The trusted source for
healthcare information and
Abstract & Commentary
Do Serum Ionized Calcium Levels Matter in the ICU?
By Andrew M. Luks, MD, Pulmonary and Critical Care Medicine, University of Washington, Seattle, is Associate Editor for Critical Care Alert.
Dr. Luks reports no financial relationship to this field of study.
Synopsis: This retrospective study of more than 7000 patients from four combined medical-surgical ICUs demonstrated that ionized calcium levels are unrelated to mortality over a broad range of values but are independent predictors of mortality when levels are severely increased or decreased.
Source: Egi M, et al. Ionized calcium concentration and outcome in critical illness. Crit Care Med 2011;39:314-321.
Although previous studies have suggested that hypocalcemia, a common problem in critical illness, is associated with increased mortality in ICU patients, and correction of hypocalcemia has been advocated to prevent neurologic and cardiovascular complications, the literature is still unclear as to the precise relationship between abnormal calcium levels both hypo- and hypercalcemia and ICU outcomes.
Egi and colleagues addressed this question using a retrospective study of a large number of patients in multiple medical/surgical ICUs. They gathered data from 7024 patients who were part of an established quality-assurance project across four hospitals in Australia, 43% of whom were surgical patients. They included all patients admitted over a 4-year period, excluding those who required renal replacement therapy using citrate-based anticoagulation and those for whom arterial blood gas data were not available. All ionized calcium (iCa) concentrations were measured by a blood gas analyzer at 37°C and were corrected for a pH of 7.40. The highest and lowest concentrations were obtained for each patient during their ICU stay. Mild, moderate, and severe hypocalcemia were defined as iCa values < 1.15, < 0.9, and < 0.8 mmol/L, respectively, while the corresponding values for hypercalcemia were > 1.25, > 1.35, and > 1.45 mmol/L, respectively.
A total of 177,578 iCa measurements were made in the study population, corresponding to one measurement every 4.5 hours on average. The incidence of at least one episode of mild, moderate, and severe hypocalcemia was 88%, 10.8%, and 3.3%, respectively. The corresponding values for mild, moderate, and severe hypercalcemia were 22.7%, 6.7%, and 2.0%, respectively. ICU mortality was 12.1%, while hospital mortality was 21.6%. The lowest and highest iCa concentrations were significantly different between survivors and non-survivors, while the incidence of hypo- and hypercalcemia was also significantly higher in the non-survivors compared to the survivors. The probability of mortality was increased with at least one episode of hypocalcemia during the ICU stay, with the odds ratio (OR) increasing as one moved from the mild (OR, 1.46) to the severe category (OR, 2.50). A similar pattern was seen with regard to the effect of hypercalcemia on ICU mortality, with ORs of 2.00, 2.62, and 2.90 for mild, moderate, and severe hypercalcemia, respectively. Similar results were seen regarding the effects of both hypo-and hypercalcemia on hospital mortality.
In multivariate analysis, the authors detected a significant relationship between the minimum and maximum iCa measurements and ICU and hospital mortality, with the results in both cases being driven by those patients with the most severely abnormal values.
One of the staggering results in this study was the exceptionally large number of calcium measurements conducted in patients during the study period more than 177,000 total measurements and an average of one measurement every 4.5 hours! What is not included in the study results, unfortunately, are the costs associated with all of these measurements as well as the costs associated with repletion of low calcium levels that would inevitably follow reporting of low values in many centers. I suspect that both types of costs would be exceptionally high.
What might justify these interventions and the ensuing expenses and risks (e.g., skin necrosis, cardiac arrhythmia, enthesopathy) would be some sense that rectifying the low calcium levels is associated with some demonstrable change in important ICU outcomes such as mortality, length of stay, or time off mechanical ventilation. At first glance, this study would appear to provide a rationale for aggressive repletion of low iCa levels and management of severely elevated levels as mortality was significantly increased in patients with severe derangements in either range. It is important to remember, however, that this study was retrospective and, more importantly, did not examine whether aggressive correction of mild-severe derangements does, in fact, improve outcomes. Specifically, no randomized studies have demonstrated a benefit in this regard, and the severity of the derangements might simply reflect the severity of the underlying illness rather than being a casual factor in the observed outcomes.
These issues could be better clarified with a prospective, randomized, multicenter trial looking at the benefits of intervening to correct hypo- and hypercalcemia. For a variety of reasons, however, such as the relatively low cost of calcium gluconate or calcium chloride compared to other possible ICU interventions and the subsequent lack of appeal to pharmaceutical companies, and the difficulties of controlling practice across multiple centers and diverse patient populations in order to isolate the effect of calcium supplementation on patient outcomes, the likelihood of such a trial is low and the clinician is left trying to decide how aggressive to be about correcting derangements in calcium levels.
Given the U-shaped nature of the relationship between calcium derangements and morality observed in this study and the fact that the results were driven largely by abnormalities that fall within the severe range (iCa < 0.8 or > 1.45), we should probably be reserving our interventions for those patients with abnormalities that fall near or within this severe range and hold off on intervening for more mild abnormalities. Efforts should also be made to decrease the frequency of iCa measurements, particularly when initial measurements repeatedly fall in the normal-mildly abnormal range. Fewer results reported by the laboratory means fewer abnormal results to react to in an unnecessary manner.