Does Fluid Choice Make Any Difference in DKA?

Setting the Scene

Your patient is a 21 year-old female with a history of type 1 diabetes mellitus who was brought to the ED by her boyfriend for diminished responsiveness. In a stupor, she is unable to give any history. Her vitals are: BP 102/66, pulse 120, respiratory rate 24, temperature 98.9 oral, and O2 saturation 98% on room air. Her finger stick glucose is >500 mg/dl. She looks dry and is somnolent (GCS 9). Pupils are equal, round, and reactive. Neck is supple. She is protecting her airway well, her lungs are clear, and you hear no murmurs. Her belly is soft, and you see no signs of trauma or exanthema. Her skin tents when you pinch it. She is moving all extremities in response to noxious stimulus. As the rest of her labs (including serum osmolality and cultures, of course!) are sent off, her boyfriend tells you that she has not been taking her medications over the past 2 weeks and has had symptoms consistent with polydipsia and polyuria most noticeably over the past few days. A rapid shock panel returns with a glucose level of >500 mg/dl, pH 7.2, bicarbonate 10, and a urine dipstick shows large ketones. These confirm your suspicion of diabetic ketoacidosis (DKA). You wait for further results to decide whether a full sepsis work-up and antibiotics are necessary. In the meantime, you look at the bag of normal saline (0.9% saline solution) that is already hanging and you wonder, “Am I sure this is really the best solution to resuscitate a patient with DKA?”


Consensus for Resuscitation in DKA

Diabetic ketoacidosis is one of the diseases for which emergency physicians are expected to have a plan to quickly put into action. The basics should be familiar: Manage the patient’s ABCs, place an IV, put the patient on a monitor to check vitals frequently, and start with an intravenous fluid bolus. There are nuances beyond what is described here, of course (frequent glucose monitoring, adding dextrose once the patient’s blood glucose reaches 200-250 mg/dl, and the like). But our focus will be on that critical first intervention: The choice of IV fluid.

There is a consensus among endocrinologists and emergentologists that the first and most critical intervention in DKA is fluid resuscitation. Different guidelines point to normal saline (0.9% saline solution, aka NS) as the initial fluid of choice. [1, 2, 3, 4] Hydration should initially proceed rapidly at a rate of 1-1.5 liters over the first hour to support hemodynamic functions (e.g., kidney and brain perfusion). The rest of the fluid is meant to replace what was lost from the intracellular and interstitial compartments and should be given over the next 24 hours. [1, 2, 3, 4] Lately, some researchers have sought to compare balanced saline solutions (Plasma-Lyte, lactated Ringer’s, and others) to normal saline in DKA, because there is no data showing that NS is truly the best choice. This is an especially salient issue if you are managing DKA patients who are boarding in the Emergency Department. We will not be discussing colloids and hypertonic solution. Generally, they should not be used in DKA since these fluids would worsen hyperosmolarity, one of the main issues related to DKA. [4 , 5, 6]

What is the Strong Ion Difference (SID)?

One of the keys to understanding acid-base balance in any patient and also why balanced solutions such as Plasma-Lyte and Ringer’s lactate may be worth reaching for in DKA, is a concept called the strong ion difference (SID) or the strong ion gap (SIG). It is presented in a simplified form here. This concept is derived from the work of Peter Stewart and is predicated upon the idea that in human body fluids, quantitative acid-base balance (pH) is determined mainly by three independent variables: Partial gas pressure of CO2, the concentration of weak acids (mostly albumin and phosphate), and the net strong ion charge. [7] Strong ion charge is defined as the difference between completely dissociated cations such as sodium and potassium, and their anionic counterpart, chloride. [8]


SID = (Na+ + K+ + Ca2+ + Mg2+) – (Cl – other strong anions)


The main ideas to grasp from this are 1) when the SID decreases, it also pulls the pH down toward acidosis, and 2) when the SID increases, it also causes the pH to shift toward alkalosis. Knowing the SID of the infusing solution rounds out what we need to know. The SID of the solution will pull the SID of the body fluid (i.e., plasma) towards its own. Normally the human body has a SID of about 38, which coincides with the amount of HCO3 and other weak anions needed to keep a normal pH. Normal saline has a SID of zero (Na and Cl are of equal charge and concentration). When it is infused, the SID is pulled toward zero (as discussed before, this tends toward acidosis). In contrast, Ringer’s solution has a SID of 24, which is closer to normal plasma SID. Therefore, the tendency of plasma toward acidemia is less. (See table below. Please note that it incorrectly lists 154 mEq/L of K instead of Cl in NS) [8] Armed with this basic knowledge of SID, we see how NS can cause hyperchloremic acidosis.


Got that IV Flow

From a physiologic standpoint, the concern is that during resuscitation, one will be trading acidosis from excess ketones for acidosis from excess chloride. [9] In a small, multi-center, randomized study [10] on ICU patients admitted from the ED Chua et al gave either PL or NS for the first 12 hours and found that patients resuscitated with PL had faster resolution of metabolic acidosis (primary outcome) and less hyperchloremia during these 12 hours. This was a very small study (n = 23) and it was not known which fluids the subjects received prior to ICU admission and after the 12-hour study period, but at the end of the study period, the only statistically significant difference was a greater change in chloride in the NS group. Although, there was significant room for confounding in this study, this finding of hyperchloremia with NS reflects those of other studies.

For instance, a separate small, single-center, blinded, randomized study compared NS to Plasma-Lyte (PL). Subjects received an infusion of each study fluid with an insulin drip until their glucose reached 250, and then dextrose was added to their study fluid. Their blood chemistry was followed until their gap closed to 12. Those in the PL arm had significantly lower mean chloride and higher mean bicarbonate levels at the end of the study – hyperchloremic acidosis was not observed. [9] We will explore why avoiding hyperchloremia may be important for your patient later in this post.

Research has also focused on Lactated Ringer’s (LR) solution compared to NS in DKA. Van Zyl et al. [11] conducted a small, randomized, double-blinded comparison of these two fluids starting from the ED and extending until subjects reached a glucose level ≤14 mmol/l (~250 mg/dl). Thereafter, patients were managed according to the attending physician preference, but their electrolytes were followed until clinical resolution of DKA occurred per American Diabetes Association criteria (2006). This was already a small study by design, but slow recruitment and expiry of consumables led to the study being stopped early and while underpowered. Notwithstanding, their analysis showed the LR group took a significantly longer time to reach the target glucose level and also had a higher number of subjects with potassium >5.2 mmol/l after one hour of fluid resuscitation (the observed hyperkalemia was not sustained beyond the one hour period). Although not statistically significant, their findings also showed a trend toward normalization of pH and bicarbonate favoring LR. [11]


The finding of elevated glucose is consistent with the physiology of lactate metabolism (also known as Cori Cycle).



The majority of lactate in LR is metabolized in the liver through gluconeogenesis with consumption of hydrogen ions. The rest is oxidized with the consumption of more hydrogen ions resulting in the creation of CO2 and H2O. The net effects from these processes would be an increase in glucose and an improvement in acidosis due to the generation of bicarbonate from the H2O and CO2. Recall that the Van Zyl study showed an increased time to control hyperglycemia in the LR group. However, a study by O’Malley et al. on subjects undergoing renal transplant compared NS to LR and showed less hyperkalemia and fewer episodes of metabolic acidosis in subjects on LR. [13] Although, these are not DKA patients, large volumes of fluids were infused (mean volumes NS and LR were 6.1 +/-1.2 and 5.6 +/-1.4, respectively – a non-significant difference), and this may provide a model for large volume resuscitation in patients with some underlying level of kidney injury. These volumes are commonly infused when treating hyperglycemic crises.

Aside from the transient increase in potassium in the Van Zyl study, the risk for hyperkalemia with LR is not reported by the other studies reviewed here. On the other hand, hyperkalemia was documented in the O’Malley study within the NS group. Some theorize that hyperchloremic acidosis may cause potassium elevations. [1, 13, 14, 15] In addition, a delay in the correction of hypernatremia and hyperosmolality are also a concern when using NS due to its higher normal plasma concentration of sodium. [14] Of these physiologic parameters the hyperkalemia is most important to avoid as it can lead to dysrhythmias and quickly change the clinical course from something routine to a true crisis.

What About Acetated Ringer’s?

Although research on acetated Ringer’s solution is scant, other solutions such as Plasma-Lyte, used in the Chua study, has 27 mmol of acetate. [(See table above.)] Cardiotoxicity is a concern with infusion of acetate. But this occurs at higher concentrations than are available in IV solutions and at much faster infusion rates than would be used to resuscitate DKA patients.  [16, 17]


Hyperchloremia and the Kidney

A final concern about NS is that the high chloride doses have been associated with renal dysfunction. Some have shown excess chloride is an independent predictor of acute kidney injury. [18] We always want to avoid iatrogenic harm especially in something that is supposed to be beneficial or at least mostly harmless, such as fluid resuscitation. When the possible outcome is something like need for renal replacement therapy (RRT), we want to be even more fastidious in our choice of fluid. A prospective, open-label, sequential pilot study was done on consecutively admitted patients to a single ICU over two 6-month periods separated by a 6-month wash-out period. During the first 6-month study period, fluids were given according to physician preference. During the second 6-month study period only chloride restricted fluids (i.e. balanced saline solutions and chloride-poor albumin) were allowed. Primary outcomes included incidence of acute kidney injury (AKI) and secondary post-hoc analyses included requirement of RRT. In both of these measures, there were significant differences favoring use of fluids with less chloride. Time to kidney injury was longer, incidence of kidney injury was 6% lower, and fewer required RRT while in the ICU. The Kaplan-Meier curves for these two outcomes show separation even within the first day in the ICU. While there were no long-term differences in these outcomes between the groups, even short-term avoidance of AKI and need for RRT is desirable. The main limitation the authors concede is that their study was not blinded, but they attempt to explain how they believe their study may have mitigated some of the inherent bias. [19]

There is data that fails to replicate these findings. Specifically, the SPLIT trial was a randomized, blinded, multi-center, double crossover study conducted in 4 medical and surgical ICUs in New Zealand that compared NS to PL. The primary outcome was the proportion of patients who had AKI at the end of the 90-day study period. Study outcomes were measured in five pre-defined subgroups that included sepsis, surgical procedures, and an APACHE II score greater than or equal to 25. No difference in rate of AKI was found between subjects receiving NS or PL. The study had a number of shortcomings that warrant pointing out. First, about 90% of subjects were exposed to PL before being enrolled in the study (which only infused about 2 L of fluid per group), and they may have received different fluids outside the ICU during the study period. Second, mean age was 60, and 2/3 of subjects were male, which makes this data less applicable to a broader population. Third, data on the primary outcome was also missing for 7.5% of subjects To overcome this, two analyses were conducted: One assuming all these subjects got AKI, the other assuming they did not. These did not significantly alter primary outcome results, which showed no difference between the two groups. Third, and kind of an interesting caveat, about half of doctors who participated in the study answered a survey that asked them to guess which blinded study fluid was which. Sixty-three percent guessed right. The authors note that they did not see any significant difference in co-intervention despite this flaw, but it is uncertain in what way this may have affected patient treatment by clinicians. Last, and arguably most important, no sample size calculations were performed. This means we can’t tell what implicit bias might be present simply due to this design. [20] Considering that there was no difference between the two groups and that other studies may show harm with NS, the benefit of balanced solutions remains a possibility.


What is the cost difference of these fluids?

Regional and systems differences exist in price and cost of these fluids, and these may be factors in fluid choice at your institution. However, the most important question of all leads back to the patient. What is the cost to the patient in terms of outcomes? Is any of this clinically significant? I argue that it is. Too much chloride seems to be associated with some level of kidney injury. Even just an excessive change from baseline chloride may cause harm. Clinically, AKI opens the patient up to further risks, such as electrolyte derangements, especially considering that diabetics may already have underlying kidney dysfunction. More concerning are the findings that too much chloride may lead to RRT. Although these were ICU patients, separation between groups started within the first 24 hours from admission on the Kaplan-Meier curve. [19] Hyperchloremic acidosis may also accompany the patient’s progression out of DKA. Although this is a physiologic parameter, it has clinical implications in that significant acidosis by any mechanism carries consequences (e.g., cardiac depression) that may predispose patients to worse outcomes. [14] Surprisingly, the risk of hyperkalemia was seen not only with a balanced salt solution, but also with NS, perhaps resulting from the hyperchloremic acidosis. [11, 13] Finally, although an extended period to reach a blood glucose level of 250 mg/dl may also be undesirable in DKA due to cost and bed space, this delay is not likely to be clinically harmful. Therefore, my choice for initial fluid resuscitation in DKA is a balanced saline solution, such as lactated Ringer’s. There is not enough data to suggest this over Plasma-Lyte or other solutions in this vein (pun sincerely intended), but they seem physiologically advantageous without an undue financial burden. Hopefully in the near future, we will have a study that helps answer some of the clinical questions related to this topic of resuscitation fluids with data geared toward measuring the clinical effects of our choice of therapy. [21]




  1. Kitabchi AE, Umpierrez GE, Miles JM, Fisher JN. Hyperglycemic Crises in Adult Patients With Diabetes. Diabetes Care. 2009;32(7):1335-1343.
  1. Savage MW, Kilvert A. ABCD guidelines for the management of hyperglycaemic emergencies in adults. Practical Diabetes Int. 2006;23(5):227-231.
  1. Savage MW, Dhatariya KK, Kilvert A, Rayman G, Rees JA, Courtney CH, Hilton L, Dyer PH, Hamersley MS; Joint British Diabetes Societies.. Joint British Diabetes Societies guideline for the management of diabetic ketoacidosis. Diabet Med. 2011 May;28(5):508-15. PubMed PMID: 21255074.
  1. Nyenwe EA, Kitabchi AE. Evidence-based management of hyperglycemic emergencies in diabetes mellitus. Diabetes Res Clin Pract. 2011 Dec;94(3):340-51. PubMed PMID: 21978840.
  2. Perel P, Roberts I, Ker K. Colloids versus crystalloids for fluid resuscitation in critically ill patients. Cochrane Database Syst Rev. 2013 Feb 28;(2):CD000567. PubMed PMID: 23450531.
  3. Finfer S, Bellomo R, Boyce N, French J, Myburgh J, Norton R; SAFE Study Investigators.. A comparison of albumin and saline for fluid resuscitation in the intensive care unit. N Engl J Med. 2004 May 27;350(22):2247-56. PubMed PMID: 15163774
  4. Stewart PA. How to understand acid-base A quantitative acid-base primer for biology and medicine. New York: Elsevier; 1981.
  5. Langer T, Ferrari M, Zazzeron L, Gattinoni L, Caironi P. Effects of intravenous solutions on acid-base equilibrium: from crystalloids to colloids and blood components. Anaesthesiol Intensive Ther. 2014 Nov-Dec;46(5):350-60. PubMed PMID: 25432555.
  6. Mahler SA, Conrad SA, Wang H, Arnold TC. Resuscitation with balanced electrolyte solution prevents hyperchloremic metabolic acidosis in patients with diabetic ketoacidosis. Am J Emerg Med. 2011 Jul;29(6):670-4. PubMed PMID: 20825879.
  7. Chua HR, Venkatesh B, Stachowski E, Schneider AG, Perkins K, Ladanyi S, Kruger P, Bellomo R. Plasma-Lyte 148 vs 0.9% saline for fluid resuscitation in diabetic ketoacidosis. J Crit Care. 2012 Apr;27(2):138-45. PubMed PMID: 22440386.
  8. Van Zyl DG, Rheeder P, Delport E. Fluid management in diabetic-acidosis–Ringer’s lactate versus normal saline: a randomized controlled trial. QJM. 2012 Apr;105(4):337-43. PubMed PMID: 22109683.
  9. Rodwell VW, Bender DA, Botham KM, Kennelly PJ, Weil P. Harper’s Illustrated Biochemistry, 30e; 2015 Available at: Accessed: April 17, 2017.
  10. O’Malley CM, Frumento RJ, Hardy MA, Benvenisty AI, Brentjens TE, Mercer JS, Bennett-Guerrero E. A randomized, double-blind comparison of lactated Ringer’s solution and 0.9% NaCl during renal transplantation. Anesth Analg. 2005 May;100(5):1518-24, table of contents. PubMed PMID: 15845718.
  11. Cardoso L, Vicente N, Rodrigues D, Gomes L, Carrilho F. Controversies in the management of hyperglycaemic emergencies in adults with diabetes. Metabolism. 2017 Mar;68:43-54. PubMed PMID: 28183452.
  12. Jivan D. Management of diabetic ketoacidosis. Journal of Endocrinology, Metabolism and Diabetes of South Africa. 2014 Aug 14:16(1):10-14.
  13. Myburgh JA, Mythen MG. Resuscitation fluids. N Engl J Med. 2013 Sep 26;369(13):1243-51. PubMed PMID: 24066745.
  14. Neavyn MJ, Boyer EW, Bird SB, Babu KM. Sodium Acetate as a Replacement for Sodium Bicarbonate in Medical Toxicology: a Review. Journal of Medical Toxicology. 2013;9(3):250-254.
  15. Suetrong B, Pisitsak C, Boyd JH, Russell JA, Walley KR. Hyperchloremia and moderate increase in serum chloride are associated with acute kidney injury insevere sepsis and septic shock patients. Crit Care. 2016 Oct 6;20(1):315. PubMed PMID: 27716310; PubMed Central PMCID: PMC5053142.
  16. Yunos NM, Bellomo R, Hegarty C, Story D, Ho L, Bailey M. Association between a chloride-liberal vs chloride-restrictive intravenous fluid administration strategy and kidney injury in critically ill adults. JAMA. 2012 Oct;20(1):315. PubMed PMID: 27716310; PubMed Central PMCID: PMC5053142.
  17. Young P, Bailey M, Beasley R, Henderson S, Mackle D, McArthur C, McGuinness S, Mehrtens J, Myburgh J, Psirides A, Reddy S, Bellomo R; SPLIT Investigators.; ANZICS CTG.. Effect of a Buffered Crystalloid Solution vs Saline on Acute Kidney Injury Among Patients in the Intensive Care Unit: The SPLIT Randomized Clinical Trial. JAMA. 2015 Oct 27;314(16):1701-10. Erratum in: JAMA. 2015 Dec 15;314(23):2570. PubMed PMID: 26444692.
  18. Self WH, Semler MW, Wanderer JP, Ehrenfeld JM, Byrne DW, Wang L, Atchison L, Felbinger M, Jones ID, Russ S, Shaw AD, Bernard GR, Rice TW. Saline versus balanced crystalloids for intravenous fluid therapy in the emergency department: study protocol for a cluster-randomized, multiple-crossover trial. Trials. 2017 Apr 13;18(1):178. PubMed PMID: 28407811; PubMed Central PMCID: PMC5390477.










The following two tabs change content below.

Raul Hernandez

PGY III EM Resident

Leave a Reply

Your email address will not be published.