Twenty-four-hour urinary trace element excretion: reference intervals and interpretive issues

Abstract

Background

Introduction of inductively coupled plasma mass spectrometry (ICP-MS) into clinical laboratories has led to an increasing application of analyses to risk assessment for toxicity from environmental exposure to trace elements, and in occupational monitoring. Interpretation of results from random urine samples may be problematic and measurement of excretion over 24 h is sometimes preferable. Recent reference data are sparse.

Methods

Twenty-four-hour urine samples from 111 healthy adults from the renal stones clinic in Southampton, UK, were analysed for 31 trace elements using ICP-MS and for zinc using atomic absorption spectroscopy. Non-parametric 0.95 coverage intervals were determined for trace element excretion per 24 h and as a ratio to creatinine, for the full study cohort and separately for men (n = 77) and women (n = 34).

Results

Beryllium was undetectable in 95% of samples, bismuth in 87% and uranium in 75%. In comparison with published ranges, reference intervals for this cohort were higher for molybdenum, tin and vanadium, and for arsenic due to inclusion of fish arsenicals. Aluminium, chromium, iron, lead and mercury were lower. In our cohort, 24-h excretion of 17 elements was significantly higher in men than in women. However, when expressed as trace element to creatinine ratios, the situation reversed strikingly. Because of their lower creatinine excretion, ratios for 18 elements were significantly higher for women.

Conclusions

New adult reference intervals were obtained for 24-h urine trace element excretion. Trace element:creatinine ratios must be used cautiously, with separate ranges for men and women.

Introduction

Trace elements are defined as elements with a concentration of less than 100 μg/g or 100 μg/mL. Depletion of those which are essential for life may cause deficiency syndromes. Excessive accumulation of these and of other, non-essential, elements may cause toxicity. The range of elements that can be measured reliably in clinical chemistry laboratories increased dramatically following the introduction of atomic absorption spectroscopy in the mid-1960s. This has extended further to over 40 elements with inductively coupled mass spectrometry (ICP-MS),¹ which is now used widely. This increased versatility has been accompanied by an extension of the applications of trace element analyses. Detection of gross abnormalities associated with overt toxicity or deficiency remains a priority. However, there is now also increasing concern that excessive environmental exposure of the general public to trace elements may have long-term consequences.^2,3 Appropriate reference data are essential to identify potential risk and to monitor trends, and thereby guide public health measures to reduce exposure.^4–7

Many trace elements are excreted predominantly in urine. Exceptions are copper, zinc, iron and manganese, with predominant faecal loss.⁸ For practical reasons, random urine samples are the preferred samples for most occupational monitoring, environmental and paediatric studies. Until recently, patient results and reference data were reported most often as concentrations per litre. These are strongly influenced by fluctuations in hydration status. While adequate for diagnosis of overt toxicity, in which urinary concentrations are generally raised unequivocally, milder disturbances may be masked. Hence, in recent studies, concentrations have been related to urine creatinine.^1,2,7,9–12 However, this is also problematic. Creatinine excretion is influenced by gender,^13–15 age,^1,16,17 nutrition and urine flow rate,^18,19 and it cannot be assumed that it is concordant with trace element excretion over 24 h. An alternative has been to relate excretion to urine specific gravity.¹²

Estimation of 24-h excretion gives a more representative overview of trace element status and has been advised for detecting iodine deficiency, investigating individuals for possible subclinical abnormalities and for research studies.^20,21 However, reference data are lacking. The most substantial were from a compilation of good studies published between 1985 and 1995.⁸ More recent studies have provided data for only selected elements,^1,22 or have included children.²³

As part of their routine metabolic investigations, patients attending the renal stones clinic of the Clinical Biochemistry Department in Southampton, UK, collect an accurately timed 24- h urine sample. They are healthy outpatients seen at least four weeks after their stone episode. We have taken the opportunity to measure trace elements in their collections to provide new population reference intervals for 24-h excretion, and to examine the correlation with trace element: creatinine ratios.

Patients, materials and methods

Patients

From September 2005 to 2006, 112 subjects were recruited consecutively from the renal stones clinic after referral for investigation for biochemical risk factors for kidney stones, at least four weeks after their stone event. All were well and leading a normal life on their normal diet. Patients were excluded if they had occupational exposure to trace elements, were taking trace element supplements or drugs known to affect trace element metabolism, had joint prostheses or significant renal impairment. All those invited to participate accepted and gave written informed consent to the study. Data from one woman aged 68 y were subsequently rejected because the 24-h collection was incomplete. The remaining 111 subjects were aged 21–85 y (median 51.5 y); 77 were men (aged 22–71 y, median 51) and 34 were women (aged 21–85 y, median 54.5).

Urine samples

After careful instruction by a clinic doctor, subjects collected a 24-h urine sample into new trace element-free polyethylene containers with 5 g of thymol as preservative. In preliminary studies, there was no evidence that thymol interfered with any of the trace element analyses: 1 L of water was added to five 24-h urine containers containing thymol. After thorough mixing, 20-mL aliquots were removed, stored and subsequently analysed for each trace element. In all cases, the elements were below detection limits, defined as more than 3 standard deviations of the method blank. Men urinated directly into the container or through a clean polythene funnel. Women used a clean polythene jug. They were instructed not to use detergents or sterilizing agents. As is the routine practice in clinical laboratories, the urine volumes were recorded by weighing using the approximation 1 L = 1 kg, without correction for specific gravity. Urine samples (40–60 mL) were aliquoted for trace element analyses and stored at −20°C. A further aliquot was taken for creatinine measurement.

Reagents

Rhodium and certified single-standard preparations of trace elements (1000 mg/L) were from BDH (British Drug Houses, Poole, UK) and Fisher Scientific (Loughborough, UK). Nitric acid (70%, trace analysis grade), chemical reagents, analytical reagent-grade ammonia solution (35%) and butan-1-ol were from Fisher Scientific. Methane (100%) and ammonia gas (vapour pressure 6.2 bar at 15°C) were from British Oxygen Company (Manchester, UK); water was de-ionized (Milli-Q system; Millipore, Watford, UK).

Sample preparation and analyses

The urine samples for the study were analysed for trace elements over approximately 12 months, slotted into the diagnostic work schedule of the laboratory, with the validated procedures in routine use. For zinc, this was still with atomic absorption spectroscopy and for the other 31 elements with ICP-MS, using published procedures which have been modified and validated in our laboratory.^24–28 Although ICP-MS has the capacity to measure groups of elements simultaneously, the elements were measured individually, in accordance with the current laboratory policy to restrict diagnostic analyses to requested tests.

After thawing and thorough mixing, samples for most elements were diluted in 0.01% nitric acid (1 in 7.5 to 1 in 30 dilutions). For iodine estimation, urine samples were diluted 1 in 15 in 0.3% ammonia. For arsenic, nickel and selenium analyses, samples were diluted 1 in 15 in 1% butan-1-ol in order to increase ionization and hence sensitivity.²⁷ Rhodium (final concentration 15 μg/L) was added as internal standard for most analyses since, despite the wide mass range of the elements measured, it was shown to be suitable in the initial validation studies. For historic reasons, for mercury and iodine the internal standards were thallium (15 μg/L) and tellurium (150 μg/L), respectively, which had been used to set up the assays in routine use during the study. For flame atomic spectroscopy analysis of zinc, samples were aspirated without dilution.

Individual elements were measured in batches of around 50 patient samples together with internal quality controls by ICP-MS using a PerkinElmer Elan 6100 DRC Plus, inductively coupled plasma mass spectrometer (PerkinElmer SCIEX, Beaconsfield, UK), with a closed-loop water chiller (Polyscience, Niles, IL, USA) for cooling, an AS90plus autosampler (PerkinElmer), cross-flow nebulizer (PerkinElmer) and Dell OptiPlex GX620 computer (Dell Computer corporation, Bracknell, UK) for data processing. The nebulizer flow rate was 0.95–1.00 mL/min. In order to eliminate argon-based polyatomic interferences, a dynamic reaction cell (DRC) pressurized with methane (flow rate 0.5 mL/min) for selenium analysis and with ammonia gas (flow rate 0.4 mL/min) for aluminium, chromium, iron, manganese, silicon and vanadium analyses was used. The instrument settings were radio frequency power: 1250 W; argon gas flow rates: nebulizer 0.9–1.0 L/min, auxiliary 1.2 L/min and plasma 15 L/min; data acquisition parameters: scanning mode: peak hopping; mass: masses of analyte and internal standard; dwell time: 20–100 ms (200 ms in DRC mode); 100 sweeps per reading (20 in DRC mode); one reading per replicate; three replicates (Table 1). Calibrants were prepared from certified analytical standard solutions (1000 mg/L) in 0.1% nitric acid with addition of fresh urine as a matrix modifier.

Table 1

ICP-MS method parameters, internal quality controls (QC) and External Quality Assurance (EQA) schemes

Element	Ion mass	Internal standard	DRC gas	Dwell time (ms)	Sweeps per reading	Internal quality control material	EQA schemes
Aluminium	27	¹⁰³Rh	Ammonia 0.4 mL/min	100	20	Biorad 69091 Biorad 69092	TEQAS Quebec
Antimony	121	¹⁰³Rh		50	100	Biorad 69091Biorad 69092	Quebec
Arsenic	75	¹⁰³Rh		50	100	Biorad 69091Biorad 69092	Quebec
Barium	138	¹⁰³Rh		50	100	Seronorm NO2525Seronorm OK4336	Quebec
Beryllium	9	¹⁰³Rh		50	100	Seronorm NO2525	Quebec
Bismuth	109	¹⁰³Rh		50	100	Seronorm NO2525	Quebec
Boron	10	¹⁰³Rh		50	100	Seronorm NO2525
Cadmium	111	¹⁰³Rh		50	100	Seronorm NO2525	TEQASQuebec
Chromium	52	¹⁰³Rh	Ammonia 0.35 mL/min	200	20	Biorad 69091 Biorad 69092	Quebec
Cobalt	59	¹⁰³Rh		50	100	Biorad 69091Biorad 69092	Quebec
Copper	65	¹⁰³Rh		50	100	Biorad 69091Biorad 69092	TEQASQuebec
Gold	197	¹⁰³Rh		50	100	Seronorm JL4409 (1 in 100 dilution)
Iodine	127	¹²⁶Te		50	100	Seronorm NO2525	Quebec
Iron	54	¹⁰³Rh	Ammonia 0.3 mL/min	200	20	Seronorm 0511545	TEQAS
Lead	208	¹⁰³Rh		50	100	Biorad 69091Biorad 69092	Quebec
Lithium	7	¹⁰³Rh		50	100	Seronorm 0511545
Manganese	55	¹⁰³Rh	Ammonia 0.3 mL/min	200	20	Biorad 69091Biorad 69092	Quebec
Mercury	202	²⁰⁵Tl		50	100	Quebec QC	TEQASQuebec
Molybdenum	98	¹⁰³Rh		100	100	Seronorm 0511545	Quebec
Nickel	62	¹⁰³Rh		100	100	Biorad 69091Biorad 69092	Quebec
Platinum	195	¹⁰³Rh		50	100	Quebec QC	Quebec
Selenium	78	¹⁰³Rh	Methane 0.5 mL/min	200	20	Seronorm NO2525Seronorm OK4336	Quebec
Silicon	28	¹⁰³Rh	Ammonia 0.4 mL/min	200	20	Seronorm 0511545
Silver	109	¹⁰³Rh		50	100	Spiked urine	Quebec
Strontium	88	¹⁰³Rh		50	100	Seronorm 0511545	Quebec
Tellurium	125	¹⁰³Rh		50	100	Seronorm 0511545	Quebec
Thallium	205	¹⁰³Rh		50	100	Seronorm OK4336Biorad 69091	Quebec
Tin	118	¹⁰³Rh		50	100	Seronorm 0511545	Quebec
Tungsten	182	¹⁰³Rh		50	100	Seronorm 0511545
Uranium	238	¹⁰³Rh		100	100	Seronorm 0511545Quebec QC	Quebec
Vanadium	51	¹⁰³Rh	Ammonia 0.4 mL/min	200	20	Seronorm 0511545Seronorm OK4336	Quebec
Zinc	Analysed by AAS					Biorad 69091Biorad 69092	TEQASQuebec

AAS, atomic absorption spectroscopy; TEQAS, United Kingdom Trace Element Quality Assurance Scheme; Quebec, Quebec Multielement External Quality Assurance Scheme; DRC, dynamic reaction cell

Zinc was measured by atomic absorption spectrophotometry using a PerkinElmer AAnalyst 200 atomic absorption spectrometer (PerkinElmer). Calibrants were prepared by diluting working zinc stock standard (100 μmol/L of water) in 14 mmol/L sodium chloride containing 5 mmol/L hydrochloric acid. Creatinine was analysed using a Beckmann Coulter Unicel DXC 800 autoanalyser (Beckman, High Wycombe, UK) by a kinetic Jaffe reaction.

Accuracy and imprecision

Commercially available certified reference materials were analysed 10 times to determine accuracy and imprecision: Bio-Rad Lyphochek urine metal control levels 1 and 2 (product numbers 400 and 405; Bio-Rad Laboratories, Irvine, CA, USA); Seronorm urine and Seronorm urine blank (Sero AS, Billingstad, Norway). For silver, certified reference material was not available and urine from the Quebec Multielement External Quality Assurance Scheme (QMEQAS; Institut National de Sante Publique du Quebec, Canada) with assigned concentration was used. For all analyses, the analytical results were within the target ranges (Table 2). This table also presents data for imprecision at levels close to those found in the study. The lower limits of detection (LODs) were calculated as 3 standard deviations of the mean of 10 replicate analyses of the reagent blanks. Accuracy and imprecision were monitored by analysis of controls at the beginning of each batch and after every five samples of each batch, and by participation in external quality assurance schemes (QMEQAS; United Kingdom Trace Element Quality Assurance Scheme: UKTEQAS) (Table 1).

Table 2

Detection limits, analytical imprecision and accuracy

		Imprecision			Comparison with certified reference materials*
Element Units	Detection limit per litre	Mean concentration per litre	Within batch CV %	Between batch CV %	Mean (SD)	Target (range)
Aluminium (nmol)	30	1000	3.2	5.1	1112 (33.4)	955 (764–1146)
					2005 (66.7)	1839 (1471–2206)
Antimony (nmol)	0.7	85	2.0	7.6	85 (1.6)	(61–92)
					846 (19.7)	(813–862)
Arsenic (μmol)	0.001	1.00	1.2	6.6	1.00 (0.012)	0.84 (0.67–1.01)
					2.28 (0.064)	2.03 (1.65–2.47)
Barium (nmol)	0.8	175	2.4	4.7	68 (5.8)	70 (40–100)
Beryllium (nmol	1.1	519	2.8	3.3	520 (11.1)	557 (539–575)
		52	4.9	14.7
Bismuth (nmol)	0.1	11	2.7	3.3	0.19 (0.096)	<0.5
					104.3 (3.35)	100 (96–104)
Boron (μmol)	0.46	70	3.3	3.7	66 (2.2)	64 (59–69)
Cadmium (nmol)	0.2	15	2.3	5.1	45 (1.2)	45 (40–50)
		44.5	2.9	1.9	88 (3.7)	(65–95)
					140 (3.9)	(123–185)
Chromium (nmol)	0.4	11.9	6.5	8.4	13.5 (0.06)	12.7 (10.2–15.3)
					479 (19.2)	386 (309–463)
Cobalt (nmol)	0.1	130	1.2	5.2	131 (1.7)	122 (98–147)
					345 (10.2)	320 (256–384)
Copper (nmol)	7	500	2.2	9.7	157 (14.2)	170 (140–200)
		900	3.6	6.3	768 (33.0)	850 (680–1020)
Gold (nmol)	0.1	2	1.9	5.3	469 (17.8)	(432–584)
Iodine (μmol)	0.01	0.7	2.5	4.6	2.13 (0.07)	2.22 (2.08–2.36)
		2.2	1.2	2.9
Iron (nmol)	20	500	4.9	7.8	265 (27.4)	220
		1000	1.9	2.8
Lead (nmol)	0.3	66	4.4	7.5	67 (2.9)	76 (61–91)
					321 (13.9)	340 (270–410)
Lithium (μmol)	0.01	1.44	2.5	2.1	1.44 (0.043)	1.47
Manganese (nmol)	2	150	3.7	6.3	157 (5.5)	152 (121–182)
					664 (20.6)	594 (475–713)
Mercury (nmol)	0.5	20	2.8	23	18 (0.6)	24 (14–34)
		130	1.6	4.2
Molybdenum (nmol)	0.6	170	1.7	No data	445 (9.4)	441 (427–454)
		450	2.1	2.1
Nickel (nmol)	15	852	6.4	8.9	51 (6.8)	55 (44–66)
					405 (23.8)	407 (325–488)
Platinum (nmol)	0.05	5.0	2.6	2.7	5.1 (0.15)	(4.6–5.6)
Selenium (nmol)	10	300	2.0	3.6	271 (5.1)	270 (230–310)
					849 (15.2)	847 (757–937)
Silicon (μmol)	1.18	40.0	4.3	14.3	183 (12)	191
Silver (nmol)	0.2	9.0	4.4	2.6	46 (2.8)	46 (41–52)^†
Strontium (μmol)	0.001	1.01	2.7	5.4	1.05 (0.03)	1.04 (0.97–1.11)
Tellurium (nmol)	0.2	3.0	13	10.6	198 (4.7)	194 (182–205)
		200	2.4	3.9
Thallium (nmol)	0.08	1.0	4.5	7.3	42 (0.98)	45 (36–53)
Tin (nmol)	0.2	30	5.5	2.3	28 (1.7)	(25.3–28.6)
					226 (16.0)	261 (178–345)
Tungsten (nmol)	0.05	0.71	4.0	13.7	0.71 (0.05)	0.92 (0.65–1.20)
		13.76	0.9	7.10
Uranium (nmol)	0.04	1.0	10.0	5.9	0.12 (0.009)	<0.21
					0.95 (0.095)	(0.63–1.05)
Vanadium (nmol)	1.0	500	3.5	2.6	10.1 (2.0)	(7.9–11.8)
					493 (17.7)	493 (471–514)
Zinc (μmol)	0.09	13.0	1.0	3.6	7.2 (0.4)	7.5 (6.0–9.0)
					13.0 (0.1)	14 (12–17)

Imprecision, 10 replicate analyses of control materials at concentrations close to physiological; CV%, coefficient of variation (%)

Detection limits: mean + 3 standard deviations of 10 reagent blank analyses

*Ten aliquots of a certified reference preparation for each element were analysed

^†Assigned values for urine from the Quebec Multielement External Quality Assurance Scheme; certified reference material was not available

Statistical analysis and calculation of reference intervals

Analytical data and urine volumes were entered onto Microsoft Excel spread sheets (Microsoft Office 97–2003). Samples with undetectable trace element concentrations were assigned a concentration of LOD/2 in order to include them in statistical evaluations.^2,6,11 Excretion per 24 h and the ratios of trace element to creatinine concentration were calculated for all subjects and for men and women separately. The distributions of the 24-h data were examined for each element (Analyse-It statistical package, Leeds, UK) and clear outliers identified visually. These samples were excluded from further statistical evaluations for that element. In view of the wide range of concentrations found, mathematical criteria were not used to trim more outliers from the data, because of the risk of producing a range that was too exclusive and inappropriate for the general population. After removing outliers, none of the data had a Gaussian distribution and the Log-transformed data were then tested for normality with the Kolmogorov–Smirnov test using online statistical package (http://www.physics.csbsju.edu/stats/descriptive2.html).

Non-parametric 0.95 reference intervals (between the 0.025 and the 0.975 fractiles) were calculated as advised by the International Federation of Clinical Chemistry for datasets of less than 120 samples.²⁹ For the women, the smallest group in the study (up to 34 values), the 2.5th and 97.5th percentiles were calculated using Microsoft Excel. For the larger datasets, up to 111 samples from the entire group, or 77 samples from men, the non-parametric 0.95 coverage intervals with confidence 0.95 and coverage uncertainty at 0.95 were calculated, using the guidelines of the International Union for Pure and Applied Chemistry (IUPAC).³⁰ The Mann-Whitney U test was used to compare values for men and women, and Spearman's rank test to investigate correlation of trace element:creatinine ratios to 24-h excretion. For statistical computations, Analyse-it and GraphPad Prism 5 (GraphPad, La Jolla, CA, USA) packages were used.

Results

Urine volume and 24-h creatinine excretion

The 24-h urine volume for all 111 subjects was 0.47–5.25 L, median 1.98 L, and was not significantly different for men (0.62–3.62 L, median 1.98 L, n = 77) and women (0.47–5.25 L, median 2.01 L, n = 34).

The 24-h creatinine excretion for all subjects was 3.15–24.88 mmol, median 13.23 mmol. It was significantly higher for men: 8.97–24.88 mmol, median 15.00 mmol, than women: 3.15–14.52 mmol, median 9.27 mmol (P < 0.0001, Mann-Whitney U test).

Trace element concentrations below the detection limits and outliers

For 111 samples, beryllium was below detection limits in 105 (95%; LOD 1.1 nmol/L), bismuth in 97 (87%; LOD 0.1 nmol/L) and uranium in 83 (75%; LOD 0.04 nmol/L). The other elements were detectable in all, or the majority, of samples (Table 3). No statistical processing of the beryllium and bismuth results was undertaken. From scatter plots of 24-h excretion of the other 30 elements, clear outliers were identified in 22 samples. There were three outliers for copper, two each for cadmium, iodine, lead, lithium and tungsten and one for each of antimony, barium, boron, chromium, manganese, mercury, nickel, platinum and selenium. Intravenous urography using ¹²⁷I explained increased iodine excretion by two patients. There were no obvious explanations for the other abnormalities. Raised lithium was not attributable to reported therapy. There were no clinical suspicions of Wilson's disease in the three individuals with raised copper excretion, and all had normal liver function tests. Serum caeruloplasmin was measured in two, and was normal.

Table 3

Twenty-four-hour excretion of trace elements: 111 adults

Element units	Number* (number <LOD)	Per 24 h Geometric mean^†	Median	0.95 coverage interval^‡	Per mmol creatinine Geometric mean^†	Median	0.95 coverage interval^‡
Aluminium (nmol)	111 (17)		200	ND–816		16.4	ND–58.1
Antimony (nmol)	110 (16)		2.4	ND–6.5	0.18	0.17	ND–0.52
Arsenic (μmol)	111 (0)		0.48	0.10–6.1		0.04	0.01–0.77
Barium (nmol)	110 (0)	19.0	18.5	3.3–95.9	1.5	1.6	0.29–7.1
Beryllium (nmol)	111 (105)		ND	ND–2.7^§		ND	ND–0.2^§
Bismuth (nmol)	111 (97)		ND	ND–0.72^§		ND	ND–0.1^§
Boron (μmol)	110 (0)		138	59.9–434		10.7	5.1–38.5
Cadmium (nmol)	109 (0)		8.4	4.1–29.0		0.67	0.35–2.0
Chromium (nmol)	110 (1)		4.8	1.9–17.5		0.38	0.17–1.2
Cobalt (nmol)	111 (0)	27.0	26.3	13.1–76.0	2.1	2.2	0.97–4.2
Copper (nmol)	108 (0)	340	340	208–717		24.7	17.1–72.1
Gold (nmol)	111 (0)		0.56	0.18–5.1		0.09	0.01–0.58
Iodine (μmol)	109 (0)		1.4	0.60–4.3	0.12	0.11	0.05–0.36
Iron (nmol)	111 (0)	210	210	91.8–533	15.1	15.1	7.0–53.7
Lead (nmol)	109 (1)		11.5	1.3–39.4		0.85	0.11–3.3
Lithium (μmol)	109 (0)		2.7	1.1–5.8	0.21	0.21	0.11–0.42
Manganese (nmol)	110 (6)	12.4	12.6	ND–46.8	0.97	0.96	ND–4.0
Mercury (nmol)	110 (5)	3.3	3.3	0.68–13.4	0.26	0.26	ND–1.1
Molybdenum (nmol)	111 (0)		692	167–1850	51.8	52.9	18.2–134
Nickel (nmol)	110 (19)		86.7	ND–275		6.2	ND–18.9
Platinum (nmol)	110 (5)		0.13	ND–0.73	0.02	0.02	ND–0.07
Selenium (nmol)	110 (0)		298	154–667		22.8	14.1–49.6
Silicon (μmol)	111 (0)	449	457	179–1077	35.3	34.9	16.0–74.9
Silver (nmol)	111 (0)	3.8	3.8	0.82–15.1	0.30	0.31	0.07–1.8
Strontium (μmol)	111 (0)		2.3	0.55–5.6		0.18	0.05–0.39
Tellurium (nmol)	111 (0)	5.7	6.2	1.3–21.8		0.48	0.10–1.4
Thallium (nmol)	111 (0)		1.5	0.56–3.0		0.11	0.05–0.34
Tin (nmol)	111 (0)	19.5	19.9	4.8–77.9		3.1	0.28–13.7
Tungsten (nmol)	109 (0)	48.3	48.7	29.4–83.2		3.6	2.9–8.3
Uranium (nmol)	111 (83)		ND	ND–3.5^§		ND	ND–0.27^§
Vanadium (nmol)	111 (9)		38.0	ND–317		2.6	ND–27.6
Zinc (μmol)	111 (0)	7.0	7.0	3.0–19.3		0.59	0.19–1.3

LOD: lower limit of detection; ND: below detection limits

*Number of samples analysed after excluding outliers

^†For elements with log-normal distribution

^‡Confidence 0.95; coverage uncertainty 0.95 ± 0.04

^§Undetectable in most samples; observed ranges shown

Trace element excretion of the full study cohort

Table 3 presents the data for 24-h excretion of all 111 subjects. Geometric means are presented for 12 elements which were normally distributed after log transformation. However, for all elements except beryllium, bismuth and uranium, the 95% coverage intervals were derived from non-parametric analyses, and these are tabulated with median values. Table 3 also presents ratios of trace element to creatinine concentration for the 24-h urine samples. The log-transformed data of the ratios for 12 elements were normally distributed.

Trace element excretion of men and women compared

Table 4 presents the separate data for 77 men and 34 women. Twenty-four-hour excretion of 17 elements was higher for men, with the most significant statistical differences (P < 0.001 or P < 0.0001, Mann-Whitney U test) observed for cadmium, cobalt, lithium, molybdenum, nickel, selenium, silicon, strontium, tungsten and zinc. Women had a small increase in gold (P < 0.05). Comparison with results for the trace element:creatinine ratios showed a striking reversal of the situation, with significantly higher values for 18 elements in women and none increased in men. This anomaly was particularly apparent from the paired data for boron, chromium, cobalt, iron, lithium, selenium, silicon, thallium and tungsten.

Table 4

Trace element excretion of 77 men and 34 women

	Per 24 h			Per mmol creatinine
Element (units)	Men^† Range Median	Women^‡ Range Median		Men^† Range Median	Women^‡ Range Median
Aluminium (nmol)	ND–817	28.3–405	NS	ND–63	2.8–46.8	F*
	204	206		14.0	21.3
Antimony (nmol)	ND–7.1	ND–4.7	NS	ND–0.43	ND–0.55	F**
	2.4	2.6		0.16	0.25
Arsenic (μmol)	0.10–10.3	0.08–3.3	NS	0.01–0.67	0.01–0.78	F****
	0.45	0.57		0.04	0.11
Barium (nmol)	4.1–96.6	3.0–76.5	NS	0.29–6.8	0.28–7.1	NS
	17.7	19.7		1.3	2.2
Boron (μmol)	79.0–462	39.7–425	M*	5.1–25.6	5.3–45.4	F*
	144	116		10.1	11.9
Cadmium (nmol)	4.6–29.3	3.9–12.2	M****	0.31–2.3	0.39–1.3	NS
	9.2	7.0		0.63	0.72
Chromium (nmol)	1.9–18.6	2.1–8.5	M*	0.16–1.4	0.21–1.1	F*
	5.1	4.0		0.40	0.36
Cobalt (nmol)	15.1–77.5	9.7–42.1	M***	0.10–5.0	0.93–3.8	F*
	28.2	22.1		2.1	2.7
Copper (nmol)	220–675	160–720	NS	16.4–47.0	19.6–76.2	F****
	350	290		22.5	33.3
Gold (nmol)	0.19–3.5	0.18–6.3	F*	0.01–0.29	0.02–0.71	F****
	0.51	0.94		0.07	0.25
Iodine (μmol)	0.69–4.3	0.55–4.5	M**	0.05–0.24	0.06–0.44	NS
	1.7	1.2		0.11	0.14
Iron (nmol)	91.8–626	89.3–469	M*	6.1–43.6	10.4–59.4	F**
	230	180		14.8	16.2
Lead (nmol)	1.4–40.9	1.2–26.0	NS	0.08–3.8	0.17–3.1	NS
	11.8	10.7		0.81	1.1
Lithium (μmol)	1.7–7.0	0.82–4.5	M****	0.11–0.41	0.11–0.50	F*
	2.8	2.2		0.20	0.21
Manganese (nmol)	ND–46.8	ND–41.1	NS	ND–2.8	ND–4.4	F**
	12.2	16.7		0.78	1.3
Mercury (nmol)	ND–16.9	ND–7.7	M*	ND–1.2	ND–0.94	NS
	3.6	2.7		0.26	0.27
Molybdenum (nmol)	265–1948	150–1235	M****	16.1–187	20.0–119	NS
	737	456		52.9	52.8
Nickel (nmol)	ND–288	ND–138	M****	ND–21.6	ND–18.6	NS
	98.7	47.82		4.4	8.4
Platinum (nmol)	ND–0.67	ND–0.81	NS	ND–0.05	ND–0.07	F***
	0.13	0.143		0.02	0.03
Selenium (nmol)	176–667	77.2–555	M****	14.1–41.6	13.9–55.0	F*
	332	233		21.9	25.2
Silicon (μmol)	229–1170	177–716	M***	15.3–92.0	20.3–73.2	F*
	496	368		34.3	41.8
Silver (nmol)	0.82–17.2	1.2–13.6	NS	0.07–0.87	0.13–2.2	F***
	3.8	3.9		0.26	0.42
Strontium (μmol)	0.95–5.6	0.42–4.7	M***	0.06–0.33	0.04–0.52	NS
	2.51	1.65		0.17	0.19
Tellurium (nmol)	1.5–28.3	1.3–12.2	M*	0.10–1.4	0.12–1.2	NS
	6.6	4.2		0.44	0.54
Thallium (nmol)	0.71–3.0	0.21–3.2	M*	0.06–0.17	0.04–0.37	F***
	1.6	1.3		0.10	0.14
Tin (nmol)	5.0–81.1	4.7–57.4	NS	0.24–13.5	0.74–15.8	F*
	21.5	15.9		2.6	4.1
Tungsten (nmol)	34.8–83.2	7.7–78.6	M****	2.9–5.0	2.3–10.3	F***
	51.2	36.7		3.4	4.0
Vanadium (nmol)	ND–410	ND–153	NS	ND–27.6	ND–16.9	NS
	38.0	36.4		2.5	3.9
Zinc (μmol)	3.2–23.7	1.8–9.4	M****	0.19–1.3	0.21–1.15	NS
	8.3	5.0		0.60	0.56

ND, below detection limits

^†Ranges for men (n= 76 or 77 samples after excluding outliers): 0.95 coverage interval; confidence 0.95; coverage uncertainty 0.95 ± 0.049

^‡Ranges for women (n= 32–34 samples after excluding outliers): 2.5th to 95th percentiles

Difference between men and women (Mann-Whitney U test): NS: not significant – P> 0.05

*P< 0.05; **P< 0.01; ***P< 0.001; ****P< 0.0001; M = higher in men; F = higher in women

Scatter plots showed that trace element:creatinine ratios and 24-h trace element excretion were linearly related. Correlation was generally good. For men, Spearman's correlation coefficient r was ≥0.95, 0.94–0.85 or 0.84–0.75 for 8, 14 and 4 elements, respectively, and 0.74–0.65 for copper, lithium and selenium. For women, r was ≥0.95, 0.94–0.85 or 0.84–0.75 for 7, 8, and 10 elements, respectively, 0.74–0.65 for cadmium, cobalt and silicon and 0.64 for selenium. There was poor correlation for tungsten (r = 0.30 for men; 0.53 for women).

Discussion

This study has provided badly needed reference intervals for urinary trace elements which are representative of healthy adults with normal trace element exposure. Excretion per 24 h was measured because this is currently the most informative laboratory indicator of the body status of many trace elements. Trace element:creatinine ratios were also calculated since these are often applied to random urine samples. As previously reported, beryllium, bismuth and uranium were undetectable in most samples.^1,11

The suitability of using patients from the renal stones clinic is a key issue. It could be conjectured that stone formers have a primary or acquired abnormality of the urinary tract which affects renal handling of trace elements. Neither of these possibilities is likely. Individuals with significant renal impairment were excluded from the study, and investigation was at least four weeks after the last stone episode. The majority of renal stone formers have structurally normal kidneys when investigated by sensitive imaging techniques. Further, published studies have not found any consistent, persisting urinary trace element abnormalities among stone formers. Analysis of 14 trace elements found decreased urinary nickel, manganese and lithium within two weeks after passage or removal of a stone, but no differences from controls 12 months after a stone episode.³¹ Other studies have produced conflicting results with reports of low, increased or no abnormality of zinc excretion and raised urinary copper.³² There is no evidence that trace elements at biological concentrations have a primary role in stone formation.³² The similarity of trace element excretion found in this study to published data (Tables 5 and 6) is strong supporting evidence that the cohort was suitable. In all other respects, the group was representative of the local adult population with typical non-occupational environmental exposure. Our exclusion criteria were less stringent than those proposed for the EURO TERVIHT (European Trace Element Reference Values in Human Tissues) project to collect reference data across the European Union,³⁴ and the strict criteria applied in other studies.^4,6,35 However, there is a risk that with excessive exclusions, the findings will be unrepresentative for the general population,⁹ with healthy people worried unnecessarily by apparently abnormal results.

Table 5

Excretion of trace elements per 24 h: study reference intervals compared with published data

Element (units)	Study0.95 coverage interval	Iyengar⁸	Komaromy-Hiller et al.¹ 2.5–97.5 percentiles	Dlugaszek et al.²³ 5–95 percentiles*	Rasmussen et al.³³ Mean ± SD (median)
Aluminium (nmol)	ND–816	370–3706	ND–2254
Antimony (nmol)	ND–6.5	<0.8–41?^†
Arsenic (μmol)	0.10–6.1	0.13–0.40	ND–0.94
Barium (nmol)	3.3–95.9	14.6–29.1
Boron (μmol)	59.9–434	46.3–278
Cadmium (nmol)	4.1–29.0	8.9–44.0	ND–18.7	2.7–42.7
Chromium (nmol)	1.9–17.5	3.8–38.5	9.6–128.9
Cobalt (nmol)	13.1–76.0	8.5–33.9
Copper (nmol)	208–717	472–944	ND–2329	63–1904
Gold (nmol)	0.18–5.1	0.005–3.0?^†
Iodine (μmol)	0.60–4.30	0.79–1.58			1.10 ± 0.46 (1.01)
Iron (nmol)	92–533	1791–3581		340–7090
Lead (nmol)	1.3–39.4	48.3–96.5	0–38.6	4.8–412
Lithium (μmol)	1.1–5.8	3.6–14.4
Manganese (nmol)	ND–46.8	9.1–36.4
Mercury (nmol)	0.68–13.4	24.9–99.7	ND–67.8
Molybdenum (nmol)	167–1850	208–312
Nickel (nmol)	ND–275	34–136	0–182
Platinum (nmol)	ND–0.73	<5.1?^†
Selenium (nmol)	154–667	317–633
Silicon (μmol)	179–1077	107–427
Silver (nmol)	0.82–15.1
Strontium (μmol)	0.55–5.6
Tellurium (nmol)	1.3–21.8
Thallium (nmol)	0.56–2.98	0.49–4.89	undetectable
Tin (nmol)	4.8–77.9	8.4–16.9
Tungsten (nmol)	29.4–83.2
Uranium (nmol)	ND–3.5	0.02–0.29
Vanadium (nmol)	ND–317	3.9–19.6
Zinc (μmol)	3.0–19.3	6.1–9.2	1.5–25.4	2.6–40

ND, below detection limits

*n = 74; 22 (30%) were aged <18 y

^†?=uncertain (as published)

Table 6

Excretion of trace elements per mmol of creatinine: study reference intervals compared with published data

Element (units)	Study0.95 coverage interval	Komaromy-Hiller et al.¹ 2.5–97.5 percentiles	White and Sabbioni⁹ 2.5–97.5 percentiles	Paschal et al.² 95th percentile	Beneš et al.⁷ 95th percentile	MiscellaneousBatáriová et al.¹⁰ 95th percentileIversen et al.¹³ 2.5–97.5 percentiles
Aluminium (nmol)	ND–58.1	ND–210	6.2–145
Antimony (nmol)	ND–0.52	ND		2.6
Arsenic (μmol)	0.01–0.77	ND–0.09	0.002–0.020
Barium (nmol)	0.29–7.1	0.49–7.7		6.2
Beryllium (nmol)	ND–0.2*	ND–1.3		7.8
Bismuth (nmol)	ND–0.1*	ND–0.05
Boron (μmol)	5.1–38.5
Cadmium (nmol)	0.35–2.0	ND–1.4	0.05–2.1		2.4	1.3¹⁰
Chromium (nmol)	0.17–1.2	0.65–9.8	0.05–2.6	1.3
Cobalt (nmol)	0.97–4.2	ND–3.8	0.15–4.1	9.1
Copper(nmol)	17.1–72.1	2.3–155			160
Gold (nmol)	0.01–0.58	ND–1.03
Iodine (μmol)	0.05–0.36
Iron (nmol)	7.0–53.7
Lead (nmol)	0.11–3.3	ND–3.3	1.6–27.2	3.0	8.8
Lithium (μmol)	0.11–0.42
Manganese (nmol)	ND–4.0		0.17–3.14	4.98
Mercury (nmol)	ND–1.1	ND–4.6			2.1	1.0¹⁰
Molybdenum (nmol)	18.2–134	10.8–81.3		158		4.6–150¹¹
Nickel (nmol)	ND–18.9	ND–27.4	0.45–10.8
Platinum (nmol)	ND–0.07			2.1
Selenium (nmol)	14.1–49.6	67.2–188	13.6–54.8		63.6
Silicon (μmol)	16.0–74.9
Silver (nmol)	0.07–1.8	ND
Strontium (μmol)	0.05–0.39	0.03–0.38
Tellurium (nmol)	0.10–1.4	ND
Thallium (nmol)	0.05–0.34	ND		0.42
Tin (nmol)	0.28–13.7	ND–38.0		15.3
Tungsten (nmol)	2.9–8.3			1.7
Uranium (nmol)	ND–0.27*	ND
Vanadium (nmol)	ND–27.6
Zinc (μmol)	0.19–1.3	0.02–1.6			4.6

ND, below detection limits

*Observed ranges; beryllium undetectable in 95% of samples, bismuth in 87% and uranium in 75% of samples

Several large studies have produced reference data from random urine samples for concentrations of trace elements per litre of urine,^2,4,35–37 or as a ratio to creatinine^1,7,9–12 or for both.^2,11 In contrast, published data for 24-h urine samples are sparse. The strongest were compiled from carefully scrutinized studies published from 1985 to 1995, to compute the total body content of selected trace elements for a Reference Man: a Caucasian, weight 70 kg, height 170 cm, taking a Western diet.⁸ Data for 24-h excretion of 10 trace elements were reported from a large study in the USA;¹ however, this included samples analysed to monitor occupational exposure and used questionable statistics to derive reference intervals. Another large study included children and measured only zinc, copper, iron, lead and cadmium.²³ Ranges have been reported for zinc²² and iodine.³³ Generally, the ranges from our cohort agree closely with those proposed for the Reference Man and the other reported data (Table 5). The higher arsenic excretion is explained by inclusion of non-toxic organic compounds, mainly arsenobetaine, from fish and other sea foods in our ICP-MS assay, which measures total arsenic. The study patients were on an unrestricted diet. Urine silicon levels were also high, as reported by Iyengar.⁸ Possible analytical interferences in the measurement of Si28 by ICP-MS cannot be excluded; however, these were reduced by use of a DRC pressurized with ammonia and the results for Seronorm-certified reference material were within the target range. Improved analytical sensitivity probably explains the higher levels found for barium and vanadium, which were previously difficult elements to measure. There are no obvious explanations for the higher ranges found for molybdenum or tin. The low chromium levels are almost certainly due to reduced sample contamination by this ubiquitous element. The lower levels of lead may reflect decreased environmental exposure as a result of legislation.⁹ This may also explain the lower mercury levels found here and observed in a German survey.³⁷ However, mercury has poor stability in urine and pre-analytical losses cannot be excluded. When related to creatinine, with the exception of higher arsenic ratios, trace element excretion in this study was also generally similar to reported ranges (Table 6).

Men excreted significantly larger amounts of 17 elements per 24 h than women (Table 4). Increased dietary intake may account for the higher urinary iodine excretion.³³ However, iodine is concentrated in the thyroid, salivary glands and stomach, and larger organ mass and body stores may be contributory. Similarly, larger size probably explains the higher zinc (muscle mass), molybdenum, selenium and possibly cobalt (liver), silicon (skin and arterial walls), strontium (bone), lithium (thyroid and bones) and perhaps cadmium (liver and kidneys). Cigarette smoking, not recorded, may also increase cadmium excretion. Only gold was higher in women (P < 0.05). The increase was small and of dubious significance.

Expression of trace element excretion as ratios to creatinine produced an anomalous and potentially misleading situation (Table 4). Although the actual excretion per 24 h of nine elements was significantly higher in men, when expressed as a creatinine ratio, women had significantly higher values. For other elements, the ratio masked the 24-h gender difference. This is explained by the lower creatinine excretion of women, reflecting smaller muscle bulk. Others have reported higher trace element to creatinine ratios for women,^{10,13–16,33} which may lead to misinterpretation.^23,33 Ratios also increase with ageing and poor protein nutrition, and are higher in children.^{1,11,14,16,21}

Use of creatinine ratios to adjust for variation in hydration status assumes that the rate of creatinine excretion is constant over 24 h and that excretion of trace elements and creatinine change in parallel. From this study, this appears to be a reasonable assumption. For both women and men, creatinine ratios were linearly related to 24-h excretion, generally with good or moderately good correlation, except for tungsten. However, estimation of the ratio for a 24-h urine pool masks fluctuations during the day. Urine creatinine excretion varies with urine flow rate^18,19 as does excretion of mercury.¹⁹ A circadian variation of urinary excretion has been reported for zinc,³⁸ iodine,³³ mercury,¹⁹ and iron, copper, zinc and nickel.³⁹ Excretion of toxic trace elements may vary during the day in occupationally exposed workers.¹⁹ Hence, correlation of trace element:creatinine ratio of a single random urine sample with 24-h excretion is likely to be weaker than that in this study. Proposals to reduce these problems include reporting results in concentration per litre with no correction for hydration, restricting analyses to samples with creatinine concentrations of 3–30 mmol/L,²⁰ avoidance of the overnight fasting urine and collection of a morning specimen or other appropriately timed sample,³³ adjusting for urine flow rate using a timed collection and correction factor,¹⁸ and correction for hydration status by relating trace element concentrations to an age and sex-adjusted value for 24-h creatinine,^21,33 urine specific gravity¹² or osmolality.⁴⁰ Clearly, when trace element:creatinine ratios are used for random urine samples, they must be interpreted with appropriate reference intervals for gender, age and nutritional status, if from a poorly nourished cohort. When these variables differ, ratios for different study populations must be compared with caution.

Conclusion

The reference intervals presented are representative for a healthy British population with normal exposure to trace elements. They should be applicable to other similar populations. Of necessity, random urine samples will be used for most large population studies and occupational monitoring. Adjustment of concentrations to correct for hydration status is then important. Creatinine ratios are useful if applied cautiously, but an alternative might be preferable. Urine osmolality is a possible candidate worth re-evaluation.

DECLARATIONS

Competing interests: None.

Funding: No external funding.

Ethical approval: The study was approved by the Southampton & South West Hampshire Research Ethics Committee (study number 05/Q1704/59).

Guarantor: VW.

Contributorship: CES supervised the analytical work, did some analyses and helped to process the data; LCJ did most of the analyses and tabulated the results; RO helped with the analyses, reviewed the literature and the clinical records; and VW conceived the study, recruited patients, processed the data and wrote the manuscript.

Acknowledgements: We are grateful to Dr Emmanuel Abu and Dr Robert Walton for recruiting clinic patients for the study. RO submitted his contribution to the study as his dissertation for membership of the Royal College of Pathologists.

References

Komaromy-Hiller

, Ash

, Costa

, Howerton

. Comparison of representative ranges based on U.S. patient population and literature reference intervals for urinary trace elements. Clin Chim Acta 2000;296:71–90

Paschal

, Ting

, Morrow

, Trace metals in urine of United States residents: reference range concentrations. Environ Res Section A 1998;76:53–9

Dongarrà

, Lombardo

, Tamburo

, Varrica

, Cibella

, Cuttitta

. Concentration and reference interval of trace elements in human hair from students living in Palermo, Sicily (Italy). Environ Toxicol Pharmacol 2011;32:27–34

Minoia

, Sabbioni

, Apostoli

, Trace element reference values in tissues from inhabitants of the European Community I. A study of 46 elements in urine, blood and serum of Italian subjects. Sci Total Environ 1990;95:89–105

Hamilton

, Sabbioni

, Van der Venne

. Element reference values in tissues from inhabitants of the European Community. VI. Review of elements in blood, plasma and urine and a critical evaluation of reference values for the United Kingdom population. Sci Total Environ 1994;158:165–90

Alimonti

, Petrucci

, Krachler

, Bocca

, Caroli

. Reference values for chromium, nickel and vanadium in urine of youngsters from the urban area of Rome. J Environ Monit 2000;2:351–4

Beneš

, Spĕváčková

, Šmíd

, Determination of normal concentration levels of Cd, Pb, Hg, Cu, Zn and Se in urine of the population in the Czech Republic. Cent Eur J Public Health 2002;10:3–5

Iyengar

. Reevaluation of the trace element content in reference man. Radiat Phys Chem 1998;51:545–60

White

, Sabbioni

. Trace element reference values in tissues from inhabitants of the European Union. X. A study of 13 elements in blood and urine of a United Kingdom population. Sci Total Environ 1998;216:253–70

10.

Batáriová

, Spĕváčková

, Beneš

, Čejchanová

, Šmíd

, Černá

. Blood and urine levels of Pb, Cd, and Hg in the general population of the Czech Republic and proposed reference values. Int J Hyg Environ Health 2006;209:359–66

11.

Heitland

, Köster

. Biomonitoring of 30 trace elements in urine of children and adults by ICP-MS. Clin Chim Acta 2006;365:310–8

12.

Ohashi

, Fukui

, Takada

, Moriguchi

, Ezaki

, Ikeda

. Reference values for cobalt, copper, manganese, and nickel in urine among women of the general population in Japan. Int Arch Occup Environ Health 2006;80:117–26

13.

Iversen

, Menné

, White

, Kristiansen

, Christensen

, Sabbioni

. Inductively coupled mass spectrometric determination of molybdenum in urine from a Danish population. Analyst 1998;123:81–5

14.

Zeiner

, Ovari

, Zaray

, Steffan

. Reference concentrations of trace elements in urine of the Budapestian population. Biol Trace Elem Res 2004;101:107–15

15.

Zeiner

, Ovari

, Zaray

, Steffan

. Selected urinary metal reference concentrations of the Viennese population – urinary metal reference values (Vienna). J Trace Elem Med Biol 2006;20:240–4

16.

Konno

, Yuri

, Miura

, Kumagai

, Murakami

. Clinical evaluation of the iodide/creatinine ratio of casual urine samples as an index of daily iodide excretion in a population study. Endocr J 1993;40:163–9

17.

Kristiansen

, Christensen

, Iversen

, Sabbioni

. Toxic trace element reference levels in blood and urine: influence of gender and lifestyle factors. Sci Total Environ 1997;204:147–60

18.

Greenberg

, Levine

. Urinary creatinine excretion is not stable: a new method for assessing urinary toxic substance concentrations. J Occup Med 1989;31:832–8

19.

Trachtenberg

, Barregård

, McKinlay

. The influence of urinary flow rate on mercury excretion in children. J Trace Elem Med Biol 2010;24:31–5

20.

Thomson

, Smith

, Butler

, Packer

. An evaluation of urinary measures of iodine and selenium status. J Trace Elem Med Biol 1996;10:214–22

21.

Knudsen

, Christiansen

, Brandt-Christensen

, Nygaard

, Perrild

. Age-and sex-adjusted iodine/creatinine ratio. A new standard in epidemiological surveys? Evaluation of three different estimates of iodine excretion based on casual urine samples and comparison to 24h values. Eur J Clin Nutr 2000;54:361–3

22.

Ilich

, Blanusa

, Orlić

, Orct

, Kostial

. Comparison of calcium, magnesium, sodium, potassium, zinc, and creatinine concentration in 24-h and spot urine samples in women. Clin Chem Lab Med 2009;47:216–21

23.

Dlugaszek

, Kaszczuk

, Mularczyk-Oliwa

. Magnesium, calcium, and trace elements excretion in 24-h urine. Biol Trace Elem Res 2011;142:1–10

24.

Delves

, Sieniawska

, Fell

, Determination of antimony in urine, blood and serum in liver and lung tissues of infants by inductively coupled plasma mass spectrometry. Analyst 1997;122:1323–9

25.

Schramel

, Wendler

, Angerer

. The determination of metals (antimony, bismuth, lead, cadmium, mercury, palladium, platinum, tellurium, thallium, tin and tungsten) in urine samples by inductively coupled plasma-mass spectrometry. Int Arch Occup Environ Health 1997;68:219–23

26.

Fatemian

, Allibone

, Walker

. Use of gold as a routine and long term preservative for mercury in potable water, as determined by ICP-MS. Analyst 1999;124:1233–6

27.

Sieniawska

, Meniskov

, Delves

. Determination of total selenium in serum, whole blood and erythrocytes by ICP-MS. J Anal At Spectrom 1999;14:109–12

28.

Vanhoe

, Damms

, Versieck

. Use of inductively coupled plasma mass spectrometry for the determination of ultra-trace elements in human serum. J Anal At Spectrom 1994;9:23–31

29.

Solberg

: International Federation of Clinical Chemistry (IFCC). Approved recommendation (1987) on the theory of reference values. Part 5. Statistical treatment of collected reference values. Determination of reference limits. J Clin Chem Clin Biochem 1987;25:645–56

30.

Poulsen

, Holst

, Christensen

. Calculation and application of coverage intervals for biological reference values. Pure Appl Chem 1997;69:1601–11

31.

Hofbauer

, Steffan

, Höbarth

, Trace elements and urinary stone formation: new aspects of the pathological mechanism of urinary stone formation. J Urol 1991;145:93–6

32.

Słojewski

. Major and trace elements in lithogenesis. Cent Eur J Urol 2011;64:58–60

33.

Rasmussen

, Ovesen

, Christiansen

. Day-to-day and within-day variation in urinary iodine excretion. Eur J Clin Nutr 1999;53:401–7

34.

Sabbioni

, Minoia

, Pietra

, Fortaner

, Gallorini

, Saltelli

. Trace element reference values in tissues from inhabitants of the European community. II. Examples of strategy adopted and trace element analysis of blood, lymph nodes and cerebrospinal fluid of Italian subjects. Sci Total Environ 1992;120:39–61

35.

Batista

, Rodrigues

, Tormen

, Curtius

, Barbosa

Jr . Reference concentrations for trace elements in urine for the Brazilian population based on q-ICP-MS with a simple dilute-and-shoot procedure. J Braz Chem Soc 2009;20:1406–13

36.

Iyengar

, Woittiez

. Trace elements in human clinical specimens: evaluation of literature data to identify reference values. Clin Chem 1988;34:474–81

37.

Wilhelm

, Ewers

, Schulz

. Revised and new reference values for some trace elements in blood and urine for human biomonitoring in environmental medicine. Int J Hyg Environ Health 2004;207:69–73

38.

Franco

, Fidanza

, Casale

. Biological monitoring and chronobiology: preliminary time-quantified reference values of urinary zinc. Chronobiologia 1987;14:49–52

39.

Kanabrocki

, Sothern

, Ryan

, Circadian characteristics of serum calcium, magnesium and eight trace elements and of their metallo-moieties in urine of healthy middle-aged men. Clin Ter 2008;159:329–46

40.

MacFarlane

. Measurement of trace elements in urine. Proc R Soc Med 1976;69:476–7