Association of blood cadmium level with nonmelanoma skin cancer: a cross-sectional study

doi:10.21203/rs.3.rs-2357139/v1

Download PDF

Research Article

Association of blood cadmium level with nonmelanoma skin cancer: a cross-sectional study

https://doi.org/10.21203/rs.3.rs-2357139/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Purpose

Cadmium is widely used in industries and found in cigarette smoke. While animal studies have shown some potential effects of cadmium exposure on the skin, little has been known about the role of chronic cadmium exposure on the development of nonmelanoma skin cancer (NMSC). We conducted this study to examine whether blood cadmium was related to the risk of NMSC.

Methods

An analysis of cross-sectional data from the National Health and Nutrition Examination Survey from 1999–2000 to 2017–2018 was conducted. NMSC was ascertained based on self-reported physician diagnoses. Taking the cadmium level as either a continuous variable or a quartile variable, we examined the association between the blood cadmium level and the odds of NMSC with a multivariable logistic regression model. Restricted cubic splines (RCS) were added to the regression model the estimate the non-linear relationship between cadmium and NMSC.

Results

A total of 41577 participants were included, of which 952 were reported to have NMSC and 40625 did not have any caners. Cadmium was not associated with NMSC when it was taken as a continuous or quartile variable. RCS showed an inverted U-shaped curve between the blood cadmium level and the odds of NMSC. The result remained similar when limiting the participants to non-Hispanic White.

Conclusions

Blood cadmium level was associated with the risk of NMSC. Further studies are needed to explore the causality and underlying mechanisms.

Cadmium

Nonmelanoma skin cancer

Restricted cubic splines

National Health and Nutrition Examination Survey

Nonmelanoma skin cancer (NMSC) is the most common malignancy in people with fair skin, including basal cell carcinoma, squamous cell carcinoma, and other rarer skin tumors[2]. The incidence of NMSC is increasing globally with the aging population[14]. According to the World Health Organization, approximately 2 to 3 million NMSC occur worldwide every year recently[21]. Due to the high prevalence, the cost of treatment for NMSC is a great financial burden for patients and society. In the United States (US), the average cost of treating NMSC was estimated at 4.8 billion dollars each year[1]. Understanding the risk factors is critical for the prevention or early detection of NMSC in order to reduce its burden.

Cadmium is an element of low concentrations in the earth’s crust. There are a number of environmental cadmium pollution sources, including fossil fuels, nonferrous industries, fertilizers, and sewage sludge[24]. The general population is exposed to cadmium through polluted water and food, and cigarettes are another significant source of exposure[17]. Cadmium is highly toxic and classified as a group I carcinogen by the International Agency for Research on Cancer. It was associated with tumors in different parts of the human body such as the lung, liver, kidney, blood, prostate, etc[28]. Animal studies have indicated that there might be an interaction between cadmium and keratinocytes[7, 19]. However, to our knowledge, there are no epidemiologic studies investigating the association between cadmium and NMSC. Thus, we used data from the National Health and Nutrition Examination Survey (NHANES) to examine whether the blood cadmium was related to the risk of NMSC.

Study population

This study was conducted using cross-sectional data of NHANES from 1999–2000 to 2017–2018. The NHANES data was openly available online[3]. In brief, it is an ongoing complex sample survey that includes in-home interviews, physical examinations, and laboratory tests. It collects information on the demographics, health, and nutritional status of the US civilian noninstitutionalized population. Since 1999, data was collected continuously and released in 2-year cycles. The study protocols of NHANES were approved by the National Center for Health Statistics Institutional review board and written informed consent was obtained from each participant[4]. Our participants were limited to those aged 20 years or older. Those without information on the measurement of blood cadmium level or the diagnosis of cancers were excluded.

Exposure and outcome measurements

The exposure of interest is whole-blood cadmium level. During physical examinations, blood samples were collected from participants. The samples were processed and sent to a central laboratory for testing. A detailed description of the laboratory methodology of whole-blood cadmium measurement can be found on the NHANES website. In brief, whole-blood cadmium concentration was determined using inductively coupled plasma mass spectrometry (ICP-MS). This multi-element analytical technique is based on quadrupole ICP-MS technology. For values below the detection limit, the square root of two was substituted for the detection limit.

The outcome was based on the self-reported personal interview data collected from the medical condition questionnaire. If participants answered “yes” to the question “Have you ever been told by a doctor or other health professional that you had a cancer or a malignancy of any kind”, they would further answer the question “What kind of cancer”. We created a binary variable for participants with NMSC or without any cancers.

Covariates

Important covariates were determined in the literature on NMSC, and included were: age, gender, race/ethnicity, education, income, body mass index (BMI), hypertension, diabetes, dyslipidemia, current smoke, alcohol use, hours of physical activity per week, and sunscreen or protective clothing. There were four categories of race/ethnicity: non-Hispanic white, non-Hispanic black, Mexican American, and others Education included less than high school, high school graduate, some college, and college graduate or higher. Income was determined by the income-to-poverty ratio, which was defined as annual family income divided by the poverty threshold after adjusting for inflation and family size. BMI was calculated as the weight (kilograms) divided by height (meters). Hypertension was defined as systolic BP of 130 mm Hg or higher, diastolic BP of 80 mm Hg or higher, or the use of antihypertensive medications. Diabetes was defined as hemoglobin A1c of 6.5% or higher, or the use of antidiabetic medications. Dyslipidemia was defined as total cholesterol of 240 mg/dL or higher, or the use of lipid-lowering medications. Current smoke was evaluated based on questions about whether participants were currently smoking or not. Alcohol use was classified as equal to or less than 1 drink per week, between 1 to 5 drinks per week, and more than 5 drinks per week. By a drink, it means a 12 oz. beer, a 5 oz. glass of wine, or one and half ounces of liquor. Physical activity was calculated as hours of moderate-intensity activity plus twice the minutes of vigorous-intensity activity[6]. Sunscreen or protective clothing was based on questions of whether participants use sunscreen or wear a long-sleeved shirt.

Statistical analysis

All statistical analyses were conducted in R version 4.1.3 using the “Survey” package. The complex sampling design was considered during the analysis. Our analysis followed NHANES recommendations for combining survey cycles and generating sample weights for subsamples to ensure accurate association and variance estimation. All statistical tests were 2-sided and P < 0.05 was considered statistically significant.

Continuous variables were presented as means and categorical variables as percentages. The levels of cadmium were right-skewed and natural log-transformed for the analysis. The difference in continuous variables between those with NMSC and without cancer was examined with a t-test and the difference in categorical variables was examined with a Chi-squared test. Binomial logistic regression analysis was applied to investigate the association between cadmium level and the odds of NMSC. After natural log-transformation, the cadmium level was first modeled as a continuous variable in the regression analysis. Second, it was divided into quartiles and modeled as an ordinal variable with the first quartile being the reference. Third, to identify the possible non-linear relationship between cadmium level and the odds of NMSC, restricted cubic splines (RCS) were used in the regression analysis. Three knots were applied in the RCS in order to better fit the model while avoiding overfitting. The P-value for the non-linear trend was determined by the Wald test for the coefficient of RCS. In addition, to test whether the association between cadmium and NMSC differed in different subgroups by age, gender, race, education, and income, a two-way interaction term between the quartile of cadmium level and subgroup status was added to the regression model. Further stratified analysis would be conducted if a difference was detected in a specific subgroup.

We developed 3 models for the regression analysis. No covariables were adjusted in model 1, and age, gender, race/ethnicity, education, and income were adjusted in model 2. In model 3, BMI, hypertension, diabetes, dyslipidemia, current smoke, alcohol use, hours of physical activity, and sunscreen or protective clothing were further adjusted based on model 2.

A total of 112339 participants were included in NHANES from 1999–2000 to 2017–2018. Among them, 55513 were adults aged 20 years or older. Of the 55513 participants, 13936 were further excluded due to the missing information on the cadmium level or the diagnosis of cancers. This resulted in a total of 41577 subjects available for analysis, of which 952 were reported to have NMSC and 40625 did not have any cancers.

The differences between participants with NMSC and without any cancers were reported in Table 1. Compared to those without any cancers, individuals with NMSC were older and more likely to be men and non-Hispanic White, received less education, had lower income, and were more likely to have hypertension, diabetes, and dyslipidemia. They were also less likely to engage in current smoking and physical activity. In addition, they were more likely to use sunscreen or had protective clothing compared to their controls.

Table 1

Characteristics of participants with and without NMSC
Characteristics	With NMSC	Without any cancers	P-value
Total, n	952	40625
Mean age, year	63.4 (0.6)	45.2 (0.2)	< 0.001
Men, %	53.3 (2.1))	48.7 (0.3)	0.038
Non-Hispanic White, %	97.5 (0.5)	67.0 (1.1)	< 0.001
Less than high school education, %	10.6 (1.1)	18.1 (0.5)	< 0.001
Income-to-poverty ratio < 1.3, %	7.7 (1.1)	20.5 (0.5)	< 0.001
BMI	28.4 (0.3)	28.6 (0.1)	0.476
Hypertension, %	68.1 (2.0)	44.9 (0.5)	< 0.001
Diabetes, %	13.3 (1.3)	9.3 (0.2)	< 0.001
Dyslipidemia, %	40.6 (2.2)	25.6 (0.4)	< 0.001
Current smoke, %	11.5 (1.4)	22.2 (0.4)	< 0.001
Alcohol use, %			0.227
<= 1 drinks per week	53.8 (2.8)	56.6 (0.8)
>1, <=5 drinks per week	20.3 (2.0)	21.2 (0.5)
> 5 drinks per week	25.9 (2.3)	22.1 (0.5)
Hours of physical activity per week	10.1 (0.8)	16.4 (0.3)	< 0.001
Sunscreen or protective clothing, %	76.9 (4.6)	62.7 (0.6)	0.008
Cadmium, mg/dL	0.37 (0.02)	0.36 (0.01)	0.415
All the data were presented as mean or percentage with standard error in parenthesis. *The level of cadmium was presented as geometric mean. NMSC: nonmelanoma skin cancer; BMI: body mass index

When taken as a continuous variable, the association between cadmium level and the odds of NMSC was insignificant in model 1[Odds ratio (OR), 1.04; 95% confidence interval (95% CI), 0.95–1.14], model 2 (OR, 0.944; 95% CI, 0.83–1.08), and model 3 (OR, 1.02; 95% CI, 0.54–1.93). When taken as a quartile, the association between cadmium level and the odds of NMSC was significant only in model 1 (P for trend, 0.011) and those in the second and third quartiles had higher odds of NMSC compared to those in the first quartile (Table 2). The results of regression analysis with RCS were presented in Fig. 1. We found that cadmium was significantly associated with NMSC in an inverted U-shaped manner in model 1, model 2, and model 3 (all P values for non-linearity were less than 0.05).

Table 2

Odds ratio (95% confidence interval) of NMSC associated with quartiles of cadmium levels
	Quartile 1 (≤ 0.20 ug/L)	Quartile 2 (0.20–0.33 ug/L)	Quartile 3 (0.33–0.60 ug/L)	Quartile 4 (> 0.60 ug/L)	P for trend
NMSC (n)	152	215	390	195
Model 1	Reference	1.89 (1.46–2.44)	2.13 (1.65–2.74)	1.26(0.96–1.66)	0.011
Model 2	Reference	1.27 (0.96–1.68)	1.17 (0.89–1.55)	0.99 (0.74–1.33)	0.849
Model 3	Reference	1.25 (0.50–3.14)	1.51 (0.65–3.50)	0.90 (0.20–4.08)	0.797
NMSC: nonmelanoma skin cancer. No covariates were adjusted in model 1. Age, gender, race, education, and income were adjusted in model 2. Age, gender, race, education, income, body mass index, hypertension, diabetes, dyslipidemia, current smoke, alcohol use, hours of physical activity, and sunscreen or protective clothing were adjusted in model 3.

The association between cadmium and NMSC was broadly similar across different subgroups except for race (Table 3). According to our result, most patients with NMSC were non-Hispanic White (97.5%). When limiting our participants to non-Hispanic White, RCS showed that cadmium was still associated with NMSC in a similar non-linear manner in all three models (Fig. 2).

Table 3

P values for the interaction of cadmium levels by quartiles with different subgroups
	Age	Gender	Race	Education	Income
Model 1	0.919	0.831	< 0.001	0.694	0.626
Model 2	0.779	0.346	0.043	0.353	0.556
Model 3	0.278	0.831	0.210	0.571	0.476
No covariates were adjusted in model 1. Age, gender, race, education, and income were adjusted in model 2. Age, gender, race, education, income, body mass index, hypertension, diabetes, dyslipidemia, current smoke, alcohol use, hours of physical activity, and sunscreen or protective clothing were adjusted in model 3.

Our study investigated the relationship between low blood cadmium levels and the odds of NMSC by using data from a representative sample of the US population. While the cadmium was not associated with NMSC when it was taken as a continuous or quartile variable, we did find a non-linear relationship between them even after adjusting for the potential covariates. In addition, this relationship remained significant when limiting the participants to the non-Hispanic White.

As a toxic heavy metal, cadmium has significant adverse effects on both occupational and environmental health [9]. As a cumulative toxin, cadmium has an extremely long biological half-life[9]. It has been recognized as a human carcinogen by the World Health Organization's International Agency for Research on Cancer and the US National Toxicology Program [16, 20]. According to most research, occupational exposure to cadmium increases the risk of lung cancer, making lung the most established site for cadmium-induced cancer [5, 25]. Cadmium's recognition as a human carcinogen by regulatory agencies is largely due to the established relationship between occupational exposure to cadmium and lung cancer [16, 20]. Cadmium has also been linked to cancer of the liver, prostate, kidney, bladder, hematopoietic system, and stomach in humans [11, 12, 15, 27, 28]. Nevertheless, the link between cadmium exposure and cancers at other specific sites needs further epidemiological and experimental investigation.

Only a few studies have investigated the potential carcinogenesis of cadmium exposure in human skin. In Austria, Weinlich et al. enrolled 760 patients with primary cutaneous melanoma in a prospective cohort study. They found that the increased level of metallothionein, a protein to which cadmium was highly attracted, was linked to a higher risk of disease progression[29]. In Italy, Vinceti et al enrolled 58 controls with cutaneous melanoma and 58 controls, and they measured the concentration of cadmium in the toenails of these subjects. However, the toenail level of cadmium was not related to melanoma[26]. In another study conducted by Vinceti et al., they evaluated the intake of cadmium among 380 participants with melanoma and 719 controls and found that there was a positive non-linear association between cadmium intake and odds of melanoma[8]. These studies together suggested that there might be a possible but not very strong link between cadmium exposure and the risk of melanoma.

Currently, there were no human studies that had ever investigated cadmium and NMSC. However, cadmium has been seen as a co-mutagen with ultraviolet radiation, which can lead to the damage of keratinocytes and melanocytes [13]. In addition, the results from experimental studies might also indicate the potential correlation between cadmium and NMSC. In an in vitro study, Wester et al. demonstrated that human skin absorbs more cadmium from contaminated soil and water than plasma[30]. Later, Lansdown et al. applied cadmium chloride solution to the shaved skin of rats and mice for 10 days. Dose-related damages of dermal hyperkeratosis and acanthosis were observed with ulcerative changes. There was also an increase in the mitotic index of skin cells[19]. The animal study strongly suggested that cadmium can be absorbed by the skin while the actual mechanisms require further investigation. Perhaps it can bind to sulfhydryl radicals of cysteine in epidermal keratins or by the induction of complexing with metallothionein[7]. All these studies together suggested that cadmium might be a potent carcinogen of the skin.

Our study investigated the relationship between blood cadmium and the risk of NMSC among the general population for the first time. We detected an inverted U-shaped dose-effect curve between the levels of blood cadmium and the odds of NMSC, typically when no covariates were adjusted or with only demographic and socioeconomic factors were adjusted. This inverted U-shaped curve meant that a moderately increased level of blood cadmium was associated with an increased risk of NMSC while a significantly increased level of blood cadmium was inversely associated with the risk of NMSC. This unexpected result was hard to explain, especially when a positive non-linear association between dietary cadmium intake and the risk of melanoma was observed in a population study[8]. However, considering the complex metabolism of cadmium in human organs[18, 23], the concentration of cadmium in the diet does not necessarily equal its concentration in blood and the concentration in the blood does not necessarily mean the concentration in the skin either. Moreover, our result was not unprecedented. The inverse association between blood cadmium level in blood and the risk of breast cancer was also indicated in European Prospective Investigation into Cancer and Nutrition-Italy cohort and in a Meta-analysis[10, 22]. Further studies are needed to explore the underlying mechanisms of the inverse relationship between blood cadmium and the risk of NMSC.

The limitations of this study should be noted. First, the diagnosis of NMSC was based on self-report and it was not confirmed by medical or pathological reports. Thus, there might be some outcome misclassifications. Second, only one routine blood test was conducted to collect information on the level of blood cadmium and serial blood tests might be more reliable. Third, NHANES was a cross-sectional design and the casual inference between cadmium and NMSC should not be made. Nonetheless, this study has some strengths including the reliability of the NHANES data as well as its large sample size.

In conclusion, our study showed that blood cadmium level was significantly associated with the risk of NMSC. Further prospective studies are warranted to determine the causality and underlying mechanisms.

Authors’ Contributions: All authors contributed to the study conception and design. Jigang Chen and Rui Feng collected the data and performed the statistical analysis. Yanni Wang prepared the manuscript. All authors read and approved the final manuscript.

Data availability: The data are from NHANES, which is openly available at https://www.cdc.gov/nchs/nhanes/index.htm

Funding: This work was supported by the Natural Science Foundation of China (No. 82101433).

Compliance With Ethical Standards

Competing interests: The authors declare no competing interests.

Ethics approval:

The study protocols of NHANES were approved by the National Center for Health Statistics Institutional review board (Approval ID: Protocol #98-12; Protocol #2005-06; Protocol #2011-17; Protocol #2018-01).

Consent to Participate: Written informed consent was obtained from each participant.

Consent to Publish: Not applicable. There is no individual level data in our publication.

Aggarwal P, Knabel P, Fleischer Jr AB (2021) United States burden of melanoma and non-melanoma skin cancer from 1990 to 2019 Journal of the American Academy of Dermatology 85: 388-395
Amaral T, Garbe C (2017) Non-melanoma skin cancer: new and future synthetic drug treatments Expert Opinion on Pharmacotherapy 18: 689-699
Centers for Disease Control and Prevention National Health and Nutrition Examination Survey. In: Editor (ed)^(eds) Book National Health and Nutrition Examination Survey, City.
Centers for Disease Control and Prevention National Health and Nutrition Examination Survey, Ethics Review Board (ERB) Approval. In: Editor (ed)^(eds) Book National Health and Nutrition Examination Survey, Ethics Review Board (ERB) Approval, City.
Chen C, Xun P, Nishijo M, He K (2016) Cadmium exposure and risk of lung cancer: a meta-analysis of cohort and case–control studies among general and occupational populations Journal of exposure science & environmental epidemiology 26: 437-444
Divney AA, Murillo R, Rodriguez F, Mirzayi CA, Tsui EK, Echeverria SE (2019) Diabetes prevalence by leisure-, transportation-, and occupation-based physical activity among racially/ethnically diverse US adults Diabetes care 42: 1241-1247
Fasanya-Odewumi C, Latinwo LM, Ikediobi CO, Gilliard L, Sponholtz G, Nwoga J, Stino F, Hamilton N, Erdos G (1998) The genotoxicity and cytotoxicity of dermally-administered cadmium: effects of dermal cadmium administration International journal of molecular medicine 1: 1001-1007
Filippini T, Malagoli C, Wise LA, Malavolti M, Pellacani G, Vinceti M (2019) Dietary cadmium intake and risk of cutaneous melanoma: An Italian population-based case-control study Journal of Trace Elements in Medicine and Biology 56: 100-106
Friberg LT, Elinder G-G, Kjellstrom T, Nordberg GF (2019) Cadmium and health: A toxicological and epidemiological appraisal: Volume 2: Effects and response. CRC press
Gaudet MM, Deubler EL, Kelly RS, Ryan Diver W, Teras LR, Hodge JM, Levine KE, Haines LG, Lundh T, Lenner P (2019) Blood levels of cadmium and lead in relation to breast cancer risk in three prospective cohorts International journal of cancer 144: 1010-1016
Goyer RA, Liu J, Waalkes MP (2004) Cadmium and cancer of prostate and testis Biometals 17: 555-558
Hartwig A (2013) Cadmium and cancer Cadmium: from toxicity to essentiality: 491-507
Hartwig A, Schwerdtle T (2002) Interactions by carcinogenic metal compounds with DNA repair processes: toxicological implications Toxicology letters 127: 47-54
Hu W, Fang L, Ni R, Zhang H, Pan G (2022) Changing trends in the disease burden of non-melanoma skin cancer globally from 1990 to 2019 and its predicted level in 25 years BMC cancer 22: 1-11
Il'yasova D, Schwartz GG (2005) Cadmium and renal cancer Toxicology and applied pharmacology 207: 179-186
International Agency for Research on Cancer (1993) Beryllium, cadmium, mercury, and exposures in the glass Apresentado em: IARC Working Group on the Evaluation of Carcinogenic Risks to Humans: Beryllium, Lyon
Järup L, Åkesson A (2009) Current status of cadmium as an environmental health problem Toxicology and applied pharmacology 238: 201-208
Kjellström T, Nordberg GF (2019) Kinetic model of cadmium metabolismCadmium and health: A toxicological and epidemiological appraisal. CRC Press, pp. 179-198.
Lansdown A, Sampson B (1996) Dermal toxicity and percutaneous absorption of cadmium in rats and mice Laboratory animal science 46: 549-554
National Toxicology Program (2002) 10th Report on Carcinogens. U.S. Department of Health and Human Services, Public Health Service, National Toxicology Program
Organization WH (2017) Radiation: Ultraviolet (UV) radiation and skin cancer. In: Editor (ed)^(eds) Book Radiation: Ultraviolet (UV) radiation and skin cancer, City.
Palli D, Berrino F, Vineis P, Tumino R, Panico S, Masala G, Saieva C, Salvini S, Cerati M, Pala V (2003) A molecular epidemiology project on diet and cancer: the EPIC-Italy Prospective Study. Design and baseline characteristics of participants Tumori Journal 89: 586-593
Rahimzadeh MR, Rahimzadeh MR, Kazemi S, Moghadamnia A-a (2017) Cadmium toxicity and treatment: An update Caspian journal of internal medicine 8: 135
Satarug S, Garrett SH, Sens MA, Sens DA (2010) Cadmium, environmental exposure, and health outcomes Environmental health perspectives 118: 182-190
Stayner L, Smith R, Thun M, Schnorr T, Lemen R (1992) A quantitative assessment of lung cancer risk and occupational cadmium exposure IARC Scientific Publications: 447-455
Vinceti M, Bassissi S, Malagoli C, Pellacani G, Alber D, Bergomi M, Seidenari S (2005) Environmental exposure to trace elements and risk of cutaneous melanoma Journal of Exposure Science & Environmental Epidemiology 15: 458-462
Waalkes MP (2000) Cadmium carcinogenesis in review Journal of inorganic biochemistry 79: 241-244
Waalkes MP (2003) Cadmium carcinogenesis Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 533: 107-120
Weinlich G, Bitterlich W, Mayr V, Fritsch P, Zelger B (2003) Metallothionein‐overexpression as a prognostic factor for progression and survival in melanoma. A prospective study on 520 patients British Journal of Dermatology 149: 535-541
Wester RC, Maibach HI, Sedik L, Melendres J, Dizio S, Wade M (1992) In vitro percutaneous absorption of cadmium from water and soil into human skin Toxicological Sciences 19: 1-5

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Association of blood cadmium level with nonmelanoma skin cancer: a cross-sectional study

Status:

Version 1

Abstract

Purpose

Methods

Results

Conclusions

Figures

Introduction

Methods

Statistical analysis

Results

Discussion

Declarations

References

Additional Declarations

Status:

Version 1