**4. Discussion**

None of the participants in the present study had been exposed to metals in the workplace, and their urinary concentrations of Cd and Pb were thus presumed to reflect environmental sources, notably diet [5,49]. The urinary Cd and Pb concentrations were similar to the data obtained from

population-based studies in the U.S. [9–12], Canada [13], Taiwan [14] and Korea [15]. The eGFR was calculated by using CKD-EPI equations, which are considered to be the most accurate equations for the eGFR [20,22]. The CKD-EPI equations have been validated by using inulin clearance [22]. The overall mean eGFR for participants was 105 mL/min/1.73 m2, ranging between 70 and 139 mL/min/1.73 m2. The wide range of variation in the eGFR was congruen<sup>t</sup> with the notion that a normal level of the GFR could vary widely [20]. Both physiological and pathological conditions are known to a ffect the GFR, and the mean GFR in young adult Caucasians is approximately 125 mL/min/1.73 m<sup>2</sup> [20].

In a regression model in which all subjects were included, the eGFR levels varied inversely with age, BUN and ECd/Ccr, but they did not vary with EPb/Ccr (Table 2). The inverse association between the eGFR and BUN was expected, as urea is one of the metabolic waste products that is eliminated through kidneys. Of note, however, while the inverse associations of the eGFR with age and BUN were present in all subgroups, the association between the eGFR and ECd/Ccr reached statistical significance levels in non-smokers (men and women). In addition, the eGFR levels in women varied inversely with both ECd/Ccr (β= −0.170, *p* = 0.008) and EPb/Ccr (β= −0.132, *p* = 0.043). These data could be interpreted to sugges<sup>t</sup> that levels of Cd and Pb intake from the diet were su fficient to produce adverse e ffects on nephrons and subsequently reduce the elimination rate of urea. They also suggested that women may be more susceptible than men to renal e ffects of elevated Cd and Pb intake levels. Supporting this argumen<sup>t</sup> is the fact that blood Cd and Pb both were associated with BUN in a prospective study in the U.S. that included 259 premenopausal women, where one-third of them had eGFR levels <90 mL/min/1.73 m<sup>2</sup> and stage 1 CKD [43]. In addition, urinary Cd and Pb were associated with increased BUN in adolescents enrolled in NHANES 2009–2014 [42].

In a multivariable regression analysis that was done by using urinary Cd and Pb that were normalized to the excretion of creatinine (Table S1), the eGFR did not show significant relationships with ECd/Ecr or EPb/Ecr in any subgroups. Normalizing to the excretion of creatinine was done to correct for urine dilution. This practice, however, inevitably introduces confounders and often creates gender bias. In general, men have a higher muscle mass than women, and, consequently, the mean ECd/Ecr was lower in men than women (Table 1). In e ffect, the health risks associated with Cd exposure in men could have been underestimated. The mean ECd/Ccr in men was almost identical to that of women (Table 1). It is increasingly recognized that creatinine adjustment is problematic, and urine specific gravity has been used to correct for dilution e ffects [9,44,45,50–52]. Herein, we have demonstrated the utility of normalizing excretion rate of metals to creatinine clearance that only required simultaneous urine and blood sampling together with the equations, given in Section 2.3 [33,35].

In a quantitative analysis for an e ffect of Pb, a non-significant association between the eGFR and EPb/Ccr was evident (Figure 1D). Furthermore, the POR for the reduced eGFR did not increase with EPb/Ccr in a dose-dependent manner (Table 3). Though urinary Pb is not a good indicator of Pb body burden, it does not rule out the possibility for a glomerular e ffect of Pb. Blood Pb levels ≥2.4 μg/dL were associated with a 1.56-fold increase in risk of eGFR levels of <60 mL/min/1.73 m<sup>2</sup> in adults who enrolled in NHANES 1999–2006 [26]. However, the absence of a dose–response relationship between the EPb/Ccr quartiles and POR for the reduced eGFR may sugges<sup>t</sup> that the levels of environmental exposure to Pb that were experienced by participants in this study were below a nephrotoxicity threshold limit for Pb.

In a quantitative analysis for an e ffect of Cd (Figure 1C), an increment of ECd/Ccr to the median level (0.38 μg/L) was associated with a significant decrease in the eGFR (an approximate of 5 mL/min/1.73 m2). An increment of ECd/Ccr to the 75th percentile level (0.62 μg/L) was not associated with a further decrease in the eGFR level. In the logistic regression analysis (Table 3), the POR for the reduced eGFR rose by 2.51 and 2.87 fold as the ECd/Ccr rose to the median level and the 75th percentile, respectively. The median ECd/Ccr level could thus be considered to represent the lowest urinary Cd level that was associated with observed adverse e ffect among the participants in the present study. The median and 75th percentile levels of ECd/Ecr corresponded to 0.44 and 0.76 μg/g of creatinine, respectively.

Of relevance, urinary Cd level of 0.8 μg/g of creatinine has been found to be associated with a significant eGFR decrease in Swedish women, aged 53–64 years [39]. In addition, U.S. population studies (NHANES) have provided a rich data source that links Cd and Pb exposure indicators to increased risk of CKD. In NHANES 1999–2006, urinary Cd levels <sup>≥</sup>1μg/<sup>L</sup> and blood Cd levels ≥0.6 μg/<sup>L</sup> were associated, respectively, with 1.48- and 1.32-fold increases in the risk of a low eGFR, defined as an eGFR <60 mL/min/1.73 m<sup>2</sup> [26,27]. In NHANES 2011–2012, blood Cd levels >0.53 μg/<sup>L</sup> were associated with 2.21-fold increases in risk of low eGFR [53].

In NHANES 2007–2012, blood Cd levels >0.61 μg/<sup>L</sup> were associated with a 1.80-fold increase in the risk of a low GFR [28]. In addition, the mean eGFR in women with hypertension and blood Cd in the highest quartile was 5.77 mL/min/1.73 m<sup>2</sup> lower than that of normotensive women who had blood Cd in the lowest quartile [28]. Of interest, an additional e ffect of hypertension on Cd-related GFR reduction has been seen in residents of an area of Thailand with Cd pollution: Those with hypertension, on average, had a 4.6 mL/min/1.73 m<sup>2</sup> lower eGFR compared with the mean eGFR of normotensive subjects who had similarly high urinary Cd levels [33].

In a Swedish women study, urinary Cd levels associated with a glomerular e ffect (eGFR decline) of 0.8 μg/g of creatinine were close to urinary Cd levels of 0.67 μg/g of creatinine that were associated with tubular injury based on urinary NAG levels [39]. These data challenge a long-held view that tubular e ffects occur long before the glomerular e ffect becomes apparent. A recent quantitative analysis of excreted Cd in relation to levels of the eGFR, urinary NAG, and β2MG suggested a decrease in the GFR to be an early e ffect, given that excreted Cd emanates from injured tubular cells and that the injury leads to nephron atrophy, a decreased eGFR, and impaired reabsorption of filtered β2MG [35]. Accordingly, it can be hypothesized that su fficient tubular injury induced by Cd disables glomerular filtration, destroys nephrons, and causes glomerulosclerosis, interstitial inflammation, fibrosis and CKD [35,54].

A lack of association between the eGFR and urinary Cd was reported in a cross-sectional study in Japan [55] in which eGFR values were derived for 1200 women by using a serum creatinine-based eGFR estimating equation for Japanese women. In this Japanese study, 222 women who were aged 42–79 years (mean 61.9) were drawn from a control area without Cd pollution, based on the Cd content in rice, which is a dietary staple, while 636 and 355 women of the similar age range were drawn from two areas with Cd pollution. Though the Cd intake levels from rice in the two Cd pollution areas were higher than those of the control group, the mean eGFR values in these three areas were similar. In addition, the eGFR levels were unrelated to urinary Cd, but they were related to age. These data suggested that the eGFR was not associated with Cd body burden in the Japanese population with relatively high levels of Cd intake [55]. A longitudinal study is required to dispute or confirm the observation made in this Japanese study. It is noteworthy, however, that a prospective study in a Cd pollution area in Thailand reported a progressive decrease in the eGFR over a five-year observation period [31].

Distinct from the Japanese study, an inverse association between the eGFR and urinary Cd was observed in the present study. The study Thai women were younger (the mean age of 36.9) and had lower urinary Cd levels (mean urinary of 0.51 μg/g of creatinine) compared to the Japanese women in the control area (mean age of 61.9) and mean urinary Cd of 3.03 μg/g of creatinine. The mean eGFR (range) was 106 (72−139) and 79.8 (30.6−130) mL/min/1.73 m<sup>2</sup> in the Thai and Japanese studies, respectively. The eGFR-urinary Cd association in the Thai study could have been due to younger age and lower Cd intake levels compared with the Japanese study. In addition, it is conceivable that the functional expressions of Cd-induced nephrotoxicity in low-dose and high-dose exposure conditions are di fferent. For instance, a low-level environmental exposure to Cd has been implicated in the pathogenesis of hypertension, a known cause and consequence of CKD, in a longitudinal study in the U.S. [56], while blood Cd as low as 0.4 μg/<sup>L</sup> was associated, respectively, with 1.54- and 2.38-fold increases in risk of hypertension in Caucasian women and Mexican-American women who were aged ≥20 years and enrolled in NHANES 1999–2006 [57]. In stark contrast, a chronic high-dose Cd exposure has not been found to be associated with hypertension in Japanese population studies [58,59].

In the present study, chronic low-level exposure to Cd, indicated by a urinary Cd level as low as 0.44 μg/g of creatinine was associated with a decrease in the eGFR by 5 mL/min/1.73m2. This low urinary Cd level was also associated with a 2.5 fold increase in the prevalence odds of eGFR levels <96 mL/min/1.73 m2.These findings support the large number of population-based studies that have suggested that low-environmental exposure to Cd may increase the risk of CKD, thereby raising the possibility for a role of Cd exposure in current epidemics of CKD. It is noteworthy that the reported toxic urinary Cd levels did not exceed the toxicity threshold limit of 5.24 μg/g of creatinine that was established by the Food and Agriculture Organization of the United Nations (FAO) World Health Organization (WHO) [49,60]. We sugges<sup>t</sup> that the current urinary Cd threshold limit does not a fford health protection and should be lowered.

A small observable decrease in the eGFR that is attributable to long-term environmental exposure to Cd was expected, as the participants in our study were relatively young, with an overall mean age of 34.9 years. However, because dietary exposure to Cd and Pb is inevitable for most people, exposure to these toxic metals is likely to continue, leading to a further reduction in the GFR. In addition, the GFR may continue to decrease presumably due to mobilization of liver Cd and bone Pb to kidneys [5]. Even a small increase in the risk of CKD can result in many a ffected people, given that environmental exposure to Cd and Pb is widespread.

In conclusion, our analysis of archived data provides evidence that links environmental exposure to Cd to GFR decline, even when dietary intake levels of Cd are low. This glomerular e ffect of low-level environmental exposure to Cd was demonstrable only when urinary Cd concentrations that were normalized to creatinine clearance. Women appeared to be more susceptible than men to toxicity due to an elevated dietary intake of Cd and Pb.

#### **5. Strengths and Limitations**

The strengths of this study include the community-based recruitment of apparently healthy women and men who were relatively young, as well as the fact that simultaneous blood and urine sampling was undertaken at the same time of the day, thereby reducing diurnal variation of the GFR. The normalization of excretion rates of Cd and Pb to creatinine clearance was an additional strength because confounding e ffects of muscle mass and urine flow rate were both eliminated. Furthermore, environmental sources of Cd and Pb were relatively homogenous, as none of the participants had occupational exposure to metals. The limitations of this study were archived data with no availability to access the same people for long term comparisons, a modest sample size, and a cross-sectional design, which limited a causal inference of Cd and Pb exposure on the observed GFR reduction.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2305-6304/8/1/18/s1, Figure S1: Comparing e ffects of ECd/Ecr and EPb/Ecr on eGFR change; Table S1: Multivariable regression analysis for association of eGFR with ECd/Ecr and EPb/Ecr; Table S2: Prevalence odds ratios for reduced eGFR across ECd/Ecr quartiles and EPb/Ecr quartiles.

**Author Contributions:** S.S., D.A.V., P.U. and G.C.G. formulated study protocols. P.U. obtained ethical institutional clearance, recruited participants and organized the collection of biologic specimens and their shipment for metal analysis by the ICP/MS in Australia. S.S. prepared an initial draft of a manuscript. D.A.V. and G.G. provided intellectual input and revised the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research project received no external funding.

**Acknowledgments:** This work was partially supported by the Chulalongkorn University Medical Faculty, Bangkok, Thailand and the Commission for Higher Education, Thailand Ministry of Education. Additionally, it was supported with resources of the Department of Nephrology, Princess Alexandra Hospital, and the Kidney Disease Research Centre, the University of Queensland Faculty of Medicine and Translational Research Institute.

**Conflicts of Interest:** The authors have declared no potential conflicts of interest.
