Elsevier

Clinica Chimica Acta

Volume 432, 15 May 2014, Pages 122-126
Clinica Chimica Acta

Common decision limits — The need for harmonised immunoassays

https://doi.org/10.1016/j.cca.2013.11.023Get rights and content

Highlights

  • Common decision limits are recommended in many clinical guidelines.

  • Impact of method-related differences for use of common decision limits may be severe.

  • Improving method comparability is a pre-requisite for use of common decision limits.

Abstract

The main aim of clinical guidelines is to encourage the best clinical outcome for patients and the best use of resources, no matter where patients are investigated or managed. Where guidelines incorporate decision limits based on levels of analytes in serum, plasma or urine these may determine whether or not to treat or may be used to tailor further treatment. Consideration should be given to the effect of method-related differences in results when implementing common decision limits. Available evidence suggests that for some analytes the implications for the patient may be serious, e.g. in terms of missed biopsies or unnecessary prostatic biopsies when prostate specific antigen is measured. Major causes of between-method differences are reviewed and means of addressing them considered.

Introduction

One of the major benefits of the practice of Evidence Based Medicine has been the development of guidelines for diagnosis and management. The aims of such guidelines are to encourage the best clinical outcomes for patients and best use of resources, no matter where patients are investigated and managed. Often, guidelines incorporate quantitative decision criteria based on levels of analytes in serum, plasma or urine (e.g. prostate specific antigen (PSA) in men with urinary symptoms or human chorionic gonadotrophin (hCG) in women with suspected ectopic pregnancy). These quantitative criteria based on population studies may determine whether to treat or not, or as a further refinement may be used to tailor treatment in individual patients on the basis of their individual biomarker result. This move towards personalised medicine is now becoming a clinical reality [1].

Implicit in the implementation of guidelines using biomarker data is the assumption that results across all testing sites are comparable, but regrettably this important assumption is rarely if ever questioned. Laboratory results may differ between different testing sites for a variety of reasons. This may be due to imprecision, or more seriously may reflect systematic differences in method results, which can arise from inherent differences in method design. This is particularly the case for the complex and heterogeneous analytes measured by immunoassay, for which significant method-related differences in results are frequently observed. The magnitude of these differences, reasons for them and assessment of their potential clinical implications are briefly described here and several means of improving method comparability reviewed.

Section snippets

Are method-related differences in immunoassay results clinically important?

As illustrated in Table 1, analytes measured by immunoassay contribute to routine medical decision-making in a number of specialities, and particularly in endocrinology and oncology. The applications are varied, ranging from aiding diagnosis (e.g. PSA in symptomatic men) to assessing ovarian reserve (e.g. anti-Müllerian hormone (AMH) in candidates for assisted conception) to monitoring cancer progression (e.g. α-fetoprotein (AFP) and hCG measurement in patients with germ cell tumours). Decision

How can the clinical impact of method-related differences be estimated?

Assessing the implications of method-related differences after implementing recommended guideline decision limits requires careful clinical audit, and ideally a direct and measurable relationship between the test result and outcome. This is particularly challenging for diagnostic tests as they usually constitute a relatively small part of a complex clinical pathway. While outcome-related analytical performance goals can be established in a variety of ways [2], demonstrating the impact of

What factors contribute to differences in immunoassay results?

As for all laboratory tests, variable pre-analytical sample handling or storage will contribute to differences observed, as will variation in procedures in different laboratories (e.g. with respect to internal quality control or analyser maintenance). Performance can occasionally also be influenced by the operator. However, four major factors contributing to differences in results for analytes measured by immunoassay are listed in Table 3 and their implications considered in more detail here.

How can adverse effects of method differences on decision limits be decreased?

From the above discussion, it is clear that method-related differences in immunoassay results pose some serious potential difficulties when attempting to apply common clinical decision limits. The severity of the consequences will depend on both the analyte and the clinical application, as will possible approaches to ameliorate them. Examples of several approaches that can be taken with the ultimate aim of improving the comparability of results and hence decision points are briefly described

Conclusions

There is an urgent requirement to improve the comparability of results for analytes for which common decision limits are recommended if laboratorians are to provide results that fulfil the needs of patients and clinicians. The approaches described above all have merit but none is likely to be perfect for every analyte. With the recent establishment of a Harmonisation Oversight Group charged with coordinating and prioritising international efforts to improve between-method agreement [11], the

Acknowledgements

I would like to thank William Egner, Finlay Mackenzie, Alan Reid and Gwen Wark for permission to quote data from the UK NEQAS centres in Sheffield, Birmingham, Glasgow and Guildford, my close colleagues Andy Ellis, Ruth Al-Sadie and Neil Syme at UK NEQAS [Edinburgh] and John Seth, for his much appreciated comments and suggestions.

References (27)

  • J.A. Gray

    The shift to personalised and population medicine

    Lancet

    (2013)
  • C.M. Sturgeon et al.

    Improving the comparability of immunoassays for prostate-specific antigen (PSA): progress and problems

    Clin Chim Acta

    (2007)
  • G.G. Klee

    Establishment of outcome-related analytic performance goals

    Clin Chem

    (2010)
  • J.S. Madsen et al.

    More studies on outcomes using biochemical diagnostic tests are needed: findings from the Danish Society of Clinical Biochemistry

    Clin Chem

    (2008)
  • A.W. Roddam et al.

    Assessing the clinical impact of prostate-specific antigen assay variability and nonequimolarity: a simulation study based on the population of the United Kingdom

    Clin Chem

    (2004)
  • J. Melia et al.

    Survey of the rate of PSA testing in general practice

    Br J Cancer

    (2001)
  • A. Almond et al.

    Current parathyroid hormone immunoassays do not adequately meet the needs of patients with chronic kidney disease

    Ann Clin Biochem

    (2012)
  • S. Steddon et al.

    Renal Association Clinical Practice Guideline in mineral and bone disorders in CKD

    Nephron Clin Pract

    (2011)
  • W.G. Miller et al.

    Proficiency testing/external quality assessment: current challenges and future directions

    Clin Chem

    (2011)
  • W.G. Miller et al.

    Why commutability matters

    Clin Chem

    (2006)
  • W.G. Miller et al.

    Roadmap for harmonization of clinical laboratory measurement procedures

    Clin Chem

    (2011)
  • L.M. Thienpont

    Accuracy in clinical chemistry — who will kiss Sleeping Beauty awake?

    Clin Chem Lab Med

    (2008)
  • W.G. John

    Use of HbA1c in the diagnosis of diabetes mellitus in the UK. The implementation of World Health Organization guidance 2011

    Diabet Med

    (2012)
  • Cited by (14)

    • Cascadion™ SM Clinical Analyzer: Evaluation of the whole blood immunosuppressants quantification and routine usability

      2023, Clinica Chimica Acta
      Citation Excerpt :

      Nevertheless, due to the lack of automation, it requires a specialized trained staff, a high hand-on-time and the use of batch mode which can increase the turnaround-time (TAT) [10]. Moreover, handling errors remain possible and LC-MS/MS standardization is challenging due to the high inter-laboratory variation [11–14]. Partial automation and commercial kits have yet improved the practicability but it remains lower than automated immunoassays that offers 24/7 random accessibility and don’t require specialized staff and equipment [15,16].

    • Comparison of two immunoassays for the measurement of serum HE4 for ovarian cancer

      2021, Practical Laboratory Medicine
      Citation Excerpt :

      In meta-analyses reporting the pooled sensitivity and specificity of HE4 for ovarian cancer detection, there is marked heterogeneity in assay method and cut-offs applied[4,5]. Analytes measured by immunoassay show method-related differences in concentration, yet decision limits for biomarkers including CA125 are fixed across laboratories and guidelines[6]. It is important to compare new analytical methods with existing ones to ensure that results are comparable and are appropriately interpreted.

    • Generating method-specific Reference Ranges – A harmonious outcome?

      2017, Practical Laboratory Medicine
      Citation Excerpt :

      For other chemistry tests, method specific RRs are justified e.g. LDH, where there is a range of commercially available assays and sub-types e.g. Lactate to Pyruvate and vice versa giving very different results (two-fold). When assays use calibration material that is traceable to certified standards, through a reference measurement system, result comparability across manufacturers’ methods may be augmented [10,17]. Such tests are thereby potentially suitable candidates for RR harmonisation.

    • The molecular relationship between antigenic domains and epitopes on hCG

      2016, Molecular Immunology
      Citation Excerpt :

      Non-standardized Abs’ specificities still leads to differences in results either due to selective over-specificity in variant recognition, undesired cross-reactivity or interference of non-measured metabolic variants like hCGβcf that are only detected by one of the two Abs in sandwich assays (hook-effect) (Berger and Sturgeon, 2014; Madersbacher and Berger, 2000; Grenache et al., 2010). Thus identification and harmonization of epitopes recognized by diagnostic Abs will in future improve method comparability and consequently patient management (Sturgeon, 2014). Unfortunately a comprehensive array of anti-hCG-antibody:antigen complexes with accurate data on the molecular location of recognized epitopes is not available, due in no small part to the difficulty of crystallizing hCG and other glycoprotein hormones.

    • Pregnancy testing with hCG - future prospects

      2014, Trends in Endocrinology and Metabolism
      Citation Excerpt :

      UK NEQAS recovery experiments in which known amounts of the Fourth IS 75/589 for hCG are added to base pools of either serum or urine confirm near quantitative recovery for the all laboratory (consensus) mean although serum recoveries for individual methods vary from 84% to 133% [53]. While results for specimens containing patient sera are in reasonably good accord, with mean between-method coefficients of variation of approximately 15% [60], results vary significantly for some pregnancy urines (e.g., from 970 U/l to 23,120 U/l) [53]. This may reflect different concentrations of hCGβcf in different pregnancy urines, since assays (e.g., category ii), which do not recognise purified hCGβcf, give lower results.

    View all citing articles on Scopus
    View full text