Edmund J Lamb1, Susan Vickery1and Andy R Ellis2
1Department of Clinical Biochemistry, East Kent Hospitals NHS Trust, Canterbury, Kent CT1 3NG
2UK NEQAS for Peptide Hormones, Department of Laboratory Medicine, Royal Infirmary, Edinburgh EH16 4SA, UK
Ann Clin Biochem 2007; 44: 1–4.
Parathyroid hormone (PTH) is released from the parathyroid glands predominantly as an 84 amino-acid peptide (PTH (1--84)). Classical biological activity, mediated via interaction with the PTH receptor PTH1R, resides in the N-terminus of the molecule. C-terminal (CPTH) fragments are also directly secreted by the parathyroid gland. Following release, PTH (1--84) undergoes proteolytic degradation in Kup¡er cells in the liver: resulting CPTH, but not the N-terminal fragments, may re-enter the circulation.1 The variously derived PTH fragments and PTH (1--84) are then excreted renally.
In the setting of chronic kidney disease (CKD), assessment of plasma PTH concentration is used as a surrogate measure of underlying renal bone disease. Measurement of PTH in the presence of renal impairment has always been problematic. In the 1970s, it became apparent that PTH radioimmunoassays crossreacted with CPTH fragments that accumulated in the presence of renal failure. The development of a two-site immunoassay by Nichols Institute Diagnostics (Nichols Allegro immunoradiometric assay [IRMA]) in the 1980s, described as an ‘intact’ PTH assay, seemed to resolve this issue.2 This assay became the bench mark against which other assays were developed. Further, the somewhat limited evidence correlating renal bone disease and plasma PTH concentration, and upon which guidelines have been based, relates largely to experience with this prototype second-generation assay. Currently a range of manufacturers o¡er intact PTH assays on a variety of automated platforms. Notably, Nichols Institute Diagnostics can no longer be included in that list following the demise of the company in 2006.
In the mid-1990s, a Canadian research group discovered that all was not as it might seem.3--5 In fact, intact PTH assays seemed to cross-react to a large degree with a CPTH fragment thought to be PTH (7--84) and other less prevalent fragments of PTH that accumulated in renal failure. It would appear that cross-reactivity with this fragment had been accounting all along for approximately 50% of the immunoreactive PTH we had been measuring in such patients, although the exact degree of cross-reactivity varies between methods. Putting aside the perhaps more trivial matter of whether you should continue to market your assay as an intact PTH assay when it is clear that it is not (many manufacturers still do so), the question of how well di¡erent PTH assays agree with each other and whether measuring PTH (7--84) has clinical relevance in kidney disease needs addressing.
It is evident that second-generation PTH assays do not agree among kidney disease patients, either with each other or with the original Nichols intact PTH assay. Among dialysis patients, a recent study has shown that commonly used commercial assays may over- or underestimate PTH by up to 50% compared with the Nichols Allegro IRMA.6 Such betweenmethod variability is also con¢rmed on a regular basis by external quality assessment data (Figure 1). This is due to a variety of factors including lack of common standardization, variation in recovery of PTH (1--84) (Figure 2) and di¡erences in the extent of crossreactivity with PTH (7--84). For example, data from a United Kingdom National External QualityAssessment Service (UK NEQAS) study in July 2004 estimated approximately 69% cross-reactivity with PTH (7--84) for the Diagnostic Products Corporation (DPC) method, 73% with the Roche ElecSys method, 102% with the Bayer Advia method and 95% with the former Nichols Advantage method.7 Indeed all 14 ‘intact PTH’assays tested appeared to cross-react with PTH (7--84).7
PTH results may be interpreted in the setting of a range of clinical recommendations by nephrologists. The UK Renal Association standard states: ‘PTH concentration should be less than four times the upper limit of normal of the assay used in patients being managed for chronic renal failure or after transplantation and in patients who have been on haemodialysis or peritoneal dialysis for longer than three months’.8 Using a cut-o¡ based on the upper limit of normal requires that the upper limit of normal is accurate and appropriate for the assay used.There is evidence to suggest that this may not always be the case. A survey conducted by UK NEQAS in July 2005 indicated that the majority of participants have an upper limit of normal close to 65 ng/L, irrespective of the assay used (Figure 3).9 Is there evidence that variations in assays contribute to clinically meaningful di¡erences? Compliance with the Renal Association standard is audited on an annual basis by the UK Renal Registry and was achieved by 66% of UK dialysis patients in 2004, but this varied between renal units from 48 to 87% compliance.10 However, it was not possible to demonstrate relationships between performance in different centres and the known assay bias or PTH (7--84) cross-reactivity of the method used. Until standardization and speci¢city issues are resolved, it will remain di⁄cult to ascertain the true contribution of analytical variation to clinical performance. If a concentration is to be used as a target value, then it would seem desirable to recognize the e¡ect of assay bias and to set assay-speci¢c cut-o¡s. However, in the USA, the National Kidney Foundation have proposed a single target for dialysis patients of o300 ng/L, irrespective of the PTH assay used.11
Other management thresholds may be a¡ected. For example, cinacalcet hydrochloride, a calcimimetic agent that increases the sensitivity of the calciumsensing receptor, thereby inhibiting the release of PTH, is under review by the National Institute for Health and Clinical Excellence (NICE) for use in secondary hyperparathyroidism. NICE’s preliminary recommendations state that cinacalcet may be used in patients with refractory (tertiary) hyperparathyroidism if surgical parathyroidectomy is contraindicated.12 To qualify for this treatment, plasma concentrations of PTH must exceed 800 ng/L. Patients’eligibility for this therapy could be compromised by the characteristics of the local PTHassay in a‘post-code’ lottery of exactly the kind that NICE guidelines are intended to circumvent. UK guidelines for the management of renal osteodystrophy among patients with stage 3 CKD,13 partly in£uenced by the earlier US guidelines,11 have also indirectly used PTH targets linked to the second-generation Nichols Allegro IRMA. For example, evaluation of vitamin D status and correction of any underlying de¢ciency is recommended when PTH exceeds 70 ng/ L in patientswith stage 3 CKD. Clearly the patient pathway will be determined by the level of agreement of the laboratory’s PTH assay with the Nichols Allegro IRMA. Lack of concordance between PTH assays undoubtedly means that consistent application of the same decision limit does not equate to consistent care. Does PTH (7--84) have biological activity and is its measurement relevant in kidney disease? Increased resistance to N-terminal PTH may in part re£ect accumulation of CPTH fragments (mainly PTH [7--84]) as kidney disease progresses. A range of in vitro and in vivo studies have demonstrated biological activities of CPTHs including PTH (7--84) that would tend to oppose those of the intact hormone.1 These actions are presumably mediated via PTH receptors other than PTH1R. The importance of these actions in the evolution of adynamic bone disease is still unclear, but it has been suggested that ratios of PTH (1--84) to PTH (7--84) could be used to assess and stage low bone turnover as an alternative to bone biopsy. Recent studies have failed to con¢rm initial enthusiasm for this approach.14--16 Although PTH (7--84) may possess biological activity as an antagonist to PTH (1--84), any clinical role in the assessment of renal osteodystrophy remains to be con¢rmed.
In the last few years, PTH assays that claim to measure only PTH (1--84) have become available. These third-generation assays utilize polyclonal antibodies targeted to epitopes on the ¢rst few N-terminal amino acids. Resolution of some of the issues described above could be achieved using such assays. Unfortunately, these third-generation assays are not yet widely available on automated platforms suitable for the majority of laboratories. There is a tight correlation between second- and third-generation assays across a range of PTH concentrations, and the clinical advantages of using third-generation assays have yet to be clearly demonstrated.15,17,18 It is, however, becoming apparent that second-generation assays cannot provide acceptable tools for the application of clinical guidelines. Furthermore, a sound evidence base to guide future practice cannot be assembled. Until we use assays accurately calibrated in terms of an appropriate International Standard and measure a de¢ned moiety, the situation is unlikely to improve.