(CSEMP015) THE USE OF ADJUSTED CALCIUM IN ALBERTA: A POPULATION-BASED STUDY

Background: Adjustment of total calcium is commonly performed in clinical settings in order to account for states of altered protein-binding. However, studies have shown that albumin-adjusted calcium correlates poorly with ionized calcium, and yet still remains a common practice. Therefore, we aimed to describe the epidemiology of calcium measurements in the province of Alberta by evaluating the correlation between adjusted calcium and total calcium against contemporaneously measured ionized calcium as a reference standard.

METHODS AND RESULTS: We assembled a population-based cohort of all adults who were tested for serum total calcium and ionized calcium within a 24-hour interval in Alberta between January 1, 2013, to October 31, 2019. When available, we additionally extracted data for serum albumin, serum creatinine, and pH, as these variables may be incorporated into calcium adjustment formulas. We then analyzed the correlation between total calcium and adjusted calcium using the Payne (total calcium (mg/100mL) – albumin (g/100mL) + 4), Payne simplified (total calcium (mmol/L) + 0.02 [40 – albumin (g/L)]), and six other formulas against ionized calcium. Overall agreement for each pair of comparisons were assessed with correlation coefficients and linear regression methods. Furthermore, we determined the frequency of serum total calcium, ionized calcium and albumin that were measured within 24 hours to estimate ordering pattern of serum calcium level in Alberta.

A total of 74 971 patients had total calcium and ionized calcium sampled within 24 hours, among whom 52 469 individuals also had a coinciding serum albumin measurement. Unadjusted total calcium correlated modestly with ionized calcium (R2=51.1%) as well as all correction formulas (R2=40.2-66.%). The simplified Payne formula, which is the most commonly used method in clinical practice, performed similarly to total calcium (R2=50.9%), while the original Payne formula had a weaker correlation (R2=44.8%). Antonio formula (equation 2) had the best correlation overall (R2=66.4%) and Pekar’s formula had the weakest correlation with ionized calcium (R2=40.2%) despite being the most complex formula, adjusting for albumin, kidney function, and pH (see Table 1). In the same study period, we identified 7,356,939 tests for total calcium amongst which the majority (58.0 %, N=4,263,000) also had an albumin result within 24-hour interval (see Figure 2).

Conclusion: Most total calcium tests were ordered within 24-hours of a serum albumin suggesting that albumin-adjusted calcium is still commonly used in clinical practice, however our preliminary data showed that total calcium correlates similarly to most adjustment formulas with ionized calcium, including the commonly used simplified Payne formula.