Diabetic ketoacidosis (DKA) is an acute, major, life-threatening complication of diabetes characterized by hyperglycemia, ketoacidosis, and ketonuria.1 DKA can lead to intensive care unit (ICU) admission due to the major hydroelectrolytic disorders.2 Correction of the underlying cause, rehydration with intravenous fluids, insulin therapy, and electrolyte replacement are the cornerstones of DKA management.1 Although there is widespread agreement about replacing insulin and electrolytes,3–5 it is still unclear which fluid is best for managing DKA.

For several decades, the most widely used crystalloid solution was normal saline (0.9% sodium chloride solution).6,7 Furthermore, current international guidelines suggest that 0.9% sodium chloride (0.9% SC) should be the preferred as replacement fluid for DKA.8 However, recent research raises questions about the possible negative effects of sodium chloride administration, mainly the increased acidemia by hyperchloremia.9 Hence, balanced crystalloids, which replace chloride anions with lactate or acetate, are becoming increasingly popular as alternatives to 0.9% SC because of the potential negative effects of its high chloride content.10 These substitutes are more similar in chemical composition to human plasma than to sodium chloride. Reduced chloride levels and increased strong ion differential in vivo are implicated in this compositional alignment.11

The debate between SC and RL in the management of DKA is far from being resolved; as indicated in several recent reports.12–16 The recent meta-analysis that encompassed 11 trials (including one in pediatric population) not found a significant difference between balanced crystalloids and normal saline for the time to resolution of DKA, major adverse kidney events, and incidence of hypokalemia.12 However, there was a significant reduction in the post-resuscitation chloride among patients received balanced crystalloids.12 While both crystalloids are safe for treating DKA, balanced solutions were associated with improved post-resuscitation electrolyte balance and preventing hyperchloremic metabolic acidosis according to the recent meta-analysis of Gupta et al.13 Other meta-analysis indicates that the use of balanced solutions resolves DKA faster than 0.9% SC with a mean difference of 5 h,14 reduced post resuscitation hyperchloremia and hypernatremia, and increased bicarbonate level.14,15 The prospective intervention trial aiming to compare the effectiveness of Sterofundin as balanced solution versus NS in the management of DKA showed a superiority of sterofundin (shorter mean time to DKA resolution, less total intravenous fluid, less parenteral insulin, and shorter hospital stays).16

We hypothesized that RL may result in a faster DKA resolution and less electrolyte imbalances, namely chloride, potassium, sodium, and bicarbonate. Therefore, clinical guidelines could be directed toward how to treat DKA with RL instead of NS. Herein, we sought to evaluate the effect of RL comparatively to 0.9% SC on the resolution of severe DKA as well as on the electrolytes variations.

Our present trial including a double cohort of 88 ICU patients with similar baseline characteristicsdid not reach statistical significancein theresolution of DKA when treated with RL compared with 0.9% SC as fluid therapy. Nevertheless, RL administration was superior in terms of potassium level which was better at H24 and H48 and a lower chloride level at H48 suggesting better control of acidemia. The subgroups analysis according to age, diabetes kind, DKA severity, and SAPS II not found any difference.

Fluid resuscitation is a crucial step in managing DKA because hyperglycemia induces osmotic diuresis, resulting in significant fluid and electrolyte losses. Fluid resuscitation aims to reverse these detrimental effects with replenishing intravascular volume and improving tissue perfusion and oxygen delivery. It also dilutes the concentration of glucose and ketones in the blood, helping to correct hyperglycemia, acidosis and electrolyte disorders.19 Initially, isotonic saline is favored to reestablish circulation but expose to the risk of hyperchloremia associated with cerebral edema.19 The comparison between balanced solutions and serum chloride in terms of efficacy in treating DKA was subject to multiple analyses and incorporating various populations and fixing a several outcome measures.

Our results adhere to those of several studies in this topic. There was no significant difference between balanced crystalloids and normal saline group for the time to resolution of DKA (MD = −1.49, 95%CI: −4.29 to 1.31, p = 0.30, I2 = 65%).12 Similarly, Gupta et al. in their recent updated systematic review and meta-analysis didn’t observe a significant difference between balanced electrolyte solution and normal saline in the time to DKA resolution (MD: −1.63; 95% CI: −7.66 to 4.41; p = 0.60) or length of hospital stay (mean difference (MD): −0.07; 95% CI: −0.44 to 0.31; p = 0.73).13 Tamzil et al.15 not reported, also, in their meta-analysis (8 RCTs, 595 patients) any difference in the duration of DKA resolution. (MD: −4.73, 95%CI −2.72 to 4.92; I2 = 92%; P = 0.180). The time to reach overall DKA endpoints was comparable in both groups in the systematic review of clinical trials of Jahangir et al.20 The absence of effect on insulin intake and mortality was also reported in the meta-analysis of Szabó et al.14 (MD: 0.16 [−3.03, 3.35] h) and (OR: −0.67 [0.12, 3.68] respectively). No difference was also demonstrated on the ICU stay or mortality in a second report.15

Conversely, another relevant meta-analysis (7 RCTs and 3 observational studies with 1006 participants) showed superiority for balanced solutions which resolved DKA faster than 0.9% saline with a mean difference (MD) of −5.36 [95% CI: −10.46, −0.26] h.14 Considering the evidence from pooled small randomized trials with moderate overall certainty of evidence, the use of balanced electrolytes solutions in DKA was associated with faster rates of DKA resolution compared to SC in the systematic review of Catahay et al.21 that showed a pooled hazard ratio at 1.46 [1.10–1.94] (p = 0.009) with 12% heterogeneity while MD was −3.02 (95% CI −6.78 to 0.74; p = 0.12) with heterogeneity of 85%.

Concerning the change in key biomarkers, our main findings are summarized in the better serum potassium levels, less hyperchloremia and better bicarbonate levels at H12 without risking hyperlactatemia with the administration of RL. Our findings go in the trends of the main recent meta-analyses.12–15 There was a significant reduction in the post-resuscitation chloride (MD −3.16, 95%CI −5.82 to −0.49, p = 0.02, I2 = 73%) among patients received balanced crystalloids.12 Balanced solutions resulted in significantly higher post-resuscitation bicarbonate levels (MD: 1.63; 95% CI: 0.86–2.39; p < 0.001) and lower post-resuscitation chloride levels (MD: −2.37; 95 % CI: −3.56 to −1.19; p < 0.001).13 Post-resuscitation chloride (MD: −4.26 [−6.97, −1.54] mmol/l) and sodium (MD: −1.38 [−2.14, −0.62] mmol/l) levels were significantly lower. In contrast, levels of post-resuscitation bicarbonate (MD: 1.82 [0.75, 2.89] mmol/l) were significantly elevated in the balanced solutions-group compared to the saline-group.14 There was a significantly lower post resuscitation chloride concentration in the balanced electrolytes solutions (MD = 2.96, 95%CI −4.86 to −1.06; I2 = 59 %; p = 0.002) and a reduction in duration for normalization of bicarbonate (MD = 3.11, 95% CI −3.98 to 2.23; I2 = 5%; p = 0.0004).15 Aditianingsih et al.22 in their randomized, single blind controlled trial showed that fluid resuscitation of DKA patients with RL resulted in slightly but not significantly higher baseexcess and anion gap than normal saline. Thus, balanced electrolyte solution as an alternative fluid resuscitation may prevents hyperchloremic acidosis in diabetic ketoacidosis patients. In our series, despite pH didn’t differ, chloremia was reduced by 8 mmol/l at average after 48 H of managing DKA with RL.

The effect of balanced crystalloids was more beneficial in the largest American trial designed as secondary analysis of cluster trials (n = 172, 94 were assigned to balanced crystalloids and 78 to saline).23 Indeed, the time to DKA resolution in the balanced crystalloids group was shorter by 4 h on average.23 Also, the same investigators revealed shorter time to insulin infusion discontinuation in the balanced crystalloids group (9.8 versus 13.4 h as medians; p = 0.03).23 The Phase 2 trial of Ramanan et al.18 found that plasmalyte (PL) resulted in faster resolution of metabolic acidosis without increasing ketosis (at 24H, DKA resolution occurred in 69% (PL) and 36% (SC) of patients; OR = 4.24, 95% CI [1.68–10.72], p = 0.002). Also, the length of ICU stay was slightly shorter for the PL group (49 h vs 55 h).18 The authors recommend confirming this hypothesis in a larger phase 3 trial.

In our series, the taken fluid volume and the cumulative received dose of insulin within 48 h (IU) were less in the RL group but without reaching statistical significance (respectively 6200 ml vs 6440 ml, p = 0.36 and 205 vs 225 IU, p = 0.2). In the Emergency department with a different primary outcome which was the rate of ICU admission, Attokaran et al.24 showed a higher rate in the SC group (50% vs 39.1%). However,this differencewas not found after adjustment for pH at presentation and diabetes type (p = 0.71).

Additionally, we found that serum potassium was higher with RL without exceeding the standards while it was low with SC at H24 and H48. Conversely, in the above cited meta-analysis of Liu et al.,12 balanced crystalloids as compared to normal saline has no effect on the incidence of hypokalemia. This result can be explained by the type of balance crystalloid used, and therefore the potassium content, which differed from one trial to another.We indicate here that the RL we used provided 0.40 g per 1000 ml of solution.

The result we showed about the better bicarbonate level at H12 in the RL group (15.4 vs 12.9 mmol/l, p < 0.005) was similar to the result of the Australian multicenter retrospective examination of patients admitted to ICU for DKA.25 In this study, PL group experienced faster metabolic acidosis resolution (median serum bicarbonate correction was higher in the PL vs NS groups at 4–6 h (8.4 vs 1.7 mmol/l) and 6–12 h (12.8 vs 6.2 mmol/l) from baseline (p < 0.05), reduced hyperchloremia, and improved blood pressure and urine output.25 The authors also reported a result that has not been highlighted in our trial concerning the change in base excess (median standard base excess improved by 10.5 vs 4.2 meq/l at 4–6 h and by 16.0 vs 9.1 meq/l at 6–12 h in the PL and NS groups, respectively (p < 0.05).25

Our results about the possible favorable effect on preventing hyperchloremia and hypokalemia should be of particular interest given the harm of these disorders as found in the meta-analysis of Maharjan et al.26 In fact, hyperchloremia was associated with longer in-hospital length of stay and longer time to resolution of DKA and potassium replacement at < 10 mmol/l was associated with higher mortality.26 Hyperchloremic acidosis also has important clinical implications. Indeed, the compensation of acidosis by tachypnea can lead to respiratory muscle fatigue and, if not appropriatelymanaged, respiratory arrest. That include also the risk of arrhythmias, which can increase mortality if left untreated. A pH below 7.20 can increase the risk of ventricular fibrillation and heart failure by impairing myocardial contraction. This last complication is strengthened by associated hypokalemia as potassium plays a crucial role in regulating cardiac electrical activity. Hypokalemia, grafted on hyperchloremic acidosis, can lead to cardiac membrane potential alterations and repolarization delay, predisposing to atrial fibrillation, and potentially life-threatening ventricular tachycardia or fibrillation. Regarding the topic of diabetic decompensation, hypokalemia can reduce insulin receptor sensitivity, leading to insulin resistance and potentially contributing to the delay in correcting hyperglycemia.27 This is what is being tried to prevent, among other things, by balanced electrolyte solutions.

Finally, we can conclude that in adult patients with DKA, balanced solutions may be superior in preventing hyperchloremia and hypokaliemia compared to 0.9% SC. The clinical recommendation that can be raised is as follows: the use of balanced electrolyte solutions should be considered as the preferred resuscitation fluid in the management of DKA given their lower predisposition to cause hyperchloremic metabolic acidosis and hypokalimia. Nevertheless, we estimated that the unavailability of serum ketone assay, the blinding deficiency (no concealment of group assignment), the restricted sample size, and the absence of renal function assessment were the weaknesses in this trial. Thus, we propose future research directions to address these identified limitations and to reach statistical significance for the hypothesis.

Ahlem Trifi: Conceptualization, Methodology, Analysis, Software, Writing – original draft, Supervision, Software, Validation. Ikram Ben Braik: Data curation, Investigation. Hounaida Galai: Data curation, Investigation. Noussair Azzouz: Data curation, Investigation. Badis Tlili: Data curation, Investigation. Asma Mahdi: Visualization, Investigation. Linda Messaoud: Data curation. Eya Seghir: Visualization, Investigation. Asma Ouhibi: Data curation. Sami Abdellatif: Writing – review & editing, Validation.

During the preparation of this work the author(s) used ChatGPT in order to translate the abstract into Spanish. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.