Sage Journals: Discover world-class research

Abstract

Background

Serum copper and zinc are measured to assess deficiency and toxicity. Atomic absorption spectrophotometry and mass spectrometry methods are expensive and require highly trained staff. Colorimetric assays are available from Randox which are inexpensive and can be automated. We validated serum copper and zinc colorimetric assays on the Binding Site Optilite analyser including comparison with flame atomic absorption spectrophotometry (FAAS) and inductively coupled plasma-mass spectrometry (ICP-MS).

Methods

Accuracy, imprecision, lower limit of quantitation, and linearity were ascertained. The impact of triglycerides, bilirubin, nickel, and iron on assay performance was also investigated. Comparison of results from colorimetric analysis of patient and external quality assurance samples with those obtained by FAAS and ICP-MS was undertaken.

Results

Intra-, and inter-assay imprecision was <9%. Serum copper and zinc assays were linear between 1.8–35.6 and 2.3–45.7 µmol/L, respectively. Agreement was good between colorimetry and FAAS (intercept = -0.7, slope = 1.04) and ICP-MS (intercept = 0.6, slope = 0.99) for serum copper in patients’ samples. For serum zinc, agreement was poor between colorimetry and FAAS (intercept = 2.2, slope = 0.87) and ICP-MS (intercept = 1.9, slope = 0.98) in patients’ samples. There was a poor concordance in assessment of hypozincaemia between colorimetry and FAAS/ICP-MS.

Conclusion

The Randox colorimetric assay for serum copper on the Optilite is simple to perform, has a short analysis time, and measured concentrations compare well with FAAS and ICP-MS. Due to poor agreement with FAAS and ICP-MS, colorimetry is not suitable for the measurement of serum zinc.

Keywords

Copper zinc colorimetry

Introduction

Copper and zinc are essential trace elements that are measured to assess deficiency and toxicity. Copper acts primarily as a co-factor in enzyme systems and within plasma is largely (> 95%) bound to caeruloplasmin.¹ Copper deficiency can be present in those on long-term parenteral/enteral nutrition or can occur from reduced absorption (e.g. coeliac disease, excessive zinc intake). Copper toxicity may arise from acute/chronic copper ingestion or in those individuals with Wilson disease.² Zinc deficiency arising from inadequate diets is a significant public health concern throughout much of the world and is associated with increased morbidity, delayed wound healing, and altered immune function. Severe zinc deficiency causes major dermal, gastrointestinal, and neurological disorders.¹ Conversely, zinc toxicity is uncommon and manifests as a variety of symptoms, including abdominal pain, nausea, vomiting, lethargy, anaemia, and dizziness.³

Copper and zinc are most frequently measured in plasma or serum. This is despite concentrations of these elements fluctuating in these matrices for reasons besides deficiency or toxicity, such as the acute phase response and diurnal variation. Within children and adults, serum copper concentrations are approximately 3.0–25.0 µmol/L and 12.0–18.0 µmol/L, respectively.¹ In Wilson disease, serum copper is typically <4 µmol/L but can be elevated if severe liver disease is present.¹ Serum zinc is typically 11.0–22.0 µmol/L for all ages.¹ Copper is also measured in 24-hour urine collections or before and after administration of D-penicillamine to assist in the diagnosis of Wilson disease.⁴ Within a clinical setting, urinary zinc is occasionally measured to assess zinc status, although interpretation can be difficult.⁵

Measurement is most commonly performed using ‘gold standard’ Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) or Flame Atomic Absorption Spectrometry (FAAS), as these methods are both sensitive and selective. However, such methods require highly trained staff, and instrumentation is expensive and may not be practical for small or geographically isolated laboratories.⁶ Colorimetric methods for copper and zinc are commercially available (Randox, Parramatta, Australia), which can be undertaken manually or semi-automated on appropriate analysers. The copper assay is based on the release of copper bound from caeruloplasmin by addition of ascorbic acid and then the formation of a stable colour complex between copper and 4-(3,5-dibromo-2-pyridylazo)-N-ethyl-N-(3-sulfopropyl)aniline (3,5-di-Br-PAESA). However, the zinc assay is based on the formation of a stable colour complex of zinc with 2-(5-bromo-2-pyridylazo)-5-(N-propyl-N-sulphopropylamino)-phenol (5-Br-PAPS) in an alkaline solution (pH = 9.8).⁷ The intensity of the colour complex for both assays is proportional to the copper and zinc concentrations and can be measured at wavelengths of 575 and 575/700 nm. The advantages of colorimetric assays are ease of use and that they are relatively inexpensive. However, instrument-specific assay parameters are not always available, and concern has been raised over the accuracy of colorimetric assays, especially when using the aqueous calibration solution supplied with the assay.⁶

The Optilite is a benchtop analyser available from the Binding Site (Edgbaston, UK) which allows the use of third-party reagents and is well suited for small to medium throughput laboratories. Parameters for the use of the colorimetric copper and zinc assays on the Optilite are not available, and published evaluation data for these assays on chemistry analysers in serum and urine are scant.^6,8–10 Therefore, we created Optilite-specific assay parameters and fully validated the performance of the copper and zinc assays for serum samples, including comparison to FAAS and ICP-MS methods. We also retrospectively reviewed imprecision data for serum copper and zinc from the Royal College of Pathologists Australia, Quality Assurance Programs (RCPAQAP) for Serum Trace Elements. We also briefly investigated the use of the colorimetric assays for copper and zinc in urine by ascertaining accuracy, imprecision, and comparison with samples from the RCPAQAP scheme for urine trace elements.

Methods

Materials and reagents

The colorimetric copper and zinc assays were purchased from Randox and consisted of deproteinisation solution (370 mmol/L aqueous trichloroacetic acid), chromogens, buffers, and an aqueous calibration solution. Water (resistivity: >18.2 MΩ) was purified by reverse osmosis and ion-exchange using a Merck Millipore Direct-Q 3 UV water purification system (Melbourne, Australia). Copper nitrate [Cu(NO₃)₂, 1 g/L], zinc nitrate [Zn(NO₃)₂, 1 g/L], nickel nitrate [Ni(NO₃)₂, 1 g/L], and iron nitrate [Fe(NO₃)₂, 1 g/L] had a purity of 99.999% (Merck). All glass and plastic vessels were soaked in 10% v/v (volume/volume) aqueous nitric acid (ACS reagent grade, Merck) for at least two hours and rinsed three times with water before use. Phosphate buffered saline (PBS, pH 7.4) was from Merck.

Instrumentation and assay

The colorimetric assays were set-up on a Binding Site, Optilite benchtop analyser. Specific instrument parameters were not available for the Optilite. However, parameters for the Thermo Scientific Konelab 20 Clinical Chemistry Analyser, which is made by the same manufacturers as the Optilite were available from Randox. The incubation times, sample, and reagent volumes provided for the Konelab 20 were increased and decreased in a stepwise manner and the absorbance noted. Internal quality controls (IQC) were analysed with each change, and the parameters that provided the smallest sample volume and analysis time without a reduction in absorbance were chosen (Table 1).

Table 1.

Colorimetric assay parameters used on the Binding Site, Optilite.

Matrix	Analyte	Volume (µL)			Incubation time (s)	Wavelength (nm)
Matrix	Analyte	Sample	Reagent	Chromogen
Serum	Copper	20	125	30	300	575
Serum	Zinc	10	100	-	300	575 and 700
Urine	Copper	30	125	30	300	575
Urine	Zinc	20	100	-	300	575 and 700

For the serum assay, the deproteinisation solution (100 µL) was added to the sample (100 µL), vortex-mixed for 10 sec, and centrifuged (10,000 g, 10 min). The supernatant was then analysed against a calibration consisting of a reagent blank (water) and a reconstituted lyophilized human serum calibration standard (Randox, Clinical Chemistry Calibration Serum Level 3, copper/zinc concentrations: 25.3/37.2 µmol/L, respectively). Internal Quality control was undertaken at 2 concentrations using reconstituted Seronorm (Billingstad, Norway) human (copper/zinc concentration: 12.7/13.0 µmol/L, respectively) and human high (copper/zinc concentration: 28.6/18.1 µmol/L, respectively) lyophilized serum controls which were analysed in duplicate with every batch of samples.

For the urine assay, the sample was centrifuged (10,000 g, 10 min) and directly analysed against a calibration consisting of a reagent blank and a reconstituted lyophilized human urine calibration standard, ClinCal calibrator, copper/zinc concentration 2.3/10.2 µmol/L, respectively (Recipe, Munich, Germany). Bio-Rad Laboratories (South Granville, Australia) Lyphocheck, Urine Metals Control levels 1 (copper/zinc concentration: 0.9/2.9 µmol/L, respectively) and 2 (copper/zinc concentration: 1.8/9.1 µmol/L, respectively) were used as IQC material.

The FAAS assay has been previously described.⁶ Briefly, the sample (300 µL) was diluted with deionized water (600 µL) and measured using a Varian Spectra 880 (Agilent Technologies, Mulgrave, Australia) at a wavelength of 324.8 and 213.9 nm, for copper and zinc, respectively. Calibration was by standard addition using calibration solutions (n = 4) prepared from zinc [Zn(NO₃)₂, 1 g/L] and copper nitrate [Cu(NO₃)₂, 1 g/L] in 5% (v/v) aqueous nitric acid, added to excess pooled human serum. The ICP-MS assay involved the sample (80 µL) being diluted with 1% (v/v) aqueous nitric acid containing gallium (10 mg/L) as internal standard (3920 µL) and measured using an Agilent 7900 (Agilent Technologies, Mulgrave, Australia) in standard mode (i.e. without a collision cell). Calibration was by aqueous solutions prepared from zinc [Zn(NO₃)₂, 1 g/L] and copper nitrate [Cu(NO₃)₂, 1 g/L] in 10% (v/v) aqueous nitric acid.

Total bilirubin and triglycerides were both measured spectrophotometrically using a Cobas 6000 automated analyser (Roche, North Ryde, Australia).

Accuracy and imprecision

Intra- and inter-assay imprecision and accuracy for the serum assay were ascertained by the analysis of a patient serum sample, 1 external quality assurance (EQA) sample from the RCPAQAP scheme for Serum Trace Elements, and Seronorm human and human high serum controls in quintuplicate on the same day and on different days (n = 5). For the urine assay, 1 EQA sample from the RCPAQAP Urine Trace Elements scheme, and Bio-Rad Lyphocheck Urine Metals Control levels 1 and 2 were analysed in quintuplicate on the same day and on different days (n = 5). Accuracy and imprecision <15% were considered acceptable.¹¹

Carry-over, linearity, and lower limit of quantitation

Carry-over was determined by the analysis of serum samples containing copper/zinc at low (L) and high (H) concentrations (5.8/5.8 and 35.6/35.7 µmol/L, L and H, respectively) followed by a blank [water, (B)] in the following sequence: L, L, L, B, B, B, and H, H, H, B, B, B. Carry-over <1% was considered insignificant.

Linearity was assessed by non-serial dilutions (1 + 1, 1 + 4, 1 + 9, 1 + 19) of a patient serum sample (initial copper/zinc concentrations: 35.6/45.7 µmol/L, respectively) with PBS. Differences were calculated between the nominal and measured concentrations, with acceptance criteria being within 10% (mean of triplicates).¹¹

The lower limit of quantitation (LLoQ) was assessed by non-serial dilutions (1 + 1, 1 + 4, 1 + 9, 1 + 19) of a patient serum sample (initial copper/zinc concentration approximately 13.3/13.9 µmol/L, respectively) with PBS using an acceptance criteria of <20% inaccuracy and imprecision (mean of triplicates).¹¹

Interference studies

The effect of total bilirubin and triglycerides was ascertained by the non-serial dilution (i.e. 1 + 1, 1 + 4, 1 + 9, 1 + 19) of 3 separate serum samples (copper and zinc; 12.5/16.9/21.6, and 12.1/15.6/25.5 µmol/L, respectively) that contained low concentrations of total bilirubin (median: 3, range: 1–6 µmol/L) and triglycerides (median: 0.6, range: 0.4–1.0 µmol/L) with a separate sample (copper and zinc: 13.5 and 13.2 µmol/L, respectively) containing high concentrations of total bilirubin (184 µmol/L) and triglycerides (8.89 mmol/L). All samples had no discernible haemolysis. Differences were calculated between the nominal and measured concentrations of copper and zinc, with acceptance criteria being within 10% (mean of triplicates).

Interference from nickel and iron in both copper and zinc assays, and copper and zinc in zinc and copper assays, respectively, was ascertained by addition (fortification) of aqueous copper, zinc, iron, and nickel nitrate (final concentration: 100 µmol/L) to aqueous solutions separately containing copper and zinc at 5, 20, and 30 µmol/L. Differences in copper and zinc concentrations between fortified and non-fortified samples were calculated, with acceptance criteria being within 10% (mean of triplicates).

Method comparison

For method comparison, excess serum patients’ samples submitted to the laboratory for the assessment of copper and zinc were used. Whole blood was collected into a polypropylene Vacuette serum separator blood collection tube (Greiner Bio-One, Kremsmünster, Austria) and transported to the laboratory on the day of collection. Once received in the laboratory, samples were centrifuged (12,000 g, 10 min), and portions (approximately 0.5 mL) of serum were stored in 1.5 mL natural microcentrifuge tubes (Eppendorf, Macquarie Park, Australia) for up to three days at 2–8°C before analysis by colorimetric, FAAS, and ICP-MS assays (microcentrifuge tubes were soaked in 10% (v/v) aqueous nitric acid for at least two hours and rinsed three times with water before use). Quality assurance samples from the RCPAQAP schemes for serum and urine trace elements were also analysed and stored at -18 to -20°C for up to 3 days from receipt in the laboratory until analysis. Serum and urine samples are prepared from bovine and human matrices, respectively, and distributed as frozen liquid samples. Retrospective precision data (relative standard deviation %) for serum copper and zinc 2017–2022 by analytical technique was obtained from RCPAQAP.¹²

Statistical analysis

Statistical analysis was performed using Analyse-it (Analyse-it Software Ltd, Leeds, UK) for Microsoft Excel. Normality was assessed by the Shapiro–Wilk test and data were found to be parametric. Data were expressed as means and standard deviations (SD). Difference between assays was evaluated using the t-test for matched pairs, correlation was assessed using Pearson’s correlation (r), and method comparison undertaken using Bland–Altman and Passing–Bablok regression analyses.

Results

Accuracy, imprecision, linearity, lower limit of quantitation, and carry-over

Intra- and inter-assay imprecision and accuracy data are given in Table 2. Although accuracy and imprecision for serum copper, serum zinc, and urine copper were largely acceptable (intra-assay accuracy for the high urine copper QC was unacceptable), accuracy of urine zinc was unacceptable, particularly at the low concentration (2.9 µmol/L). Both serum copper and zinc assays were linear over the range 1.8–35.6 (r² = 0.99, y = 0.9146x - 0.0607), and 2.3–45.7 (r² = 0.99, y = 1.0923x - 1.2542) µmol/L, respectively. The LLoQ was 1.3 (accuracy 115%, imprecision 8%) and 1.4 µmol/L (accuracy 119%, imprecision 11%) for serum copper and zinc assays, respectively, and no significant carry-over was detected (<1%).

Table 2.

Summary accuracy and imprecision data for serum copper (A), serum zinc (B), urine copper (C), and urine zinc (D) measured using Randox colorimetric assays on the Binding Site, Optilite analyser.

(A)
Intra-assay (n = 5)	IQC material		EQA material^a	Patient sample
Nominal concentration (range), µmol/L	12.7 (10.6-14.7)	28.6 (23.9-33.2)	13.0	—
Mean measured concentration (µmol/L)	13.2	27.5	13.3	21.1
Relative standard deviation (%)	3.5	5.5	7.1	2.9
Accuracy (% nominal)	103	96	102	—
Inter-assay (n = 25)
Mean measured concentration (µmol/L)	13.4	28.9	13.3	20.9
Relative standard deviation (%)	2.7	5.5	1.6	5.3
Accuracy (% nominal)	106	101	102	—

(B)
Intra-assay (n = 5)	IQC material		EQA material^a	Patient sample
Nominal concentration (range), µmol/L	13.0 (9.6-16.4)	18.1 (13.3-22.9)	30.2	—
Mean measured concentration (µmol/L)	11.9	19.3	27.9	9.9
Relative standard deviation (%)	7.6	8.7	6.1	7.4
Accuracy (% nominal)	92	107	92	—
Inter-assay (n = 25)
Mean measured concentration (µmol/L)	12.1	17.5	28.7	9.7
Relative standard deviation (%)	5.5	6.6	2.4	6.1
Accuracy (% nominal)	93	94	95	—

(C)
Intra-assay (n = 5)	IQC material		EQA material^b
Nominal concentration (range), µmol/L	0.9 (0.7-1.0)	1.8 (1.4-2.1)	1.7
Mean measured concentration (µmol/L)	0.8	2.2	1.7
Relative standard deviation (%)	<1.0	5.7	6
Accuracy (% nominal)	89	122	100
Inter-assay (n = 25)
Mean measured concentration (µmol/L)	0.8	2.0	1.6
Relative standard deviation (%)	<1.0	5.2	3.4
Accuracy (% nominal)	89	111	94

(D)
Intra-assay (n = 5)	IQC material		EQA material^b
Nominal concentration (range), µmol/L	2.9 (2.4-3.6)	9.1 (7.3-11.0)	11.6
Mean measured concentration (µmol/L)	4.1	11.0	12.9
Relative standard deviation (%)	1.2	4.6	6
Accuracy (% nominal)	141	121	111
Inter-assay (n = 25)
Mean measured concentration (µmol/L)	4.0	10.5	13.1
Relative standard deviation (%)	3.4	8.7	9.6
Accuracy (% nominal)	137	115	112

^aNominal concentration for EQA samples is the consensus median value (n = 115) for all methods (colorimetry, atomic absorption spectroscopy, ICP-MS, and inductively coupled plasma-optical emission spectroscopy).

^bNominal concentration for EQA samples is the consensus median value (n = 70) for all methods (atomic absorption spectroscopy, ICP-MS, and inductively coupled plasma-optical emission spectroscopy).

Interference studies

Total bilirubin concentrations up to 184 µmol/L and triglyceride concentrations up to 8.89 mmol/L did not interfere with the measurement of either copper (accuracy: 93–105%) or zinc (accuracy: 95–109%), neither did the addition of nickel, iron, copper, or zinc (copper/zinc accuracy: 96–104%).

Method comparison

Serum

Forty serum samples were analysed from 39 patients [13 male, mean age (range) 48 (28–74) years; 26 female, mean age (range) 47 (27–75) years] and 31 EQA samples by colorimetry and FAAS (Supplemental Table 1, Supplemental Table 2). Twenty separate serum patients' samples were also analysed by ICP-MS (Supplemental Table 3). Results of comparison studies are provided in Table 3 and Figures 1, 2, 3, and 4. There was a good correlation and agreement for serum copper in both patients’ samples and EQA samples between colorimetric, FAAS, and ICP-MS assays. In patients’ samples, there was no statistically significant difference in mean serum copper concentrations measured by colorimetric (13.7 µmol/L) vs. FAAS (14.0 µmol/L) assays or colorimetric (18.8 µmol/L) vs. ICP-MS (18.4 µmol/L) assays. However, the mean serum copper concentration measured in EQA samples was 0.3 µmol/L higher (P = <0.05) when measured by colorimetry (20.4 µmol/L) than FAAS (20.1 µmol/L). In patients’ samples, there was no statistically significant difference in mean serum zinc concentrations measured by colorimetric (13.6 µmol/L) vs. FAAS (13.8 µmol/L) assays. However, the mean serum zinc concentration measured in samples by ICP-MS (13.6 µmol/L) was 1.4 µmol/L higher (P = <0.0001) when measured by colorimetry (15.0 µmol/L).

Table 3.

Summary comparison data for serum copper (A), serum zinc (B), urine copper (C), and urine zinc (D) measured by Randox colorimetric assays compared to Flame Atomic Absorption Spectrophotometry (FAAS) and Inductively Coupled Plasma-Mass Spectrometry (ICP-MS). Results of Pearson’s correlation (r) and Passing–Bablok regression analysis. Statistical significance (P) calculated using Student’s t-test.

(A)
	Colorimetry vs. FAAS		Colorimetry vs. ICP-MS
	Patient samples (n = 40)	EQA samples (n = 31)	Patient samples (n = 20)
r (95% CI)	0.99 (0.98–0.99)	0.99 (0.99–0.99)	0.99 (0.97–0.99)
Mean bias (95% CI), µmol/L	-0.3 (-0.7–0.1), P = .11	0.3 (0.0–0.54), P = <0.05	0.4 (-0.2–0.9), P = .15
Mean bias (95% CI), %	-2.1 (-3.9– -0.4)	0.3 (-1.3–1.9)	-2.0 (-0.85–4.9)
Intercept (95% CI)	-0.7 (-1.2– -0.14)	-0.7 (-1.4– -0.34)	0.6 (-1.8–1.5)
Slope (95% CI)	1.04 (0.98–1.07)	1.01 (0.99–1.04)	0.99 (0.91–1.12)

(B)
	Colorimetry vs. FAAS		Colorimetry vs. ICP-MS
	Patient samples (n = 40)	EQA samples (n = 31)	Patient samples (n = 20)
r (95% CI)	0.90 (0.81–0.94)	0.99 (0.98–0.99)	0.96 (0.91–0.99)
Mean bias (95% CI), µmol/L	-0.2 (-0.9–0.5), P = .59	-2.2 (-2.6– -1.9), P = <0.0001	1.4 (0.7–2.1), P = <0.0001
Mean bias (95% CI), %	-0.4% (-6.5–5.7)	-6.4 (-8.6– -4.2)	11.2 (5.1–17.1)
Intercept (95% CI)	2.2 (-0.6–4.4)	-0.8 (-2.3–1.8)	1.9 (0.6–5.3)
Slope (95% CI)	0.87 (0.71–1.00)	0.93 (0.83–1.00)	0.98 (0.67–1.06)

(C)
	Colorimetry vs. FAAS
	EQA samples (n = 21)
r (95% CI)	0.99 (0.98–0.99)
Mean bias (95% CI), µmol/L	0.01 (-0.04– 0.07), P = .63
Mean bias (95% CI), %	-3.3 (-9.1– 2.4)
Intercept (95% CI)	-0.11 (-0.2–0.0)
Slope (95% CI)	1.05 (1.01–1.07)

(D)
	Colorimetry vs. FAAS
	EQA samples (n = 21)
r (95% CI)	0.99 (0.99–0.99)
Mean bias (95% CI), µmol/L	1.21 (0.63–1.79), P = <0.05
Mean bias (95% CI), %	17.8 (8.2–27.4)
Intercept (95% CI)	3.7 (3.5–4.1)
Slope (95% CI)	0.76 (0.74–0.79)

Figure 1.

Scatterplot (a) and Bland–Altman plot (b) for serum copper measured by colorimetry compared to Flame Atomic Absorption Spectrophotometry (FAAS) in 40 patient’s samples. Scatterplot shows line of identity (solid line) and linear regression analysis (dashed line); Bland–Altman plot shows the line of identity (solid line) and the mean (95% limits of agreement) bias (dotted line and dashed lines, respectively). Mean bias (95% limits of agreement); -0.30 (2 to -2.6) µmol/L.

Figure 2.

Scatterplot (a) and Bland–Altman plot (b) for serum copper measured by colorimetry compared to Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) in 20 patient’s samples. Scatterplot shows line of identity (solid line) and linear regression analysis (dashed line); Bland–Altman plot shows the line of identity (solid line) and the mean (95% limits of agreement) bias (dotted line and dashed lines, respectively). Mean bias (95% limits of agreement); 0.40 (2.6 to -1.9) µmol/L.

Figure 3.

Scatterplot (a) and Bland–Altman plot (b) for serum zinc measured by colorimetry compared to Flame Atomic Absorption Spectrophotometry (FAAS) in 40 patient’s samples. Scatterplot shows line of identity (solid line) and linear regression analysis (dashed line); Bland–Altman plot shows the line of identity (solid line) and the mean (95% limits of agreement) bias (dotted line and dashed lines, respectively). Mean bias (95% limits of agreement); -0.20 (4.2 to -4.6) µmol/L.

Figure 4.

Scatterplot (a) and Bland–Altman plot (b) for serum zinc measured by colorimetry compared to Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) in 20 patient’s samples. Scatterplot shows line of identity (solid line) and linear regression analysis (dashed line); Bland–Altman plot shows the line of identity (solid line) and the mean (95% limits of agreement) bias (dotted line and dashed lines, respectively). Mean bias (95% limits of agreement); 1.40 (4.4 to -1.7) µmol/L.

Classification of copper and zinc status based on methodology is provided in Table 4. However, classification of serum copper status compared well between FAAS, ICP-MS, and colorimetry; for zinc, there was considerable disagreement between methodologies when the serum zinc concentration was within or below the target range.

Table 4.

Comparability of colorimetry with Flame Atomic Absorption Spectrophotometry (FAAS) and Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) for classification of serum copper (A, B) and zinc (C, D) status.

(A)
		Serum copper (µmol/L) FAAS
		<12.0	12.0-18.0	>18.0
Serum copper (µmol/L) colorimetry	<12.0	6	0	0
	12.0-18.0	0	16	0
	>18.0	0	0	18

(B)
		Serum copper (µmol/L) ICP-MS
		<12.0	12.0-18.0	>18.0
Serum copper (µmol/L) colorimetry	<12.0	4	1	0
	12.0-18.0	1	4	0
	>18.0	0	0	10

(C)
		Serum zinc (µmol/L) FAAS
		<11.0	11.0-22.0	>22.0
Serum zinc (µmol/L) colorimetry	<11.0	7	5	0
	11.0-22.0	2	23	1
	>22.0	0	0	2

(D)
		Serum zinc (µmol/L) ICP-MS
		<11.0	11.0-22.0	>22.0
Serum zinc (µmol/L) colorimetry	<11.0	4	0	0
	11.0-22.0	3	11	0
	>22.0	0	0	2

Retrospective RCPAQAP imprecision data for serum copper and zinc by analytical technique is provided in Table 5. In 2021–2022, ICP-MS, FAAS, and colorimetry comprised 43, 35, and 10% of methods used for copper. For zinc, ICP-MS, FAAS, and colorimetry comprised 42, 35, and 15% of methods, respectively. For copper and zinc, ICP-MS had the lowest imprecision, and imprecision for copper measured by colorimetry and FAAS was relatively similar. However, for zinc, colorimetry had a lower imprecision over the 5-year period than FAAS and was even comparable to ICP-MS for some years (i.e. 2017–2018, 2021–2022).

Table 5.

Summary end of year imprecision data (relative standard deviation% [RSD%]) for serum copper (A), and serum zinc (B) from RCPAQAP scheme for serum trace elements, 2017–2022 (Value in parenthesis: Number of participants).

(A)
Analytical technique	RSD%
Analytical technique	2017–2018	2018–2019	2019–2020	2020–2021	2021–2022
Colorimetry	9.6 (9)	5.6 (8)	7.9 (10)	9.5 (11)	11.1 (12)
Electrothermal atomic absorption spectrometry	5.8 (15)	5.6 (11)	4.0 (9)	6.4 (11)	7.8 (9)
Flame atomic absorption spectrophotometry	6.6 (55)	7.9 (53)	8.1 (48)	8.2 (45)	8.5 (43)
Inductively coupled plasma-mass spectrometry	4.9 (43)	5.4 (48)	5.6 (54)	5.2 (53)	5.4 (53)
Inductively coupled plasma-optical emission spectrometry	4.1 (5)	9.6 (4)	5.6 (5)	7.7 (5)	8.4 (6)

(B)
Analytical technique	RSD%
Analytical technique	2017–2018	2018–2019	2019–2020	2020–2021	2021–2022
Colorimetry	4.6 (21)	5.4 (19)	8.3 (21)	7.4 (22)	5.5 (20)
Electrothermal atomic absorption spectrometry	9.2 (7)	6.6 (5)	4.4 (5)	9.1 (5)	5.5 (5)
Flame atomic absorption spectrophotometry	7.4 (56)	7.3 (57)	7.8 (51)	7.7 (48)	8.2 (45)
Inductively coupled plasma-mass spectrometry	4.5 (44)	5.7 (46)	6.8 (54)	5.9 (54)	5.4 (54)
Inductively coupled plasma-optical emission spectrometry	5.4 (5)	8.5 (4)	9.1 (5)	7.0 (5)	6.6 (5)

Urine

Twenty-one urine EQA samples were analysed. Results of comparison studies are provided in Table 3 and Figures 5 and 6. For urine copper in EQA samples, there was a good correlation and agreement between colorimetric and RCPAQAP consensus median values. Despite a good correlation for urine zinc measured by colorimetry and the RCPAQAP consensus median values, Passing–Bablok and Bland–Altman analysis identified a poor agreement and bias (1.21 µmol/L, P = <0.05).

Figure 5.

Scatterplot (a) and Bland–Altman plot (b) for urine copper measured by colorimetry compared to the consensus median values in 21 urine external quality assurance samples. Scatterplot shows line of identity (solid line) and linear regression analysis (dashed line); Bland–Altman plot shows the line of identity (solid line) and the mean (95% limits of agreement) bias (dotted line and dashed lines, respectively). Mean bias (95% limits of agreement); 0.01 (0.25 to -0.23) µmol/L.

Figure 6.

Scatterplot (a) and Bland–Altman plot (b) for urine zinc measured by colorimetry compared to the consensus median values in 21 urine external quality assurance samples. Scatterplot shows line of identity (solid line) and linear regression analysis (dashed line); Bland–Altman plot shows the line of identity (solid line) and the mean (95% limits of agreement) bias (dotted line and dashed lines, respectively). Mean bias (95% limits of agreement); 1.21 (3.71 to -1.29) µmol/L.

Discussion

We found the developed instrument parameters to provide acceptable accuracy and sensitivity for the routine measurement of serum copper in clinical samples. A good correlation and agreement were identified for copper in serum patient and EQA samples for colorimetry vs. FAAS. To date, a comparison of serum copper measured by colorimetry and ICP-MS has not been published, and in this study, we found a good agreement between these methods. Furthermore, there was good comparability between methods in assessing copper status. Although intra- and inter-assay imprecision for serum copper was deemed acceptable, we used acceptance criteria from Bioanalytical Method Validation Guidelines¹¹ which is likely to be too broad for clinical purposes, instead, analytical performance specifications using biological variation data may be more appropriate.¹³ The values obtained here are greater than those stated by other investigators (3.5 and 5.4%, respectively),⁶ which may be due to different analysers used. Furthermore, both intra- and inter-assay imprecision only meets the desirable (3.8%) and minimum (5.6%) analytical performance specifications suggested by the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM)¹³ based on biological variation data for IQC and patient samples. However, the intra-assay imprecision for the EQA material is outside of the minimum requirements, and intra-assay imprecision is markedly greater for EQA samples than IQC or patient samples. Furthermore, for EQA material, intra-assay imprecision is far worse than inter-assay imprecision, which is an unusual finding. Matrix effects maybe the cause of the results, and we did not investigate the imprecision of other QAP samples. It is advisable for laboratories to make their own assessments as to whether the imprecision data presented here meets the need of the service they provide. There was also some variability at concentrations above the target range, but this is likely to have minimal clinical significance.

For serum zinc, accuracy and sensitivity were acceptable, and although imprecision was deemed acceptable and broadly comparable to those published by other investigators,⁶ the values obtained were outside of the desirable (4.3%) and minimum (6.4%) analytical performance specifications suggested by the EFLM and unlikely to be acceptable for clinical use. As with copper in the EQA material, intra-assay imprecision is far worse than inter-assay imprecision and matrix effects may be a cause. Although there was a positive correlation between colorimetry and FAAS for serum zinc, there was considerable variability in the measured concentration and random error, which would be unacceptable for clinical use – as indicated by the lack of comparability between methods in assessing zinc status (Table 4). Interestingly, there was a better agreement between colorimetry and ICP-MS but still considerable variability which would be unacceptable. The poor agreement between colorimetry and FAAS reported here has also been identified by other investigators,⁶ who in-part attributed it to the use of an aqueous calibration solution provided with the kit. Indeed, the manufacturer recommends against using the supplied aqueous calibrator but instead a matrix matched human serum calibration solution should be used. However, we did not investigate the aqueous calibration solution provided, and still found that a matrix-matched, human serum calibrator did not improve agreement between methods – a comparable finding to other investigators.⁸

We achieved a similar linear range for serum copper and zinc to other investigators who were using large Roche Cobas 8000 analysers (copper: 4.08–33.8 µmol/L, zinc: 1.15–34.7 µmol/L).^9,10 With regards to the upper limit of quantitation, Randox states values of 158 and 459 µmol/L, for copper and zinc, respectively, on the Thermo Scientific Konelab 20. Given such high concentrations are not present in human serum, we did not investigate this claim on the Optilite.

Imprecision data for serum copper and zinc from the RCPAQAP scheme for serum trace elements was available for a 5-year period (2017–2022; Table 5). Colorimetry performance was broadly comparable to FAAS for serum copper, except for two years (2017–2018, and 2021–2022) where colorimetric performance was markedly worse than FAAS; similar data was found by other investigators.¹⁴ Again performance for serum zinc by colorimetry was broadly comparable to that by FAAS. It is important to note that the relatively few participants using colorimetric methods compared to FAAS and ICP-MS may influence the apparent imprecision, and also the matrix used for preparation of external quality samples is bovine serum, and therefore not directly comparable to human serum. Nonetheless, it allows some comparisons to be made between methodologies. Unfortunately, data on instrumentation, reagents, or calibrators (i.e. commercial or prepared in-house) was not available for any analytical technique. Therefore, the impact of different instrumentation or calibrators cannot be ascertained.

For urine copper, there was a good agreement and correlation between colorimetry and the median consensus values from EQA samples, and overall performance was promising. However, we did not investigate the assay any further as requests for urine copper are minimal in our institution and there was a lack of patient urine samples for method comparison. With regards to urine zinc, accuracy and imprecision were unacceptable, and agreement with the median consensus values from EQA samples was poor. The cause of this poor performance in part, could be due to contamination, interference from endogenous compounds, or inappropriate instrument parameters or calibration material. However, this was not investigated further as measurement of urine zinc is infrequently requested and is of limited clinical value. Subsequently, the use of colorimetry for urine zinc is deemed unsuitable.

A concern with the use of colorimetric assays for serum copper and zinc is potential interference from lipids, bilirubin, proteins in the sample, or from haemoglobin.⁵ Deproteinisation of the sample with aqueous trichloroacetic acid and analysis of the resulting supernatant helps limit the potential impact of some of these concerns, and we did not identify any deleterious effects of high bilirubin or triglyceride concentrations on serum copper or zinc measurements. However, it would be advisable for laboratories to be cautious of measuring copper or zinc in samples with excessive lipid or bilirubin concentrations, and measurement of these parameters beforehand to exclude inappropriate samples should be undertaken. We did not investigate potential interference from haemoglobin, as zinc, in-particular, is located within erythrocytes and haemolysis would falsely elevate serum zinc concentrations. Therefore, the presence of haemolysis would exclude the measurement of serum copper or zinc. Furthermore, we did not observe interference from divalent cations (iron, nickel) investigated on either copper or zinc assays at concentrations greater than what they are present in human serum. This is of especial concern for the zinc assay, as the reagents contain trisodium citrate, dimethylglyoxime, and salicylaldoxine to mask reactions with iron, copper, and nickel.⁵ With regards to interfering substances for the colorimetric urine copper assay, the manufacturer reports no interference from uric acid, glutathione, or D-penicillamine at concentrations of up to 250, 200, and 100 mg/L, respectively – the later compound being of importance in urine samples from patients suspected of having Wilson disease and having undergone a D-penicillamine challenge.

As to limitations of the study, trace element-free whole blood collection tubes were not used for the collection of patients’ samples, therefore, copper and zinc contamination may have been present. However, samples analysed by colorimetric, FAAS, and ICP-MS methods would have all been subject to the same degree of contamination. Furthermore, plasticware (e.g. microcentrifuge) tubes could be a source of contamination for both copper and zinc, but all were thoroughly washed with aqueous nitric acid to reduce the impact of any potential contamination. We also did not ascertain inter- or intra-assay accuracy or imprecision data for serum copper at concentrations below the target range, and for serum zinc, the concentration investigated was marginally below the target range. However, patients’ samples used for comparison studies did cover a wider range of clinically relevant concentrations.

For both serum assays, analysis time on the Optilite was short, with a result available within 5 minutes of loading the supernatant onto the analyser. Furthermore, assay calibration was stable for 61 days as indicated by IQC results being within RCPAQAP analytical performance specifications, allowing savings in terms of reduced reagent consumption and analysis time.

Conclusions

The Randox colorimetric assay for serum copper compares well with FAAS and ICP-MS assays. Due to poor agreement with FAAS and ICP-MS, the Randox colorimetric assay for serum zinc is not suitable for use. The Randox colorimetric assay for urine copper requires further investigation, especially comparison studies using patients’ urine samples, while the urine zinc assay is unsuitable for routine clinical use.

Supplemental material

Supplemental Material - Validation of the Randox colorimetric assays for serum copper and zinc

Supplemental Material for Validation of the Randox colorimetric assays for serum copper and zinc by Simon A Handley, Troy Wanandy, and Louise Prentice in Annals of Clinical Biochemistry.

Footnotes

Acknowledgements

We thank RCPAQAP for permission to quote EQA data, Dr DS Handley for constructive criticism of the draft manuscript and Ms J Houston for analysis of samples by FAAS.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

Ethics approval was not required as work undertaken does not include studies on humans.

Guarantor

LP.

Contributorship

SH researched literature and conceived the study. SH and TW were involved in protocol development and data analysis. SH, TW, and LP drafted the manuscript. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

ORCID iDs

Simon A Handley

Troy Wanandy

Supplemental material

Supplemental material for this article is available online.

References

Roberts

Taylor

Sodi

. Vitamins and Trace Elements. In: Rifai

Horvath

Wittwer

, eds. Tietz Textbook of Clinical Chemistry and Molecular Diagnostics. 6th ed. Missouri: Elseiver; 2018:639-718.

Nelson

. Copper. In: Nelson

Lewin

Howland

, et al., eds. Goldfrank’s Toxicologic Emergencies. 9th ed. New York: McGraw Hill; 2011:1256-1265.

Majelsi

. Zinc. In: Nelson

Lewin

Howland

, et al., eds. Goldfrank’s Toxicologic Emergencies. 9th ed. New York: McGraw Hill; 2011:1339-1344.

Müller

Koppikar

Taylor

, et al. Re-evaluation of the penicillamine challenge test in the diagnosis of Wilson's disease in children. J Hepatol. 2007;47:270-276.

Taylor

. Measurement of zinc in clinical samples. Ann Clin Biochem. 1997;34:142-150.

Beckett

Hartley

Ball

. Evaluation of the Randox colorimetric serum copper and zinc assays against atomic absorption spectroscopy. Ann Clin Biochem. 2009;46:322-326.

Homsher

Zak

. Spectrophotometric investigation of sensitive complexing agents for the determination of zinc in serum. Clin Chem. 1985;31:1310-1313.

Monge

Enrique

Carmen

, et al. Comparison study between colorimetric method and flame atomic absorption spectrophotometry in serum zinc status. Nutr. clín. diet. hosp. 2018;38:133-138.

Kraus

Mischereit

Eller

, et al. Characterization and Validation of the LT-SYS Copper Assay on a Roche Cobas 8000 c502 Analyzer. Lab Med. 2017;48:10-17.

10.

Kraus

Ludwig-Kraus

. Measuring zinc on the Roche cobas c502 analyzer-Validation, comparison, and pre-analytic aspects. J Clin Lab Anal. 2018;32:e22169.

11.

FDA/CDER . Guidance for industry. Bioanalytical method validation; 2022. Available from: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/m10-bioanalytical-method-validation-and-study-sample-analysis. Accessed 14 March 2023.

12.

Royal college of Pathologists, Quality Assurance Programs . https://rcpaqap.com.au/. https://rcpaqap.com.au/. Accessed 20th July 2023.

13.

Aarsand

Fernandez-Calle

Webster

, et al. The EFLM Biological Variation Database. Available from: https://biologicalvariation.eu/(accessed 11th September 2023).

14.

Arnaud

Weber

Weykamp

, et al. Quality specifications for the determination of copper, zinc, and selenium in human serum or plasma: evaluation of an approach based on biological and analytical variation. Clin Chem. 2008;11:1892-1899.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.15 MB