Chapter 11: Diabetic Ketoacidosis in Pregnancy

Despite recent advances in the evaluation and medical treatment of diabetes in pregnancy, diabetic ketoacidosis (DKA) remains a matter of significant concern. The fetal loss rate in most contemporary series has been estimated to range from 10% to 25%. Fortunately, since the advent and implementation of insulin therapy, the maternal mortality rate has declined to 1% or less. In order to favorably influence the outcome in these high-risk patients, it is imperative that the obstetrician/provider be familiar with the basics of the pathophysiology, diagnosis, and treatment of DKA in pregnancy.

DKA is characterized by hyperglycemia and accelerated ketogenesis. Both a lack of insulin and an excess of glucagon and other counter-regulatory hormones significantly contribute to these problems and their resultant clinical manifestations. Glucose normally enters the cell secondary to the effects of insulin. The cell then may use glucose for nutrition and energy production. When insulin is lacking, glucose fails to enter the cell. The cell responds to this starvation by facilitating the release of counter-regulatory hormones, including glucagon, catecholamines, and cortisol. These counter-regulatory hormones are responsible for providing the cell with an alternative substrate for nutrition and energy production. By the process of gluconeogenesis, fatty acids from adipose tissue are broken down by hepatocytes to ketones (acetone, acetoacetate, and β-hydroxybutyrate [BHB] = ketone bodies), which are then used by the body cells for nutrition and energy production (Fig. 11-1). The lack of insulin also contributes to increased lipolysis and decreased reutilization of free fatty acids, thereby providing more substrate for hepatic ketogenesis. A basic review of the biochemistry involving DKA is presented in Fig. 11-1.

Now that we have an understanding of how and why ketone bodies are produced during DKA, what are the maternal consequences resulting from excessive ketogenesis? In general, ketone bodies are considered to be moderately strong acids. In response to the fall in pH in most body fluids created by an accumulation of these acids, the body reacts physiologically to correct the resultant metabolic acidosis. The respiratory rate and depth increase (Kussmaul respirations) in an attempt to blow off carbon dioxide, initiating a corrective trend toward compensatory respiratory alkalosis. Serum bicarbonate levels decline, and as a result, the anion gap becomes abnormally elevated. In addition to increasing fatty acid production, poor glucose utilization results in severe hyperglycemia. Untreated hyperglycemia leads to marked glycosuria, initiating a significant osmotic diuresis. As a result, dehydration, electrolyte depletion, and, if left untreated, cardiac failure and death may follow.

A vicious cycle is created by an increase in dehydration-mediated serum hyperosmolarity and catabolism, propagated by Kussmaul respiration, leading to a further production of glucose counter-regulatory hormones, lipolysis, and subsequent hyperketonemia. An algorithm for this clinical pathophysiologic response is presented in Fig. 11-2.

The fetus appears to be at significant risk of sudden intrauterine death during an episode of maternal DKA. The mechanism for this sudden death is not completely understood; however, it appears to be related to a combination of factors. Alterations in fetal fluid and electrolyte balance, poor uterine perfusion resulting from maternal hypovolemia, and increased acid load in the form of fatty acids and lactate, all favor a reduction in fetal oxygenation and metabolic acid clearance. When caring for a patient in DKA who is carrying a potentially viable fetus, careful fetal monitoring should be judiciously used. Often, signs of fetal stress become apparent, reflecting the degree of maternal metabolic derangement. Decreased variability and late decelerations in the fetal heart tracing as well as abnormal umbilical artery Doppler values are among the changes that can be observed. Delivery of a compromised baby should be prudently delayed until the mother is metabolically stable. Correction of maternal metabolic abnormalities generally results in a rapidly improved fetal condition. Therefore, efforts should be directed at improving maternal deficits reserving emergency operative intervention for unresponsive persistent fetal compromise. As BHB can cross the placenta and accumulate in the fetal brain, with reports of neuronal injury in some neonates, the pediatric team should be informed of current or previous DKA during pregnancy.

In pregnancy, DKA may occur at a lower plasma glucose value as compared to the nonpregnant patient. DKA has been observed at plasma glucose levels as low as 180 mg/dL. It appears that a relative insulin resistance of pregnancy combined with a greater tendency toward ketosis reduces the threshold for DKA during pregnancy. The insulin resistance during pregnancy is related to an increased production of placental hormones, progesterone, insulinase, and cortisol (Fig. 11-3). Respiratory alkalosis of pregnancy and a decreased renal buffering capacity of bicarbonate serve to worsen the complications associated with DKA.

The maternal and fetal concerns resulting from DKA emphasize the importance of a rapid and reliable diagnosis. The diagnosis of DKA should be based on clinical examination and supported by an evaluation of biochemical parameters. Table 11-1 summarizes the clinical presentation, the biochemical definition, and additional laboratory findings associated with DKA.

Diabetic ketoacidosis (DKA) during pregnancy is a medical emergency. When available, the patient should be admitted to an intensive care facility and consultations obtained from maternal-fetal medicine, endocrinology, and neonatology. A detailed history and physical examination should be performed to search for underlying precipitating factors for DKA such as noncompliance with insulin administration, insulin pump malfunction, or infection (urine, skin, lungs, dental, and amniotic cavity). Glucocorticoid administration in the setting of preterm labor has also been implicated in hyperglycemia leading to DKA. Remember that any process that results in dehydration, starvation, or a catabolic state (stress) with insulin antagonism predisposes the patient to DKA. Defining the underlying etiology is also a priority, not only for the treatment of the current episode of DKA, but also for preventing a future recurrence.

Fetal monitoring should be initiated if there is a potentially viable fetus (see Chap. 24). Intervention on behalf of the fetus should be withheld until the maternal metabolic condition is stabilized. Oxygen therapy and maternal position changes, however, should be initiated to help improve fetal perfusion while correcting maternal biochemical and plasma volume abnormalities (see Chap. 22). A detailed flow sheet including a comprehensive recording of serial laboratory values, at appropriate time intervals, should be started at bedside to facilitate patient assessment following therapy. The use of invasive hemodynamic monitoring should be reserved for the patient with severe renal compromise in an effort to properly guide rehydration while avoiding iatrogenic pulmonary edema. Other less invasive measures such as the initiation of an arterial line to follow serial arterial blood gases, a Foley catheter to monitor strict urinary output, and continuous peripheral pulse oximetry should be initiated as an early adjunct to beginning therapy. The basic premise for the treatment of DKA in the pregnant patient is the simultaneous correction of fluid and electrolyte imbalance and treatment of hyperglycemia and acidosis (Fig. 11-4).

Hypovolemia during DKA results primarily from hyperglycemia-induced osmotic diuresis (see Fig. 11-2). Because the restoration of intravascular volume improves perfusion and augments systemic insulin delivery to peripheral tissues, repletion of the circulating intravascular volume is the number 1 treatment priority. The estimated total water deficit is calculated to be 100 mL/kg of actual body weight (4-10 L deficit). Once renal competence is established (urinary output of at least 0.5 cc/kg/h), fluid replacement may be initiated. Keeping in mind that many of the patients encountered with DKA during pregnancy may have preexisting renal compromise (pregestational diabetes with nephropathy), a baseline evaluation of serum blood urea nitrogen (BUN) and creatinine would be prudent to avoid fluid overload in a patient with markedly reduced creatinine clearance. Most authorities recommend isotonic normal saline as the intravenous fluid of choice for volume replacement instead of lactated Ringer solution. The reason behind this recommendation is that the use of hypotonic solutions (0.45 NS [normal saline] or lactated Ringer) as initial treatment can lead to a rapid decline in plasma osmolarity that may lead to cellular swelling and resultant cerebral edema. Therefore, the recommended approach is to use isotonic NS and replace 75% of the total calculated fluid deficit within the first 24 hours of therapy. The remaining 25% of the deficit is replaced over the remainder of the patient’s hospitalization. At our institution, 1 L of isotonic NS is given over the first hour, and 500 cc NS per hour is given over second and third hours. Isotonic NS, therefore, is used for the first 3 hours of therapy. Crystalloid infusions such as Plasmalyte are increasingly gaining favor for initial volume resuscitation in the presence of severe hypovolemia associated with DKA. Thereafter, however, lactated Ringer or 0.45 NS is given at a rate of 250 cc/h until 75% of the total deficit is replaced over the first 24 hours. Lactated Ringer solution is used to avoid further iatrogenic contributions to the fall in serum pH as the pH of lactated Ringer is 6.5 compared to a pH of 5.0 for isotonic NS. In addition, the sodium load with isotonic saline may create hypernatremia fostering the recommendations to switch from isotonic saline to a more hypotonic solution (lactated Ringer or 0.45 NS) when the patient’s serum sodium increases above 150 to 155 mEq/L (Table 11-2). Aggressive hydration with at least 200 cc/h should continue even after the patient has been transitioned to subcutaneous insulin for the next 24 hours to prevent the recurrence of DKA.

Intravenous insulin administration is the mainstay treatment for the pregnant patient in DKA. An initial intravenous loading dose of 0.1 U/kg (8-10 units is a good starting point) followed by a constant infusion of 0.1 U/kg/h effectively inhibits lipolysis and ketogenesis leading to suppression of hepatic glucose output with resultant lower serum glucose levels. The insulin infusion is prepared by adding 100 units of regular insulin to 100 mL of NS (1 cc = 1 unit). Alternatively, 0.4 U/kg of rapid acting insulin (Lispro) is preferred and may be administered as a subcutaneous (SC) or intramuscular (IM) bolus until parenteral (IV) access is achieved. However, the SC and IM routes may be fraught with unreliable insulin absorption due to poor perfusion associated with severe dehydration. Serum glucose levels should be measured every 1 to 2 hours during insulin infusion. General principles of an intravenous insulin infusion are listed in Table 11-3.

The intravenous insulin infusion should be titrated to reduce the serum glucose at a rate of 60 to 75 mg/dL/h or less to avoid rapid changes in the serum osmolarity that may precipitate cerebral edema and risk cerebellar pontine myelinolysis. A good rule of thumb is that if the plasma glucose does not decrease by 10% in the first hour or 20% by the second hour of therapy, repeat the intravenous loading dose or double the current continuous infusion rate. As the patient’s blood glucose approaches 250 mg/dL, add 5% dextrose to the IV infusion and reduce the hourly insulin infusion by one half (0.05-0.1 U/kg/h). Maintain serum glucose levels between 150 and 200 mg/dL while hypovolemia and electrolyte abnormalities are corrected. Be aware that while monitoring serum ketones in response to insulin administration, a paradoxical increase in serum acetone should be anticipated. While BHB is the predominant ketone in DKA, the ketotest primarily measures serum acetone. During insulin therapy, the overall production of ketones clearly diminishes and there is a shift of ketone production from BHB to acetone (oxidation) and acetoacetate. This phenomenon results in the apparent paradoxical worsening of ketoacidemia at the onset of insulin therapy. When testing is available, BHB should be measured directly in lieu of the ketotest and the anion gap should be used to assess response to therapy. The intravenous insulin infusion should be continued until the serum bicarbonate (18-31 mEq/L) and anion gap (<12) normalize. Table 11-4 and Fig. 11-5 summarize the mechanics of changing from an intravenous insulin infusion to SC insulin or SC insulin pump, respectively. Table 11-5 provides additional helpful information regarding the insulin sensitivity factor. An emerging body of evidence suggests that the early initiation of long-acting insulin (Glargine), concurrently with parenteral insulin, is not only considered safe, but it may also decrease rates of rebound hyperglycemia following discontinuation of insulin infusion. A reasonable initial dose of insulin glargine is 0.3 mg/kg administered in a onetime daily dose in the first 12 hours of the insulin infusion.

The anticipated potassium deficit in a pregnant patient with DKA is 5 to 10 mEq/kg. Potassium replacement, however, is most often delayed for the first 2 to 4 hours of therapy as the initial serum potassium is usually normal to mildly elevated and adequate diuresis has yet to be established. Once fluid and insulin therapy have been instituted and corrections of the metabolic acidosis are underway, serum potassium may precipitously fall as a result of urinary loss and intracellular shift. When the patient’s plasma potassium has fallen below 5 mEq/L and an adequate diuresis has been established (at least 0.5 cc/kg/h), potassium administration should be initiated. The recommended method of potassium replacement is summarized as follows:

1. Mix 40 to 60 mEq KCl/L in isotonic saline.

2. If plasma K⁺ is

   • ≥4 mEq/L, give 10 to 20 mEq

   • <4 mEq/L, give 30 to 40 mEq

3. Replace K⁺ cautiously, monitoring urinary output and serum K⁺ frequently (every 2-4 hours).

4. Do not exceed K⁺ supplementation to greater than 20 mEq/h owing to concern for cardiotoxicity/-arrhythmias.

5. Replace entire K⁺ deficit over the span of the patient’s entire hospitalization.

6. Alternatively, in the face of DKA-induced maternal phosphate deficiency (<2 mg/dL), K₂PO₄ (K-Phos) may be given as potassium replacement instead of KCl.

The use of sodium bicarbonate therapy in patients with severe DKA remains a controversial practice and should be used only in severe refractory cases of metabolic acidosis resulting from DKA. Recent studies of pregnant patients with DKA have failed to show an improvement in outcomes among those receiving bicarbonate therapy. Bicarbonate is usually reserved for those patients who have an arterial pH of less than 6.9 to 7.0 and/or a HCO₃^– less than 5 mEq/L. Rapid undiluted correction of metabolic acidosis with sodium bicarbonate is unwarranted and may lead to severe hypokalemia, hypernatremia, impaired oxygen delivery, and a paradoxical fall in cerebrospinal fluid pH. Administration of one ampule (44 mEq sodium bicarbonate) diluted in 1000 mL of 0.45 NS is a reasonable dose in a patient with severe acidemia (pH <7.0). The total deficit of bicarbonate may be calculated (obtain base deficit on arterial blood gas):

Bicarbonate (mEq) regained to fully correct metabolic acidosis.

Because oxygen hemoglobin affinity is augmented in the presence of an alkalotic shift of the oxygen-hemoglobin disassociation curve to the left, it is prudent not to fully correct the patient’s metabolic acidosis, ensuring better oxygen delivery to the fetus.

Please refer to Table 11-4 for an algorithmic summary of the treatment of DKA in pregnancy.