13,19,20].

Concerningly, however, the health risk of Cd exposure has long been underappreciated as the result of using an inappropriate toxic endpoint in health risk estimation. Most frequently, an elevation in β₂-microglobulin (β₂M) excretion was used for such purposes [21,22]. As another concern, at-risk subpopulations, namely children, women of childbearing age, and those with marginal Zn intake levels and low body Fe stores, have not been considered in computing the Cd health risk. Because of enhanced intestinal absorption of metals and immaturity of the blood brain barrier, neonates, infants, and young children are more susceptible than adults to adverse effects of food-born toxicants [23,24]. Moreover, relative to their body size, an amount of food intake is typically larger than that of adults; consequently, they are exposed to a higher dose of ingested toxicants [25,26].

Through the interference with mother-to-fetus Zn transport, maternal Cd exposure has been associated with infant low-birth weight and neurodevelopment deficits even though the passage of Cd from mother to fetus is minimal [27,28,29,30]. The ability of Cd to inhibit calcium (Ca) secretion by mammary glands has also been observed [31] along with the secretion of Cd into breast milk [31,32,33], raising concern about the potential for an early life Cd exposure, which can impact disease risk in adulthood [34,35,36] (Figure 1).

At present, around 15% of the world’s cultivation soils is contaminated with toxic metals, especially Cd, which is particularly prevalent in south and east Asia and parts of the Middle East and Africa [37]. Hence, there is a widespread concern that dietary Cd exposure in the population will gradually be rising because Cd, being non-biodegradable, can persist indefinitely in the soils, facilitating their entry into the human food chain and their ubiquitous presence in our diet, especially staple foods. As the Cd soil-to-plant transference continues as does human exposure to the metal in the diet, the prevalence of Cd-related diseases will eventually reach an epidemic proportion.

The present review provides information relevant to public health policy regarding “safe’ Cd exposure levels that are much lower than previously estimated. It argues strongly for a paradigm shift of the criteria on which toxicity of Cd is based. It highlights the inadequacy of existing food standards and exposure guidelines and underscores a conceptual flaw in using the β₂M excretion in toxicological risk assessment practice.

A special emphasis is given to the use of advanced benchmark dose (BMD) modeling [38,39,40], especially in determining the BMD limit (BMDL) values, using eGFR decline as an endpoint [3,41,42]. Cd-induced eGFR loss reliably indicates CKD and its progression toward kidney failure. The advanced BMD modeling overcomes some shortcomings of the conventional toxicological risk assessment, involving identification a point of departure (POD), from a dose-response curve [43,44].

Additionally, mechanism-based mitigation for Cd toxicity to its target cells, involving exogenous heme oxygenase -1 (HO-1) inducers is explored as potential strategies to delay CKD progression.

2. Cadmium Exposure Limits and Toxicity Thresholds

In this Section, environmental standards for Cd are highlighted along with a summary of toxicological risk assessment methodology that was used to define “safe” exposure level. Presumably, intake of Cd from drinking water was insignificant, given that Cd levels in most portable water is below standard by Australia/New Zealand at 2 µg/L and US EPA and WHO at 5 µg/L [43]. Therefore, exposure limits were determined based on its existence in the human diet, a principal source of Cd for most people. For workers, exposure limits were determined based on inhaled Cd that can enter the circulation from the lungs [43]. Figure 2 depicts pathways for Cd in food to its targets, e.g., bones and the proximal tubular cells of kidneys.

2.1. Food and Environmental Safety Standards

Cd is present in most foodstuffs; as such foods that are frequently consumed in large quantities are the major sources of dietary Cd [1,2]. The cumulative feature of Cd means that a safe exposure level may not exist; consequently, Cd levels in food crops and feeds should thus be set at the lowest level achievable e.g., 0.05-0.10 mg/kg for rice, potatoes, and wheats. However, the permissible maximum level (ML) of Cd in rice set by the Codex Alimentarius was as high as 0.4 mg/kg dry grain weight [53], while European Food Safety Authority (EFSA) set the ML for Cd in rice at 0.2 mg/kg [54].

Potatoes contributed most to Cd intake by 1- and 2-year-old Dutch children [55]. In Portugal, bread was identified as the main dietary source of Cd [56]. Rice contributed up to 90% of Cd exposure in an area of Vietnam with Cd contamination [57]. Rice and its products contributed 40-50% to dietary Cd exposure among women living in Cd-contaminated areas of Japan [58].

Data from the Chinese National Diet and Nutrition Survey and the National Food Contamination Monitoring Program supported the ML for rice below 0.2 mg/kg [59]. Severe damage to kidneys and bones as those in Itai-Itai disease patients may have occurred following ingestion of rice with Cd content of 0.27 mg/kg [60].

2.1. Toxicity Threshold Definition

A toxicity threshold is referred to as the highest dose that does not produce an adverse effect in the organ most sensitive to a health hazardous substance of concern [43,44]. Thus, for a food contaminant, like Cd, that impacts multiple organs, a health-protective exposure guideline must be based on the most sensitive endpoint.

2.2. Official Dietary Exposure Guidelines

“Safe” dietary Cd exposure levels were derived by various international agencies, including the Food and Agriculture Organization and World Health Organization (FAO/WHO) Joint Expert Committee on Food Additives and Contaminants (JECFA) [61,62], EFSA [63], the US Food and Drug Administration (US FDA) [64] and the Agency for Toxic Substances and Disease Registry (ATSDR) [43].

Notably, however, different terms were used to describe the acceptable dietary exposure levels of Cd; nevertheless, most countries have adopted JECFA’s tolerable daily intake (TDI) and a threshold at urinary Cd excretion rate of 5.24 μg/g creatinine (Table 1).

2.2.1. JECFA/TDI

JEFCA described exposure guideline for any food contaminant as a provisional tolerable weekly intake (PTWI), meaning an estimate of the amount of a chemical with no intended function that can be ingested weekly over a lifetime without appreciable health risk [61,62]. Original JECFA’s PTWI for Cd was 400–500 µg/person/wk, revised to 7 µg/kg bw/wk [53] before being 25 µg/kg bw/m (TMI), equivalent to 0.83 μg/kg bw/d(TDI) [62]. The PTWI, TMI, and TDI values for Cd were based on a permissible lifetime intake of 2 g of Cd, and β₂M excretion rate at 300 µg/g creatinine as a cut-off value for departure from normalcy.

Food Safety Commission of Japan has adopted the TWI for Cd at 7 µg/kg bw/wk (doi: 10.14252/foodsafetyfscj.D-24-00011) for the reason that an estimate of dietary Cd intake in Japan in 2022 was 2.03 μg/kg bw/wk, 29% of a regulatory level at 7 μg/kg bw/wk.

The JECFA’s assumption of “safe” lifetime intake level of Cd of 2 g was challenged by a prospective cohort study showing that a lifetime Cd intake of 1 g may have caused a 49% increase in mortality from kidney failure among women, who were residents of Cd-contaminated area of Japan [65]. This finding was adjusted for potential confounders.

The proportion of Portugal population, 18-74 years of age with Cd exposure levels higher than JECFA’s TDI was 5.4% even though mean dietary Cd intake was low (0.19 µg/ kg b.w./d) [56]. In comparison, mean dietary Cd intake in Chinese adult population was 34.3 µg/day (range: 22.6-54 μg/d) of which 15.4% had dietary Cd exposure levels exceeding JECFA’s TDI [66]. Fungi and algae had the highest Cd contents, followed by aquatic foods, nuts, cereals, beans, vegetables, meats, eggs, milk, and fruits.

Like Cd, lead (Pb) is another ubiquitous food contaminant for which PTWI was derived as 25 μg/kg bw/wk [61]. However, this guideline was withdrawn because no toxicity threshold level was identified for the neurotoxicity endpoint; consequently, no intake amount of Pb carries a negligible health risk [62].

2.2.2. EFSA/RfD

EFSA considered the kidneys to be the critical Cd toxicity target, and used β₂M excretion at a rate of 300 µg/g creatinine as a cut-off point, similar to the JECFA PTWI model. However, EFSA designated Cd excretion at a rate of 1 μg/g creatinine as a threshold level after an uncertainty factor (a safety margin) was included in a model to compensate for the variation in dietary Cd exposure levels among people [63]. EFSA designated dietary exposure to Cd at 0.36 μg/kg bw/d for 50 years as an acceptable dietary Cd exposure level and described it as a reference dose (RfD) [63].

2.2.3. US FDA/TRV

Dietary Cd exposure limits derived by US FDA ranged between 0.21 and 0.36 μg/kg bw/d, assuming Cd excretion at a rate of 0.5 μg/g creatinine as a threshold level for both bone and kidney targets [64]. These US FDA’s acceptable exposure levels for Cd were called toxicological reference values (TRV), obtained through reverse dosimetry methodology with the physiologically-based pharmacokinetics model [64]

Among US children aged 1–6 years, the food groups contributing most to Cd exposure were grains/baking, dairy and fruit and grains/baking and vegetables [67]. Respective mean value and 90th-percentile level for dietary Cd exposures were 0.43 and 0.71 μg/kg bw/d, both were higher than US FDA’s TRV but were within the JECFA’s TDI.

Because a threshold could not be determined for Pb neurotoxicity, discussed earlier, US FDA has proposed a dietary Pb intake level of 12.5 μg/d as an interim dietary exposure guideline for the general population of adults [68]. This Pb exposure rate corresponds to a blood Pb concentration of 0.5 μg/dL, which has not been found to be associated with an adverse effect in adults in any epidemiologic studies [69].

2.2.4. ATSDR/MRL

ATSDR derived Cd exposure guidelines, known as minimal risk levels (MRL) using data from experimental animal dosing regimens [43]. With the bone target, an MRL of 0.5 µg/kg bw/d was obtained for oral Cd exposure in an intermediate exposure duration (15-365 d). Data were from Wistar rats exposed to Cd as CdCl₂ in drinking water at 0, 1, 5, or 50 mg/L for 6, 9, or 12 months [70,71,72].

With the lung target, an MRL value of Cd as Cd oxide (CdO) of 0.03 μg/m³ was obtained for acute inhalational exposure duration between 1 and 14 days [43]. Data were from Fisher F344 rats exposed to CdO at 0, 0.1, 0.3, 1, 3, or 10 mg CdO/m³ for 6.2 h/, 5 d/wk, for 2 wks [73].

2.3. Other Exposure Limits and Thresholds Derived for Oral Cd

Reported Cd exposure limits and thresholds are abundant and they are highly variable, depending on methodology, demographic characteristic of study populations, toxicity targets and endpoint measures. A few studies are summarized below.

Sweden: Using data from 794 Swedish women, aged 53-64 years, Suwazona et al. identified urinary Cd excretion rate of 1 µg/g creatinine as a threshold for a disease (osteoporosis) endpoint [74]. This Cd excretion rate is the same as EFSA’s threshold for kidney effects (β₂M endpoint) [63].

France: Leconte et al. derived oral Cd of 0.35 μg/kg bw/d as an exposure limit figure for French population with Cd excretion at a rate of 0.5 μg/g creatinine being a threshold for adverse bone effects [75].

China: Qing et al. derive dietary Cd exposure limits, using Chinese population data with average Cd exposure of 30.6 μg/d. For bone target, they identified Cd exposure of 0.64 μg/kg bw/d as TDI and Cd excretion of rate 1.71 μg/g creatinine as a threshold [76]. For kidney target, Cd exposure of 0.28 μg/kg bw/d (16.8 µg/d for a 60 kg person) as TDI figure and Cd excretion rates of 3.07 and 2.93 μg/g creatinine were threshold levels for the β2M and NAG endpoints, respectively [77].

Experimentation: Wu et al. derived human TDI value for oral Cd exposure of 0.2 μg/kg bw/d [78], using data from inbred pigs, exposed to varying amounts of Cd at 0, 0.5, 2, 8, or 32 mg Cd/kg of feed for 100 d. This TDI was obtained after inclusion of an uncertainty factor of 100 to extrapolate data from pigs to humans. In this empirical study, excretion of several biomarkers of kidney effects were quantified along with β₂M. Unexpectedly, abnormal β₂M excretion occurred at the highest Cd feeding dose, while abnormal excretion of retinal binding protein (RBP) was observed at the lowest Cd-dose level. The oral Cd dose levels resulting in abnormality in excretion rate of RBP, N-acetyl-β-D-glucosaminidase (NAG), Cd complexed with metallothionein (CdMT), and β₂M were 0.67, 0.88, 1.00, and 3.08 mg/kg feed, respectively. Thus, an increase in β₂M excretion was least sensitive to Cd, compared to RBP and NAG, casting considerable doubts in its use as a basis to derive Cd exposure limits.

2.4. Cadmium Inhalational Exposure Limits for Workers

Similarly, the kidney target and β₂M endpoint are used in assessment of health risk from workplace exposure, which involves mostly an inhalational route [79]. As revealed by studies from Japan [80,81] and Korea [82], workers’ exposure limits at blood Cd concentration of 5 µg/L, and Cd excretion at a rate of 5 μg/g creatinine did not provide sufficient protection against the Cd toxicity to kidneys. The authors concluded that current workers’ Cd exposure limits should be lowered, while monitoring and management of exposures among workers remain to be necessary [80,81].

Nogawa et al. [80] used data from 326 male and 114 female Japanese workers and they estimated the BMDL value for a 40-year cumulative inhalational exposure to Cd to be 17.7 µg/m³. In addition, they found warning blood Cd concentration to be between 1.8 and 2.0 µg/L, less than half of current exposure limit of 5 µg/L [80]. Hoshino et al. analyzed data from 238 workers of two nickel-Cd battery plants in Japan; they observed that the risk for abnormal β₂M excretion was increased 17% even though the geometric mean for blood Cd among workers was 1.97 μg/L [81]. This blood Cd level associated with abnormal β₂M excretion in workers was like those reported by Nogawa et al. [80].

Choi et al. analyzed data from Korean workers of a small-scale silver soldering company who were exposed to air Cd concentrations of 6-15 µg/m³ [82]. They observed alarmingly high Cd excretion rates [mean (range) of 22.15 (3.23-62.97) μg/g creatinine] together with elevated urinary concentrations of β₂M and total protein [82]

In a case report, an Indian jewelry male worker developed hypophosphatemic osteomalacia following exposure to a high dose of Cd in fumes [83,84]. His blood Cd level was 6 times higher than the permissible workplace exposure level at 5 µg/L, while his 24 h urinary Cd excretion was 51 µg [83]. He also had hypochromic microcytic anemia, most likely from Cd-induced functional iron deficiency [5,20,85] plus an elevation in circulating levels of the bone derived hormone, fibroblast growth factor 23 (FGF23), which suppresses erythropoietin synthesis in the kidneys, while reducing kidney tubular reabsorption of phosphate [86,87].

2.5. Summary on Official Cd Exposure Limits

Current practice to evaluate adverse effects of environmental Cd exposure employed bone mineral density and β₂M excretion rate as toxicity endpoints for respective bone and kidney targets. Accordingly, official dietary Cd exposure guidelines vary from 0.21 to 0.83 μg/kg bw/d, with toxicity threshold levels of Cd excretion ranging between 0.5 and 5.24 µg/g creatinine. Data from metal workers [80,81,82] and an empirical feeding study [78] have questioned the utility of β₂M excretion to define exposure limit for Cd.

Risk of osteoporosis in postmenopausal women rose 1.95-fold and 1.99-fold, respectively, comparing Cd excretion ≥ 0.5 μg/g creatinine versus < 0.5 μg/g, and ≥ 5 μg/g creatinine versus < 5 μg/g, respectively [88]. A nearly two-fold increase in risk of having osteoporosis in both low- and high-Cd exposure groups means that a threshold for bone toxicity of Cd may not exist.

Woo et al. analyzed data from 13 publications that reported Cd excretion threshold levels to be 4.88, 3.13, and 1.9 μg/g creatinine, depending on the β₂M excretion cut-off values [89]. Excretion rate of Cd of 1.9 μg/g creatinine was a threshold level at β₂M excretion rate as high as 400 µg/g creatinine [89]. These data challenge the JECFA’s threshold figure for Cd at 5.24 µg/g creatinine which was based on a cut-off value of β₂M excretion at 300 µg/g creatinine.

Another notable result from a dose-response study in pigs is that Cd excretion itself, can, indeed, reflect its nephrotoxicity [78]. Urinary Cd complexed with MT, denoted as CdMT emanated from the proximal tubular cells (PTCs) of the kidneys following cell death from any cause, including Cd-induced ferroptosis [17]. Urinary CdMT was misinterpreted as the filtered CdMT, which was not reabsorbed by PTCs; and consequently, it is excreted in urine. Such misconception and its consequences are discussed in publication by Thévenod and Lee along with the cutting-edge knowledge on Cd nephrotoxicity [90].

3. Cd Excretion Threshold Levels Based on Disease Endpoints

Data from studies on Cd exposure and adverse health effects in the general populations and metal workers (Section 2) have repeatedly indicated the inadequacy of existing food standards and official exposure guidelines to protect populations’ health.

In this section, derivation of the critical exposure level/thresholds, known as benchmark dose (BMD) limit (BMDL) is discussed, focusing on kidney disease manifestation of Cd exposure in comparison to nephrotoxicity indicators, like β₂M, NAG and RBP.

3.1. Excretion of Low-Molecular Weight Proteins: Tubular Proteinuria

Excretion of β₂M and other low-molecular-weight proteins, namely α1-microglobulin (α1M), and RBP, have been used to indicate nephrotoxicity of Cd [1,2]. These proteins with the molecular weight <20 kDa, readily pass through the glomerular membrane into tubular lumen, and is reabsorbed, and degraded within kidney proximal tubular cells [91].

In theory, the excretion rate of any of these proteins is a function of its synthesis, glomerular filtration rate, kidney tubular reabsorption, and degradation. Thus, a rise in its production for any reason, a fall of eGFR due to nephron loss, defective tubular reabsorption or degradation all can lead to an increase in its excretion (Figure 3).

A close examination of the parameters influencing β₂M homeostasis in Thai subjects, exposed to low levels of Cd, has revealed that β₂M excretion could not be used as a measure of tubular dysfunction [92]. The variation among people in the influx of β₂M from cells into plasma is so large that β₂M excretion is minimally related to its reabsorption and degradation by kidney tubules. There is no basis to estimate permissible dietary Cd exposure levels using the β₂M excretion at a rate as high as 300 µg/g cr.

In Cd feeding trials [78], elevated excretion levels of RBP, NAG and Cd were observed at Cd at 0.67, 0.88, 1.00, and 3.08 mg/kg feed, respectively. Thus, the Cd-dose levels, causing β₂M excretion to rise were 4.6-, 3.5- and 3.1-fold higher, compared to the dose levels inducing the increment in excretion rates of RBP, NAG and Cd, respectively. Increased excretion rates of RBP, NAG and Cd all appear to occur at the Cd body burden lower than those induced an increase in β₂M excretion, thereby casting considerable doubt on TDI, RfD and TRV figures, derived from the β₂M endpoints. The mean dietary Cd exposure among Australian children, aged 8 years, was 60% of JEFCA’s TDI, while exceeding the RfD for Cd exposure derived by EFSA [93].

3.2. From NOAEL to BMDL

The practice of toxicological risk assessment involves determining from a dose-response curve, a point of departure (POD), which represents the dose at which a specific adverse effect is first observed, or at which a response deviates from a baseline or control [43,44]. POD servs as a starting point to evaluate the potential health impact of exposure to a health hazardous substance.

Typically, the POD figure is determined from a dose-response curve, constructed from experimental dosing, which often involves daily administration of 4-5 different doses for 90 days or longer [94,95,96]. From a dose-response curve, a POD may be established from the lower bound “no-observed-adverse-effect level” (NOAEL) and the upper bound “lowest-observed-adverse-effect level” (LOAEL).

The NOAEL value is referred to as the highest dose tested that produces an insignificant effect, compared to controls. To convert NOAEL to the BMD lower limit (BMDL), the NOAEL is divided by an uncertainty factor which accounts for species differences and human variability [43,44]. The use of uncertainty factor of 100 in the extrapolation of data from pig dosing experimentation obtained Cd exposure limit of 4-time lower than the JECFA’s figure [78] (Section 2.2.5).

A clear dose-response relationship is prerequisite to estimate the NOAEL(BMDL) figure for exposure to any health hazardous substance. However, reliance on a single mathematical dose–response model (equation), such as Hill model, can lead to erroneous estimates of the NOAEL(BMDL) values of Cd excretion as is the use of cut-off values to define abnormal excretion rates of nephrotoxicity biomarkers.

3.3. Multiple Mathematic Dose-Response Models: The Akaike Information Criterion (AIC)

Advanced BMD method involves fitting an entire exposure–effect dataset to multiple mathematical dose–response models, in which a specific effect size, termed benchmark response (BMR), is pre-defined [43,44,96]. In the PROAST BMD software, four mathematical dose–response (MDR) models and ten MDR models are used for fitting continuous dose-effect data and dichotomous data, respectively [38,39,40,96].

The Akaike information criterion (AIC) is used to evaluate how well each individual dose–response model fits the data [96]. The AIC assesses both goodness of fit and the model complexity; as such each model is weighted relative to an amount of information lost; the higher the weight, the better data fitting. The shape and steepness of the slope provide additional insight into exposure-effect pairs [38,39,40].

The BMDL value of any exposure indicator is defined as the lower 95% confidence bound of the BMD, computed at a 5% BMR. The BMDL value derived in this manner has replaced NOAEL, which can be the representative of a critical exposure level [44,96].

The difference between the lower bound (BMDL) and upper bound (BMDU) of the 95% confidence interval (CI) of the BMD reflects the statistical uncertainties in the BMD estimates. A narrow difference indicates a high degree of certainty of the estimated BMD figures. Conversely, a wide difference, e.g., a BMDU/BMDL ratio ≥ 100, indicate unreliable BMD estimates [44,96].

3.3.1. Comparing Cd excretion benchmarks for the β₂M and other endpoints

A population of people, exposed to a wide range of Cd doses is ideal to establish a clear dose–response relationship, from which benchmark Cd excretion levels can be identified with certainty from various nephrotoxicity endpoints. The Mae Sot District in western Thailand is a geographic area with endemic environmental Cd pollution [97] that has provided a well-circumscribed population of people with the same level of exposure that would enable one to discern the health impact of excessive dietary Cd exposure.

The Cd concentration of the paddy soil samples from the Mae Sot District exceeded the Thailand standard of 0.15 mg/kg, and the rice samples collected from household storage contained four times the amount of Thailand permissible Cd level of 0.1 mg/kg [98]. An inverse association was observed between excretion of β₂M and eGFR only in those with eGFR values within the CKD signified range, and risk of CKD rose 4.7-fold as β₂M excretion rose from 100 to 300 μg/g cr [99]. Thus, an elevation of β₂M excretion could be indicative of nephron loss which caused a fall in eGFR [99]. In a dose–response analysis, CKD risk rose 4.7-fold, 6.2-fold, and 10.5-fold at β₂M excretion rates of 100–299, 300–999, and ≥ 1000 μg/g cr, respectively [100].

To reassess the Cd excretion threshold based on β₂M excretion, in comparison with the eGFR reduction, Satarug et al. employed data from 799 Thai nationals, 18–87 years of age, selected from a cohort (n = 1189) to which both residents of low-exposure areas and moderate-to-high exposure locations of the Mae Sot District were enrolled [101].

Cd excretion rates among cohort participant ranged between 0.03 and 106 µg/g creatinine (geometric mean 2.15 µg/g creatinine). Age and BMI distributions conformed to a normal distribution [42].

The dose-response curves and behcmark Cd excretion rates using β₂M and eGFR as endpoints are providded in Figure 4 and Figure 5, respectively.

Cd excretion benchmark could not be reliably determined, when a 5% increase in β₂M excretion was an endpoint, evident from the BMDU/BMDL ratio > 100 (Figure 4). In comparison, the benchmark Cd excretion rate of 0.17 µg/g creatinine was obtained, when a 5% reduction in eGFR is used as an endpoint measure (Figure 5). The BMDU/BMDL ratio of the benchmark Cd excretion was 16.9, meaning a high degree of certainty in the estimates. The use of the eGFR endpoint is recommended for future deriving health-protective exposure limits for C because eGFR decline at a high rate signifies kidney disease.

3.3.2. Comparing Benchmark Cd Excretion Rates Derived from Human Population Data

Table 2 provides Cd excretion rates identified as Cd exposure benchmarks for the general Chinese, Japanese and Thai population.

Conventional BMD: Wang et al. employed a conventional BMD method to identify the BMDL values of Cd excretion at 5% and 10% prevalences of abnormal excretion of RBP, NAG and β₂M [102]. They used data were from 934 (469 men, 465 women), aged 10–71 years who were residents of Jiangshan City, Zhejiang, China [102]. For men, BMDL₅ (BMDL₁₀) values of Cd excretion were 0.89 (1.59), 0.62 (1.30), 0.49 (1.04) μg/g creatinine for the RBP, β₂M, and NAG, respectively. Corresponding benchmark Cd excretion rates in women were 0.76 (1.53), 0.64 (1.34), 0.65 (1.37) μg/g creatinine.

Suwazono et al. identified Cd excretion rates of 0.6–1.2 and 0.6–2.3 µg g creatinine as BMD values for abnormal β₂M excretion in men and women, respectively [103]. They used data from 410 men and 418 women, aged 40–59 years, who lived in the areas of Japan without Cd pollution [103]. The lower BMD estimates for abnormal β₂M excretion in Japanese men and Japanese women were close to BMDL₅ values of Cd excretion in Chinese study [102].

Advanced BMD: Satarug et al. used advanced BMD method to determine Cd excretion benchmarks with NAG excretion increase by 5% [104]. They used data from 734 Thai nationals (289 men and 445 women), 18-87 years of age (mean 48.1 years). Cd excretion benchmarks with 5% increase in NAG excretion in men and women were 0.060 and 0.069 µg/g cr, respectively.

Using β₂M excretion rates ≥ 300 μg/g creatinine to define abnormality, the BMDL₁₀ values of Cd excretion were 0.469 and 0.733 µg/g creatinine in men and women, respectively. A higher Cd excretion benchmark in women was a result of their universally smaller muscle mass than men; consequently, they have lower creatinine excretion rates than men. The Thai BMDL₁₀ values of Cd excretion for β₂M endpoint were 36-55% lower, compared to a Chinese study, reporting the BMDL₁₀ values of 1.30 and 1.34 µg/g cr in men and women, respectively [103]. This may be due to differences in age profiles of target population and the cut-off value to define an abnormal β₂M excretion in additional to shortcomings of reliance on a single dose-response model in conventional BMD practice [106,107].

Advantage of using seven dose-response models to determine BMDL5, BMDL10 values from dichotomous/disease prevalence data is illustrated in Figure 7.

Satarug et al. used advanced BMD method to determine Cd excretion benchmarks at 5% prevalence of proteinuria and 5% prevalence of CKD [105]. They used data from 405 Thai subjects (197 men and 208 women), aged 19-87 years (mean 44.6 years). Protein excretion rate of 100 mg/g cr. was a cut-off figure to define proteinuria, while eGFR value ≤ 60 mL/min/1.73 m² was a cut-off value to designate CKD development. The BMDL₅ value of Cd excretion at 5% prevalence of proteinuria was 1.86 µg/g creatinine while the BMDL₅ of Cd excretion at 5% prevalence of CKD was 1.19 µg/g creatinine.

The BMDL value of Cd excretion at 5% CKD prevalence was 36% lower, compared to the figure for 5% proteinuria prevalence. Apparently, a falling eGFR was a more sensitive indicator of Cd effects than proteinuria. Even a slight increase in Cd excretion (a body burden indicator of Cd) can induce a large drop in eGFR because of the relationship between eGFR and Cd excretion was exponential (Figure 7D). This finding is consistent with a prospective cohort study from Switzerland that causally linked a fall of eGFR at high rate (≥3 mL/min/1.73 m²/y) to Cd exposure [108].

Because proteinuria predicts continued eGFR deterioration [109,110,111], its clinical relevance is apparent. A health survey reported the prevalence of severe proteinuria (excretion of protein ≥ 200 mg/g cr) among residents of Mae Sot district to be as high as 24.1% with 66.7% of them had Cd excretion rates ≥ 2 µg/g cr [97].

The prevalence of proteinuria rose from 5% to 10% as Cd excretion rates rose from 1.86 to 4.47 µg/g creatinine. In comparison, the prevalence of low eGFR rose from 5% to 10% as Cd excretion rates increased from 1.19 to 1.35 µg/g creatinine; a 1.13-fold increase in Cd exposure resulted in 5% more people with CKD. In effect, the prevalence of CKD (low eGFR) was a more sensitive marker for Cd nephrotoxicity than proteinuria.

Collectively, benchmark Cd excretion data suggested that Cd-induced nephron destruction that causes eGFR to fall occurred at Cd body burden lower than those induced an increase in protein excretion to a rate of ≥100 mg/g cr and a rise in β₂M excretion rate above 300 µg/g cr.

3.4. Summary on Cd Toxicity Thresholds for the Kidney Target

Because the cytotoxicity of Cd involves the same basic biochemical mechanism, an amount of Cd causing such toxicity can be expected to be similar across human populations. Accordingly, the BMDL (NOAEL equivalent) values of Cd excretion rate, estimated for various populations, using the same endpoint, in theory, should be comparable.

As enlisted in Table 2, for the nephrotoxicity endpoint, the lowest BMDL₅ value of Cd excretion rate in a Chinese study using the NAG endpoint was 0.49 µg/g cr. In a Thai study using the NAG excretion ≥ 4 U/g cr as a cut-off value, the lowest BMDL₁₀ of Cd excretion was 0.469 µg/g cr. The lowest Cd excretion benchmark identified in a Japanese study from the β₂M endpoint was 0.6 µg/g cr. These Cd excretion benchmarks are approximately 9-12% of the JECFA’s threshold.

Using a 5% decrease in the eGFR, a clinical measure of kidney function as an endpoint, the benchmark Cd excretion was 0.17 µg/g cr. The benchmark Cd excretion at 5% increases in protein excretion was 0.054 µg/g cr. In contrast, the Cd excretion benchmark could not be reliably determined, when 5% increase in β₂M excretion was used as an endpoint. Hence, the eGFR/CKD endpoint provides a logical basis to derive new exposure guidelines for dietary Cd exposure. In addition, consideration (safety margin) should be given to also to factors influencing internal doses (blood Cd concentrations), and thus toxic manifestation of Cd [112,113,114].

4. Mechanism-Based Strategies to Mitigate Cd Toxicity

In this Section, the current view around how Cd reaches the kidney proximal tubular cells and mitochondrial target is highlighted, while accentuating the potential role for albumin in delivery of Cd to kidney tubules. A summary of the results from recent investigations advancing our knowledge on how Cd induces tubular cell death in chronic low-dose exposure conditions in humans together with the strategies that can be developed to mitigate its toxic manifestations.

4.1. Why Most Acquired Cadmium Accumulates in the Proximal Tubules?

As depicted in Figure 2, Cd is absorbed by enterocytes via several transport proteins for Fe, Zn, Ca, and Cu. Similarly, Cd enters osteoblasts via multiple metal transport proteins and voltage-gated Ca²⁺ channels [115].

In distinction from enterocytes and osteoblasts, the kidney tubular cells are well equipped with the mechanisms to reabsorb proteins, namely receptor-mediated endocytosis (RME) [116,117,118], additional to those for metals, glucose, amino acids and all other nutrients. Evidence that tubular internalization of proteins other than Cd/MT provides an additional entry route for Cd is increasingly recognized [90,119,120].

Using mice exposed to Cd in drinking water for 1, 2, or 4 months, Fujishiro et al. observed a preferential accumulation of Cd in the kidney cortex, while noting MT was more abundant in the proximal tubules than in the distal tubules [121], similar to a study using human kidney sections [122]. Given that MT expression is induced by Cd, the high abundance of MT in proximal tubules than the distal tubules can thus be expected if PTCs absorbed most of Cd.

4.2. Cd-Induced Albuminuria: Glomerular or Tubular Cause

With a large molecular weight (66 kDa) and negative charges, albumin is not filtered by glomeruli [1234,124]. In normal health, 1–10 g of albumin, (40-50 g of plasma protein) may reach the tubular lumen each day by means of transcytosis through endothelial cells and podocyte foot processes [125,126]. A small fraction of albumin reaching tubular lumen is reabsorbed through RME, while most albumin in the ultrafiltrate is reabsorbed and returned to blood circulation by fluid-phase endocytosis and transcytosis [123,124].

Glomerular Membrane Permeability: A 4-fold increase in the glomerular filtration of albumin was noted in female Sprague-Dawley rats, exposed to Cd in drinking water for up to 18 months [127,128]. This glomerular membrane effect of Cd appeared to occur before GFR and tubular effects, evident from the excretion of enzymes, NAG, alanine aminopeptidase and lactate dehydrogenase [127]. Apparently, the glomerular membrane was particularly sensitive to Cd toxicity resulting in increased membrane permeability to albumin. A non-cytotoxic concentration of Cd (1 µM) increased the permeability of human renal glomerular endothelial cells in monolayers and caused the redistribution of the adherens junction proteins vascular endothelial-cadherin and β-catenin [129,130].

Megalin/Cubilin RME: In a study using data from 519 Thai subjects with moderate-to-high Cd exposure, it was estimated that Cd may reduce fractional reabsorption of albumin and β₂M by 18 and 21%, assuming the glomerular sieving coefficients of 10⁻⁴ and 10⁻² for albumin and β₂M, respectively [131]. Because RME of β₂M requires megalin but not cubilin; RME of albumin requires both proteins, it can be inferred that Cd had disrupted megalin-mediated endocytosis of both proteins by a single “shared” mechanism. Cd reduced expression levels of CUBN (cubilin) and LRP2 (megalin) in the kidney tubular cells was observed in another study using Sprague Dawley male rats were given Cd in drinking water at 0, 50, or 75 mg/L CdCl₂ for 1 and 6 months [132].

In summary, filtered albumin not subjected to transcytosis in the S1 segment of the proximal tubule is processed identically to β₂M (Figure 8).

Because the excreted Cd is primarily in MT-bound form, which is a detoxified storage form of the metal [133,134], the Cd released from lysosomal degradation of reabsorbed albumin (Figure 8), is retained within PTC as MT-bound. The binding of Cd to MT in this manner is viewed as a detoxification mechanism. Notably, however, as the influx of Cd into PTCs continues, the metal binding sites of MT become saturated [17,135], leaving the metal in free ionic form (Cd²⁺) that can reach various cell organelles, especially the mitochondrion which is characteristically abundant in PTCs [17,136].

Because of the large number of mitochondria, the homeostasis and survival of PTCs rely heavily on autophagy [137,138,139]. As demonstrated in Cd intoxicated rats, inhibition of autophagy and interference with the function of lysosomes by Cd resulted in acute kidney injury [140], while kidney fibrosis has been linked to impairment in tubular protein endocytosis [109,110,111].

4.3. Albuminuria/Proteinuria and CKD

Albuminuria, defined as albumin-to-creatinine ratio (ACR) of 20 mg/g cr in men and 30 mg/g cr in women, like eGFR, is a diagnostic criterion for CKD. However, its utility in CKD diagnosis is based mostly on circumstantial evidence; albumin excretion at a rate of 7 mg/g creatinine predicted incident CKD within 10 years [141].

The overall impact of Cd on tubular reabsorption of albumin and β₂M was assessed using data from 519 Thai subjects with moderate-to-high Cd exposure [131] (Table 3).

Table 3 provide estimated amounts of albumin and β₂M that are reabsorbed through RME and subjected to lysosomal degradation, assuming glomerular sieving coefficients for albumin (GSC_alb) and β₂M (GSC_β2M) to be 0.01 and 1, respectively. Up to 3 g of albumin enters PTCs daily through meglin/cubilin RME.

Recently, the role for oxidized albumin in causing kidney tubular cell death through ferroptosis has been demonstrated [142]. Prior to this observation, using a cell culture model of PTC, Fels et al. have demonstrated unambiguously the cytotoxicity of Cd complexed with albumin [118]. These findings underscore the significance of meglin/cubilin RME that mediates the internalization of Cd-albumin complexes/oxidized albumin by PTCs (Figure 8 and Figure 9) [90,119,142], thereby providing additional evidence that links impaired tubular protein endocytosis and/or impaired lysosomal protein degradation to albuminuria/proteinuria and kidney fibrosis [143].

Kidney fibrosis after chronic exposure to a low-dose Cd has been demonstratable in experimental studies [144,145]. Evidence from the synchrotron imaging of metals in human kidney tissue samples is in line with Cd-induced kidney fibrosis [146]. The degree of tubular atrophy correlated with Cd accumulation levels in a histopathological examination of kidney biopsies from healthy kidney transplant donors [147].

Notably, however, studies deriving BMDL values using disease markers such as albuminuria/ proteinuria are limited [107,148.149]. Most reported BMDL figures of Cd excretion were based on nephrotoxicity biomarkers (Section 2), which were not indicative of Cd early effects nor were clinically relevant.

4.4. CKD and Its Progression Toward Kidney Failure

CKD is a major public health problem worldwide because it causes high morbidity and mortality, especially from cardiovascular disease [150,151], and kidney failure, signified by a fall of eGFR to 15 mL/min/1.73 m² or below. This ailment has now reached epidemic proportions in many parts of the world, and it is projected to be the fifth leading cause of years of life lost by 2040 [152,153]. The cost associated with treatment involving dialysis and kidney transplant is escalating.

Indisputably, exposure to Cd pollution has contributed to the development of CKD in a significant proportion of adult population worldwide [11,12], while the Bradford-Hill Criteria provide evidence that strengthens a causal relationship between CKD and Cd exposure [3].

Concerningly, there is no theoretical reason to believe that a falling eGFR due to nephron destruction by Cd is reversible nor is an effective chelation therapy to reduce the kidney Cd burden. Hence, developing strategies to prevent CKD development and to mitigate its nephrotoxicity is of great public health significance.

4.5. Cd and Cellular Stress Responses

Hijacking transport proteins for essential metals, Cd can enter any cell and reaches its organelles, and subcellular structures [136]. Thus, Cd can impact the function of nearly all cell types in the body [1,2], including erythrocytes [154]. Through MT and transport proteins for Ca and Fe, i.e., the metal coupling unit (MCU) and DMT1, Cd reaches the inner membrane of mitochondria [155]. There, it reduces the synthesis of adenosine triphosphate (ATP), suppresses the electron transport chain, and promotes the formation of reactive oxygen species (ROS) [136,156]. Accordingly, the most frequently identified sign of Cd toxicity is related to oxidative stress damage to lipids and proteins such as oxidized low-density lipoprotein, oxidized albumin, discussed in Section 4.3.

In addition to mitochondrial source, ROS can be produced, normally in the peroxisomes, and the endoplasmic reticulum [157]. To maintain cellular redox states and cell function as well as to protect against potential damage from excessive ROS formation many mechanisms have been evolved. Herein, antioxidant defenses involving the upregulation of the MT gene and the heme oxygenase-1 (HO-1) gene by cells in response to Cd are highlighted (Figure 9).

4.5.1. Cd-Induced Upregulation of the MT Gene

Upregulation of the MT gene by PTCs in response to Cd is a detoxification mechanism because sequestering of Cd in MT (Cd-MT complexes) prevents acute toxicity due to “free” Cd²⁺ ion which is the chemically reactive form of the metal. Each MT molecule can carry up to 7 atoms of Cd²⁺, 7 atoms of Zn²⁺ or 12 atoms of Cu²⁺, and the metal-MT complexes are denoted as Cd₇MT, Zn₇MT, and Cu₁₂MT [158]. Complexes of mixed metals such as Cd₃Cu₃ZnMT, Cd₄CuZn₂MT, and Cd₆CuMT, are formed in vivo [158].

There are at least 16 MT isoforms, and they belong to four major families, MT-1–MT-4 [159]. MT-1 and MT-2 are ubiquitously expressed by most cells including leucocytes and kidney tubular cells [122,160,161]. MT-3 has a particularly high binding affinity for Cu, and high levels of MT-3 isoform are found kidneys and neurons [134,162,163]. Because free Cu ion is redox active, Cu-MT-3 complexes may account for the nephrotoxicity and neurotoxicity of Cd [163].

A study from Taiwan (n = 2447, mean age 55.1 yrs) [164], noted that subjects with proteinuria had mean urinary Cd concentration 27.3% higher than those without and that risk of proteinuria was increased 2.67-fold and 1.94-fold in those with elevated urinary Cd and Cu levels, respectively [164]. Proteinuria in Taiwanese subjects was detected at a low exposure level, reflected by the mean urinary Cd in subjects with proteinuria of 1.1 μg/L. This exposure level was in line with the BMDL value of Cd excretion, determined for proteinuria endpoint.

A direct correlation between 24-hour proteinuria and urinary copper levels was observed together with an inverse correlation of eGFR and serum Cu in study that included 313 kidney disease patients and 19 healthy controls [165]. Cd exposure was not investigated in this study; nonetheless, it explains a connection between Cu and proteinuria observed in the Taiwanese study [164].

Because absorbed Cd reaches liver first, Cd-MT complexes formed in the liver presumably contain oral Cd, while those synthesized in lungs contain inhaled Cd. Liver and lungs serve as endogenous reservoirs from which Cd-MT complexes are released as cells die. Cd-MT complexes are redistributed to kidneys that are equipped with protein internalization capability. There is little evidence that the complex of Cd-MT is reabsorbed by megalin/cubilin RME in the proximal tubule [90]. However, it could be retrieved in the distal tubule and collecting duct by NGAL/lipocalin 2 REM [118,120]; consequently, filtered Cd-MT complexes all are absorbed and retained within those regions of nephrons.

Sequestering Cd by MT prevents acute cytotoxicity; however, it may increase the risk of long-term toxicity because Cd²⁺ ions can be released under certain conditions, leading to an increased synthesis of nitric oxide (NO) that liberates the Cd and Cu previously bound to MT [166,167,168]. Moreover, upregulation of MT in response to Cd can impact cellular redox state, vital to maintaining normal protein structures, intermediary metabolism, and cell function [169,170,171]. For example, increased sequestration of Zn and Cu in MT could lead to a reduction in the activity of the antioxidant enzyme superoxide dismutase 1 (SOD1) that requires Zn and Cu as cofactors [172].

4.5.2. Cd-Induced Upregulation of the HO-1 Gene

In response to any stressor, the HO-1 gene is upregulated [176]; unexpectedly, however, the HO-1 gene activation by Cd appeared to be different from physiological HO-1 activators, notably prostaglandin D2 (PGD2) [173]; as such, induction of HO-1 expression by Cd did not result in a concomitant increase in bilirubin synthesis [174]. The reason for this phenomenon remains unclear.

Using a cell culture model of human retinal epithelial cells, Sataug et al. have shown that PGD2 activated the HO-1 gene through the D-prostanoid 2 (DP2) receptor [173]. In contrast, Cd activated the HO-1 gene via two enhancers; the Cd response element (CdRE, TGCTAGATTTT) and Maf recognition antioxidant response element (MARE, GCTGAGTRTGACNNNGC), also known as stress response element (StRE) [175]. Moreover, it suppressed the lysosomal degradation of Nrf2 [176] and caused nuclear export of the repressor Bach1, thereby allowing the transactivation of the HO-1 gene by the Nrf2/small Maf complex [177].

Table 4 provide data on the potency of various metals to induce HO-1 gene expression in various human cell lines, determined by the ARE reporter gene assay [178].

Through the enhancers, CdRE and ARE, Cd can induce a massive increase in HO-1 enzyme activity, resulting in anthe degradation of heme from which Fe is released and carbon monoxide is generated. Nonetheless, there is a little change in bilirubin concentrations in cells treated with Cd [174]. Bilirubin by virtue of its lipophilic properties, protects lipids from oxidation more effectively than the water-soluble antioxidants, such as glutathione and vitamin C [179]. As the consequence, induction of HO-1 gene expression by Cd leads to loss of PTCs through ferroptosis [180,181,182].

4.6. Can Exogenous HO-1 Inducers Mitigate Cd-Induced Oxidative Stress?

As discussed above, Cd appears to cause oxidative stress through lowering levels of endogenous antioxidants, especially bilirubin. This raises the possibility to replete such insufficient bilirubin by exogenous HO-1 inducers. A wide range of antioxidants from plant foods, such as curcumin, quercetin, tert-butylhydroquinone, and caffeic acid phenethyl ester, are known to be HO-1 inducers, as are catechin (in green tea), α-lipoic acid (in broccoli, spinach), resveratrol (in red wine, grapes), carnosol, sulforaphane (cruciferous vegetable), coffee diterpenes cafestol, and kahweol [183]. Beneficial effects of consumption of these plant antioxidants could thus be mediated in part through the induction of HO-1 expression.

Diet high in anti-oxidative and anti-inflammatory nutrients was associated with increased serum bilirubin levels and reduced oxidative stress and systemic inflammation associated with Cd exposure [184]. Furthermore, consumption of plant-based diets may provide a viable option to prevent and manage CKD that continues to rise worldwide [185,186,187]. The investigation into the potency of plant-derived compounds to activate the HO-1 gene using the ARE reporter gene assay or similar constructs should be encouraged [188].

5. Conclusions

Current dietary Cd exposure guidelines vary fourfold, ranging from 0.21 to 0.83 μg/kg bw/d with a tenfold difference in threshold levels of Cd excretion, varying between 0.5 and 5.24 µg/g cr. These exposure guidelines derived by JECFA, EFSA, US FDA and ATSDR, respectively described as TDI, RfD, TRV and MRL all were based on the premise that there is a critical exposure level below which adverse effects of Cd on bone/and or kidney target are discernable. Notably, however, the NOAEL/BMDL values of Cd excretion derived for the kidney target were based on nephrotoxicity indicators, not kidney disease, which is diagnosed when eGFR falls to a third of normal value or ACR rises to 20 mg/g cr in men and 30 mg/g cr in women for at least 3 months.

The benchmark Cd excretion at 5% drop in the eGFR, identified from Thai population data was 0.17 µg/g cr, while the benchmark Cd excretion at 5% increase in protein excretion was as little as 0.054 µg/g cr. In comparison, the Cd excretion benchmark could not be reliably determined, when 5% increase in β₂M excretion was used as an endpoint. Thus, there is no basis for using β₂M excretion in computing the health risk due to Cd.

Supporting the threshold for Cd excretion at a rate below 0.17 μg/g cr is the BMDL₅/BMDL₁₀) values of 0.198 (0.365) μg/g cr, determined by Shi et al., using data from 4530 US adults and diabetes as a disease endpoint [189].