" % 7 " / $ & . & / 5 0 ' 5 ) & SCIENCE
Heavy Metals in Bottled Natural Spring Water Michael J. Sullivan, PhD, CIH, REA Shannon Leavey
"CTUSBDU
New regulations regarding the presence of contaminants in bottled water went into effect in California in January 2009. These requirements include testing, reporting, and notification to regulate the presence of heavy metals in bottled natural spring water sold in California. In the study described in this article, six sources of bottled natural spring water were purchased and analyzed for silver, arsenic, barium, beryllium, cadmium, cobalt, chromium, copper, mercury, molybdenum, nickel, lead, antimony, selenium, thallium, vanadium, and zinc. All of these metals except beryllium, mercury, and thallium were detected in at least one of the bottled natural spring water sources. No concentrations were above either federal or California maximum contaminant levels but arsenic concentrations exceeded California public health goals in all six sources. Improving the California notification requirements for bottled water contaminants would result in a process more similar to the notification process for tap water and would result in better-informed consumers.
Introduction In 2008, 8.67 billion gallons of bottled water were sold in the United States resulting in an average annual per capita consumption of 28.5 gallons (Rodwin, 2009). The state of California has consistently been the top consumer of bottled water products (International Bottled Water Association, 2004), which translates into even higher per capita bottled water consumption rates. With bottled water supplying an increasing percentage of drinking water to California residents, the regulation of bottled water in California compared to tap water has been a concern and is discussed in a recent article (Sullivan, 2009). On January 1, 2009, new regulations went into effect in California regarding bottled water sold in the state (Sullivan, 2009).
8
7PMVNFr/VNCFS
These regulations were passed as Senate Bill 220 and codified in HSC Section 111070 (Bottled, Vended, Hauled, and Processed Water, 2008). These regulations require the testing for contaminants in bottled water, the reporting of the results of the testing to the state and, in some cases, the notification to consumers through labeling requirements on bottled water packaging. While these bottled water regulations are not as stringent as the regulation of tap water by the state of California (Sullivan, 2009), they do represent an increase in the oversight of bottled water as well as increasing the information available to consumers regarding the presence of chemicals in bottled water and may serve as a model for similar regulation in other states. Other researchers have published the concentrations of heavy metals in bottled water
(Ikem, Odueyungbo, Egiebor, & Nyavor, 2002; Olsen, 1999; Shotyk & Krachler, 2007), but these studies were limited in the number of metals evaluated and did not focus on natural spring water. One of the co-authors of this article recently reviewed these studies (Sullivan, 2009). In addition, others have reviewed bottled water regulations in California (Allen & Darby, 1994); however, the recent changes in California bottled water regulations need to be reviewed as they will likely be used as a model for other states. This article reports on the presence of 17 heavy metals in bottled water and how the results would be addressed under these new regulations. Six sources of water were tested. Each source was labeled as “natural spring water” on the bottled water packaging. The heavy metals included in this evaluation were silver, arsenic, barium, beryllium, cadmium, cobalt, chromium, copper, mercury, molybdenum, nickel, lead, antimony, selenium, thallium, vanadium, and zinc. The resultant reporting and consumer notification requirements associated with the heavy metal concentrations in the six bottled water sources are presented. Recommendations for improving the reporting and communication of bottled water contaminants are discussed.
Methods Sample Procurement A local grocery store served as the purchasing location for the bottled water. The store carries over 25 brands of bottled water including its own store brand. The label on each of these brands was reviewed to determine if the source of the water was a “natural spring.” A
total of six natural spring bottled water brands, or sources, were selected. The selection favored brand names over generic or store brands and included spring sources located in California as well as other states and other countries. The packaging information regarding the product code and date of bottling, if provided, were recorded prior to opening the bottles for sampling. A total volume of at least 3 L of water of each selected source was purchased. This provided enough volume for three 1-liter samples to be prepared for analysis. The following six bottled natural spring water sources were evaluated in this study: Ethos (Ethos), Identification Codes (IC)181854031413 and 18195403148; Icelandic Glacial (Icelandic), IC-L072680112-002, -011, -012 and -038; Fiji (Fiji), IC-PRD13FEB0803:03 L3; Evian (Evian), IC-PRD111407 15:40; Arrowhead (Arrowhead), IC-051108132WF038 and 00038B05/2010; and Crystal Geyser (Crystal), IC-60516080307. Sample Preparation A total of three samples of each of the six selected sources, for a total of 18 samples of water, were prepared for analysis. Two trip blank quality assurance/quality control samples provided by the laboratory and included in the shipping coolers were the only additional samples. The laboratory provided pre-acid preserved, 1-liter plastic sample bottles. Since some of the water was sold in 1 L bottles and some in 0.5 L bottles, a sufficient number of bottles of each brand was purchased to have a total volume of at least 3 L. Each 1 L sample was prepared as a mixture of the water from all the purchased bottles for that brand. For brands sold in 1 L bottles, three bottles were opened and approximately 300 to 400 mL were poured from the water bottles into each of the three sample bottles until each was full. For brands sold in 0.5 L bottles, six water bottles were opened and approximately 150 to 200 mL were poured from the water bottles into each of the three sample bottles until each was full. This process resulted in three identical samples of bottled natural spring water for each source. All sample bottles were given a random number that was recorded on a sample table along with the brand name, bottle size, packaging information, and date/time that the samples were collected. Random sample numbers
were recorded on a chain-of-custody form and the sample bottles were placed on ice in plastic coolers. Coolers were delivered to the laboratory within three hours of the samples being collected. Analytical Each water sample was analyzed for the 17 heavy metals listed above. The analyses were for “total metals” and consisted of U.S. Environmental Protection Agency (U.S. EPA) Methods 200.8 (U.S. EPA, 1994a) and 245.1 (U.S. EPA, 1994b). The method detection limits (MDL) for these 17 metals in these analyses are presented in Table 1 where nondetect or “ND” values are reported. Statistical Evaluation We calculated the arithmetic mean of the metal concentrations in each water bottle brand. We did not calculate statistical measures of variance due to the limited number of samples. When a sample result was ND, half of the reported MDL was used as a surrogate concentration for that sample. If all three samples for a water bottle brand were ND for a particular metal, then we also reported the mean as ND and we concluded that no analytical evidence existed for the presence of that metal in the bottled water. When at least one reported detect occurred of a metal in a brand of bottled water, then we assumed half the ND values to represent likely metal concentrations and we calculated the means without additional comment. When samples were removed for blank contamination and only one sample remained, we presented that sample as the average. Comparison Criteria The mean and detected concentrations of metals in the six bottled water brands were evaluated by comparison to three drinking water criteria. These criteria are the federal maximum contaminant level (MCL) (California Department of Public Health, 2009), state of California MCL (CA-MCL) (California Department of Public Health, 2009), and the nonenforceable state of California public health goals (PHGs) (Office of Environmental Health Hazard Assessment, 2009). These comparison criteria are presented in Table 1 for each metal as available. Both the federal and California criteria for copper are action levels and the criteria for
zinc are secondary MCLs (Secondary Water Standards, 2006). Bottled water concentrations are only required to be compared to MCLs. As stated in HSC Section 111080: “Bottled water and vended water shall meet all maximum contaminant levels set for public drinking water that the department determines are necessary or appropriate so that bottled water may present no adverse effect on public health.” PHGs are also included as a comparison value since tap water in California is required to be compared to both MCLs and PHGs (Sullivan, 2009).
Results Overview A total of 14 of the 17 metals were detected in at least one of the source brands of bottled natural spring water. The three metals not detected in any of the samples were beryllium, mercury, and thallium. The concentrations of arsenic in at least one sample in all six source brands as well as all six source brand means were above the PHG of 0.004 µg/L. No other individual metal concentrations or means were above either of their respective comparison criteria. All six of the source brands of bottled spring water had detected concentrations of heavy metals. The minimum number of detected metals in a source brand was 10 and the maximum was 13. Table 1 presents a summary of the results. Data Quality Laboratory quality assurance/quality control samples, including matrix spikes and matrix spike duplicates, were all reported to be within acceptable ranges. Therefore, no issues occurred with the analytical data that required any samples to be discarded. Due to the very low detection limits achieved in the analyses, a total of seven metals had “estimated” or J-qualified reported laboratory blank concentrations. The metals and their estimated laboratory blank concentrations are barium (0.036 µg/L), chromium (0.017 µg/L), nickel (0.014 µg/L), copper (0.032 µg/L), antimony (0.11 µg/L), silver (0.069 µg/L), and zinc (1.4 µg/L). Analytical results for metals with reported detections in laboratory blank samples were handled in the following manner. Any bottled
June 2011r+PVSOBMPG&OWJSPONFOUBM)FBMUI
9
" % 7 " / $ & . & / 5 0 ' 5 ) & SCIENCE
TABLE 1 Heavy Metals Detected in Bottled Natural Spring Water
Comparison Values
Silver
Arsenic
Barium
Beryllium
Cadmium
Cobalt
Chromium
Copper
a
d
N/A
50
1000
4
5
N/A
100
1300j
CA-MCLb
N/Ad
10
1000
4
5
N/A
50
1300j
PHG
N/A
0.004
2000
1
0.04
N/A
N/A
300
Result
Result
Result
Result
Result
Result
Result
Result
Ethos
Re
0.4
25
ND
ND
0.037
0.42
0.25
MCL
c
Source brand
Ethos
0.26
f
0.39
25
ND
ND
0.035
1.8
0.35
Ethos
0.25
0.42
26
ND
ND
0.038
0.41
0.29
Mean
0.25
0.40
25.33
ND
ND
0.04
0.88
0.30
Icelandic
NDg
0.1 h
0.16
ND
ND
0.008
0.96
0.33
Icelandic
ND
0.12
0.17
ND
ND
0.012
1.1
0.37
Icelandic
ND
0.12
0.22
ND
ND
0.007
0.89
0.35
Mean
ND
0.12
0.18
ND
ND
0.01
0.98
0.35
Fiji
ND
0.81
4.7
ND
ND
0.036
0.71
R
Fiji
ND
0.078
4.7
ND
ND
0.035
0.62
1.1
Fiji
ND
0.87
4.8
ND
ND
0.039
0.85
0.11
Mean
ND
0.59
4.73
ND
ND
0.04
0.73
0.61
Evian
ND
0.64
110
ND
ND
0.1
0.74
0.074
Evian
ND
0.65
110
ND
ND
0.1
0.47
0.073
Evian
ND
0.58
110
ND
ND
0.1
0.38
R
Mean
ND
0.62
110.00
ND
ND
0.10
0.53
0.07
Arrowhead
R
0.072
9.8
ND
0.017
0.067
2.4
0.17
Arrowhead
ND
0.074
9.6
ND
ND
0.067
1.1
0.21
Arrowhead
ND
0.079
9.6
ND
0.018
0.068
0.78
0.21
Mean
ND
0.08
9.67
ND
0.01
0.07
1.43
0.20
Crystal
ND
2
0.24
ND
ND
0.054
1.3
0.11
Crystal
ND
1.8
0.25
ND
ND
0.052
0.95
0.092
Crystal
ND
2
0.24
ND
ND
0.051
0.69
ND
Mean
ND
1.93
0.24
ND
ND
0.05
0.98
0.07
Blank
0.0691 h
ND
0.0362
ND
ND
ND
0.0171
0.0323
MDL
0.008
0.014
0.024
0.022
0.013
0.005
N/A
0.022
i
i
b
j
MCL is the federal maximum contaminant level. CA-MCL is the California maximum contaminant level. c PHG is the California public health goal. d N/A is not applicable. e R notation for data removed due to blank contamination. f All reported concentrations in units of μg/L. g ND is nondetect concentration at reported MDL. h Results in italics are J-qualified.
10
MDL is reported analytical method detection limit California action level k Draft proposed PHG l Secondary MCL
a
7PMVNFr/VNCFS
Continued on page 11
TABLE 1
continued from page 10
Heavy Metals Detected in Bottled Natural Spring Water
Comparison Values
Mercury
Molybdenum
Nickel
Lead
Antimony
Selenium
Thallium
Vanadium
Zinc
a
2
N/A
N/A
15
6
50
2
N/A
5000l
CA-MCLb
2
N/A
100
15
6
50
2
N/A
5000l
1.2
N/A
12
0.2
20
30
0.1
N/A
N/A
Result
Result
Result
Result
Result
Result
Result
Result
Result
ND
0.16
0.25
ND
0.32
ND
ND
2.8
2.9
MCL PHG
c
Source brand Ethos
k
Ethos
ND
0.21
0.27
ND
0.31
0.053
ND
2.7
2.9
Ethos
ND
0.13
0.2
0.023
R
0.059
ND
3
R
Mean
ND
0.17
0.24
0.01
0.32
0.06
ND
2.83
2.90
Icelandic
ND
0.13
0.22
ND
0.32
0.062
ND
14
R
Icelandic
ND
0.13
0.29
ND
0.35
0.17
ND
13
R
Icelandic
ND
0.12
0.41
ND
0.37
0.16
ND
13
R
Mean
ND
0.13
0.31
ND
0.35
0.13
ND
13.33
N/A
Fiji
ND
0.63
0.24
ND
0.62
0.026
ND
46
R
Fiji
ND
0.63
0.79
ND
0.64
0.064
ND
46
R
Fiji
ND
0.65
0.27
0.26
0.65
0.038
ND
47
3.8
Mean
ND
0.64
0.43
0.26
0.64
0.04
ND
46.33
3.8
Evian
ND
0.61
0.89
ND
0.58
0.45
ND
ND
6.6
Evian
ND
0.59
0.87
ND
0.51
0.62
ND
ND
6.2
Evian
ND
0.65
0.92
ND
0.55
0.54
ND
ND
5.5
Mean
ND
0.62
0.89
ND
0.55
0.54
ND
ND
6.10
Arrowhead
ND
12
0.61
ND
R
0.46
ND
1.9
3.7
Arrowhead
ND
12
0.75
ND
R
0.45
ND
1.9
3.5
Arrowhead
ND
12
0.41
ND
R
0.39
ND
2
R
Mean
ND
12.00
0.59
ND
N/A
0.43
ND
1.93
3.6
Crystal
ND
0.76
0.49
ND
0.58
0.21
ND
ND
3.4
Crystal
ND
0.61
0.27
ND
0.56
0.22
ND
ND
R
Crystal
ND
0.71
0.24
ND
0.57
0.2
ND
ND
R
Mean
ND
0.69
0.33
ND
0.57
0.21
ND
ND
3.4
Blank
ND
ND
0.014
ND
0.106
ND
ND
ND
1.41
MDL
0.025
N/A
N/A
0.017
N/A
0.017
0.02
0.009
N/A
i
MDL is reported analytical method detection limit California action level k Draft proposed PHG l Secondary MCL
a
i
b
j
MCL is the federal maximum contaminant level. CA-MCL is the California maximum contaminant level. c PHG is the California public health goal. d N/A is not applicable. e R notation for data removed due to blank contamination. f All reported concentrations in units of μg/L. g ND is nondetect concentration at reported MDL. h Results in italics are J-qualified.
June 2011r+PVSOBMPG&OWJSPONFOUBM)FBMUI
11
" % 7 " / $ & . & / 5 0 ' 5 ) & SCIENCE
water concentration that was less than two times the reported blank concentration was not considered to be due to the bottled water and that sample was removed from the dataset. A total of 17 samples from four metals were removed from the data for this reason. The metals and the number of removed samples were silver, 2; copper, 2; antimony, 4; and zinc, 9. For silver and copper, the removed samples left some bottled water sources with only two results. For antimony and zinc, data removals left three sources with two samples remaining, two sources with only one result remaining, and two sources with no results remaining. Removed samples are designated with an “R” in Table 1. Evaluation by Metal Three metals were not detected in any of the source brands: beryllium, mercury, and thallium. The MDLs for these three metals were 0.022, 0.025, and 0.02 µg/L, respectively, and these MDLs are all below both their respective comparison criteria. Fourteen metals were detected. The results are presented in decreasing frequency of detection. Six metals—arsenic, barium, cobalt, chromium, molybdenum and nickel—had 18 detects out of 18 samples and no samples were removed due to blank concentrations. Selenium was detected in 17 of 18 samples with no removals. Copper was detected in 17 samples but two were removed due to blank concentration. Antimony had 18 reported detections but four samples were removed due to blank concentration. Vanadium was detected in 12 samples with no removals. Zinc was detected in all 18 samples but 9 were removed due to blank concentration. Silver had four detects but two removals due to blank concentration. Only two detects of cadmium and lead occurred. Evaluation by Source All six source brands of natural spring water contained detectable concentrations of heavy metals. Ethos contained 13 metals and had the only detected concentrations of silver. Fiji contained 12 metals and had the highest concentrations of four metals: copper, lead, antimony, and vanadium. Arrowhead contained 11 metals and had the highest concentrations of three metals: cadmium, chromium, and molybdenum.
12
7PMVNFr/VNCFS
Icelandic, Evian, and Crystal each had 10 metals. Icelandic had no maximum concentrations, Crystal had the highest arsenic concentration, and Evian had the maximum concentrations of five metals: barium, cobalt, nickel, selenium, and zinc.
Discussion Unlike tap water in California, which is regulated by the Health and Human Services Agency of the California Department of Public Health (CDPH), bottled water is regulated by the CDPH Food and Drug Branch (Sullivan, 2009). New California bottled water regulations require the testing and the reporting of chemical concentrations in bottled water. Testing reports are available to the general public, but only if requested. No requirements exist for proactively providing the test results to the general public or to provide notifications on water bottle packaging. The results for the six sources of bottled natural spring water would not trigger any special reporting requirement or bottled water package labeling requirements. The only language that consumers would see on their packaging is specified in HSC Section 111170, Subsection (f)(3)(B) (Bottled, Vended, Hauled, and Processed Water, 2008): “For more information and to obtain additional consumer information relating to water quality, including a bottled water report, contact (name of bottled water company) at (telephone number or toll-free telephone number) and (at least one of the following: mailing address, e-mail address, or the bottled water company’s Web site).” The only metal to be close to triggering any notification requirement was arsenic. The highest level of arsenic in bottled natural spring water of 2 µg/L is below the federal MCL and CA-MCL. If concentrations were between 5 and 10 µg/L, however, the following notification would have been required only on the analytical report provided to consumers who request it: “Arsenic levels above 5 ppb (parts per billion) and up to 10 ppb are present in your drinking water. While your drinking water meets the current EPA standard for arsenic, it does contain low levels of arsenic. The standard balances the current understanding of arsenic’s possible health effects against the costs of removing arsenic from drinking water. The State Department of Public Health
continues to research the health effects of low levels of arsenic, which is a mineral known to cause cancer in humans at high concentrations and is linked to other health effects, including, but not limited to, skin damage and circulatory problems.” This 2 µg/L concentration of arsenic, however, if found in tap water in California would have been compared to both the PHG and the CA-MCL (Sullivan, 2009). Had this 500-fold exceedance of the arsenic PHG of 0.004 µg/L by the bottled water concentration of 2 µg/L been detected in tap water, it would have to be noted in the mandated Consumer Confidence Reports (CCR) that large water system owners must provide, proactively, to their customers. Consumers of tap water containing 2 µg/L of arsenic in California have the CCR provided to them as a means of notification allowing them to evaluate the safety of their water. Consumers of bottled natural spring water must request the analytical report from the water supplier in order to become informed regarding the safety of their bottled water.
Conclusion We recommend that changes be made to the way that bottled water is regulated in regards to the two issues highlighted in this article: comparison criteria and reporting. Bottled water should be compared to the same criteria as tap water. PHGs, as the health-based criteria from drinking water developed by the state of California, should be used along with MCLs in the evaluation of the concentration of chemicals in bottled water. For reporting, a more transparent process that allows for easier and quicker consumer access to the testing results for bottled water should be developed. For example, bottled water packaging could contain the phone number or web address where test results could be immediately accessed. Web sites of bottled water producers could contain the results of the required analytical testing with explanation for consumers just as found on many Web sites for tap water providers. These improvements would increase consumer knowledge and lead to betterinformed decisions regarding consumer choices on water consumption. These concentrations of heavy metals in bottled natural spring water and the reporting requirements triggered by detected metals are
representative of the information now available to California consumers of bottled water. Unless the bottled water consumers are aware of the availability of analytical reports, however, the usefulness of this information, likely received long after the water has been consumed, is limited. Other states considering how to regulate bottled water should consider the appropriateness of comparison
criteria and the lack of consumer notification in the California law and require, similar to tap water, health-based comparisons and proactive notification of consumers regarding the presence of contaminants in their bottled water. These changes would result in a reporting process more similar to the reporting process for tap water and would result in better-informed consumers.
Corresponding Author: Michael J. Sullivan, Professor, Department of Environmental and Occupational Health, California State University at Northridge, 18111 Nordhoff Street, Northridge, CA 91330. E-mail:
[email protected].
References Allen, L. & Darby, J.L. (1994). Quality control of bottled and vended water in California: A review and comparison to tap water. Journal of Environmental Health, 56(8), 17–22. Bottled, Vended, Hauled, and Processed Water. (2008). California Health and Safety Code, Sections 111070–111195. California Department of Public Health (2009). List of federal and California maximum contaminant levels. Retrieved July 28, 2010, from http://www.cdph.ca.gov/certlic/drinkingwater/pages/ chemicalcontaminants.aspx Ikem, A., Odueyungbo, S., Egiebor, N., & Nyavor, K. (2002). Chemical quality of bottled waters from three cities in eastern Alabama. Science of the Total Environment, 285(1–3), 165–175. International Bottled Water Association. (2004). U.S. bottled water consumerage by leading states 1994–1999. Retrieved June 22, 2009, from http://www.bottledwater.org/public/statistics_main.htm Office of Environmental Health Hazard Assessment. (2009). Public health goals. Retrieved June 22, 2009, from http://www.oehha. ca.gov/water/phg/allphgs.html Olsen, E.D. (1999). Bottled water: Pure drink or pure hype? New York: Natural Resources Defense Council. Rodwin, J.G., Jr. (2009, April/May). Confronting challenges: U.S. and international bottled water developments and statistics for 2008. Bottled Water Reporter, 12–18. Retrieved June 22, 2009,
from http://www.bottledwater.org/public/2008%20Market%20 Report%20Findings%20reported%20in%20April%202009.pdf Secondary Water Standards. (2006). California Code of Regulations, Title 22, Division 4, Chapter 15, Article 16. Retrieved July 28, 2010, from http://www.cdph.ca.gov/certlic/drinkingwater/ Documents/Recentlyadoptedregulations/R-21-03-finalregtext.pdf Shotyk, W., & Krachler, M. (2007). Contamination of bottled waters with antimony leaching from polyethylene terephthalate (PET) increase upon storage. Environmental Science & Technology, 41(5), 1560–1563. Sullivan, M.J. (2009). Regulating tap water and bottled water in California. California Journal of Environmental Health, 24(1), 15–19. U.S. Environmental Protection Agency. (1994a). Method 200.8: Determination of trace elements in waters and wastes by inductively coupled plasma—Mass spectrometry (Revision 5.4.). Retrieved January 29, 2011, from http://water.epa.gov/scitech/swguidance/ methods/methods_index.cfm U.S. Environmental Protection Agency. (1994b). Method 245.1: Determination of mercury in water by cold vapor atomic absorption spectrometry (Revision 3.0.). Retrieved January 29, 2011, from http://water.epa.gov/scitech/swguidance/methods/methods_ index.cfm
NEHA Credentials Protecting human health and the environment since 1937
Why should your employees hold a NEHA credential? BECAUSE YOU WANT THE BEST WORKING TO PROTECT YOUR COMMUNITY! Professional credentials such as the Registered Environmental Health Specialist/Registered Sanitarian (REHS/RS) and Certified Professional – Food Safety
Copyright 2011, Journal of Environmental Health (www.neha.org)
(CP-FS) have been rigorously developed to insure that those who successfully pass the credentialing exams have the knowledge, skills, and abilities to competently practice environmental health. For more information on NEHA credentials, please visit our Web site at neha.org/credential or contact the credentialing department at (303) 756-9090, ext. 337.
June 2011r+PVSOBMPG&OWJSPONFOUBM)FBMUI
13
