Determination of a Beckman Coulter AU turbidometric method-specific caeruloplasmin reference interval

Abstract

Background

Method-dependent variation of caeruloplasmin measurement is significant and necessitates the requirement for assay-specific reference intervals. Local determination becomes necessary in the absence of suitable published or manufacturer-quoted reference intervals.

Methods

Applicability of the Beckman Coulter AU-quoted reference interval was determined by assay of 20 surplus serum samples from patients attending the medical renal stones clinic at University Hospital Southampton. Subsequently, 60 additional samples were collected for local reference interval determination. Samples were analysed for caeruloplasmin using the Beckman Coulter AU turbidometric kit on an AU680 analyser. Outliers were removed, and non-parametric rank analysis of the results was performed.

Results

The Beckman Coulter-quoted reference interval of 200–600 mg/L was unsuitable with 40% of the verification samples falling below 200 mg/L. A caeruloplasmin reference interval of 150–320 mg/L was established.

Conclusion

Users should be aware the quoted Beckman Coulter AU turbidometric reference interval may not be appropriate. We have established a method-specific adult reference interval for routine use.

Introduction

Caeruloplasmin is an abundant α₂-glycoprotein that, in health, contains more than 95% of serum copper. Laboratory measurement of caeruloplasmin is performed primarily in the investigation of Wilson’s disease, an inherited disorder of copper metabolism characterized by reduced excretion of copper into bile and pathological accumulation of copper in the liver, brain, kidneys and eyes. In addition, there is impaired incorporation of copper into caeruloplasmin reducing its half-life. The European Association for the Study of the Liver guidelines recommend serum caeruloplasmin as a routine screening test for Wilson’s disease. ¹

However, despite the availability of the International Federation of Clinical Chemistry (IFCC) standard CRM 470 for serum proteins, comparability of caeruloplasmin results produced by different assays is a longstanding issue. Agreement studies and external quality assurance schemes have demonstrated large quantitative variation between caeruloplasmin methods, including those traceable to the same reference standard.^2,3 It follows that method-specific reference intervals are required for effective result interpretation.

The optimal method to determine a reference interval is to assay a minimum of 120 samples from an appropriate reference population. Clinical and Laboratory Standards Institute (CLSI) guidelines EP28-A3c recognize barriers in the clinical laboratory that may preclude local reference range determination and emphasize the utility of reference interval verification. Where appropriate published or manufacturer-quoted intervals are available, these may be adopted following verification using 20 samples from a suitable reference population. ⁴ For our method of interest, the Beckman Coulter AU immunoturbidometric assay, there was a paucity of suitable reference intervals in the literature. In this present study, we therefore sought to establish the suitability of the manufacturer-quoted range of 200–600 mg/L prompted by the observation that the frequency of low caeruloplasmin concentrations seemed too high. This is a widely quoted reference interval despite the issues of comparability of caeruloplasmin results described above. Method-specific adult reference intervals were subsequently established.

Methods

From May 2015 to April 2016, surplus blood samples were collected from patients attending the medical renal stones clinic at University Hospital Southampton. All were well and leading a normal life. Samples from patients with conditions known to impact on serum caeruloplasmin and/or copper concentrations were excluded using available clinical data. Blood was collected into BD Vacutainer™ SST™ II Advance Tubes (Becton Dickinson, Plymouth, UK) and stored for up to 72 h at 4°C. Following anonymization, samples (n = 20 for reference interval verification; n = 60 for local reference interval determination) were assayed for caeruloplasmin using the Beckman Coulter AU immunoturbidometric kit on a Beckman Coulter AU680 analyser (Beckman Coulter, Brea, USA). Analytical data were entered onto a Microsoft Excel spreadsheet. For statistical computations, AnalyseIT (Microsoft Corporation, WA, USA) was used. For the verification study, the percentage of results falling outside of Beckman AU reference interval was calculated. For the local reference interval determination, the Shapiro Wilk test was used to assess the normality of the sample population. Outliers 1.5-fold greater than the interquartile range identified using an outlier box plot were excluded (n = 2). Non-parametric rank analysis was performed on the final sample (n = 58) to determine the 2.5th and 97.5th percentiles for the reference interval.

Results

Manufacturer-quoted reference interval verification

Forty per cent of samples for the verification reference population had results less than the manufacturer-quoted lower limit of normal (200 mg/L). No samples were found to have results greater than the quoted upper limit of normal (600 mg/L).

Reference interval determination

The Shapiro Wilk test statistic (P = 0.013) indicated the reference population data to be non-normally distributed, with some skew and kurtosis evident on visual examination of the histogram plotted. The data were evaluated using the non-parametric rank method to produce a reference interval of 150–320 mg/L. Confidence intervals were not determined due to the population size being less than 120 samples.

Discussion

The importance of reference intervals appropriate for the methodology is of particular relevance to caeruloplasmin where poor quantitative standardization has long been recognized. Transference of externally derived reference intervals using verification with 20 samples from a carefully selected ‘normal’ population is acceptable to obviate the need for local reference range determination. A reference interval is considered verified if 10% or less of results for the reference population falls outside of the interval. ⁴ In the case of the Beckman Coulter AU turbidometric method, 40% of results were found to be below the lower limit of normal of the manufacturer-quoted reference interval, and thus the interval was considered invalid for our method and population. The Beckman-quoted interval is a consensus reference interval based on nephelometric analysis that was intended for interim use following introduction of IFCC CRM 470. ⁵ UK National External Quality Assessment Scheme (NEQAS) data demonstrate turbidometric caeruloplasmin methods consistently run up to around 30% lower than nephelometric methods. Therefore, it is perhaps unsurprising that the reference interval was not successfully verified.

The absence of suitable published or manufacturer-quoted reference intervals necessitated in-house determination of method-specific ranges. The medical renal stones clinic patients are ideal reference population subjects typically being free from acute illness or other conditions relevant to copper or caeruloplasmin metabolism. In alignment with the Human Tissue Act 2004, surplus specimen can be utilized for performance assessment activities without informed consent if they are appropriately anoymized. ⁶ Although a minimum of 120 samples for reference intervals comprising 95% of the normal population is the ideal, the time to collect this number of specimens was prohibitive and CLSI EP28-A3c guidelines state a minimum of 39 specimens is acceptable. Following exclusion of outliers, non-parametric rank analysis was performed on a final sample size of 58 to establish an adult caeruloplasmin 95% reference interval of 150–320 mg/L. The interval is considerably lower than the range we initially sought to verify and appears to reflect the bias evident on UK NEQAS data. This work reiterates the importance of method-specific reference intervals for effective clinical use and demonstrates the relative simplicity with which these can be generated in-house. Application of the new interval has resulted in fewer patients misclassified as having a low caeruloplasmin concentration and reduced unnecessary further investigations for Wilson’s disease.