**3. Results**

#### *3.1. Mortality Risk Ratios and Cd Exposure Markers in Men*

Hazard ratios for all causes, major causes of deaths, and renal diseases for a 10-μg/g Cr increase of urinary Cd and those for a 1-g increase of LCd after controlling for age in men are shown in Table 2. For men, the hazard ratio for all causes of deaths was significantly increased for a 10-μg/g Cr increase of urinary Cd, albeit no major causes of deaths were found to contribute to increased mortality. For a 1-g increase of LCd, no significantly increased mortality risks for all causes and no causes of death were noted in men (Table 2).


**Table 2.** Dose–response relationships between hazard ratios and Cd exposure markers in men.

D: number of deaths, HR: hazard ratio, CI: confidence interval, N: number of subjects *n* = 1121 for urinary Cd, *n* = 1135 for lifetime Cd intake.

#### *3.2. Mortality Risk Ratios and Cd Exposure Markers in Women*

In women, hazard ratio was significantly increased for all causes for 10 μg/g Cr increase of urinary Cd. For renal diseases, particularly for renal failure, hazard ratios were significantly increased for10 μg/g Cr increase of urinary Cd in women (Table 3). Although hazard ratio for all causes was not significantly increased for 1 g increase of LCd, mortality risks for renal diseases and renal failure were significantly increased for 1 g increase of LCd in women (Table 3). These results in women indicate that increasing dose of Cd exposure may increase mortality for renal diseases, particularly renal failure. At the same time, however, they rise a question which exposure marker, urinary Cd or LCd, is a better marker to contribute to increasing risk for renal disease deaths.


**Table 3.** Dose–response relationships between hazard ratios and Cd exposure markers in women.

D: number of deaths, HR: hazard ratio, CI: confidence interval, N: number of subjects. *n* = 1333 for urinary Cd, *n* = 1361 for lifetime Cd intake.

#### *3.3. Comparisons of Mortality From Renal Diseases between Men and Women at Di*ff*erent Cd Exposure Levels*

Crude mortality rates for renal diseases in six groups with a 1-g di fference of LCd (a) and in three groups with a 10-μg di fference of urinary Cd (b) in men and women are illustrated in Figure 1.

**Figure 1.** Crude mortality rates (%) for renal diseases and Cd exposure levels indicated by (**a**) LCd (+1 g) and (**b**) UCd (urinary Cd; +10 μg/g Cr). D (M, W): number of deaths for men and women; *n* (M, W): number of subjects of men and women.

Crude mortality from renal diseases was higher in men in the lowest exposure group, indicated by LCd (Figure 1a) or urinary Cd (Figure 1b), suggesting that factors other than Cd exposure, such as smoking and hypertension, relevant to chronic kidney disease (CKD), might influence mortality. While we found more deaths for renal diseases in the highest LCd group (LCd ≥ 5 g) in women compared men (Figure 1a), no stable result was obtained in the highest urinary Cd group (urinary Cd ≥ 10 μg/gCr), because of the small number of male subjects. These results sugges<sup>t</sup> that mortality risk ratios for renal diseases might be changed, particularly in men, if relevant factors other than age could be controlled for in the analysis using Cox's proportional model.

#### *3.4. Comparisons of E*ff*ects on Mortality between Urinary Cd and LCd in Women*

To investigate which Cd exposure marker contributed to increased mortality from all causes and renal diseases, including renal failure, the stepwise elimination method was used for mortality risk analysis with three explanatory valuables, namely, urinary Cd, LCd, and age in women (Table 4). For all causes of deaths, age and urinary Cd were selected as factors that significantly increased mortality risks in women. However, age and LCd were selected as factors increasing mortality from renal diseases and renal failure in women, suggesting that LCd may be a good marker to show a better dose–effect relationship with mortality risk for renal diseases, including renal failure, compared with urinary Cd.


**Table 4.** Dose–response relationships between hazard ratios and Cd exposure markers in women.

D: number of deaths, HR: hazard ratio, CI: confidence interval, Number of subjects = 1333.
