Abstract
Background:
Epidemiologic studies suggest cadmium exposure is associated with cardiovascular disease risk, including heart failure. However, prior findings may be influenced by tobacco smoking, a dominant source of cadmium exposure and risk factor for heart failure. The present study leverages up to 20 years of follow-up in the Danish Diet, Cancer and Health cohort to examine the relationship between urinary cadmium and incident heart failure among people who never smoked.
Methods:
Between 1993 and 1997, 19,394 never-smoking participants (ages 50–64 years) enrolled and provided a urine sample. From this sample, we randomly selected a subcohort of 600 men and 600 women and identified 958 incident heart failure cases occurring between baseline and 2015. Using a case–cohort approach, we estimated adjusted hazard ratios (aHR) for heart failure in Cox proportional hazards models with age as the time scale.
Results:
Participants had relatively low concentrations of urinary cadmium, as expected for never smokers (median= 0.20; 25th, 75th= 0.13, 0.32 μg cadmium/g creatinine). In adjusted models, we found that higher urinary cadmium was associated with a higher rate of incident heart failure overall (aHR=1.1 per interquartile range difference; 95% CI= 1.0 – 1.2). In sex-stratified analyses the association seemed restricted to men (aHR= 1.5; 95% CI= 1.2 – 1.9).
Conclusion:
In this cohort of people who never smoked tobacco, environmental cadmium was positively associated with incident heart failure, especially among men.
Keywords: cadmium, incident heart failure, urine biomarker, case–cohort study design
INTRODUCTION
Cadmium is a naturally occurring, non-essential, toxic metal that is transferred into the environment from industrial and agricultural activities. Human exposure to cadmium through diet and inhalation is a public health concern given its documented associations with cancer, impaired renal function, and cardiovascular morbidity and mortality (1–5). Three recent systematic reviews have concluded that cadmium exposure is a risk factor for cardiovascular disease (6–8), especially coronary heart disease. However, there remains some concern that the associations observed in prior studies may be influenced by tobacco smoking, a dominant source of cadmium exposure and major risk factor for heart failure and cardiovascular disease (9). Furthermore, a majority of the prior epidemiologic studies have evaluated composite cardiovascular disease outcomes, and the few studies that have considered incident heart failure were underpowered to evaluate this association independent of tobacco smoking (10–12). These studies reported an association between higher concentrations of cadmium biomarkers and heart failure that decreased in precision when the analyses were limited to never smokers.
Heart failure is a prevalent condition that is characterized by left ventricular dysfunction which progresses into alterations in cardiac structure and the development of clinical symptoms (13). Previous research suggests that cadmium exposure may increase risk for heart failure by directly altering the myocardium and potentially contributing to systolic left ventricular dysfunction (14–16). The association between cadmium and heart failure may also be mediated through secondary mechanisms including hypertension and kidney dysfunction (17–19). Determining whether cadmium exposure is a risk factor for heart failure can inform policy-makers who are considering health-based regulations to limit environmental cadmium exposure among the general population (5,20).
The objective of this study was to evaluate the relationship between urinary cadmium and heart failure in a prospective case–cohort study of people who never smoked tobacco. In our analyses we adjusted for urinary cotinine concentrations, an indicator of secondhand tobacco smoke exposure. We conducted analyses in the full sample and separate sex strata. Our hypothesis was that higher urinary cadmium concentrations would be positively associated with incident heart failure. Furthermore, we hypothesized that this association may vary by sex due to potential differences in cadmium absorption and heart failure etiology.
Previously, we published analyses examining the association of urinary cadmium with acute myocardial infarction and stroke among participants who self-identified at baseline as never tobacco smokers in the Danish Diet, Cancer and Health (DCH) cohort (21,22). In these analyses, we did not find strong evidence to suggest that low levels of cadmium exposure were associated with incident stroke or acute myocardial infarction.
METHODS
We used existing data and urine samples from the prospective DCH cohort to evaluate the relationship between urinary cadmium concentrations and incident heart failure among men and women who never smoked tobacco. Details of sampling, recruitment, and follow-up procedures for the DCH cohort are provided elsewhere (23) and in the supplemental material. Briefly, between December 1993 and May 1997, 57,053 individuals (ages 50–64 years) living in the area surrounding Copenhagen or Aarhus consented to participate. The DCH cohort was established in accordance with the Helsinki Declaration and approved by the local Ethics Committees. Written informed consent was provided by all participants.
Only participants who self-identified at baseline as never tobacco smokers (n=19,394; n= 6,821 men and n=12,573 women) were eligible to be included in this analysis. Among eligible participants, we identified incident heart failure cases (ICD-10 codes I50.0-I50.9 and I11.0) and date of diagnosis in the Danish National Patient Registry (NPR) by means of the civil registry number, a unique personal identification number for every Danish citizen (24). We also used the Danish Cause of Death Registry to identify fatal heart failure cases that were not included in the NPR and occurred after enrollment through 2013. We included all incident cases recorded in the NPR that occurred through December 31st, 2015. We subsequently created a referent subcohort, by randomly sampling 600 men and 600 women (total n=1,200) from the 19,394 participants who were never smokers. Of the 958 cases of incident heart failure identified in the full cohort, 64 were from the referent subcohort. We censored participants at the emigration date recorded in the Danish Central Person Registry (n=3), the date of death documented in the Danish Cause of Death Registry (n=154), or December 31st, 2015, whichever date occurred first.
At baseline, we collected spot urine samples using trace metal-free techniques as previously described (23). Samples from cases and subcohort members were mixed and anonymized before shipment on dry ice to the Trace Metals Laboratory at RTI International (Research Triangle Park, NC). At RTI, we used an iCAP Q ICP-MS system (Thermo Scientific, Waltham, MA) equipped with a helium gas collision cell to quantify cadmium concentrations in urine samples (see eAppendix). Additionally, we quantified urinary creatinine colorimetrically by the Jaffe reaction with a Cayman Chemicals (Ann Arbor, MI) Creatinine Assay Kit following the manufacturer’s instructions. We measured urinary osmolality by the freezing point depression method using a Model 3320 Micro-Osmometer by Advanced Instruments, Inc. (Norwood, MA).
For the main analyses, we used the machine-read cadmium values for all participants including the 18 heart failure cases and 22 subcohort members with values below the limit of detection (LOD; median LOD [25th, 75th] = 0.003 [0.001, 0.003] μg/L)). We calculated creatinine-standardized cadmium concentrations by dividing urinary cadmium concentrations (μg) by creatinine (g).
We used a cotinine ELISA bioassay kit (Abnova Corporation, Taipei, Taiwan) to measure urinary cotinine, an indicator of tobacco smoke exposure. Participants self-reported age, sex, education, and physical activity on questionnaires administered at baseline, as previously described (23). Women reported menopausal status and parity. To be considered post-menopausal, women either had surgically removed ovaries, reported an age of menopause, or reported no natural menstruation during the last 12 months. We considered women who did not fulfill these criteria but reported sequential hormone replacement therapy as probably menopausal. Pre-menopausal women reported at least one menstruation within the 12 months before entry to the cohort and had no hormone replacement therapy use. At baseline, lab technicians also measured height and weight, and calculated body mass index (BMI).
Statistical Analyses
We used a case–cohort approach to estimate the relationship between creatinine-standardized urinary cadmium concentrations and incident heart failure (25). In this time-to-event analysis, follow-up for subcohort members started at their age on the DCH cohort enrollment date and ran until the age on the date of heart failure diagnosis or the censoring date. Based on the Prentice method, we estimated hazard ratios and robust 95% confidence intervals for an interquartile range difference in cadmium using Cox proportional hazards models with age as the time scale (25). We used age as the time scale because age is strongly predictive of the outcome and using time-on-study may bias results when age is related to the outcome (26). We selected covariates a priori (27) and examined minimally adjusted models including only sex, as well as fully adjusted models additionally including BMI (continuous–linear term), education (categorical), and urinary cotinine concentrations (continuous–linear term). We considered quartiles of creatinine-standardized cadmium to assess the dose–response relationship and conducted a linear test for trend by assigning observations in each cadmium quartile the median concentration of the quartile, treating the variable as continuous, and assessing the p-value for this term (28). Moreover, we visually evaluated linearity of the dose–response relationship by modeling creatinine-standardized cadmium using a natural cubic spline with three degrees of freedom (knots at 33rd and 67th percentiles). We used analysis of variance to compare this spline model to the model with a linear term and assessed the p-value as the non-linearity p-value.
We added a sex by cadmium (continuous variable) product term to adjusted models, as well as estimated the association between creatinine-standardized urinary cadmium concentrations and heart failure in sex-stratified analyses. Among the strata of women, we included menopausal status as an additional covariate in all models. We repeated the main analyses using cadmium quartiles and splines to assess the dose–response relationship in each sex stratum. We also conducted secondary analyses examining the association between urinary cadmium and heart failure among post-menopausal women only.
We conducted a series of sensitivity analyses (see eAppendix). In these sensitivity analyses, we examined multiple statistical approaches with urinary creatinine and osmolality measures to account for differences in urine density. We conducted these analyses because it is important to account for urine density when using urinary biomarkers, but there is little consensus on the optimal approach. We also conducted a series of analyses adjusting models for or excluding participants with health conditions that could be risk factors for heart failure. We did not include hypertension, diabetes, or hypercholesterolemia as covariates in our main analyses because it is plausible that these conditions could be on the causal pathway between cadmium exposure and heart failure (29). Since secondhand tobacco smoke exposure is a well-documented confounder, we also examined different approaches adjusting for cotinine concentrations and history of secondhand tobacco smoke exposure.
Lastly, we estimated the proportion of incident heart failure cases attributable to the highest quartile of urinary cadmium by calculating the relative measure of effect ([adjusted hazard ratio comparing highest versus lowest cadmium quartile-1]/adjusted hazard ratio comparing highest versus lowest cadmium quartile) and multiplying it by the proportion of heart failure cases in the highest cadmium quartile for the full study sample and each sex strata (30). All analyses were conducted in R version 3.6.2 using the survival package and functions to account for the case–cohort study design.
RESULTS
Participants who developed heart failure tended to be older, more obese, and have fewer years of formal education compared with the summary characteristics for the subcohort (Table 1). Participants who developed heart failure also had a slightly higher median urinary cadmium concentration [0.22 (0.10, 0.41) μg/L] compared with the subcohort members [0.19 (0.09, 0.34) μg/L]. Overall, women had higher median creatinine-standardized cadmium concentrations [0.30 (0.20, 0.44) μg/g] than men [0.16 (0.11, 0.23) μg/g; combined median= 0.20; 25th, 75th= 0.13, 0.32 μg/g; eTable 1]. Compared to the subcohort, a higher proportion of the female heart failure cases were among post-menopausal women. The median age at incident heart failure was similar among men [71 (66, 75)] and women [73 (68, 77) years].
Table 1.
Baseline characteristics of heart failure cases and subcohort members.
| Baseline characteristics | Subcohort (n=1200) | HF cases (n=958) | ||
|---|---|---|---|---|
| No (%) | Median (25th, 75th) | No (%) | Median (25th, 75th) | |
| Age at enrollment (years) | 1200 | 55.8 (52.5, 59.9) | 958 | 58.7 (54.5, 61.9) |
| Sex | ||||
| Male | 600 (50) | 481 (50) | ||
| Female | 600 (50) | 477 (50) | ||
| Education | ||||
| Low (<8 years) | 344 (29) | 361 (38) | ||
| Medium (8–10 years) | 549 (46) | 415 (43) | ||
| High (>10 years) | 307 (25) | 182 (19) | ||
| Parity (women only) | ||||
| 0 | 77 (13) | 68 (14) | ||
| 1 to 2 | 360 (60) | 260 (55) | ||
| 3 to 8 | 163 (27) | 149 (31) | ||
| Menopausal status (women only) | ||||
| Post-menopausal | 349 (58) | 335 (70) | ||
| Pre-menopausal | 103 (17) | 37 (8) | ||
| Probably menopausal | 148 (25) | 105 (22) | ||
| Cotinine concentration | ||||
| <= 20 ng/mL | 690 (58) | 478 (50) | ||
| >20 ng/ml - <50 ng/mL | 354 (30) | 303 (32) | ||
| >=50 ng/mL - <200ng/mL | 134 (11) | 153 (16) | ||
| >=200 ng/mL | 22 (1) | 24 (2) | ||
| Body mass index | ||||
| < 25 | 479 (40) | 240 (25) | ||
| 25 to < 30 | 523 (44) | 397 (41) | ||
| 30 + | 198 (16) | 321 (34) | ||
| Leisure-time physical activity | ||||
| No | 465 (39) | 444 (46) | ||
| Yes | 735 (61) | 514 (54) | ||
| Urine Cd concentration (μg/L) | 0.19 (0.09, 0.34) | 0.22 (0.10, 0.41) | ||
| Urine creatinine concentration (g/L) | 0.99 (0.47, 1.64) | 1.05 (0.49, 1.65) | ||
| Urine osmolality (mOsM) | 583 (312, 784) | 596 (325, 784) | ||
Abbreviation definitions: Cd = cadmium; HF= heart failure
We found a positive trend in the relationship between quartiles of urinary cadmium and incident heart failure (Table 2). Specifically, the fully-adjusted hazard ratio (aHR) for heart failure was 1.3 (95% CI= 0.95 – 1.8) comparing participants in the highest versus lowest quartile of creatinine-standardized cadmium concentrations, with no evidence of departure from linearity (linearity p=0.66; eFigure 1). An interquartile range difference in urinary cadmium (interquartile range width [IQRw]= 0.19 μg/g) yielded a fully adjusted hazard ratio for heart failure of 1.1 (95%CI= 1.0 – 1.2; Table 2).
Table 2.
Adjusted hazard ratio (HR) for incident heart failure per interquartile range difference and quartile of urinary cadmium concentration (creatinine standardized)
| U-Cd (μg/g creatinine) | Total | Cases (n) | U-Cd Median (μg/g creatinine) | Minimally adjusted HRa (95% CI) | Fully adjusted HRb (95% CI) | Trend Test P-value |
|---|---|---|---|---|---|---|
| Continuousc | 2094 | 958 | 1.1 (1.0 – 1.2) | 1.1 (1.0 – 1.2) | 0.048 | |
| Quartile 1: ≤ 0.13 | 495 | 211 | 0.10 | Ref | Ref | |
| Quartile 2: >0.13 – 0.20 | 471 | 195 | 0.17 | 0.89 (0.69 – 1.2) | 0.93 (0.70 – 1.2) | |
| Quartile 3: >0.20 – 0.32 | 544 | 257 | 0.25 | 1.2 (0.90 – 1.5) | 1.2 (0.93 – 1.6) | |
| Quartile 4: > 0.32 | 584 | 295 | 0.45 | 1.2 (0.92 – 1.6) | 1.3 (0.95 – 1.8) |
Minimally adjusted model includes sex.
Fully adjusted model includes: sex (categorical), BMI (continuous), education (categorical), and cotinine (continuous).
Continuous HR for an IQRw =0.19 μg/g difference in creatinine standardized urinary cadmium concentration.
Abbreviation definitions: U-Cd = urinary cadmium; HR = hazard ratio; CI= confidence interval
We found strong evidence suggesting the association between cadmium and incident heart failure varied by sex (sex by cadmium product term p-value=0.012). In sex-stratified analyses, we found a positive trend in the relationship between urinary cadmium and incident heart failure among men. The fully adjusted hazard ratio for heart failure was 2.0 (95% CI= 1.2 – 3.2) comparing men in the highest versus lowest quartile of creatinine-standardized cadmium concentrations (Table 3). Using splines to model cadmium, we report slight evidence of a non-linear U-shape relationship among men (linearity p-value=0.16; eFigure 1). An interquartile range difference (IQRw= 0.19 μg/g) in cadmium was associated with a hazard ratio of 1.5 (95% CI= 1.2 – 1.9) among men.
Table 3.
Adjusted hazard ratio (HR) for incident heart failure per interquartile range increase and quartile of urinary cadmium concentration in sex-stratified analyses
| Male | Female | |||||
|---|---|---|---|---|---|---|
| U-Cd (μg/g creatinine) | Total (n) | Cases (n) | Fully adjusted HR (95% CI) | Total (n) | Cases (n) | Fully adjusted HR (95% CI) |
| Continuousa | 1041 | 481 | 1.5 (1.2 – 1.9) | 1053 | 477 | 1.1 (0.97 – 1.2) |
| Quartile 1: ≤ 0.13 | 385 | 170 | Ref | 110 | 41 | Ref |
| Quartile 2: >0.13 – 0.20 | 313 | 134 | 0.93 (0.67 – 1.3) | 158 | 61 | 0.93 (0.52 – 1.7) |
| Quartile 3: >0.20 – 0.32 | 240 | 113 | 1.1 (0.79 – 1.6) | 304 | 144 | 1.4 (0.81 – 2.3) |
| Quartile 4: > 0.32 | 103 | 64 | 2.0 (1.2 – 3.2) | 481 | 231 | 1.1 (0.68 – 1.8) |
Continuous HR for an IQRw=0.19 μg/g difference in creatinine standardized urinary cadmium concentration.
Adjusted model includes: BMI (continuous-linear), education (categorical-linear), and cotinine (continuous-linear). Female model also adjusted for menopause status.
P-value for sex by cadmium product term in the main fully adjusted model: p= 0.012
Abbreviation definitions: U-Cd = urinary cadmium; HR = hazard ratio; CI= confidence interval
We did not find strong evidence to suggest a relationship between higher urinary cadmium and a higher rate of heart failure among women, with little indication of departure from linearity (linearity p=0.27; eFigure 1). The fully-adjusted hazard ratio for heart failure was 1.1 (95% CI= 0.68 – 1.8) comparing women in the highest versus lowest quartile of creatinine-standardized cadmium concentrations (Table 3). The results were not materially different when we limited the secondary analyses to women who were post-menopausal (eTable 5).
In sensitivity analyses we found that these results were robust to a number of analytic choices. Specifically, both censoring values below the LOD and excluding participants beyond the 5th and 95th percentile of creatinine-standardized cadmium concentrations resulted in slightly higher fully-adjusted hazard ratio estimates and consistent exposure-response relationships compared with the main results (eTable 2). Furthermore, using creatinine adjustment, covariate adjusted creatinine standardization, osmolality adjustment, or osmolality standardization did not meaningfully alter our results (eTable 3). Using these different methods to account for urine dilution, the relationship between urinary cadmium and incident heart failure ranged from aHR=1.1 per IQRw (95%CI= 1.0 – 1.2) to aHR=1.2 per IQRw (95%CI= 1.0 – 1.3). In the sex-stratified analyses accounting for urine dilution with osmolality adjustment or standardization, we found similar, but less pronounced, patterns in the relationship between cadmium and incident heart failure compared with our main results (eTable 4). Our results were not materially different across other sensitivity analyses adjusting for diabetes, hypertension, and hypercholesteremia or excluding participants with high urinary cotinine (≥200 ng/ml) or extreme urinary creatinine levels (<0.03 g/L and >3 g/L) concentrations (eTable 6, eTable 7, and eTable 8). Moreover, the results were consistent across different approaches adjusting for tobacco smoke exposure (eTable 9).
To facilitate interpretation of our results, we estimated the proportion of total heart failure cases in this population that may have been attributable to the highest quartile of urinary cadmium (concentrations above 0.32 μg/g creatinine; Table 4). Approximately 6.9% (95% CI= −1.6 – 13) of all incident heart failure cases, and 6.6% (95% CI= 2.6 – 9.1) of all cases among men, were attributable to the highest quartile of urinary cadmium. We did not find strong evidence that heart failure cases among women are attributable to the highest quartile of urinary cadmium (4.8% (95% CI= −23 – 22).
Table 4.
The proportion of total heart failure cases that may have been attributable to the highest quartile of urinary cadmium (>0.32 μg/g creatinine)
| % of Heart Failure Cases (95% CI) | |
|---|---|
| All Participants | 6.9 (−1.6 – 13) |
| Men | 6.6 (2.6 – 9.1) |
| Women | 4.8 (−23 – 22) |
Abbreviation definition: CI= confidence interval
DISCUSSION
In this cohort of individuals who reported never smoking tobacco, we found that higher urinary cadmium was associated with incident heart failure, primarily among men. We observed this positive association at levels of urinary cadmium quite common in the general population. Our results were robust across a variety of sensitivity analyses. These results differ from our prior studies in which we found no strong evidence of an association between higher urinary cadmium and acute myocardial infarction or stroke among self-reported non-smokers in the DCH cohort (21,22). However, our results are consistent with evidence suggesting that cadmium adversely impacts myocardial structure and function to plausibly increase the risk of heart failure (14–16).
Heart failure is among the most prevalent and costly health conditions in middle- and high-income countries (9,31,32). Given widespread cadmium exposure driven by a number of regulatable sources, the recognition of environmental cadmium as a risk factor for heart failure in the absence of tobacco use has potentially important public health implications.
People are primarily exposed to cadmium through diet and tobacco use, largely due to anthropogenic contamination of land (5). While background levels of cadmium are naturally found in soil, agricultural application of phosphate fertilizers is a major driver of soil cadmium accumulation on croplands (33,34). Additionally, emissions from industrial facilities, such as solid waste incinerators, iron and steel manufacturers, zinc mines, and metal finishing facilities, can contribute to cadmium emissions in the air and deposition on agricultural soils (35–37). Cadmium that accumulates in the soil is subsequently taken up by vegetation, including tobacco plants and virtually all food crops—with about 80% of food cadmium coming from cereals, vegetables, and potatoes (38). To protect human health and the environment, some European countries, including Denmark, have enacted national policies to limit cadmium levels in phosphate fertilizers. Furthermore, the European Union recently adopted regulations (effective in 2022) to expand similar protections across all member states (39,40). Both regulatory approaches are more restrictive than regulations in the United States that have only been implemented in three states (5,20).
Cadmium exposure may directly contribute to left ventricular dysfunction, as well as indirectly impair cardiovascular health by disrupting the function of other organs, such as the kidneys (14,16–19). Decades of experimental and observational research suggest that cadmium exposure can contribute to endothelial dysfunction, promote atherosclerosis, and disrupt cardiac function by inducing subtle alterations within cardiovascular tissue (14,15,41–47). Cadmium exposure may be linked to cardiovascular disease risk through mechanisms involving inflammatory pathways, epigenetic modifications, and endocrine disruption (7,45,48–50).
Our results are similar to, but more precise than, prior epidemiologic studies suggesting that cadmium exposure is associated with incident heart failure (10–12). For example, in a study of tobacco smokers and non-smokers, Tellez-Plaza et al. (2013) found a higher rate of heart failure among participants in the highest urinary cadmium quartile compared with the lowest quartile (aHR=1.61; 95% CI=1.10 – 2.36) while adjusting for tobacco use. However, since tobacco smoking is a major source of cadmium exposure, these studies report substantially higher concentrations of blood or urine cadmium for former and current tobacco users compared with non-users, driving concerns that statistical adjustment may not sufficiently control for cardiovascular risk associated with tobacco use (11,51,52). The prior studies that have estimated relationships between cadmium and incident heart failure among subgroups of never smokers report imprecise results (10–12). For example, in the same study, Tellez-Plaza et al. (2013) report an attenuated relationship of cadmium concentrations with incident heart failure among never smokers (aHR=1.18 comparing 80th and 20th percentiles of cadmium; 95% CI 0.68 – 2.05). The consistency of the results across these prior studies and our analysis suggests that the association between cadmium and heart failure is not due to residual confounding by tobacco smoking. Therefore, our results help to further establish environmental cadmium exposure, at the relatively low levels typically seen in non-smoking individuals, as a risk factor for heart failure among the general population.
In general, risk factors for and incidence of heart failure, and its etiologic subtypes, have been found to vary by sex (9,53–56). For example, among the Danish population less than age 74 years, the incidence of heart failure was higher among men than women between 1995 and 2012 (54). Studies across Europe and the United States also suggest that men have a higher risk specifically of heart failure with reduced ejection fraction, primarily due to artery disease and myocardial infarction, while women tend to have a higher risk of heart failure with preserved ejection fraction (57).
Our sex-stratified analyses suggest that, among men, higher cadmium concentrations were associated with a higher rate of incident heart failure which is consistent with some prior research (10). However, not all studies have shown sex differences in the relationship between cadmium and incident heart failure (11,12) and this question remains unsettled. Compared with men, women tend to absorb cadmium at higher rates due to lower iron levels, primarily during reproductive ages; therefore, women may have higher cadmium concentrations in urine, blood, and the kidneys (58,59). However, the sex-specific relations of cadmium with different health outcomes could vary based on the underlying physiology and etiological involvement of hormones (60,61). Cadmium-associated heart failure risk, in particular, may vary by sex due to differences in heart failure etiology and pathophysiology, cardiac structure, and prevalent comorbidities (53,55). Our results warrant further investigation with additional consideration of factors that may contribute to sex-specific confounding.
Our study population is predominantly non-Hispanic white. Therefore, our results may not be generalizable to other racial and ethnic groups who have different cadmium exposure patterns or risk factors for heart failure. We expect our results to be generalizable to other Scandinavian and non-Hispanic white populations of never smokers with similar cadmium exposure levels.
Our study has several additional limitations. First, although we restricted our study to self-reported never smokers and adjusted for several potential confounders, we cannot exclude the possibility of unmeasured confounding factors. For example, we did not adjust for potential confounding by dietary factors. However, it seems unlikely that diet is an important confounder since cadmium is found in a wide variety of foods, including many that are nutritionally beneficial for cardiovascular health. Residual confounding by income is also possible and we did not consider other metal biomarkers that could be correlated with cadmium concentrations and associated with heart failure. We also were unable to control for chronic kidney disease (CKD) or estimated glomerular filtration rate as a marker of progressive kidney disease, which may be potential confounders. It is possible that cadmium could impact renal function to subsequently increase cardiovascular disease risk, as well as affect urinary cadmium biomarker concentrations. However, our results were similar across sensitivity analyses accounting for several other related health conditions. Second, although the use of the Danish NPR for identification of heart failure cases is specific and has been shown to have a high positive predictive value (62,63), we could have missed incident heart failure cases during the follow-up period because our approach was less sensitive (64). However, we do not expect this underreporting of heart failure cases to be related to cadmium exposure; therefore, we expect that if we included more cases with the same exposure distribution it would only increase the power and precision of the associations. Third, the lack of information extracted from the NPR on the pathophysiologic etiology limited our ability to evaluate sex-specific results in more detail. Among the women, we observed a small association with wide confidence intervals. Therefore, with a larger sample of women it is plausible that we would have made different inferences about the association. Fourth, we were unable to identify prevalent heart failure among subcohort members who did not have an additional diagnosis of heart failure in the NPR during the follow-up period. However, the prevalence of heart failure at baseline was less than 2%; therefore, the impact of this source of misclassification is likely to be small. Last, it is possible that for some participants in the subcohort and larger DCH cohort, death may have precluded development of clinically-recognized heart failure, and therefore could be considered a competing risk. However, we do not expect this to substantially impact our results since few subcohort members were censored due to death.
Our study has several notable strengths. First, we used data from a Danish prospective, population-based study with a long follow-up period for our analysis. Second, we quantified cadmium concentrations in urine samples collected at baseline, which is a proxy measure of long-term exposure. Urinary cadmium is considered a biomarker for long-term cadmium exposure because concentrations reflect the level of cadmium that has bioaccumulated in the kidney, where it has an estimated half-life of 10–30 years (5,33). Third, we assessed both urinary creatinine and osmolality and showed little difference across methods accounting for urinary dilution.
In conclusion, our study suggests that higher urinary cadmium concentration, a biomarker of long-term exposure, is positively associated with incident heart failure, especially among men. Our results extend findings from prior epidemiologic studies by providing evidence of an association between environmental cadmium exposure and incident heart failure, independent of tobacco smoking. Assuming that the observed associations are causal and minimally biased, our results estimate that approximately 6.9% of heart failure cases, primarily among men, were attributable to environmental cadmium exposure resulting in urinary concentrations greater than 0.32 μg/g creatinine. Our results increase confidence that exposure to environmental cadmium is a risk factor for heart failure.
Supplementary Material
Sources of Funding:
The results reported herein correspond to specific aims of grant R01ES026614 to investigator Jaymie Meliker from the National Institute of Environmental Health Sciences. The establishment and running of The Diet, Cancer and Health cohort was funded by the Danish Cancer Society.
Footnotes
Conflicts of Interest: Dr. Wellenius has received consulting fees from the Health Effects institute (Boston, MA) and recently served as a paid visiting scientist at Google LLC (Mountain View, CA). The other authors declare that there is no conflict of interest.
Data availability:
The data underlying this article cannot be shared publicly for the privacy of individuals that participated in the study. The data underlying this article were provided by The Diet, Cancer and Health Cohort with permission for this article.
REFERENCES
- 1.Navas-Acien A, Tellez-Plaza M, Guallar E, Muntner P, Silbergeld E, Jaar B, et al. Blood cadmium and lead and chronic kidney disease in US adults: A joint analysis. Am J Epidemiol. 2009;170(9):1156–1164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Tellez-Plaza M, Navas-Acien A, Crainiceanu CM, Guallar E. Cadmium exposure and hypertension in the 1999–2004 National Health and Nutrition Examination Survey (NHANES). Environ Health Perspect. 2008;116(1):51–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Tellez-Plaza M, Navas-Acien A, Menke A, Crainiceanu CM, Pastor-Barriuso R, Guallar E. Cadmium exposure and all-cause and cardiovascular mortality in the U.S. general population. Environ Health Perspect. 2012;120(7):1017–1022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.International Agency for Research on Cancer (IARC). Cadmium and Cadmium Compounds. In: Beryllium, Cadmium, Mercury, and Exposures in the Glass Manufacturing Industry. IARC Monographs on the Evaluation of Carcinogenic risk to Humans, No. 58. Lyon, FR: IARC; 1993. [PMC free article] [PubMed] [Google Scholar]
- 5.Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for Cadmium. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service; 2012. [PubMed] [Google Scholar]
- 6.Tellez-Plaza M, Jones MR, Dominguez-Lucas A, Guallar E, Navas-Acien A. Cadmium exposure and clinical cardiovascular disease: A systematic review. Curr Atheroscler Rep. 2013;15(10):356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Tinkov AA, Filippini T, Ajsuvakova OP, Skalnaya MG, Aaseth J, Bjørklund G, et al. Cadmium and atherosclerosis: A review of toxicological mechanisms and a meta-analysis of epidemiologic studies. Environ Res. 2018;162:240–260. [DOI] [PubMed] [Google Scholar]
- 8.Chowdhury R, Ramond A, O’Keeffe LM, Shahzad S, Kunutsor SK, Muka T, et al. Environmental toxic metal contaminants and risk of cardiovascular disease: Systematic review and meta-analysis. BMJ. 2018;k3310. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Benjamin EJ, Muntner P, Alonso A, Bittencourt MS, Callaway CW, Carson AP, et al. Heart disease and stroke statistics—2019 update: A report from the American Heart Association. Circulation. 2019;139(10). Available from: https://www.ahajournals.org/doi/10.1161/CIR.0000000000000659 [DOI] [PubMed] [Google Scholar]
- 10.Borne Y, Barregard L, Persson M, Hedblad B, Fagerberg B, Engstrom G. Cadmium exposure and incidence of heart failure and atrial fibrillation: A population-based prospective cohort study. BMJ Open. 2015;5(6):e007366–e007366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Peters JL, Perlstein TS, Perry MJ, McNeely E, Weuve J. Cadmium exposure in association with history of stroke and heart failure. Environ Res. 2010;110(2):199–206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Tellez-Plaza M, Guallar E, Howard BV, Umans JG, Francesconi KA, Goessler W, et al. Cadmium exposure and incident cardiovascular disease. Epidemiology. 2013;24(3):421–429. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Bui AL, Horwich TB, Fonarow GC. Epidemiology and risk profile of heart failure. Nat Rev Cardiol. 2011;8(1):30–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kopp SJ, Perry M, Glonek T, Erlanger M, Perry EF, Barany M, et al. Cardiac physiologic-metabolic changes after chronic low-level heavy metal feeding. Am J Physiol-Heart Circ Physiol. 1980;239(1):H22–30. [DOI] [PubMed] [Google Scholar]
- 15.Kopp SJ, Bárány M, Erlanger M, Perry EF, Perry HM. The influence of chronic low-level cadmium and/or lead feeding on myocardial contractility related to phosphorylation of cardiac myofibrillar proteins. Toxicol Appl Pharmacol. 1980;54(1):48–56. [DOI] [PubMed] [Google Scholar]
- 16.Yang W, Zhang Z, Thijs L, Cauwenberghs N, Wei F, Jacobs L, et al. Left ventricular structure and function in relation to environmental exposure to lead and cadmium. J Am Heart Assoc. 2017;6(2):e004692. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nishijo M, Nakagawa H, Morikawa Y, Tabata M, Senma M, Miura K, et al. Mortality of inhabitants in an area polluted by cadmium: 15 year follow up. Occup Environ Med. 1995;52(3):181–184. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Nishijo M Causes of death and renal tubular dysfunction in residents exposed to cadmium in the environment. Occup Environ Med. 2006;63(8):545–550. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Staessen JA, Amery A, Lauwerys RR, Roels HA, Ide G, Vyncke G. Renal function and historical environmental cadmium pollution from zinc smelters. The Lancet. 1994;343(8912):1523–1527. [DOI] [PubMed] [Google Scholar]
- 20.Ulrich AE. Cadmium governance in Europe’s phosphate fertilizers: Not so fast? Sci Total Environ. 2019;650:541–545. [DOI] [PubMed] [Google Scholar]
- 21.Sears CG, Poulsen AH, Eliot M, Howe CJ, James KA, Harrington JM, et al. Urine cadmium and acute myocardial infarction among never smokers in the Danish Diet, Cancer and Health cohort. Environ Int. 2021;150:106428. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Poulsen AH, Sears CG, Harrington J, Howe CJ, James KA, Roswall N, et al. Urinary cadmium and stroke - a case–cohort study in Danish never-smokers. Environ Res. 2021;200:111394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Tjønneland A, Olsen A, Boll K, Stripp C, Christensen J, Engholm G, et al. Study design, exposure variables, and socioeconomic determinants of participation in Diet, Cancer and Health: A population-based prospective cohort study of 57,053 men and women in Denmark. Scand J Public Health. 2007;35(4):432–441. [DOI] [PubMed] [Google Scholar]
- 24.Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Health. 2011;39(7):30–33. [DOI] [PubMed] [Google Scholar]
- 25.Barlow WE, Ichikawa L, Rosner D, Izumi S. Analysis of case–cohort designs. J Clin Epidemiol. 1999;52(12):1165–1172. [DOI] [PubMed] [Google Scholar]
- 26.Thiébaut ACM, Bénichou J. Choice of time-scale in Cox’s model analysis of epidemiologic cohort data: A simulation study. Stat Med. 2004. Dec 30;23(24):3803–20. [DOI] [PubMed] [Google Scholar]
- 27.Hernán MA. Causal knowledge as a prerequisite for confounding evaluation: An application to birth defects epidemiology. Am J Epidemiol. 2002;155(2):176–184. [DOI] [PubMed] [Google Scholar]
- 28.Rothman KJ, Greenland S, Lash TL. Modern Epidemiology. Third edition. Philadelphia, PA: Wolters Kluwer/Lippincott Williams & Wilkins; 2008. [Google Scholar]
- 29.VanderWeele TJ. Principles of confounder selection. Eur J Epidemiol. 2019;34(3):211–219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Rothman K Epidemiology: An Introduction. 2nd Edition. New York, NY: Oxford University Press; 2012. [Google Scholar]
- 31.Heidenreich PA, Albert NM, Allen LA, Bluemke DA, Butler J, Fonarow GC, et al. Forecasting the impact of heart failure in the United States: A policy statement from the American Heart Association. Circ Heart Fail. 2013;6(3):606–619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.van Riet EES, Hoes AW, Wagenaar KP, Limburg A, Landman MAJ, Rutten FH. Epidemiology of heart failure: The prevalence of heart failure and ventricular dysfunction in older adults over time. A systematic review. Eur J Heart Fail. 2016;18(3):242–252. [DOI] [PubMed] [Google Scholar]
- 33.Järup L, Åkesson A. Current status of cadmium as an environmental health problem. Toxicol Appl Pharmacol. 2009;238(3):201–208. [DOI] [PubMed] [Google Scholar]
- 34.Singh BR. Trace element availability to plants in agricultural soils, with special emphasis on fertilizer inputs. Environ Rev. 1994;2(2):133–146. [Google Scholar]
- 35.Exposure Kjellström T. and accumulation of cadmium in populations from Japan, the United States, and Sweden. Environ Health Perspect. 1979;28:169–197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Mobley J, Rackley K, Pope A, Brown D. Persistent bioaccumulative toxic emissions in the US. Ann Arbor, MI: National Exposure Research Laboratory Office of Air Quality Planning and Standards, US EPA; 2003. p. 41. [Google Scholar]
- 37.Plachy J Cadmium recycling in the United States in 2000. Reston, VA: U.S. Geological Survey; 2003. p. 14. [Google Scholar]
- 38.Olsson I-M, Eriksson J, Öborn I, Skerfving S, Oskarsson A. Cadmium in food production systems: A health risk for sensitive population groups. Ambio. 2005;34(4):344–51. [DOI] [PubMed] [Google Scholar]
- 39.European Council (EC). Regulation (EU) 2019/1009 of the European Parliament and of the Council of 5 June 2019 laying down rules on the making available on the market of EU fertilising products and amending regulations (EC) no 1069/2009 and (EC) no 1107/2009 and Repealing Regulation (EC) no 2003/2003. Official Journal of the European Union. 2019. [Google Scholar]
- 40.Marini M, Caro D, Thomsen M. The new fertilizer regulation: A starting point for cadmium control in European arable soils? Sci Total Environ. 2020;745:140876. [DOI] [PubMed] [Google Scholar]
- 41.Angeli JK, Cruz Pereira CA, de Oliveira Faria T, Stefanon I, Padilha AS, Vassallo DV. Cadmium exposure induces vascular injury due to endothelial oxidative stress: the role of local angiotensin II and COX-2. Free Radic Biol Med. 2013;65:838–848. [DOI] [PubMed] [Google Scholar]
- 42.Fagerberg B, Barregard L, Sallsten G, Forsgard N, Östling G, Persson M, et al. Cadmium exposure and atherosclerotic carotid plaques –Results from the Malmö diet and Cancer study. Environ Res. 2015;136:67–74. [DOI] [PubMed] [Google Scholar]
- 43.Fujiwara Y, Watanabe S, Kaji T. Promotion of cultured vascular smooth muscle cell proliferation by low levels of cadmium. Toxicol Lett. 1998;94(3):175–180. [DOI] [PubMed] [Google Scholar]
- 44.Kaji T, Ohkawara S, Inada M, Yamamoto C, Sakamoto M, Kozuka H. Cadmium stimulation of glycosaminoglycan synthesis by cultured vascular endothelial cells: Comparison of various cell types. Biol Pharm Bull. 1994;17(3):454–457. [DOI] [PubMed] [Google Scholar]
- 45.Knoflach M, Messner B, Shen YH, Frotschnig S, Liu G, Pfaller K, et al. Non-toxic cadmium concentrations induce vascular inflammation and promote atherosclerosis. Circ J. 2011;75(10):2491–2495. [DOI] [PubMed] [Google Scholar]
- 46.Messner B, Knoflach M, Seubert A, Ritsch A, Pfaller K, Henderson B, et al. Cadmium is a novel and independent risk factor for early atherosclerosis mechanisms and in vivo relevance. Arterioscler Thromb Vasc Biol. 2009;29(9):1392–1398. [DOI] [PubMed] [Google Scholar]
- 47.Yamamoto C, Kaji T, Sakamoto M, Kozuka H. Cadmium stimulation of plasminogen activator inhibitor-1 release from human vascular endothelial cells in culture. Toxicology. 1993;83(1–3):215–223. [DOI] [PubMed] [Google Scholar]
- 48.Kluxen FM, Höfer N, Kretzschmar G, Degen GH, Diel P. Cadmium modulates expression of aryl hydrocarbon receptor-associated genes in rat uterus by interaction with the estrogen receptor. Arch Toxicol. 2012;86(4):591–601. [DOI] [PubMed] [Google Scholar]
- 49.Stoica A, Katzenellenbogen BS, Martin MB. Activation of estrogen receptor-α by the heavy metal cadmium. Mol Endocrinol. 2000;14(4):545–553. [DOI] [PubMed] [Google Scholar]
- 50.Wang B, Li Y, Shao C, Tan Y, Cai L. Cadmium and its epigenetic effects. Curr Med Chem. 2012;19(16):2611–2620. [DOI] [PubMed] [Google Scholar]
- 51.Adams SV, Newcomb PA. Cadmium blood and urine concentrations as measures of exposure: NHANES 1999–2010. J Expo Sci Environ Epidemiol. 2014;24(2):163–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Berglund M, Larsson K, Grandér M, Casteleyn L, Kolossa-Gehring M, Schwedler G, et al. Exposure determinants of cadmium in European mothers and their children. Environ Res. 2015;141:69–76. [DOI] [PubMed] [Google Scholar]
- 53.Beale AL, Meyer P, Marwick TH, Lam CSP, Kaye DM. Sex differences in cardiovascular pathophysiology: Why women are overrepresented in heart failure with preserved ejection fraction. Circulation. 2018;138(2):198–205. [DOI] [PubMed] [Google Scholar]
- 54.Christiansen MN, Køber L, Weeke P, Vasan RS, Jeppesen JL, Smith JG, et al. Age-specific trends in incidence, mortality, and comorbidities of heart failure in Denmark, 1995 to 2012. Circulation. 2017;135(13):1214–23. [DOI] [PubMed] [Google Scholar]
- 55.Cleland J The EuroHeart Failure survey programme—a survey on the quality of care among patients with heart failure in Europe Part 1: patient characteristics and diagnosis. Eur Heart J. 2003;24(5):442–63. [DOI] [PubMed] [Google Scholar]
- 56.Levy D, Kenchaiah S, Larson MG, Benjamin EJ, Kupka MJ, Ho KKL, et al. Long-term trends in the incidence of and survival with heart failure. N Engl J Med. 2002;347(18):1397–402. [DOI] [PubMed] [Google Scholar]
- 57.Lam CSP, Arnott C, Beale AL, Chandramouli C, Hilfiker-Kleiner D, Kaye DM, et al. Sex differences in heart failure. Eur Heart J. 2019;40(47):3859–3868c. [DOI] [PubMed] [Google Scholar]
- 58.Vahter M, Akesson A, Lidén C, Ceccatelli S, Berglund M. Gender differences in the disposition and toxicity of metals. Environ Res. 2007;104(1):85–95. [DOI] [PubMed] [Google Scholar]
- 59.Nishijo M, Satarug S, Honda R, Tsuritani I, Aoshima K. The gender differences in health effects of environmental cadmium exposure and potential mechanisms. Mol Cell Biochem. 2004;255(1–2):87–92. [DOI] [PubMed] [Google Scholar]
- 60.Ali I, Engström A, Vahter M, Skerfving S, Lundh T, Lidfeldt J, et al. Associations between cadmium exposure and circulating levels of sex hormones in postmenopausal women. Environ Res. 2014;134:265–269. [DOI] [PubMed] [Google Scholar]
- 61.Bochud M, Jenny-Burri J, Pruijm M, Ponte B, Guessous I, Ehret G, et al. Urinary cadmium excretion is associated with increased synthesis of cortico- and sex steroids in a population study. J Clin Endocrinol Metab. 2018;103(2):748–758. [DOI] [PubMed] [Google Scholar]
- 62.Mard S, Nielesen FE. Positive predictive value and impact of misdiagnosis of a heart failure diagnosis in administrative registers among patients admitted to a University Hospital cardiac care unit. Clin Epidemiol. 2010;235. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 63.Thygesen SK, Christiansen CF, Christensen S, Lash TL, Sørensen HT. The predictive value of ICD-10 diagnostic coding used to assess Charlson comorbidity index conditions in the population-based Danish National Registry of Patients. BMC Med Res Methodol. 2011;11(1):83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 64.Kümler T, Gislason GH, Kirk V, Bay M, Nielsen OW, Køber L, et al. Accuracy of a heart failure diagnosis in administrative registers. Eur J Heart Fail. 2008;10(7):658–660. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article cannot be shared publicly for the privacy of individuals that participated in the study. The data underlying this article were provided by The Diet, Cancer and Health Cohort with permission for this article.