- Email: [email protected]
Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults
Environmental Pollution xxx (2017) 1e8
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults* Junghoon Kim a, Esther Garcia-Esquinas b, Ana Navas-Acien c, d, Yoon-Hyeong Choi a, e, * a
Department of Preventive Medicine, Gachon University College of Medicine, Incheon, Republic of Korea noma de Madrid/IdiPaz, and Ciber of Epidemiology and Public Health (CIBERESP), Departamento de Medicina Preventiva y Salud Pública, Universidad Auto Madrid, Spain c Department of Environmental Health Sciences, Columbia Mailman School of Public Health, New York, NY, USA d Department of Environmental Health Sciences, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA e Gachon Advanced Institute for Health Sciences and Technology, Gachon University, Incheon, Republic of Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 December 2016 Received in revised form 21 March 2017 Accepted 3 September 2017 Available online xxx
Reduced physical performance is an important feature of aging, and walking speed is a valid measure of physical performance and mobility in older adults. Previous epidemiological studies suggest that cadmium exposure, even at low environmental levels, may contribute to vascular, musculoskeletal, and cognitive dysfunction, which may all be associated with reductions in physical performance. To this end, we investigated the associations of blood and urine cadmium concentrations with walking speed in middle-aged and older adults in the U.S. general population. We studied U.S. adults from the National Health and Nutrition Examination Survey 1999 to 2002 who were 50 years of age, who had determinations of cadmium in blood or in urine, and who had measurements of the time taken to walk 20 feet. Walking speed (ft/sec) was computed as walked distance (20 ft) divided by measured time to walk (in seconds). The weighted geometric means of blood and urine cadmium were 0.49 [95% confidence interval (CI): 0.47, 0.52] mg/L and 0.37 (95% CI: 0.34, 0.42) ng/mL, respectively. After adjusting for sociodemographic, anthropometric, health-related behavioral, and clinical risk factors and inflammation markers, the highest (vs. lowest) quintile of blood cadmium was associated with a 0.18 (95% CI: 0.10, 0.25) ft/sec reduction in walking speed (p-Trend <0.001). No association was observed for urine cadmium levels with walking speed. Cadmium concentrations in blood, but not in urine, were associated with slower gait speed. Our findings add to the growing volume of evidence supporting cadmium's toxicity even at low levels of exposure. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Cadmium exposure Environmental exposure Physical function Walking speed Epidemiology
1. Introduction A decline in physical performance is an important characteristic of aging. Walking speed is a reliable measure of physical
Abbreviations: Cr, Creatinine; CVD, Cardiovascular disease; DSST, Digit Symbol Substitution Test; GFR, Glomerular Filtration Rate; LOD, Limit of Detection; MEC, Mobile Examination Center; METs, Metabolic Equivalent; NHANES, National Health and Nutrition Examination Survey. * This paper has been recommended for acceptance by David Carpenter. * Corresponding author. Department of Preventive Medicine, Gachon University College of Medicine, 155 Gaetbeol-ro, Yeonsu-gu, Incheon 21999, Republic of Korea. Tel.: þ82 32 899 6586. E-mail addresses: [email protected] (J. Kim), [email protected] (E. Garcia-Esquinas), [email protected] (A. Navas-Acien), yoonchoi@ gachon.ac.kr (Y.-H. Choi).
performance and mobility in older adults and has been recommended as a practical and informative “vital sign” for assessing functional status and overall health (Middleton et al., 2015). Previous studies have reported that slow walking speed may be associated with increased mortality (Studenski et al., 2011), disability (Kuo et al., 2006), hospital admissions (Penninx et al., 2000), and poor quality of life in older adults (Ekstrom et al., 2011). Identifying and reducing the risk factors that contribute to a decline in walking speed is therefore important for healthy aging, yet few epidemiological studies have been conducted on the association between environmental toxicants (lead and cobalt) and walking speed (Ji et al., 2013; Lang et al., 2009). Cadmium is a ubiquitous environmental toxicant, with exposure primarily through cigarette smoke, dietary sources (mainly shellfish, offal, and vegetables) and ambient air in urban or industrial
http://dx.doi.org/10.1016/j.envpol.2017.09.022 0269-7491/© 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
2
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
€rup et al., 1998; Olmedo et al., 2016). Cadmium is known to areas (Ja be cardiotoxic and carcinogenic after inhalation, and its accumulation in the body results in the development of chronic disease (IARC, 1993; Nordberg et al., 2007). In this sense, there is epidemiologic evidence that environmental exposures to cadmium, even at low levels, contribute to the development of hypertension (Tellez-Plaza et al., 2008), cardiovascular disease (Tellez-Plaza et al., 2013), peripheral arterial disease (Fagerberg et al., 2013), diabetes (Schwartz et al., 2003), kidney disease (Navas-Acien et al., 2009), osteoporosis (Gallagher et al., 2008), and hearing loss (Choi et al., 2012). In addition, a recent study suggests a putative adverse relationship between urine cadmium at high exposure levels and frailty in U.S. older adults (Garcia-Esquinas et al., 2015). Given the ubiquitous exposure of cadmium and its cardiotoxic and musculoskeletal effects, which are themselves associated with motor dysfunction (Dumurgier et al., 2010; Hausdorff et al., 2005), we hypothesized that exposure to cadmium may be a risk factor affecting poor physical performance in older adults. Both blood and urine cadmium are biomarkers of ongoing and long-term exposure, although blood cadmium more readily reflects biologically active cadmium (recent exposure) than does urine cadmium (Nordberg et al., 2007). We therefore examined the association of blood and urine cadmium concentrations with walking speed in a representative U.S. sample of middle-aged and older adults who participated in the National Health and Nutrition Examination Survey (NHANES) 1999 to 2002, while controlling for important potential confounding factors including demographic, anthropometric, behavioral, and clinical factors and inflammation markers.
hand-held stopwatch. Timing begun when the participant's foot first touched the floor beyond the start line; stop time was obtained when the foot first touched the floor beyond the 20-ft finish line (CDC, 2000). This method of measuring walking speed was previously proved to have test-retest reliability (Studenski et al., 2003). We computed walking speed (ft/sec) by dividing the walked distance (20 ft) by measured time to walk (seconds). 2.3. Cadmium concentrations in blood and urine Whole blood and spot urine specimens were processed, were frozen at 20 C, and shipped to the Division of Laboratory Sciences, National Center for Environmental Health, CDC (Atlanta, GA, USA) for analysis (CDC, 2001). Blood cadmium concentrations were measured using a simultaneous multi-element atomic absorption spectrometer (model SIMMA 6000; PerkinElmer, Norwalk, CT, USA) with Zeeman background correction. The limit of detection (LOD) for blood cadmium concentrations was 0.3 mg/L (CDC, 2015); blood cadmium concentrations were below the LOD in 14.07% of the study participants. The interassay coefficients of variation ranged between 4.1% and 9.4%. Urine cadmium concentrations were measured using inductively coupled plasma mass spectrometry (model PerkinElmer/SCIEX 500, PerkinElmer, Norwalk, Connecticut) (CDC, 2003). The LOD for urine cadmium concentrations was 0.06 ng/mL; urine cadmium concentrations were below the LOD in 3.78% of the study participants. The interassay coefficients of variation ranged between 1.2% and 4.7%. For blood and urine cadmium levels below the LOD, NHANES reported a value equal to the LOD/√2.
2. Methods 2.1. Study population NHANES is an ongoing cross-sectional survey of a nationally representative U.S. population conducted by the CDC's National Center for Health Statistics. The survey includes an initial extensive interview at home with a subsequent physical examination and additional interviews at a mobile examination center (MEC) (CDC, 2000). The present analysis used data from NHANES 1999e2002 with the time to walk 20 feet measured in participants 50 years or older. Survey participants who were not able to walk alone without holding onto someone, were excluded from the timed walk component, or who had a history of chest or abdominal surgery within the prior three weeks, myocardial infarction within the prior six weeks, knee surgery, severe back pain, or a history of brain aneurysm or stroke (CDC, 2002); the initial sample eligible for a timed walk examination was 4449 participants. From these, we excluded participants with unavailable measures on walking speed (n ¼ 489), as well as those who used assistive devices during walking speed measures (n ¼ 169), or who had unreliable high walking speed of >6 ft/s (n ¼ 4), yielding 3787 participants. Of those 3787 adults eligible for evaluation of walking time, 3671 and 1157 adults had cadmium measures in blood and urine, respectively. After further excluding participants with missing data for covariates listed in Table 1, a total of 3226 and 1003 adults were eligible for analyses of blood and urine cadmium, respectively. 2.2. Walking speed Time to complete a 20-ft walk test was measured by certified health technicians intensively trained in NHANES examination protocols (CDC, 2002, 2004). Participants were asked to walk 20 feet at their usual pace, and the time to walk was recorded using a
2.4. Other covariates We used a number of variables as confounding factors: demographic information (age, sex, race/ethnicity, and education level), anthropometric measurements (height and weight), healthrelated behaviors (physical activity, alcohol consumption, and cumulative cigarette smoke), clinical factors (hypertension, diabetes, arthritis, and cardiovascular disease (CVD)), and serum biomarkers of inflammation (C-reactive protein and homocysteine). Race/ethnicity was categorized as Non-Hispanic White, NonHispanic Black, Mexican American, and Other. Education level was categorized ashigh school. Cumulative cigarette smoke was categorized as never, <20 and 20 cumulative cigarette pack-years. Alcohol consumption was categorized as never, <1 and 1 day per week. Physical activity was self-reported for moderate-to-vigorous leisure-time activities performed over the previous 30 days; metabolic equivalents per week (METs-hours/week) were computed based on the total volume of physical activity intensity and frequency (Ainsworth et al., 1993). Physical activity was categorized as <7.5, 7.5e21, and >21 METshours/week (Haskell et al., 2007; US Institute of Medicine, 2002). Definition of hypertension was based on a self-reported physician diagnosis, current use of anti-hypertensive medication, or a clinical blood pressure reading 140/90 mmHg. Definition of type 2 diabetes mellitus was based on a self-reported physician diagnosis or current use of anti-hyperglycemia medication (Bainbridge et al., 2008). Arthritis was defined as self-reported physician diagnosis. CVD was defined as a self-reported presence in any one of following diseases: coronary heart disease, congestive heart failure, heart attack, or angina/angina pectoris. Because of their reported association with lower walking speed, we controlled our analyses for plasma concentrations of C-reactive protein and homocysteine (Kuo et al., 2007). Because walking speed may be directly affected by height through a long stride, we controlled for measured height
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
2.5. Statistical analysis
Table 1 Participants characteristics (n ¼ 3226). Characteristics
3
Distribution
Age [years] 63.29 Walking speed [ft/sec] 3.41 Blood cadmium a [mg/L] 0.49 Urine cadmium a,b [ng/mL] 0.37 Height [cm] 167.35 Weight [kg] 79.78 C-reactive protein [mg/dL] 0.48 Homocysteine [mmol/L] 9.63 Sex [n (%)] Men 1621 Women 1605 Race/ethnicity [n (%)] Non-Hispanic White 1918 Non-Hispanic Black 481 Mexican American 614 Other 213 Education [n (%)]High School 1311 Moderate-to-vigorous physical activity [METs-hours/week] <7.5 2087 7.5 to <21 524 21 615 Cumulative cigarette pack-years [n (%)] Never 1576 <20 823 20 827 Alcohol consumption [days/week; n (%)] 0 1373 1 1179 >1 674 Hypertension [n (%)] No 1312 Yes 1914 Diabetes mellitus [n (%)] No 2822 Yes 404 Cardiovascular diseases [n (%)] No 2720 Yes 506 Arthritis [n (%)] No 1898 Yes 1328
±0.22 ±0.02 ±0.02 ±0.08 ±0.16 ±0.39 ±0.02 ±0.08 (46.6) (53.4) (81.8) (7.1) (3.2) (7.9) (24.2) (26.3) (49.5) (60.1) (17.9) (22.0) (47.4) (25.4) (27.2) (38.6) (37.4) (24.0) (45.6) (54.4) (90.5) (9.5) (85.1) (14.9) (59.3) (40.7)
Data in tables are weighted means ± SE for continuous variables and sample size (weighted percentages) for categorical variables. a Geometric means ± SE are presented. b Subsample of participants with urine cadmium information available (n ¼ 1003).
and weight, separately, instead of body mass index (BMI; an obesity index of height divided by weight). Urine creatinine (Cr) concentrations (milligrams per deciliter) were further considered as a confounder when modeling urine cadmium (Barr et al., 2010). In sensitivity analyses, serum cotinine concentrations, cognitive function, and BMI were considered as potential confounders. Serum cotinine concentrations, a biomarker of exposure to active and passive tobacco smoke, was measured by isotope dilution-high performance liquid chromatography/atmospheric pressure chemical ionization tandem mass spectrometry (ID HPLC-APCI MS/MS), and was available in a subsample of participants (n ¼ 3161). Cognitive function was measured in participants 60 years of age using Digit Symbol Substitution Test (DSST) score, and was available in a subsample of participants (n ¼ 2041). BMI was computed as measured weight (kg) divided by measured height squared (m2). Serum cotinine and DSST were normally distributed and therefore controlled as continuous, while BMI was log-transformed to normalize the distribution.
We performed all statistical analyses using SAS survey procedures (ver. 9.4, SAS Institute, Cary, NC) to account for the complex survey design. We used sample weights for 4-year cycles between 1999 and 2002. We used PROC SURVEYREG to estimate the mean change (95% confidence interval (CI)) in walking speed per quintile increase in blood and urine cadmium concentrations. Tests for trend were conducted for ordinal blood and urine cadmium quintiles in regression models using integer values (0e4). We developed sequential models to assess the influence of potential confounding factors: model A was adjusted for age, sex, race/ethnicity, education level, height, and body weight; model B was additionally adjusted for physical activity, smoking status, and alcohol consumption; model C was additionally adjusted for current diagnosis of hypertension, diabetes, arthritis, and CVD; model D was additionally adjusted for C-reactive protein and homocysteine concentrations. When examining associations between walking speed and urine cadmium, we additionally adjusted for urinary Cr concentration as a covariate (Barr et al., 2010). In addition, we examined stratified models by sex, age groups (50e59, 60e69, and 70 years), and cigarette smoke in order to evaluate sex-, age-, smoking-related differences in the associations between cadmium exposure and walking speed. 2.6. Sensitivity analysis First, we conducted sensitivity analyses after further adjusting for serum cotinine and for the DSST score as potential confounders. These sensitivity analyses were performed 1) because cadmium and cotinine are known to share exposure sources, i.e., tobacco €rup et al., 1998), and smoking is related to walking ability smoke (Ja (Fritschi et al., 2013), and 2) because cadmium exposure is known to be associated with cognitive function (Ciesielski et al., 2013), and cognitive function is connected to walking speed (Sheridan et al., 2003). Second, we ran an alternative model after adjusting for BMI instead of adjusting for height and weight. Third, in order to rule out the potential role of variations in cadmium excretion affected by renal function, we performed a sensitivity analysis by running the models after excluding individuals with abnormal glomerular filtration rate (GFR >60 ml/min/1.73 m2), estimated using Chronic Kidney Diseases Epidemiology Collaboration equation (Levey et al., 2009). In our sensitivity analysis for urinary cadmium, we examined alternative modeling by using Cr-corrected cadmium concentrations instead of adjusting for urine creatinine as a covariate. Cr was corrected by dividing cadmium concentration by creatinine concentration (microgram per gram) in each urine sample using a method suggested in order to capture variation in urine concentration (Barr et al., 2005). 3. Results Table 1 shows descriptive characteristics of the entire study population (n ¼ 3226). Overall, mean age (SE) was 63.29 (0.22) years and mean walking speed (SE) was 3.41 (0.02) ft/sec. The geometric means (SE) for blood and urine cadmium concentrations were 0.49 (0.02) mg/L and 0.37 (0.08) ng/mL, respectively. Analyses for urine cadmium concentrations were restricted to participants with urine data (n ¼ 1003). Blood cadmium concentrations were significantly higher in women, older participants, participants with lower height, lower weight, lower education, lower physical activity levels, higher cumulative cigarette smoke, non-diabetes, arthritis, higher C-reactive protein, and higher homocysteine values (Fig. 1 (A)). Urine
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
4
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
cadmium concentrations were significantly higher in participants with higher cumulative cigarette smoke, CVD, higher C-reactive protein, and lower education, and differ across race-ethnicity (Fig. 1 (B)). Blood cadmium concentrations were moderately correlated with urine cadmium concentrations (ng/mL) and Cr-corrected urine cadmium concentrations (mg/g) (Spearman correlation coefficient r ¼ 0.470, p < 0.001 and r ¼ 0.663, p < 0.001, respectively). Table 2 shows the participants’ distribution by quintiles of blood cadmium concentrations. Participants with higher blood cadmium tend to be older and female and to have lower walking speed and less physical activities. For urine cadmium, participants with higher blood cadmium tend to be male and non-Hispanic black and to have less physical activities (Table 3). Table 4 shows associations between blood cadmium concentrations and walking speed. Blood cadmium concentrations were significantly associated with slower walking speed in all sequential models (all p-Trends <0.001). In fully adjusted models (Model D) there was a 0.18 ft/s (95% CI: 0.25, 0.10) decrease in walking speed among participants in the highest quintile (0.8 mg/L) compared to the lowest quintile (<0.4 mg/L) of blood cadmium. Given our model of mean reduction in walking speed, this 0.18 ft/s reduction approximates the estimated reduction in walking speed with a 7-year increase in age, a 12-cm decrease in height [0.19 ft/s], a 25-kg increase in weight [0.18 ft/s], and a 13 mmol/L increase in homocysteine [0.13 ft/s]. This same 0.18 ft/s reduction in walking speed is greater than the reduction found in sex difference [male vs female; 0.08 ft/s], physical activity [low (<7.5 METs-hr/wk) vs. moderate activities (7.5e21 MET hr/wk); 0.15 ft/s], current diabetes [0.14 ft/s], and arthritis [0.11 ft/s] (see Table 4 (Model D) and Table S1). Table 5 shows no significant associations between urine cadmium concentrations and walking speed. Similarly, when running alternative models using Cr-corrected urine cadmium concentrations (microgram per gram), results were not significant (Table S2). In stratified analyses, neither effect modification by sex, age, nor smoking was observed for the associations between blood and urine cadmium concentrations with walking speed (Table S3 for blood and data not shown for urine). We conducted sensitivity analyses after additional adjustment for serum cotinine and cognitive function. We found that associations of blood cadmium with walking speed remained and were even stronger after adjusting for cotinine (Table S4), and these associations remained unchanged after adjusting for cognitive function (Table S5). Also, results were similar (data not shown) when we ran an alternative model after adjusting for BMI instead of adjusting for height and weight. For urine cadmium, results were not changed after adjusting for cotinine, cognitive function, or BMI (data not shown). 4. Discussion In a representative sample of U.S. middle-aged and older adults who participated in NHANES 1999e2002, cadmium concentrations in blood, but not in urine, were associated with declines in walking speed. To our knowledge, this is the first investigation to examine the association between blood cadmium and walking speed. Two previous epidemiological studies (Garcia-Esquinas et al., 2015; Lang et al., 2009) have studied the association between urine cadmium and walking speed, and both have shown similar results as those reported in this article. In the first study, based on NHANES 1999e2004, no association between urine cadmium (as logged concentration) and self-reported problem of walking speed was observed, after controlling for urinary creatinine as a confounder (Lang et al., 2009). In the second, based on NHANES III, no
association was observed between urine cadmium concentrations (as tertile) and slowness (through the eight-foot walking speed test) with adjusted for urinary creatinine (Garcia-Esquinas et al., 2015). Different associations for blood and urine cadmium have been previously reported for other health outcomes (e.g., hypertension) (Staessen et al., 2000; Tellez-Plaza et al., 2008), and may be explained due to differences in metabolism. Although overlap is substantial, blood cadmium is a better indicator of biologically active cadmium from recent exposure while urine cadmium is a better biomarker for long-term exposure (i.e., several decades) (Nordberg et al., 2007). Based on previous studies, we suggest that our results reflect short-term effects of this metal on walking speed. Additionally, null findings for urine cadmium could be influenced by age-related variations in renal function. In this sense, recent studies have suggested that cadmium may accumulate in the kidney until age 60, when variations in renal function would result in changes in urine cadmium excretion (Chaumont et al., 2013). An association between blood cadmium and walking speed could be explained through several mechanisms. First, cadmium could affect walking speed through its effects in the motor nervous system. In this sense, there is evidence from case-control studies suggesting that cadmium may induce motor neuron disease (Pamphlett et al., 2001) and Parkinson's-like motor symptoms (Okuda et al., 1997). Also, in vivo animal studies show changes in the structure of the cerebellar vermis, which has an important role of body posture and motor control, after exposure to cadmium in albino rats (Wahdan et al., 2014). Second, cadmium could affect walking speed through its effects on cognitive function (Ciesielski et al., 2013), which is linked to motor performance (Hausdorff et al., 2005). Additional adjustment for the DSST score did not change our results (Table S5), however, suggesting that changes in cognitive function do not mediate the observed associations. Recently, Mielke et al. have reported an interesting finding from a longitudinal study: increased walking speed at baseline was associated with less cognitive decline, while cognitive function at baseline did not predict changes in walking speed (Mielke et al., 2013). Based on those findings and our results, we suggest that a decline in cognitive function attributable to cadmium exposure does not causally precede a slower walking speed. Third, cadmium is associated with a higher risk of musculoskeletal disorders including bone fragility and fractures (Staessen et al., 1999) or osteoporosis (Gallagher et al., 2008; Jin et al., 2004), all of which are known to be associated with poor motor performance (Gregg et al., 2000; Kwon et al., 2007). A previous experimental study using a rat model of human environmental exposure to cadmium showed that low lifetime exposure to cadmium is associated with demineralization of bone, and may increase the risk of osteoporosis and fractures (Brzoska and Moniuszko-Jakoniuk, 2004). Moreover, a review article of epidemiological studies provides an evidence of exposure to cadmium significantly affecting to the reduced bone mineral density and increased risk for fracture even at lower exposure level (Bhattacharyya, 2009). Then, the bone disorders caused by cadmium exposure may contribute to a decline in walking speed. Fourth, cadmium is a known risk factor for cardiovascular disease (Tellez-Plaza et al., 2013) and hypertension (Tellez-Plaza et al., 2008), which in turn can influence walking speed (Bouillon et al., 2013; Dumurgier et al., 2010). In the current study, however, adjustment for these conditions did not alter the observed associations (see Table 4, Model C). Fifth, oxidative stress and inflammation could also explain the associations between cadmium and walking speed, either as one pathway of overall frailty or in a step throughout various
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
5
Fig. 1. Geometric means (95% CIs) of cadmium concentrations in blood and urine by participant's characteristics. a Cut-off points based on median from total population (n ¼ 3226); median from subjects (n ¼ 1003) assigned to urinary cadmium analysis is height 166.5 cm, weight 77.7 kg, Creactive protein 0.27 mg/dL, and homocysteine 8.77 mmol/L. * Statistical significance (p < 0.05). We used survey t-test for binominal groups and the Wald F-test for categorical groups.
mechanisms of central nervous system neurotoxicity, vascular diseases, and other chronic diseases described above. In this regard,
there is evidence that chronic exposures to cadmium are associated with higher levels of oxidative stress and inflammation (Colacino
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
6
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
Table 2 Characteristics of study population by blood cadmium concentrations (n ¼ 3226). Characteristics
Cadmium Quintiles Q 1 (<0.4 mg/L) (n ¼ 759)
P-Trend Q 2 (0.4e0.5 mg/L) (n ¼ 683)
63.42 Age [years] 61.25 ±0.35a Walking speed [ft/sec] 3.55 ±0.03 3.47 Sex [n (%)] Men 446 (57.5)b 325 Women 313 (42.5) 358 Race/ethnicity [n (%)] Non-Hispanic White 456 (82.2) 410 Non-Hispanic Black 144 (8.8) 86 Mexican American 117 (2.9) 140 Other 42 (6.0) 47 Moderate-to-vigorous physical activity [METs-hours/week] <7.5 452 (53.0) 408 7.5 to <21 139 (21.0) 121 21 168 (26.0) 154 a b
Q 3 (0.5e0.6 mg/L) (n ¼ 572)
Q 4 (0.6e0.8 mg/L) (n ¼ 560)
Q 5 (0.80e7.4 mg/ L) (n ¼ 652)
±0.49 ±0.04
64.65 3.41
±0.40 ±0.04
65.78 3.32
±0.62 ±0.05
62.62 3.25
±0.47 ±0.04
0.002 <0.001
(43.8) (56.2)
264 308
(41.7) (58.3)
254 306
(40.7) (59.3)
332 320
(44.8) (55.2)
<0.001
(82.3) (5.9) (3.5) (8.4)
338 77 117 40
(82.5) (6.0) (3.3) (8.2)
330 70 130 30
(82.9) (6.3) (3.9) (7.0)
384 104 110 54
(79.5) (7.6) (2.8) (10.1)
0.137
(54.0) (20.3) (25.7)
362 87 123
(59.3) (17.3) (23.3)
375 90 95
(62.8) (18.0) (19.2)
490 87 75
(73.6) (11.9) (14.5)
<0.001
Weighted mean ± SE from surveymean (all such values). Weighted percentage from surveyfrequency (all such values).
Table 3 Characteristics of study population by urine cadmium concentrations (n ¼ 1003). Characteristics
Cadmium Quintiles Q1 (0.18 ng/mL) (n ¼ 200)
P-Trend Q 2 (0.18e0.32 ng/ mL) (n ¼ 201)
Age [years] 65.12 ±10.12a 67.21 Walking speed [ft/sec] 3.37 ±0.79 3.26 Sex [n (%)] Men 88 (41.3)b 102 Women 112 (58.7) 99 Race/ethnicity [n (%)] Non-Hispanic White 129 (87.0) 134 Non-Hispanic Black 18 (4.4) 20 Mexican American 47 (4.7) 37 Other 6 (4.0) 10 Moderate-to-vigorous physical activity [METs-hours/week] <7.5 113 (51.1) 108 7.5 to <21 40 (23.6) 37 21 47 (25.2) 56 a b
Q 3 (0.32e0.51 ng/ mL) (n ¼ 201)
Q 4 (0.52e0.85 ng/ mL) (n ¼ 201)
Q 5 (0.86 e36.78 ng/mL) (n ¼ 200)
±9.91 ±0.81
66.32 3.20
±10.41 ±0.81
66.04 3.17
±9.95 ±0.73
64.67 3.23
±9.65 ±0.74
0.865 0.061
(46.2) (53.8)
109 92
(45.9) (54.1)
105 96
(44.6) (55.4)
118 82
(56.1) (43.9)
0.030
(87.5) (4.1) (3.0) (5.4)
123 24 34 20
(80.5) (6.6) (2.3) (10.7)
106 34 44 17
(78.2) (7.6) (3.6) (10.6)
102 49 34 15
(75.8) (11.1) (2.9) (10.2)
0.009
(49.0) (20.6) (30.5)
118 36 47
(56.2) (16.8) (27.0)
140 30 31
(68.7) (16.2) (15.1)
141 29 30
(64.8) (15.2) (19.9)
0.013
Weighted mean ± SE from surveymean (all such values). Weighted percentages from surveyfrequency (all such values).
Table 4 Change (95% CIs) in walking speed (ft/sec) by blood cadmium concentrations (n ¼ 3226). Cadmium Quintile (mg/L)
No.
Model A
Q 1 (<0.4) Q 2 (0.4e0.5) Q 3 (0.5e0.6) Q 4 (0.6e0.8) Q 5 (0.80e7.4) P-Trend
759 683 572 560 652
0 0.01 0.01 0.08 0.24 <0.001
Model Model Model Model
Model B (Reference) (0.08, 0.09) (0.10, 0.08) (0.17, 0.02) (0.31, 0.17)
0 0.01 0.00 0.05 0.20 <0.001
Model C (Reference) (0.07, 0.08) (0.08, 0.09) (0.15, 0.05) (0.27, 0.12)
0 0.00 0.00 0.06 0.20 <0.001
Model D (Reference) (0.07, 0.08) (0.08, 0.08) (0.16, 0.04) (0.28, 0.13)
0 0.00 0.00 0.05 0.18 <0.001
(Reference) (0.08, 0.08) (0.08, 0.09) (0.15, 0.04) (0.25, 0.10)
A: adjusted for age, sex, race/ethnicity, education, height, and weight. B: adjusted for Model A covariates plus physical activity, pack-years of cigarette smoke, and alcohol consumption. C: adjusted for Model B covariates plus hypertension, diabetes, arthritis, and cardiovascular diseases. D: adjusted for Model C covariates plus C-reactive protein and homocysteine.
et al., 2014; Lin et al., 2009) that play an important role in physical function decline (McClure et al., 2014). In our study, when adjusting for inflammation marker of C-reactive protein and homocysteine, the associations between blood cadmium and walking speed were similar but slightly reduced (Table 4, Model D). Finally, we could not rule out the potential role of cadmium accumulation and excretion related to renal function. One U.S. cross-sectional study reported that estimated GFR levels are related
to lower blood cadmium and higher urine cadmium, and suggested measured cadmium levels as a possible consequence of kidney function rather than an effect of cadmium on kidney function (Buser et al., 2016). One study of older patients observed that reduced renal function, as estimated GFR, is associated with reduced walking speed (Lattanzio et al., 2012). When we examined models after excluding individuals with reduced GFR, however, our associations between blood or urine cadmium and walking speed
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
7
Table 5 Changes (95% CIs) in walking speed (ft/sec) by urinary cadmium concentrations (n ¼ 1003). Cadmium Quintile (ng/mL)
No.
Model A
Q 1 (0.18) Q 2 (0.18e0.32) Q 3 (0.32e0.51) Q 4 (0.52e0.85) Q 5 (0.86e36.78) P-Trend
200 201 201 201 200
0 0.06 0.10 0.13 0.08 0.236
Model Model Model Model
Model B (Reference) (0.23, 0.11) (0.25, 0.06) (0.27, 0.01) (0.27, 0.12)
0 0.07 0.08 0.07 0.01 0.798
Model C (Reference) (0.23, 0.09) (0.22, 0.05) (0.21, 0.07) (0.22, 0.19)
0 0.08 0.10 0.09 0.03 0.658
Model D (Reference) (0.24, 0.09) (0.24, 0.04) (0.23, 0.05) (0.24, 0.18)
0 0.09 0.11 0.11 0.05 0.545
(Reference) (0.25, 0.07) (0.26, 0.03) (0.25, 0.04) (0.27, 0.17)
A: adjusted for age, sex, race/ethnicity, education, height, and weight. B: adjusted for Model A covariates plus physical activity, pack-years of cigarette smoke, and alcohol consumption. C: adjusted for Model B covariates plus hypertension, diabetes, arthritis, and cardiovascular diseases. D: adjusted for Model C covariates plus C-reactive protein and homocysteine. All models were adjusted for urinary creatinine concentrations.
were not changed (Tables S6 and S7). The general population is primarily exposed to cadmium through cigarette smoke, dietary intake (shellfish, grain, and vegetables), and ambient air pollution in urban or industrial regions, and cadmium accumulation in the human body (biological half-life: 10e30 years) could contribute to the development of many dis€rup et al., 1998; Nordberg et al., 2007). There is growing eases (Ja evidence that cadmium exposures even at levels currently observed in the general population have adverse effects on other health outcomes including neurocognitive dysfunction (Ciesielski et al., 2013), cardiovascular disease (Tellez-Plaza et al., 2013), diabetes (Schwartz et al., 2003), hypertension (Tellez-Plaza et al., 2008), kidney disease (Navas-Acien et al., 2009), peripheral arterial disease (Fagerberg et al., 2013), osteoporosis (Gallagher et al., 2008) and hearing loss (Choi et al., 2012). Our findings in walking speed support previous observations that indicate adverse health effects of current environmental levels of cadmium exposures. An ongoing trial of repeated chelation with edetate disodium compared to placebo could provide the opportunity to assess the health impact of removing cadmium from the body (Lamas et al., 2016). The trial is measuring the effect of edetate disodium in changing urine and blood cadmium levels over time and the potential benefits for cardiovascular disease. Additional outcomes, relevant for cadmium-related health effects, such as walking speed could be added to this trial. Strengths of the present study include the use of a representative sample of the middle and older general adults, the availability of objectively-measured walking speed, as well as various potential confounders evaluated. Also, our study had limitations. Because NHANES is cross-sectional, we could not infer temporality. In addition, estimating cadmium exposures using single blood and urine measures may raise an issue of within-individual variability. Uncertainties in exposure assessment, however, are likely to be non-differential bias and may lead to an underestimate of risk. 5. Conclusion Cadmium concentrations in blood, but not in urine, are associated with reduced walking speed, which may reflect short-term effects of recent exposures as well as reflect biologically active cadmium. Our findings add to the growing volume of evidence supporting cadmium's toxicity even at low levels of exposure and support the need to further reduce current exposure levels in order to effectively prevent motor function decline and promote healthy aging in the general population. Further research using longitudinal data are needed to confirm these results. Acknowledgment This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded
by the Korea Ministry of Education (grant number 2013R1A6A3A04059556) and the Korea Ministry of Science, ICT and Future Planning (grant number 2015R1C1A2A01054768). Role of the funding source The funders had no role in this study design, data collection and analysis and prepared all results. Conflict of interests None. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2017.09.022. References Ainsworth, B.E., et al., 1993. Compendium of physical activities: classification of energy costs of human physical activities. Med. Sci. Sports Exerc 25, 71e80. Bainbridge, K.E., et al., 2008. Diabetes and hearing impairment in the United States: audiometric evidence from the national health and nutrition examination survey, 1999 to 2004. Ann. Intern Med. 149, 1e10. Barr, D.B., et al., 2010. Urinary concentrations of metabolites of pyrethroid insecticides in the general U.S. population: national Health and Nutrition Examination Survey 1999-2002. Environ. Health Perspect. 118, 742e748. Barr, D.B., et al., 2005. Urinary creatinine concentrations in the U.S. population: implications for urinary biologic monitoring measurements. Environ. Health Perspect. 113, 192e200. Bhattacharyya, M.H., 2009. Cadmium osteotoxicity in experimental animals: mechanisms and relationship to human exposures. Toxicol. Appl. Pharmacol. 238, 258e265. Bouillon, K., et al., 2013. Cardiovascular disease risk scores in identifying future frailty: the Whitehall II prospective cohort study. Heart 99, 737e742. Brzoska, M.M., Moniuszko-Jakoniuk, J., 2004. Low-level exposure to cadmium during the lifetime increases the risk of osteoporosis and fractures of the lumbar spine in the elderly: studies on a rat model of human environmental exposure. Toxicol. Sci. 82, 468e477. Buser, M.C., et al., 2016. Urinary and blood cadmium and lead and kidney function: NHANES 2007-2012. Int. J. Hyg. Environ. Health 219, 261e267. CDC, 2000. National Health and Nutrition Examination Survey Muscle Strength Procedures. Manual Available: http://www.cdc.gov/nchs/data/nhanes/ms.pdf (accessed Dec 7, 2015). CDC, 2001. National Health and Nutrition Examination Survey 1999-2000, Laboratory Procedure Manual. Available: http://www.cdc.gov/nchs/data/nhanes/ nhanes_99_00/lab06_met_lead_and_cadmium.pdf (accessed Dec 7, 2015). CDC, 2002. National Health and Nutrition Examination Survey 1999-2000. Data Documentation, Codebook, and Frequencies: Muscle Strength. Available: http:// wwwn.cdc.gov/Nchs/Nhanes/1999-2000/MSX.htm (accessed Dec 7, 2015). CDC, 2003. Laboratory Procedure Manual: Multiple Toxic Elements, NHANES 19992000. Available: http://www.cdc.gov/nchs/data/nhanes/nhanes_99_00/lab06_ met_hm.pdf (accessed Dec 7, 2015). CDC, 2004. National Health and Nutrition Examination Survey 2001-2002. Data Documentation, Codebook, and Frequencies: Muscle Strength. Available: http:// wwwn.cdc.gov/Nchs/Nhanes/2001-2002/MSX_B.htm (accessed Dec 7, 2015). CDC, 2015. Fourth National Report on Human Exposure to Environmental Chemicals. Available: https://www.cdc.gov/biomonitoring/pdf/fourthreport_ updatedtables_feb2015.pdf (accessed Dec 14, 2016).
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
8
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
Chaumont, A., et al., 2013. Associations of urinary cadmium with age and urinary proteins: further evidence of physiological variations unrelated to metal accumulation and toxicity. Environ. Health Perspect. 121, 1047e1053. Choi, Y.-H., et al., 2012. Environmental cadmium and lead exposures and hearing loss in u.s. Adults: the national health and nutrition examination survey, 1999 to 2004. Environ. Health Perspect. 120, 1544e1550. Ciesielski, T., et al., 2013. Associations between cadmium exposure and neurocognitive test scores in a cross-sectional study of US adults. Environ. Health 12, 13. Colacino, J.A., et al., 2014. Dietary antioxidant and anti-inflammatory intake modifies the effect of cadmium exposure on markers of systemic inflammation and oxidative stress. Environ. Res. 131, 6e12. Dumurgier, J., et al., 2010. Hypertension and lower walking speed in the elderly: the Three-City study. J. Hypertens. 28, 1506e1514. Ekstrom, H., et al., 2011. Effects of walking speed and results of timed get-up-and-go tests on quality of life and social participation in elderly individuals with a history of osteoporosis-related fractures. J. Aging Health 23, 1379e1399. Fagerberg, B., et al., 2013. Cadmium exposure, intercellular adhesion molecule-1 and peripheral artery disease: a cohort and an experimental study. BMJ Open 3. Fritschi, C., et al., 2013. The effects of smoking status on walking ability and healthrelated quality of life in patients with peripheral arterial disease. J. Cardiovasc Nurs. 28, 380e386. Gallagher, C.M., et al., 2008. Urinary cadmium and osteoporosis in U.S. Women >or¼ 50 years of age: NHANES 1988-1994 and 1999-2004. Environ. Health Perspect. 116, 1338e1343. Garcia-Esquinas, E., et al., 2015. Association of lead and cadmium exposure with frailty in US older adults. Environ. Res. 137, 424e431. Gregg, E.W., et al., 2000. Diabetes and physical disability among older U.S. adults. Diabetes Care 23, 1272e1277. Haskell, W.L., et al., 2007. Physical activity and public health: updated recommendation for adults from the american college of sports medicine and the american heart association. Circulation 116, 1081e1093. Hausdorff, J.M., et al., 2005. Walking is more like catching than tapping: gait in the elderly as a complex cognitive task. Exp. Brain Res. 164, 541e548. IARC, 1993. Cadmium and cadmium compounds, IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Human, 1993/01/01 ed, pp. 119e237. €rup, L., et al., 1998. Health effects of cadmium exposure e a review of the literJa ature and a risk estimate. Scand. J. Work Environ. Health 24, 1e51. Ji, J.S., et al., 2013. Association between blood lead and walking speed in the national health and nutrition examination survey (NHANES 1999-2002). Environ. Health Perspect. 121, 711e716. Jin, T., et al., 2004. Osteoporosis and renal dysfunction in a general population exposed to cadmium in China. Environ. Res. 96, 353e359. Kuo, H.K., et al., 2006. Exploring how peak leg power and usual gait speed are linked to late-life disability: data from the National Health and Nutrition Examination Survey (NHANES), 1999-2002. Am. J. Phys. Med. Rehabil. 85, 650e658. Kuo, H.K., et al., 2007. Relationship of homocysteine levels to quadriceps strength, gait speed, and late-life disability in older adults. J. Gerontol. A Biol. Sci. Med. Sci. 62, 434e439. Kwon, J., et al., 2007. Association between change in bone mineral density and decline in usual walking speed in elderly community-dwelling Japanese women during 2 years of follow-up. J. Am. Geriatr. Soc. 55, 240e244. Lamas, G.A., et al., 2016. Heavy metals, cardiovascular disease, and the unexpected benefits of chelation therapy. J. Am. Coll. Cardiol. 67, 2411e2418. Lang, I.A., et al., 2009. Age-related impairments of mobility associated with cobalt and other heavy metals: data from NHANES 1999-2004. J. Toxicol. Environ.
Health A 72, 402e409. Lattanzio, F., et al., 2012. Relationship between renal function and physical performance in elderly hospitalized patients. Rejuvenation Res. 15, 545e552. Levey, A.S., et al., 2009. A new equation to estimate glomerular filtration rate. Ann. Intern Med. 150, 604e612. Lin, Y.S., et al., 2009. Cadmium exposure is associated with elevated blood Creactive protein and fibrinogen in the U. S. population: the third national health and nutrition examination survey (NHANES III, 1988-1994). Ann. Epidemiol. 19, 592e596. McClure, C.K., et al., 2014. Prospective associations between inflammatory and hemostatic markers and physical functioning limitations in mid-life women: longitudinal results of the Study of Women's Health across the Nation (SWAN). Exp. Gerontol. 49, 19e25. Middleton, A., et al., 2015. Walking speed: the functional vital sign. J. Aging Phys. Act. 23, 314e322. Mielke, M.M., et al., 2013. Assessing the temporal relationship between cognition and gait: slow gait predicts cognitive decline in the Mayo Clinic study of aging. J. Gerontol. A Biol. Sci. Med. Sci. 68, 929e937. Navas-Acien, A., et al., 2009. Blood cadmium and lead and chronic kidney disease in US adults: a joint analysis. Am. J. Epidemiol. 170, 1156e1164. Nordberg, G., et al., 2007. Cadmium. In: Nordberg, G., Fowler, B., Nordberg, M. (Eds.), Handbook on the Toxicology of Metals. Elsevier, Amsterdam, the Netherlands, pp. 446e486. Okuda, B., et al., 1997. Parkinsonism after acute cadmium poisoning. Clin. Neurol. Neurosurg. 99, 263e265. Olmedo, P., et al., 2016. Dietary determinants of cadmium exposure in the strong heart family study. Food Chem. Toxicol. 100, 239e246. Pamphlett, R., et al., 2001. Blood levels of toxic and essential metals in motor neuron disease. Neurotoxicology 22, 401e410. Penninx, B.W., et al., 2000. Lower extremity performance in nondisabled older persons as a predictor of subsequent hospitalization. J. Gerontol. A Biol. Sci. Med. Sci. 55, M691eM697. Schwartz, G.G., et al., 2003. Urinary cadmium, impaired fasting glucose, and diabetes in the NHANES III. Diabetes Care 26, 468e470. Sheridan, P.L., et al., 2003. Influence of executive function on locomotor function: divided attention increases gait variability in Alzheimer's disease. J. Am. Geriatr. Soc. 51, 1633e1637. Staessen, J.A., et al., 2000. Exposure to cadmium and conventional and ambulatory blood pressures in a prospective population study. Public health and environmental exposure to cadmium study group. Am. J. Hypertens. 13, 146e156. Staessen, J.A., et al., 1999. Environmental exposure to cadmium, forearm bone density, and risk of fractures: prospective population study. Public health and environmental exposure to cadmium (PheeCad) study group. Lancet 353, 1140e1144. Studenski, S., et al., 2011. Gait speed and survival in older adults. JAMA 305, 50e58. Studenski, S., et al., 2003. Physical performance measures in the clinical setting. J. Am. Geriatr. Soc. 51, 314e322. Tellez-Plaza, M., et al., 2013. Cadmium exposure and clinical cardiovascular disease: a systematic review. Curr. Atheroscler. Rep. 15, 356. Tellez-Plaza, M., et al., 2008. Cadmium exposure and hypertension in the 1999e2004 national health and nutrition examination survey (NHANES). Environ. Health Perspect. 116, 51e56. US Institute of Medicine, 2002. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients). National Academies Press, Washington, DC. Wahdan, M.H., et al., 2014. Effect of cadmium exposure on the structure of the cerebellar vermis of growing male albino rat. Int. Res. J. Appl. Basic Sci. 8, 163e176.
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
Contents lists available at ScienceDirect
Environmental Pollution journal homepage: www.elsevier.com/locate/envpol
Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults* Junghoon Kim a, Esther Garcia-Esquinas b, Ana Navas-Acien c, d, Yoon-Hyeong Choi a, e, * a
Department of Preventive Medicine, Gachon University College of Medicine, Incheon, Republic of Korea noma de Madrid/IdiPaz, and Ciber of Epidemiology and Public Health (CIBERESP), Departamento de Medicina Preventiva y Salud Pública, Universidad Auto Madrid, Spain c Department of Environmental Health Sciences, Columbia Mailman School of Public Health, New York, NY, USA d Department of Environmental Health Sciences, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA e Gachon Advanced Institute for Health Sciences and Technology, Gachon University, Incheon, Republic of Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 December 2016 Received in revised form 21 March 2017 Accepted 3 September 2017 Available online xxx
Reduced physical performance is an important feature of aging, and walking speed is a valid measure of physical performance and mobility in older adults. Previous epidemiological studies suggest that cadmium exposure, even at low environmental levels, may contribute to vascular, musculoskeletal, and cognitive dysfunction, which may all be associated with reductions in physical performance. To this end, we investigated the associations of blood and urine cadmium concentrations with walking speed in middle-aged and older adults in the U.S. general population. We studied U.S. adults from the National Health and Nutrition Examination Survey 1999 to 2002 who were 50 years of age, who had determinations of cadmium in blood or in urine, and who had measurements of the time taken to walk 20 feet. Walking speed (ft/sec) was computed as walked distance (20 ft) divided by measured time to walk (in seconds). The weighted geometric means of blood and urine cadmium were 0.49 [95% confidence interval (CI): 0.47, 0.52] mg/L and 0.37 (95% CI: 0.34, 0.42) ng/mL, respectively. After adjusting for sociodemographic, anthropometric, health-related behavioral, and clinical risk factors and inflammation markers, the highest (vs. lowest) quintile of blood cadmium was associated with a 0.18 (95% CI: 0.10, 0.25) ft/sec reduction in walking speed (p-Trend <0.001). No association was observed for urine cadmium levels with walking speed. Cadmium concentrations in blood, but not in urine, were associated with slower gait speed. Our findings add to the growing volume of evidence supporting cadmium's toxicity even at low levels of exposure. © 2017 Elsevier Ltd. All rights reserved.
Keywords: Cadmium exposure Environmental exposure Physical function Walking speed Epidemiology
1. Introduction A decline in physical performance is an important characteristic of aging. Walking speed is a reliable measure of physical
Abbreviations: Cr, Creatinine; CVD, Cardiovascular disease; DSST, Digit Symbol Substitution Test; GFR, Glomerular Filtration Rate; LOD, Limit of Detection; MEC, Mobile Examination Center; METs, Metabolic Equivalent; NHANES, National Health and Nutrition Examination Survey. * This paper has been recommended for acceptance by David Carpenter. * Corresponding author. Department of Preventive Medicine, Gachon University College of Medicine, 155 Gaetbeol-ro, Yeonsu-gu, Incheon 21999, Republic of Korea. Tel.: þ82 32 899 6586. E-mail addresses: [email protected] (J. Kim), [email protected] (E. Garcia-Esquinas), [email protected] (A. Navas-Acien), yoonchoi@ gachon.ac.kr (Y.-H. Choi).
performance and mobility in older adults and has been recommended as a practical and informative “vital sign” for assessing functional status and overall health (Middleton et al., 2015). Previous studies have reported that slow walking speed may be associated with increased mortality (Studenski et al., 2011), disability (Kuo et al., 2006), hospital admissions (Penninx et al., 2000), and poor quality of life in older adults (Ekstrom et al., 2011). Identifying and reducing the risk factors that contribute to a decline in walking speed is therefore important for healthy aging, yet few epidemiological studies have been conducted on the association between environmental toxicants (lead and cobalt) and walking speed (Ji et al., 2013; Lang et al., 2009). Cadmium is a ubiquitous environmental toxicant, with exposure primarily through cigarette smoke, dietary sources (mainly shellfish, offal, and vegetables) and ambient air in urban or industrial
http://dx.doi.org/10.1016/j.envpol.2017.09.022 0269-7491/© 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
2
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
€rup et al., 1998; Olmedo et al., 2016). Cadmium is known to areas (Ja be cardiotoxic and carcinogenic after inhalation, and its accumulation in the body results in the development of chronic disease (IARC, 1993; Nordberg et al., 2007). In this sense, there is epidemiologic evidence that environmental exposures to cadmium, even at low levels, contribute to the development of hypertension (Tellez-Plaza et al., 2008), cardiovascular disease (Tellez-Plaza et al., 2013), peripheral arterial disease (Fagerberg et al., 2013), diabetes (Schwartz et al., 2003), kidney disease (Navas-Acien et al., 2009), osteoporosis (Gallagher et al., 2008), and hearing loss (Choi et al., 2012). In addition, a recent study suggests a putative adverse relationship between urine cadmium at high exposure levels and frailty in U.S. older adults (Garcia-Esquinas et al., 2015). Given the ubiquitous exposure of cadmium and its cardiotoxic and musculoskeletal effects, which are themselves associated with motor dysfunction (Dumurgier et al., 2010; Hausdorff et al., 2005), we hypothesized that exposure to cadmium may be a risk factor affecting poor physical performance in older adults. Both blood and urine cadmium are biomarkers of ongoing and long-term exposure, although blood cadmium more readily reflects biologically active cadmium (recent exposure) than does urine cadmium (Nordberg et al., 2007). We therefore examined the association of blood and urine cadmium concentrations with walking speed in a representative U.S. sample of middle-aged and older adults who participated in the National Health and Nutrition Examination Survey (NHANES) 1999 to 2002, while controlling for important potential confounding factors including demographic, anthropometric, behavioral, and clinical factors and inflammation markers.
hand-held stopwatch. Timing begun when the participant's foot first touched the floor beyond the start line; stop time was obtained when the foot first touched the floor beyond the 20-ft finish line (CDC, 2000). This method of measuring walking speed was previously proved to have test-retest reliability (Studenski et al., 2003). We computed walking speed (ft/sec) by dividing the walked distance (20 ft) by measured time to walk (seconds). 2.3. Cadmium concentrations in blood and urine Whole blood and spot urine specimens were processed, were frozen at 20 C, and shipped to the Division of Laboratory Sciences, National Center for Environmental Health, CDC (Atlanta, GA, USA) for analysis (CDC, 2001). Blood cadmium concentrations were measured using a simultaneous multi-element atomic absorption spectrometer (model SIMMA 6000; PerkinElmer, Norwalk, CT, USA) with Zeeman background correction. The limit of detection (LOD) for blood cadmium concentrations was 0.3 mg/L (CDC, 2015); blood cadmium concentrations were below the LOD in 14.07% of the study participants. The interassay coefficients of variation ranged between 4.1% and 9.4%. Urine cadmium concentrations were measured using inductively coupled plasma mass spectrometry (model PerkinElmer/SCIEX 500, PerkinElmer, Norwalk, Connecticut) (CDC, 2003). The LOD for urine cadmium concentrations was 0.06 ng/mL; urine cadmium concentrations were below the LOD in 3.78% of the study participants. The interassay coefficients of variation ranged between 1.2% and 4.7%. For blood and urine cadmium levels below the LOD, NHANES reported a value equal to the LOD/√2.
2. Methods 2.1. Study population NHANES is an ongoing cross-sectional survey of a nationally representative U.S. population conducted by the CDC's National Center for Health Statistics. The survey includes an initial extensive interview at home with a subsequent physical examination and additional interviews at a mobile examination center (MEC) (CDC, 2000). The present analysis used data from NHANES 1999e2002 with the time to walk 20 feet measured in participants 50 years or older. Survey participants who were not able to walk alone without holding onto someone, were excluded from the timed walk component, or who had a history of chest or abdominal surgery within the prior three weeks, myocardial infarction within the prior six weeks, knee surgery, severe back pain, or a history of brain aneurysm or stroke (CDC, 2002); the initial sample eligible for a timed walk examination was 4449 participants. From these, we excluded participants with unavailable measures on walking speed (n ¼ 489), as well as those who used assistive devices during walking speed measures (n ¼ 169), or who had unreliable high walking speed of >6 ft/s (n ¼ 4), yielding 3787 participants. Of those 3787 adults eligible for evaluation of walking time, 3671 and 1157 adults had cadmium measures in blood and urine, respectively. After further excluding participants with missing data for covariates listed in Table 1, a total of 3226 and 1003 adults were eligible for analyses of blood and urine cadmium, respectively. 2.2. Walking speed Time to complete a 20-ft walk test was measured by certified health technicians intensively trained in NHANES examination protocols (CDC, 2002, 2004). Participants were asked to walk 20 feet at their usual pace, and the time to walk was recorded using a
2.4. Other covariates We used a number of variables as confounding factors: demographic information (age, sex, race/ethnicity, and education level), anthropometric measurements (height and weight), healthrelated behaviors (physical activity, alcohol consumption, and cumulative cigarette smoke), clinical factors (hypertension, diabetes, arthritis, and cardiovascular disease (CVD)), and serum biomarkers of inflammation (C-reactive protein and homocysteine). Race/ethnicity was categorized as Non-Hispanic White, NonHispanic Black, Mexican American, and Other. Education level was categorized as
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
2.5. Statistical analysis
Table 1 Participants characteristics (n ¼ 3226). Characteristics
3
Distribution
Age [years] 63.29 Walking speed [ft/sec] 3.41 Blood cadmium a [mg/L] 0.49 Urine cadmium a,b [ng/mL] 0.37 Height [cm] 167.35 Weight [kg] 79.78 C-reactive protein [mg/dL] 0.48 Homocysteine [mmol/L] 9.63 Sex [n (%)] Men 1621 Women 1605 Race/ethnicity [n (%)] Non-Hispanic White 1918 Non-Hispanic Black 481 Mexican American 614 Other 213 Education [n (%)]
±0.22 ±0.02 ±0.02 ±0.08 ±0.16 ±0.39 ±0.02 ±0.08 (46.6) (53.4) (81.8) (7.1) (3.2) (7.9) (24.2) (26.3) (49.5) (60.1) (17.9) (22.0) (47.4) (25.4) (27.2) (38.6) (37.4) (24.0) (45.6) (54.4) (90.5) (9.5) (85.1) (14.9) (59.3) (40.7)
Data in tables are weighted means ± SE for continuous variables and sample size (weighted percentages) for categorical variables. a Geometric means ± SE are presented. b Subsample of participants with urine cadmium information available (n ¼ 1003).
and weight, separately, instead of body mass index (BMI; an obesity index of height divided by weight). Urine creatinine (Cr) concentrations (milligrams per deciliter) were further considered as a confounder when modeling urine cadmium (Barr et al., 2010). In sensitivity analyses, serum cotinine concentrations, cognitive function, and BMI were considered as potential confounders. Serum cotinine concentrations, a biomarker of exposure to active and passive tobacco smoke, was measured by isotope dilution-high performance liquid chromatography/atmospheric pressure chemical ionization tandem mass spectrometry (ID HPLC-APCI MS/MS), and was available in a subsample of participants (n ¼ 3161). Cognitive function was measured in participants 60 years of age using Digit Symbol Substitution Test (DSST) score, and was available in a subsample of participants (n ¼ 2041). BMI was computed as measured weight (kg) divided by measured height squared (m2). Serum cotinine and DSST were normally distributed and therefore controlled as continuous, while BMI was log-transformed to normalize the distribution.
We performed all statistical analyses using SAS survey procedures (ver. 9.4, SAS Institute, Cary, NC) to account for the complex survey design. We used sample weights for 4-year cycles between 1999 and 2002. We used PROC SURVEYREG to estimate the mean change (95% confidence interval (CI)) in walking speed per quintile increase in blood and urine cadmium concentrations. Tests for trend were conducted for ordinal blood and urine cadmium quintiles in regression models using integer values (0e4). We developed sequential models to assess the influence of potential confounding factors: model A was adjusted for age, sex, race/ethnicity, education level, height, and body weight; model B was additionally adjusted for physical activity, smoking status, and alcohol consumption; model C was additionally adjusted for current diagnosis of hypertension, diabetes, arthritis, and CVD; model D was additionally adjusted for C-reactive protein and homocysteine concentrations. When examining associations between walking speed and urine cadmium, we additionally adjusted for urinary Cr concentration as a covariate (Barr et al., 2010). In addition, we examined stratified models by sex, age groups (50e59, 60e69, and 70 years), and cigarette smoke in order to evaluate sex-, age-, smoking-related differences in the associations between cadmium exposure and walking speed. 2.6. Sensitivity analysis First, we conducted sensitivity analyses after further adjusting for serum cotinine and for the DSST score as potential confounders. These sensitivity analyses were performed 1) because cadmium and cotinine are known to share exposure sources, i.e., tobacco €rup et al., 1998), and smoking is related to walking ability smoke (Ja (Fritschi et al., 2013), and 2) because cadmium exposure is known to be associated with cognitive function (Ciesielski et al., 2013), and cognitive function is connected to walking speed (Sheridan et al., 2003). Second, we ran an alternative model after adjusting for BMI instead of adjusting for height and weight. Third, in order to rule out the potential role of variations in cadmium excretion affected by renal function, we performed a sensitivity analysis by running the models after excluding individuals with abnormal glomerular filtration rate (GFR >60 ml/min/1.73 m2), estimated using Chronic Kidney Diseases Epidemiology Collaboration equation (Levey et al., 2009). In our sensitivity analysis for urinary cadmium, we examined alternative modeling by using Cr-corrected cadmium concentrations instead of adjusting for urine creatinine as a covariate. Cr was corrected by dividing cadmium concentration by creatinine concentration (microgram per gram) in each urine sample using a method suggested in order to capture variation in urine concentration (Barr et al., 2005). 3. Results Table 1 shows descriptive characteristics of the entire study population (n ¼ 3226). Overall, mean age (SE) was 63.29 (0.22) years and mean walking speed (SE) was 3.41 (0.02) ft/sec. The geometric means (SE) for blood and urine cadmium concentrations were 0.49 (0.02) mg/L and 0.37 (0.08) ng/mL, respectively. Analyses for urine cadmium concentrations were restricted to participants with urine data (n ¼ 1003). Blood cadmium concentrations were significantly higher in women, older participants, participants with lower height, lower weight, lower education, lower physical activity levels, higher cumulative cigarette smoke, non-diabetes, arthritis, higher C-reactive protein, and higher homocysteine values (Fig. 1 (A)). Urine
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
4
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
cadmium concentrations were significantly higher in participants with higher cumulative cigarette smoke, CVD, higher C-reactive protein, and lower education, and differ across race-ethnicity (Fig. 1 (B)). Blood cadmium concentrations were moderately correlated with urine cadmium concentrations (ng/mL) and Cr-corrected urine cadmium concentrations (mg/g) (Spearman correlation coefficient r ¼ 0.470, p < 0.001 and r ¼ 0.663, p < 0.001, respectively). Table 2 shows the participants’ distribution by quintiles of blood cadmium concentrations. Participants with higher blood cadmium tend to be older and female and to have lower walking speed and less physical activities. For urine cadmium, participants with higher blood cadmium tend to be male and non-Hispanic black and to have less physical activities (Table 3). Table 4 shows associations between blood cadmium concentrations and walking speed. Blood cadmium concentrations were significantly associated with slower walking speed in all sequential models (all p-Trends <0.001). In fully adjusted models (Model D) there was a 0.18 ft/s (95% CI: 0.25, 0.10) decrease in walking speed among participants in the highest quintile (0.8 mg/L) compared to the lowest quintile (<0.4 mg/L) of blood cadmium. Given our model of mean reduction in walking speed, this 0.18 ft/s reduction approximates the estimated reduction in walking speed with a 7-year increase in age, a 12-cm decrease in height [0.19 ft/s], a 25-kg increase in weight [0.18 ft/s], and a 13 mmol/L increase in homocysteine [0.13 ft/s]. This same 0.18 ft/s reduction in walking speed is greater than the reduction found in sex difference [male vs female; 0.08 ft/s], physical activity [low (<7.5 METs-hr/wk) vs. moderate activities (7.5e21 MET hr/wk); 0.15 ft/s], current diabetes [0.14 ft/s], and arthritis [0.11 ft/s] (see Table 4 (Model D) and Table S1). Table 5 shows no significant associations between urine cadmium concentrations and walking speed. Similarly, when running alternative models using Cr-corrected urine cadmium concentrations (microgram per gram), results were not significant (Table S2). In stratified analyses, neither effect modification by sex, age, nor smoking was observed for the associations between blood and urine cadmium concentrations with walking speed (Table S3 for blood and data not shown for urine). We conducted sensitivity analyses after additional adjustment for serum cotinine and cognitive function. We found that associations of blood cadmium with walking speed remained and were even stronger after adjusting for cotinine (Table S4), and these associations remained unchanged after adjusting for cognitive function (Table S5). Also, results were similar (data not shown) when we ran an alternative model after adjusting for BMI instead of adjusting for height and weight. For urine cadmium, results were not changed after adjusting for cotinine, cognitive function, or BMI (data not shown). 4. Discussion In a representative sample of U.S. middle-aged and older adults who participated in NHANES 1999e2002, cadmium concentrations in blood, but not in urine, were associated with declines in walking speed. To our knowledge, this is the first investigation to examine the association between blood cadmium and walking speed. Two previous epidemiological studies (Garcia-Esquinas et al., 2015; Lang et al., 2009) have studied the association between urine cadmium and walking speed, and both have shown similar results as those reported in this article. In the first study, based on NHANES 1999e2004, no association between urine cadmium (as logged concentration) and self-reported problem of walking speed was observed, after controlling for urinary creatinine as a confounder (Lang et al., 2009). In the second, based on NHANES III, no
association was observed between urine cadmium concentrations (as tertile) and slowness (through the eight-foot walking speed test) with adjusted for urinary creatinine (Garcia-Esquinas et al., 2015). Different associations for blood and urine cadmium have been previously reported for other health outcomes (e.g., hypertension) (Staessen et al., 2000; Tellez-Plaza et al., 2008), and may be explained due to differences in metabolism. Although overlap is substantial, blood cadmium is a better indicator of biologically active cadmium from recent exposure while urine cadmium is a better biomarker for long-term exposure (i.e., several decades) (Nordberg et al., 2007). Based on previous studies, we suggest that our results reflect short-term effects of this metal on walking speed. Additionally, null findings for urine cadmium could be influenced by age-related variations in renal function. In this sense, recent studies have suggested that cadmium may accumulate in the kidney until age 60, when variations in renal function would result in changes in urine cadmium excretion (Chaumont et al., 2013). An association between blood cadmium and walking speed could be explained through several mechanisms. First, cadmium could affect walking speed through its effects in the motor nervous system. In this sense, there is evidence from case-control studies suggesting that cadmium may induce motor neuron disease (Pamphlett et al., 2001) and Parkinson's-like motor symptoms (Okuda et al., 1997). Also, in vivo animal studies show changes in the structure of the cerebellar vermis, which has an important role of body posture and motor control, after exposure to cadmium in albino rats (Wahdan et al., 2014). Second, cadmium could affect walking speed through its effects on cognitive function (Ciesielski et al., 2013), which is linked to motor performance (Hausdorff et al., 2005). Additional adjustment for the DSST score did not change our results (Table S5), however, suggesting that changes in cognitive function do not mediate the observed associations. Recently, Mielke et al. have reported an interesting finding from a longitudinal study: increased walking speed at baseline was associated with less cognitive decline, while cognitive function at baseline did not predict changes in walking speed (Mielke et al., 2013). Based on those findings and our results, we suggest that a decline in cognitive function attributable to cadmium exposure does not causally precede a slower walking speed. Third, cadmium is associated with a higher risk of musculoskeletal disorders including bone fragility and fractures (Staessen et al., 1999) or osteoporosis (Gallagher et al., 2008; Jin et al., 2004), all of which are known to be associated with poor motor performance (Gregg et al., 2000; Kwon et al., 2007). A previous experimental study using a rat model of human environmental exposure to cadmium showed that low lifetime exposure to cadmium is associated with demineralization of bone, and may increase the risk of osteoporosis and fractures (Brzoska and Moniuszko-Jakoniuk, 2004). Moreover, a review article of epidemiological studies provides an evidence of exposure to cadmium significantly affecting to the reduced bone mineral density and increased risk for fracture even at lower exposure level (Bhattacharyya, 2009). Then, the bone disorders caused by cadmium exposure may contribute to a decline in walking speed. Fourth, cadmium is a known risk factor for cardiovascular disease (Tellez-Plaza et al., 2013) and hypertension (Tellez-Plaza et al., 2008), which in turn can influence walking speed (Bouillon et al., 2013; Dumurgier et al., 2010). In the current study, however, adjustment for these conditions did not alter the observed associations (see Table 4, Model C). Fifth, oxidative stress and inflammation could also explain the associations between cadmium and walking speed, either as one pathway of overall frailty or in a step throughout various
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
5
Fig. 1. Geometric means (95% CIs) of cadmium concentrations in blood and urine by participant's characteristics. a Cut-off points based on median from total population (n ¼ 3226); median from subjects (n ¼ 1003) assigned to urinary cadmium analysis is height 166.5 cm, weight 77.7 kg, Creactive protein 0.27 mg/dL, and homocysteine 8.77 mmol/L. * Statistical significance (p < 0.05). We used survey t-test for binominal groups and the Wald F-test for categorical groups.
mechanisms of central nervous system neurotoxicity, vascular diseases, and other chronic diseases described above. In this regard,
there is evidence that chronic exposures to cadmium are associated with higher levels of oxidative stress and inflammation (Colacino
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
6
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
Table 2 Characteristics of study population by blood cadmium concentrations (n ¼ 3226). Characteristics
Cadmium Quintiles Q 1 (<0.4 mg/L) (n ¼ 759)
P-Trend Q 2 (0.4e0.5 mg/L) (n ¼ 683)
63.42 Age [years] 61.25 ±0.35a Walking speed [ft/sec] 3.55 ±0.03 3.47 Sex [n (%)] Men 446 (57.5)b 325 Women 313 (42.5) 358 Race/ethnicity [n (%)] Non-Hispanic White 456 (82.2) 410 Non-Hispanic Black 144 (8.8) 86 Mexican American 117 (2.9) 140 Other 42 (6.0) 47 Moderate-to-vigorous physical activity [METs-hours/week] <7.5 452 (53.0) 408 7.5 to <21 139 (21.0) 121 21 168 (26.0) 154 a b
Q 3 (0.5e0.6 mg/L) (n ¼ 572)
Q 4 (0.6e0.8 mg/L) (n ¼ 560)
Q 5 (0.80e7.4 mg/ L) (n ¼ 652)
±0.49 ±0.04
64.65 3.41
±0.40 ±0.04
65.78 3.32
±0.62 ±0.05
62.62 3.25
±0.47 ±0.04
0.002 <0.001
(43.8) (56.2)
264 308
(41.7) (58.3)
254 306
(40.7) (59.3)
332 320
(44.8) (55.2)
<0.001
(82.3) (5.9) (3.5) (8.4)
338 77 117 40
(82.5) (6.0) (3.3) (8.2)
330 70 130 30
(82.9) (6.3) (3.9) (7.0)
384 104 110 54
(79.5) (7.6) (2.8) (10.1)
0.137
(54.0) (20.3) (25.7)
362 87 123
(59.3) (17.3) (23.3)
375 90 95
(62.8) (18.0) (19.2)
490 87 75
(73.6) (11.9) (14.5)
<0.001
Weighted mean ± SE from surveymean (all such values). Weighted percentage from surveyfrequency (all such values).
Table 3 Characteristics of study population by urine cadmium concentrations (n ¼ 1003). Characteristics
Cadmium Quintiles Q1 (0.18 ng/mL) (n ¼ 200)
P-Trend Q 2 (0.18e0.32 ng/ mL) (n ¼ 201)
Age [years] 65.12 ±10.12a 67.21 Walking speed [ft/sec] 3.37 ±0.79 3.26 Sex [n (%)] Men 88 (41.3)b 102 Women 112 (58.7) 99 Race/ethnicity [n (%)] Non-Hispanic White 129 (87.0) 134 Non-Hispanic Black 18 (4.4) 20 Mexican American 47 (4.7) 37 Other 6 (4.0) 10 Moderate-to-vigorous physical activity [METs-hours/week] <7.5 113 (51.1) 108 7.5 to <21 40 (23.6) 37 21 47 (25.2) 56 a b
Q 3 (0.32e0.51 ng/ mL) (n ¼ 201)
Q 4 (0.52e0.85 ng/ mL) (n ¼ 201)
Q 5 (0.86 e36.78 ng/mL) (n ¼ 200)
±9.91 ±0.81
66.32 3.20
±10.41 ±0.81
66.04 3.17
±9.95 ±0.73
64.67 3.23
±9.65 ±0.74
0.865 0.061
(46.2) (53.8)
109 92
(45.9) (54.1)
105 96
(44.6) (55.4)
118 82
(56.1) (43.9)
0.030
(87.5) (4.1) (3.0) (5.4)
123 24 34 20
(80.5) (6.6) (2.3) (10.7)
106 34 44 17
(78.2) (7.6) (3.6) (10.6)
102 49 34 15
(75.8) (11.1) (2.9) (10.2)
0.009
(49.0) (20.6) (30.5)
118 36 47
(56.2) (16.8) (27.0)
140 30 31
(68.7) (16.2) (15.1)
141 29 30
(64.8) (15.2) (19.9)
0.013
Weighted mean ± SE from surveymean (all such values). Weighted percentages from surveyfrequency (all such values).
Table 4 Change (95% CIs) in walking speed (ft/sec) by blood cadmium concentrations (n ¼ 3226). Cadmium Quintile (mg/L)
No.
Model A
Q 1 (<0.4) Q 2 (0.4e0.5) Q 3 (0.5e0.6) Q 4 (0.6e0.8) Q 5 (0.80e7.4) P-Trend
759 683 572 560 652
0 0.01 0.01 0.08 0.24 <0.001
Model Model Model Model
Model B (Reference) (0.08, 0.09) (0.10, 0.08) (0.17, 0.02) (0.31, 0.17)
0 0.01 0.00 0.05 0.20 <0.001
Model C (Reference) (0.07, 0.08) (0.08, 0.09) (0.15, 0.05) (0.27, 0.12)
0 0.00 0.00 0.06 0.20 <0.001
Model D (Reference) (0.07, 0.08) (0.08, 0.08) (0.16, 0.04) (0.28, 0.13)
0 0.00 0.00 0.05 0.18 <0.001
(Reference) (0.08, 0.08) (0.08, 0.09) (0.15, 0.04) (0.25, 0.10)
A: adjusted for age, sex, race/ethnicity, education, height, and weight. B: adjusted for Model A covariates plus physical activity, pack-years of cigarette smoke, and alcohol consumption. C: adjusted for Model B covariates plus hypertension, diabetes, arthritis, and cardiovascular diseases. D: adjusted for Model C covariates plus C-reactive protein and homocysteine.
et al., 2014; Lin et al., 2009) that play an important role in physical function decline (McClure et al., 2014). In our study, when adjusting for inflammation marker of C-reactive protein and homocysteine, the associations between blood cadmium and walking speed were similar but slightly reduced (Table 4, Model D). Finally, we could not rule out the potential role of cadmium accumulation and excretion related to renal function. One U.S. cross-sectional study reported that estimated GFR levels are related
to lower blood cadmium and higher urine cadmium, and suggested measured cadmium levels as a possible consequence of kidney function rather than an effect of cadmium on kidney function (Buser et al., 2016). One study of older patients observed that reduced renal function, as estimated GFR, is associated with reduced walking speed (Lattanzio et al., 2012). When we examined models after excluding individuals with reduced GFR, however, our associations between blood or urine cadmium and walking speed
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
7
Table 5 Changes (95% CIs) in walking speed (ft/sec) by urinary cadmium concentrations (n ¼ 1003). Cadmium Quintile (ng/mL)
No.
Model A
Q 1 (0.18) Q 2 (0.18e0.32) Q 3 (0.32e0.51) Q 4 (0.52e0.85) Q 5 (0.86e36.78) P-Trend
200 201 201 201 200
0 0.06 0.10 0.13 0.08 0.236
Model Model Model Model
Model B (Reference) (0.23, 0.11) (0.25, 0.06) (0.27, 0.01) (0.27, 0.12)
0 0.07 0.08 0.07 0.01 0.798
Model C (Reference) (0.23, 0.09) (0.22, 0.05) (0.21, 0.07) (0.22, 0.19)
0 0.08 0.10 0.09 0.03 0.658
Model D (Reference) (0.24, 0.09) (0.24, 0.04) (0.23, 0.05) (0.24, 0.18)
0 0.09 0.11 0.11 0.05 0.545
(Reference) (0.25, 0.07) (0.26, 0.03) (0.25, 0.04) (0.27, 0.17)
A: adjusted for age, sex, race/ethnicity, education, height, and weight. B: adjusted for Model A covariates plus physical activity, pack-years of cigarette smoke, and alcohol consumption. C: adjusted for Model B covariates plus hypertension, diabetes, arthritis, and cardiovascular diseases. D: adjusted for Model C covariates plus C-reactive protein and homocysteine. All models were adjusted for urinary creatinine concentrations.
were not changed (Tables S6 and S7). The general population is primarily exposed to cadmium through cigarette smoke, dietary intake (shellfish, grain, and vegetables), and ambient air pollution in urban or industrial regions, and cadmium accumulation in the human body (biological half-life: 10e30 years) could contribute to the development of many dis€rup et al., 1998; Nordberg et al., 2007). There is growing eases (Ja evidence that cadmium exposures even at levels currently observed in the general population have adverse effects on other health outcomes including neurocognitive dysfunction (Ciesielski et al., 2013), cardiovascular disease (Tellez-Plaza et al., 2013), diabetes (Schwartz et al., 2003), hypertension (Tellez-Plaza et al., 2008), kidney disease (Navas-Acien et al., 2009), peripheral arterial disease (Fagerberg et al., 2013), osteoporosis (Gallagher et al., 2008) and hearing loss (Choi et al., 2012). Our findings in walking speed support previous observations that indicate adverse health effects of current environmental levels of cadmium exposures. An ongoing trial of repeated chelation with edetate disodium compared to placebo could provide the opportunity to assess the health impact of removing cadmium from the body (Lamas et al., 2016). The trial is measuring the effect of edetate disodium in changing urine and blood cadmium levels over time and the potential benefits for cardiovascular disease. Additional outcomes, relevant for cadmium-related health effects, such as walking speed could be added to this trial. Strengths of the present study include the use of a representative sample of the middle and older general adults, the availability of objectively-measured walking speed, as well as various potential confounders evaluated. Also, our study had limitations. Because NHANES is cross-sectional, we could not infer temporality. In addition, estimating cadmium exposures using single blood and urine measures may raise an issue of within-individual variability. Uncertainties in exposure assessment, however, are likely to be non-differential bias and may lead to an underestimate of risk. 5. Conclusion Cadmium concentrations in blood, but not in urine, are associated with reduced walking speed, which may reflect short-term effects of recent exposures as well as reflect biologically active cadmium. Our findings add to the growing volume of evidence supporting cadmium's toxicity even at low levels of exposure and support the need to further reduce current exposure levels in order to effectively prevent motor function decline and promote healthy aging in the general population. Further research using longitudinal data are needed to confirm these results. Acknowledgment This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded
by the Korea Ministry of Education (grant number 2013R1A6A3A04059556) and the Korea Ministry of Science, ICT and Future Planning (grant number 2015R1C1A2A01054768). Role of the funding source The funders had no role in this study design, data collection and analysis and prepared all results. Conflict of interests None. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2017.09.022. References Ainsworth, B.E., et al., 1993. Compendium of physical activities: classification of energy costs of human physical activities. Med. Sci. Sports Exerc 25, 71e80. Bainbridge, K.E., et al., 2008. Diabetes and hearing impairment in the United States: audiometric evidence from the national health and nutrition examination survey, 1999 to 2004. Ann. Intern Med. 149, 1e10. Barr, D.B., et al., 2010. Urinary concentrations of metabolites of pyrethroid insecticides in the general U.S. population: national Health and Nutrition Examination Survey 1999-2002. Environ. Health Perspect. 118, 742e748. Barr, D.B., et al., 2005. Urinary creatinine concentrations in the U.S. population: implications for urinary biologic monitoring measurements. Environ. Health Perspect. 113, 192e200. Bhattacharyya, M.H., 2009. Cadmium osteotoxicity in experimental animals: mechanisms and relationship to human exposures. Toxicol. Appl. Pharmacol. 238, 258e265. Bouillon, K., et al., 2013. Cardiovascular disease risk scores in identifying future frailty: the Whitehall II prospective cohort study. Heart 99, 737e742. Brzoska, M.M., Moniuszko-Jakoniuk, J., 2004. Low-level exposure to cadmium during the lifetime increases the risk of osteoporosis and fractures of the lumbar spine in the elderly: studies on a rat model of human environmental exposure. Toxicol. Sci. 82, 468e477. Buser, M.C., et al., 2016. Urinary and blood cadmium and lead and kidney function: NHANES 2007-2012. Int. J. Hyg. Environ. Health 219, 261e267. CDC, 2000. National Health and Nutrition Examination Survey Muscle Strength Procedures. Manual Available: http://www.cdc.gov/nchs/data/nhanes/ms.pdf (accessed Dec 7, 2015). CDC, 2001. National Health and Nutrition Examination Survey 1999-2000, Laboratory Procedure Manual. Available: http://www.cdc.gov/nchs/data/nhanes/ nhanes_99_00/lab06_met_lead_and_cadmium.pdf (accessed Dec 7, 2015). CDC, 2002. National Health and Nutrition Examination Survey 1999-2000. Data Documentation, Codebook, and Frequencies: Muscle Strength. Available: http:// wwwn.cdc.gov/Nchs/Nhanes/1999-2000/MSX.htm (accessed Dec 7, 2015). CDC, 2003. Laboratory Procedure Manual: Multiple Toxic Elements, NHANES 19992000. Available: http://www.cdc.gov/nchs/data/nhanes/nhanes_99_00/lab06_ met_hm.pdf (accessed Dec 7, 2015). CDC, 2004. National Health and Nutrition Examination Survey 2001-2002. Data Documentation, Codebook, and Frequencies: Muscle Strength. Available: http:// wwwn.cdc.gov/Nchs/Nhanes/2001-2002/MSX_B.htm (accessed Dec 7, 2015). CDC, 2015. Fourth National Report on Human Exposure to Environmental Chemicals. Available: https://www.cdc.gov/biomonitoring/pdf/fourthreport_ updatedtables_feb2015.pdf (accessed Dec 14, 2016).
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022
8
J. Kim et al. / Environmental Pollution xxx (2017) 1e8
Chaumont, A., et al., 2013. Associations of urinary cadmium with age and urinary proteins: further evidence of physiological variations unrelated to metal accumulation and toxicity. Environ. Health Perspect. 121, 1047e1053. Choi, Y.-H., et al., 2012. Environmental cadmium and lead exposures and hearing loss in u.s. Adults: the national health and nutrition examination survey, 1999 to 2004. Environ. Health Perspect. 120, 1544e1550. Ciesielski, T., et al., 2013. Associations between cadmium exposure and neurocognitive test scores in a cross-sectional study of US adults. Environ. Health 12, 13. Colacino, J.A., et al., 2014. Dietary antioxidant and anti-inflammatory intake modifies the effect of cadmium exposure on markers of systemic inflammation and oxidative stress. Environ. Res. 131, 6e12. Dumurgier, J., et al., 2010. Hypertension and lower walking speed in the elderly: the Three-City study. J. Hypertens. 28, 1506e1514. Ekstrom, H., et al., 2011. Effects of walking speed and results of timed get-up-and-go tests on quality of life and social participation in elderly individuals with a history of osteoporosis-related fractures. J. Aging Health 23, 1379e1399. Fagerberg, B., et al., 2013. Cadmium exposure, intercellular adhesion molecule-1 and peripheral artery disease: a cohort and an experimental study. BMJ Open 3. Fritschi, C., et al., 2013. The effects of smoking status on walking ability and healthrelated quality of life in patients with peripheral arterial disease. J. Cardiovasc Nurs. 28, 380e386. Gallagher, C.M., et al., 2008. Urinary cadmium and osteoporosis in U.S. Women >or¼ 50 years of age: NHANES 1988-1994 and 1999-2004. Environ. Health Perspect. 116, 1338e1343. Garcia-Esquinas, E., et al., 2015. Association of lead and cadmium exposure with frailty in US older adults. Environ. Res. 137, 424e431. Gregg, E.W., et al., 2000. Diabetes and physical disability among older U.S. adults. Diabetes Care 23, 1272e1277. Haskell, W.L., et al., 2007. Physical activity and public health: updated recommendation for adults from the american college of sports medicine and the american heart association. Circulation 116, 1081e1093. Hausdorff, J.M., et al., 2005. Walking is more like catching than tapping: gait in the elderly as a complex cognitive task. Exp. Brain Res. 164, 541e548. IARC, 1993. Cadmium and cadmium compounds, IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Human, 1993/01/01 ed, pp. 119e237. €rup, L., et al., 1998. Health effects of cadmium exposure e a review of the literJa ature and a risk estimate. Scand. J. Work Environ. Health 24, 1e51. Ji, J.S., et al., 2013. Association between blood lead and walking speed in the national health and nutrition examination survey (NHANES 1999-2002). Environ. Health Perspect. 121, 711e716. Jin, T., et al., 2004. Osteoporosis and renal dysfunction in a general population exposed to cadmium in China. Environ. Res. 96, 353e359. Kuo, H.K., et al., 2006. Exploring how peak leg power and usual gait speed are linked to late-life disability: data from the National Health and Nutrition Examination Survey (NHANES), 1999-2002. Am. J. Phys. Med. Rehabil. 85, 650e658. Kuo, H.K., et al., 2007. Relationship of homocysteine levels to quadriceps strength, gait speed, and late-life disability in older adults. J. Gerontol. A Biol. Sci. Med. Sci. 62, 434e439. Kwon, J., et al., 2007. Association between change in bone mineral density and decline in usual walking speed in elderly community-dwelling Japanese women during 2 years of follow-up. J. Am. Geriatr. Soc. 55, 240e244. Lamas, G.A., et al., 2016. Heavy metals, cardiovascular disease, and the unexpected benefits of chelation therapy. J. Am. Coll. Cardiol. 67, 2411e2418. Lang, I.A., et al., 2009. Age-related impairments of mobility associated with cobalt and other heavy metals: data from NHANES 1999-2004. J. Toxicol. Environ.
Health A 72, 402e409. Lattanzio, F., et al., 2012. Relationship between renal function and physical performance in elderly hospitalized patients. Rejuvenation Res. 15, 545e552. Levey, A.S., et al., 2009. A new equation to estimate glomerular filtration rate. Ann. Intern Med. 150, 604e612. Lin, Y.S., et al., 2009. Cadmium exposure is associated with elevated blood Creactive protein and fibrinogen in the U. S. population: the third national health and nutrition examination survey (NHANES III, 1988-1994). Ann. Epidemiol. 19, 592e596. McClure, C.K., et al., 2014. Prospective associations between inflammatory and hemostatic markers and physical functioning limitations in mid-life women: longitudinal results of the Study of Women's Health across the Nation (SWAN). Exp. Gerontol. 49, 19e25. Middleton, A., et al., 2015. Walking speed: the functional vital sign. J. Aging Phys. Act. 23, 314e322. Mielke, M.M., et al., 2013. Assessing the temporal relationship between cognition and gait: slow gait predicts cognitive decline in the Mayo Clinic study of aging. J. Gerontol. A Biol. Sci. Med. Sci. 68, 929e937. Navas-Acien, A., et al., 2009. Blood cadmium and lead and chronic kidney disease in US adults: a joint analysis. Am. J. Epidemiol. 170, 1156e1164. Nordberg, G., et al., 2007. Cadmium. In: Nordberg, G., Fowler, B., Nordberg, M. (Eds.), Handbook on the Toxicology of Metals. Elsevier, Amsterdam, the Netherlands, pp. 446e486. Okuda, B., et al., 1997. Parkinsonism after acute cadmium poisoning. Clin. Neurol. Neurosurg. 99, 263e265. Olmedo, P., et al., 2016. Dietary determinants of cadmium exposure in the strong heart family study. Food Chem. Toxicol. 100, 239e246. Pamphlett, R., et al., 2001. Blood levels of toxic and essential metals in motor neuron disease. Neurotoxicology 22, 401e410. Penninx, B.W., et al., 2000. Lower extremity performance in nondisabled older persons as a predictor of subsequent hospitalization. J. Gerontol. A Biol. Sci. Med. Sci. 55, M691eM697. Schwartz, G.G., et al., 2003. Urinary cadmium, impaired fasting glucose, and diabetes in the NHANES III. Diabetes Care 26, 468e470. Sheridan, P.L., et al., 2003. Influence of executive function on locomotor function: divided attention increases gait variability in Alzheimer's disease. J. Am. Geriatr. Soc. 51, 1633e1637. Staessen, J.A., et al., 2000. Exposure to cadmium and conventional and ambulatory blood pressures in a prospective population study. Public health and environmental exposure to cadmium study group. Am. J. Hypertens. 13, 146e156. Staessen, J.A., et al., 1999. Environmental exposure to cadmium, forearm bone density, and risk of fractures: prospective population study. Public health and environmental exposure to cadmium (PheeCad) study group. Lancet 353, 1140e1144. Studenski, S., et al., 2011. Gait speed and survival in older adults. JAMA 305, 50e58. Studenski, S., et al., 2003. Physical performance measures in the clinical setting. J. Am. Geriatr. Soc. 51, 314e322. Tellez-Plaza, M., et al., 2013. Cadmium exposure and clinical cardiovascular disease: a systematic review. Curr. Atheroscler. Rep. 15, 356. Tellez-Plaza, M., et al., 2008. Cadmium exposure and hypertension in the 1999e2004 national health and nutrition examination survey (NHANES). Environ. Health Perspect. 116, 51e56. US Institute of Medicine, 2002. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients). National Academies Press, Washington, DC. Wahdan, M.H., et al., 2014. Effect of cadmium exposure on the structure of the cerebellar vermis of growing male albino rat. Int. Res. J. Appl. Basic Sci. 8, 163e176.
Please cite this article in press as: Kim, J., et al., Blood and urine cadmium concentrations and walking speed in middle-aged and older U.S. adults, Environmental Pollution (2017), http://dx.doi.org/10.1016/j.envpol.2017.09.022