Article, Endocrinology

Looking at diabetic ketoacidosis through electrocardiogram window!

a b s t r a c t

Background: Initial serum potassium (K+) in Diabetic ketoacidosis often does not reflect the true amount of total body K+ storage, and it is not a good predictor of subsequent hypokalemia. In this study, we tested the hypoth- esis that a deficiency of the total body K+ storage can be detected initially on surface Electrocardiography . Method: Medical records of 350 patients with a diagnosis of DKA were reviewed. Data regarding serial basic metabolic panels, arterial blood gases, serum ketones, and total K+ replacement that patient received during admission were collected. We compared biochemical findings for patients with and without QTU corrected (QTUc) prolongation by using the t test. Patients who were taking medications known to affect QTUc or cause ST-T changes were excluded. Results: After exclusion criteria, 61 patients were enrolled in this study. In 38 patients (62.9%), QTUc was more than or equal to 450 milliseconds. Patients with prolonged QTc received statistically more K+ supplementation during ad- mission (P = .014). They also had lower serum K+ level during their hospital course (P = .002) compared to patients with normal QTUc intervals. No significant difference was found between initial serum K+, calcium, glucose, anion gap, acidosis, age, or heart rate between these 2 groups.

Conclusion: The significant relationship between K+ depletion and the ECG changes observed in this study deserves further consideration. Our findings confirm the concept that the ECG is an easy and reliable tool for early diagnosis of hypokalemia in patients with DKA.

(C) 2015


Diabetic ketoacidosis (DKA) remains one of the most serious acute met- abolic complications of diabetes mellitus with a mortality rate of 2% to 10% [1,2]. Deficits of total body K+ as great as 300 to 600 milliequivalents can be seen in DKA. Patients with DKA usually present with a serum K+ level at or above the upper limits of normal. However, this initial serum level is a poor predictor of K+ depletion in these patients [3-5].

The association between hypokalemia and a prolongation of the QT interval raises the question of whether electrocardiographic changes reflect a low total K+ storage. In this study, we tested the hy- pothesis that QT prolongation can be the first clue of a depleted total body K+ and a good guide for emergency physicians initiating electro- lyte management of patient presenting with DKA.


The study was approved by the institutional review board of the par- ticipating institution. Patients were eligible for this study if they had an

* Corresponding author at: Emergency Department, Metropolitan Hospital, New York Medical College, 1901 1st Avenue, New York, NY 10029. Tel.: +1 212 423 6262.

E-mail address: [email protected] (G.W. Hassen).

admitting diagnosis of diabetes mellitus complicated by DKA. Diabetic ketoacidosis was defined as a Serum glucose level of greater than 250 mg/dL, venous pH of less than 7.25, and/or a serum bicarbonate level of less than 15 mEq/L and a positive test for urine ketones [6]. Patients were ineligible if they were taking medications or urine was positive for drugs known to affect the QT interval. In addition, we excluded pa- tients with a wide complex QRS, abnormal ST segment, or T-wave changes. We tried to eliminate all confounding factors that could cause QTc prolongation or interfere in our measurement of QTc. The purpose following the strict inclusion criteria was to have homogenous patient pool with reliable and complete data.

Patients were treated according to a standardized DKA protocol [7]. Basic metabolic panel (BMP), ECG, urine ketone, and a venous blood gas were obtained before initiating of treatment. Every 2 hours, the basic metabolic panel was repeated and patients’ hypokalemic status was treated with K+ supplements. The total K+ supplementation received during the admission was calculated. Initial anion gap was calculated [AG = Na – (Cl + HCO3)]. A standard 12-lead ECG was recorded at 25 mm/s. The QT or QTU was calculated manually by a cardiologist and corrected by Bazett formula, who was blinded to the patients’ bio- chemical data. The QTU corrected (QTUc) or QTc interval was measured from the onset of the QRS complex to the end of the U wave or T wave.

0735-6757/(C) 2015

264 S. Talebi et al. / American Journal of Emergency Medicine 34 (2016) 263265


Average of initial biochemical findings, the lowest serum K+ level and total K+ supple- ment during admission and compression of data in patients with and without QTUc pro- longation by using the t test

QTc or QTUc >= 450


QTc or QTUc b450


t value


Patient number









Age (y)





QTUc or QTc





Potassium (mmol/L)





Lowest potassium (mmol/L)





Bicarbonate (mmol/L)





Magnesium (mg/dL)





Calcium (mg/dL)





Anion gap










Heart rate (beat/min)





Glucose (mg/dL)





Potassium supplement (mEq)





Hospital stay (d)





Bold values indicate significance at P Value, b 0.05.

The end of the QT or QTU interval was defined as the intersection of a tan- gent to the steepest downslope of the dominant repolarization wave with the isoelectric line. When T and U waves were fused, the U component was included (QTU) for measurement purposes, if a discrete U wave was seen after the T wave it was excluded from measurement (QTc). Three separate measurements by the same cardiologist were obtained for each ECG, and the mean of these measurements was used as the value. Lead II, V2, or V3 was preferentially used for our measurements. If

a T or U wave could not be clearly discerned in these leads, then other leads with a clearly discernible T or U wave were used. Prolongation of QTc or QTUc was defined as being at least 450 milliseconds [8-11].

To calculate intrareader variability 10% of the ECGs were randomly

reviewed again by the same cardiologist without knowledge of prior measurements. We compared clinical findings, biochemical findings, and total K+ replacement between patients with and without QTUc or QTc prolongation. P N .05 was considered statistical significance [12].


After reviewing 350 patient charts with a diagnosis of DKA, 61 pa- tients met inclusion criteria and were enrolled in this study. Prolonga- tion of QTUc (>= 450 milliseconds), flat ST, and a prominent U wave that merged into a T wave occurred in 38 patients (62.9%). In comparing patients with normal and prolonged QTUc, we found a significant corre- lation existed between total supplemental K+ required during the ad- mission and the QTUc prolongation (P = .014). We noted that the serum K+ level had a greater decrease after starting treatment in pa- tients with a prolonged QTU interval when compared to patients with a normal QTU interval (P = .002). This finding suggested a strong asso- ciation between K+ storage depletion and prolongation of the QTU in- terval in the initial ECG. We did not find a significant correlation between with QTUc prolongation with age (P = .27), initial serum Blood sugar (P = .12), K+ level (P = .07), Ca level (P = .40), Mg level (P = .35), pH (P = .17), anion gap (P = .14), bicarbonate level (P =

.07), duration of hospitalization (P = .07), and heart rate (P = .13) (Table). No significant differences in biochemical or clinical variables were noted between patients with or without QTU prolongation. By analysis of variance, intrareader SD of QTUc was 7.

Figure. A 19-year-old woman with diagnosis of DKA. Initial laboratory results were pH 7.17, K+ 3.5 mmol/L, bicarbonate 5.4 mmol/L, anion gap 40, blood sugar 432 mg/dL, and urine pos- itive for ketone. Patient received 580 mEq KCL during admission. Initial ECG revealed normal sinus rhythm, depression of the T wave, sagging of the ST segment, and a prominent U wave, flat T wave merges with a positive U wave, and QTUc interval 700 milliseconds.

S. Talebi et al. / American Journal of Emergency Medicine 34 (2016) 263265 265


In spite of a severe total body K+ deficiency (3-10 mEq/kg of body weight) in patients with DKA, the initial the serum K+ concentration may be normal or elevated for several reasons. In acidotic patients, there is a shift of K+ from the intracellular to extracellular fluid. In DKA, depletion of intracellular glycogen and increased protein catabo- lism may release K+ to the extracellular space. In addition, a loss of 200 mEq or more of K+ may not lower extracellular concentrations sig- nificantly. Initiation of therapy included hydration and insulin which decreased extracellular K+, regardless of its starting point. In the usual patient with initial K+ values at a relatively high level, 4 to 8 hours of in- tensive therapy may be undertaken before hypokalemia will be detect- ed in serum [3,4,13-15].

Studies on the serum K+ and ECG changes in DKA showed that the ECG was more likely to reflect intracellular potassium changes. Cardiac effects of hypokalemia are usually minimal until serum K+ concentra- tions are less than 3 mEq. Hypokalemia causes depression of the T wave, sagging of the ST segment, and a prominent U wave. With marked hypokalemia, the T wave becomes progressively smaller and the U wave becomes larger. Sometimes, a flat or positive T wave merges with a pos- itive U wave (Figure). When T and U waves are fused, the U component must be included for QT interval measurement [16-22].

Kuppermann et al [23] in their study found that QTc prolongation is present in almost half of children diagnosed with DKA. Soler et al [24] showed that serial ECGs may also be used as a guide to the rate of K+ replacement in DKA. Maintaining of a normal ECG was suggested as an important goal during replacement of the potassium deficit of DKA. During the rapid phases of abnormal K+ metabolism such as during DKA, the serum potassium does not always reflect the intracellular K+ concentration. Particular attention to QTU interval should be part of the initial evaluation of all patients with DKA. Parenteral K+ administra- tion may be deferred until the serum K+ concentration is determined if the initial ECG shows hyperkalemic changes. Parenteral K+ infusion should be considered as initial management if hypokalemic changes are evident in the initial ECG because rapid life-threatening hypokale- mia may develop after starting of therapy. The ECG is a readily available tool for detecting total K+ deficits, although not a substitute for serum electrolyte determinations, but it can assist us making early decisions

in the management of DKA [23-25].

To the best of our knowledge, this is the first study to examine the interrelationship between total body K+ deficit and initial QT or QTU prolongation. On ECG, attention to subtle changes on the ECG is impor- tant during initial evaluation of the patient with DKA. These changes may potentially be the first sign of significant K+ depletion and fore- shadow the risk of a cardiac arrhythmia. As relatively small alterations in serum K+ concentration can have significant clinical manifestations,

these patients always benefit from close observation and monitoring. This study attempted to test the hypothesis that an ECG which is a read- ily available tool could be useful in assessing total K+ deficit. Definitely, larger prospective studies are necessary to get a better and in depth in- sight into this concept.


  1. McMullen JK. Asystole and hypomagnesaemia during recovery from diabetic ketoacidosis. Br Med J 1977;1(6062):690.
  2. McCormack P. Cardiac arrest complicating diabetes. Lancet 1977;2(8028):87-8.
  3. Burnell JM, Scribner BH, Uyeno BT, Villamil MF. The effect in humans of extracellular ph change on the relationship between Serum potassium concentration and intra- cellular potassium. J Clin Invest 1956;35:935-9.
  4. Halperin ML, Bear R, Goldstein MB, Richardson RM, Robson WL. Interpretation of the serum potassium concentration in metabolic acidosis. Clin Invest Med 1979;2(1):55-7.
  5. Adrogue HJ, Lederer ED, Suki WN, Eknoyan G. Determinants of plasma potassium levels in diabetic ketoacidosis. Medicine (Baltimore) 1986;65(3):163-72.
  6. Wolfsdorf JI, Allgrove J, Craig ME, Edge J, Glaser N, Jain V, et al. ISPAD Clinical Practice Consensus guidelines 2014. Diabetic ketoacidosis and hyperglycemic hyperosmolar state. Pediatr Diabetes 2014;15(Suppl. 20):154-79.
  7. Kitabchi AE, Umpierrez GE, Miles JM, Fisher JN. Hyperglycemic crises in adult pa- tients with diabetes. Diabetes Care 2009;32(7):1335-43.
  8. Surawicz B, Knilans T. Chou’s electrocardiography in clinical practice. 6th ed. Saunders; 2008 569-85[Chapter 24].
  9. Toivonen L. More light on QT interval measurement. Heart 2002;87(3):193-4.
  10. Viskin S. The QT, interval: too long, too short or just right. Heart Rhythm 2009;6(5): 711-5.
  11. Bazett H. An analysis of the time relationships of electrocardiograms. Heart 1920;7: 353-70.
  12. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33(1):159-74.
  13. Bernstein LM, Foley EF, William S, Hoffman WS. Renal function during and after di- abetic coma. J Clin Invest 1952;31(7):711-6.
  14. Fenn WO. The deposition of potassium and phosphate with glycogen in rat livers. J Biol Chem 1939;128:297-307.
  15. Martin HE, Wertman M. Serum potassium, magnesium and calcium levels in diabet- ic acidosis. J Clin Invest 1947;26(2):217-28.
  16. Surawicz B, Lepeschkin E. The electrocardiographic pattern of hypopotassernia with and without hypocalcemia. Circulation 1953;8(6):801-28.
  17. Weaver WF, Burchel H. Serum potassium and the electrocardiogram and hypokalaemia. Circulation 1973;47:408-18.
  18. Helfant RH. Hypokalaemia and arrhythmias. Am J Med 1986;80(Suppl. 4A):13-22.
  19. Reddy GV, Schamroth I, Schamroth CL. Tall and peaked U waves in hypokalaemia. Chest 1987;91:605-7.
  20. Kunin AS, Surawicz B, Sims EAH. Decrease in serum potassium concentrations and appearance of cardiac arrhythmias during infusion of potassium with glucose in po- tassium depleted patients. N Eng J Med 1962;266:228-33.
  21. Henderson CB. Potassium and the cardiographic changes in diabetic acidosis. Br Heart J 1953;15(1):87-94.
  22. Nadler CS, Bellet S, Lanning M. Influence of the serum potassium and other electro- lytes on the electrocardiogram in diabetic acidosis. Am J Med 1948;5(6):838-48.
  23. Kuppermann N, Park J, Glatter K, Marcin JP, Glaser NS. Prolonged QT interval corrected for heart rate during diabetic ketoacidosis in children. Arch Pediatr Adolesc Med 2008;162(6):544-9.
  24. Soler NG, Bennett MA, Fitzgerald MG, Malins JM. Electrocardiogram as a guide to po- tassium replacement in diabetic ketoacidosis. Diabetes 1974;23(7):610-5.
  25. Danowski TS, Peters JH, Rathbun JC, Quashnock JM, Greenman L. Studies in diabetic acidosis and coma, with particular emphasis on the retention of administered potas- sium. J Clin Invest 1949;28(1):1-9.