Trace element analysis of human urine collected after administration of Gd-based MRI contrast agents: characterizing spectral interferences using inorganic mass spectrometry

Biomonitoring is the direct measurement of environmental chemicals or their metabolites in physiological specimens (e.g., blood, urine, serum, etc.) to assess internal dose. 1 When assessing human exposure to essential (e.g., Cu, Zn, Se) and non-essential (Pb, Cd, Pt) trace elements, urine is often analyzed by quadrupole-based inductively coupled plasma mass spectrometry (Q-ICP-MS). In some respects, urine may be preferred as a biomarker of recent exposure, since it is easily collected (i.e., non-invasive) and many analytes of interest for biomonitoring are easily detected. There are several advantages to using Q-ICP-MS for biomonitoring, including its high sensitivity, multi-element capability, its dynamic linear range and its high throughput for quantifying trace elements in various biological matrices. Despite these advantages, Q-ICP-MS is prone to numerous spectral interferences that may produce erroneously elevated results. Selenium (Se) is an essential trace element that is necessary for normal enzymatic and cellular function. The main source of Se for humans is through dietary intake. There is a narrow critical physiological range for Se with detrimental health consequences being associated with both deficiency status and excessive accumulation. Diets deficient in Se may cause a form of cardiomyopathy called Keshan disease. 2 In recent years, dietary Se supplementation has grown in popularity due to potential cancer preventive properties. 3 However, the efficacy of enhanced Se intake on human health is difficult to interpret since diets supplemented with Se have also been associated with type II diabetes. 4 Although human sera are routinely analyzed to assess Se body burden, urine may also be a useful biomarker for assessing excessive Se intake. Urinary Se excretion reflects dietary absorption 5 and is the main physiological elimination pathway, with greater than 50% of ingested Se excreted in urine. 6 Concentrations of Se in human urine are quite variable with a reported reference range of 20–200 µg/L. 7 In contrast, platinum (Pt) is a non-essential element that is known to cause platinosis, 8 a severe asthma like reaction 9 to hexachoroplatinate salts in some sensitive individuals. There is a growing need for Pt measurements in biological matrices due to the increased therapeutic use of Pt coordination compounds as antitumor agents. 10 Although Pt containing anti-neoplastic drugs have been effective in the treatment of certain cancers, unintentional human exposure to background levels of Pt through environmental contamination from hospital effluent may be a growing area of concern. 11 Occupational settings and anthropogenic environmental contamination from sources such as vehicle exhaust catalysts are also potential sources of human exposure to Pt. Many studies have used urine to assess Pt body burden. Although urine Pt is a useful biomarker of exposure, excretion and toxicity may be largely dependent upon the chemical form and route of exposure. 12 Urinary Pt levels in the population are generally very low and less than the ICP-MS detection limit, i.e., typically <10 ng/L Pt. 13 Here we describe some select results from a recent epidemiologic study 14 that included the analysis of 619 urine specimens for 21 trace elements including Se and Pt, determined by Q-ICP-MS operated in both the Standard mode and Dynamic Reaction Cell (DRC) mode. During the analysis, highly anomalous results were observed for Se and Pt that were several orders of magnitude greater than the upper limit of the population reference ranges. These observations led to an investigation of potential isobaric interferences as the reason for the anomalous Se and Pt results. Sector-Field (SF-) ICP-MS was used in both medium and high-resolution modes as part of this investigation.

Women, 18 to 44 years in age, were recruited for participation in the Endometriosis: Natural History, Diagnosis, and Outcomes (ENDO Study), 2007–2009. The primary purpose of the ENDO Study was to assess environmental chemicals and the odds of an endometriosis diagnosis in two cohorts of women – operative and population. Complete details related to the ENDO Study design have been described elsewhere. 14 Respective Institutional Review Board approvals or reliances for the conduct of the ENDO Study were obtained from the National Institute of Child Health and Human Development (NICHD), New York State Department of Health (NYSDOH), the University of Utah, and the University of California San Francisco. All women were fully consented before data or biospecimen collection. Spot urine specimens were to be collected by protocol from women upon enrollment into the study and before surgery or Magnetic Resonance Imaging (MRI) utilizing containers determined to be free of the chemicals under study. Study participants were instructed on appropriate procedures for collection to minimize contamination potential. Urine was transferred into 2-mL Nalgene ® cryovials (Thermo Fisher Scientific, Rochester, NY, USA) and stored at −20°C. Specimens were shipped frozen on dry ice to the Trace Elements Section of the Laboratory of Inorganic and Nuclear Chemistry at the New York State Department of Health’s (NYSDOH) Wadsworth Center (Albany, NY, USA) for multi-element analysis. Specimens were stored at approximately −80°C until batched analysis could be arranged. All storage containers and materials used were determined to be trace element free. All specimens were prepared for analysis in Class II, Type A2 Biological Safety Cabinet (BSC) (The Baker Company, Sanford, MA, USA), that was validated as meeting Class 100 specifications. Specimens were sampled in a positive pressure Class 100 Clean Room (Terra Universal, Fullerton, CA, USA). Class 100 clean conditions or better for the BSC and clean room were ensured throughout via periodic monitoring with a laser particle counter (Fluke Corporation, Everette, WA, USA). Urine specimens were thawed to room temperature and diluted 1+19 for analysis with 2% (v/v) double distilled HNO 3 produced in-house using a DuoPUR sub-boing acid still (Milestone, Shelton, CT, USA), 0.005% (v/v) Triton ® X-100 (Sigma Ultra, Sigma-Aldrich Inc., St Louis, MO, USA) as a surfactant, and 10 µg/L Ga, Y, Rh and Ir (High-Purity Standards, Charleston, SC, USA) as internal standards in ≥18 MΩ•cm doubly-deionized water (DI) obtained using a NANOpure Diamond™ mixed bed deionization system (Barnstead, Waltham, MA, USA). Matrix-matched calibration curves were established using six standards prepared by serial dilution of custom multi-element stock solution (High-Purity Standards, Charleston, SC, USA) traceable to the US National Institute of Standards and Technology (NIST, Gaithersburg, MD, USA). For each calibration standard, 500 µL of base urine was added for matrix matching purposes, along with the same reagents used to prepare patient specimens. Reagent blank samples, n=3 for each analytical run, were subtracted from each specimen result. Aqueous solutions of Gd were prepared from a 1000 mg/L stock solution (GFS Chemicals, Powell, OH, USA) in 2% (v/v) double distilled HNO 3 and DI water. A solution containing chloride ions was prepared at a concentration of approximately 200 mmol/L from molecular biology grade NaCl (MP Biomedicals United States, Solon, OH, USA). The Gd and Cl solutions, along with select participant urine specimens, were diluted 1+19 in 2% (v/v) HNO 3 for SF-ICP-MS spectroscopic interference studies. Quadrupole-based (Q-) and sector field (SF-) ICP-MS instrumentation were utilized for analysis. Typical instrumental operating parameters are shown in Table 1 . Total elemental concentrations in urine were determined using a Perkin Elmer ELAN DRC II ICP-MS (PerkinElmer Life and Analytical Sciences, Shelton, CT, USA) with Dynamic Reaction Cell (DRC) technology. The instrument was equipped with a CETAC ASX-520 Autosampler (Omaha, NE, USA), a Meinhard ® High Sensitivity (Type A) concentric quartz nebulizer (Meinhard Glass Products, Golden, CO), and a Quartz Cyclonic Spray Chamber. Urine specimens were analyzed for 21 trace elements using a well-validated method that has been in routine operation in the Trace Elements Laboratory since 2002. The method has been described in more detail elsewhere. 15 Selenium ( m/z =78) was determined in the DRC mode with 10% hydrogen (H 2 ) in argon (Ar 2 ) as a collision gas at a flow rate optimized at 0.3 L min −1 . Platinum ( m/z =195) was determined in standard mode. A Thermo Fisher Element 2 Sector Field (SF-) ICP-MS instrument (Thermo Fisher Scientific, Bremen, Germany) was used for more detailed interference studies. The instrument was equipped with a CETAC ASX-520 autosampler, a Conikal ® concentric borosilicate glass nebulizer (Glass Expansion, Pocasset, MA, USA), and a quartz double pass spray chamber (Thermo Fisher Scientific, Rochester, NY, USA). Mass spectral scans were collected in high resolution mode ( R = m/Δm = 10,000) for all isotopes of Se and in medium resolution mode ( R = m/Δm = 4000) for all isotopes of Pt. Instrument parameters are given in Table 1 . Method detection limits (MDL) were calculated according to International Union of Pure and Applied Chemistry (IUPAC) guidelines. 16 The MDL is defined as three times the standard deviation of concentrations measured in a base urine pool or low-level sample for a minimum of 10 independent analytical runs. The calculated MDL for Se and Pt in urine were 5 and 0.02 µg/L respectively, which incorporates a 1+19 dilution factor for urine samples. Limits of Quantitation (LOQ), defined as 10 times the calculated standard deviation, were determined for each element using the same material and analyses as those used to calculate the MDLs. The LOQ for Se and Pt in urine was 16 and 0.08 µg/L respectively.

Our data suggest that administration of Gd-based MRI contrast agents are associated with major spectral interferences for the determination of Se and Pt in urine by Q-ICP-MS. The specific interferences were characterized using SF-ICP-MS by analyzing urine specimens obtained from 2 study subjects, who had received a Gd-chelate contrast agent. For each of the Se isotopes, we also analyzed a solution containing 1 mg/L Gd and, for the Pt isotopes, we analyzed solutions containing 10 mg/L Gd, 100 mg/L Gd, and 100 mg/L Gd (aq) supplemented with Cl. The large amount of Gd present in some of the ENDO Study participants’ specimens was sufficient to result in erroneously elevated values for urinary Se, measured at m/z =78, and urinary Pt, measured at m/z =195 using Q-ICP-MS. SF-ICP-MS can be used in high and medium-resolution modes to resolve the Gd-based interferences from 78 Se + and 195 Pt + . It may be reasonable to employ SF-ICP-MS instrumentation for some purposes when Gd based interferences are suspected in a study population; however, use of SF-ICP-MS may be cost prohibitive for large-scale epidemiological and biomonitoring studies. A spectral interference caused by 156 Gd + was previously reported when analyzing a serum specimen for Se at m/z =78; 33 however, to our knowledge the observations reported herein are the first to document Gd interferences in urine specimens, and first to identify GdAr, GdCl, and GdO as interferences on Pt. Data from this study are particularly unique with regard to the number of participants’ specimens affected by the Gd interferences. The prevalence of interferences in argon plasma from medical formulations when analyzing biomonitoring study specimens may increase as Gd-based contrast agents become more widely used and new formulations make their way into common usage. This study serves as a cautionary reminder of the importance of maintaining effective communication between the analytical, medical, and epidemiological team members regarding strict adherence to study protocols in human biomonitoring studies.

The Trace Elements laboratory operates with repeat thresholds for elevated specimens based on information obtained from internal biomonitoring studies or from data provided by the Centers for Disease Control and Prevention related to the National Health and Nutrition Examination Survey (NHANES) studies. 1 Repeat thresholds are used within our laboratory to confirm elevated results compared to an established reference range. During the course of the ENDO Study, we were surprised to observe anomalous results for Se and Pt that were several orders of magnitude greater than the upper limit of established population reference ranges. Figure 1 illustrates the distribution of Se ( Fig 1a ) and Pt ( Fig 1b ) values in the ENDO Study cohort. It is clear from Figure 1 that a “cluster” of participants appear to have extremely elevated results for both Se and Pt. Measured urinary Se results for the ENDO Study are shown on a log scale. From these data, we identified 38 ENDO Study participants with extremely high urinary Se that were several orders of magnitude above the laboratory’s established repeat threshold. Oddly, the same participants had elevated urinary Pt concentrations as well. Compared to reference ranges cited in literature for Se (i.e., 20–200 µg/L), 7 concentrations measured in the ENDO Study population ranged from below our method detection limit (MDL) to ~600 mg/L. We expected that most of the ENDO Study participants would have urinary Pt concentrations that were below our MDL, since urinary Pt is barely detectable in the US national biomonitoring study. 1 However, urine Pt concentrations for the ENDO Study ranged from non-detectable to 27.3 µg/L. These unexpected data were compared to select urinary Se and Pt concentrations for exposed populations that have been published in literature. 17 , 18 Elevated urinary Se values in the range 20–1,000 µg/L have been reported for individuals from seleniferous regions in the Midwestern US 17 and Pt concentrations for occupationally exposed individuals have been reported as 0.021–2.90 µg/L. 18 It became clear from our data that extremely elevated levels of “Se” and “Pt” in urine specimens from the ENDO Study cohort were highly unusual even when compared to those reported for exposed populations, and would require further investigation. These elevated urinary “Se” and “Pt” results triggered internal quality assurance actions within the laboratory due to the potential adverse health implications associated with such exposures. We assessed PT and EQA performance for Se and Pt during the period of time the ENDO Study analyses were conducted within the laboratory and found them all to be satisfactory. We also re-examined method accuracy for Se and Pt via analysis of NIST Standard Reference Materials (SRMs) 2670a Toxic Elements in Urine (Freeze-Dried) and 2668 Toxic Elements in Frozen Human Urine, and confirmed that it was acceptable. We then investigated potential spectral interferences in inorganic mass spectrometry that might explain the anomalous data. The ENDO Study protocol called for a random sample of women in the operative cohort to undergo the same pelvic MRI as did all women in the population cohort who were not having surgery to inspect the pelvis for endometriosis and to assess abdominal fat distribution. By design, all MRIs were to be performed after enrollment and completion of the baseline interview, anthropometric assessment, and collection of blood and urine samples for various environmental analyses, including trace elements. MRI is a powerful tool for visualizing soft tissues, and Gd-based contrast agents are frequently used to improve image quality. For MRI purposes, Gd is administered as a chelate, such as Gd-DTPA (diethylenetriaminepentacetate) and Gd-DOTA (tetraazacyclododecanetetraacetic acid), to increase solubility and decrease toxicity from “free” or unbound Gd 3+ ion. 19 Once administered, approximately 90% of the chelated Gd is eliminated through urinary excretion within 24 hours 19 in subjects with normal renal function. Exceptions are observed for patients with renal insufficiency where the half-life is prolonged and may exceed 30 hours. 20 Although Gd contrast agents have been shown to be safe in patients with normal renal function through extensive clinical trials 19 and to have a low incidence of allergic-like reactions, 21 in renal insufficiency, Gd-based agents may contribute to nephrogenic systemic fibrosis (NSF). 19 , 21 – 23 Toxicity associated with Gd contrast agents is due to the dissociation of the chelated Gd complex into the “free” metal ion and ligand. Free Gd is very toxic and accumulates in the liver and in lymph nodes. 24 It also accumulates in bone 19 where it may become irreversibly incorporated. 25 For this reason it is desirable that the specific ligand used to chelate the Gd 3+ results in a complex with a high thermodynamic stability constant that is therefore kinetically inert. 25 Adding to the complexity of MRI contrast agent behavior after administration, the chelate may undergo a competitive ligand reaction resulting in Gd metal ion exchange 26 , 27 with endogenous metals such as Zn. The process by which a Gd ion is exchanged by a chelate for an in vivo metal ion is referred to as transmetallation. These phenomena have been reported to increase human urinary excretion of zinc and copper. 28 MRI contrast agents have been known to interfere with several clinical analyses including a negative interference indicating apparent pseudohypocalcemia in serum 29 and plasma 30 specimens. Serum creatinine 31 and iron 32 analyses have also been undesirably impacted by Gd-based MRI contrast agent administration. In a recent report in the clinical literature, Gd was identified as an interference on Se at m/z =78 in a serum specimen analyzed by Q-ICP-MS. 33 While investigating an unusual high serum Se result obtained by ICP-MS, Walter 33 et al., discovered that the patient had received a Gd-based MRI contrast agent prior to the serum specimen being collected. Based on this report, we considered the possibility that administration of Gd-based MRI contrast agents might be responsible for the unusual results observed for Se, and for Pt during the ENDO Study. To date there have been no published reports of Gd interferences observed in urine specimens analyzed by Q-ICP-MS. Thus, the ENDO Study may be considered particularly unique with regard to the large number of participants’ urine specimens affected by presumed spectral interferences from an unknown source. Based on a review of the clinical information obtained from the ENDO Study, it was found that a small number of study participants had provided urine specimens for analysis after, rather than before, undergoing an MRI. This was largely attributed to challenges in scheduling MRIs at convenient times for the participants. For other subjects, the clinical records were unclear. Therefore, we categorized the ENDO Study data for urine Se and urine Pt based on the date of participant urine collection relative to the date of MRI test as indicated in the clinical record. Figure 2 shows the urinary Se ( Fig 2a ) and Pt ( Fig 2b ) concentrations stratified by information on the timing of MRIs. These six categories include: 1) “No MRI Information” for participants that did not have MRI information available; 2) “MRI No Date” for subjects that received an MRI, but for which the MRI date in relation to urine collection was not available; 3) “Urine pre MRI” for those participants that provided urine specimens prior to having an MRI; 4) “Urine + MRI Same Day” for subjects that provided a urine specimen on the same day as undergoing an MRI; 5) “Urine 1–2 Days Post MRI” for subjects whose urine was collected 1–2 days after administration of the Gd-based MRI contrast agent; and 6) “Urine >2 Days Post MRI” for subjects that provided urine more than 2 days after having had an MRI. Measured “Se” values are shown on a log scale in Figure 2a . It is apparent from the data that there are two distinct groups within the ENDO cohort, a grossly elevated “Se” group and an unexposed group. Urinary Pt data are shown in Figure 2b , along with the MDL for Pt. Again, it is apparent that there are two groups present within the ENDO Study, one with highly elevated urine “Pt” and, as expected, a majority of subjects with Pt levels close to the MDL. It is interesting to note that for some ENDO Study participants, elevated “Se” and “Pt” are still apparent in 1–2 Days after having undergone the MRI group. The only data irregularity or outlier is for an individual woman in the “Urine pre MRI” group with “elevated Se and Pt” levels. While this specific woman did undergo MRI after donating urine as part of the ENDO Study, it is possible that another MRI may have been performed previously with a Gd-based MRI contrast agent. The possibility of unknown isobaric or polyatomic interferences was considered as a potential explanation for unusually elevated urine Se and Pt results. Table 2 lists these interferences based on information obtained from various literature sources 5 , 10 , 33 , 34 and other theoretical possibilities. It is interesting to note that doubly charged Gd could interfere with all of the major isotopes of Se, except 82 Se + , which is subject to the well known 40 Ar 2 H 2 + polyatomic interference. In the case of Pt, all of the isotopes are subject to spectral overlap that may be attributed to gadolinium-containing polyatomic species. The determination of Se in complex biological matrices by Q-ICP-MS is an analytical challenge due to numerous potential spectral interferences. Argon based polyatomic interferences are prevalent for all the major isotopes of Se. However, the introduction of DRC technology for Q-ICP-MS has opened up new approaches for attenuating argide interferences associated with measurement of Se. For example, the argon dimer 40 Ar 38 Ar + that interferes with 78 Se + can be attenuated by using 10% hydrogen (by volume) in Ar as the DRC gas. 35 Since 38 Ar + has an abundance of just 0.06%, it follows that the interference will be small at m/z =78. According to published reports, collision processes sufficiently reduce 40 Ar 38 Ar + interference on the 78 Se + isotope to negligible levels. 35 , 36 Based upon our satisfactory performance for serum and urine Se in various PT and EQA schemes, problems with Se measurements were not anticipated for the ENDO Study. Doubly charged species are not generally problematic for the measurement of 78 Se + by Q-ICP-MS, and the 138 Ba 2+ / 138 Ba + ratio is typically optimized to be less than 2%, since Ba is a good indicator of doubly charged ion formation due to it’s low second ionization potential. 37 Previous studies of doubly charged ions (M 2+ ) indicate that those elements with second ionization potentials that are below that of Ar (i.e., 15.76 eV) are likely to become doubly ionized. 38 With a second ionization potential of 12.09 eV, Gd is a good candidate for M 2+ production that could affect Se measurements at m/z =78. Walter 33 et al., has previously noted an interference on 78 Se + when an elevated measurement was observed in a clinical serum specimen. The resultant investigation indicated that doubly charged species present from administration of a Gd-based MRI contrast agent may have caused an isobaric interference and hence, an erroneously elevated serum 78 Se + value. In order to explore the source of interference in the serum specimen, the authors added an aqueous Gd solution to archived EQA scheme samples with Se target values established by inter-laboratory comparison. An observed increase in the measured Se values following addition of Gd confirmed the 156 Gd 2+ interference on 78 Se + . With access to sector field (SF) instrumentation, we are able to explore the 156 Gd 2+ interference on 78 Se + in more detail for an aqueous Gd solution and a clinical urine specimen from the ENDO Study. In Figure 3 , a mass scan is shown of an aqueous solution containing 1 mg/L Gd (dashed line) and a participant urine specimen (solid line), measured in high-resolution mode ( R = m/Δm = 10,000) by SF-ICP-MS. A small peak identified as 78 Se + ( m/z =77.92), is clearly evident in the urine specimen and is consistent with expected levels in humans. A peak that may be attributed to a combination of 40 Ar 38 Ar + , 78 Kr + , and 62 Ni 16 O + is observed for both the urine specimen and the aqueous solution. The large peak at m/z =77.96 is doubly charged 156 Gd 2+ and confirms the presence of large amounts of the contrast agent in the participant’s urine specimen. For the participant urine, an unidentified peak appears in the mass scan at m/z =77.95, and a signal at this mass is not observed for the aqueous solution. The analyte signal at m/z =77.95 is therefore presumably due to an endogenous matrix component that is only present in the participant urine. Given the doubly charged interference observed at m/z =78, we elected to investigate potential interferences on each of the other isotopes of Se. However, to accomplish this using the limited residual volume of participant urine available to us required collection of fewer data points per scan. Figure 4(a–f) shows the SF-ICP-MS scans in high resolution for a second participant urine (solid line) compared to 1 mg/L Gd (aq) solution (dashed line) for each of the 6 isotopes for Se: 74 Se, 76 Se, 77 Se, 78 Se, 80 Se, and 82 Se. As expected, doubly charged Gd ions were confirmed in the second participant urine specimen by comparison to a 1 mg/L aqueous Gd solution, and are observed for all isotopes of Se except for 74 Se + and 82 Se + . Interesting additional features were observed for many of the Se isotopes when explored in high resolution by SF-ICP-MS. For 74 Se + ( Fig 4a ) an identified peak that is barely detectable occurs at m/z =73.920 that may be an endogenous matrix component in the study participant’s specimen and 74 Se + is apparent at m/z =73.922. In Figure 4a , a small peak is visible for the Gd (aq) solution at m/z =73.930 that may be 36 Ar 38 Ar + . By contrast, a peak observed at m/z =73.932 in the participant urine specimen could be the species 37 Cl 2 + . The well-characterized argon dimer interferences were observed for 76 Se + ( Fig 4b , 36 Ar 40 Ar + ), 78 Se + ( Fig 4d , 38 Ar 40 Ar + ), and 80 Se + ( Fig 4e , 40 Ar 40 Ar + ). Additional polyatomic interferences can also be discerned under high resolution for 77 Se + ( Fig 4c , 36 Ar 40 ArH + ), and 82 Se + ( Fig 4f , 40 Ar 40 ArH 2 + ). Argon interferences cannot be resolved from other potential interferences, such as nickel oxides, even in high-resolution. For example, the argon dimers observed for 76 Se + ( Fig 4b ) and 78 Se + ( Fig 4d ) may also include 60 Ni 16 O + ( m/z =75.930) and 62 Ni 16 O + ( m/z =77.923). Potential contributions from krypton (i.e., 78 Kr + and 82 Kr + ) due to contamination of liquid Ar supplies are identified in Figs. 4d and 4e , however these have not been confirmed independently. Additional peaks visible in the mass scans have been tentatively attributed to various chloride and oxide species. These include 37 Cl 40 Ar + on 77 Se + ( Fig 4c ) and 66 Zn 16 O + on 82 Se + ( Fig 4f ). For Se isotopes where doubly charged Gd interferences were expected, 76 Se + , 77 Se + , 78 Se + , and 80 Se + ( Fig 4b–e ), these species are clearly evident and confirmed using an aqueous Gd solution. Indeed the relative intensities observed in the participant urine for each of the four Gd 2+ isotopes matching the major Se isotopes, closely match the natural abundance for Gd ( Table 3 ). Figure 5 shows a mass scan from 194.85 to 195.00 amu of an aqueous solution containing 1 mg/L Gd (dashed line) and a participant urine specimen (solid line), measured in medium-resolution mode ( R = m/Δm = 4,000) by SF-ICP-MS. A distinct peak, identified as 195 Pt + ( m/z =194.96), is evident in the participant urine specimen, and is consistent with very low background levels of Pt found in human urine. Analysis of the aqueous Gd solution (i.e., without Pt) under medium resolution shows a small peak at m/z =194.89, which could be assigned as one or more GdAr species, e.g., 155 Gd 40 Ar + and 157 Gd 38 Ar + . By contrast, the presence of a much larger peak at m/z =194.89 in the participant urine suggests some other species might be present, and thus further investigation was undertaken with a second participant urine. In this follow up investigation, we explored potential interferences on each of the isotopes of Pt: 190 Pt, 192 Pt, 194 Pt, 195 Pt, 196 Pt, and 198 Pt, as we had done with Se. Fig 6 (a–f) shows SF-ICP-MS scans (in medium resolution) that were obtained for each of the following: a participant urine; 10 mg/L Gd; 100 mg/L Gd; and 100 mg/L Gd + 10 mmol/L NaCl, to identify select polyatomic species. Note the NaCl concentration was selected to match human urine following dilution. Trace amounts of Pt are clearly evident in the participant urine at all of the isotopes of Pt ( Fig 6c–f ), except 190 Pt + and 192 Pt + , which are <1% relative abundance. It is interesting to note that a number of Gd-containing polyatomic species are present at all isotopes of Pt. Potential polyatomic species were identified using the “Interferences Workshop” feature in the Element2 software. For example, Fig 6c shows the mass scan in the region of m/z =194, under medium resolution. A peak is clearly visible m/z =193.89 in the participant urine but it is absent in both the 10 mg/L and 100 mg/L Gd solutions. Since this peak cannot be due to a 154 Gd 40 Ar + species, the only other possibility is that it must be another Gd polyatomic, perhaps GdCl. Adding sufficient Cl to the 100 mg/L Gd (aq) solution to match that found in typical human urine provided convincing evidence that the peak at m/z =193.89 must be due to a 157 Gd 37 Cl + species. Indeed, other GdCl species could well be causing interferences on 190 Pt + , 192 Pt +, and 196 Pt + ( Fig 6a,b,e ). For 195 Pt + , the interference is an unresolved mixture of several Gd argide species and 160 Gd 35 Cl + . For 196 Pt + , there is a potential interference from 158 Gd 35 ClH 3 + . Additional interferences identified by the Element2 program included gadolinium oxide species (GdO). Evidence for a GdO species was obtained by comparing 10 mg/L to 100 mg/L Gd (aq) . Potential interferences include: 158 Gd 16 O 2 + on 190 Pt; 160 Gd 16 O 2 + on 192 Pt; and 160 Gd 16 O 2 H 2 + on 194 Pt. Many of the Gd-containing polyatomic species identified under medium resolution, would remain unresolved even under high-resolution. Polyatomic interference production in an argon plasma, as is the case with GdAr species, is likely to be instrument specific. 39 The precise mechanism for polyatomic species formation in ICP-MS is still unclear. Ion extraction investigations propose formation of argides and oxides in the region between sampler and skimmer cones due to collisions or in the sheath in the inner boundary layer of the sampler cone. 37 If this is the case, then these interferences might be attenuated with the use of DRC technology.