Science of the Total Environment 408 (2010) 1014–1020

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r.com/ locate /sc i totenv

Plasma levels of copper, manganese and selenium in an adult population in southernSpain: Influence of age, obesity and lifestyle factors

Cristina Sánchez, María López-Jurado, Pilar Aranda, Juan Llopis ⁎Institute of Nutrition and Food Technology and Department of Physiology, University of Granada, E-18071 Granada, Spain

⁎ Corresponding author. Departamento de Fisiología,de Cartuja, Universidad de Granada, E-18071 Granada,fax: +34 958 248959.

E-mail address: jllopis@ugr.es (J. Llopis).

0048-9697/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.scitotenv.2009.11.041

a b s t r a c t

a r t i c l e i n f o

Article history:Received 1 September 2009Received in revised form 13 November 2009Accepted 20 November 2009

Keywords:Mediterranean regionAdultPlasmaCopperManganeseSeleniumLifestyle factors

Copper, manganese and selenium are elements involved in protecting the body against oxidative stress.Determining their plasma level may contribute to assessing the health and nutritional status of populations.The aim of this study was to assess factors influencing copper, manganese and selenium plasma levels in anadult Mediterranean population and to identify groups at risk of deficiency. A cross-sectional survey wascarried out in Andalusia, a region in southern Spain. Blood samples were obtained in a random subsample of340 subjects. Food consumption was assessed by 48-h recall. Height, weight, skinfolds, waist and hipcircumferences were measured. Copper, manganese and selenium were measured in plasma. Informationabout physical exercise, educational level, alcohol and smoking habits was obtained with a structuredquestionnaire. Plasma copper was found to be higher in women than among men. Hypocupraemia was foundin 4.4% of the population, while 9.7% presented hypomanganesemia. Moreover, 86.5% presented plasmaselenium values below 125 µg/L (cutoff for optimal glutathione peroxidase activity). No association wasfound between plasma elements, anthropometric indices and lifestyle factors; there were tendencies, nomore. Copper tended to decrease in obese and increase in sedentary, while selenium tended to decreaseamong smokers. Plasma Cu was positively correlated with the consumption of monounsaturated andpolyunsaturated fats. Plasma Mn was directly correlated with the consumption of dairy products. Levels of Sewere positively correlated with age, the consumption of fruit, vegetables, energy obtained fromcarbohydrates, and the consumption of fibre, and inversely correlated with the consumption of meat andsweets. Our results provide an estimate of the copper, manganese and selenium status in the adultpopulation of southern Spain. The correlations found for Se suggest that there is a tendency for Se levels to bebetter maintained among the population that shows a stronger preference for the traditional diet.

Facultad de Farmacia, CampusSpain. Tel.: +34 958 246676;

ll rights reserved.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Economic development in Spain during recent decades hasfavoured the appearance of social, cultural and dietary changes inthis Mediterranean country. The influence of the mass media, tourismand international trade has led to the appearance of new dietaryhabits characterized by a decrease in the consumption of grainproducts, green vegetables and legumes, and an increase in theconsumption of meat, fruit and dairy products (Serra-Manjem et al.,1995). Abundant information is available on the dietary habits of theSpanish population and on the intakes of different foods and nutrients(Varela et al., 1995; MAPA, 2000). However, data on the nutritionalstatus regarding copper, manganese and selenium in Spain are not soabundant, and is mainly focused on selenium (Navarro et al., 1995;

Alegría et al., 1996; Ferrer et al., 1999; Henriquez et al., 2000; Díaz-Romero and López-Blanco, 2001; Sabé et al., 2002; Navarro-Alarcónand Cabrera-Vique, 2008). Moreover, information on the influence ofage, obesity, educational level, smoking habit, alcohol consumptionand physical exercise on trace elements status in the adult Spanishpopulation is even scarcer, especially for manganese.

Copper, manganese and selenium are required in small amounts ascomponents of antioxidant enzymes; they are actively involved inprotecting the body against oxidative stress (Northrop and Thurham,2007). Therefore, determining their plasma level may contributeto better assessing the health and nutritional status of certainpopulations.

The aims of this study were to evaluate the influence ofdemographic characteristics and lifestyle factors on plasma levelsof Cu, Mn and Se in the adult population in Andalusia, a westernMediterranean region in southern Spain. Our findings highlightgroups at risk of deficiency, and suggest factors that may influencethe status of these nutrients. It is hoped that this information will beuseful in designing future health interventions aimed at modifyingdietary habits.

1015C. Sánchez et al. / Science of the Total Environment 408 (2010) 1014–1020

2. Methods

2.1. Participants

The data reported here were obtained within the framework of alarge-scale study in the region of Andalusia (southern Spain) (Mataixet al., 2003). A cross-sectional epidemiological survey was conductedwith a representative random sample of adults aged between 25 and60 years living in Andalusia. Sampling was probabilistic and stratified,and took place in several stages. The primary sampling unit was citiesand towns (municipalities), the secondary unit was homes, and thetertiary unit was individuals of either gender.

Blood samples were taken for biochemical analysis from a randomsample of 340 subjects (167 men, 173 women). Participants wereasked whether they had any acute or chronic illness, and wereincluded if they were (or appeared to be) in good health; pregnantand lactating women were excluded. The population of participantswho consumed mineral or vitamin–mineral supplements (3.9%) wasalso excluded.

The demographic characteristics of the sample, the samplingsystem, methods of food consumption assessment (48-h recall)(Cameron and Van Staveren, 1988), blood sample collection forbiochemical analysis, and recording of level of education, smoking anddrinking habits, and physical exercise are described in detailelsewhere (Mataix et al., 2005, 2006). The study protocol wasapproved by the Medical-Ethical Committee of the Health Council ofthe Andalusian Regional Government, and informed consent wasobtained from each subject.

For anthropometric measurements, the subjects were shoeless anddressed in their underwear and a disposable gown. Body weight wasmeasured with a portable digital scale (Tefal, Sensitive Computer9202 series 2/0, France) with a precision of 0.1 kg, and height wasmeasured with a portable stadiometer (Holtain Portable, London, UK)with a precision of 1 mm. Skinfolds of the triceps, biceps, subscapularand suprailiac weremeasuredwith skinfold calipers (Holtain, London,UK) with a precision of 0.1 mm. Measurements of the circumferenceof the arm, waist (WC) and hips were taken with a plastic tapemeasure (Holtain, London, UK) with a precision of 1 mm. Allmeasurements were obtained following the techniques and recom-mendations of the International Biological Programme by personnelsuitably trained for this task.

Total body density was calculated using the equations of Durninand Womersley (1974). The percentage of body fat was calculatedfrom the density with Siri's equation (Siri, 1961).

2.2. Analytical methods

In all samples, the content of Cu, Mn and Se in plasma wasanalyzed by FAAS (Perkin Elmer AAnalyst 300 spectrometer, Norwalk,CT, USA). The background correction used was based on the D2-method which required a single element hollow cathode lamp for theelement-specific absorption and a deuterium lamp for the backgroundabsorption. The deuterium arc background correction providedsimultaneous correction for molecular absorption and light-scatter-ing. The source of energy for free atom production was heat in theform of an air–acetylene flame (Air Liquide, Spain). The mixture wasignited in a flame whose temperature ranged from 2100 to 2800 °C.Plasma levels of copper and manganese were determined in samplesthat were diluted in water (quality Milli Q) before the analysis.Selenium levels in plasma were determined by FAAS with the hydridegeneration of samples that had been previously wet-ashed with nitricacid, quality suprapur (Merck) and perchloric acid according to thetechnique of Palacios et al. (1985). All samples were analyzed intriplicate.

The accuracy of the method was evaluated by the analysis of ahuman serum certified reference material (Seronorm™ Trace Ele-

ments ref. MI0181 SERO AS, Billingstad, Norway), and studies ofrecovery in samples of plasma enriched with multielementarystandards. The value obtained for Cu was 1.19±0.05 mg/L (certifiedvalue 1.14–1.22 mg/L), for Mn it was 10.38±0.17 µg/L (certifiedvalue, 10.2–11.2 µg/L), and for Se it was 82±1.99 µg/L (certified value80–86 µg/L). For each element we used the mean of five separatedeterminations.

Levels of Cu in plasma samples were analyzed at the wavelength of324.8 nm (Slit 0.7 nm), using a flow rate (Air/C2H2) of 10/1.9 L/min.For the quantification, we used a five point calibration curve of Cu in aconcentration range of 0–1 mg/L, obtaining r2=0.9997. Plasma levelsof Mn were analyzed at the wavelength of 297.5 nm (Slit 0.2 nm),using a flow rate (Air/C2H2) of 10/1.9 L/min. For the quantification, weused a five point calibration curve ofMn in a concentration range of 0–10 µg/L, obtaining r2=0.9995. Se levels in plasma samples wereanalyzed at the wavelength of 196.0 nm (Slit 2.0 nm), using a flowrate (Air/C2H2) of 8/1.9 L/min. For the quantification, we used a fivepoint calibration curve of Se in a concentration range of 0–1 mg/L,obtaining r2=0.99935.

2.3. Statistical analysis

The crude experimental data were subjected to Student's t test forindependent samples.

Differences between percentage values were verified with anasymptotic test to compare independent binomial proportions. Linearregression analysis was used to find bivariate correlations andPearson's correlation coefficient was calculated for 95% confidencelevels. After controlling by age, gender and anthropometric variables,correlations between analytical values, food and nutrient intake wereobtained with a partial correlation procedure. Linear regressionanalysis was used to estimate the degree of association betweenanalytical values (dependent variable) and anthropometric variable,sex, age, educational level, smoking, drinking and physical exercise.All analyses were done with version 15.0 of the Statistical Package forSocial Sciences (SPSS Inc., Chicago, IL). Differences were consideredsignificant at the 5% probability level.

3. Results

Table 1 shows the characteristics of the study population,including the demographic data (age and gender), the results of theanthropometric study and body composition, lifestyle factors anddaily consumption of energy and macronutrients (always for the totalpopulation and for each gender).

The lifestyle data show that 56% of the population werenonsmokers, and that among the smokers, men were the majority.Alcohol consumption was greater among men, as was the perfor-mance of physical activity. There was a high percentage of inactivityamong the women and among the population in general.

In general terms, our population was characterized as presenting alow energy intake, but with high consumption levels of proteins(which provided almost 16% of the total energy supply) and fat (about39% of energy supply). The consumption of carbohydrates was low(providing only 41% of the energy supply).

Women consumed less energy and nutrients, and there weresignificant differences in all the parameters, except in the percentageof energy provided by protein and carbohydrates, and by fibre.

Thewomen presented plasma copper levels that were significantlyhigher than among the men. However, no differences between thesexes were observed in the mean values of Mn and Se, although thelatter tended to present higher values among the men. These meanvalues were within the range of normality (Sauberlich, 1999).

Table 2 shows themeanplasma values of Cu,Mn and Se for the totalpopulation, by age groups and taking into account anthropometric and

Table 1Characteristics of the participants.

Total population(N=340)

Men(N=167)

Women(N=173)

Age (years) 43.9±10.7 43.7±11.6 44.1±10.2Weight (kg) 70.9±13.5 75.4±13.3 66.5±12.4⁎Height (cm) 162.6±8.9 168.6±7.6 156.8±6.2⁎BMI (kg/m2) 26.8±4.8 26.5±4.2 27.0±5.1Overweight: BMIN25b30 (kg/m2) (%) 36.8 35.9 37.6Obese: BMI≥30 (kg/m2) (%) 19.1 16.8 21.4⁎Waist circumference (WC) (cm) 88.3±14.2 94.0±12.1 82.8±13.8⁎Waist/hip circumference ratio (WHR) 0.86±0.11 0.93±0.08 0.79±0.09⁎Body fat (%) 30.6±8.6 24.4±6.8 36.6±5.8⁎Lifestyle factors

Physical exercise: Sedentarya (%) 66.5 55.7 76.9⁎Active (%) 33.5 44.3 23.1⁎Active (h/week of exercise) 4.4±4.5 5.1±5.3 3.8±3.9⁎

Smoking: Current smokers (%) 44.2 53.9 34.7⁎Nonsmokersb (%) 55.8 46.1 65.3⁎

Drinking: Drinkers (%) 48.0 61.1 35.3 ⁎

Nondrinkersc (%) 52.0 38.9 64.7⁎Educational level: University (%) 17.7 19.2 16.2⁎

Secondary (%) 22.2 26.9 17.3⁎Primary/no schooling(%) 60.1 53.9 66.5⁎

Energy, macronutrient and alcohol intakeEnergy (MJ/day) 9.3±3.2 10.5±3.8 8.2±2.6⁎Protein (g/day) 90.7±34.1 101.6±38.6 80.2±27.2⁎Energy from protein (%) 16.3±3.4 16.2±3.2 16.5±3.5Total fat (g/day) 96.9±41.2 105.5±46.4 88.7±35.5⁎Energy from total fat (%) 39.0±8.2 37.5±7.9 40.5±8.4⁎Saturated fat (g/day) 33.7±14.1 37.3±16.1 30.2±11.4⁎Monounsaturated fat (g/day) 48.7±18.3 53.2±18.3 44.4±18.3⁎Polyunsaturated fat (g/day) 11.4±5.1 12.8±5.4 10.1±4.7⁎Carbohydrates (g/day) 242.2±96.7 270.4±110.6 215.0±77.8⁎Energy from carbohydrates (%) 41.4±9.2 40.9±9.1 41.9±9.3Alcohol (g/day) 11.9±19.4 19.7±28.2 4.4±10.9⁎Energy from alcohol (%) 3.5±5.5 5.5±7.6 1.6±3.6⁎Cholesterol (g/day) 365.5±243.4 422.4±319.7 310.6±169.9⁎Fibre (g/day) 18.5±8.3 19.8±8.3 17.2±8.1

Plasma levelsCu (mg/L) 1.37±0.48 1.31±0.47 1.42±0.49⁎Mn (µg/L) 0.65±0.26 0.64±0.26 0.67±0.26Se (µg/L) 82.7±48.3 85.0±52.7 80.5±44.1

Date are expressed as mean±standard deviation or N (%); Men vs. Women ⁎Pb0.05.a Sedentary: devoted less than 1 h/week to leisure-time physical exercise.b Nonsmokers: “Never” and “Former”.c Nondrinkers: Never drink, or drink only on special occasions.

1016 C. Sánchez et al. / Science of the Total Environment 408 (2010) 1014–1020

lifestyle variables.With greater age, themean plasma values of Cu andSe tended to rise.

In general, the anthropometric parameters measured did notsignificantly modify the plasma levels of the elements studied, exceptthat there was a slight tendency for Cu levels to fall with rising BMI.

Neither significant changes were observed as a consequence ofdifferences in lifestyles, except as slight tendencies. Thus, the sedentarypopulation presented plasma levels of Cu and Se that were slightlyhigher than those of the more active population. Se levels were loweramong populations with a lower educational level. Smokers had lowermean levels of Se than did nonsmokers.

The Pearson correlation coefficients among plasma levels of Cu,Mn and Se for gender, anthropometric variables, nutrient and energyintake and lifestyle were, in general low and not significant.

Table 3 shows the Pearson correlation coefficients among theplasma levels of copper, Mn and Se with age and food and nutrientsconsumed. Plasma levels of Cu were positively correlated with thepercentage of energy provided by fat, the intake of monounsaturatedand polyunsaturated fats, and with the energy provided bycarbohydrates.

Plasma levels of Mn were only correlated with the intake of dairyproducts. However, Se levels were positively correlated with age, withthe intake of fruit vegetables, and with the energy provided bycarbohydrates and fibre. Plasma levels of Se were inversely correlated

with the consumption of meat, sweets and the percentage of energyprovided by fat.

However, when the correlations among the plasma levels of theelements and the intake of nutrients and foodwere adjusted by age, sexand anthropometric variables, no significant correlation remained.

Linear regression analysis did not disclose significant associationsbetween analytical values (dependent variable) and age, gender,anthropometric variable and lifestyle factors.

Fig. 1 shows the study area and the distribution of themean resultsobtained for the three elements examined, in the eight provinces thatconstitute the region of Andalusia.

4. Discussion

Plasma or serum copper is most often used to assess coppernutritional status (Sauberlich, 1999; Gibson, 2005). The mean valuesof plasma Cu in our population were higher than those observed in astudy carried out in the Canary Isles (Henriquez et al., 2000) andsimilar to those recorded in previous studies (Sauberlich, 1999).Although Cu deficiency is not common among humans, our studyrevealed that 4.4% (15 persons) of the total population (340 persons)had hypocupraemia (plasma concentrationsb75 µg/dL) (Sauberlich,1999). This situation might be because in our geographic area(Andalusia), the consumption of foods rich in Cu such as whole

Table 2Plasma levels of Cu, Mn and Se by age group, anthropometric variables and lifestyle factors.

N Cu plasma mg/L Mn plasma μg/L Se plasma μg/L

Age groups (years)25–39 152 1.30±0.47 0.64±0.28 76.7±52.140–49 80 1.42±0.44 0.66±0.26 79.1±44.250–60 108 1.43±0.52 0.66±0.24 89.0±48.2

Anthropometric variablesLevels of BMINonobesea 150 1.39±0.47 0.66±0.27 81.0±49.2Overweightb 125 1.40±0.49 0.65±0.26 84.8±51.3Obesec 65 1.27±0.51 0.68±0.25 79.6±44.5

WHR (waist/hip ratio)WHR (men≤0.95, women≤0.8) 214 1.39±0.45 0.67±0.28 79.2±47.2WHR (menN0.95, womenN0.8) 126 1.34±0.49 0.64±0.24 86.4±49.4

Waist circumferenceWaist (men≤102, women≤88 cm) 249 1.38±0.47 0.65±0.27 82.1±49.1Waist (menN102, womenN88 cm) 91 1.35±0.45 0.66±0.25 84.4±46.3

Body fatBody fat (menb25%, womenb30%) 136 1.38±0.42 0.67±0.31 83.2±54.4Body fat (menN25%, womenN30%) 204 1.36±0.49 0.66±0.24 83.3±45.2

Lifestyle factorsPhysical exerciseSedentaryd 226 1.40±0.50 0.65±0.24 83.5±49.9Active 114 1.31±0.41 0.67±0.36 78.2±44.5

Educational levelUniversity 60 1.35±0. 46 0.63±0.21 89.1±52.6Secondary 75 1.34±0.41 0.64±0.29 84.5±49.2Primary/no schooling 205 1.39±0.49 0.66±0.26 80.5±48.0

Alcohol consumptionNondrinkerse 177 1.35±0.51 0.69±0.24 80.0±49.3Drinkers 163 1.39±0.42 0.62±0.27 84.4±47.6

SmokingNonsmokersf 190 1.34±0.54 0.67±0.25 85.0±49.4Smokers 150 1.44±0.49 0.63±0.29 74.1±45.2

Data are expressed as mean±standard deviation.a Nonobese: BMIb25.b Overweight: BMI≥25, b30.c Obese: BMI≥30.d Sedentary: devoted less than 1 h/week to leisure-time physical exercise.e Nondrinkers: Never drink or drink only on special occasions.f Nonsmokers: “Never” and “Former”.

1017C. Sánchez et al. / Science of the Total Environment 408 (2010) 1014–1020

cereals, dried fruits, and legumes (Mataix and Llopis, 2009), tends tobe low (Planells et al., 2003).

Among the study population there was a tendency for obesepersons to present lower plasma levels of Cu. This could be related tothe fact that in our population the obese tend to consume fewerpolyunsaturated fats (Mataix et al., 2005). The positive correlationbetween the two parameters (see Results) supports these hypotheses.

Various factors have been proposed as possibly affecting plasmalevels of Cu, including age. Plasma Cu levels are higher in the elderlythan among younger populations (Fischer et al., 1990; Sauberlich,1999). In our study, there was found to be a tendency to higher levelsof Cu in plasma with increased age. However, the differences amongage groups were not significant. This might be due to the fact that theages in our population ranged from 25 to 60 years, and a wider rangeis needed to reveal significant differences.

Gender also influences plasma Cu. In our study, the womenpresented mean plasma Cu levels that were significantly higher thanthose of the men, a fact that as also been observed in other studies(Fischer et al., 1990; Johnson et al., 1992). This finding could be due tothe fact that adult women (aged 20–59 years) present higher levels ofabsorption than do men, and thus have a more rapid turnover of Cuonce it is absorbed (Johnson et al., 1992).

There have been found to be higher concentrations of Cu amongmaleuniversity athletes, in relation to controls (Lukaski et al., 1983). The resultsobtained in thepresent study revealed theopposite trend, however; therewere lower concentrations of Cu in plasma among the active population.This might be because, on average, the number of hours per weekdedicated to physical exercise among the active population (Table 1) ismarkedly lower than that corresponding to the athletes.

It has been reported that smoking increases plasma levels of Cu, andthat these levels are positively correlated with CuZn-SOD activity. Bothplasma Cu and CuZn-SOD activities have been shown to increase inresponse to the chronic inflammation of the respiratory tract found insmokers (Northrop and Thurham, 2007). In our case, there was only aslight trend towards suchan increase, possibly because our smokersmaybe classed as only “moderate smokers”, with an average consumption of16 cigarettes per day.

Although to date no method has been established for evaluatingthe nutritional status of Mn among humans, the values in serum orplasma are commonly used (Sauberlich, 1999).

The mean values obtained in this study, for our population, aresimilar to those observed among post menopausal women in France(Bureau et al., 2002) and are within the reference limits (0.36 µg/L–0.96 µg/L) (Aggett, 1991). However, we observed that 9.7% of thepopulation (33 persons) presented plasma values that were lowerthan the reference values. The mean values reported here for Mn arebelow those observed among adult males inWisconsin (USA) (Gregeret al., 1990), Turkish children (Kocyigit et al., 2004) and among thepopulation of the Canary Isles (Henriquez et al., 2000), althoughamong the latter population there were wide ranges of variation.There were also found to be wide ranges of variation in a study carriedout in India (Sharma and Pervez, 2005).

Although it has been reported that sports activities raise levels ofMn (Nasolodin and Gladkikh, 2007), in our study mean levels of Mnwere almost unaffected by age, anthropometric variables or lifestyle.

A current object of study is the use of different selenoproteins, Se-dependent enzymes, thyroid hormones, etc. as markers of Se status.However, the concentration of Se in plasma, and also in other body

Table 3Pearson correlation coefficients among plasma levels of Cu, Mn and Se with age, andfood and macronutrient intake.

Cu plasma Mn plasma Se plasma

Age r=0.043 r=0.039 r=0.134⁎Foods

Total grain products r=0.036 r=0.035 r=0.108Meat r=0.066 r=−0.006 r=−0.114⁎Fish r=0.039 r=−0.044 r=0.057Eggs r=−0.064 r=0.057 r=0.061Dairy products r=−0.096 r=0.117⁎ r=−0.013Fruits r=−0.027 r=0.078 r=0.139⁎Vegetables r=0.025 r=−0.30 r=0.164⁎Pulses r=0.059 r=−0.020 r=0.051Sweets r=0.041 r=−0.088 r=−0.113⁎Nuts r=−0.002 r=−0.001 r=0.035

Energy and macronutrientEnergy r=0.051 r=0.062 r=0.006Protein r=0.010 r=0.050 r=−0.001Energy from protein r=−0.095 r=−0.055 r=−0.031Total fat r=0.094 r=0.059 r=−0.061Energy from total fat r=0.138⁎ r=0.022 r=−0.136⁎Saturated fat r=0.053 r=0.044 r=−0.086Monounsaturated fat r=0.123⁎ r=0.095 r=−0.030Polyunsaturated fat r=0.124⁎ r=0.024 r=0.014Carbohydrates r=−0.010 r=0.076 r=0.076Energy from carbohydrates r=−0.123⁎ r=0.051 r=0.127⁎Energy from total alcohol r=0.066 r=0.066 r=0.066Cholesterol r=0.010 r=0.064 r=−0.038Fibre r=−0.016 r=0.080 r=0.236⁎

⁎Pb0.05.

1018 C. Sánchez et al. / Science of the Total Environment 408 (2010) 1014–1020

fluids and tissues, is an indicator that is frequently used to evaluatethe Se status (Sauberlich, 1999; Gibson 2005).

The mean values of plasma Se found in our population (Table 1)are close to those reported by other authors for other parts of Spain(Navarro et al., 1995; Alegría et al., 1996; Ferrer et al., 1999;Henriquez et al., 2000; Díaz-Romero and López-Blanco, 2001; Sabé

Fig. 1. Location of the study area, with mean plasma

et al., 2002). Our values are also within the range of values observed indifferent regions of Europe, including Antwerp (Belgium), the CzechRepublic, Hannover (Germany), Central Italy and Vienna (Austria)(Sabé et al., 2002; Navarro-Alarcón and Cabrera-Vique, 2008) andGreece (Letsiou et al., 2009). However, they are notably lower thanthose found among the population of NHANES III (USA) and Taiwan(China) (Navarro-Alarcón and Cabrera-Vique, 2008).

There is a certain difficulty in establishing reference values forplasma Se, given the wide variety of results reported, depending onthe geographic area in which the study was carried out (Gibson, 2005;Navarro-Alarcón and Cabrera-Vique, 2008). We believe it might be ofinterest to note that our values are within the range of 28–200 µg/L(the values measured for 95% of the population were in the range of40–165 µg/L). This is similar to the range described in previous studiescarried out in Spain (Navarro et al., 1995; Romero et al., 2001).Furthermore, we observed that 86.5% of the population presentedvalues below 125 µg/L, the cutoff considered to be required foroptimal glutathione peroxidase activity (Thomson, 2004), and that53.8% presented values below 82 µg/L, the cutoff considered to berequired for optimal iodothyronine deiodinase activity (Gibson,2005), which suggests that Se intake in southern Spain is below thatrecommended for optimal health. Similar results have been obtainedin Greece (Letsiou et al., 2009).

The dietary intake of Se is possibly the major determinant factor inthe plasma concentration of Se. Moreover, it is known that the plasmacontent of Se presents important regional variations. These differ-ences are mainly due to the content and availability of Se in foods. TheSe content in food of animal origin reflects the levels of Se present inthe diet consumed, while that found in food of vegetable origin isdirectly affected by the content and physical and chemical form of Sein the soil. Moreover, the Se content of foods may also be influencedby their processing and preparation. However, the Se content of wateris relatively unimportant, as drinking water is generally a poor sourceof Se (Navarro et al., 1995; Gibson, 2005; Navarro-Alarcón andCabrera-Vique, 2008).

levels of Cu, Mn and Se distributed by provinces.

1019C. Sánchez et al. / Science of the Total Environment 408 (2010) 1014–1020

In our study, it was observed that plasma levels of Se werepositively correlated with the intake of fruit and vegetables and theenergy obtained from carbohydrates; this accounts for the positivecorrelation with fibre intake (see Results). In southern Spain,carbohydrates, vegetables and fruit constitute the most importantsource of fibre (MAPA, 2000). However, plasma Se is inverselycorrelated with the consumption of meat and sweets (Letsiou et al.,2009).

Unlike other authors (Burri et al., 2008) we found no significantdifferences by gender (Romero et al., 2001; Van Cauwenbergh et al.,2007; Letsiou et al., 2009) (Table 1). Neither were there significantvariations with regard to age; this finding has also been reported, forthe adult population, by other authors (Sabé et al., 2002; Burri et al.,2008). However, we did find a positive correlation between plasma Seand age (see Results). In our study, the more elderly population wasmore likely to maintain the traditional diet, with a significantly higherconsumption of fruit, vegetables and carbohydrates, and a lowerintake of sweets (Planells et al., 2003). Nevertheless, at advanced ages(N70 years), there was a negative association between Se and age(Gibson, 2005; Rousseau et al., 2006; Arnaud et al., 2007).

As in previous studies, no relation was observed between plasmalevels of Se and the anthropometric variables studied (Letsiou et al.,2009). However, among the older population, it was noted thatdecreased levels of plasma Se were associated with a higherproportion of obesity (Arnaud et al., 2007).

In elderly people, serum Se was associated with higher levels ofphysical activity (González et al., 2007). However, other authors havefound no differences in levels of plasma Se between sedentary personsand those taking physical exercise (Rousseau et al., 2006; Letsiou etal., 2009). We did not observe any variations in plasma Se related tophysical activity. Neither was plasma Se found to be associated withthe level of education, which is in accordance with the resultsobtained by other authors (Rousseau et al., 2006).

It is well known that high levels of alcohol consumption canprovoke a fall in levels of plasma Se, and that such a decrease seems tobe caused by liver disorders more than by intake deficiencies (Maillotet al., 2001; Bergheim et al., 2003). In our study, no significantdifferences were observed between drinkers and nondrinkers. Themoderate-to-low alcohol intakes in the population we studied (1–2drinks/day) (Sánchez et al., 2009) (Table 1) may explain why wefound no impairment in Se status among drinkers.

It has been reported that smoking has a negative impact on plasmaSe (Gibson, 2005; Letsiou et al., 2009). In our study, there was onlyseen to be a tendency for this level to fall (Table 2). Earlier studieshave also failed to record any significant association between smokingand levels of plasma Se (Arnaud et al., 2007) or of Se in total blood(Chen et al., 2007).

Although soil and water composition varies within our study area,an analysis of the geographic distribution of the mean plasma valuesof the three elements analyzed did not reveal major differencesbetween the provinces of Andalusia (Fig. 1). However, there wereseen to be higher levels of Cu in western Andalusia (Huelva province),an area that has historically been known for copper deposits; thismight have produced a contamination of the soil and water, and thusprovoked higher plasma Cu levels among the population. The soil inthe study area is, on average, poor in Se and Mn (Aguilar et al., 1999).The highest plasma levels of Mn were found in the central area ofAndalusia (Córdoba province), while Se levels were highest in theeastern provinces of Almería, Granada and Jaén.

In conclusion, the mean plasma levels of Cu, Mn and Se among thehealthy adult population living in Andalusia are similar to thosemeasured among comparable populations in earlier studies, andwithin the range of normality. We found significant differences bygender in the case of Cu, but not for Mn or Se. In our population, nosignificant associations were found between the anthropometricindices and lifestyle factors and plasma levels of Cu, Mn and Se,

although there were found to be tendencies. However, the positivecorrelation between plasma Se and age suggests that there is atendency for Se levels to rise among the population with a strongerpreference for the traditional diet. Although the interpretation of datafrom survey studies can be complex, our results provide a preciseestimate of the status for copper, manganese and selenium among theadult population of southern Spain. It is hoped that this informationwill be useful in designing future studies and in healthcare interven-tions aimed at assessing the health and nutritional status of thepopulation.

Acknowledgments

We thank the Escuela Andaluza de Salud Pública in Granada, Spain,the Dirección General de Salud Pública and the Health Council of theAndalusian Regional Government for their support.

References

Aggett PJ. Manganese. In: Fidanza F, editor. Nutritional status assessment. London:Chapman & Hall; 1991. p. 412.

Aguilar J, Galán E, Gómez JL. Estudio de los elementos traza en los suelos de Andalucía (I).ConsejeríadeMedioAmbiente. JuntadeAndalucía; 1999. http://www.juntadeandalucia.es/medioambiente/site/web/Bloques_Tematicos/Estado_YCalidad_De_Los_Recursos_Naturales/Suelo/Suelo.pdf.

Alegría A, Barberá R, Clemente G, Farré R, García MJ, Lagarda MJ. Selenium andglutathione peroxidase reference values in whole blood and plasma of a referencepopulation living in Valencia, Spain. J Trace Elem Med Biol 1996;10:223–8.

Arnaud J, Akbaraly NT, Hininger AMR, Berr C. Factors associated with longitudinalplasma selenium decline in the elderly: the EVA study. J Nutr Biochem2007;18:482–7.

Bergheim I, Parlesak A, Dierks C, Bode JC, Bode C. Nutritional deficiencies in Germanmiddle-class male alcohol consumers: relation to dietary intake and severity ofliver disease. Eur J Clin Nutr 2003;57:431–8.

Bureau I, Anderson RA, Arnaud J, Raysiguier Y, Favier AE, Roussel AM. Trace elements inpost menopausal women: impact of hormonal replacement therapy. J Trace ElemBiol 2002;16:9-13.

Burri J, Haldimann, Dudler V. Selenium status of the Swiss population: assessment andchange over a decade. J Trace Elem Med Biol 2008;22:112–9.

CameronME, Van StaverenWA. Manual onmethodology for food consumption studies.Oxford: Oxford University Press; 1988. 83 pp.

Chen Y, Hall M, Graziano JH, Slavkovich V, van Green A, Parvez F, et al. A prospectivestudy of blood selenium levels and the risk of arsenic-related premalignant skinlesions. Cancer Epidemiol Biomarkers Prev 2007;16:207–13.

Díaz-Romero CD, López-Blanco FL, Henriquez-Sánchez P, Rodríguez-Rodríguez E, Serra-Manjem LS. Serum selenium concentration in a representative sample of theCanary Island population. Sci Total Environ 2001;269:65–73.

Durnin JVGA,Womersley J. Body fat assessed from total body density and its estimationfrom skinfold thickness: measurements on 481 men and women aged 16 to72 years. Br J Nutr 1974;32:77–84.

Ferrer E, Alegría A, Barberá R, Farre R, Lagarda MJ, Monleon J. Whole blood seleniumcontent in pregnant women. Sci Total Environ 1999;227:139–43.

Fischer PWF, L'Abbe MR, Giroux A. Effects of age, smoking, drinking, exercise andestrogen use on indices of copper status in healthy adults. Nutr Res 1990;10:1081–90.

Gibson R. Principles of nutritional assessment. Oxford: Oxford University Press; 2005.683–796 pp.

González S, Huerta JM, Fernández S, Patterson AM, Lasheras C. Life-quality indicators inelderly people are influenced by selenium status. Aging Clin Exp Res 2007;19:10–5.

Greger JL, Davis CD, Suttie JW, Lyle BJ. Intake, serum concentrations, and urinaryexcretion of manganese by adult males. Am J Clin Nutr 1990;51:457–61.

Henriquez P, Díaz C, Rodríguez E, López F, Álvarez E, Díaz J, et al. Biochemicalassessment of nutritional status in the Canary Island population. Arch LatinoamNutr 2000;50:43–53.

Johnson PE, Milne DB, Lykken GI. Effects of age and sex on copper absorption, biologicalhalf-life, and status in humans. Am J Clin Nutr 1992;56:917–25.

Kocyigit A, Armutcu F, Gurel A, Ermis B. Alterations in plasma essential trace elementsselenium, manganese, zinc, copper, and iron concentrations and the possible role ofthese elements on oxidative status in patients with childhood asthma. Biol TraceElem Res 2004;97:31–41.

Letsiou S, Nomikos T, Panagiotakos D, Pergantis S, Fragopoulou E, Antonopoulou S, et al.Serum total selenium status in Greek adults and its relation to age. The ATTICAstudy cohort. Biol Trace Elem Res 2009;128:8-17.

Lukaski HC, Bolonchuk WW, Klevay LM, Milne DB, Sandstead HH. Maximal oxygenconsumption as related to magnesium, copper, and zinc nutriture. Am J Clin Nutr1983;37:407–15.

Maillot F, Farad S, Lamisse F. Alcool et nutrition. Pathol Biol 2001;49:683–8.Mataix J, Llopis J. In:Mataix J, editor.Minerales. Nutrición y AlimentaciónHumanaMadrid:

Ergón; 2009. p. 265-231.

1020 C. Sánchez et al. / Science of the Total Environment 408 (2010) 1014–1020

Mataix J, Aranda C, Sánchez C, Montellano M, Planells E, Llopis J. Assessment of thiamin(vitamin B1) and riboflavin (vitamin B2) status in an adult Mediterraneanpopulation. Br J Nutr 2003;90:661–6.

Mataix J, López-Frías M, Martínez de Victoria E, López-Jurado M, Aranda P, Llopis J.Factors associated with obesity in an adult Mediterranean population: influence onplasma lipid profile. J Am Coll Nutr 2005;24:456–65.

Mataix J, Aranda C, López-Jurado, Sánchez C, Planells E, Llopis J. Factors influencing theintake and plasma levels of calcium, phosphorus andmagnesium in southern Spain.Eur J Nutr 2006;45:349–54.

Ministerio de Agricultura Pesca y Alimentación (MAPA). La alimentación en España.Ministerio de Agricultura Pesca y Alimentación. Dirección General de Alimentación.Madrid; 2000.

Nasolodin VV, Gladkikh IP. Manganese provision in trained and untrained school-children and student in different seasons. Gig. Sanit. 2007;1:59–61.

NavarroM, López H, Ruiz ML, González S, Pérez V, LópezMC. Determination of seleniumin serum by hydride generation atomic absorption spectrometry for calculation ofdaily dietary intake. Sci Total Environ 1995;175:245–53.

Navarro-Alarcón M, Cabrera-Vique C. Selenium in food and the human body: a review.Sci Total Environ 2008;400:115–41.

Northrop CA, Thurham DI. Trace elements: selenium, copper, zinc, and iron. Clin ChimActa 2007;377:14–38.

Palacios MA, Arévalo I, Cámara C. Determinación de selenio en muestras biológicas porabsorción atómica mediante generación de hidruros. Quim Anal 1985;4:320–5.

Planells E, Sánchez C, Montellano MA, Mataix J, Llopis J. Vitamin B6 and B12 and folatestatus in an adult Mediterranean population. Eur J Clin Nutr 2003;57:777–85.

Romero CD, López F, Henríquez P, Rodríguez E, Serra L. Serum selenium concentrationin a representative sample of the Canarian population. Sci Total Environ 2001;269:65–73.

Rousseau AS, Margaritis I, Arnaud J, Faure H, Roussel AM. Physical activity altersantioxidant status in exercising elderly subjects. J Nutr Biochem 2006;17:463–70.

Sabé R, Rubio R, Gárcia-Beltran L. Reference values of selenium in plasma populationfrom Barcelona. Comparison with several pathologies. J Trace Elem Med Biol2002;16:231–7.

Sánchez C, López-Jurado M, Planells E, Llopis J, Aranda P. Assessment of iron and zincintake and related biochemical parameters in an adult Mediterranean populationfrom southern Spain: influence of lifestyle factors. J Nutr Biochem 2009;20:125–31.

Sauberlich HE. Laboratory tests for the assessment of nutritional status. Boca Raton, FL:CRC Press; 1999. 313 pp.

Serra-Manjem L, Ribas L, Tresserras R, Ngo J, Salleras L. How could changes in dietexplain changes in coronary heart diseasemortality in Spain? The Spanish paradox.Am J Clin Nutr 1995;61:S1351–9 Suppl.

Sharma R, Pervez S. Toxic metals status in human blood and breast milk samples in anintegrated steel plant environmental in Central India. Environ Geochem Health2005;27:39–45.

Siri WB. Body composition from fluid spaces and density: analysis of the methods.Techniques formeasuring body composition. Washington DC: National Academy ofSciences, National Research Council; 1961. p. 223–44.

Thomson CD. Assessment of requirements for selenium and adequacy of seleniumstatus: a review. Eur J Clin Nutr 2004;58:391–402.

Van Cauwenbergh R, Robberecht H, van Vlaslaer V, De Smet A, Emonds MP, Hermans N.Plasma selenium levels in healthy blood donors in the central eastern part ofBelgium. J Trace Elemen Med Biol 2007;21:225–33.

Varela G, Moreiras O, Carbajal A, CampoM. Encuestas de presupuestos familiares 1990–1991: Estudio nacional de nutricional y alimentación 1991. Madrid: InstitutoNacional de Estadística; 1995.