HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Extract Full Text (PDF) Submit an electronic Letter to the Editor about this paper Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Dewitte, K. Articles by Thienpont, L. M. Search for Related Content PubMed PubMed Citation Articles by Dewitte, K. Articles by Thienpont, L. M. Related Collections Endocrinology and Metabolism

Reliability of Measurement of Ionized Magnesium in Ultrafiltrate

Universiteit Gent, Faculteit Farmaceutische Wetenschappen, Laboratorium voor Analytische Chemie, Harelbekestraat 72, B-9000 Gent, Belgium
^a Author for correspondence. Fax 32-9-264 81 98; e-mail Linda.Thienpont{at}rug.ac.be

Because of a printer error, the Table that should have appeared with the Letter by Dewitte et al. in the January issue of this Journal (Clinical Chemistry 1999;45:157–8) was printed on the wrong page. Following is the Letter with the Table included. The printer apologizes for any confusion this may have caused.

To the Editor:

Measurements of ionized magnesium (Mg²⁺) in serum and ultrafiltrate (S-cMg²⁺ and UF-cMg²⁺) are expected to give identical results, provided that (a) ultrafiltration does not disturb the equilibrium of S-cMg²⁺; (b) the complexation behavior of Mg²⁺ is identical in both matrices; and (c) matrix effects in sensing Mg²⁺ therein are absent. In a recent article in this Journal, however, Zoppi and Cristalli (1), referring to data of a prior study (2), reported an unexpectedly low value for UF-cMg²⁺ (Table 1 ). Because of an increase of the pH in the ultrafiltrate to 8.3 attributed to the loss of dissolved CO₂ and the absence of proteins, they recalculated S-cMg²⁺ to a pH of 8.3. Because this correction could not fully compensate for the difference (Table 1 ), they also corrected S-cMg²⁺ with a bicarbonate factor. With these changes, UF-cMg²⁺ was 10% higher than the overall corrected S-cMg²⁺.

In our opinion, the approach raises several questions. Zoppi and Cristalli (1) corrected only S-cMg²⁺ for bicarbonate, and moreover, they used a factor that was derived from aqueous solutions at the actual pH (2). If used at all, such a correction should be based on the difference of the bicarbonate effect in serum and ultrafiltrate. Furthermore, recalculating S-cMg²⁺ to a pH of 8.3 and comparing it with UF-cMg²⁺ is problematic. It assumes, for serum with an actual pH of 7.4, that the serum water at the membrane interface is equilibrated to a pH of 8.3 before it passes through the membrane. Alternatively, it assumes that for measurement of cMg²⁺ in serum and ultrafiltrate at the same pH, the ultrafiltrate may be adjusted to a pH of 7.4. This approach assumes that UF-cMg²⁺ corresponds with S-cMg²⁺ at the actual pH.

Here, we discuss the approach of Zoppi and Cristalli (1) on the basis of some recent unpublished experiments, also performed with the AVL 988–4 analyzer (AVL List GmbH). Because we used 57 frozen serum samples with a mean pH of 7.9 (range, 7.6–8.2) instead of fresh samples, we investigated whether the pH dependence of S-cMg²⁺ was comparable in the serum panel used by Zoppi and Cristalli and the panel used in our laboratory. Indeed, we found a slope (0.110) similar to that of Zoppi and Cristalli (0.117) for the pH dependence of S-cMg²⁺. In other words, the ratio of S-cMg²⁺ at pH 8.3/pH 7.4 was 0.796 in our case (Note: We use the ratio when we compare the same quantity, e.g., S-cMg²⁺ at different pH values.) and 0.785 in the work of Zoppi and Cristalli (1). In addition, we adjusted the pH (at room temperature) of our samples (per two) to 7.4 with CO₂ by placing them in a box, the bottom of which was covered with solid CO₂. Afterward, the vials were closed and kept an additional 15 min at room temperature before measurement. This produced values of S-cMg²⁺ that were, on average, 72% of the total magnesium, being identical to the value reported in the original publication by Zoppi et al. (2). We concluded from these experiments that the pH dependence of S-cMg²⁺ was identical in the two serum panels.

Interestingly, the pH equilibrium in our frozen sera was completely reversible. This contrasts somewhat with the observations of Zoppi et al. (2), who used pooled sera diluted with water, supplemented with electrolyte, and tonometered with CO₂ to pH 7.1, and Altura et al. (3), who used one repeatedly frozen and thawed patient plasma pool; neither of these groups observed a decrease of cMg²⁺ with increasing pH. We can only speculate that the difference between the results reflects the different natures of the samples used by those groups and our group.

We performed ultrafiltration on 12 of the samples somewhat differently than Zoppi and Cristalli (1), however. We used a VIVASPIN concentrator (molecular mass cutoff, 5000 Da; Vivascience) to centrifuge 500 µL of serum at 8000g for 8 min at room temperature without any special anaerobic protection, and recovered 120 µL of ultrafiltrate. In addition, we measured the cMg²⁺ and pH in the original and the remaining serum and in the ultrafiltrate. The S-cMg²⁺ was virtually unchanged after ultrafiltration, and the serum pH increased by a maximum of 0.2 units. Using our pH-adapted sera, we observed much higher values for UF-cMg²⁺ than Zoppi and Cristalli (Table 1 ). Consequently, when we recalculated S-cMg²⁺ to a pH of 8.3, the value dropped below that of the UF-cMg²⁺ (Table 1 ), prohibiting an additional bicarbonate correction. Because of the observed discrepancy, we increased the ultrafiltration time to 45 min and recovered 420 µL of ultrafiltrate. The fraction of UF-cMg²⁺ then decreased from 90% to 80%, which is much closer to the fraction of 71% observed by Zoppi and Cristalli (see Table 1 ). Interestingly, the total magnesium in our ultrafiltrates (at a serum pH of 7.4) was ~10% higher than was reported in the original publication (2). We conclude from these experiments that the ultrafiltration procedure used by Zoppi and Cristalli accounted for the low UF-cMg²⁺ they observed, and not a bicarbonate effect on S-cMg²⁺.

To investigate whether serum at pH 7.4 equilibrates to pH 8.3 during ultrafiltration, we ultrafiltered our untreated as well as our pH-adjusted sera and adjusted the pH of the ultrafiltrates of the latter to 7.4 with CO₂. In this method, the pH values of the sera and the respective ultrafiltrates were similar (Table 1 ). Because of the small pH difference between untreated serum and its ultrafiltrate, we did not adjust the pH of the latter. We found for both cases 14–15% less cMg²⁺ in the ultrafiltrates than in the respective sera (Table 1 ). Moreover, the ratio of cMg²⁺ in the ultrafiltrates at pH 8.1 and 7.4 (0.86) was very similar to the ratio for the respective sera (0.84). Both findings indicate that, in our procedure, the ultrafiltrate indeed was representative for S-cMg²⁺ at actual pH and that equilibration of serum to pH 8.3 did not occur at the membrane interface. Consequently, recalculation of S-cMg²⁺ to pH 8.3 and comparison with UF-cMg²⁺ is an invalid approach.

One should keep in mind, however, that under aerobic conditions, the pH of the serum will continue to increase as the ultrafiltration time increases. Therefore, when performing ultrafiltration under aerobic conditions, one should try to keep the centrifugation times short to keep the serum pH constant. In addition, to avoid disturbance of the equilibrium of S-cMg²⁺ during ultrafiltration (e.g., by the Donnan effect), one should aim for low fractions of recovered ultrafiltrate.

The remaining difference that we observed between measured cMg²⁺ in serum and ultrafiltrate may have several explanations. Even with short centrifugation times and low recovery, the ultrafiltration procedure might disturb the equilibrium of S-cMg²⁺ to a certain extent. Furthermore, complexation of magnesium might be different in serum and ultrafiltrate. Unfortunately, investigations of this problem gave confusing results in the past [see discussion in Ref. (1)]. Interestingly, when we decreased the pH in the ultrafiltrate from 8.3 to 7.4 by equilibration with CO₂, the UF-cMg²⁺ also decreased (compare rows 1 and 2 in Table 1 ). This is the reverse of the behavior of cMg²⁺ in aqueous solutions containing physiological salt concentrations, including bicarbonate and phosphate (personal observations). For example, when we adjusted the pH of a solution containing 20 mmol/L bicarbonate (pH 7.9) with CO₂ to a pH of 7.4, the cMg²⁺ increased by ~2%. This observation does not resolve the problem, but it indicates that simplified aqueous models may not be appropriate models for serum water. Another reason might be calibration differences of the AVL instrument between aqueous and serum solutions. Such differences should exist (similar to the ones observed for ionized calcium), but we have no knowledge about the magnitude of the effects.

References

1. Zoppi F, Cristalli C. Ionized magnesium in serum and ultrafiltrate: pH and bicarbonate effect on measurements with the AVL 988–4 electrolyte analyzer. Clin Chem 1998;44:668-671. [Free Full Text]
2. Zoppi F, De Gasperi A, Gaugnellini E, Marocchi A, Mineo E, Pazzucconi F, et al. Measurement of ionized magnesium with AVL 988/4 electrolyte analyzer: preliminary analytical and clinical results. Scand J Clin Lab Investig 1996;56(Suppl 224):259-274. [ISI][Medline] [Order article via Infotrieve]
3. Altura BT, Shirey TL, Young CC, Dell'Orfano K, Hiti J, Welsh R, et al. Characterization of a new ion selective electrode for ionized magnesium in whole blood, plasma, serum, and aqueous samples. Scand J Clin Lab Investig 1994;54(Suppl 217):21-36.

This Article Extract Full Text (PDF) Submit an electronic Letter to the Editor about this paper Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Dewitte, K. Articles by Thienpont, L. M. Search for Related Content PubMed PubMed Citation Articles by Dewitte, K. Articles by Thienpont, L. M. Related Collections Endocrinology and Metabolism