Analysis of bromate and bromide in blood

Analysis of bromate and bromide in blood

Toxicology 221 (2006) 229–234 Analysis of bromate and bromide in blood Oscar Qui˜nones a , Shane A. Snyder a,∗ , Joseph A. Cotruvo b , Jeffrey W. Fis...

135KB Sizes 1 Downloads 27 Views

Toxicology 221 (2006) 229–234

Analysis of bromate and bromide in blood Oscar Qui˜nones a , Shane A. Snyder a,∗ , Joseph A. Cotruvo b , Jeffrey W. Fisher c a


Water Quality Research and Development Department, Southern Nevada Water Authority, 1350 Richard Bunker Ave., Henderson, NV 89015, USA b Joseph Cotruvo & Associates, LLC, 5015 46th St. NW, Washington, DC 20016, USA Department of Environmental Health Science, The University of Georgia, 206 Environmental Health Science Building, Athens, GA 30602-2102, USA Received 15 November 2005; received in revised form 5 January 2006; accepted 12 January 2006 Available online 14 February 2006

Abstract Bromate is a regulated disinfection byproduct primarily associated with the ozonation of water containing bromide, but also is a byproduct of hypochlorite used to disinfect water. To study the pharmacokinetics of bromate, it is necessary to develop a robust and sensitive analytical method for the identification and quantitation of bromate in blood. A critical issue is the extent to which bromate is degraded presystemically and in blood at low (environmentally relevant) doses of ingested bromate as it is delivered to target tissue. A simple isolation procedure was developed using blood plasma spiked with various levels of bromate and bromide. Blood proteins and lipids were precipitated from plasma using acetonitrile. The resulting extracts were analyzed by ion-chromatography with inductively-coupled plasma mass spectrometry (IC-ICP/MS), with a method reporting limit of 5 ng/mL plasma for both bromate and bromide. Plasma samples purchased commercially were spiked with bromate and stored up to 7 days. Over the 7 day storage period, bromate decay remained under 20% for two spike doses. Decay studies in plasma samples from spiked blood drawn from live rats showed significant bromate decay within short periods of time preceding sample freezing, although samples which were spiked, centrifuged and frozen immediately after drawing yielded excellent analytical recoveries. © 2006 Aww Research Foundation. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: Bromate; Bromide; ICP-MS; Blood; Toxicity; Analysis; Method

1. Introduction Bromate is found in some finished drinking waters as the result of chlorination and ozonation disinfection processes, which are essential to water treatment in order to maintain microbiological safety suitable for human consumption (Krasner et al., 1995; Siddiqui et al., 1995; Song et al., 1997). Ozonation of water that contains bromide produces bromate directly in the drinking water.

Corresponding author. Tel.: +1 702 856 3668; fax: +1 702 856 3647. E-mail address: [email protected] (S.A. Snyder).

Electrolysis of seawater or brine to produce hypochlorite will also form some bromate as a byproduct. The amount of bromate in the finished water is a function of the amount generated during ozonation of bromide containing water and the degree of bromate contamination in hypochlorite used for disinfection. Bromate exists in solution as the anion (BrO3 − ), as in metal salts (e.g. sodium or potassium bromate) or as hypobromous acid (Krasner et al., 1995). Bromate is thermodynamically a strong oxidizing agent and it will react readily with reducing agents such as sulfide or sulfite. The reaction kinetics are highly pH dependent. At very low pH (∼1) it will even slowly oxidize chloride to chlorine, and bromide to bromine, or bromide and chlo-

0300-483X/$ – see front matter © 2006 Aww Research Foundation. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.tox.2006.01.017


O. Qui˜nones et al. / Toxicology 221 (2006) 229–234

ride to bromine chloride, a mixed halogen. The presence of reducing agents such as hydrogen sulfide at low pH gives rapid reduction of bromate to bromide, probably passing through several intermediates including BrO2 (Zager and Burkhart, 1998). When administered orally, bromate is partially excreted in the urine within 2 hr in concentrations proportional to dose (Fujii et al., 1984). Bromate has been shown to degrade rapidly in various tissues, although decay was relatively slow in blood plasma and saliva (Tanaka et al., 1984). The degradation of bromate has been attributed to glutathione (Tanaka et al., 1984). While these studies have begun to unravel the decomposition pathway of bromate in vivo, the concentrations tested were much greater (i.e., mg/L) than those generally present in drinking water (i.e., ␮g/L). The measurement of bromide in blood is also of importance in bromate metabolism studies. Bromide has been measured in blood previously, with human blood averaging 5–7 mg/L (Kage et al., 2005; Olszowy et al., 1998). The decomposition of bromate to bromide, at doses that would be derived from the high doses used in chronic studies in animals, would be expected to make a significant contribution to the body burden of bromide, potentially affecting the production of pathology, particularly in the thyroid. Inversely, in vivo conversion of the low doses derived from drinking water will not add to the body burden as the process merely converts bromate to its original bromide form in the source water (i.e., non-disinfected water). Consequently, no additional risk of brominism is anticipated as the result of consuming ozonated drinking water. The analysis of bromate in blood plasma was first described by Dunn and McIntyre (1949); however, the colorimetric method was relatively non-selective and lacked sufficient sensitivity with a detection limit of approximately 500 ␮g/mL plasma. A later publication also demonstrated the successful analysis of bromate in blood plasma, yet again at levels much greater than those relevant to drinking water exposure (Tanaka et al., 1984). This study was undertaken to develop a sensitive and selective analytical method to be applied to investigations of in vitro and in vivo chemistry and metabolism of bromate at levels relevant to bromate exposure via drinking water. With bromate formation during ozonation of drinking water typically not exceeding the few tens of micrograms per liter, and the established USEPA (1998) maximum contaminant limit of 10 ␮g/L, expected bromate consumption from drinking water would generally represent a dose of less than 1 ␮g/kg per day. In addition to low drinking water concentrations of bromate, any presystemic degradation following consumption would further increase the need for improved ana-

lytical sensitivity if reflective in vivo inferences are sought. Ion-chromatography coupled to inductively coupled plasma mass spectrometry (IC-ICP/MS) was employed for the analysis of bromate and bromide. Ion-chromatography was utilized to selectively separate the compounds of interest from the complex biological matrix. Detection was accomplished by simultaneously measuring two atomic bromine isotopes. Blood plasma from rat species was obtained both through commercial sources and from centrifuged fresh rat blood. Plasma was treated with acetonitrile to precipitate bulk proteins and lipids. The resulting actetonitrile extract was further concentrated using nitrogen and then diluted in water. The aqueous extract was injected directly into the IC-ICP/MS system. The method reporting limit for this method is 5 ng/mL plasma assuming a 1 mL plasma sample. Spike/recovery experiments using commercial plasma was found to be nearly 100% through several experiments. However, fresh plasma from rat blood spiked immediately at sampling with bromate was found to rapidly degrade spiked bromate. The same plasma once frozen and rethawed, was spiked further, and recovery of bromate was again nearly 100%. These data show that aged blood plasma can be analyzed for bromate with essentially 100% recovery of bromate and bromide, while fresh blood plasma appears to rapidly degrade bromate. 2. Experimental 2.1. Materials Bromate and bromide standards were obtained from Ultra Scientific (North Kingstown, RI). Unfiltered Sprague–Dawley rat plasma samples with sodium heparin were obtained from Bioreclamation, Inc. (Hicksville, NY). Reagent grade acetonitrile was purchased from Burdick & Jackson (Muskegon, MI). Stock and calibration standards were made by dissolving neat standards into deionized water, which had been previously determined free of the compounds of interest at the method detection limits. Trace analysis grade methanol and acetonitrile was obtained from Burdick and Jackson (Muskegon, MI). Formic acid was purchased from EM Science (Gibbstown, NJ). Sodium carbonate was obtained from Mallinckrodt Baker (Paris, KI). Syringe filters (0.45 ␮m polytetrafluoroethylene (PTFE) membrane) were obtained from Nalge Nunc (Rochester, NY). 2.2. Sample preparation For method development and storage stability studies, 1 mL plasma samples were spiked with bromate and/or bromide and extracted in 13 mL centrifuge vials by adding 3 mL acetonitrile

O. Qui˜nones et al. / Toxicology 221 (2006) 229–234


Table 1 Separation and detection parameters for the analysis of bromate and bromide IC parameters Analytical column Guard column Elution Fig. 1. Chromatogram of a 10 ␮g/L bromate and bromide standard in water.

and vortexing. The resulting extract was filtered using a 3 mL syringe equipped with a 0.45 ␮m PTFE syringe filter into a calibrated 15 mL centrifuge vial. The precipitate was rinsed with an additional 1 mL acetonitrile, the rinse solution filtered, and combined with the original filtrate. The resulting extract was then evaporated in a 40 ◦ C water bath under a gentle nitrogen stream to a 1 mL volume, and reconstituted to a 5 mL volume using reagent water. For evaluation of fresh plasma, extraction was adjusted to accommodate limited blood availability from live rat specimens. Whole blood (0.3 mL) was collected from live rats and centrifuged. The plasma aliquots were then precipitated as described previously using a 1 mL extraction volume and a 1 mL rinse volume. The resulting acetonitrile extracts were evaporated to 0.3 mL and diluted to 1.5 mL with reagent water. 2.3. Ion chromatography-inductively-coupled plasma mass spectrometry All analyses were conducted using IC-ICP/MS. The separation method was similar to that employed in EPA method 317.0 for the determination of bromate in drinking water. A Dionex (Sunnyvale, CA) DX 500 ion chromatography (IC) system equipped with a GP50 gradient pump was connected to a 250 mm × 4 mm (IonPac® AS9-HC) analytical column and 4 mm (AG-9-HC) guard column. A 15-min isocratic elution with 9 mM sodium bicarbonate was utilized for all chromatographic separations. A Dionex AS40 autosampler with 0.5 mL cartridge was used for the injection of 250 ␮L of sample into the IC column. For detection, an inductively-coupled plasma mass spectrometer (ICP/MS) from Agilent Technologies (model 7500c, Palo Alto, CA) was used under hot plasma conditions for elemental analysis. The eluent from the IC column was introduced directly into the spray chamber of the ICP/MS. A time resolved elemental signal was obtained for bromine masses m/z 79 and m/z 81; however, m/z 79 was used for quantitation since it yielded a higher signal to noise ratio and improved sensitivity. The time resolved signal was converted into chromatographic data and analytes quantitated from extrapolation of peak areas against those of calibration standards. Separation and detection parameters are outlined in Table 1. Bromate and bromide showed excellent chromatographic resolution in both water and plasma samples using this method (Figs. 1 and 2).

Flow (mL/min) Injection loop (␮L) Bromate retention time (min) Bromide retention time (min) Analysis time (min) ICP/MS parameters Mode Quantitation mass (m/z) Acquisition dwell time (s/point) Sampling frequency (s)

Dionex 250 mm × 4 mm IonPac® AS9-HC 50 mm × 4 mm IonPac® AG9-HC Isocratic, 0.9 mM sodium bicarbonate 1.3 250 4.6 8.9 20 Time-resolved analysis, hot plasma conditions 79 0.5 1

2.4. Detection limit studies and calibration Instrument detection limits (IDL) were established from eight replicate measurements of bromate and bromide standards in water at 1 ␮g/L. The IDL was calculated by multiplying the standard deviation of the replicate measurements by the appropriate Student’s t-value for 7 degrees of freedom. Because plasma samples were extracted and diluted five-fold, the established method detection limit (MDL) in plasma was five times the IDL, in this case 1.3 and 1.6 ␮g/L plasma for bromate and bromide, respectively (Table 2). A conservative method reporting limit (MRL) of 5 ␮g/L was established for both bromate and bromide. A calibration curve was constructed from standards in reagent water with a linear range from 1 to 50 ␮g/L for bromate and 1 to 1000 ␮g/L for bromide. Linearity (r2 ) of all calibration curves exceeded 0.995.

Fig. 2. Chromatogram of rat plasma sample spiked with 0.32 ␮g/L bromate.


O. Qui˜nones et al. / Toxicology 221 (2006) 229–234

Table 2 Method detection limit studies at 1.0 ␮g/L Bromate (␮g/L)

Bromide (␮g/L)

Replicate #1 Replicate #2 Replicate #3 Replicate #4 Replicate #5 Replicate #6 Replicate #7 Replicate #8

1.13 0.97 1.00 0.86 1.04 0.98 1.09 1.10

0.94 0.90 1.04 0.91 1.03 1.12 1.13 1.18

Standard deviation Student’s t-value IDL MRL/lowest calibration point

0.09 3.00 0.26 1.00

0.11 3.00 0.32 1.00

Detection limits in plasma MDL in plasma MRL in plasma

1.3 5.0

1.6 5.0

ice. Samples were then stored at −20 ◦ C until extraction and analysis. Reagent water blanks, blood blanks (unspiked rat plasma), rat food and water, and heparin also were analyzed for bromide and bromate with each experiment. Results for bromate were below detection limits in every case, while bromide was found in nearly every sample except purified water blanks. Additionally, bromide and bromate stock solutions used for spiking rat blood and plasma were analyzed and the nominal concentration verified to be within 10% of the IC-ICP/MS calibration curve.

3. Results and discussion 3.1. Analytical observations

2.5. Plasma and whole blood spike studies Storage stability studies for bromate were carried out with purchased plasma, spiking samples at 0.03 mg/L and 0.30, aliquoting 1 mL samples and immediately freezing at −20 ◦ C. Frozen aliquots were thawed, extracted, and analyzed after 1, 3, and 7 days of storage. Bromate decay studies were carried out using 0.3 mL fresh blood aliquots from male Fischer 344 rats. Bromate stock solution was added to sample sets to yield one set with a spike concentration at 0.30 mg/L, and a second at 0.03 mg/L. Bromate was spiked into whole blood immediately after sampling and vortexed for 15 s. Samples were centrifuged at 0, 10, and 20 min after spiking, and at each time interval the plasma fraction was blast frozen in liquid nitrogen. Samples were stored at −80 ◦ C after freezing and expressed shipped on dry

The IC-ICP/MS method described here performed well for the detection of bromate and bromide in standards and rat plasma extracts following protein precipitation using acetonitrile. Extracts of bromate spikes in reagent water and rat plasma using a 1:5 final dilution of the acetonitrile matrix in water yielded average recoveries of over 85% for both analytes of interest (Tables 3 and 4). More extensive concentration of samples was investigated in order to achieve lower detection limits; however, recoveries suffered. Therefore, a conservative 1:5 dilution factor was used throughout this study. Bromate was not detected in unspiked rat plasma samples during method development, but bromide was consistently detected at 700–800 ␮g/L. Due to the high concentration of bromide in unspiked plasma, conversion of bromate into bromide at low levels was difficult to assess. Storage stability studies showed remarkable bromate stability over 7 days for both spike concentrations.

Table 3 Average results and spike recoveries for bromate (␮g/L) Sample


Spike amount

Bromate result

Avg. recovery (%)


Reagent water standard Extracted reagent water standard Unspiked rat plasma Rat plasma spike

6 4 3 6

10 5 0 5

9.54 5.16 n.d. 4.73

95.4 103.15 n/a 94.7

3.5 2.5 – 4.7

Table 4 Average results and spike recoveries for bromide (␮g/L) Sample


Spike amount

Bromide result

Avg. recovery (%)


Reagent water standard Extracted reagent water standard Unspiked rat plasma Rat plasma spike

6 4 3 2

10 100 0 100

10.1 109.5 805 893

101 109.5 n/a 88.4

5.5 1.6 2.4 1.2

O. Qui˜nones et al. / Toxicology 221 (2006) 229–234


at the analytical facility and plasma subsequently spiked with bromate, recoveries were consistently near 100% (Table 5). 3.2. Summary

Fig. 3. Bromate decay in spiked rat plasma samples during storage at −20◦ C.

Fig. 4. Bromate decay in fresh whole blood.

Analytical results established bromate loss under 20% for all spiked plasma samples analyzed over 7 days of storage (Fig. 3). In commercially obtained rat plasma samples, spiked bromate recovery exceeded 85% in all experiments (Table 1). Plasma fractions of whole blood samples spiked without delay at sampling and centrifuged immediately thereafter yielded bromate results close to 100% (Fig. 4). Conversely, whole blood spikes held for 10 and 20 min revealed rapid decay, with bromide nondetectable at 20 min for both spike concentrations (Fig. 4). However, when these samples were received

The method presented here provides a fast and sensitive technique for the analysis of bromate and bromide in blood. Bromide and bromate are more commonly analyzed using IC with electro-conductivity detection or colorimetric detection; however, these methods have limited sensitivity and are prone to false positives due to matrix interferences typical of blood plasma. To the best knowledge of the authors, no other method has been previously published on bromate and bromide measurements in blood using IC-ICP/MS. A 15 min isocratic elution of bromate and bromide combined with the high sensitivity and selectivity of the ICP-MS detector for elemental bromine yielded a simple yet robust method for analysis of bromate and bromide at trace levels. The large sample loading capacity of the IC column allows for the injection of increased sample volumes to achieve environmentally relevant detection limits. A simple precipitation-evaporation procedure without further sample cleanup worked adequately for this technique, yielding a total sample preparation and analysis time of about 30 min. Matrix spikes of purchased plasma and water samples showed excellent recovery and reproducibility, and no matrix effects were observed. Stored plasma samples (commercial and from fresh blood after storage) spiked with bromate remained relatively stable for 7 days when stored at −20 ◦ C; however, fresh plasma from blood spiked with bromate showed rapid decay of bromate to non-detectable levels when freezing was delayed at sampling. Therefore, to effectively apply this method, care must be taken at sampling to perform separation and freezing immediately to prevent bromate degradation. The discrepancy between fresh versus aged blood plasma in regards to bromate stability raises important issues about the mechanisms of bromate decay in blood. It also emphasizes the need for adequate sample preparation procedures and potential preservation approaches to

Table 5 Recovery results for bromate spikes in rat plasma blanks from live sampling events (␮g/L) Event no.

Plasma blank result

Spike concentration

Plasma spike result

Recovery (%)

1 2 3

<5.0 <5.0 <5.0

320.00 100.00 100

338.67 92.60 107.25

105 93 107


O. Qui˜nones et al. / Toxicology 221 (2006) 229–234

accurately measure levels in biological tissues reflective of conditions at sampling. While extensive work remains on methodologies and bromate metabolism research in general, the sampling and analytical strategies detailed in this method provide a viable and sensitive analytical tool to measure bromate in blood at levels that would yield environmentally relevant inferences. Future efforts will focus on determination of bromate metabolism in blood and definition of those factors which influence bromate loss during sample handling and extended storage. It is possible that the addition of a quenching agent may prevent decomposition after sampling. It is critical to determine the mechanisms of bromate decay in blood in order to develop pharmacokinetic models necessary for determining toxicological impacts of the relatively low levels of bromate that may occur in drinking water. Acknowledgements The authors are grateful to Libby Myers and Margarita Ortiz-Serrano at the University of Georgia for their extensive contributions to the in vivo portion of this project. We also thank Brett Vanderford, Janie Holady, Ira Racoma, and Anthony Dippuccio at the Southern Nevada Water Authority’s Water Quality Research & Development Division for their valuable analytical support to this research.

References Dunn, A.L., McIntyre, A.R., 1949. Detection of bromate in blood and urine. J. Lab. Clin. Med. 34, 425–427. Fujii, M., Oikawa, K., Saito, H., Fukuhara, C., Onosaka, S., Tanaka, K., 1984. Metabolism of potassium bromate in rats. Part 1. In vivo studies. Chemosphere 13, 1207–1212. Kage, S., Kudo, K., Ikeda, H., Tsujita, A., Ikeda, N., 2005. Determination of bromide in whole blood and urine from humans using gas-chromatography-mass spectrometry. J. Chromatogr. B 817, 335–339. Krasner, S.W., Gramith, J.T., Coffey, B.M., Yates, R.S., 1995. Impact of water quality and operational parameters on the formation and control of bromate during ozonation. Water Supply 13, 145– 156. Olszowy, H.A., Rossiter, J., Hegarty, J., Geoghegan, P., HaswellElkins, M., 1998. Background levels of bromide in human blood. J. Anal. Toxicol. 22, 225–230. Siddiqui, M.S., Amy, G.L., Rice, R.G., 1995. Bromate ion formation: a critical review. J. AWWA 87, 58–70. Song, R., Westerhoff, P., Minear, R., Amy, G., 1997. Bromate minimization during ozonation. J. AWWA 89, 69–78. Tanaka, K., Oikawa, K., Fukuhara, C., Saito, H., Onosaka, S., Min, K.-S., Fujii, M., 1984. Metabolism of potassium bromate in rats. Part II. In vitro studies. Chemosphere 13, 1213– 1219. United States Environmental Protection Agency (USEPA), 1998. Stage 1 Disinfectants and Disinfection Byproducts Rule. United States Environmental Protection Agency, Washington. Zager, R.A., Burkhart, K.M., 1998. Differential effects of glutathione and cysteine on Fe2+ , Fe3+ , H2 O2 and myoglobin-induced proximal tubular cell attack. Kidney Int. 53, 1661–1672.