Survey of the chemical composition of 571 European bottled mineral waters

Journal of Food Composition and Analysis 24 (2011) 376–385

Contents lists available at ScienceDirect

Journal of Food Composition and Analysis journal homepage: www.elsevier.com/locate/jfca

Original Article

Survey of the chemical composition of 571 European bottled mineral waters Daniela Bertoldi a,*, Luana Bontempo a, Roberto Larcher a, Giorgio Nicolini a, Susanne Voerkelius b, Gesine D. Lorenz b, Henriette Ueckermann c, Heinz Froeschl d, Malcolm J. Baxter e, Jurian Hoogewerff c, Paul Brereton e a

FEM-IASMA Fondazione Edmund Mach – Istituto Agrario di San Michele all’Adige, via E. Mach 1, 38010 San Michele all’Adige, Italy Hydroisotop GmbH, Woelkestrasse 9, 85301 Schweitenkirchen, Germany UEA - Centre for Forensic Provenancing, School of Chemistry, University of East Anglia, Norwich NR4 7TJ, United Kingdom d Seibersdorf Labor GmbH, A-2444 Seibersdorf, Austria e FERA - The Food and Environment Research Agency, Sand Hutton, York YO41 1LZ, United Kingdom b c

A R T I C L E I N F O

A B S T R A C T

Article history: Received 19 January 2010 Received in revised form 13 July 2010 Accepted 31 July 2010 Available online 27 November 2010

As part of the European TRACE project (Tracing Food commodities in Europe, VI FP, Contract N. 006942), this paper provides a wide-ranging survey of the chemical composition of 571 mineral waters bottled and marketed in 23 European countries, and discusses 39 compositional parameters (specific electric conductivity, pH, hardness, total alkalinity, ammonia, chloride, fluoride, nitrate, nitrite, sulphate, Ca, K, Mg, Na, Al, B, Ba, Cd, Ce, Co, Cs, Cu, La, Li, Lu, Mn, Mo, Nd, Ni, Pb, Rb, Se, Sm, Sr, Tl, U, V, Yb, Zn) mainly referring to legal limits and nutritional implications. According to European legislation 58.1% of samples could be defined as ‘suitable for a low-sodium diet’ while 8.1% could be defined as ‘containing sodium’, 13.7% could be labelled as ‘containing magnesium’, 10.2% as ‘containing fluoride’, 4.9% as ‘containing chloride’, 13.5% as ‘containing sulphate’ and 17.5% as ‘containing calcium’. 2.8% of samples did not conform with European Community limits for at least one parameter (Se, NO2 , Mn, Ni, Ba, F and NO3 ). About 9% of samples had boron, nitrate or nitrite levels above the legal limit existing in individual European countries. ß 2010 Elsevier Inc. All rights reserved.

Keywords: Mineral waters Product analysis Inorganic composition Trace elements Regulation Nutrition Health concern Food composition Food analysis

1. Introduction Natural mineral water is defined as a ‘microbiologically wholesome water originating in an underground water table or deposit, and emerging from a spring tapped at one or more natural or bore exits’ (Council Directive 1980/777/EC and Commission Directive 1996/70/EC and 2009/54/EC), with a distinguishing and constant chemical composition. The global bottled water market reached a value of $60,938 million (115,393.5 million L) in 2006 and is forecast to increase in the future (King, 2008). Nowadays, many people prefer bottled mineral water, deeming it to be more carefully controlled, safer and healthier, and even therapeutic because of its mineral content. Oligomineral waters are considered useful for people with hyperuricemia, gout, urolithiasis and hypertension whereas calcium-rich mineral waters are useful for growing children, menopausal and pregnant women and elderly people, mainly

* Corresponding author. Tel.: +39 0461 615139; fax: +39 0461 615288. E-mail address: [email protected] (D. Bertoldi). 0889-1575/$ – see front matter ß 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jfca.2010.07.005

contributing towards the daily intake of this mineral. The list recognised by the Member States of the European Union includes more than 2000 natural mineral water sources (OJEC, 2002a,b,c; OJEU, 2005, 2006; http://ec.europa.eu/food/food/labellingnutrition/water/mw_eulist_en.pdf, retrieved 19/10/09). Legal limits are fixed for elemental composition (Table 1). Although several minerals (e.g. Ca, Mg, Co, Cu, Mo, Se, Zn) are required for human health, others – such as As, Cd, Pb but also essential micronutrients when present in excessive levels – could be harmful (Pais and Jones, 1997). A mineral water can only be sold if the concentration of constituents is within the limits set by the Regulations and the maximum allowable concentrations are calculated by considering the nature of the constituent, its possible degree of toxicity and the long-term maximum daily intake. In a few cases and in some countries, specific age brackets are taken into account, e.g. some maximum limits are defined for consumption by infants in Austria, Germany and Italy. In the last few years many studies have focused on the elemental composition of bottled and some mineral waters in individual countries – such as Nigeria (Nkono and Asubiojo, 1997), Canada (Dabeka et al., 2002), Sweden (Rosborg et al., 2005), Greece

D. Bertoldi et al. / Journal of Food Composition and Analysis 24 (2011) 376–385

377

Table 1 Maximum allowable content (mg/L) of mineral elements in mineral water established by different European Countries, the European Community, World Health Organisation and United States Environmental Protection Agency. Components

Aluminium Ammonium Barium Boron Cadmium Calcium Chlorides Copper Fluorides Fluorides declaration Hydrogencarbonate Lead Magnesium Manganese Molybdenum Nickel Nitrates Nitrites Potassium Selenium Sodium Sulphates Thallium Uranium Zinc

ECc

Austriad Austriaa,d Francee Germanyf

2003

2006

1.0

1.0

0.003

0.003

1999

2007

Germanya,f Italyg

2006

1.0

1.0

0.003

0.003

1.0 5.0

1.0 (5.0 since 2008) 1.5

0.010

0.010

Italya,g Netherlandsh Spaini Englandj WHOb,k US-EPAb,l

2003

2007

1.0 5.0 0.003

1.0

1.0 5.0

1.0 5.0

0.010

2006 (0.2) (0.5) 1.0 1.0 0.003

2008

(0.05–0.2)

0.003

0.7 0.5 0.003

0.005

1.0 5.0 1.5

1.0 5.0

2 1.5

(250) 1.3 (1) 4 (2)

0.010

0.010

0.010

0.01

0.015

0.50

0.50

(0.05)

0.50

(0.05)

0.020

0.020

0.020

0.020

0.4 0.07 0.07

0.003

1.0

175 50 1.0 5.0

1.0 5.0 1.5

0.010

0.010

1.5

0.7

2009

1.5

2

550 50 0.50

0.50

0.50

0.50

0.020

0.020

0.020

0.050 (0.020 since 2008) 50 0.1

50 0.1 0.010

25 0.1

10 0.02 10

0.010

50 0.1 0.010

0.05

10 0.02

0.010

45 0.02

10

0.010

20 240

50 0.1 0.010

50 0.1

50 0.1

0.010

50 3

0.010

10 (as N) 1 (as N)

0.01

20 240 0.002

0.015

0.05 (250) 0.002 0.03 (5)

EC = European Community; WHO = World Health Organisation; US-EPA = United States Environmental Protection Agency (US-EPA). The maximum contaminant level is given. a Waters suitable for preparation of infant food. b Refers to drinking waters. c Commission Directive 2003/40/EC of 16/05/2003. d Mineralwasser- und Quellwasserverordnung. e Arreˆte´ du 14/03/2007. f Mineral- und Tafelwasserverordnung vom 01/08/84. g Decreto 29/12/2003. h Besluit van 6/11/2003. i Real decreto 1744/2003 de 19/12/2003. Values in brackets are indicative parameters. j The Natural Mineral Water, Spring Water and Bottled Drinking Water. (England) Regulation 2007. k WHO (2008), no. 1/4. l US-EPA, 2009. 816-F-09-004. Values in brackets are non-enforceable guidelines (National Secondary Drinking Water Regulation; NSDWR). Please refer to ‘‘References’’ section for full bibliographic citation.

(Soupioni et al., 2006; Karamanis et al., 2007) and Turkey (Baba et al., 2008) – or of a limited number of samples from several countries (Misund et al., 1999; Lau and Luk, 2002; Krachler and Shotyk, 2009) or simply derived from the compositional information given on bottle labels (Naddeo et al., 2008). Several studies have dealt with the influence of leaching from different packaging materials on water composition (Misund et al., 1999; Shotyk and Krachler, 2007a,b; Krachler and Shotyk, 2009). Results for mineral waters collected in 23 countries and analysed within the context of the TRACE project (Tracing Food commodities in Europe, VI FP, Contract N. 006942) are presented in this paper. With regard to product analysis, it provides the widest overview of the chemical composition of European bottled mineral waters currently available.

2. Materials and methods 2.1. Sampling Five hundred and seventy-one bottled mineral waters (159 in glass, 407 in PET and 5 in Tetra Pak) from different brands and sources were bought on the market in 23 European countries during 2005 and 2006 (Table 2), stored at 10–15 8C in the dark and analysed within 2 years after bottling. The geographical origin of the mineral waters was gathered from the label of each bottle and, when possible, confirmed from the EU lists (OJEU, 2005). The sampling was organised to get a widespread overview of the European market and to obtain a relatively complete geologic and geographic coverage. The unequal number of samples available in

Table 2 Mineral waters sampled for each country. Country

Country code

No. of samples

Country

Country code

No. of samples

Country

Country code

No. of samples

Austria Belgium Bosnia i Herzegovina Croatia Czech Republic Denmark Finland

A B BiH HR CZ DK FIN

17 1 1 3 4 3 4

France Germany Great Britain Greece Hungary Ireland Italy Netherlands

F D GB GR H IRL I NL

32 185 34 2 9 3 186 3

Norway Poland Portugal Russia Slovenia Spain Switzerland Ukraine

N PL P RUS SLO E CH UA

4 17 10 1 3 42 6 1

D. Bertoldi et al. / Journal of Food Composition and Analysis 24 (2011) 376–385

378

Table 3 ICP-MS analytical conditions for each laboratory involved. Laboratory

FEM-IASMAa

UEAb

Seibersdorf Labor GmbHc

FERAd

ICP-MS model

7500ce, Agilent Technologies, Tokyo, Japan

7500ce, Agilent Technologies, Tokyo, Japan

Nebuliser Spray chamber Cones Collision/reaction gas IS added pre-measurement IS added on-line Dilution of samples

Concentric micromist Scott-double pass Ni H2, He – Sc, Rh, Tb –

Concentric micromist Scott-double pass Ni H2, He Rh, Pt – 10 times

Sciex Elan 6100, Perkin Elmer, Waltham, USA Cross-flow

Sciex Elan 6000, Perkin Elmer, Waltham, USA Concentric Cyclonic Pt – Rh, In – 100 times

a b c d

Ni – Rh – –

FEM-IASMA = Fondazione Edmund Mach – Istituto Agrario di San Michele all’Adige, via E. Mach 1, 38010 San Michele all’Adige, Italy. UEA = Centre for Forensic Provenancing, School of Chemistry, University of East Anglia, Norwich, NR4 7TJ, United Kingdom. Seibersdorf Labor GmbH = Seibersdorf Labor GmbH, A-2444 Seibersdorf, Austria. FERA = The Food and Environment Research Agency, Sand Hutton, York YO41 1LZ, United Kingdom; IS = internal standard.

different countries reflects differences in drinking cultures, the mineral water markets, and groundwater reserves. 2.2. Analysis and elaboration The vials and the bottlenecks were cleaned beforehand with 5% HNO3 and rinsed with MilliQ1 water. An aliquot (50 mL) of sample was transferred into a polypropylene vial, and degassed and acidified by adding 0.5 mL of HNO3 (Superpure 69.5%, Carlo Erba, Milano, Italia). 25 elements (Al, B, Ba, Cd, Ce, Co, Cs, Cu, La, Li, Lu, Mn, Mo, Nd, Ni, Pb, Rb, Se, Sm, Sr, Tl, U, V, Yb, Zn) were quantified by quadrupole Inductively Coupled Plasma Mass Spectrometry (ICPMS) by four laboratories participating in the TRACE project (Table 3). All the operations were carried out at the maximum level of cleaning achievable in labs not equipped with a clean room. For quality control purposes, each analytical sample sequence included analysis of 3 blank samples (MilliQ1 water), one spiked procedural blank (containing known amounts of the analytes), and one standard reference material (NIST 1640 ‘Natural water’; National Institute of Standard and Technologies, Gaithersburg, MD, USA), thus ensuring the accuracy of the analytical technique and the comparability of the results of the different laboratories. Only analytical batches with recoveries of NIST certified elements between 80 and 120% for at least 75% of the elements were accepted; however, recoveries between 60 and 140% were required for all the certified elements. The elemental composition of the samples was corrected for the biases according to NIST recovery for each element and lab (Table 4). The content of the non-certified elements was recalculated on the basis of the mean value of the 4

labs. Failure to meet the accuracy criteria in even in one of the four labs would have meant that it was not possible to give data regarding legally important elements (e.g. As, Sb). The instruments were calibrated daily against external standard solutions and a regression fit value of more than 0.995 was required for each element. Preparation and analysis of both samples and blanks were carried out in triplicate. The detection limit (DL) of each element was calculated as 3 times the standard deviation of the signal of the blank sample analysed 10 times (IUPAC, 1997). For each element, we chose as DL the highest DL among those determined by the four laboratories involved in the project (Table 5). The main cations Na, K, Ca and Mg (DIN EN ISO 14911, E34) and the anions sulphate, nitrate, and chloride (DIN EN ISO 10304-1, D19) were analysed by ion chromatography. The analysis by ion chromatography is done in several dilution factors due to the concentration of each ion. The plausibility of the results was tested against ion balance, which was better than 2% for 97% of samples. Specific electric conductivity (EN 27888:1993-11) and pH (DIN 38404-5:1984-01, C5) were analysed straight after opening the bottle and corrected to 25 8C and 20 8C, respectively. Total alkalinity was analysed according to the method DIN 38409-7:2004-03. Ammonia and nitrite contents were analysed photometrically. Fluoride was measured using ion sensitive electrode (IES, EN ISO 10304-1:1995-04). These parameters were analysed by the Hydroisotop GmbH laboratory (Schweitenkirchen, Germany). Total hardness, in French degrees (8f), was calculated by summing the concentrations (mg/L) of Ca and Mg, both expressed as CaCO3 (Spellman, 2004).

Table 4 Results (mg/L) for NIST 1640 measurements ( 95% confidence interval) in the laboratories. Element

NIST1640

FEM-IASMA

UEA

Seibersdorf Labor GmbH

FERA

Li B Al V Mn Co Ni Cu Zn Se Rb Sr Mo Cd Ba Pb

50.7  1.4 301.1  6.1 52.0  1.5 12.99  0.37 121.5  1.1 20.28  0.31 27.4  0.8 85.2  1.2 53.2  1.1 21.96  0.51 2.00  0.02 124.2  0.7 46.75  0.26 22.79  0.96 148.0  2.2 27.89  0.14

48.1  3.4 294.0  19.8 49.5  2.1 12.73  0.24 116.0  4.4 19.90  0.45 25.7  1.0 80.4  3.3 51.2  3.5 22.36  0.87 1.46  0.60 124.4  2.2 48.84  2.29 22.43  0.91 147.8  3.0 25.73  1.11

48.0  3.7 282.0  20.7 48.3  3.1 12.92  1.39 122.9  5.6 20.38  0.96 26.1  2.2 86.2  5.6 56.4  2.2 23.55  1.90 2.00  0.04 132.5  1.3 46.64  1.09 24.46  0.34 160.1  3.9 30.42  0.80

55.9  2.1 330.0  13.6 52.9  1.5 13.51  0.42 125.4  1.8 20.95  0.22 28.0  0.2 88.5  1.3 61.3  3.8 23.53  0.78 2.03  0.03 124.4  1.9 47.50  0.75 23.12  0.23 150.0  1.0 29.46  0.65

51.9  0.7 312.4  13.0 52.4  1.5 13.22  0.83 124.0  4.8 20.08  0.46 24.0  0.3 89.2  2.8 54.8  1.8 24.53  2.79 1.88  0.04 120.4  6.1 46.58  1.22 22.85  0.80 147.4  3.6 27.48  0.62

NIST 1640 = National Institute of Standard and Technologies Standard Reference Material 1640 ‘‘Trace Elements in Natural Water’’. Please, refer to Table 3 for the definition of laboratory names.

D. Bertoldi et al. / Journal of Food Composition and Analysis 24 (2011) 376–385

379

Table 5 Statistical distribution of the parameters analysed in 571 European mineral waters. Parameters Conductivity (mS/cm) pH Total alkalinity (mequiv./L) Macro-elements (mg/L) Ca Cl F K Mg Na NH4+ NO3 NO2 SO42 Micro-and trace elements (mg/L) Al B Ba Cd Ce Co Cs Cu La Li Mn Mo Nd Ni Pb Rb Se Sm Sr Tl U V Yb Zn Total mineral content (mg/L)

DL

%>DL

Minimum

10th percentile

25th percentile

Median

75th percentile

90th percentile

3

99.1

19 3.66
140 4.93 41

329 5.23 151

585 6.27 274

1050 7.43 393

0.5 0.5 0.1 0.5 0.5 0.5 0.1 0.5 0.01 0.5

99.6 99.5 69.5 98.1 97.5 99.8 11.0 67.8 7.9 98.9

9.5 2.1

33.0 5.2

67.2 13.5 0.2 2.1 16.4 14.5

1.5 7 0.37 0.05 0.05 0.11 0.05 1.6 0.05 0.06 0.6 0.34 0.05 1.9 0.05 0.05 5.5 0.05 1.7 0.05 0.05 1 0.05 4.7

34.3 80.6 97.9 2.5 2.3 43.1 50.1 6.5 1.8 98.4 34.3 54.5 2.5 11.7 10.7 97.9 3.5 0.7 99.3 11.9 72.3 12.1 0.7 18.6

271

484


95

Maximum

Spread

2045 7.72 858

6540 9.28 4300

344 3 1433

117 36.2 0.5 5.6 33.4 47.6

221 94.3 1.1 17.0 60.5 161 0.1 9.1

715 988 8.8 225 350 1660 3.2 52.0 2.04 1820

1430 1976 88 450 700 3320 32 104 204 3640

2.9 145 75.6

9.6 494 160

854

1900

420 5313 1873 0.69 0.70 3.79 519 112 1.15 5456 1938 37.8 0.82 30.3 0.44 1010 49.3 0.13 22,763 1.32 72.2 66 0.31 260 6642

280 782 5131 13 14 34 10,380 68 23 87,524 3429 113 16 16 8.7 20,195 9.0 2.6 13,509 26 1443 66 6.2 55 604

DL = detection limit; %>DL = percentage of samples with detectable content; spread = maximum/minimum values ratio.

Distribution statistics for pH and specific electric conductivity values and for the content of elements and the relevant DLs are shown in Table 5. The total mineral content, calculated by summing the main ions, and spread (maximum/minimum ratio) are also given. Applying the same conservative approach used by Misund et al. (1999), the spread of elements not analytically detectable in all samples was calculated by considering the DL. The remarks per Country reported thereinafter allow us to better describe our dataset that, anyway, does not exhaust the mineral waters available on the market in each Country.

mountainous areas in Portugal, Scandinavia and Scotland. Mountainous areas of France and Germany provide mostly mineral waters with high conductivities (>1500 mS/cm). Considering each Country separately, the lowest median conductivity was for Norwegian waters (57 mS/cm) followed by the Portuguese samples (95 mS/cm). High conductivity median value (1412 mS/cm), as well as the maximum, was found in waters from France. Values higher than 3500 mS/cm were measured in waters from France (5 samples), Germany (3) and Spain (1). The pH values ranged from 3.66 to 9.28, detected respectively in a German and a Portuguese sample and pHs higher than 8.00 were found also in Italy (2 samples), Great Britain (1), Ireland (1), France (1) and Spain (1). Focusing on the countries with a high number of samples, the pH median value was low (5.18) in Germany and equal to 7.50 in Great Britain. Of course, these values are affected by the possible presence of CO2 artificially added to fulfil the customer taste.

3.1. Conductivity and pH

3.2. Macro-elements

Conductivity values ranged from 19 to 6540 mS/cm. The widespread of conductivity values, related to total dissolved solid, presents the high variability of mineral waters. The majority of samples with low conductivity (<150 mS/cm) which was analysed in 10% of all samples is produced from Alpine areas in Italy,

The chemical composition of the samples (Table 5) is discussed here, both in comparison with previous data referring to European mineral and bottled waters and also considering the limits and recommendations set for adults by the National Academy of Science (NAS, 2004) as regards nutritional considerations (Table 6).

Box plots and distribution statistics were obtained using Statistica v 8.0 (StatSoft Inc., Tulsa, OK, USA) whereas the Piper diagram was prepared with RockWorks/14 rev. 2009.2.5 (RockWare Inc., Golden, CO, USA). 3. Results and discussion

D. Bertoldi et al. / Journal of Food Composition and Analysis 24 (2011) 376–385

380 Table 6 Nutritional guidelines. Element B Ca Cl Cu F K Mg Mn Mo Na Ni Se V Zn

AI (mg/day)

RDA (mg/day)

EAR (mg/day)

1000–1300 1800–2300 0.7 3–4 4700 310–420

255–350

0.045

0.034

0.055

0.045

8–11

6.8–9.4

1.8–2.3 1200–1500

UL (mg/day) 20 2500 3600 10 10

11 2 2300 1 0.4 2 40

% median to AI or RDAa 5.2–6.7 0.59–0.75 5–7 0.045 3.9–5.3 < 0.025–0.031 0.84 1–1.2 < 10 < 0.04–0.06

% median to ULa 0.17 2.69 0.38 < 0.016 2.00

< 0.01 0.02 0.63 < 0.19 < 1.36 < 0.05 < 0.012

% maximum o AI or RDAa 55–72 43–55 220–293 4.8 83–113 84–108 84 111–138 90 2.4–3.2

% maximum to ULa 27 29 27 1.1 88

18 1.9 72 3.1 12 3.3 0.6

AI = Adequate Intake set for adults by the National Academy of Science (2004). RDA = Recommended Dietary Allowance set for adults by the National Academy of Science (2004). EAR = Estimated Average Requirement set for adults by the National Academy of Science (2004). UL = Tolerable Upper Intake Level set for adults by the National Academy of Science (2004). Please refer to ‘‘References’’ section for full bibliographic citation. a The percentages of AI, RDA, EAR and UL taken up in drinking 1 L of mineral water per day are given, calculated by referring to the median and maximum values for each element. Where median was below detection limit, the latter was used. Please refer to Table 5 for detection limit values.

The overall median of the total mineral content was 484 mg/L. In relation to specific countries, the highest value was found in the unique analysed sample from Bosnia and Herzegovina (3067 mg/ L). High median values were found in Swiss (1233 mg/L), Finnish (1176 mg/L), French (1149 mg/L) and Czech (1019 mg/L) waters whereas the lowest were in the Portuguese (61 mg/L) and Norwegian (43 mg/L) samples. Calcium and magnesium levels ranged between trace concentrations lower than DL and 715 mg/L and 350 mg/L respectively. Roughly a quarter of the samples could be considered as soft water (<12.5 8f), whereas about half could be considered hard (>25 8f; Mutschmann and Stimmelmayr, 2007). Maximum calcium levels of about 500 mg/L are generally reported in the literature (Misund et al., 1999; Lau and Luk, 2002; Krachler and Shotyk, 2009). Naddeo et al. (2008) showed for Italian waters a maximum level of 864 mg/L, very close to the maximum level found in this study in an Italian sample. An intake of 1 L of these waters gives about half the Ca Adequate Intake (AI) and roughly one third of the tolerable upper intake level (UL) set by NAS (2004) for adults. Maximum magnesium levels lower than 215 mg/L were reported by Misund et al. (1999), Lau and Luk (2002), Rosborg et al. (2005) and Krachler and Shotyk (2009). On the other hand, our maximum value (determined in an Italian water) was close to that noted by Naddeo et al. (2008; 328 mg/L) and an intake of 1 L of this water could represent a very large part of the Mg recommended dietary allowance (RDA) values set for adults according to gender and age. The chloride concentration range was smaller than previously reported by Misund et al. (1999; 0.46–3187 mg/L) and Naddeo et al. (2008; 0.15–8056 mg/L). The maximum fluoride content was in agreement with the maximum value reported by Naddeo et al. (2008; 8.40 mg/L), but higher than the maximum values – ranging between 2.25 and 3.05 mg/L – shown in other studies (Misund et al., 1999; Lau and Luk, 2002; Rosborg et al., 2005). An intake of 1 L of water with the maximum content observed by us for chloride and fluoride could respectively cover a very large part or significantly exceed the AI set by NAS for adults according to gender and age. As regards potassium, 5.8% and 3.2% of the samples exceeded the concentration range reported by Misund et al. (1999) and Lau and Luk (2002) respectively, but were however lower than the maximum values registered by Rosborg et al. (2005; 268 mg/L) or Naddeo et al. (2008; 300 mg/L). For a healthy adult a minimum requirement of 2000 mg/day is recognised by the USA National

Research Council (NRC, 1989) and the AI is 4700 mg/L (NAS, 2004), so the contribution of water to K intake can be considered negligible. Sodium concentration was higher than 500 mg/L in 12 samples: this means that a consumption of 1–2 L/day of these waters would contribute significantly to reaching the sodium AI set for adults (NAS, 2004) and the recommended maximum daily intake of Na (2.4 g) suggested by the NRC (1989) or the UL (2.3 g/day) set by NAS (2004). It is well known that a high intake of Na is detrimental for human health, increasing the incidence of hypertension (Du et al., 2002). Ammonium, that can have both geogenic and anthropogenic origin, was detectable in 11% of samples with a maximum level lower than the maximum value reported by Naddeo et al. (2008; 4.5 mg/L). The maximum nitrate value was similar to that reported by Naddeo et al. (2008) and Lau and Luk (2002), 47.49 mg/L and 38.1 mg/L respectively. On the other hand, it was decidedly higher than the maximum levels reported by Misund et al. (1999) and Rosborg et al. (2005), 18.8 mg/L and 5.5 mg/L respectively. About 30% of samples did not have detectable nitrate levels (<0.5 mg/L). According to Misund et al. (1999) the waters with high nitrates (>2.62 mg/L), probably due to crop fertilisation, come from the south of Europe (France, Italy and Greece). In the present survey, 213 samples (37% of our dataset) had nitrate content higher than 2.6 mg/L and the waters with nitrates higher than 20 mg/L samples were found in Italy (N = 13; 7% of the analysed national samples), Great Britain (6; 18%), Spain (1; 2.4%) and Germany (1; 0.5%). Nitrite content was detectable (>0.01 mg/L) in less than 8% of samples, with a maximum level lower than that reported by Naddeo et al. (2008; 3 mg/L). The sulphate concentration was in agreement with the data reported by Naddeo et al. (2008) and slightly higher than the data shown by Misund et al. (1999) and Lau and Luk (2002), whereas, in their work on Swedish waters, Rosborg et al. (2005) found a lower maximum content (174 mg/L). 3.3. Micro- and trace elements Aluminium concentrations observed in this study (overall mean = 5.78 mg/L) were lower than the values reported by Naddeo et al. (2008) for Italian waters (mean = 254 mg/L). Actually, the second highest concentration observed by us was in an Italian water (274 mg/L), albeit ten times lower than the maximum reported by the aforementioned authors. 98% of Italian waters

D. Bertoldi et al. / Journal of Food Composition and Analysis 24 (2011) 376–385

381

et al., 2007) and from trace to 9800 mg/L for Italian waters (Naddeo et al., 2008). Intermediate levels are reported by Misund et al. (1999) for European waters (0.1–200 mg/L), and by Krachler and Shotyk (2009) for samples from 28 countries worldwide (0.025– 310 mg/L). Eleven of our samples had higher levels than the maximum level found by Misund et al. (1999) and intake (1 L) could provide a large part of AI (1.8 and 2.3 mg/day for adult females and males, respectively) but not more than 18% of UL. The median molybdenum value was comparable with that generally reported in the literature (Misund et al., 1999; Rosborg et al., 2005; Krachler and Shotyk, 2009) whereas the maximum (37.8 mg/L in an Italian sample) was higher than the maximum set by Misund et al. (1999) for European waters (14.7 mg/L). Nickel was above DL in less than 12% of samples, reaching the maximum in a German sample (31 mg/L). This value, although being two to four times the values given respectively by Rosborg et al. (2005; 7.52 mg/L) and Misund et al. (1999; 14.4 mg/L), covers only a limited percentage of the UL. Lead was always far below the EU limit for mineral water, reaching the maximum, 0.44 mg/L, in an Austrian sample. Other maximum levels for European waters ranged between 0.51 mg/L and 2.34 mg/L (Misund et al., 1999; Rosborg et al., 2005; Karamanis et al., 2007; Krachler and Shotyk, 2009). Naddeo et al. (2008) reported a very high mean concentration for Italian waters (350 mg/L), but this is the result of including 2 curative waters sold in chemists with content of up to 3500 mg/L in the sampling. The median rubidium value was in line with previous data (Misund et al., 1999; Rosborg et al., 2005). We found a maximum of 1010 mg/L in a French sample, about eight times higher than the level found by Misund et al. (1999; 108 mg/L) for European waters but only slightly higher than the level found by Krachler and Shotyk (2009; 840 mg/L) for worldwide samples. Content higher than 100 mg/L was found in 5.6% of total samples: i.e. Italy (10; 5.4%), France (9 samples; 28% of analysed national samples), Germany (7; 3.8%), Czech Republic (2; 50%), Austria (1; 5.9%), Slovenia (1; 33%) and Spain (1; 2.4%). We detected a measurable concentration of selenium in only 3.5% of samples. The maximum level registered was 49.3 mg/L, as measured by Misund et al. (1999). The median strontium level was similar to that reported in the literature for European waters (Misund et al., 1999, 448 mg/L). Levels higher than 7000 mg/L were found only in Germany (17 samples; 9.2% of analysed national samples, up to 22,763 mg/L), Italy (2; 1.1%), France (2; 6.3%) and Switzerland (1; 16.7%). The highest Sr value found by Lau and Luk (2002) in a survey of over 60 European and Asian waters was for a French sample (4940 mg/L). Thallium was found in detectable amounts in about 12% of samples, with a maximum content of 1.32 mg/L, only slightly higher than the maximum found by Misund et al. (1999; 0.9 mg/L).

showed concentration levels in agreement with the values given by Plessi and Monzani (1995) for 43 Italian mineral waters (from <1.9 to 24.9 mg/L). Our results covered the ranges given for waters from Sweden (Rosborg et al., 2005) and from European and extraEuropean countries (Misund et al., 1999; Krachler and Shotyk, 2009) with only 4 samples higher than the maximum levels reported by these authors (147 mg/L). Boron amounts were consistent with the range reported by Rosborg et al. (2005), at all events lower than the maximum values reported by Misund et al. (1999) and Naddeo et al. (2008). An intake of 1 L of the water with the highest B content found by us would represent, at most, about one quarter of UL (Table 5). Barium content, both as regards median and maximum values, was in agreement with that reported by Misund et al. (1999). Cadmium was detectable in 2.5% of samples, with a maximum level of 0.69 mg/L in an Italian sample, a level which is however far below the legal limit set by the European Community (3 mg/L). Naddeo et al. (2008) reported concentrations of up to 2 mg/L for Italian samples. Other authors found maximum levels under 0.21 mg/L (Misund et al., 1999; Rosborg et al., 2005; Karamanis et al., 2007). The data found here for cobalt were in agreement with the range indicated by Misund et al. (1999) and by Krachler and Shotyk (2009). Rosborg et al. (2005) and Soupioni et al. (2006) reported values which were decidedly lower for Swedish (maximum = 0.6 mg/L) and Greek (maximum = 0.012 mg/L) bottled waters respectively. The median caesium value was higher than reported in the literature (0.02–0.029 mg/L) as was the maximum level (32.5– 36.4 mg/L; Misund et al., 1999; Rosborg et al., 2005). About 3% of our samples exceeded the maximum values mentioned. Of the 57 samples above the 90th percentile (6.59 mg/L), 27 were from Germany (about 15% of analysed national samples) and 10 from France (31%), whereas the others were distributed over 8 different countries. Copper was detectable in only 6.5% of samples, with a maximum of 112 mg/L in one sample from Great Britain, well under the legal limits set by all countries (Table 1). Excluding this sample, we observed values which were always lower than 30.7 mg/L, the maximum indicated for European waters by Misund et al. (1999), while only 3 samples (from Poland, Portugal and Spain) were close to the maximum reported by Rosborg et al. (2005; 22.3 mg/L). Karamanis et al. (2007) gave a decidedly lower maximum level of 0.96 mg/L for 16 Greek waters. Lithium concentrations agreed with the results in other European (Misund et al., 1999) and worldwide studies (Krachler and Shotyk, 2009). The highest content was in a French sample (5456 mg/L), confirming the previous data of Krachler and Shotyk (2009; maximum = 5460 mg/L). As regards manganese, wide concentration ranges are given in the literature: e.g. 0.61–2.57 mg/L in 16 Greek waters (Karamanis

Table 7 Number of waters per country that could be labelled with the indications provided for by European Commission Directive 2009/54/EC. Indications

Country codea

Contains sodium Suitable for a low-sodium diet Contains bicarbonateb Contains sulphate Contains chloride Contains fluoride Contains magnesium Contains calcium

4 7

A

a b

5 5

4 4

B

BiH

CH

CZ

D

3

1 1

23 88

2

24 45 17 22 35 50

1 1 1 1

2 3

1

2 2 4

1 2 1

DK

E

F

FIN

2

1 25

8 14

2 2

1 3 2 2 1

13 9 4 12 11 14

2 1 2

Please refer to Table 2 for Country codes. Assuming total alkalinity as an assessment of the bicarbonate content.

3

GB 23

1 1

GR 2

H

HR

I

1

2

3 135

3 1

1

3 3 3

IRL

15 9 1 12 11 15

1

N 3

NL 2

P

PL

7

1 11

RUS

SLO

1

1 1

6

1

5 7

1 1 1

1 2

UA

[()TD$FIG]

382

D. Bertoldi et al. / Journal of Food Composition and Analysis 24 (2011) 376–385

Fig. 1. Piper diagram showing the chemical composition of the 571 mineral waters.

The elements of the rare earth group (La, Ce, Nd, Sm, Yb, Lu) were detectable in less than 3% of samples and Lu was never found in amounts above the DL (0.49 mg/L). These elements are only rarely reported in the literature, nevertheless, maximum data were often similar or slightly lower than those reported by Misund et al. (1999), whereas Yb was higher. High uranium levels (>15 mg/L) were found in some waters from Spain (N = 4), France (2) and Germany (1) confirming, at least for Spain, the data of Krachler and Shotyk (2009). Other studies on bottled water report maximum U levels ranging from 0.38 mg/L in Greece (Soupioni et al., 2006) to 9.45 mg/L (Misund et al., 1999) and 27.5 mg/L (Krachler and Shotyk, 2009) in European and worldwide waters respectively, and up to 72 mg/L in a Swedish sample (Rosborg et al., 2005). Vanadium was present above the DL in about 12% of samples and our maximum agreed with the levels found by other authors (Misund et al., 1999; Rosborg et al., 2005; Krachler and Shotyk, 2009). For zinc, the median and maximum values observed by us corresponded with the literature (Misund et al., 1999; Rosborg et al., 2005) with the exception of Naddeo et al. (2008). These authors reported a mean value of 45 mg/L for Italian waters, decidedly higher than in our 186 Italian samples (<4.7 mg/L) whereas the maximum levels (180 mg/L and 153 mg/L respectively) were similar. Both Karamanis et al. (2007) and Soupioni et al. (2006) found low content (max. 2.35 and 6.7 mg/L respectively) in Greek waters, similar to our results (below the DL) for the same Country. The concentration spread (calculated as maximum/minimum value) of the elements varied between 101 and 105, being larger

for Li, Rb and Sr and smaller for Cd, Co, Ni, Pb, Se and rare earths (Table 5). 3.4. Substances with nutritional physiological characteristics As regard some substances having nutritional physiological characteristics, the European legislation allows the use of specific terms in label (e.g. ‘containing calcium’) if the concentrations of these substances meet defined criteria (European Commission Directive 09/54/EC).

Table 8 Number and percentage distribution of main types of mineral waters based on their predominant hydrochemical composition. Natural mineral water type

No. of samples

%

Ca–Mg–HCO3 Ca–HCO3 Ca–HCO3–SO4 Ca–HCO3–SO4 Ca–Mg–SO4–HCO3 Ca–Na–HCO3 Na–Ca–HCO3 Na–HCO3 Ca–SO4–HCO3 Ca–Mg–Na–HCO3 Ca–Na–HCO3–Cl Na–HCO3–Cl Others

127 104 29 27 23 20 17 16 14 12 11 10 161

22.2 18.2 5.1 4.7 4.0 3.5 3.0 2.8 2.5 2.1 1.9 1.8 28.2

HCO3 = assessed as equal to total alkalinity.

D. Bertoldi et al. / Journal of Food Composition and Analysis 24 (2011) 376–385

According to the cited Directive, 58.1% of samples could be defined as ‘suitable for a low-sodium diet’, containing less than 20 mg/L, whereas 8.1% of samples could be defined as ‘containing sodium’ having more than 200 mg/L. 13.7% of mineral waters could be labelled as ‘containing magnesium’ (>50 mg/L), 10.2% as ‘containing fluoride’ (F > 1 mg/L), 4.9% as ‘containing chloride’ [()TD$FIG](Cl > 200 mg/L), 13.5% as ‘containing sulphate’ (SO42 > 200 mg/

383

L) and 17.5% as ‘containing calcium’ (>150 mg/L). Assuming the measured total alkalinity as an acceptable assessment of the bicarbonate content, approximately the 13% of the waters could be defined as ‘containing bicarbonate’ (HCO3 > 600 mg/L). The number of waters that could be labelled according to these criteria in each Country is shown in Table 7 and the main characteristics are summarised using a classical Piper diagram (Piper, 1944;

Fig. 2. Content distribution per country of the elements regulated by the European Commission Directive 2003/40/EC. Black lines indicate the legal limits (please refer to Table 2 for the number of sample in each Country).

384

D. Bertoldi et al. / Journal of Food Composition and Analysis 24 (2011) 376–385

Fig. 1). Van der Aa (2003), classifying 291 worldwide mineral waters, found a similar distribution, with 55% of waters ‘suitable for a low-sodium diet’, 15% of samples characterised as ‘containing sodium’, 16% ‘containing magnesium’, 6% ‘containing fluoride’, 8% ‘containing chloride’, 20% of samples ‘containing bicarbonate’, 12% ‘containing sulphate’ and 16% ‘containing calcium’. 3.5. Classification of natural mineral waters The Piper diagram of Fig. 1 illustrates the high variability of hydrochemical composition in mineral waters which can be found on the European mineral water market. The majority of the samples is waters with high content of calcium, magnesium and bicarbonate (Ca–Mg–HCO3, Table 8), but bottled mineral waters of Europe are provided in any sort of hydrochemical composition. According to traditional characterisation which is commonly used for medicinal and mineral waters, the hydrochemical composition of the water type is expressed with the main elements of more than 20 mequiv.%. The sample set presents 76 different water types whereas not only Ca, Mg, Na, HCO3 , SO42 and Cl are characterising but also K and NO3 . Additional differences and variabilities of the mineral waters are founded by the composition of minor and trace elements which allows the consumer to choice what ever he likes, provided the information is available. 3.6. Relevant legislation Table 1 shows the limits set in European countries and the USA, according to available information. EC limits were exceeded in only 7 samples for Se (4 French, 1 Norwegian, 1 German and 1 British), 4 for nitrite (2 Polish, 1 Italian and 1 British), 3 for Mn (2 French and 1 German), 2 for Ni (Austrian and German), 1 for Ba (French), 1 for fluoride (French) and 1 for nitrate (British). All the samples complied with the maximum tolerable level established for Cd, Cu and Pb (Fig. 2). Overall, 16 samples did not comply with EC limits. Some European countries have established further or more restrictive limits for specific elements. To date, no legal B limit has been established by the EU but 21 samples from several countries (6 French, 5 German, 3 Austrian, 3 Italian, 1 British, 1 Czech, 1 Slovenian and 1 Spanish) could be considered anomalous taking into account the limits set by Spain whereas 1 sample from France could be anomalous in the light of Italian limits (Table 1). Twentyseven samples had a fluoride content higher than 1.5 mg/L, considered problematical as it can induce dental fluorosis (WHO, 2008) and established as a limit for fluoride declaration on label by Spanish and Austrian legislations. 479 samples fulfilled the fluoride German requirement (<0.7 mg/L) for labelling as ‘suitable for preparation of infant food’. Thirteen samples had nitrate levels of over 25 mg/L, the limit set by Austria, and 1 sample exceeded the Italian limit (45 mg/L). As regards nitrite, 14 samples exceeded the more restrictive limit (0.02 mg/L) adopted in Italy. The European Community has not set a legal limit for Tl, but for toxicological reasons this element should be monitored. All the samples complied with the Environmental Protection Agency (USEPA, 2009) guidelines, having Tl concentrations lower than 2 mg/L. Similarly, there is no EU limit for U, however 7 samples (4 Spanish, 2 French, 1 German) did not comply with the guidelines issued for drinking waters by World Health Organisation (WHO, 2008) and 3 of them (2 Spanish and 1 French) did not comply with the US-EPA limit (2009). In total, 52 out of the 571 samples could exceed at least one parameter for the most restrictive limits (Table 1) set in Europe by the different countries (1 mg/L for Ba, B and Cu; 0.003 mg/L for Cd; 5 mg/L for F; 0.5 mg/L for Mn; 0.020 mg/L for Ni; 25 mg/L for NO3 ; 0.02 mg/L for NO2 ; 0.010 mg/L for Pb; 0.010 mg/L for Se).

4. Conclusion This paper provides one of the most extensive and up-to-date surveys of the mineral composition of bottled mineral waters available on the European market. In addition to the most frequently analysed parameters, e.g. macro elements, it also provides a picture of micro and trace element concentration, some of these (e.g. Mo, U, Tl) having toxicological implications. Our survey showed that 2.8% of samples did not comply with European Community limits for at least one parameter (Se, NO2 , Mn, Ni, Ba, F and NO3 ) and about 9% of samples had boron, nitrate or nitrite levels above the legal limit existing in individual European countries. Disclaimer The information reflects the author’s views; the European Commission is not liable for any use of the information contained therein. Acknowledgments This work is funded by the European Commission, under the FP6 Food Quality and Safety Priority, within the framework of the Integrated Project TRACE – 006942 – entitled ‘‘Tracing Food Commodities in Europe’’. This project is carried out by a consortium coordinated by FERA, The Food and Environment Research Agency (UK) and includes 54 partners. References Arreˆte´ du 14 mars 2007 relatif aux crite`res de qualite´ des eaux conditionne´es, aux traitements et mentions d’e´tiquetage particuliers des eaux mine´rales naturelles et de source conditionne´es ainsi que l’eau mine´rale naturelle distribue´e en buvette publique. Journal Officiel de la Re´publique franc¸aise, n881 Texte 46 du 05/04/2007 (in French). ¨ zdilek, H.G., 2008. An assessment of the ¨ ., Cam, S., O Baba, A., Erees, F.S., Hic¸so¨nmez, U quality of various bottled mineral water marketed in Turkey. Environmental Monitoring and Assessment 139, 277–285. Besluit van 6/11/2003, houdende wijziging van het Warenwetbesluit Verpakte waters en van het Warenwetbesluit bestuurrlijke boeten. Staatsblad, n8 473 van 25/11/2003, p. 1 (in Dutch). Commission Directive 1996/70/EC of 28 October 1996 amending Council Directive 80/777/EEC on the approximation of the laws of the Member States relating to the exploitation and marketing of natural mineral waters. Official Journal of the European Union L299 of 23/11/1996. Commission Directive 2003/40/EC of 16 May 2003 establishing the list, concentration limits and labelling requirements for the constituents of natural mineral waters and the conditions for using ozone-enriched air for the treatment of natural mineral waters and spring waters. Official Journal of the European Union L126/34 of 22/05/2003. Commission Directive 2009/54/EC of 18 June 2009 on the exploitation and marketing of natural mineral waters. Official Journal of the European Union L164/45 of 26/06/2009. Council Directive 1980/777/EC of 15 July 1980 on the approximation of the laws of the Member States relating to the exploitation and marketing of natural mineral waters. Official Journal of the European Union L229 of 30/08/1980. Dabeka, R.W., Conacher, H.B.S., Lawrence, J.F., Newsome, W.H., McKenzie, A., Wagner, H.P., Chadha, R.K.H., Pepper, K., 2002. Survey of bottled drinking waters sold in Canada for chlorate, bromide, bromate, lead, cadmium and other trace elements. Food Additives and Contaminants 19, 721–732. Decreto 29 dicembre 2003 del Ministero della salute. Attuazione della direttiva n8 2003/40/CE della Commissione nella parte relative ai criteri di valutazione delle caratteristiche delle acqua minerali naturali di cui al decreto ministeriale 12 novembre 1992, n8 542, e successive modificazioni, nonche´ alle condizioni di utilizzazione dei trattamenti delle acque minerali naturali e delle acque di sorgente. Gazzetta Ufficiale Italiana, n8 302 del 31 Dicembre 2003 (in Italian). DIN, 2004. Bestimmung der Sa¨uren – und Basenkapazitat. DIN 38409-7. Deutsches Institut fu¨r Normung, Berlin, Germany (in German). DIN, 1984. German Standard Methods for Examination of Water, Waste Water and Sludge; Physical and Physical–Chemical Parameters (Group C); Determination of pH-value (C5). DIN 38404-5. Deutsches Institut fu¨r Normung, Berlin, Germany. Du, C.G., Ribstein, J., Mimran, A., 2002. Dietary sodium and target organ damage in essential hypertension. American Journal of Hypertension 15, 222–229. EN, 1993. Water Quality. Method for the Determination of Electrical Conductivity. EN 27888. European Committee for standardization, Brussels, Belgium.

D. Bertoldi et al. / Journal of Food Composition and Analysis 24 (2011) 376–385 EN ISO, 1995. Water Quality. Determination of Dissolved Anions by Liquid Chromatography of Ions. Determination of Fluoride, Chloride, Nitrite, Orthophosphate, Bromide, Nitrate and Sulfate Ions. Method for Water with Low Contamination. EN ISO 10304-1. European Committee for standardization/International Organization for Standardization, Brussels, Belgium/Geneva, Switzerland. EN ISO, 1999. Water Quality. Determination of Dissolved Li+, Na+, NH4+, K+, Mn2+, Ca2+, Mg+2, Sr+2 and Ba2+ Using Ion Chromatography. Method for Water and Waste Water. EN ISO 14911-1. European Committee for Standardization/International Organization for Standardization, Brussels, Belgium/Geneva, Switzerland. IUPAC, 1997. Compendium of Chemical Terminology. The Gold Book, 2nd ed. Blackwell Science, Oxford, UK. , http://old.iupac.org/publications/books/author/ mcnaught.html. Karamanis, D., Stamoulis, K., Ioannides, K.G., 2007. Natural radionuclides and heavy metals in bottled water in Greece. Desalinisation 213, 90–97. King, M., 2008. Bottled Water—Global Industry Guide, New Research Report on Companies and Markets. Retrieved July 7, 2008 from http://www.companiesandmarkets.com/Summary-Market-Report/bottled-water-global-industryguide-44377.asp. Krachler, M., Shotyk, W., 2009. Trace and ultratrace metals in bottled waters: survey of sources worldwide and composition with refillable metal bottles. Science of the Total Environment 407, 1089–1096. Lau, O.-W., Luk, S.-F., 2002. A survey on the composition of mineral water and identification of natural mineral water. International Journal of Food Science and Technology 37, 309–317. Mineral- und Tafelwasserverordnung vom 01 August 1984 (BGBl. I, S.1036) i.d.F. v. 01. Dezember 2006. Bundesgesetzblatt (BGBl), I. n8 56 vom 11. Dezember 2006, S.2762 ff. (in German). Mineralwasser- und and Quellwasserverordnung. Mineralwasser- und Quellwasserverordnung. Verordnung der Bundesministerin fu¨r Frauenangelegenheiten und Verbraucherschutz u¨ber natu¨rliche Mineralwa¨sser und Quellwa¨sser. Bundesgesetzblatt (BGB l), II n8 309/1999 idF: BGB l. II n8 500/2004 (in German). Misund, A., Frengstad, B., Siewers, U., Reimann, C., 1999. Variation of 66 elements in European bottled mineral waters. The Science of the Total Environment 243/ 244, 21–41. Mutschmann, J., Stimmelmayr, F., 2007. 4. Wasseraufbereitung. Taschenbuch der Wasserversorgung 14. Auflage 2007. Vieweg Teubner Verlag, Wiesbaden, Germany, p. 206 (in German). Naddeo, V., Zarra, T., Belgiorno, V., 2008. A comparative approach to the variation of natural elements in Italian bottled waters according to the national and international standard limits. Journal of Food Composition and Analysis 21, 505–514. National Academy of Science (NAS), 2004. Dietary Reference Intakes (DRIs): Recommended Intakes for Individuals. Food and Nutrition Board, Institute of Medicine, National Academy of Science Retrieved September 14, 2009 from http://fnic.nal.usda.gov. Nkono, N.A., Asubiojo, O.I., 1997. Trace elements in bottled and soft drinks in Nigeria – a preliminary study. The Science of the Total Environment 208, 161–163. National Research Council (NRC), 1989. Water and electrolytes. In: The National Academies Press (Eds.), Recommended Dietary Allowances. 10th ed. Subcommittee on the Tenth Edition of the Recommended Dietary Allowances, Food and Nutrition Board, Commission on Life Sciences, National Research Council, Washington, DC, pp. 247–262.

385

OJEC C41/01 of 14/02/2002, 2002a. Commission: List of Natural Mineral Waters Recognised by the Member States. . OJEC C143/03 of 15/06/2002, 2002b. List of Mineral Waters Recognised by the United Kingdom. . OJEC C165/11 of 11/07/2002, 2002c. Standing Committee of the EFTA States. List of Natural Mineral Waters Recognised by Iceland. . OJEU C59/06 of 09/03/2005. List of Natural Mineral Waters Recognised by Belgium: Denmark: Germany: Greece Spain: France, Italy: Netherlands: Austria, Portugal: Sweden and United-Kingdom. OJEU C7/08 of 12/01/2006. EFTA Standing Committee. List of Mineral Waters in Iceland and Norway Pursuant to Article 1 of Council Directive 80/777/EEC of 15 July 1980 on the Approximation of the Laws of the Member States Relating to the Exploitation and Marketing of Natural Mineral Waters. Pais, I., Jones, J.B., 1997. The Handbook of Trace Elements. St. Lucie Press, Boca Raton, New York. Piper, A.M., 1944. A graphic procedure in the geochemical interpretation of wateranalyses. Transaction of the American Geophysical Union 25, 914–923. Plessi, M., Monzani, A., 1995. Aluminium determination in bottled mineral waters by electrothermal atomic absorption spectrometry. Journal of Food Composition and Analysis 8, 21–26. Real decreto 1744/2003 de 19/12/2003, por el que se modifica el Real Decreto 1074/ 2002, de 18 de octubre, por el que se regula el proceso de elaboracio´n, circulacio´n y comercio de aguas de bebida envasadas. Boletı´n Oficial del Estado, n8312 de 30/12/2003 (in Spanish). RockWare Inc., 2009. RockWorks/14 rev 2009.2.5. RockWare Inc., Golden, CO, USA. Rosborg, I., Nihlgard, B., Gerhardsson, L., Gernersson, M.-L., Ohlin, R., Olsson, T., 2005. Concentration of inorganics elements in bottled waters on the Swedish market. Environmental Geochemistry and Health 27, 217–227. Shotyk, W., Krachler, M., 2007a. Contamination of bottled waters with antimony leaching from polyethylene terephthalate (PET) increases upon storage. Environmental Science & Technology 41 (5), 1560–1563. Shotyk, W., Krachler, M., 2007b. Lead in bottled waters: contamination from glass and comparison with pristine groundwater. Environmental Science & Technology 41 (10), 3508–3513. Soupioni, M.J., Symeopoulos, B.D., Papaefthymiou, H.V., 2006. Determination of trace elements in bottled water in Greece by instrumental and radiochemical neutron activation analyses. Journal of Radioanalytical and Nuclear Chemistry 268, 441–444. Spellman, F.R., 2004. Water softening. In: Mathematic Manual for Water and Wastewater Treatment Plants Operators, CRC Press, Boca Raton, Florida, USA, pp. 211–214 (Chapter 21). StatSoft Inc., 2008. STATISTICA (Data Analysis Software System), Version 8.0. StatSoft Inc., Tulsa, OK, USA. , In: www.statsoft.com. The Natural Mineral Water, Spring Water and Bottled Drinking Water (England). Regulations 2007. Statutory Instruments 2007 No. 2785. US-EPA, 2009. National Primary and Secondary Drinking Water Regulations. EPA 816-F-09-004. United States Environmental protection Agency. Retrieved September 14, 2009 from http://www.epa.gov/safewater/consumer/pdf/mcl.pdf. Van der Aa, M., 2003. Classification of mineral water types and comparison with drinking water standards. Environmental Geology 44, 554–563. WHO, 2008. Guidelines for Drinking-water Quality, vol. 1, Annex 4, 3rd ed. World Health Organization, Geneva, Switzerland, pp. 491–493.