- Email: [email protected]
Whole blood selenium determination by inductively coupled plasma mass spectrometry
CLB-09493; No. of pages: 4; 4C: Clinical Biochemistry xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Clinical Biochemistry journal homepage: www.elsevier.com/locate/clinbiochem
Whole blood selenium determination by inductively coupled plasma mass spectrometry Dustin R Bunch, Wendy Cieslak, Sihe Wang ⁎ Department of Laboratory Medicine, Cleveland Clinic, Cleveland, OH, USA
a r t i c l e
i n f o
Article history: Received 11 November 2016 Received in revised form 24 January 2017 Accepted 26 January 2017 Available online xxxx Keywords: Selenium ICP-MS Whole blood Reference interval
a b s t r a c t Objective: To develop a sensitive method for accurately measuring whole blood selenium and determining an appropriate reference interval for the local Cleveland population. Design and methods: The assay was developed and validated on an inductively coupled plasma mass spectrometry (ICP-MS) with a collision cell. Whole blood trace element free EDTA tubes were used to collect samples for the reference interval study (n = 50). Samples were collected after at least 8 h fast from healthy adults (76% females) with ages between 19 and 64 yr. Whole blood aliquots (1 mL) in acid washed cryogenic vials were stored at −70 °C until analysis. Results: The method passed the matrix effect, interference (except for Gd), and carryover tests. The method had a linear range of 0.2–7.1 μmol/L with accuracies of 87.1–118.1%. The total assay imprecision (CV) was b2.5% across the concentration levels tested. Comparison to another ICP-MS assay offered by an independent clinical lab yielded a Deming regression with a slope of 0.98, an intercept of 0.1 μmol/L, a standard error of estimate of 0.1 μmol/L, a correlation coefficient of 0.9846, and an average difference of 0.8%. The whole blood Se reference interval using a transformed parametric method was 2.2–3.5 μmol/L. Conclusions: This whole blood Se ICP-MS methodology is sensitive and acceptable for patient testing. © 2017 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
1. Introduction Selenium (Se) is a micronutrient, important for cardiovascular and fertility health as well as thyroid function and possibly cancer prevention [1–4]. Se is integrated into numerous selenoproteins and has antioxidant effects [5]. Se concentrations vary greatly in human blood [6] and are dependent on either the regional availability of Se which dictates the element content in foods or supplementation by the individual persons or farms [6–9]. The richest sources of Se include seafood, meat, cereals and grains. Both excessive and insufficient intake of Se can have health implications. Low Se intake may increase certain cancers, increase incidence of cardiovascular disease (e.g. Keshan), weaken the immune system, impair growth, increase thyroid dysfunction, and impair fertility in men [1–5]. Deficiency is not only due to insufficient intake but can also be due to loss through procedures such as hemodialysis
Abbreviations: CV, coefficient of variation; ICP-MS, inductively coupled plasma mass spectrometry; IS, internal standard. ⁎ Corresponding author at: Department of Laboratory Medicine, Mail Code LL3-140, 9500 Euclid Ave, Cleveland, OH 44195, USA. E-mail address: [email protected] (S. Wang).
[10,11]. Conversely, Se toxicity includes gastrointestinal upsets, hair and nail loss, tooth decay, liver failure, skin lesions, fatigue and damage to the nervous system [12,13]. Currently, there are controversies over the appropriate amounts of Se to consume [14,15]. Historically, atomic absorption spectrometry was used widely to determine metal elements in various matrices. In recent years, inductively coupled plasma mass spectrometry (ICP-MS) has become the method of choice for measuring trace elements [16,17]. Initially, Se measurement by ICP-MS was plagued with polyatomic interferences, which are an artifact of the ionization techniques employed. Polyatomic interferences are usually due to the sample matrix, sample diluent, and/or the inert gas used during ionization. An example of a polyatomic interference is argon (40Ar+ 2 ) which interferes with the most abundant selenium isotope (80Se+). Technology improvements such as magnetic sector filter [18], increased mass spectrometry resolution, and reaction cells, have decreased the occurrence of interferences [19]. However, not all interferences can be eliminated. For example gadolinium, a contrast agent for MRI studies, is known to cause interference for Se measurement by ICP-MS [20–23]. The primary goal of this study was to develop a highly sensitive and specific ICP-MS assay to monitor Se levels in whole blood. Polyatomic interferences were minimized by precise control of the collision cell
http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013 0009-9120/© 2017 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
Please cite this article as: D.R. Bunch, et al., Whole blood selenium determination by inductively coupled plasma mass spectrometry, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013
2
D.R. Bunch et al. / Clinical Biochemistry xxx (2017) xxx–xxx
and kinetic energy discrimination parameters. The secondary goal was to determine a reference interval for the local population. 2. Materials and methods 2.1. Chemical, reagents and solutions Optima Nitric Acid was purchased from Fisher Scientific (Pittsburgh, PA, USA). Type 1 water was produced by a Millipore Synergy System (Billerica, MA, USA). Se (12.7 mmol/L), the internal standard (IS) germanium (1.38 mmol/L; Ge), and gadolinium (Gd) were purchased from VHG Labs (Manchester, NH, USA). The Se concentration was traceable to the NIST SRM 3149. The argon (4.8 ICP grade) and helium (6.0 research grade) were purchased from Praxair (Danbury, CT, USA). ClinChek quality controls were purchased from Iris Technologies (Munich, Germany). Stocks and calibration standards for Se were prepared in 0.1% nitric acid. Calibration standards were prepared at 0.2 μmol/L, 0.3 μmol/L, 0.6 μmol/L, 1.3 μmol/L, 2.5 μmol/L, and 3.8 μmol/L, while IS solution contained 0.7 μmol/L Ge in 0.1% nitric acid. Acid washed polypropylene containers were used to store all solutions at room temperature until use. 2.2. Sample preparation Into a 15 mL acid-washed polypropylene conical tube, 1 mL of patient whole blood, quality control, or calibration standard and 9 mL 0.1% nitric acid were added and vortex mixed. After centrifugation at 3500 ×g for 5 min, the supernatant was infused to the ICP-MS system via a 1.0 mL continuous sample loop configured with an Elemental Scientific Autosampler (Omaha, NE). IS solution was introduced through a high-flow tee connection (0.25 mm i.d.) prior to the ICP-MS. 2.3. ICP-MS method Instrumentation and instrument settings were similar to a previous publication [24]. A Thermo Fisher X Series2 ICP-MS system configured with a collision cell was used to develop this assay. Instrument control software was PlasmaLab (ver. 2.6.1.355, ThermoFisher). Sample was introduced through a quartz cyclonic spray chamber cooled to 3 °C and was nebulized through a peristaltic pump at 13% of max speed and a Teflon/carbon fiber probe (150 cm capillary; 1.0 mm i.d.). The plasma (RF power 1400 W) was formed using a quartz torch with a Ni sampler and skimmer cone. The 78Se and 72Gd ions were chosen for quantification. Instrument tuning was performed to reach the following: N 100,000 counts for indium, b 1500 counts for Ar (nebulizer gas; 1.02 mL/min), and a b 0.0200 ratio for cerium oxide over cerium. A flow rate of 3.5– 5.0 mL/min of 100% helium was used in the collision cell. The kinetic energy discrimination was set at 4.5 V and standard mass resolution was set at 0.02 amu. Peak jump scan mode was used for data acquisition with a dwell time of 100.00 ms and 1500.00 ms for Ge and Se, respectively. Each sample had 3 readings with 20 sweeps. 2.4. Method validation Evaluation of matrix effects was performed through extraction and injection of 6 patient samples, a candidate matrix of 0. 1% nitric acid solution spiked with Se at 1.3 μmol/L, and 1:1 mixtures of the patient samples with the spiked candidate matrix. The passing criterion was that each measured 1:1 mixture value had to be within 20% of the mean of the corresponding patient and spiked candidate matrix solution. This matrix was determined to be appropriate for preparing calibrators and specimen dilution. Next, common endogenous ions were also investigated for interference using two distinct patient pools with Se levels at a low (1.4 μmol/L)
and high (2.8 μmol/L) concentration. Extractions were performed for a 1:1 mix of each patient pool and the dilution matrix, and a 1:1 mix of each patient pool and dilution matrix spiked with calcium, iron, sodium, chloride, carbon, aluminum and magnesium at 7.5 mmol/L, 4.5 mmol/L, 28.3 mmol/L, 56.4 mmol/L, 16.7 mmol/L, 3.7 mmol/L and 4.1 mmol/L, respectfully. Interference was considered to be within acceptable limits if the result for the 1:1 mix of the patient pool and the spiked dilution matrix was within 20% of the 1:1 mix of patient pool and the un-spiked dilution matrix for both levels tested. Using the multicomponent mix, the sodium and chlorine levels were below the physiological ranges. An additional study was performed with sodium and chloride and Gd due to information available in the literature [20–23]. Gd was spiked at 6.3, 62.5, and 3125 μmol/L in 0.1% nitric acid and was analyzed for Se. Sodium (43.4 mmol/L) and chlorine (282.0 mmol/L) in 0.1% nitric acid were analyzed undiluted for Se. The high linearity sample was a spiked whole blood pool, which was serially diluted using 0.1% nitric acid to the appropriate levels. The blank pool was at 1.9 μmol/L. There were totally nine levels in this experiment: 6.3, 5.4, 3.6, 1.8, 1.4, 0.9, 0.5, 0.4, and 0.2 μmol/L. The resulting samples were extracted and analyzed in triplicate. Both analytical recovery and imprecision were evaluated for each level. Lower limit of quantification was defined by the lowest concentration within the linearity study that met the following two parameters: 1. Accuracy between 80% and 120%; 2. Coefficient variation (CV) was b20%. The ideal total error for the assay should be within 14.1% based on biological variations [25]. To evaluate the assay carryover, a high sample and a low sample were extracted in triplicate and analyzed in the sequence of low1high-low2. Low2 was a re-injection of low1. Significant carryover was defined by the mean low2 value ≥ 20% from the mean low1 value or the mean low2 value is N 3 standard deviations of the mean low1. The standard deviation was calculated from the low1 results. The EP10-A3 guideline from Clinical Laboratory and Standards Institute (Wayne, PA, USA) was modified to perform precision evaluation. Intra-assay and total assay CV were calculated by assaying 3 levels of patient derived samples twice a day for 5 days in the sequence of mid-hi-low-midmid-low-low-hi-hi-mid. This method was also compared with an ICPMS method offered by an independent clinical lab using leftover patient specimens (n = 30). 2.5. Statistics Statistical analysis was performed using EP Evaluator Release 10 (Data Innovations, South Burlington, VT, USA) and Excel 2010 (Microsoft, Redmond WA, USA). 2.6. Reference interval sample collection The Cleveland Clinic Institutional Review Board approved the collection of blood samples for determining the reference interval. Briefly, whole blood samples (n = 50) were collected after a minimum of 8 h fasting in EDTA vacutainer tubes free of trace elements from healthy adults (76% females) with ages between 19 and 64 yr (mean: 38.8). Stainless steel metal needle was used for the collection; however, approximately 35 mL of blood was collected prior to the trace metal tubes. Exclusion criteria for the patients were [24]: body mass index below 15 or above 30; having a cold, flu, virus or other infection in the past two weeks; chemotherapy in the past year or current immunosuppressant therapy; pregnancy; a diabetes diagnosis; diagnosis of malabsorption; gastric or intestinal surgery, or frequent diarrhea. Aliquots of the blood samples were placed in acid washed cryogenic vials and stored at − 70 °C until analysis. Additionally, patients samples (n = 20), typically analyzed in the lab, were collected. Through a patient chart review, it was identified that these typical patient samples all had some types of malnutrition. Central 95% were calculated for reference range samples and the patient samples, separately.
Please cite this article as: D.R. Bunch, et al., Whole blood selenium determination by inductively coupled plasma mass spectrometry, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013
D.R. Bunch et al. / Clinical Biochemistry xxx (2017) xxx–xxx
3
3. Results 3.1. Assay validation With a mean difference of −6.7 ± 1.1% between the measured and calculated mixture values (n = 6), the matrix effect test was passed for this assay. This allowed 0.1% nitric acid to be used as a diluent and the matrix for preparing calibrators. Observed interference was b4% for the following components: calcium (7.5 mmol/L), iron (4.5 mmol/L), sodium (434.5 mmol/L), chloride (2820.6 mmol/L), carbon (16.7 mmol/L), aluminum (3.7 mmol/L), or magnesium (4.1 mmol/L). Gd significantly interfered with the assay with measured Se concentrations double the Gd concentrations spiked. The determined linearity was 0.2–7.1 μmol/L and had accuracies from 87.1–118.1%. No significant carryover was detected at 9.2 μmol/L. The intra-assay CV was b 1.6% and total assay CV was b2.5% at concentrations of 0.9 to 2.4 μmol/L, which were well within the 6.0% desirable imprecision based on biological variation. Deming regression of the patient results (n = 30) between this method and an independent ICP-MS assay showed a slope of 0.98 (0.91 to 1.04), an intercept of 0.1 μmol/L (− 0.1 to 0.2), a standard error of estimate of 0.1 μmol/L, a correlation coefficient of 0.9846, and an average difference of 0.8% (Fig. 1). If we use the difference from our method comparison (0.8%) as the method bias and the maximal imprecision (2.5%) found in this study, the total error at 2 standard deviations was 5.8%, which is well within the desirable total error of 14.1% calculated based on biological variations [25]. 3.2. Reference intervals A log transformation was performed for the reference interval sample results (n = 50) followed by parametric statistical calculations to determine the central 95% for the reference interval. The reference interval was determined to be 2.2–3.5 μmol/L with a median of 2.7 μmol/L (Fig. 2). The typical patient samples (n = 20) had a significantly lower and broader range (0.7–3.0 μmol/L) with a median of 1.8 μmol/L. 4. Discussion Se is typically measured to monitor the nutritional status for patients with transplants and/or on total parenteral nutrition. The Se reference interval (2.2–3.5 μmol/L) determined in this study is slightly higher than both the 0.41–3.2 μmol/L range reported for the world in 1996 [6] and the 0.74–2.97 μmol/L range as reported in Tietz [26]. However,
Fig. 2. Histogram of whole blood Se from a reference population in Cleveland, Ohio.
the patient samples with malnutrition status included in this study had a range of 0.7–3.0 μmol/L, which would fit within the 1996 reported world range. It was also reported in 1996 that central USA had higher Se levels than other regions of the world [6] and may explain the general higher level seen in this study. Also, it is worth noting that in recent years there has been an increased health awareness that includes consuming more whole grains [27], a Se rich food. As a whole, the broader distribution found in the malnourished patients versus the normal patients is to be expected. In our reference study we did not exclude participants based on supplement consumption. The required sample volume for this assay was 1 mL, which would make it difficult to get enough whole blood from specific patient populations such as children and the elderly. To address this issue, 0.5 mL at the same dilution was assessed and found to be acceptable without further modification of the assay. There is a possibility of improving this further through modification of the tubing internal diameter for sample delivery. The major limitations of this study were the reference interval determination and the significant interference from Gd. When computing the Se reference interval, age and sex were examined but no statistical differences were identified. The reference interval population was limited in number (n = 50) due to reference subject recruiting difficulties, possessed a skewed male to female ratio (12 males and 38 females), and included only adults. Gd, a common MRI contrast reagent, caused significant interference for this method. Gd interference may be eliminated by using 100% hydrogen gas in the collision cell [21]. However, at this time our facilities are not equipped to work with hydrogen gas.
Fig. 1. Method comparison. A. Scatter plot showing correlation of this ICP-MS method and the independent ICP-MS method [YICP-MS = 0.98(0.91 to 1.04) XInd + 0.1(0.1 to 0.2)]. B. Difference plot illustrating absolute individual specimen difference. C. Percent difference plot.
Please cite this article as: D.R. Bunch, et al., Whole blood selenium determination by inductively coupled plasma mass spectrometry, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013
4
D.R. Bunch et al. / Clinical Biochemistry xxx (2017) xxx–xxx
To work-around the Gd interference, physicians are requested to refrain from ordering a whole blood Se within 3 days of a Gd infusion. Future work could take into account a recent publication, which compared online and offline addition of internal standard and showed that both were acceptable; however, imprecision and bias were improved with the offline addition [28]. For this assay the comparison was not performed, but this latter approach will be studied for new trace metal assays in our laboratory. 5. Conclusions Total whole blood Se was successfully developed with a sample preparation involving only 1:9 dilution with 0.1% nitric acid. This ICPMS methodology is sensitive with the ability to quantify Se from 0.2 to 7.1 μmol/L. The assay was also used to determine the Se reference interval in a local reference population, which was found to be 2.2–3.5 μmol/ L. Ethical approval Collection of samples for this study was approved by the Institutional Review Board. Author contributions All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. Authors' disclosures of potential conflicts of interest No authors declared any potential conflicts of interest. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.clinbiochem.2017.01.013. References [1] W. Saliba, R. El Fakih, W. Shaheen, Heart failure secondary to selenium deficiency, reversible after supplementation, Int. J. Cardiol. 141 (2010) e26–e27. [2] M.R. Safarinejad, S. Safarinejad, Efficacy of selenium and/or n-acetyl-cysteine for improving semen parameters in infertile men: a double-blind, placebo controlled, randomized study, J. Urol. 181 (2009) 741–751. [3] K. Ashton, L. Hooper, L.J. Harvey, R. Hurst, A. Casgrain, S.J. Fairweather-Tait, Methods of assessment of selenium status in humans: a systematic review, Am. J. Clin. Nutr. 89 (2009) 2025S–2039S. [4] M.B. Zimmermann, J. Kohrle, The impact of iron and selenium deficiencies on iodine and thyroid metabolism: biochemistry and relevance to public health, Thyroid 12 (2002) 867–878. [5] V.M. Labunskyy, D.L. Hatfield, V.N. Gladyshev, Selenoproteins: molecular pathways and physiological roles, Physiol. Rev. 94 (2014) 739–777.
[6] G. Alfthan, J. Neve, Reference values for serum selenium in various areas evaluated according to the tracy protocol, J. Trace Elem. Med. Biol. 10 (1996) 77–87. [7] G. Alfthan, M. Eurola, P. Ekholm, E.R. Venalainen, T. Root, K. Korkalainen, et al., Effects of nationwide addition of selenium to fertilizers on foods, and animal and human health in Finland: from deficiency to optimal selenium status of the population, J. Trace Elem. Med. Biol. 31 (2015) 142–147. [8] J.L. Slavin, D. Jacobs, L. Marquart, K. Wiemer, The role of whole grains in disease prevention, J. Am. Diet. Assoc. 101 (2001) 780–785. [9] K.H. Schulpis, T. Karakonstantakis, S. Gavrili, G. Chronopoulou, G.A. Karikas, G. Vlachos, et al., Maternal—neonatal serum selenium and copper levels in Greeks and Albanians, Eur. J. Clin. Nutr. 58 (2004) 1314–1318. [10] M. Tonelli, N. Wiebe, B. Hemmelgarn, S. Klarenbach, C. Field, B. Manns, et al., Trace elements in hemodialysis patients: a systematic review and meta-analysis, BMC Med. 7 (2009) 25. [11] M.D. Churchwell, D.A. Pasko, I.F. Btaiche, J.C. Jain, B.A. Mueller, Trace element removal during in vitro and in vivo continuous haemodialysis, Nephrol. Dial. Transplant. 22 (2007) 2970–2977. [12] H. Steinbrenner, H. Sies, Selenium homeostasis and antioxidant selenoproteins in brain: implications for disorders in the central nervous system, Arch. Biochem. Biophys. 536 (2013) 152–157. [13] M. Vinceti, E.T. Wei, C. Malagoli, M. Bergomi, G. Vivoli, Adverse health effects of selenium in humans, Rev. Environ. Health 16 (2001) 233–251. [14] M. Vinceti, T. Maraldi, M. Bergomi, C. Malagoli, Risk of chronic low-dose selenium overexposure in humans: insights from epidemiology and biochemistry, Rev. Environ. Health 24 (2009) 231–248. [15] E. Jablonska, M. Vinceti, Selenium and human health: witnessing a Copernican revolution? J. Environ. Sci. Health C Environ. Carcinog. Ecotoxicol. Rev. 33 (2015) 328–368. [16] S. Chan, B. Gerson, R.E. Reitz, S.A. Sadjadi, Technical and clinical aspects of spectrometric analysis of trace elements in clinical samples, Clin. Lab. Med. 18 (1998) 615–629. [17] T.M.T. Sheehan, D.J. Halls, Measurement of selenium in clinical specimens, Ann. Clin. Biochem. Int. J. Biochem. Med. 36 (1999) 301–315. [18] A.M. Featherstone, A.T. Townsend, G.A. Jacobson, G.M. Peterson, Comparison of methods for the determination of total selenium in plasma by magnetic sector inductively coupled plasma mass spectrometry, Anal. Chim. Acta 512 (2004) 319–327. [19] J.M. Harrington, D.J. Young, A.S. Essader, S.J. Sumner, K.E. Levine, Analysis of human serum and whole blood for mineral content by icp-ms and icp-oes: development of a mineralomics method, Biol. Trace Elem. Res. 160 (2014) 132–142. [20] L. Liu, K. Tamama, O.M. Peck Palmer, Selenium overload? Clin. Chem. 62 (2016) 658. [21] C.F. Harrington, A. Walter, S. Nelms, A. Taylor, Removal of the gadolinium interference from the measurement of selenium in human serum by use of collision cell quadrupole inductively coupled plasma mass spectrometry (q-icp-ms), Ann. Clin. Biochem. 51 (2014) 386–391. [22] G. Lippi, M. Daves, C. Mattiuzzi, Interference of medical contrast media on laboratory testing, Biochem. Med. (Zagreb). 24 (2014) 80–88. [23] A. Walter, S. Nelms, C.F. Harrington, A. Taylor, Interference of gadolinium on the measurement of selenium in human serum by inductively coupled plasma-quadrupole mass spectrometry, Ann. Clin. Biochem. 48 (2011) 176–177. [24] W. Cieslak, K. Pap, D.R. Bunch, E. Reineks, R. Jackson, R. Steinle, et al., Highly sensitive measurement of whole blood chromium by inductively coupled plasma mass spectrometry, Clin. Biochem. 46 (2013) 266–270. [25] J. Westgard, Desirable Biological Variation Database Specifications, https://www. westgard.com/biodatabase1.htm2014. [26] C.A. Burtis, E.R. Ashwood, D.E. Bruns, N.W. Tietz, Tietz Textbook of Clinical Chemistry and Molecular Diagnostics, fifth ed. Saunders, St. Louis, Mo., 2013 [27] A.M. Albertson, M. Reicks, N. Joshi, C.K. Gugger, Whole grain consumption trends and associations with body weight measures in the United States: results from the cross sectional national health and nutrition examination survey 2001–2012, Nutr. J. 15 (2015) 8. [28] J.T. Collie, R.J. Massie, O.A. Jones, P.D. Morrison, R.F. Greaves, A candidate reference method using icp-ms for sweat chloride quantification, Clin. Chem. Lab. Med. 54 (2016) 561–567.
Please cite this article as: D.R. Bunch, et al., Whole blood selenium determination by inductively coupled plasma mass spectrometry, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013
Contents lists available at ScienceDirect
Clinical Biochemistry journal homepage: www.elsevier.com/locate/clinbiochem
Whole blood selenium determination by inductively coupled plasma mass spectrometry Dustin R Bunch, Wendy Cieslak, Sihe Wang ⁎ Department of Laboratory Medicine, Cleveland Clinic, Cleveland, OH, USA
a r t i c l e
i n f o
Article history: Received 11 November 2016 Received in revised form 24 January 2017 Accepted 26 January 2017 Available online xxxx Keywords: Selenium ICP-MS Whole blood Reference interval
a b s t r a c t Objective: To develop a sensitive method for accurately measuring whole blood selenium and determining an appropriate reference interval for the local Cleveland population. Design and methods: The assay was developed and validated on an inductively coupled plasma mass spectrometry (ICP-MS) with a collision cell. Whole blood trace element free EDTA tubes were used to collect samples for the reference interval study (n = 50). Samples were collected after at least 8 h fast from healthy adults (76% females) with ages between 19 and 64 yr. Whole blood aliquots (1 mL) in acid washed cryogenic vials were stored at −70 °C until analysis. Results: The method passed the matrix effect, interference (except for Gd), and carryover tests. The method had a linear range of 0.2–7.1 μmol/L with accuracies of 87.1–118.1%. The total assay imprecision (CV) was b2.5% across the concentration levels tested. Comparison to another ICP-MS assay offered by an independent clinical lab yielded a Deming regression with a slope of 0.98, an intercept of 0.1 μmol/L, a standard error of estimate of 0.1 μmol/L, a correlation coefficient of 0.9846, and an average difference of 0.8%. The whole blood Se reference interval using a transformed parametric method was 2.2–3.5 μmol/L. Conclusions: This whole blood Se ICP-MS methodology is sensitive and acceptable for patient testing. © 2017 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
1. Introduction Selenium (Se) is a micronutrient, important for cardiovascular and fertility health as well as thyroid function and possibly cancer prevention [1–4]. Se is integrated into numerous selenoproteins and has antioxidant effects [5]. Se concentrations vary greatly in human blood [6] and are dependent on either the regional availability of Se which dictates the element content in foods or supplementation by the individual persons or farms [6–9]. The richest sources of Se include seafood, meat, cereals and grains. Both excessive and insufficient intake of Se can have health implications. Low Se intake may increase certain cancers, increase incidence of cardiovascular disease (e.g. Keshan), weaken the immune system, impair growth, increase thyroid dysfunction, and impair fertility in men [1–5]. Deficiency is not only due to insufficient intake but can also be due to loss through procedures such as hemodialysis
Abbreviations: CV, coefficient of variation; ICP-MS, inductively coupled plasma mass spectrometry; IS, internal standard. ⁎ Corresponding author at: Department of Laboratory Medicine, Mail Code LL3-140, 9500 Euclid Ave, Cleveland, OH 44195, USA. E-mail address: [email protected] (S. Wang).
[10,11]. Conversely, Se toxicity includes gastrointestinal upsets, hair and nail loss, tooth decay, liver failure, skin lesions, fatigue and damage to the nervous system [12,13]. Currently, there are controversies over the appropriate amounts of Se to consume [14,15]. Historically, atomic absorption spectrometry was used widely to determine metal elements in various matrices. In recent years, inductively coupled plasma mass spectrometry (ICP-MS) has become the method of choice for measuring trace elements [16,17]. Initially, Se measurement by ICP-MS was plagued with polyatomic interferences, which are an artifact of the ionization techniques employed. Polyatomic interferences are usually due to the sample matrix, sample diluent, and/or the inert gas used during ionization. An example of a polyatomic interference is argon (40Ar+ 2 ) which interferes with the most abundant selenium isotope (80Se+). Technology improvements such as magnetic sector filter [18], increased mass spectrometry resolution, and reaction cells, have decreased the occurrence of interferences [19]. However, not all interferences can be eliminated. For example gadolinium, a contrast agent for MRI studies, is known to cause interference for Se measurement by ICP-MS [20–23]. The primary goal of this study was to develop a highly sensitive and specific ICP-MS assay to monitor Se levels in whole blood. Polyatomic interferences were minimized by precise control of the collision cell
http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013 0009-9120/© 2017 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
Please cite this article as: D.R. Bunch, et al., Whole blood selenium determination by inductively coupled plasma mass spectrometry, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013
2
D.R. Bunch et al. / Clinical Biochemistry xxx (2017) xxx–xxx
and kinetic energy discrimination parameters. The secondary goal was to determine a reference interval for the local population. 2. Materials and methods 2.1. Chemical, reagents and solutions Optima Nitric Acid was purchased from Fisher Scientific (Pittsburgh, PA, USA). Type 1 water was produced by a Millipore Synergy System (Billerica, MA, USA). Se (12.7 mmol/L), the internal standard (IS) germanium (1.38 mmol/L; Ge), and gadolinium (Gd) were purchased from VHG Labs (Manchester, NH, USA). The Se concentration was traceable to the NIST SRM 3149. The argon (4.8 ICP grade) and helium (6.0 research grade) were purchased from Praxair (Danbury, CT, USA). ClinChek quality controls were purchased from Iris Technologies (Munich, Germany). Stocks and calibration standards for Se were prepared in 0.1% nitric acid. Calibration standards were prepared at 0.2 μmol/L, 0.3 μmol/L, 0.6 μmol/L, 1.3 μmol/L, 2.5 μmol/L, and 3.8 μmol/L, while IS solution contained 0.7 μmol/L Ge in 0.1% nitric acid. Acid washed polypropylene containers were used to store all solutions at room temperature until use. 2.2. Sample preparation Into a 15 mL acid-washed polypropylene conical tube, 1 mL of patient whole blood, quality control, or calibration standard and 9 mL 0.1% nitric acid were added and vortex mixed. After centrifugation at 3500 ×g for 5 min, the supernatant was infused to the ICP-MS system via a 1.0 mL continuous sample loop configured with an Elemental Scientific Autosampler (Omaha, NE). IS solution was introduced through a high-flow tee connection (0.25 mm i.d.) prior to the ICP-MS. 2.3. ICP-MS method Instrumentation and instrument settings were similar to a previous publication [24]. A Thermo Fisher X Series2 ICP-MS system configured with a collision cell was used to develop this assay. Instrument control software was PlasmaLab (ver. 2.6.1.355, ThermoFisher). Sample was introduced through a quartz cyclonic spray chamber cooled to 3 °C and was nebulized through a peristaltic pump at 13% of max speed and a Teflon/carbon fiber probe (150 cm capillary; 1.0 mm i.d.). The plasma (RF power 1400 W) was formed using a quartz torch with a Ni sampler and skimmer cone. The 78Se and 72Gd ions were chosen for quantification. Instrument tuning was performed to reach the following: N 100,000 counts for indium, b 1500 counts for Ar (nebulizer gas; 1.02 mL/min), and a b 0.0200 ratio for cerium oxide over cerium. A flow rate of 3.5– 5.0 mL/min of 100% helium was used in the collision cell. The kinetic energy discrimination was set at 4.5 V and standard mass resolution was set at 0.02 amu. Peak jump scan mode was used for data acquisition with a dwell time of 100.00 ms and 1500.00 ms for Ge and Se, respectively. Each sample had 3 readings with 20 sweeps. 2.4. Method validation Evaluation of matrix effects was performed through extraction and injection of 6 patient samples, a candidate matrix of 0. 1% nitric acid solution spiked with Se at 1.3 μmol/L, and 1:1 mixtures of the patient samples with the spiked candidate matrix. The passing criterion was that each measured 1:1 mixture value had to be within 20% of the mean of the corresponding patient and spiked candidate matrix solution. This matrix was determined to be appropriate for preparing calibrators and specimen dilution. Next, common endogenous ions were also investigated for interference using two distinct patient pools with Se levels at a low (1.4 μmol/L)
and high (2.8 μmol/L) concentration. Extractions were performed for a 1:1 mix of each patient pool and the dilution matrix, and a 1:1 mix of each patient pool and dilution matrix spiked with calcium, iron, sodium, chloride, carbon, aluminum and magnesium at 7.5 mmol/L, 4.5 mmol/L, 28.3 mmol/L, 56.4 mmol/L, 16.7 mmol/L, 3.7 mmol/L and 4.1 mmol/L, respectfully. Interference was considered to be within acceptable limits if the result for the 1:1 mix of the patient pool and the spiked dilution matrix was within 20% of the 1:1 mix of patient pool and the un-spiked dilution matrix for both levels tested. Using the multicomponent mix, the sodium and chlorine levels were below the physiological ranges. An additional study was performed with sodium and chloride and Gd due to information available in the literature [20–23]. Gd was spiked at 6.3, 62.5, and 3125 μmol/L in 0.1% nitric acid and was analyzed for Se. Sodium (43.4 mmol/L) and chlorine (282.0 mmol/L) in 0.1% nitric acid were analyzed undiluted for Se. The high linearity sample was a spiked whole blood pool, which was serially diluted using 0.1% nitric acid to the appropriate levels. The blank pool was at 1.9 μmol/L. There were totally nine levels in this experiment: 6.3, 5.4, 3.6, 1.8, 1.4, 0.9, 0.5, 0.4, and 0.2 μmol/L. The resulting samples were extracted and analyzed in triplicate. Both analytical recovery and imprecision were evaluated for each level. Lower limit of quantification was defined by the lowest concentration within the linearity study that met the following two parameters: 1. Accuracy between 80% and 120%; 2. Coefficient variation (CV) was b20%. The ideal total error for the assay should be within 14.1% based on biological variations [25]. To evaluate the assay carryover, a high sample and a low sample were extracted in triplicate and analyzed in the sequence of low1high-low2. Low2 was a re-injection of low1. Significant carryover was defined by the mean low2 value ≥ 20% from the mean low1 value or the mean low2 value is N 3 standard deviations of the mean low1. The standard deviation was calculated from the low1 results. The EP10-A3 guideline from Clinical Laboratory and Standards Institute (Wayne, PA, USA) was modified to perform precision evaluation. Intra-assay and total assay CV were calculated by assaying 3 levels of patient derived samples twice a day for 5 days in the sequence of mid-hi-low-midmid-low-low-hi-hi-mid. This method was also compared with an ICPMS method offered by an independent clinical lab using leftover patient specimens (n = 30). 2.5. Statistics Statistical analysis was performed using EP Evaluator Release 10 (Data Innovations, South Burlington, VT, USA) and Excel 2010 (Microsoft, Redmond WA, USA). 2.6. Reference interval sample collection The Cleveland Clinic Institutional Review Board approved the collection of blood samples for determining the reference interval. Briefly, whole blood samples (n = 50) were collected after a minimum of 8 h fasting in EDTA vacutainer tubes free of trace elements from healthy adults (76% females) with ages between 19 and 64 yr (mean: 38.8). Stainless steel metal needle was used for the collection; however, approximately 35 mL of blood was collected prior to the trace metal tubes. Exclusion criteria for the patients were [24]: body mass index below 15 or above 30; having a cold, flu, virus or other infection in the past two weeks; chemotherapy in the past year or current immunosuppressant therapy; pregnancy; a diabetes diagnosis; diagnosis of malabsorption; gastric or intestinal surgery, or frequent diarrhea. Aliquots of the blood samples were placed in acid washed cryogenic vials and stored at − 70 °C until analysis. Additionally, patients samples (n = 20), typically analyzed in the lab, were collected. Through a patient chart review, it was identified that these typical patient samples all had some types of malnutrition. Central 95% were calculated for reference range samples and the patient samples, separately.
Please cite this article as: D.R. Bunch, et al., Whole blood selenium determination by inductively coupled plasma mass spectrometry, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013
D.R. Bunch et al. / Clinical Biochemistry xxx (2017) xxx–xxx
3
3. Results 3.1. Assay validation With a mean difference of −6.7 ± 1.1% between the measured and calculated mixture values (n = 6), the matrix effect test was passed for this assay. This allowed 0.1% nitric acid to be used as a diluent and the matrix for preparing calibrators. Observed interference was b4% for the following components: calcium (7.5 mmol/L), iron (4.5 mmol/L), sodium (434.5 mmol/L), chloride (2820.6 mmol/L), carbon (16.7 mmol/L), aluminum (3.7 mmol/L), or magnesium (4.1 mmol/L). Gd significantly interfered with the assay with measured Se concentrations double the Gd concentrations spiked. The determined linearity was 0.2–7.1 μmol/L and had accuracies from 87.1–118.1%. No significant carryover was detected at 9.2 μmol/L. The intra-assay CV was b 1.6% and total assay CV was b2.5% at concentrations of 0.9 to 2.4 μmol/L, which were well within the 6.0% desirable imprecision based on biological variation. Deming regression of the patient results (n = 30) between this method and an independent ICP-MS assay showed a slope of 0.98 (0.91 to 1.04), an intercept of 0.1 μmol/L (− 0.1 to 0.2), a standard error of estimate of 0.1 μmol/L, a correlation coefficient of 0.9846, and an average difference of 0.8% (Fig. 1). If we use the difference from our method comparison (0.8%) as the method bias and the maximal imprecision (2.5%) found in this study, the total error at 2 standard deviations was 5.8%, which is well within the desirable total error of 14.1% calculated based on biological variations [25]. 3.2. Reference intervals A log transformation was performed for the reference interval sample results (n = 50) followed by parametric statistical calculations to determine the central 95% for the reference interval. The reference interval was determined to be 2.2–3.5 μmol/L with a median of 2.7 μmol/L (Fig. 2). The typical patient samples (n = 20) had a significantly lower and broader range (0.7–3.0 μmol/L) with a median of 1.8 μmol/L. 4. Discussion Se is typically measured to monitor the nutritional status for patients with transplants and/or on total parenteral nutrition. The Se reference interval (2.2–3.5 μmol/L) determined in this study is slightly higher than both the 0.41–3.2 μmol/L range reported for the world in 1996 [6] and the 0.74–2.97 μmol/L range as reported in Tietz [26]. However,
Fig. 2. Histogram of whole blood Se from a reference population in Cleveland, Ohio.
the patient samples with malnutrition status included in this study had a range of 0.7–3.0 μmol/L, which would fit within the 1996 reported world range. It was also reported in 1996 that central USA had higher Se levels than other regions of the world [6] and may explain the general higher level seen in this study. Also, it is worth noting that in recent years there has been an increased health awareness that includes consuming more whole grains [27], a Se rich food. As a whole, the broader distribution found in the malnourished patients versus the normal patients is to be expected. In our reference study we did not exclude participants based on supplement consumption. The required sample volume for this assay was 1 mL, which would make it difficult to get enough whole blood from specific patient populations such as children and the elderly. To address this issue, 0.5 mL at the same dilution was assessed and found to be acceptable without further modification of the assay. There is a possibility of improving this further through modification of the tubing internal diameter for sample delivery. The major limitations of this study were the reference interval determination and the significant interference from Gd. When computing the Se reference interval, age and sex were examined but no statistical differences were identified. The reference interval population was limited in number (n = 50) due to reference subject recruiting difficulties, possessed a skewed male to female ratio (12 males and 38 females), and included only adults. Gd, a common MRI contrast reagent, caused significant interference for this method. Gd interference may be eliminated by using 100% hydrogen gas in the collision cell [21]. However, at this time our facilities are not equipped to work with hydrogen gas.
Fig. 1. Method comparison. A. Scatter plot showing correlation of this ICP-MS method and the independent ICP-MS method [YICP-MS = 0.98(0.91 to 1.04) XInd + 0.1(0.1 to 0.2)]. B. Difference plot illustrating absolute individual specimen difference. C. Percent difference plot.
Please cite this article as: D.R. Bunch, et al., Whole blood selenium determination by inductively coupled plasma mass spectrometry, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013
4
D.R. Bunch et al. / Clinical Biochemistry xxx (2017) xxx–xxx
To work-around the Gd interference, physicians are requested to refrain from ordering a whole blood Se within 3 days of a Gd infusion. Future work could take into account a recent publication, which compared online and offline addition of internal standard and showed that both were acceptable; however, imprecision and bias were improved with the offline addition [28]. For this assay the comparison was not performed, but this latter approach will be studied for new trace metal assays in our laboratory. 5. Conclusions Total whole blood Se was successfully developed with a sample preparation involving only 1:9 dilution with 0.1% nitric acid. This ICPMS methodology is sensitive with the ability to quantify Se from 0.2 to 7.1 μmol/L. The assay was also used to determine the Se reference interval in a local reference population, which was found to be 2.2–3.5 μmol/ L. Ethical approval Collection of samples for this study was approved by the Institutional Review Board. Author contributions All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. Authors' disclosures of potential conflicts of interest No authors declared any potential conflicts of interest. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.clinbiochem.2017.01.013. References [1] W. Saliba, R. El Fakih, W. Shaheen, Heart failure secondary to selenium deficiency, reversible after supplementation, Int. J. Cardiol. 141 (2010) e26–e27. [2] M.R. Safarinejad, S. Safarinejad, Efficacy of selenium and/or n-acetyl-cysteine for improving semen parameters in infertile men: a double-blind, placebo controlled, randomized study, J. Urol. 181 (2009) 741–751. [3] K. Ashton, L. Hooper, L.J. Harvey, R. Hurst, A. Casgrain, S.J. Fairweather-Tait, Methods of assessment of selenium status in humans: a systematic review, Am. J. Clin. Nutr. 89 (2009) 2025S–2039S. [4] M.B. Zimmermann, J. Kohrle, The impact of iron and selenium deficiencies on iodine and thyroid metabolism: biochemistry and relevance to public health, Thyroid 12 (2002) 867–878. [5] V.M. Labunskyy, D.L. Hatfield, V.N. Gladyshev, Selenoproteins: molecular pathways and physiological roles, Physiol. Rev. 94 (2014) 739–777.
[6] G. Alfthan, J. Neve, Reference values for serum selenium in various areas evaluated according to the tracy protocol, J. Trace Elem. Med. Biol. 10 (1996) 77–87. [7] G. Alfthan, M. Eurola, P. Ekholm, E.R. Venalainen, T. Root, K. Korkalainen, et al., Effects of nationwide addition of selenium to fertilizers on foods, and animal and human health in Finland: from deficiency to optimal selenium status of the population, J. Trace Elem. Med. Biol. 31 (2015) 142–147. [8] J.L. Slavin, D. Jacobs, L. Marquart, K. Wiemer, The role of whole grains in disease prevention, J. Am. Diet. Assoc. 101 (2001) 780–785. [9] K.H. Schulpis, T. Karakonstantakis, S. Gavrili, G. Chronopoulou, G.A. Karikas, G. Vlachos, et al., Maternal—neonatal serum selenium and copper levels in Greeks and Albanians, Eur. J. Clin. Nutr. 58 (2004) 1314–1318. [10] M. Tonelli, N. Wiebe, B. Hemmelgarn, S. Klarenbach, C. Field, B. Manns, et al., Trace elements in hemodialysis patients: a systematic review and meta-analysis, BMC Med. 7 (2009) 25. [11] M.D. Churchwell, D.A. Pasko, I.F. Btaiche, J.C. Jain, B.A. Mueller, Trace element removal during in vitro and in vivo continuous haemodialysis, Nephrol. Dial. Transplant. 22 (2007) 2970–2977. [12] H. Steinbrenner, H. Sies, Selenium homeostasis and antioxidant selenoproteins in brain: implications for disorders in the central nervous system, Arch. Biochem. Biophys. 536 (2013) 152–157. [13] M. Vinceti, E.T. Wei, C. Malagoli, M. Bergomi, G. Vivoli, Adverse health effects of selenium in humans, Rev. Environ. Health 16 (2001) 233–251. [14] M. Vinceti, T. Maraldi, M. Bergomi, C. Malagoli, Risk of chronic low-dose selenium overexposure in humans: insights from epidemiology and biochemistry, Rev. Environ. Health 24 (2009) 231–248. [15] E. Jablonska, M. Vinceti, Selenium and human health: witnessing a Copernican revolution? J. Environ. Sci. Health C Environ. Carcinog. Ecotoxicol. Rev. 33 (2015) 328–368. [16] S. Chan, B. Gerson, R.E. Reitz, S.A. Sadjadi, Technical and clinical aspects of spectrometric analysis of trace elements in clinical samples, Clin. Lab. Med. 18 (1998) 615–629. [17] T.M.T. Sheehan, D.J. Halls, Measurement of selenium in clinical specimens, Ann. Clin. Biochem. Int. J. Biochem. Med. 36 (1999) 301–315. [18] A.M. Featherstone, A.T. Townsend, G.A. Jacobson, G.M. Peterson, Comparison of methods for the determination of total selenium in plasma by magnetic sector inductively coupled plasma mass spectrometry, Anal. Chim. Acta 512 (2004) 319–327. [19] J.M. Harrington, D.J. Young, A.S. Essader, S.J. Sumner, K.E. Levine, Analysis of human serum and whole blood for mineral content by icp-ms and icp-oes: development of a mineralomics method, Biol. Trace Elem. Res. 160 (2014) 132–142. [20] L. Liu, K. Tamama, O.M. Peck Palmer, Selenium overload? Clin. Chem. 62 (2016) 658. [21] C.F. Harrington, A. Walter, S. Nelms, A. Taylor, Removal of the gadolinium interference from the measurement of selenium in human serum by use of collision cell quadrupole inductively coupled plasma mass spectrometry (q-icp-ms), Ann. Clin. Biochem. 51 (2014) 386–391. [22] G. Lippi, M. Daves, C. Mattiuzzi, Interference of medical contrast media on laboratory testing, Biochem. Med. (Zagreb). 24 (2014) 80–88. [23] A. Walter, S. Nelms, C.F. Harrington, A. Taylor, Interference of gadolinium on the measurement of selenium in human serum by inductively coupled plasma-quadrupole mass spectrometry, Ann. Clin. Biochem. 48 (2011) 176–177. [24] W. Cieslak, K. Pap, D.R. Bunch, E. Reineks, R. Jackson, R. Steinle, et al., Highly sensitive measurement of whole blood chromium by inductively coupled plasma mass spectrometry, Clin. Biochem. 46 (2013) 266–270. [25] J. Westgard, Desirable Biological Variation Database Specifications, https://www. westgard.com/biodatabase1.htm2014. [26] C.A. Burtis, E.R. Ashwood, D.E. Bruns, N.W. Tietz, Tietz Textbook of Clinical Chemistry and Molecular Diagnostics, fifth ed. Saunders, St. Louis, Mo., 2013 [27] A.M. Albertson, M. Reicks, N. Joshi, C.K. Gugger, Whole grain consumption trends and associations with body weight measures in the United States: results from the cross sectional national health and nutrition examination survey 2001–2012, Nutr. J. 15 (2015) 8. [28] J.T. Collie, R.J. Massie, O.A. Jones, P.D. Morrison, R.F. Greaves, A candidate reference method using icp-ms for sweat chloride quantification, Clin. Chem. Lab. Med. 54 (2016) 561–567.
Please cite this article as: D.R. Bunch, et al., Whole blood selenium determination by inductively coupled plasma mass spectrometry, Clin Biochem (2017), http://dx.doi.org/10.1016/j.clinbiochem.2017.01.013