Case Presentation: An 89-year-old year old woman with a history of type II diabetes mellitus, insulin-naïve and well-controlled on a regimen of metformin, sitagliptin, and dapagliflozin, presented with 3 days of nausea, vomiting, and poor intake of food and liquids. She had recently been prescribed cephalexin for a suspected UTI. She reported a subacute history of poor appetite, and of daily alcohol use of 2 standard drinks. At presentation, she appeared clinically volume depleted with a mild tachycardia with HR 107/min, dry mucous membranes, and decreased skin turgor, but with an otherwise unremarkable physical exam. Initial blood chemistries revealed an elevated anion gap of 21, a bicarbonate of 23, and normal venous pH of 7.34. Her initial blood glucose was 140 mg/dL. In addition to 11-20 WBC/hpf and negative nitrites and leukocyte esterase, urinalysis was significant for 2+ ketones, and serum ketones were elevated at 33.0 mg/dL (normal <2.8). Delta anion gap/delta bicarbonate ratio (21-12)/(24-23) yielded a value >2 consistent with a concomitant anion gap metabolic acidosis, and metabolic alkalosis. In the setting of a SGLT2 inhibitor use and this acute illness, a presumptive diagnosis of euglycemic diabetic ketoacidosis was made. Her oral diabetes regimen was held, and she was treated with intravenous fluid resuscitation and subcutaneous insulin. Her anion gap normalized at 12 with a bicarbonate of 27 six hours later, and her nausea responded to antiemetics. She was discharged on a revised oral diabetic regimen.
Discussion: Sodium-Glucose transporter 2 inhibitors (SGLT2i) are increasingly prescribed oral antihyperglycemic agents which act at the proximal tubule to prevent glucose reabsorption by the kidney. Since their introduction, the risk of developing diabetic ketoacidosis (DKA) following their initiation has become well defined. In one retrospective study, the hazard ratio of developing DKA within 180 days of starting any SGLT2i was 2.2 when compared with starting a DPP4 inhibitor (21 cases per 1,000 person-years). Similarly, a cohort study comparing two groups of 17,000 patients starting an SGLT2i (61% taking dapagliflozin, and 38% empagliflozin) or any GLP-1 receptor agonist showed a hazard ratio of 2.14 for developing DKA in the SGLT2i cohort. Such cases often develop in the absence of significant hyperglycemia— so-termed euglycemic DKA. Surreptitiously lower glucose levels may cause a delay in diagnosis, leading to catastrophic complications. In our patient, the underlying metabolic acidosis was further obscured by a co-existing metabolic alkalosis, likely due to vomiting. Calculation of the ratio of the change in the patient’s anion gap to the change in bicarbonate was key in demonstrating this patient’s underlying acid-base disturbance, and helped lead to the prompt initiation of appropriate, effective treatment. This case demonstrates that it is important for today’s hospitalist to maintain a high index of suspicion for euglycemic DKA in patients on an SGLT2i who present with a consistent clinical picture. This is especially true as use of these drugs becomes more commonplace. The case also underscores the importance of maintaining a systematic approach to the diagnosis of acid-base disturbances.
Conclusions: • Diagnosis of SGLT2i-associated euglycemic DKA may require a high clinical index of suspicion.• Employing a systematic approach to diagnosing acid-base disorders can help prevent diagnostic and therapeutic delays in euglycemic DKA.