Assessment and clinical course of hypocalcemia in critical illness

Abstract

Introduction: Hypocalcemia is common in critically ill patients. However, its clinical course during the early days of admission and the role of calcium supplementation remain uncertain, and the assessment of calcium status is inconsistent. We aimed to establish the course of hypocalcemia during the early days of critical illness in relation to mortality and to assess the impact of calcium supplementation on calcium normalization and mortality.

Methods: Data were collected on 1,038 admissions to the critical care units of a tertiary care hospital. One gram of calcium gluconate was administered intravenously once daily to patients with adjusted calcium (AdjCa)<2.2 mmol/L. Demographic and outcome data were compared in normocalcemic (ionized calcium, iCa, 1.1-1.3 mmol/L) and mildly and severely hypocalcemic patients (iCa 0.9-1.1 mmol/L and <0.9 mmol/L, respectively). The change in iCa concentrations was monitored during the first four days of admission and comparisons between groups were made using Repeated Measures ANOVA. Comparisons of normalization and outcome were made between hypocalcemic patients who did and did not receive calcium replacement according to the local protocol. The suitability of AdjCa to predict low iCa was determined by analyzing sensitivity, specificity and receiver operating characteristic (ROC) curves. Multivariate logistic regression was performed to determine associations of other electrolyte derangements with hypocalcemia.

Results: 55.2% of patients were hypocalcemic on admission; 6.2% severely so. Severely hypocalcemic patients required critical care for longer (P=0.001) compared to normocalcemic or mildly hypocalcemic patients, but there was no difference in mortality between groups (P=0.48). iCa levels normalized within four days in most, with no difference in normalization between those who died and survived (P=0.35). Severely hypocalcemic patients who failed to normalize their iCa by day 4 had double the mortality (38% vs. 19%, P=0.15). Neither iCa normalization nor survival were superior in hypocalcemic patients receiving supplementation on admission. AdjCa<2.2 mmol/L had a sensitivity of 78.2% and specificity of 63.3% for predicting iCa<1.1 mmol/L. Low magnesium, sodium and albumin were independently associated with hypocalcemia on admission.

Conclusions: Hypocalcemia usually normalizes within the first four days after admission to ICU and failure to normalize in severely hypocalcemic patients may be associated with increased mortality. Calcium replacement appears not to improve normalization or mortality. AdjCa is not a good surrogate of iCa in an ICU setting.

Figure 1

Flow chart outlining calcium supplementation protocol followed on units in which the study was carried out.

Figure 2

Time course of mean ionized calcium concentrations over the first 4 days of admission. The mean ionized calcium concentrations for various population subgroups at admission and at 06:00 for the following 3 days are plotted. Repeated measures ANOVA was carried out for linear changes in iCa over the period. F and P values are based on the Greenhouse-Geisser adjustment. (A) Changes in mean ionized calcium concentrations for normocalcemic, mildly hypocalcemic and severely hypoclacemic patients. Time effect F = 273.6; P < 0.001, effect of calcium status on time effect F = 102.7; P < 0.001. (B) Comparison of changes in mean iCa concentration for patients that survived and those that died. Time effect F = 82.4; P < 0.001, effect of survival group on time effect F = 1.06; P = 0.35. (C) Comparison of changes in iCa between those who did and did not have sepsis within 3 days of admission as determined by a senior consultant using APPC/SCCM consensus guidelines. Time effect F = 99.2; P < 0.001, effect of sepsis status on time effect F = 2.22; P = 0.058. (D) Comparison of changes in iCa levels in patients with ionised hypocalcemia on admission who also had adjusted calcium <2.2 mmol/L and therefore received calcium supplementation and those who did not. Time effect F = 135.1; P < 0.001, effect of adjCa status on time effect F = 1.86; P = 0.15.

Figure 3

ROC curves for albumin-adjusted calcium in predicting iCa <1.1. Receiver operating characteristic curves displaying the performance of albumin-adjusted calcium in predicting hypocalcemia. Different lines represent different formulae used, locally defined and in the literature [23]. Hypocalcemia is defined as iCa <1.1 mmol/L.