Speciation of trace mercury impurities in fish oil supplements

Accepted Manuscript Speciation of trace mercury impurities in fish oil supplements

Ni Mei, Bunhong Lai, Jixin Liu, Xuefei Mao, Guoying Chen PII:

S0956-7135(17)30395-X

DOI:

10.1016/j.foodcont.2017.08.001

Reference:

JFCO 5735

To appear in:

Food Control

Received Date:

23 May 2017

Revised Date:

31 July 2017

Accepted Date:

02 August 2017

Please cite this article as: Ni Mei, Bunhong Lai, Jixin Liu, Xuefei Mao, Guoying Chen, Speciation of trace mercury impurities in fish oil supplements, Food Control (2017), doi: 10.1016/j.foodcont. 2017.08.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT 1

1

Speciation of trace mercury impurities in fish oil

2

supplements

3

Ni Mei a, Bunhong Lai b, Jixin Liu c, Xuefei Mao c, Guoying Chen b,*

4

a Shanghai

5

b U.S.

6 7 8

9

Institute for Food and Drug Control, 1500 Zhangheng Road, Shanghai 201203, China.

Department of Agriculture, Agricultural Research Service, Eastern Regional Research

Center, 600 E. Mermaid Lane, Wyndmoor, Pennsylvania 19038, USA. c Institute

of Quality Standard and Testing Technology for Agro-Products, Chinese Academy of

Agricultural Sciences, 12 S. Zhongguancun Street, Beijing 100081, China

ARTICLE INFO

10

Article history:

11

Received

ACCEPTED MANUSCRIPT 2

12

Keywords:

13

Mercury impurity

14

Fish oil

15

Supplement

16

Speciation

17

Photochemical

18

Vapor generation

19

Atomic fluorescence spectrometry

20

Conflict of interest

21

Mention of trade names or commercial products in this article is solely for the purpose of

22

providing specific information and does not imply recommendation or endorsement by the U.S.

23

Department of Agriculture (USDA). USDA is an equal opportunity employer.

24

The authors declare no competing financial interest.

ACCEPTED MANUSCRIPT 3

25

ABSTRACT

26

Fish oil supplement is becoming increasingly popular worldwide because of beneficial long-

27

chain omega-3 polyunsaturated fatty acids. However, mercury (Hg) impurity causes considerable

28

concern because of its toxicity and bioaccumulation in the food chain. In this work, Hg

29

impurities were extracted from fish oil by liquid-liquid partitioning. The sample solution was

30

then mixed with a reductant (0.4% anthranilic acid-20% formic acid) and sequentially exposed to

31

311 and 254 nm UV radiation. The resulting Hg0 vapor was detected by atomic fluorescence

32

spectrometry. Speciation was fulfilled by solving a set of two linear equations. Recovery of

33

MeHg+ was 73%; total Hg was validated by ICP-MS. This method achieved 0.50 and 0.63 ng

34

mL-1 limits of detection for Hg++ and MeHg+, respectively. Average Hg++ and MeHg+ contents in

35

fish oil samples (n=38), 0.670.45 and 1.11.3 ng mL-1, respectively, were 2-3 orders of

36

magnitude lower than those in fish.

ACCEPTED MANUSCRIPT 4

37

1. Introduction

38

Benefits of long-chain omega-3 polyunsaturated fatty acids (LCn3PUFAs) in fish oil were

39

recognized four decades ago based on the correlation of high fish diet and low coronary artery

40

disease (CAD) rate in the Greenland Eskimo population (Dyerberg, Bang, & Hjorne, 1975 ).

41

Numerous studies since have shown that in addition to cardiovascular health (Mozaffarian & Wu,

42

2011; Nestel, et al., 2015; Kris-Etherton, Harris, & Appel, 2002). LCn3PUFAs also benefit eye

43

(Yashodhara, et al., 2009), gastrointestinal (Yashodhara, et al., 2009), brain and neurological

44

(Burca & Watson, 2014), as well as dozens of other conditions. These benefits were ascribed to

45

rich presence of eicosapentaenoic acid (EPA, 20:5n-3) and decosahexanoic acid (DHA, 22:6n-3).

46

In comparison, benefits of alpha-linolenic acid (ALA, 18:3n-3), another LCn3PUFA derived

47

from plant oils or seeds such as flaxseed, remains inconclusive (C. Wang, et al., 2006). Pathways

48

of endogenous conversion to EPA and DHA exist but at low rates: <8% to EPA vs. <4% to DHA

49

(Pawlosky, Hibbeln, Novotny, & Salem, 2001). Worldwide, fish oil production reached 1-1.25

50

million tonnes in 2010 (Pike & Jackson, 2010); LCn3PUFA supplements are becoming

51

increasingly popular reaching $1.1 billion annual sale in 2011 (FN Media Group, 2015).

52

MeHg+ is known to bioaccumulate along the aquatic food chain and accounts for 75-98% of

53

total Hg presence in fish. Fish oil supplements are produced from anchovy, mackerel, herring,

54

sardines, tuna, salmon, cod, krill, etc., among which long-life, piscivorous species can

55

accumulate Hg up to several g g-1. For human, consumption of fish and shell fish is the main

56

pathway of Hg exposure. Hg impurities in fish oil thus become legitimate concerns besides other

57

lipophilic environmental pollutants such as polychlorinated dibenzo-p-dioxins (PCDDs),

58

polychlorinated dibenzofurans (PCDFs), and polychlorinated biphenyls (PCBs) (Burca &

59

Watson, 2014; Oh, 2005). Once ingested, lipophilic and hydrophilic MeHg+ can easily pass cell

ACCEPTED MANUSCRIPT 5

60

membranes and blood-brain barrier, bind to sulfhydryl and selenohydryl groups, alter 3-D

61

structures of proteins, and interrupt cell functions (Farina, Aschner, & Rocha, 2011). It is well

62

documented that MeHg+ manifests a wide spectrum of adverse effects to mammal, collectively

63

known as Minamata disease (National Research Council, 2000). Neurotoxicity of this compound

64

is extremely harmful to children from utero stage to early childhood (Castoldi, et al., 2008;

65

Trasande, Landrigan, & Schechter, 2005). Losses of intelligence and productivity caused by Hg

66

emission from US coal-fired power plants alone, which accounts for 41% of US anthropogenic

67

emission, are estimated to reach $8.7 billion per year (Hylander & Goodsite, 2006). To protect

68

public health, the United Nations Food and Agriculture Organization/World Health Organization

69

(FAO/WHO) Joint Expert Committee on Food Additives (JECFA) has established provisional

70

tolerable weekly intakes (PTWI) at 1.6 µg MeHg+/kgbw and 4 µg iHg/kgbw for the general

71

population, but warned a greater risk for pregnant and lactating women. Hg impurities in fish oil

72

supplements is not regulated by the US Food and Drug Administration (FDA); the fish oil

73

industry follows the 100 parts per billion (ppb) total Hg safe level set by the Global Organization

74

for EPA and DHA Omega-3s (GOED), the European Pharmacopoeia, Norwegian Medicinal

75

Standard, and the Council for Responsible Nutrition, etc. To boost consumer confidence and

76

uphold public health, Hg impurities in fish oil supplement must be closely monitored.

77

Ashing coupled to atomic absorption spectrometry (AAS) achieved 6 ng mL-1 limit of

78

detection (LOD) for tHg impurity (Foran, Flood, & Lewandrowski, 2003). Such sensitivity is

79

marginal for fish oil samples. To enhance sensitivity, amalgamation was implemented to enrich

80

elemental Hg after ashing and catalytic conversion (Levine, et al., 2005). Speciation of Hg

81

impurities in fish oil has recently been fulfilled by HPLC coupled to inductively-coupled plasma

82

(ICPMS) that improved LOD to 0.5-1 ng g-1 (Yao, Jiang, Sahayam, & Huang, 2017). Speciation

ACCEPTED MANUSCRIPT 6

83

is not only important to consumers due to species-dependent toxicity, but also to quality control

84

and quality assurance. Speciation data render extra insight to guide purification practice. HPLC-

85

ICPMS is a powerful technique for sensitive speciation, but expensive instrumentation and high

86

operation cost become obstacles to its availability in smaller laboratories especially in

87

developing countries. Described in this manuscript are the speciation results of trace Hg

88

impurities in fish oil by differential photoreduction under two UV wavelengths, 311 vs. 254 nm.

89

The resulting Hg0 vapor was detected by atomic fluorescence spectrometry (AFS). Average

90

contents of Hg++ and MeHg+ (n=38) were 2-3 orders of magnitude lower than fish and hence

91

much safer.

92

2. Materials and methods

93

2.1. Chemicals and solutions

94

Hg(NO3)2 standards in 12% HNO3 (1000 g mL-1) and 99.9% solid methylmercury chloride

95

(MeHgCl) were purchased from Fluka (Milwaukee, WI, USA). HgCl2 powder (99.999%) and

96

methylmercury chloride standard in water (1000 g mL-1) were purchased from Alfa Aesar

97

(Ward Hill, MA, USA). ACS reagent grade anthranilic acid (AA), 96% formic acid (FA), and

98

nitric acid were purchased from Sigma-Aldrich (Milwaukee, WI, USA), from which a 0.4% (w/v)

99

AA20% (v/v) FA reductant solution was prepared daily. Stock standards at 0.5 g mL-1 were

100

made weekly by serial dilution in deionized water (DIW) and stored at 4 °C; working standard

101

solutions at 0-5 ng mL-1 were prepared daily. Spiked fish oil samples were prepared from

102

MeHgCl and serially diluted to 10-100 ng mL-1 using a fish oil with low (< LOQ) intrinsic Hg

103

content. Used glassware was soaked in 15% nitric acid overnight and rinsed thoroughly with

104

deionized water (DIW). A Barnstead E-pure system (Dubuque, IA) was used to make DIW.

ACCEPTED MANUSCRIPT 7

105

2.2. Liquid-liquid extraction

106

Fish oil samples in capsule or liquid form were purchased from local stores in the

107

Philadephia area (PA, USA) or online. Capsules were punctured with a clean syringe needle; the

108

fish oil was squeezed to a glass vial, from which 2 mL was pipetted to a 50 mL polypropylene

109

centrifuge tube using a displacement pipette. After addition of 40 mL of DIW, the tubes were

110

capped tightly and arranged vertically in a tube rack. The tubes were secured between a base and

111

a cover of a LC1012 vortex mixer (Glas-Col, Terre Haute, IN, USA); both were lined with a 0.5”

112

sheet of polymeric rubber. Vigorous agitation, set at motor speed 80, lasted for 10 min with

113

pulsing. Centrifugation followed at 4000 rpm for 10 min; the upper oil layer was discarded while

114

the aqueous layer was kept for analysis.

115

2.3. Photochemical vapor generation

116

A custom-made, synthetic silica coil was installed between a mixing valve and a gas/liquid

117

separator (G/L) of a Millennium Merlin atomic fluorescence spectrometer (AFS) (P S Analytical,

118

Kent, UK) as a UV-photoreactor (Chen, et al., 2017). A 254 nm low-pressure Hg lamp was

119

inserted into the center of the coil while two 311 nm fluorescent lamps were installed by the coil

120

with a 10 mm gap. The spectrometer, under the control of the Millennium software, was operated

121

in flow-injection (FI) mode. The aqueous sample solution and the 0.4% AA20% FA reductant

122

solution were delivered by a peristaltic pump at 9 and 4.5 mL min-1 flow rates, respectively, and

123

mixed in the reactor coil. The mixture was exposed to either 311 nm or 254 nm UV light. The

124

resulting Hg0 vapor was swept by high-purity argon at 300 mL min-1 to the G/L where liquid was

125

drained. The vapor reached a Perma-Pure dryer where most moisture permeated across a Nafion

126

membrane and evaporated into a counter-flowing (2500 mL min-1) nitrogen stream. The Hg0

127

vapor finally entered the AFS detector.

ACCEPTED MANUSCRIPT 8

128

2.4. Atomic fluorescence spectrometry

129

Detection of Hg0 vapor was performed in the detection chamber of the Millennium Merlin

130

AFS. Dry Hg0 vapor was excited by a 254 nm beam from a hollow cathode lamp. The resonance

131

fluorescence signal, collected at 90, passed a 254 nm filter and was detected by a

132

photomultiplier tube.

133

2.5. Hg++ vs. MeHg+ Speciation

134

Concentrations of each analytes in a sample were solved from a set of two linear equations

135

obtained at 311 or 254 nm, respectively. The AFS signal intensity was a linear function of the

136

concentrations of both analytes:

137

I = m[Hg++] + n[MeHg+]

(=254 or 311 nm)

138

where I was AFS peak height at wavelength ; m and n are slopes of standard curves at

139

wavelengths ; [Hg++] and [MeHg+] are the concentrations of Hg++ and MeHg+ expressed in Hg.

140

2.6. Microwave digestion and ICP-MS validation

141

Digestion of fish oil was performed using a MARS 6 microwave digestion system (CEM,

142

Matthews, NC, USA). Aliquots of 0.50 mL fish oil were delivered to 55 mL Xpress vessels and

143

added with 5 mL of concentrated HNO3. Digestion was carried out at 120 C for 2 min, 160 C

144

for 8 min, and 190 C for 30 min; with 2 min ramps. After cooled down to room temperature, the

145

contents were filled with DIW to 50 mL. Total Hg (tHg) was measured using an Agilent 7900

146

ICP-MS (Santa Clara, CA, USA) under the operation parameters listed in our previous report

147

(Chen, et al., 2017).

ACCEPTED MANUSCRIPT 9

148

3. Results and discussion

149

3.1. Liquid-liquid extraction

150

Hg impurities in oil matrix can be transferred to gas phase by ashing or distillation for total

151

mercury (tHg) determination; or to aqueous phase by wet extraction. Microwave-aided digestion

152

in closed vessel speeds up extraction with minimal loss. However, MeHg+ may be oxidized to

153

Hg++ by strong oxidizing acids and high temperature which are needed to decompose organic

154

matrices (Kelly, Long, & Mann, 2003). Extraction induced by emulsion breaking (EIEB) is a

155

common approach to transfer elemental analytes from oil to surfactant phase (Robaina, Brum, &

156

Cassella, 2012). Relatively polar mercury species, fortunately, can be rapidly extracted by liquid-

157

liquid partitioning (LLP). For oil matrices, EIEB is commonly applied to extract target analytes

158

to aqueous media (Cassella, Brum, de Paula, & Lima, 2010; Corazza & Tarley, 2016). A

159

surfactant solution added to the oil sample breaks oil down during agitation into a large number

160

of small oil droplets forming an oil-in-water emulsion. Large interface area enables rapid mass

161

equilibrium. This emulsion is next broken down by heating or centrifugation, allowing analytes

162

to enter the surfactant phase. However, the surfactant caused severe foaming in the G/L and thus

163

interfered with PVG. Dispersive liquid-liquid microextraction (DLLME) also achieves rapid

164

mass transfer by forming an emulsion, but the added disperser, such as alcohol or acetonitrile,

165

also interfered with the Nafion dryer.

166

Fortunately, liquid-liquid extraction (LLE) is effective to extract both Hg species to aqueous

167

phase based on favorable partitioning coefficients: Ko/w = 0.61 for Hg++ and Ko/w = 1.60.2 for

168

MeHg+ (Halbach, 1985; Major, Rosenblatt, & Bostian, 1991). Foaming didn’t occur for LLE.

169

Under 10 min agitation, MeHg+ recovery was ~73% at 10 ng mL-1 with a water-to-oil volume

170

ratio of 20. Because Hg++ is more polar than MeHg+, its recovery was expected to be higher. In

ACCEPTED MANUSCRIPT 10

171

practice, however, measurement was not carried out due to multiple difficulties to make a stable

172

and homogeneous Hg++ solution in organic media (Snell, et al., 1998). The closest commercial

173

standards, Hg++ in mineral oil, are of no use because the added stabilizing ligand disables this

174

PVG method. During optimization, it turned out that 10-min vigorous agitation was enough to

175

achieve partition equilibrium using a platform vortexer.

176

3.2. Hg speciation by PVG

177

Differential PVG behaviors at UV-B vs. UV-C were the basis of this mathematical approach.

178

This method gained green chemistry advantage by replacing unstable and harmful reductants,

179

such as SnCl2 or NaBH4 with less harmful FA and AA. Four prerequisites of this mathematical

180

approach were described previously (Chen, et al., 2017). The kinetics of Hg++ or MeHg+

181

reduction, however, can be complex under experimental conditions. The key parameters were (a)

182

reductant composition and concentration, (b) flow rates of sample and reductant solutions, (c)

183

actinic light wavelengths and output power, and (d) exposure time. Details of consideration and

184

optimization of these conditions were previously presented (Chen, et al., 2017). Though linear

185

relationship (R2 > 0.999) was observed for both Hg species at either 311 nm or 254 nm

186

wavelength under optimized conditions, it is advised to check linear dependence experimentally.

187

3.3. ICP-MS validation based on tHg

188

When suitable CRM is unavailable in elemental analysis, validation relies upon spike-

189

recovery studies (Szymczycha-Madeja & Welna, 2013). Unlike MeHg+, preparation of a stable

190

and homogeneous Hg++ standard in fish oil failed, vide supra. An able technique, ICP-MS, was

191

thus applied to validation based on tHg. Digestion was carried out according to a previously

192

reported microwave-aided extraction (MAE) protocol (Yao, et al., 2017) except H2O2 was

ACCEPTED MANUSCRIPT 11

193

excluded to avoid excess foaming. The recovery was found to be 90.4% at 80 ng mL-1 Hg in a

194

MeHgCl spiked sample. Four fish samples with relatively high tHg presence (2.18 - 6.06 ng mL-1)

195

were selected for comparison. The differences between two methods range from -0.36 to 2.18 ng

196

mL-1 (Fig. 1). To break down oil matrix into aqueous phase, 5 mL of concentrated (~70%) HNO3

197

was added to 0.5 mL of fish oil. In the process, MeHg+ was oxidized to Hg++. However, to avoid

198

corrosion to nickel cones, the sample solution should contain 10% HNO3. The digest was thus

199

diluted 10 times with water, leading to a final Hg content two orders of magnitude lower than

200

that in the original fish oil. The method LOD and LOQ were 1.6 and 5.4 ng mL-1, respectively,

201

which were much higher than the ICP-MS instrument LOD and LOQ. Obviously, trace-level Hg

202

presence in fish oil supplements in today’s market pushed both PVG-AFS and ICP-MS

203

techniques to their limits. The overlapping error bars in Fig. 1 indicate reasonable agreement

204

between these two methods.

205

3.4. Hg++ vs. MeHg+ Speciation in fish oil supplements

206

Among 38 fish oil samples (Table 1), 28 were in capsule form while the others were in liquid

207

form. Some were labelled with their original fish species or organs; some were added with lemon

208

flavor. LODs were calculated from the baseline data collected in a course of one week. LOD of

209

Hg++, 0.50 ng mL-1, was obtained under 311 nm UV-B; whereas LOD of MeHg+, 0.63 ng mL-1,

210

was obtained under 254 nm UV-C. Among 38 fish oil samples tested so far, 12 contained MeHg+

211

above LOD; of which 6 exceeded 1.4 ng mL-1 LOQ. Meanwhile, 15 contained Hg++ above LOD,

212

of which only one exceeded 1.1 ng mL-1 LOQ (Table 1). If the below-LOD data were replaced

213

by respective LODs, then the average Hg++ and MeHg+ concentrations were 0.670.45 and

214

1.11.3 ng mL-1; respectively, leading to an average tHg at 1.8 ng mL-1. These data align with

215

independent test results by LabDoor (LabDoor, 2016) and ConsumerLab (ConsumerLab, 2016).

ACCEPTED MANUSCRIPT 12

216

The former analyzed 53 fish oil products in the U.S. market; average tHg content was obtained at

217

2.4 ppb. The latter applied an unspecified method to 41 common fish oil supplements; all

218

contained Hg below detectable level while most contained trace amount of PCBs. These results

219

were 1-2 orders of magnitude lower than the 100-ppb standard in fish oil industry, and 2-3 orders

220

of magnitude lower than the FDA action level for MeHg+ in fish. In this data set, Hg++ and

221

MeHg+ account for 38% and 62% of tHg, respectively, but below three capsule samples reported

222

elsewhere (Yao, et al., 2017).

223

On the other hand, these results are slightly lower than those a decade ago (Foran, et al., 2003;

224

Levine, et al., 2005), most likely because of technology advancement in purification practice

225

(Sathivel, 2011). Several measures are especially effective to eliminate polar and hence water

226

soluble components such as proteins and minerals. In the first purification step, metals and other

227

polar components like sulphur compounds are removed by neutralization with NaOH followed

228

by discharge of soap fraction and water washing. Bleaching using solid adsorbents, such as

229

activated carbon and silicates, is mainly designed to improve color and remove off-flavor and

230

oxidation products, but also effectively eliminate metals, soap, and persistent environmental

231

contaminants. Finally, distillation achieves separation by volatility, of which molecular weight

232

plays a key role. Both atmospheric distillation and vacuum distillation require high temperature

233

and long residence time thus are not applicable to heat-sensitive bio oils. Steam distillation

234

overcomes this obstacle by using steam and operating at lower than boiling point temperatures

235

(Sullivan, 1976). Molecular distillation, also known as short-path distillation, operates under

236

0.01-torr vacuum at which the mean free path of the molecules is equal to the equipment

237

dimensions. This feature allows free-molecular flow regime and short-term exposure, and is thus

238

especially suitable for purification and enrichment of heat-sensitive LCn3PUFAs (S. Wang, et al.,

ACCEPTED MANUSCRIPT 13

239

2009). Furthermore, use of species intrinsically lower in aquatic food chain, such as anchovy,

240

herring, sardine, or krill, would lower Hg level in fish oil.

241

Of 38 samples, 19 were made from anchovy, mackerel, or sardine; the average presence:

242

0.580.14 ng mL-1 for Hg++ and 0.890.63 ng mL-1 for MeHg+, are only slightly lower than the

243

overall average (Table 2). Meanwhile, 11 are made from cod; the average presence: 0.830.80

244

ng mL-1 for Hg++ and 1.21.4 ng mL-1 and MeHg+ are only slightly higher than the overall

245

average. The differences of impurity presence among original fish species are hence statistically

246

insignificant. Such facts indicate the effectiveness of purification techniques in production.

247

Speciation data provide insight and guidance to purification practice in the industry. In terms of

248

Hg impurities, fish oil proved much safer than fish. These results would boost consumer

249

confidence and sales of LCn3PUFAs supplements; no doubt popularity of such products will

250

enhance public health.

251

Conclusions

252

Speciation of Hg impurities in fish oil supplement was fulfilled by differential PVG at two

253

wavelengths followed by AFS detection. Both Hg++ and MeHg+ impurities were found to be at

254

low single–digit ng mL-1 level, 1-2 orders of magnitude lower than the industrial safety standard

255

and 2-3 orders of magnitude lower than the FAO-WHO guild line for MeHg+ in fish. Fish oil

256

thus proves much safer than fish in terms of Hg presence.

ACCEPTED MANUSCRIPT 14

257

Table 1. Hg++ and MeHg+ concentrations in 38 fish oil samples by PVG-AFS. #

Format

Origin species

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 LOD LOQ

liquid liquid liquid liquid liquid liquid capsule liquid liquid capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule liquid capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule capsule

cod cod cod pollock cod anchovy, sardine, mackerel cod cod cod pollock, whiting anchovy, mackerel, sardine anchovy, sardine unspecified anchovy, mackerel, sardine salmon anchovy, sardine cod anchovy, mackerel, sardine anchovy, sardine cod Fish anchovy, mackerel, sardine anchovy, mackerel, sardine cod cod anchovy, mackerel, sardine anchovy, mackerel, sardine unspecified anchovy, mackerel, sardine anchovy, mackerel, sardine unspecified pollock, whiting anchovy, mackerel, sardine unspecified anchovy, mackerel, sardine anchovy, mackerel, sardine anchovy, mackerel, sardine anchovy, mackerel, sardine anchovy, mackerel, sardine

Hg++ (ng mL-1) 0.70
MeHg+ (ng mL-1)
ACCEPTED MANUSCRIPT 15

259

260

Table 2. Comparison of Hg++ and MeHg+ concentrations among fish species. #

n

Original species

average Hg++ (ng mL-1)

average MeHg+ (ng mL-1)

1

38

all

0.670.45

1.11.3

2

19

anchovy, mackerel, sardine

0.580.14

0.890.63

3

11

cod

0.830.80

1.21.4

4

8

else

0.660.19

1.62.0

ACCEPTED MANUSCRIPT 16

261 262

Fig. 1. Comparison of tHg results by PVG-AFS and ICP-MS. The error bars indicate 1

263

standard deviation (SD).

ACCEPTED MANUSCRIPT 17

264

References

265

Burca, N., & Watson, R. R. (2014). Fish oil supplements, contaminants, and excessive doses. In

266

F. d. Meester & R. R. Watson (Eds.), Omega-3 fatty acids in brain and neuological health

267

(pp. 447-454). London, UK: Elsevier.

268

Cassella, R. J., Brum, D. M., de Paula, C. E. R., & Lima, C. F. (2010). Extraction induced by

269

emulsion breaking: a novel strategy for the trace metals determination in diesel oil samples

270

by electrothermal atomic absorption spectrometry. Journal of Analytical Atomic

271

Spectrometry, 25, 1704-1711.

272

Castoldi, A. F., Johansson, C., Onishchenko, N., Coccini, T., Roda, E., Vahter, M., Ceccatelli, S.,

273

& Manzo, L. (2008). Human developmental neurotoxicity of methylmercury: Impact of

274

variables and risk modifiers. Regulatory Toxicology and Pharmacology, 51, 201-214.

275

Chen, G., Lai, B., Mei, N., Liu, J., & Mao, X. (2017). Mercury speciation by differential

276

photochemical vapor generation at UV-B vs. UV-C wavelength. Spectrochimica Acta Part B.

277

In Press.

278

ConsumerLab.

(2016).

ConsumerLab.com

Answers

279

https://www.consumerlab.com/reviews/fish_oil_supplements_review/omega3/. Last accessed

280

10.01.2017.

281

Corazza, M. Z., & Tarley, C. R. T. (2016). Development and feasibility of emulsion breaking

282

method for the extraction of cadmium from omega-3 dietary supplements and determination

283

by flow injection TS-FF-AAS. Microchemical Journal, 127, 145-151.

284 285

Dyerberg, J., Bang, H. O., & Hjorne, N. (1975 ). Fatty acid composition of the plasma lipids in Greenland Eskimos. American Journal of Clinical Nutrition, 28, 958-966.

ACCEPTED MANUSCRIPT 18

286 287 288

Farina, M., Aschner, M., & Rocha, J. B. T. (2011). Oxidative stress in MeHg-induced neurotoxicity. Toxicology and Applied Pharmacology, 256, 405-417. FN Media Group, L. (2015).

289

http://www.prnewswire.com/news-releases/billion-dollar-omega-3-fish-oil-market-

290

forecasted-to-continue-to-grow---company-receives-fda-approval-for-high-potency-omega-3-

291

recovery-supplements-296421951.html. Last accessed 01.10.2017.

292

Foran, S. E., Flood, J. G., & Lewandrowski, K. B. (2003). Measurement of mercury levels in

293

concentrated over-the-counter fish oil preparations---Is fish oil healthier than fish? Archives

294

of Pathology and Laboratory Medicine, 127, 1603-1605.

295 296 297 298

Halbach, S. (1985). The octanol/water distribution of mercury compounds. Archives of Toxicology, 57, 139-141. Hylander, L. D., & Goodsite, M. E. (2006). Environmental costs of mercury pollution. Science of the Total Environment, 368, 352-370.

299

Kelly, W. R., Long, S. E., & Mann, J. L. (2003). Determination of mercury in SRM crude oils

300

and refined products by isotope dilution cold vapor ICP-MS using closed-system combustion.

301

Analytical and Bioanalytical Chemistry, 376, 753-758.

302 303

Kris-Etherton, P. M., Harris, W. S., & Appel, L. J. (2002). Fish Consumption, Fish Oil, Omega-3 Fatty Acids, and Cardiovascular Disease. Circulation, 106, 2747-2757.

304

LabDoor. (2016). https://labdoor.com/rankings/fish-oil/report. Last accessed 11.01.2017.

305

Levine, K. E., Levine, M. A., Weber, F. X., Hu, Y., Perlmutter, J., & Grohse, P. M. (2005).

306

Determination of Mercury in an Assortment of Dietary Supplements Using an Inexpensive

307

Combustion Atomic Absorption Spectrometry Technique. Journal of Automated Methods

308

and Management in Chemistry, 2005, 211-216.

ACCEPTED MANUSCRIPT 19

309

Major, M. A., Rosenblatt, D. H., & Bostian, K. A. (1991). The octanol/water partition

310

coefficients of methylmercuric chloride and methylmecury hydroxide in puer water and salt

311

solutions. Environmental Toxicology and Chemistry, 10, 5-8.

312

Mozaffarian, D., & Wu, J. H. Y. (2011). Omega-3 Fatty Acids and Cardiovascular Disease:

313

Effects on Risk Factors, Molecular Pathways, and Clinical Events. Journal of the American

314

College of Cardiology, 58, 2047-2067.

315

National Research Council, C. o. t. T. E. o. M., Board on Environmental Studies and Toxicology.

316

(2000). Toxicological Effects of Methylmercury. Washington DC, USA: National Academy

317

Press.

318

Nestel, P., Clifton, P., Colquhoun, D., Noakes, M., Mori, T. A., Sullivan, D., & Thomas, B.

319

(2015). Indications for Omega-3 Long Chain Polyunsaturated Fatty Acid in the Prevention

320

and Treatment of Cardiovascular Disease. Heart, Lung and Circulation, 24, 769-779.

321 322

Oh, R. (2005). Practical applications of fish oil (Omega-fatty acids) in primary care. Journal of American Board Family Practice, 18, 28-36.

323

Pawlosky, R. J., Hibbeln, J. R., Novotny, J. A., & Salem, N. (2001). Physiological

324

compartmental analysis of α-linolenic acid metabolism in adult humans. Journal of Lipid

325

Research, 42, 1257-1265.

326 327

Pike, I. H., & Jackson, A. (2010). Fish oil: production and use now and in the future. Lipid Technology, 22, 59-61.

328

Robaina, N. F., Brum, D. M., & Cassella, R. J. (2012). Application of the extraction induced by

329

emulsion breaking for the determination of chromium and manganese in edible oils by

330

electrothermal atomic absorption spectrometry. Talanta, 99, 104-112.

ACCEPTED MANUSCRIPT 20

331

Sathivel, S. (2011). Fish oil extraction, purification, and its properties. In C. Alasalvar, F.

332

Shahidi, K. Miyashita & U. Wanasundara (Eds.), Handbook of Seafood Quality, Safety and

333

Health Applications (pp. 423-432). Chichester, UK: Wiley-Blackwell.

334 335 336 337

Snell, J., Qian, J., Johansson, M., Smit, K., & Frech, W. (1998). Stability and reactions of mercury species in organic solution. Analyst, 123, 905-909. Sullivan, F. E. (1976). Steam refining. Journal of the American Oil Chemists’ Society, 53, 358360.

338

Szymczycha-Madeja, A., & Welna, M. (2013). Evaluation of a simple and fast method for the

339

multi-elemental analysis in commercial fruit juice samples using atomic emission

340

spectrometry. Food Chemistry, 141, 3466-3472.

341

Trasande, L., Landrigan, P. J., & Schechter, C. (2005). Public Health and Economic

342

Consequences of Methyl Mercury Toxicity to the Developing Brain. Environmental Health

343

Perspectives, 113, 590-596.

344

Wang, C., Harris, W. S., Chung, M., Lichtenstein, A. H., Balk, E. M., Kupelnick, B., Jordan, H.

345

S., & Lau, J. (2006). n−3 Fatty acids from fish or fish-oil supplements, but not α-linolenic

346

acid, benefit cardiovascular disease outcomes in primary- and secondary-prevention studies:

347

a systematic review. The American Journal of Clinical Nutrition, 84, 5-17.

348 349

Wang, S., Gu, Y., Liu, Q., Yao, Y., Guo, Z., Luo, Z., & Cen, K. (2009). Separation of bio-oil by molecular distillation. Fuel Processing Technology, 90, 738-745.

350

Yao, C.-H., Jiang, S.-J., Sahayam, A. C., & Huang, Y.-L. (2017). Speciation of mercury in fish

351

oils using liquid chromatography inductively coupled plasma mass spectrometry.

352

Microchemical Journal, 133, 556-560.

ACCEPTED MANUSCRIPT 21

353

Yashodhara, B. M., Umakanth, S., Pappachan, J. M., Bhat, S. K., Kamath, R., & Choo, B. H.

354

(2009). Omega-3 fatty acids: a comprehensive review of their role in health and disease.

355

Postgraduate Medical Journal, 85, 84-90.

ACCEPTED MANUSCRIPT 22

356

357

Graphical Abstract

ACCEPTED MANUSCRIPT

Highlights    

Hg++/MeHg+ speciation by photochemical vapor generation at 311 nm vs. 254 nm Hg++ and MeHg+ averaged 0.670.45 and 1.11.3 ng mL-1 in fish oil Method validation by ICP-MS Low LODs: 0.50 ng mL-1 for Hg++ and 0.63 ng mL-1 for MeHg+