Search

The biochemical composition of "unstimulated" whole saliva was determined in healthy adult subjects. Based on their relative concentration, salivary analytes could be classified into three arbitrary categories: concentration lower than in serum (saliva/serum ratio < 0.5; 12 analytes), similar to serum (ratio = 0.5-1.5; five analytes), and higher than in serum (ratio > 1.5; five analytes). Consistent with local production, an elevated lactate dehydrogenase (LDH) activity in the saliva was associated with a non-serum like LDH isoenzyme pattern: LDH5 >> LDH4 > LDH3 >> LDH2 > LDH1. Compared with serum, the concentrations of hydrogen (as reflected in the pH), potassium and inorganic phosphorus were much higher (saliva/serum ratio > or = 3), whereas that of sodium, total magnesium, chloride, and total carbon dioxide were lower (saliva/serum ratio < or = 0.3). The concentration of ionized calcium was similar in saliva and serum (saliva/serum ratio = 0.8), while ionized magnesium was unmeasurable in saliva. The salivary ionized calcium fraction was higher (0.76) than previously suggested (0.51). The difference between the main salivary cations (potassium, sodium), and anions (phosphate, chloride) was similar to serum (anion gap: 4 vs. 11 meq/l). Highly significant (p < or = 0.012) correlations occurred among salivary pH, dihydrophosphate, total calcium, and potassium. Our data suggest that calcium, potassium, chloride and phosphates are the major salivary complex-forming ions. The major compositional differences between serum and saliva show that saliva is not a passive "ultrafiltrate" of serum and salivary constituents may play a distinct physiological role.

In a two-center (Academic Medical Center, The Netherlands, and National Institutes of Health, USA) study, we compared ionized magnesium (iMg2+) results in serum determined with the AVL 988/4, KONE Microlyte 6 and NOVA CRT, which are the currently available analyzers equipped with a magnesium ion-selective electrode. The comparison was performed with frozen serum samples from normal individuals and patients. Imprecision and reference intervals were established. We found the best agreement between the KONE(x) and AVL(y) magnesium ion-selective electrodes (y= 0.972x-0.013; n=138) with samples from patients. With samples from normals, all three analyzers reported significantly different results (p<0.05). Best precision was found using the NOVA; coefficients of variation established at three levels were all < 4.0%. Coefficients of variation for the AVL and KONE were <5% at normal and high iMg2+, but 10.7 and 9.4%, respectively, at iMg2+ approximately 0.30 mmol/l. The reference intervals (mean+/-standard deviation) based on measurements in fresh serum samples were different for each analyzer: 0.55-0.63 mmol/l for AVL, 0.470.57 mmol/l for KONE and 0.43-0.55 mmol/l for NOVA. Thus, significant differences among the ionized magnesium concentration obtained with the three analyzers, limit comparison of results in clinical practice, and need to be resolved (e.g. by improvement of specificity and standardization of calibrators).

Changes in serum total and ionized magnesium (Mg and Mg2+) and calcium (Ca and Ca2+) were monitored in three patients who transiently developed severe (total Mg < 0.50 mmol/l) to profound hypomagnesemia (total Mg < 0.35 mmol/l) due to cisplatin or interleukin-2 therapies. Mg2+ and Ca2+ were measured with the Nova ion-selective electrodes at 37 degrees C and all results were normalized to pH 7.40. Independent of the etiology, the Mg2+ fraction (Mg2+/total Mg) increased as the concentration of the serum total Mg decreased in all three patients. When the total Mg was around or below 0.35 mmol/l the Mg2+ approached or exceeded total Mg, suggesting an error in the measurement of Mg2+. The findings were extended by including a group of 31 additional patients whose serum total Mg, Mg2+, total Ca, and Ca2+ concentrations varied from abnormally low to above normal. The serum total and ionized concentrations strongly correlated for both Mg (r2 = 0.88) and Ca (r2 = 0.92). The Mg2+ fraction rapidly increased with a fall in the total Mg concentration (r2 = 0.76) and total Mg/total Ca ratio (r2 = 0.71). In fact, with decreasing total Mg concentrations or total Mg/total Ca ratios, the Mg2+ fraction progressively increased to 93-128% of the total, confirming an error in the Mg2+ determinations. The Ca2+ fraction showed a slight and insignificant decrease with falling total Ca concentrations and total Mg/total Ca ratios. The Mg2+ concentration was directly related (r2 = 0.62), whereas the Ca2+ concentration showed a complex relationship to the total Mg/total Ca ratio. Whether this latter relationship represents a technical artifact or a true biological phenomenon requires further study. The apparent overestimation of Mg2+ at very low total Mg concentrations, and in the presence of a very low total Mg/total Ca ratio, could be due to improper chemometric correction of the Ca effect on the Mg electrode, non-linearity, and inadequate calibration. Whatever the mechanism, the failure of this method to correctly measure very low serum Mg2+ concentrations in the sera of patients with severe hypomagnesemia, or likely in any patient with an unusually low total Mg/total Ca ratio, erodes its diagnostic usefulness.