Iatrogenic Over-replacement Thyrotoxicosis Induced DKA

Background

Diabetic ketoacidosis (DKA) in established type 1 diabetes mellitus (T1DM) is commonly precipitated by insulin omission or infection. When no clear trigger is identified, alternative causes merit consideration. Thyrotoxicosis can exacerbate hyperglycemia; increased hepatic glucose production, insulin degradation and lipolysis. Insulin-pump therapy also carries increased DKA risk when delivery is interrupted due of limited subcutaneous insulin reserves. This case highlights two distinct DKA episodes associated with iatrogenic thyrotoxicosis and insulin-pump failure.

Case Presentation

A 78-year-old woman with T1DM managed by insulin pump, coronary artery disease (CAD), hypothyroidism and peripheral arterial disease presented with a pump “high” alarm (fingerstick 542 mg/dL), nausea, vomiting, and acute substernal chest pain radiating to both shoulders. Despite crystalloid administration, her blood pressure declined from 130s/80s to 90s/40s. Laboratory evaluation confirmed DKA (glucose 547 mg/dL, anion gap 19, venous pH 7.23, bicarbonate 18 mmol/L, beta-hydroxybutyrate 42.6 mg/dL, serum osmolality 325 mOsm/kg, and 3+ urine ketones), accompanied by hypotension, worsening lactic acidosis (2.7 to 4.2 mmol/L), and mild troponin elevation (28 to 32 ng/L) with nonspecific ST changes. CT angiography excluded pulmonary embolism. Review of a hospitalization one month earlier revealed a TSH of 66.3. Without medication reconciliation, clinicians presumed inadequate replacement and increased levothyroxine from 137 to 175 mcg. Pharmacy records subsequently confirmed she had not filled levothyroxine for four months. After one month of adherence to the increased dose, she developed biochemical thyrotoxicosis (TSH 0.02, free T4 2.29) and presented with DKA. She was admitted to the ICU.

Findings

DKA resolved with intravenous insulin and fluids, and hypotension resolved without vasopressors. No alternative precipitant was identified; the clinical timeline supported iatrogenic thyrotoxicosis as the cause of the initial DKA. Echocardiography showed preserved ejection fraction (50–55%). Lexiscan showed no inducible ischemia, favoring demand-related myocardial injury. HbA1c was 7.1%, consistent with relatively well-controlled baseline diabetes. Levothyroxine was held. Eight days after discharge, she was readmitted with recurrent DKA (glucose 604 mg/dL, pH 7.27, beta-hydroxybutyrate 40 mg/dL) after her pump stopped functioning when an expired Dexcom sensor could not be replaced. This episode resolved with protocol-based treatment, and she was discharged after a replacement sensor arrived.

Discussion

Two preventable DKA episodes arose from distinct failures: inappropriate medication adjustment and interruption of insulin delivery due to device-access barriers. The first highlights the importance of verifying pharmacy dispensing history before modifying thyroid replacement. In older patients with CAD, levothyroxine over-replacement also increases concern for atrial fibrillation, myocardial ischemia, and heart failure. The second underscores the vulnerability of pump-dependent patients when monitoring supplies are unavailable.

Conclusion

Recurrent DKA in device-dependent T1DM warrants investigation beyond presumed insulin omission. Medication reconciliation, careful interpretation of thyroid function tests, evaluation for alternative precipitants, and reliable access to insulin-delivery and monitoring supplies are essential to prevention.