Chemical speciation of arsenic in serum of uraemic patients† Xinrong Zhanga, Rita Cornelis*a, Louis Meesa, Raymond Vanholderb and Norbert Lameireb a Laboratory for Analytical Chemistry, Institute for Nuclear Sciences, University of Ghent, Proeftuinstraat 86, B-9000 Ghent, Belgium b Renal Division, Department of Medicine, University Hospital, De Pintelaan 185, B-9000 Ghent, Belgium Chemical speciation of arsenic was carried out in serum of a total of 51 uraemic patients: 19 non-dialysis (ND), 18 haemodialysis (HD) and 14 continuous ambulatory peritoneal dialysis (CAPD) patients.The low molecular mass As species were separated by ion-exchange liquid chromatography and measured on-line by hydride generation atomic absorption spectrometry (HGAAS). The high molecular mass As species were separated by fast protein liquid chromatography, either size-exclusion, ion-exchange or affinity chromatography, and the fractions were digested and measured off-line with HGAAS.The mean total As concentrations in the serum of the three groups of the uraemic patients were significantly higher than the reference value (6.47 ± 4.28, 5.12 ± 5.58 and 4.67 ± 5.41 mg l21 for HD, ND and CAPD patients, respectively, versus the reference value of 0.96 ± 1.52 mg l21). The major As species in serum of the patients were dimethylarsinic acid (DMA) and arsenobetaine. The HD patients showed a significantly higher mean DMA level than ND and CAPD patients. No selective removal of different As species in serum of HD patients was observed after 4 h of haemodialysis.The inorganic As species in serum were bound to proteins, mainly transferrin (about 5–6% of total As in serum). This binding may play an important role in arsenic detoxification. Keywords: Speciation; serum arsenic; uraemic patients; dialysis; liquid chromatography; hydride generation atomic absorption spectrometry It is well known that the kidneys act as filters to remove toxic waste products from the blood via glomerular filtration.Reabsorption and secretion processes on plasma membranes of the epithelial cells allow useful metabolites and electrolytes to be conserved and waste products to be excreted. For patients who suffer from chronic renal diseases, toxic waste products cannot be excreted efficiently, and increased concentrations of trace elements in the blood of such patients has been documented. 1–9 Zhang et al.3 reported a threefold increase in the mean arsenic concentration in serum and a twofold increase in that in packed cells of uraemic patients compared with controls.3 Similar results were also reported by Giovannetti et al.8 in uraemic plasma.A higher accumulation of arsenic in plasma and packed blood cells of haemodialysis and haemodiafiltration patients have been found by De Kimpe et al.,4 Brune et al.9 and Van Renterghem et al.5 A relationship between arsenic levels and the concentration of serum creatinine in the serum of the patients has also been reported.3 Although considerable efforts have been made to study arsenic accumulation in chronic renal patients, the toxic effects of arsenic are still not entirely understood because most publications give information only about total arsenic concentrations in body fluids.3–9 The toxicology of arsenic is complicated by its ability to convert between oxidation states and different organometallic forms.10,11 These processes cause differences in the relative affinity of the various arsenic species bound to tissues, and they determine both the intoxication and detoxification mechanisms.For instance, the transformation of arsenate into dimethylarsinic acid is very important because this process is believed to be the principal detoxification mechanism. 12 It has also been suggested that during the methylation process, intermediates of arsenite and As–protein molecules are formed that, if allowed to accumulate, could be more (for arsenite) or less (for As–protein molecules) toxic than arsenate. 13,14 In order to understand the mechanism of arsenic intoxication and detoxification more clearly, speciation studies are obviously the primary and necessary step. The purpose of this work was to use different types of liquid chromatographic techniques coupled to hydride generation atomic absorption spectrometry (HGAAS) for the speciation of As species in serum of three types of uraemic patients: nondialysis (ND), haemodialysis (HD) and continuous ambulatory peritoneally dialysis (CAPD) patients.This is expected to provide a better insight into the accumulation of the different As species. Moreover, this study will allow one to compare the levels of toxic As species between patients with and without dialysis treatment. These data may be helpful in evaluating the risk of As accumulation for patients with chronic renal insufficiency. Experimental Sample collection and pre-treatment Blood samples from 19 uraemic ND, 18 HD and 14 CAPD patients were collected at the Nephrology Department of the University Hospital, Ghent.The patients gave their informed consent. They were requested to abstain from sea-food intake during the 3 d before the blood collection. The first 25 ml of blood were preserved for routine clinical laboratory tests. The subsequent samples for As analysis were collected in carefully cleaned high-purity quartz tubes, which were immediately covered with Teflon stoppers, returned to an air-tight plastic transport container and transported to a dust-free room (class 100). After clotting, serum and packed cells were separated by centrifugation of the tubes at 1000 g.Deproteinization was carried out by ultrafiltration on membranes with a 10 kDa molecular mass cut-off (Filtron Microsep, Filtron Technology, Northborough, MA, USA). The ultrafiltrate was submitted to speciation of low molecular mass (LMM) As species.The high molecular mass (HMM) As species were determined in the original serum samples without undergoing deproteinization. † Presented at The Sixth Nordic Symposium on Trace Elements in Human Health and Disease, Roskilde, Denmark, June 29–July 3, 1997. Analyst, January 1998, Vol. 123 (13–17) 13Reagents and apparatus Arsenic species standards of monomethylarsonic acid (MMA), dimethylarsinic acid (DMA), arsenobetaine (AsB) and arsenocholine (AsC) were provided by the Commission of the European Union, DGXII, Measurements and Testing Programme (Brussels, Belgium).Sodium arsenite and arsenate were obtained from Merck (Darmstadt, Germany). A 2% m/v solution of sodium tetrahydroborate (Janssen Chimica, Beerse, Belgium) in 0.05% m/v NaOH (UCB, Leuven, Belgium) was freshly prepared every day and filtered prior to use. A 2 mol l21 HCl solution was prepared from concentrated HCl (32% m/v) purified under sub-boiling conditions.Two types of HPLC columns were used for the separation of LMM anionic and cationic As species:15,16 a silica-based anionexchange column (Supelcosil LC-SAX, 250 3 4.6 mm id). (Supelco, Bellefonte, PA, USA) and a polystyrene–divinylbenzene- based cation-exchange column (Ionpac) CS 10, 250 3 4 mm id) (Dionex, Sunnyvale, CA, USA). A low-pressure UV lamp (6 W, 12 3 3 cm id) (Philips, Eindhoven, The Netherlands) was applied for on-line sample digestion. A fast protein liquid chromatographic (FPLC) system (Pharmacia, Uppsala, Sweden) with three types of columns was used for the separation and identification of As–protein binding:17 sizeexclusion (SEC) (Superose HR 10/30 analytical column and Sephadex G-50 XK 50/30 preparative column); anion-exchange (IEC) (MonoQ HR 16/10) and affinity (NHS-activated HiTrap Superose HR 10/2, coupled with goat anti-human transferrin).The HPLC mobile phase for the separation of LMM As species by cation exchange was prepared with an aqueous solution of 100 mmol l21 HCl and 50 mmol l21 NaH2PO4·H2O (Merck).The mobile phase for the separation of LMM As species in anionic form was prepared with an aqueous solution of 30 mmol l21 NaH2PO4 without pH adjustment. The mobile phases for the separation of HMM As species were 0.025 mol l21 TRIS–HCl (pH 7.4) for SEC and 0.025 mol l21 TRIS– HCl (pH 8.0) in a linear NaCl gradient (0–0.5 mol l21) for IEC. The buffer solutions used for affinity chromatography were as follows: buffer A, 0.2 mol l21 NaHCO3–0.5 mol l21 NaCl (pH 8.3) for ligand coupling; buffer B, 1 mol l21 ethanolamine in buffer A for deactivation of excess NHS groups; buffer C, 75 mmol l21 TRIS–HCl (pH 8) for affinity absorption; and buffer D, 0.5 mol l21 glycine–HCl (pH 2.0) for desorption of transferrin from the column.A Perkin-Elmer (Norwalk, CT, USA) Model 3030 atomic absorption spectrometer with an arsenic electrodeless discharge lamp, operating at a power of 8 W, was used throughout for the detection of arsenic signals.The wavelength was set at 193.7 nm with a spectral slit-width of 0.7 nm. A FIAS 200 flowinjection system (Perkin-Elmer) was used to generate the hydride. The temperature of the quartz tube atomizer was set at 900 °C. Commercial reagents (Boehringer, Mannheim, Germany) and a Model 747 analyser (Hitachi, Tokyo, Japan) were used for the determination of serum creatinine. The serum creatinine levels in individuals with normal renal function range from 57 to 93 mmol l21 (n = 63, 95% range) and from 50 to 80 mmol l21 (n = 55, 95% range) for men and women, respectively.Statistics Three determinations were performed for each sample. Reagent blanks were processed with the samples. Analytical results were obtained by calculating the mean values of three blank corrected analyses of each sample. Values lower than the detection limits were statistically considered as zero. The differences between the results were analysed by Student’s t-test (normal distribution) and the Mann–Whitney U-test (distribution-free), using the p-value as a judgment of significance.The Shapiro–Wilk test was used for goodness of fit. Stated values were considered significantly different for p < 0.05. Correlation was calculated and tested by the Pearson test. All these statistics were executed by the program UNITSTAT. Quality controls The accuracy for total As determination was tested by the determination of arsenic in a certified freeze-dried reference serum from the University of Ghent.18 The accuracy for As speciation was tested by the determination of arsenic in the BCR candidate Reference Material CRM 526 tunafish tissue.Table 1 summarizes the arsenic concentrations in the reference materials obtained in this work and the certified values. No significant differences were established by the paired t-test at the 95% confidence level for all three reference materials.The results demonstrate that the proposed method is accurate for the determination of the arsenic concentrations. The detection limits, estimated as three times the standard deviation of the blank, were 0.031 (total As), 1.0 (AsIII), 1.5 (AsV), 1.0 (MMA), 1.3 (DMA), 1.5 (AsB) and 1.5 mg l21 (AsC) in serum. The blank obtained by injection of a blank digested solution gave a mean signal of 0.003 absorbance, which corresponds to a concentration of 0.05 mg l21 total As.Results and discussion Speciation of LMM As species Mean concentrations The mean (±s) total As levels in the 19 uraemic ND, 18 HD and 14 CAPD patients are 5.12 ±5.58, 6.47 ±4.28 and 4.67 ±5.41 mg l21, respectively, all significantly higher (p < 0.001) than the reference values of 0.96 ±1.52 mg l21 in serum of healthy subjects (n = 23), indicative of a significant accumulation of arsenic in the serum of the uraemic patients. The mean (±s) DMA concentrations in serum of ND and CAPD patients are 0.82 ±1.05 mg l21 (16.0% of total As) and 0.71 ±1.05 mg l21 (15.2% of total As), respectively, but a significantly higher DMA value of 1.93 ±1.51 mg l21 (29.8% of total As) was observed in the serum of HD patients (p < 0.02), indicative of the accumulation of this As species.DMA is a methylated metabolite of inorganic As, which is less toxic than inorganic As species. The methylation of inorganic As in mammals is a detoxification mechanism. Normally, DMA is rapidly excreted in the urine through the kidney.For HD patients, this function is only partially replaced by the HD treatment. The higher DMA level in HD patients compared with ND and CAPD patients apparently indicates that this treatment is of more limited effect for the removal of this As species from the blood of HD patients. The mean (±s) AsB concentrations in serum of ND, HD and CAPD patients are 3.55 ±4.58, 3.47 ±2.89 and 3.56 ±4.27 mg l21, respectively. Compared with the level of 0.96 mg l21 of total As in healthy subjects, the mean levels of AsB in the serum Table 1 Arsenic concentrations measured in certified reference materials Mean ± CI, 95% Reference As species Reference material and units This work (n = 5) values Certified freezedried reference serum Total/mg kg21 17.5 ± 2.2 19.6 ± 4.0 CRM-526 tunafish tissue Total/mg g21 4.7 ± 0.1 4.8 ± 0.3 AsB/mmol kg21 53.3 ± 2.6 51.5 ± 2.1 DMA/mmol kg21 2.19 ± 0.14 2.04 ± 0.27 14 Analyst, January 1998, Vol. 123of the three groups were markedly increased. However, we did not observe any differences between them. These results may indicate that AsB species in the serum of uraemic patients is affected by food intake, which may be sea-food but also fish derivatives in processed food or meat of animals fed on fishmeal. All this was difficult to control in the present survey. Table 2 summarizes the results for different As species in the serum of uraemic ND, HD and CAPD patients. Unfortunately, the concentrations of As species in the serum of a reference group of healthy subjects could not be measured in the present study because they lie below the detection limits of the present coupled technique.To our knowledge, no As speciation in serum of healthy controls has ever been published. A further improvement in the methodology of As speciation is obviously needed in the future. Correlations To establish whether the accumulation of As species in serum of uraemic patients proceeds in parallel with the progression of renal failure, the correlations between As species and creatinine levels in the serum of the 51 patients were studied.As shown in Fig. 1, the concentrations of creatinine are correlated with the concentrations of DMA species (r = 0.46, p = 0.004). No correlation was observed, however, between the creatinine levels and AsB species concentrations (r = 0.065, p = 0.33). This is because the AsB concentration in serum of an individual at any given moment is related to two factors: the amount of AsB intake from foodstuffs and the rate of AsB excretion from the kidney.A very high concentration of AsB species can be detected in blood samples of an individual with normal renal function, just after ingestion of sea-food high in As. It is difficult, therefore, to obtain the correlation between serum creatinine and AsB concentration. The effects of sex and age on the As accumulation were also studied.No significant sex-specific variations in arsenic concentrations can be observed for DMA, AsB and total As (p > 0.05). Table 3 lists the mean As levels and p-values for males and females. On dividing the patients according to age, most patients belonged to the age groups 40–60 (n = 19) and 60–80 years (n = 29). Only three patients were in the group 20–40 years. No significant age-specific variations in arsenic concentrations could be observed for DMA, AsB and total As in the two major groups (p > 0.05).Table 3 lists the mean As levels and p-values in these two groups. From these data, we can conclude that As accumulation is dependent on the degree of renal insufficiency but not on sex or age. As concentration changes before versus after dialysis To evaluate whether the selective accumulation of DMA in HD patients is dependent on the selective removal of this species, the percentage removal of the different As species by haemodialysis was determined. The results show that 4 h of dialysis removed a mean (±s) of 67.7 ±4.78% of total As, 66.5 ±8.18% of AsB and 66.8 ±4.27% of DMA.No significant differences (p > 0.05) were observed among the efficiencies of removal of total As, AsB and DMA as a result of the HD treatment. As no selective removal of different As species occurs, a possible hypothesis about the DMA accumulation in HD patients could be the longer residence time of inorganic As, because HD treatment is only an intermittent process.This longer residence time of inorganic As in patients’ bodies may increases the amount of in vivo methylation and so result in increased serum DMA concentrations, as illustrated in Fig. 2, where V1 is the rate of DMA formation by the methylation of inorganic As species, V2 is the rate of renal excretion of DMA and Vt is a sum of DMA formation rates. As no renal excretion but only an intermittent treatment can be carried out for the removal of DMA in the serum of HD patients, V2?minimum and Vt?maximum.A long residence of the inorganic As species in HD patients causes an accumulation of the DMA species. Fig. 1 Correlation between DMA concentrations and creatinine levels in the serum of 51 uraemic patients. Table 2 Concentrations of As species in serum of uraemic patients (n = 51) As concentration/mg l21 Uraemic non-dialysis patients Haemodialysis patients Continuous ambulatory peritoneal (n = 19) (n = 18) dialysis patients (n = 14) As species DMA AsB Total As DMA AsB Total As DMA AsB Total As Range < DL–2.60 < DL–20.3 0.50–24.8 < DL–6.9 < DL–9.4 1.6–17.5 < DL–4.16 < DL–15.4 0.3–20.6 Median < DL 2.50 3.90 1.60 2.80 4.60 < DL 1.87 2.65 Mean 0.82 3.55 5.12 1.93 3.47 6.47 0.71 3.56 4.67 s 1.05 4.58 5.58 1.51 2.89 4.28 1.05 4.27 5.41 As (%) 16.0 69.3 100 29.8 53.6 100 15.2 76.2 100 Table 3 Comparisons of As species in the serum of male and female patients and in the serum of the two major age groups of patients Mean ± s Total DMA/mg l21 AsB/mg l21 As/mg l21 Males (n = 22) 1.23 (1.23) 4.16 (5.17) 6.12 (6.31) Females (n = 29) 1.15 (1.44) 3.04 (2.66) 4.99 (4.05) p 0.72 0.86 0.70 60–80 years (n = 29) 1.13 (1.17) 3.49 (4.85) 5.26 (5.97) 40–60 years (n = 19) 1.34 (1.65) 3.46 (2.29) 5.54 (3.92) p 0.84 0.18 0.13 Analyst, January 1998, Vol. 123 15As-inorganic As-protein via intermediate of As(lll) DMA DMA V1 V1 V2 V2 = V1– V2 Vt minimum maximum When (Vt: Sum of DMA formation rates) By comparing the efficiency of As removal from serum with that from packed cells, we found that HD treatment was of very limited clinical effectiveness for As in packed cells.Compared with a mean of 67.7% for total As removal from serum, only 15.7% of total As was removed from packed cells of six patients after 4 h of HD treatment. Speciation results were not achieved in packed cells, however, as only 5–10-fold diluted packed cells lysate could be processed on the column.The concentration of the As species in most samples became lower than the detection limit after this dilution. For this reason, it is still unclear whether DMA was selectively accumulated or removed from packed cells. A more sensitive method suitable for the speciation of As species in packed cells is obviously needed for future work. The changes in total As, AsB and DMA in the dialysate of CAPD patients before versus after 24 h dialysis treatment were also studied.A very low As level was found in fresh dialysate (0.04 mg l21). The mean (±s) concentrations were, however, increased to levels of 3.98 ±4.91 mg l21 total As, 0.59 ±0.87 mg l21 DMA and 3.06 ±3.96 mg l21 AsB in 24 h dialysate. No significant differences in As levels were found between serum and drained dialysate after 24 h of dialysis (p = 0.45, 0.60 and 0.58 for total As, DMA and AsB, respectively). Analyses for total As and As species were carried out on each of the four exchanges of CAPD dialysate.No significant differences in concentrations were observed during the day for the same patient. The calculation of the distribution of the total amount of arsenic in serum, packed cells, dialysate and urine shows that 53% of the total amount of arsenic was removed into dialysate in 24 h, but only 10% into urine in 24 h. Failure of renal function causes a dramatic decrease in urinary As excretion. Speciation of HMM As species To identify the As species that bind to proteins in the serum of uraemic patients, three sera from CAPD patients were further studied. CAPD is a continuous treatment and therefore more similar to the clearing by the human kidney compared with HD treatment.The number of patients is limited because of the difficult and delicate nature of the experiments, excluding even the feasibility of a large survey for all kinds of patients. Determination of As–protein concentration by ultrafiltration The LMM As species in the serum of the 14 CAPD patients was preliminarily separated from the serum matrix by ultrafiltration.The As–protein concentrations were calculated by subtracting the concentrations of LMM As species from the total As concentrations. The mean concentration (±s) of As bound to protein molecules was 0.26 ±0.38 mg l21, accounting for 5.57% of total As in the serum (4.67 mg l21). Speciation of As–protein binding by SEC Considering that the results obtained by ultrafiltration are based on an indirect calculation, further identification of the As– protein binding in the serum of three patients with higher As concentrations was carried out by SEC.After SEC separation, arsenic was distributed into two peaks. The first peak contained HMM As species with a molecular mass of about 80 kDa and the second peak contained the LMM As species with a molecular mass of < 1 kDa. We assumed that the protein molecules and the As, detected in the same fraction, are associated with one another.The concentrations (±s) of As bound to proteins in serum for the three patients were 0.44 ±0.12, 0.19 ±0.09 and 0.59 ±0.09 mg l21, respectively. In vitro incubation of the five As species, i.e., AsIII, AsV, MMA, DMA and AsB, with serum showed that As–protein binding can only take place between inorganic As and serum proteins. On incubating 10 ng of inorganic As species with 1 ml of serum at 37 °C for 24 h, the extents of As–protein binding were 6.5 and 5.3% for AsV and AsIII, respectively. Speciation of the proteins bound to As species by IEC The SEC used in the present study is sufficient for the separation of LMM from HMM As species, but this method does not allow an adequate separation of proteins with small differences in molecular mass.The identification of the serum proteins bound to As species was therefore carried out on an anion-exchange column (MonoQ HR 16/10), because this column shows a high resolution for the separation of serum proteins. As the LMM AsIII species, carrying the same apparent charges as the macromolecules, eluted together with the protein molecules on this column, a pre-separation of the LMM As species from HMM As species by SEC was carried out before submitting the sample to anion-exchange chromatography.This separation removes the interference of AsIII with As–protein molecules. The chromatograms showed nine distinct protein regions (Fig. 3). Isoelectric focusing of the protein peaks revealed the nature of the different proteins in each peak. Peaks 4 and 5 are transferrin which may exist in two different forms: asialo- and sialo-transferrin, with iron in different sites of the molecules. Albumin is mainly situated in peaks 6 and 7, and to a lesser extent also in peaks 8 and 9. The arsenic distribution over the different fractions containing proteins was measured by Fig. 2 Schematic illustration of DMA accumulation in the serum of uraemic patients.V1 = the rate of DMA formation from inorganic As molecules and V2 = the rate of renal excretion of DMA. For HD patients, V2 ? 0 and the DMA concentration was therefore increased in the serum. Fig. 3 Chromatogram for the separation of human serum using an anionexchange column (MonoQ HR 16/10; eluent 0.025 mol l21 TRIS–HCl at pH 8 in a linear NaCl gradient from 0 to 0.5 mol l21). Peaks 4 and 5 contain transferrin; peaks 7 and 8 contain albumin. 16 Analyst, January 1998, Vol. 123HGAAS. Almost all of the As molecules were distributed over the fractions containing transferrin (peaks 4 and 5). No arsenic was detected in the fractions containing albumin. Identification of As–transferrin binding by affinity chromatography Considering the occurrence of arsenic together with transferrin in the same fractions, it was useful to look for further proof of the statement that arsenic is bound to serum transferrin. Therefore, affinity chromatography based on the immunoassay of transferrin–antitransferrin was applied.As expected, only transferrin and As–transferrin molecules were absorbed on the columns. After desorption, the fractions obtained from five repetitive elutions were collected in quartz tubes and digested and As was measured. The results showed that arsenic is indeed bound to serum transferrin. As no arsenite and arsenate were detected in the LMM fractions in serum from uraemic patients, serum transferrin may play a role as a biological storage site or reservoir in human blood to prevent arsenite-inhibitable metabolic reactions and arsenate-substitutable reactions in ATP formation prior to methylation. It is also possible that the toxic inorganic As species, transported by transferrin, go to the target sites including bone marrow, causing adverse effects. Further study is obviously needed to understand the detoxication/intoxication of As–transferrin molecules in the serum of uraemic patients.References 1 Zhang, X., Cornelis, R., De Kimpe, J., Mees, L., and Lameire, N., Clin. Chem., 1997, 43, 406. 2 Zhang, X., Cornelis, R., De Kimpe, J., Mees, L., Vanderbiesen, V., De Cubber, A., and Vanholder, R., Clin. Chem., 1996, 42, 1231. 3 Zhang, X., Cornelis, R., De Kimpe, J., Mees, L., Vanderbiesen, V., and Vanholder, R., Fresenius’ J. Anal. Chem., 1995, 353, 143. 4 De Kimpe, J., Cornelis, R., Mees, J., Van Lierde, S., and Vanholder, R., Am. J. Nephrol., 1993, 13, 429. 5 Van Renterghem, D., Cornelis, R., and Vanholder, R., J. Trace Elem. Electrolytes Health Dis., 1992, 6, 169. 6 Astrug, A., Kuleva, V., Kiriakov, Z., Tomov, A., and Djingova, R., Trace Elem. Med., 1984, 1, 65. 7 Alvadeo, A., Minoia, C., Segagni, S., and Villa, G., Int. J. Artif. Organs, 1977, 2, 17. 8 Giovannetti, S., Maggiore, Q., and Malvano, R., Nuclear Activation Techniques in the Life Sciences, IAEA Vienna, 1967, pp. 511–515. 9 Brune, D., Samsahl, K., and Wester, P. O., Clin. Chim. Acta, 1966, 13, 285. 10 Vahter, M., in Biological Monitoring of Toxic Metals, ed. Clarkson, T. W., Friberg, L., Nordberg, G. F., and Sager, P. R., Plenum Press, New York, 1988, pp. 303–321. 11 Shiomi, K., in Arsenic in the Environment, Part II: Human Health and Ecosystem Effects, ed. Nriagu, J. O., Wiley, New York, 1994, pp. 261–282. 12 Thompson, D. J., Chem. Biol. Interact., 1993, 88, 89. 13 Vahter, M., and Envall, J., Environ. Res., 1983, 32, 14. 14 Vahter, M., and Marafante, E., in The Biological Alkylation of Heavy Elements, ed. Craig, P. J., and Glockling, F., Special Publication No. 66, Royal Society of Chemistry, London, 1988, pp. 105–119. 15 Zhang, X., Cornelis, R., De Kimpe, J., and Mees, L., J. Anal. At. Spectrom., 1996, 11, 1075. 16 Zhang, X., Cornelis, R., De Kimpe, J., and Mees, L., Anal. Chim. Acta, 1996, 319, 177. 17 Zhang, X., Cornelis, R., De Kimpe, J., Mees, L., and Lameire, N., Clin. Chem., in the press. 18 Versieck, J. Vanballenberghe, L., De Kesel, A., Hoste, J., Wallaeys, B., Vandenhaute, J., Baeck, N., Steyaert, H., Byrne, A. R., and Sunderman, F. R., Jr., Anal. Chim. Acta, 1988, 204, 63. Paper 7/04841F Received July 8, 1997 Accepted September 9, 1997 Analyst, January 1998, Vol. 123 17