Plasma calcium concentration modifies the blood sodium during hemodialysis: lessons from hard water syndrome


Several evidences suggest that a large reservoir of osmotically inactive sodium is represented by the extracellular matrix of bone, skin and musclesTitze J – 2014 [1]“.

It has been recently suggested that there is a balance between the osmotically active and inactive Na+ pool; thus, upon slow changes in body fluids, the composition of inactive Na+ pool changes accordingly. In rats, long-term salt deprivation is accompanied by a decrease in the charge density of skin GAGs, and the consequent mobilization of osmotically inactive Na+. This finding suggests that the skin and the connective tissues may serve as a Na+ storage, capable to release Na+ in response to reduced intake by changing its polyanionic character “Schafflhuber M – 2007 [2] (full text)” (Figure 1).

It remains unclear how this large Na+ reservoir is regulated. The exact knowledge of this interplay is important because it may provide an additional mechanism of sodium regulation, independent of the kidney function; in addition, this regulation might have clinical impact particularly in patients undergoing dialysis (who lack kidney regulation for extracellular sodium and rely on the dialysis system).

Recently, we have had the possibility to study the effects of an extracellular calcium increase on Na+ levels, in subjects undergoing chronic hemodialysis treatment. These subjects underwent by accident to a sharp increase in calcium levels due to a failure in the water control system (Hard Water Syndrome).

Our hypothesis is that, in presence of a large increase of extracellular calcium, the latter should compete with the Na+ bound to the extracellular matrix, thus leading to a linear increase of (unbound) plasmatic Na+.

We also expect that this effect is dampened by the buffering effect of the dialysis process on Na+ ions; however, a correlation between calcium levels and the Na+ levels should yet be detected as the rate equilibrium of Na+ would be slower if a greater amount of Na+ is mobilized by a larger amount of calcium.

Materials and methods

Fourteen patients under chronic hemodialytic treatment three times a week for end-stage renal disease (ESRD) have received, by accident, unsoftened tap water in their dialytic process.

The water conductivity in samples derived from tap water has reached 483 mS/cm, while the treated water had a mean value of 15 mS/cm. The dialytic treatment was stopped after 60 minutes due to the emergence of severe headache, vomiting, hypertension, tachycardia and nausea, typical of the “Hard Water Syndrome” “Freeman RM – 1967 [3]“. Water samples for chemical surveillance have been collected from the tubes serving the dialysis: the problem in the water filters was recognized and subsequently solved. All patients subsequently fully recovered from the accident and no patient suffered any health problem due to it.

For comparison, the blood samples from 11 patients under regular dialysis have been taken 60 minutes after the start of the treatment. (Figure 2)

Hemodialysis was delivered using a Fresenius HD machine FX 5800 (Fresenius FMC, Germany) programmed to provide a dialysis flow rate of 500ml/min at a temperature of 36°C. Standard bicarbonate buffered dialysate concentrate (Fresenius 335) was used to yield a dialysis fluid containing the following concentrations: bicarbonate 32mmol/L, glucose 5.5mmol/L, calcium 1.5mmol/L, K+ 2.0 mmol/L, Na+ 142 mmol/L. Blood flow was set in the range 200-360 ml/min.

All patients had arterovenous fistulas and arterial blood samples were collected from the fistula at the end of the dialytic process. All patients received seleparin 3000-4000U as anticoagulation each dialysis session.

At the end of the 60-min dialytic process, the blood samples were collected to measure plasma sodium, potassium, calcium, urea and creatinine.

Statistical analysis

Multivariate ANOVA has been used to verify significant differences between the two groups. Multiple post-hoc have been performed to identify significant differences. Correlation analysis between the measured variables has been conducted to observe relationships among the serum ions. The rejection value was set at p<0.05.


As shown in the table (Figure 3), after 1hr of dialytic treatment, the control and hypercalcemic patients were comparable for urea, K+ and Na+ levels. Conversely, the most remarkable result was the presence, in all patients treated with hard water, of an increase by 45% of blood calcium levels (p <0.01, test t for non paired data). Moreover, hypercalcemic patients also showed a small increase (13%), but statistically significant, of creatinine levels (p=0.011, t-test for non-paired data).

A correlation analysis on the measured variables after dialysis is done. Specifically, calcemia was not correlated with potassium, creatinine and urea levels, whereas a positive correlation was noted with the Na levels (Pearson = 0.428, p=0.032). We also showed the linear regression analysis between calcium and Na+ (Figure 3).


  • In the absence of renal function, an extracellular increase of calcium induced by high calcium levels into the dialysate is accompanied by a trend toward increased plasma Na+ levels.
  • This observation went previously unnoticed because the strength of the relationship is quite weak (for 1mg/dl of change in calcemia, the natremia changes of only 0.27 mmol/L).
  • To observe this relationship two conditions must be met: 1) large modifications in the calcemia must occur to induce a quantifiable modification in the natremia 2) linear correlation methods should be used rather than classification of subjects in two or more calcemic groups.
  • This phenomenon might be explained by the buffering activity of the extracellular matrix in response to an increased extracellular calcium concentration, leading to the release of Na+ in exchange with calcium (Figure 4).
  • Speculatively, the hypercalcemia might foster sodium release from the extracellular matrix. This might represent an additional form of sodium homeostasis, which does not necessitate kidney intervention.