Rethinking Ionized Calcium Test Utilization
New Evidence Supports Selective Rather Than Routine Measurement
By Bill Malone
Despite a lack of evidence that calcium supplementation for the critically ill affects outcomes, regular ionized calcium testing is common for hospitalized patients, with many receiving IV calcium therapy. A recent study lends credence to the growing suspicion that such testing and therapy are frequently avoidable. This issue of Strategies examines those findings.
Regulated by the parathyroid and vitamin D axes, calcium is essential for basic cell functions, neural transmission, and blood coagulation. Clinicians have preferred measuring patients’ ionized calcium (iCa) concentration over total calcium (tCa) because the ionized form is the species that is physiologically regulated and relevant in the body, with low levels causing cardiovascular and neuromuscular problems. As the understanding of iCa’s role has expanded, so too has utilization of iCa testing, especially in critically ill patients in whom hypocalcemia is common. For these patients, clinicians may initiate calcium gluconate therapy with the goal of boosting iCa levels and protecting physiologic systems that rely on a tightly-regulated iCa level.
"There is no doubt that if you want to assess the physiologically important concentration of calcium in the human body, iCa is the way to go," remarked Geoffrey Baird, MD, PhD, assistant professor in the Department of Laboratory Medicine at the University of Washington, Seattle. "That said, you then have to question what the conditions and situations are in which it is actually important to do that."
That’s exactly what Baird and his colleagues set out to do. Despite the apparent logic of assessing iCa to detect hypocalcemia, there is little evidence that calcium therapy improves outcomes in most critically ill patients, said Baird. Compared with tCa, iCa testing is also expensive and labor-intensive because no automated instrument currently runs the test. Seeking to refine utilization of iCa testing, Baird and his University of Washington colleagues designed a three-part investigation at two academic medical centers. In the newly published study, they compared iCa test data with pharmacy-derived calcium therapy data, instituted an algorithm to replace iCa with tCa in routine calcium monitoring, and scrutinized in-hospital deaths and discharge diagnoses for outcomes (Clin Chem 2009;55:533–540).
Looking for Answers
A major reason the volume of iCa testing was so high was simple: both hospitals in the study included a check box on admissions order forms for the test. According to the authors, the order form led to many hospitalized patients getting tested daily. The frequent observation of results below the reference interval seemed to justify the practice, and also led to protocols specifying automatic calcium gluconate therapy if the results of the test were below some threshold. "It snowballed to the point that on some of our wards, you could basically measure hospital occupancy by the number of calcium tests," said Baird. "Clearly there are groups of patients that need very close calcium measurement, but on some floors in our hospitals it was pretty much done on every single person at least once every day."
Based on this initial observation, the authors decided to combine the iCa test data with calcium dose data in a correlated database. They collected 58,040 iCa results from 8,726 patients over a 9-month period, then matched up this data to a query of the pharmacy database for the same time period for all instances of calcium therapy. The patients’ initial iCa concentration, the size of the calcium dose, and the change in iCa went into a multiple linear regression model, with calcium dose and initial iCa as independent variables.
The results were striking. The increase in mean iCa concentrations in hypocalcemic patients was directly proportional to the difference between the starting iCa concentration and the population mean for the sample of 1.18 mmol/L. The model showed only a tiny, clinically insignificant dependence of iCa change on the size of the calcium dose, but a larger significant dependence on the starting iCa concentration. Essentially, the high rate of calcium therapy that followed ubiquitous iCa testing was not doing much at all.
Baird theorizes that the critically ill just have a lower set point for this tightly regulated ion, meaning that the iCa test results just showed a regression to the mean. "If you have an iCa that’s just bouncing around a tightly regulated mean, when it falls below the mean at one point, most likely the next time the test is performed it’s going to be closer to the mean," he explained. "Regardless of whether we gave them calcium, the change between iCa at one time and the next was proportional to the distance that the first measurement was from the mean. It just shows that it’s being regulated."
To get iCa utilization under control, Baird had tried discussing the problem with physicians during grand rounds. They had generally agreed with his assessment that the test was overutilized, but he didn’t see any difference in orders coming into the lab. He also met with heads of service for wards that were the highest users, again without success. Finally, Baird and his team decided on a different approach that changed calcium testing routines altogether.
Getting Physicians Onboard
Instead of trying to directly challenge physicians’ entrenched dependence on iCa, the researchers looked to the less sensitive, but more efficient, tCa measurement. Baird and his colleagues worked on a screening approach using tCa and decided on a threshold of 2.00 mmol/L. The labs would offer reflexive testing for iCa only if the tCa value was <2.00 mmol/L or >2.55 mmol/L, indicating a reasonable likelihood that iCa would be low or high enough to warrant clinical concern. The researchers met with the necessary forms committees, culled all the forms with a check box for daily iCa testing, and then initiated the reflexive strategy using tCa; however, they gave physicians the option to order iCa as a write-in test. Physicians supported the intervention because they still could order an iCa test, said Baird. "There was a stark realization among almost all of our clinicians on the wards that many tests like this were being overutilized, so it didn’t take too much pushing on our side."
When they compared iCa testing in the year before the screening strategy intervention to the year after the change, the researchers found a 72% reduction in iCa testing at one hospital and 76% at the other. Parallel to this result, they noted an 81% decline in IV calcium gluconate usage at one hospital and 45% at the other.
The team also carefully followed the outcomes of the intervention, which Baird emphasized was especially important given the dramatic decrease in both iCa testing and in administration of calcium therapy. They found no significant increase in deaths, cardiac arrests, or tetany, but did discover a 51% decrease in the frequency of discharge diagnoses of hypocalcemia, in addition to a statistically significant decrease in the frequency of seizures. "You would expect that if we were ignoring hypocalcemia that people would get quite ill and there would be worse outcomes, but we could find no measurable effect, either anecdotally or statistically," said Baird. "So at a time when we started using 60% or less calcium therapy we actually saw less diagnosed hypocalcemia, suggesting to me that a lot of hypocalcemia that is seen is ‘laboratory hypocalcemia,’ but not clinically significant." He also noted that the change
undoubtedly would lead to a lot of downstream cost savings, both by using a cheaper automated screening test and by eliminating ineffective therapeutic interventions and the risks associated with these.
Baird likens hypocalcemia in critically ill patients to an acute phase reactant, an idea that aligns with other research that has shown it to be closely tied to acute illness. Julie Robertson Zivin, MD, who led a 2001 study that solidified the association between hypocalcemia and critical illness, noted that even when acutely ill patients were receiving calcium more than once a day, usually their levels wouldn’t come up until they improved clinically. "What we’re doing with replacing calcium is really unknown. The effect is so transient," she said. "But the real question is still which patients do we have to worry about, and how low is too low?" Even with the strong results of this study, Zivin stressed that it’s still too early to give up on calcium replacement, at least in some cases. Baird agreed that a prospective study is necessary to give more definitive proof to the idea that calcium replacement isn’t useful. His team is planning a prospective study of hypocalcemia in a critical care environment to focus on the outcomes associated with measuring and replacing calcium in different populations.