Key findings

The experimental setup of the study, with controlled fluid infusions and the repeated sampling of acid–base parameters, hemoglobin, and electrolytes, made it possible to reveal the relationship between the acid–base status and relative changes in PV in patients with DKA. The contribution of changes in the Na–Cl gap to metabolic acidosis upon admission appeared to be limited at first sight. Most patients even had an increased Na–Cl gap, which partly countered the acidosis caused by ketoacids. However, after correcting for hypovolemia, almost all patients had a decreased Na–Cl gap. Thus, hypovolemia masks significant deviations in the Na–Cl gap in DKA.

Clinical implications

Our findings suggest that the commonly described phenomenon of progressive hyperchloremic acidosis during the treatment of DKA may not be primarily caused by chloride-rich resuscitation fluids or by renal retention of chloride. Instead, a masked, low Na–Cl gap is already present initially and is then revealed by the restoration of the intravascular volume.

This phenomenon can be expected to be pronounced when plasma volume is expanded by infusion fluids with a Na–Cl gap of zero, irrespective of the absolute amount of chloride in the solution. According to the Stewart paradigm, an infusion of saline has the same effect on SID, BE, and pH as an infusion of only water because they have the same effect on the Na–Cl gap. In our study, an infusion of saline significantly decreased the Na–Cl gap, but its effect on BE was offset by the dilution of albumin, lactate, and OI.

A simple calculation supports our view that the surplus of chloride in 0.9% saline cannot cause acidosis without a concomitant change in extracellular fluid volume. In the present study, the initial Na and Cl concentrations were 132 and 93 mmol/L, respectively. A volume deficit of 25% indicates that the extracellular fluid volume would be approximately 11 L, and its content in terms of Na and Cl ions would then be 1452 and 1023 mmol, respectively. Using mass balance, the Na–Cl gap will remain at 39 mmol/L, regardless of how many equal amounts of Na and Cl are added, provided that the increased amounts are dispersed in 11 L of fluid.

The use of infusion fluids with a normal or even supraphysiological Na–Cl gap can be expected to mitigate or even compensate for the decrease in the Na–Cl gap caused by plasma volume expansion. This would, at least in theory, support the use of solutions such as Plasma-Lyte 148 (Baxter Inc., Springfield, Illinois, USA) for resuscitation in DKA, even in patients presenting with normal or even increased Na–Cl gaps. We used 1 L of 0.9% saline during the experiments, while Ringer’s acetate was used for resuscitation between the experimental infusions. This initial saline infusion was intended to counteract the high incidence of hyponatremia on admission. However, Ringer’s acetate has a SID of 30 which is still lower than plasma and this infusion fluid is, therefore, not able to correct a low Na–Cl gap.

This also explains why the Na–Cl gap at t3 (assuming normovolemia) was larger than the “virtual Na–Cl gap” at t1. The latter was calculated by hypothetically expanding the PV at t1 with water until it was equal to the PV at t3. This “virtual Na–Cl gap” then quantifies the impact of loss of plasma water per se on the Na–Cl gap or, in other words, the decrease of the Na–Cl gap that can be expected solely by normalizing plasma volume. However, the actual Na–Cl gap between Day 1 and Day 2 changed not only because of PV expansion, which decreased the Na–Cl gap, but also because a resuscitation fluid with a SID higher than 0 (Ringer’s acetate) mitigated the decrease of the gap. Surprisingly, the kidneys initially decreased the gap on Day 1 by excreting relatively more Na than Cl. The summary effect of these factors yielded the actual Na–Cl gap at t3 being larger than the “virtual Na–Cl gap” at t1.

The Stewart approach holds that pH is determined by SID and the plasma concentrations of weak acids and PaCO2. It provides a clear picture of the intimate relation between changes in serum electrolyte concentrations and changes in pH. This information is not provided by the traditionally used Henderson–Hasselbalch equation which, in our material, would show the reduced metabolic acidosis between Day 1 and 2 but not indicating that the acidosis would be worsened by a decreased Na–Cl gap. The Henderson–Hasselbalch equation is still valid but is, according to the Stewart approach, of minor importance since it only describes how pH and bicarbonate as dependent entities relate to changes in the independent entity PaCO2.


Most studies comparing the rate of correction of the acid–base balance in DKA have found small differences between the use of 0.9% saline and balanced crystalloid fluids for resuscitation, and many of these studies have been observational.

Van Zyl et al. reported faster normalization of pH with Ringer’s lactate as compared to saline in a randomized trial with 54 patients; however, the difference did not reach significance [8]. The time required for DKA to resolve was shorter with Plasma-Lyte 148 as compared to saline in a study with 66 pediatric patients, but here as well, the difference did not reach statistical significance [9]. Self et al. reported a speedier resolution of DKA with Ringer’s lactate or Plasma-Lyte 148 in a sub-study of a randomized trial involving 172 adult patients [10]. In a retrospective analysis, Oliver et al. found a faster increase in pH in 84 DKA patients who received fluid resuscitation with Plasma-Lyte instead of saline, but the time to resolution did not differ [11].

Two randomized trials comparing saline with Plasma-Lyte 148 in 45 and 90 patients showed faster normalization of acidosis [1, 12]. These findings, as well as a reduced length of stay in patients treated with Plasma-Lyte 148, were recently confirmed in a meta-analysis [13]. Unfortunately, specific data on the course of the Na–Cl gap are lacking in these reports, and a larger trial is needed to establish whether Plasma-Lyte 148 is beneficial in DKA.

Other contributors to the acid–base balance

The changes in acid–base balance between Day 1 and 2 are not only related to the composition of the resuscitation fluids used, but also internal factors, such as renal homeostasis. Moreover, metabolic changes affecting SID during treatment for DKA are not limited to changes in the Na–Cl gap. The concentrations of other strong ions, such as potassium and lactate, will also change, though to a lesser degree. Blood ketones (which are negative anions) were only measured on admission to the ICU but can be assumed to have decreased during treatment. This is reflected in the substantial decrease in other ions in the cohort, as well as in decreases in the strong ion gap (SIG), as reported by others [14, 15].

Most patients presented with elevated plasma albumin, although a few showed low values, which resulted in a low-normal average for the entire cohort. Hypoalbuminemia was common after the plasma volume had been restored, which at least temporarily contributes to the decrease in metabolic acidosis because albumin is a weak acid [16].


The number of included patients was small, and some patients could have received small amounts of fluids before admission to the ICU. Sodium was analyzed using a direct ion-selective electrode (ISE) method. However, this method was not available for chloride when the present study was performed, which forced us to rely on the indirect ISE method. This method involves dilution, whereby measurement errors can occur due to abnormal plasma concentrations of proteins and lipids [17]. This could also explain why our “normal” Na–Cl gap was slightly higher than those usually described. Blood gas analyzers generally report slightly higher chloride values as compared to standard electrolyte analyzers, resulting in smaller normal values for the Na–Cl gap [18].

The use of absolute numbers instead of ranges for normal values for electrolytes and albumin can be questioned. Electrolyte concentrations can fluctuate slightly around the normal value without being considered pathological. However, the actual normal range for SID (and, thus, its principal component, the Na–Cl gap) is much smaller than theoretically possible based on the normal ranges of sodium and chloride [19]. Further, according to the Stewart paradigm, every mmol/L increase or decrease in the Na–Cl gap will cause a concomitant change in BE of 1 mmol/L.

We are aware that the equation used for the calculation of OI that is described by Story [7] is an approximation. For example, this equation does not account for the fact that the albumin charge slightly changes with pH [20]. However, we have no reason to suspect that these limitations significantly affect our findings.

Our calculations assume that erythrocyte volume is constant throughout the experiment and that the plasma volume at t3 reflects intravascular normovolemia. We did not measure total urinary excretion of electrolytes such as sodium and chloride, which could have provided more information on the renal contribution of the acid–base status.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit


This article is autogenerated using RSS feeds and has not been created or edited by OA JF.

Click here for Source link (