The diagnosis of DKA should be suspected at triage, and aggressive fluid therapy should be initiated before receiving the laboratory results4 (Figure 225-2). Place patients on a cardiac monitor and begin at least one large-bore (16- to 18-gauge) IV infusion of normal saline. A second IV line with 0.45% normal saline at minimal rate to keep the IV line open can be considered. The goals of therapy are (1) volume repletion, (2) reversal of the metabolic consequences of insulin insufficiency, (3) correction of electrolyte and acid-base imbalances, (4) recognition and treatment of precipitating causes; and (5) avoidance of complications. The order of therapeutic priorities is volume first and foremost, correction of potassium deficits, and then insulin administration. Metabolic disturbances should be corrected at the approximate rate of occurrence or over 24 to 36 hours.
Timeline for the typical adult patient with suspected diabetic ketoacidosis (DKA). *IV insulin infusion <1.0 units/kg/hr may require a bolus dose of regular insulin (0.1 unit/kg)4. AG = anion gap; ARDS = acute respiratory distress syndrome; BS = blood sugar; ECG = electrocardiogram; ICU = intensive care unit; I/Os = inputs/outputs; NS = normal saline; TKO = to keep vein open; VBG = venous blood gas.
Meeting the goals of safely replacing deficits and supplying missing insulin requires monitoring every 2 hours of electrolytes (glucose, potassium, and anion gap), vital signs, level of consciousness, and volume input/output until recovery is well established. The goal of treatment is glucose <200 milligrams/dL (<11.1 mmol/L), bicarbonate ≥18 mEq/L (≥18 mmol/L), and venous pH >7.3.
Fluid helps restore intravascular volume and normal tonicity, perfuse vital organs, improve glomerular filtration rate, and lower serum glucose and ketone levels.4 Rehydration improves the response to low-dose insulin therapy.4 The average adult patient has a water deficit of 100 mL/kg (5 to 10 L) and a sodium deficit of 7 to 10 mEq/kg (7 to 10 mmol/L/kg).4 Normal saline is the most frequently recommended fluid for initial volume repletion even though the extracellular fluid of the patient is initially hypertonic.14 Normal saline does not provide "free water" to correct intracellular fluid loss, but it does prevent an excessively rapid fall in extracellular osmolarity and the potential devastating transfer of excessive water into the CNS. After initial resuscitation with normal saline, change fluids to 0.45% normal saline once the corrected serum sodium is normal or elevated.6,7
Based on clinical suspicion alone and before initial electrolyte results, administer the initial fluid bolus of isotonic saline at a rate of 15 to 20 mL/kg/h during the first hour unless there are mitigating circumstances.6 The rate of hydration should depend on hemodynamic stability, hydration status, urine output, and serum electrolytes. After the initial bolus, administer normal saline at 250 to 500 cc/h in hyponatremic patients, or give 0.45% normal saline at 250 to 500 cc/h for eunatremic and hypernatremic patients.7 In general, the first 2 L are administered rapidly over 0 to 2 hours, the next 2 L over 2 to 6 hours, and then an additional 2 L over 6 to 12 hours. This replaces approximately 50% of the total water deficit over the first 12 hours, with the remaining 50% water deficit to be replaced over the subsequent 12 hours. When the blood glucose level is 250 milligrams/dL (13.8 mmol/L), change to 5% dextrose in 0.45% normal saline. Patients without extreme volume depletion can be managed safely with a more modest fluid replacement regimen such as 250 to 500 mL/h for 4 hours. Consider central venous pressure or pulmonary artery wedge pressure monitoring in the elderly or in those with heart or renal disease. Excess fluid may contribute to the development of adult respiratory distress syndrome and cerebral edema.
Patients in DKA usually present with profound total-body potassium deficits in the range of 3 to 5 mEq/kg (3 to 5 mmol/kg).4 This deficit is created by insulin deficiency, metabolic acidosis, osmotic diuresis, and frequent vomiting. Only 2% of total-body potassium is intravascular. The initial serum concentration is usually normal or high because of the intracellular exchange of potassium for hydrogen ions during acidosis, the total-body fluid deficit, and diminished renal function. Initial hypokalemia indicates severe total-body potassium deficits, and large amounts of replacement potassium are usually necessary in the first 24 to 36 hours.
Correction of the acidosis predicts the change in serum potassium concentration. For each 0.1 decrease in pH, serum potassium concentration rises approximately 0.5 mEq/L (0.5 mmol/L), and the same relationship holds as the pH increases. This can be used as a guide for estimating the serum potassium concentration when pH balance is restored.
During initial therapy for DKA, the serum potassium concentration may fall rapidly, primarily due to the action of insulin promoting reentry of potassium into cells and, to a lesser degree, the dilution of extracellular fluid, correction of acidosis, and increased urinary loss of potassium. If these changes occur too rapidly, precipitous hypokalemia may result in fatal cardiac arrhythmias, respiratory paralysis, paralytic ileus, and rhabdomyolysis. The rapid development of severe hypokalemia is potentially the most life-threatening electrolyte derangement during the treatment of DKA.4
As a general guideline, an initial serum potassium level >3.3 mEq/L (>3.3 mmol/L) but <5.2 mEq/L (<5.2 mmol/L) (before fluid resuscitation and insulin, coupled with urine output) calls for 20 to 30 mEq/L (20 to 30 mmol/L) for at least 4 hours to keep K+ between 4 and 5 mEq/L (4 and 5 mmol/L).6 Because the most rapid changes occur during the first few hours of therapy, measure the plasma potassium level initially every 2 hours. If oliguria or renal insufficiency is present, withhold or decrease potassium replacement.
Initial hypokalemia (<3.3 mEq/L or <3.3 mmol/L) is uncommon but necessitates a more aggressive replacement before insulin therapy.4 In this setting, give potassium IV at 20 to 30 mEq/h (20 to 30 mmol/h) and hold insulin until [K+] is ≥3.5 mEq/L (≥3.5 mmol/L).6,7 There is no advantage to using potassium phosphate (K2PO4) compared to potassium chloride because K2PO4 may result in hypocalcemia and metastatic precipitation of calcium phosphate in tissues. Oral potassium replacement is safe and effective and is the preferred route of replacement as soon as the patient can tolerate oral fluids. In DKA, initial potassium replacement is usually by an intravenous line. Each institution may have specific guidelines for potassium replacement, but a general approach is a rate no faster than 10 mEq/h (10 mmol/h) via peripheral IV or 20 mEq/h (20 mmol/h) via central line access. Continuous electrocardiogram monitoring is generally recommended while replacing potassium in the severely hypokalemic patient. During the first 24 hours, 100 to 200 mEq or mmol of KCl is usually required.
Obtain an electrocardiogram immediately and check for signs of hyperkalemia once DKA is suspected.
Giving potassium to a patient in a hyperkalemic potentiating state (i.e., acidemia, insulin deficiency, volume contraction, renal insufficiency) may dangerously increase the extracellular potassium level and precipitate fatal dysrhythmias. The initial measurement of serum electrolytes, electrocardiogram review for signs of hyperkalemia, and the presence of urine output determine initial potassium therapy. An initial serum potassium level >5.2 mEq/L usually reflects a more profound acidemia and volume depletion, or renal insufficiency. Fluid and insulin therapy alone usually will lower the serum potassium level rapidly. Albuterol nebulization can provide an additional quick potassium-lowering effect. See chapter 17, Fluids and Electrolytes, for further treatment of hyperkalemia.
Low-dose regular insulin administration by an infusion pump is simple and safe, ensures a steady blood concentration of insulin, allows flexibility in adjusting the insulin dose, and promotes a gradual fall in serum glucose and ketone body levels.4 The half-life of IV insulin is 4 to 5 minutes, with an effective biologic half-life at the tissue level of approximately 20 to 30 minutes.
After the initial fluid bolus, or simultaneously in a second IV line, administer insulin at a rate of 0.1 to 0.14 unit/kg/h with no insulin bolus once hypokalemia ([K+] <3.3 mEq/L [<3.3 mmol/L]) is excluded. An alternative insulin regimen is 0.1 unit/kg bolus IM, if it is difficult to establish another IV line,9 followed by a drip rate at 0.1 unit/kg/h.7 An IV loading dose of insulin is not recommended in children and new-onset young adult diabetics and is optional in adults.9,15 Plasma glucose concentration typically decreases by 50 to 75 milligrams/dL/h (2.8 to 4.2 mmol/L/h), but if the blood glucose fails to drop by 10% 1 hour after initial therapy, or 3 mmol/L/h, (assuming adequate hydration), give a 0.14 unit/kg bolus and resume insulin drip rate.6,7 Another option is to increase the insulin infusion rate by 1 unit/h.9 The incidence of nonresponse to low-dose continuous IV insulin administration is 1% to 2%, with infection being the primary reason for failure to respond.
Resolution of hyperglycemia usually occurs earlier than resolution of the anion gap, so once the serum glucose is 200 milligrams/dL (11 mmol/L), add dextrose to the IV fluids and reduce the insulin drip rate to 0.02 to 0.05 unit/kg/h. Maintain the serum glucose between 150 and 200 milligrams/dL (8.3 and 11 mmol/L) until the resolution of DKA.7 Occasionally a 10% dextrose solution may be needed to maintain glucose levels.9 Continue the insulin infusion until the resolution of DKA—glucose <200 milligrams/dL (<11 mmol/L) and two of the following: a serum bicarbonate level >15 mEq/L, a venous pH >7.3, and/or a normal calculated anion gap.7 Monitor laboratory values every 1 to 2 hours to ensure that insulin is being administered in the desired amount.
from IV Insulin After DKA Correction A transition from the IV insulin infusion to SC insulin is necessary to avoid relapse to hyperglycemia or DKA when the insulin infusion is stopped. Relapse can occur quickly, within an hour after IV insulin is stopped, due to the short duration of action of IV insulin. The method of insulin transition varies, and there is no set protocol. Once the patient eats, the glucose infusion can be stopped. In patients who can eat, the transition should include a short-acting and long-acting insulin given when DKA has resolved. It is best to collaborate with the inpatient team or endocrinologist to develop a protocol for the transition to SC insulin. One method consists of giving 10 units of SC regular insulin 30 to 60 minutes before the insulin infusion is stopped and 80% of the usual long-acting insulin dose 1 to 2 hours before discontinuing the IV insulin infusion. Another method is to give 50% of the usual long-acting insulin dose 2 hours before the IV insulin infusion is stopped (see Figure 223-1, for onset and duration of action of long-acting insulins). If the patient is a newly diagnosed diabetic, one can estimate a starting dose of long-acting insulin at 0.1 to 0.2 unit/kg. Additional glucose coverage can be provided with short-acting insulin as needed. Continue glucose checks every hour for 2 hours. Further intervals for glucose checks and the need for additional SC regular insulin dosing depend on the patient's response and institutional protocols.
In uncomplicated mild to moderate DKA, the use of rapid-acting SC insulin may be another treatment option, although the standard treatment remains continuous IV insulin.4,7,16 The dose of SC rapid-acting insulin is an initial injection of 0.2 unit/kg followed by 0.1 unit/kg every hour, or an initial dose of 0.3 unit/kg followed by 0.2 unit/kg every 2 hours until blood glucose is <250 milligrams/dL (<13.8 mmol/L). Then, the insulin dose is decreased by half and administered every 1 or 2 hours until resolution of DKA.4,7 This can avoid intensive care admissions and lower hospital costs, but still requires close nursing monitoring that is difficult to accomplish in the ED or in a regular hospital bed.
Serum phosphate levels often are normal or increased on presentation of DKA and do not reflect the total-body phosphate deficits secondary to enhanced urinary losses.11 Phosphate (similar to glucose and potassium) reenters the intracellular space during insulin therapy, resulting in low phosphate concentrations. Hypophosphatemia is usually most severe 24 to 48 hours after the start of insulin therapy. Acute phosphate deficiency (<1.0 milligram/dL) can result in hypoxia, skeletal muscle weakness, rhabdomyolysis, hemolysis, respiratory failure, and cardiac dysfunction.
There is no established role for initiating IV K2PO4 for DKA in the ED.4,7,11 In general, do not give IV phosphate unless the serum phosphate concentration is <1.0 milligram/dL (0.323 millimol/L). Significant hypophosphatemia tends to develop many hours into therapy, after the patient is already admitted. Undesirable side effects from IV phosphate administration include hyperphosphatemia, hypocalcemia, hypomagnesemia, metastatic soft tissue calcifications, hypernatremia, and volume loss from osmotic diuresis. If absolutely necessary (a phosphate level <1.0 milligram/dL early in therapy), IV phosphate replacement should be administered as IV K2PO4, 2.5 to 5 milligrams/kg (0.08 to 0.16 millimol/kg).17 Monitor serum calcium level if giving supplemental phosphate.
Osmotic diuresis may cause hypomagnesemia and deplete magnesium stores from bone. Hypomagnesemia may inhibit parathyroid hormone secretion, causing hypocalcemia and hyperphosphatemia. If the serum magnesium concentration is <2.0 mEq/L (<1.0 mmol/L) or symptoms are suggestive of hypomagnesemia, give magnesium sulfate 2 grams IV over 1 hour. Obtain serum magnesium and calcium levels on presentation and 24 hours into therapy. Monitor levels every 2 hours if there is initial hypomagnesemia or hypocalcemia or if symptoms suggestive of hypomagnesemia or hypocalcemia occur.
Acidotic patients routinely recover from DKA without alkali therapy, as fluid and insulin therapy inhibit lipolysis and resolve ketoacidosis without added bicarbonate.4 Give bicarbonate if the initial pH is ≤6.9, but do not give bicarbonate if the pH is ≥7.0.7,16,18,19
Severe metabolic acidosis is associated with numerous cardiovascular (impaired contractility, vasodilation, and hypotension) and neurologic (cerebral vasodilation and coma) complications. Theoretical advantages of bicarbonate include improved myocardial contractility, elevated ventricular fibrillation threshold, improved catecholamine tissue response, and decreased work of breathing. The disadvantages of bicarbonate include worsening hypokalemia, paradoxical CNS acidosis, worsening intracellular acidosis, impaired (shift to left) oxyhemoglobin dissociation, hypertonicity and sodium overload, delayed recovery from ketosis, elevation of lactate levels, and possible precipitation of cerebral edema. During DKA treatment, hydrogen ion production ceases when ketogenesis stops; excessive hydrogen ions are eliminated through the urine and respiratory tract. Ketone body metabolism results in the endogenous production of alkali.
The decision to use bicarbonate in DKA patients should be based on the clinical condition and pH of the patient. Potential benefits of bicarbonate in the elderly with cardiovascular instability must be balanced against the potential disadvantages.4 There may be selected patients who benefit from cautious alkali therapy, including those with decreased cardiac contractility and peripheral vasodilatation, and patients with life-threatening hyperkalemia and coma. Patients with severe acidosis may be at higher risk for clinical deterioration, so adults with a pH <6.9 can be given 100 mEq (100 mmol) of sodium bicarbonate in 400 mL of water with 20 mEq (20 mmol) KCl at 200 mL/h for 2 hours until the venous pH >7.0. If the pH remains <7.0 despite the infusion, repeat the infusion until pH >7.0.6,7 Remember to check [K+] every 2 hours. Severe acidosis (pH <7.0) and worsening pH despite aggressive therapy for DKA should prompt investigation for other causes of metabolic acidosis (see chapter 15).