Abstract
To the Editor,
I read with interest the recent papers by Jassam et al., 1 , 2 regarding the use of population-specific adjusted calcium equations. In particular, the factors which may affect the performance of locally derived adjustment equations are investigated in more detail revealing many potential pitfalls.
I would however like to draw attention to some confusion regarding the use of the Pathology Harmony reference range for calcium. 3 Firstly, as noted in the reference, the range of 2.20–2.60 mmol/L was proposed specifically for adjusted calcium, not total calcium. The reason for this was that there were significant calibration issues relating to calcium assays at the time – a situation which thankfully appears to be improving as reported in a more recent paper by Jassam et al. 4 Secondly, the adjusted calcium range of 2.2–2.6 should only be applied to equations normalized to a mean calcium of 2.40 mmol/L. This method for generating a locally derived equation is as suggested in the 2015 ACB albumin-adjusted calcium position paper. 5
In consequence of this misunderstanding of the use of the range, it may be that some conclusions drawn in the papers are not strictly valid. To illustrate this with a specific example, in the paper comparing the population-specific adjusted calcium to ionized calcium, 2 Table 4 classifies the calcium status of healthy individuals against a gold standard of ionized calcium. However, the other three measures (total calcium and adjusted calcium using inpatient and community adjustment equations) use the Pathology Harmony range of 2.2–2.6 for assessment of calcium status. In the case of total calcium, the kit manufacturer’s range, or a reliably derived local reference range, would be more suitable; for the adjusted calcium, the regression equations need to be normalized to a mean calcium of 2.4 before applying the range (alternatively a separate reference range should be derived for each adjustment equation).
The adjustment equations quoted are
Inpatient equation (1)
Adjusted calcium = Total calcium–0.018 (Albumin–38.3)
(slope 0.018, intercept 1.579)
Community equation (2)
Adjusted calcium = Total calcium–0.014 (Albumin–44.9)
(slope 0.014, intercept 1.77)
Normalizing these equations to a mean calcium of 2.4 will result in
Inpatient equation (3)
Adjusted calcium = Total calcium–(0.018×Albumin) +(2.4–1.579)
= Total calcium–0.018 (Albumin–45.6)
Community equation (4)
Adjusted calcium = Total calcium–(0.014× Albumin) +(2.4–1.77)
= Total calcium–0.014 (Albumin–45.0)
Without access to the patient data, it is not possible to comment on exactly how the normalization will change the classification of calcium status of the healthy individuals, but comparison of the inpatient equations (1) and (3) suggests that normalizing will significantly reduce the number of healthy patients classified as hypocalcaemic by this equation, whereas the similarity between the community equations (2) and (4) suggests that there will be little change in this case.
The process of normalization suggests that the inpatient population may simply be a subpopulation of the general population of normocalcaemic individuals, who happen to have lower albumin values due to other pathology, although the variation in slope between the community and inpatient equations may indicate some differences in the binding characteristics as suggested by the authors.
Finally, it should be recognized that adjustment of calcium, as for all derived parameters, is a relatively inaccurate measure. As the authors point out, invalid results may occur in extremes of albumin concentration, acidosis, renal dialysis patients, neonates, etc. It is not clear, however, whether constructing a family of adjustment equations to accommodate these different groups is the best way forward. Increased use of ionized calcium measurements and further work to assess and validate the reference range seems a more promising strategy. The use of a well-validated locally derived calcium adjustment will still provide reassurance for postoperative patients, and others with low albumin values; also the identification and follow-up of community patients with marginally raised adjusted calcium and an inappropriately high PTH level in order to investigate possible subclinical primary hyperparathyroidism. The input of the clinical scientist will, however, remain essential in interpretation of the adjustment in some patient groups.
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.
