The ionized calcium level in blood is known to be falsely decreased due to dilution and binding effects when self-prepared liquid heparin anticoagulants are used.1 The CSLI (Clinical and Laboratory Standard Institute, formerly the National Committee for Clinical Laboratory Standards or NCCLS) recommends the use of lyophilized calcium-balanced heparin syringes2 for collecting specimens for ionized calcium measurement. Although excess heparin has been suggested to affect ionized calcium levels,3 such an effect on the measurement of ionized magnesium has not been documented. Some commercial calcium-balanced heparin syringes have been reported to affect ionized magnesium levels, although the clinical significance was believed to be minimal.4 Liquid heparin may still be used for the determination of ionized calcium or magnesium by some clinical laboratories in developing countries due to budgetary constraints. In this study, we set out to systematically determine the liquid heparin effect on both ionized calcium and magnesium in clinical samples by head-to-head comparison.

Samples were obtained from the 44 patients to evaluate the effect of heparin on ionized magnesium levels. Informed consent was obtained from the patients and the study protocol was approved by the institutional review board.

We made user-prepared syringes using sodium heparin, 1000 U/mL (Choongwae Pharma Corporation, Seoul, Korea), which concentration was specially manufactured to reduce the heparin concentration for blood sampling and was lower than the common value prescribed for therapeutic heparinization. Heparin in the amount of 0.5 to 1.0 mL was drawn into a 2-mL syringe to coat the whole inner surface of the syringe, and was then expelled back into the original vial. Before and after coating the inner surface of the blood gas syringe with heparin, we tared the syringe on the electronic balance to calculate the amount of heparin solution remaining in the syringe after expulsion. The needle was then changed to one appropriate for collecting the sample. Two types of syringes which differed in their heparin coating protocols were compared to the reference syringe (lyophilized calcium balanced lithium heparin, BD Preset™ syringe, 30 I.U., BD, Franklin Lakes, NJ, USA) according to the behavioral pattern of the users as follows. With the type 1 syringe, heparin and air was ejected once or twice such that heparin may have been left in the dead space of syringe hub. In contrast, heparin and air were ejected methodically several times from the type 2 syringe until heparin was not detectable in either the syringe or the needle. Although bias may remain for the calcium balanced lithium heparin syringe compared to whole blood without anticoagulation,4 we used it as the reference. Drawing 1.5 ml of whole arterial blood from each subject into each syringe, the samples were then recapped until analysis. Ionized calcium and magnesium were measured in all samples within 20 minutes after collection using Stat Profile M (Nova Biomedical, Waltham, MA, USA). The 'bias corrected by dilution effect' was calculated by subtracting the expected ion concentration from the actual value. The expected ion concentration was calculated as the ion concentration of the reference syringe multiplied by the following factor.

Factor = 1.5/(1.5 + V)

where '1.5' is the volume (mL) of whole blood drawn from the patient and 'V' is the volume (mL) of the liquid heparin remaining in the syringe.

We analyzed the agreement of the three kinds of blood gas syringes for both ionized calcium and magnesium by Passing and Bablok analysis, linear regression analysis, and Pearson correlation using Analyse-it software (Leeds, England, UK). The bias of type 1 or 2 syringes was calculated against the calcium balanced lithium reference.

Table 1 shows that traces of heparin were found in user-prepared syringes, with type 2 syringes containing much less heparin than type 1.

The type 1 syringe showed significant negative bias in both ionized calcium and magnesium levels (Table 2). However, the type 2 syringe, which contained less heparin, showed different results between the two tests; there was a statistically significant negative mean bias (- 0.007 mmol/L) for ionized magnesium, but no significant mean bias (0.006 mmol/L) for ionized calcium (Table 2). The user-prepared heparinized syringes exhibited more than 10% negative bias above the level of 40 U/ml of heparin for ionized calcium (Fig. 1A), and above the level of 14 U/ml of heparin for ionized magnesium (Fig. 1B) with the user-prepared heparinized syringes. Table 3 shows that significant negative bias of the ionized magnesium compared to the ionized calcium could be observed if the dilution effect was excluded.

As the ionized fraction of magnesium is the biologically active part, its accurate determination is an important issue in critically ill patients.5,6 The CLSI recommends that whole blood determinations of ionized calcium be performed on specimens transferred into commercially available, ready-to-use syringes or capillaries anticoagulated with one of the modified preparations of heparin designed to minimize its effects on the ionized calcium.3 However, the performance of these syringes with ionized magnesium has not been documented. We found a clear reduction in measured ionized magnesium levels with increasing concentrations of heparin in syringes, confirming previous reports.2,7 We postulate that ionized magnesium measurements are more greatly impacted by liquid heparin than ionized calcium values, as the absolute amount of ionized magnesium is much less than that of ionized calcium. We did not, however, confirm its mechanism. The maximum allowable heparin concentration in syringes was proposed to be 40 U/mL or 15 U/mL for ionized magnesium.7,8 For ionized calcium, the CLSI recommends that the calcium balanced heparin concentration should be 70 U/mL or less. We found a greater than 10% negative bias above the clinical prescribed level of 40 U/mL of heparin for ionized calcium, and above the level of 14 U/mL of heparin for ionized magnesium in user-prepared heparinized syringes, due to both dilution and heparin binding effects. If the liquid heparin was deliberately and thoroughly expelled from the blood gas syringe, the mean bias in values did not exceed - 0.08 mmol/L for ionized calcium and - 0.05 mmol/L for ionized magnesium, putting the values within a clinically acceptable range.

Although Chantler and Cox showed that self-prepared heparinized syringes for measuring ionized magnesium and calcium were sufficiently close to dry balanced heparinized syringes to justify their use under economically constraining conditions,9 the protocol of thoroughly expelling the heparin solution should be mandated to minimize dilution and heparin binding effects.

In conclusion, we recommend using lyophilized calcium balanced heparinized syringes for the determination of ionized magnesium as well as ionized calcium, as we observed a greater negative bias of ionized magnesium values compared with ionized calcium in the presence of residual heparin. In the case of using user-prepared syringes, a protocol for thorough evacuation of heparin solution should be standardized and mandated.