Low Serum Total Calcium Concentration as a Marker of Low Serum Ionized Calcium Concentration in Critically Ill Patients Receiving Specialized Nutrition Support
Departments of * Clinical Pharmacy and Correspondence: Roland N. Dickerson, PharmD, Professor of Pharmacy, University of Tennessee Health Science Center, 26 South Dunlap St, Room 210, Memphis, TN 38163. Electronic mail may be sent to rdickerson{at}utmem.edu.
Background: The intent of this study was to ascertain to what
extent serum total calcium concentration (tCa) <7 mg/dL reflects
hypocalcemia (defined by ionized calcium concentration [iCa] of
Aberration in calcium metabolism is an important, but often undetected,
metabolic problem in the critically ill patient. Severe hypocalcemia, if
untreated, can lead to serious neurologic and cardiovascular
complications.1,2
However, physiologic evidence of compromised cardiovascular contractility and
increased premature ventricular contractions due to asymptomatic hypocalcemia
have been described in the literature at significantly higher serum ionized
calcium concentrations (iCas) than those concentrations resulting in tetany or
torsades de
pointes.2–10
Hypocalcemia often goes unrecognized in the critically ill patient because
serum proteins, particularly albumin, decrease in response to
stress.11 It is
assumed that the serum total calcium concentration (tCa) is falsely low
because the serum albumin concentration is also depressed. We have previously
shown that 21% of critically ill trauma patients have hypocalcemia (as defined
by a serum iCa of
Adult patients (18 years or older) admitted to the trauma, surgical, medical, burn, or neurosurgical intensive care units, trauma stepdown unit, or the progressive care unit from August 2002 to December 2005 and who were referred to the nutrition support service were retrospectively identified for potential inclusion into the study. Serum chemistries, arterial blood gas measurements, nutrition markers, and iCa used in this study were simultaneously obtained from each patient. The blood was obtained at approximately 3 AM via an indwelling arterial or venous catheter while the patient lay supine in bed. Patients with a serum tCa of <8 mg/dL and a serum iCa performed on the same day were considered for evaluation. A serum tCa of <8 mg/dL was arbitrarily chosen as a potentially hypocalcemic value as it been shown to have a significant prognostic implication in the assessment of the severity of acute pancreatitis.13,14 Nutrition assessment measurements (serum albumin and prealbumin levels, height, weight, weight loss history) were also conducted. Laboratory tests were ordered either by the patient's primary service or the Nutrition Support Service and performed by the hospital laboratory as part of the patient's routine clinical care. When laboratory tests were ordered by the Nutrition Support Service, a "complete set" of tests (including levels of serum iCa and tCa, albumin, prealbumin, other electrolytes, liver function tests, triglycerides, international normalized ratio for prothrombin time (INR), and complete blood count with differential) were obtained on the day after initiation of nutrition support. Only the first simultaneous serum iCa-tCa determination was used for each patient who contributed once to the data pool. The serum tCa was measured using a sequential multiple autoanalyzer, whereas the serum iCa was determined using an ion-selective electrode method using the blood gas analyzer from the hospital laboratory.
Most patients were evaluated within the first week after admission to the
hospital. Those who received calcium, vitamin D, furosemide, bumetanide, or
therapeutic doses of heparin within 24 hours before the ionized calcium
measurement were excluded from the analysis. Additional exclusion criteria
included patients with hyperphosphatemia (>6 mg/dL), severe hypomagnesemia
( Patients were given enteral nutrition by a small-bore, nasoenteric feeding tube or jejunostomy. Parenteral nutrition (PN) was given via the subclavian or jugular vein when enteral feeding was contraindicated. The enteral formulas contained 40 mEq of calcium/L and the PN solution contained 5 mEq of calcium gluconate/L. The calcium gluconate content of the PN was increased to 10 mEq/L and IV calcium supplementation was given if the patient's iCa indicated severe hypocalcemia. Hypocalcemic patients receiving enteral nutrition received IV calcium gluconate supplementation, without additional calcium added to their feeding. Patients who required phosphorus supplementation were given an IV dosage scheme that has been previously shown to not significantly influence serum iCa or tCa.16,17
Continuous data were expressed as mean ± SD. All data analysis was
conducted using SigmaStat for Windows, version 3.10 (Systat Software Inc,
Point Richmond, CA). The data were evaluated for normality of the distribution
using the Kolmogorov-Smirnov normality test. Comparisons of interval data
between 2 independent groups were performed by the Student's t-test
for unpaired variables. One-way analysis of variance with post hoc pairwise
comparisons using the Tukey test was used for multiple group comparisons.
Nominal data were evaluated by The study was approved and conducted in accordance with the guidelines established by the University of Tennessee Health Science Center institutional review board. Because all measurements were performed as part of the routine metabolic evaluation of the patient and confidentiality procedures for the patient were maintained, the requirement for informed consent was waived.
A convenience sample of 195 patients was retrospectively enrolled into the study. According to our previous experience with assessing hypocalcemia in 100 consecutive critically ill trauma patients receiving specialized nutrition support,12 about half of the total patient census would be excluded due to a serum tCa of  8
mg/dL. Multiple trauma patients composed 91% (n = 177) of the population and
had a mean ISS of 31 ± 13. The patients were studied 4.2 ± 3.1
days after hospital admission for calcium status. One hundred eighty-two
patients (93%) were studied within the first week, with the remaining patients
studied within 8–14 days post–hospital admission. One hundred
seventy-four patients (89%) were ventilator-dependent. Specialized nutrition
support was initiated 2.8 ± 1.8 days after admission to the hospital:
84% of the patients received enteral nutrition, 15% received PN, and 1%
received combined enteral/PN at the time of calcium evaluation. Other
demographic data are given in Tables
1 and
2.
About one-third of all patients with a serum tCa <8 mg/dL experienced
hypocalcemia (iCa
Serum iCa correlated significantly with serum tCa (p < .001); however, there was substantial dispersion of the data, resulting in a poor association (r2 = 0.075; Figure 1). The preponderance of hypocalcemia in patients with a serum tCa <7 mg/dL can be observed when examining the data enclosed within the drawn perimeter box in Figure 1. Forty-three patients (22% of the population) were hypokalemic (serum potassium concentration <3.5 mEq/L) at the time of calcium evaluation. Of those patients with severe hypocalcemia (iCa <1 mmol/L), 3 had a serum potassium of 3–3.4 mEq/L and 1 had a serum potassium <3 mEq/L (2.4 mEq/L). No evidence of arrhythmias from hypokalemia or hypocalcemia was noted in the patients' medical charts.
Hypocalcemia has become an increasingly recognized complication associated with critical illness.1,12,18–25 Despite this emergence of data, hypocalcemia is still often unappreciated in the critically ill patient. It has been erroneously assumed12 that serum tCa is low but iCa levels remain normal because serum proteins, particularly albumin, decrease in concentration through transcapillary escape in response to stress.11 We have previously shown a 21% incidence of hypocalcemia in critically ill multiple trauma patients receiving specialized nutrition support.12 Of particular interest in that study, we observed that patients who had a serum tCa <7 mg/dL tended to be hypocalcemic (serum iCa  1.12 mmol/L).
Unfortunately, the number of patients in that subset of patients was small and
it was difficult to make any firm conclusions. The intent of this study was to ascertain to what extent a serum tCa <7 mg/dL reflects hypocalcemia. A secondary intent was to identify any other potential variables that may help to identify hypocalcemic patients. Patients with a low serum tCa of <8 mg/dL were chosen as potentially hypocalcemic as determined by serum iCa. This concentration was chosen as the point of demarcation for evidence of potential hypocalcemia, in part, due to its significant prognostic value in the assessment of the severity of acute pancreatitis (eg, Ranson's criteria).13,14 Within the selected population of critically ill patients receiving specialized nutrition support and a serum tCa of <8 mg/dL in this study, we found a 31% incidence of hypocalcemia in contrast to our previous observation of 21% in critically ill multiple trauma patients.12 This higher incidence is most likely due to bias in the selection process used in this study to preselect those patients with a low total serum calcium concentration as opposed to examining all trauma patients referred to our service irrespective of serum tCa as in the first study.12 Additionally, this population is reflective of a high acuity of illness as evidenced by an ISS of 31 ± 13, which is similar in acuity, if not slightly more sick, to previous studies from this institution that demonstrated the clinical benefits of enteral nutrition over PN26 (ISS, 25 ± 12 vs 25 ± 13, respectively) and examined the impact of an immune-enhancing diet vs conventional diet27 (25 ± 13 vs 28 ± 12, respectively) in multiple-trauma patients.
Our data in this study indicated that as the serum tCa declined, the
incidence of hypocalcemia (according to serum iCa) rose. With each 0.5 mg/dL
decrement in serum tCa below 8 mg/dL, the incidence of hypocalcemia
dramatically increased from 11% to 33% to 85% (p < .001;
Table 1). However, serum iCa
and tCa levels correlated poorly as there is considerable dispersion of the
data (Figure 1). Of particular
note is the relationship between serum iCa and serum tCa when the serum tCa is
<7 mg/dL. The vast majority (85%) of the data points reside within the
hypocalcemic range (iCa As anticipated, patients with severe hypocalcemia (iCa <1 mmol/L) had the lowest serum tCa (Table 2). The severely hypocalcemic patients had a significantly higher mean serum glucose concentration than the other groups. However, it is unclear whether this difference in serum glucose concentration was due to differences in stress, carbohydrate intake at the time of the measurement, or differences in the incidence of diabetes mellitus between groups. However, the white blood cell count was significantly increased and the mean arterial pH was significantly lower (Table 2). Given the abundance of literature describing the detrimental effects of acute illness upon calcium homeostasis,18,19,21,22,28–36 it may be hypothesized that these data indicate that severe hypocalcemia (iCa <1 mmol/L) may be associated with a greater severity of illness. This supposition might be plausible according to our population demographics (intensive care unit or stepdown unit patients, mostly ventilator-dependent, all requiring specialized nutrition support). However, examination of mean ISSs stratified into groups by severity of hypocalcemia would suggest a similar level of acuity of illness for each group (p = NS; Table 2). The severe hypocalcemia group's lower arterial blood pH (Table 2) would seem incongruent with hypocalcemia and with the results found in our previous study.12 In the presence of alkalemia, binding of free calcium to serum proteins is increased, which results in a lower serum iCa. It has been suggested that serum iCa be corrected by a factor of 0.0437 or 0.0538 for every 0.1-increment increase or decrease in pH outside of the normal range. As a result, patients undergoing aggressive management for acidemia or alkalemia will likely have a significant change in serum iCa upon completion of therapy for the acid-base disorder. In our previous study, 35% (6 of 17) of the alkalemic population (pH >7.45) were hypocalcemic in contrast to 19% (15 of 77) of the nonalkalemic patients.12 These data suggest that some patients in the current study may have had a lower serum iCa than measured after correction for pH. These findings may also suggest that acute illness and the systemic inflammatory response syndrome overwhelms the contrasting impact of acidemia on serum iCa.
Serum iCa is the preferred method for assessing calcium status in critically ill patients. A serum tCa of <7 mg/dL is associated with a high rate (85%) of hypocalcemia (iCa 1.12 mmol/L). According to these findings,
we have modified our current practice when the event of an unknown serum iCa
for a critically ill patient with a serum tCa of <7 mg/dL occurs. We either
obtain an emergent serum iCa or provide empiric conservative
treatment,23–25
with a follow-up serum iCa determination performed for reevaluation of calcium
status. This project was supported in part by grant D34HP01032 from Health Resources and Services Administration of the US Department of Health & Human Services and the Minority Center of Excellence of the University of Tennessee College of Pharmacy.
Nutrition in Clinical Practice, Vol. 22, No. 3,
323-328 (2007) This article has been cited by other articles:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Top
Materials and Methods
Results
2 analysis. Goodness of fit of
the linear model between 2 variables was assessed from the coefficient of
determination (r2) which was derived from linear
correlation using the Pearson product moment correlation coefficient. A
p value of 
