A DKA Presentation That May Not Mean Type 1 Diabetes
He arrived looking like a new type 1 diagnosis. His labs and phenotype tell a more specific story — one where the insulin he's on now may not be the insulin he needs a year from now.
Terrence O., a 46-year-old man, has coached youth basketball at his neighborhood recreation center for over a decade, work he kept doing even through the week that landed him in the hospital. He presented to the emergency department with several days of increasing thirst, frequent urination, and abdominal pain, found to have a glucose of 587 mg/dL and a significant anion-gap metabolic acidosis with large ketones — diabetic ketoacidosis, treated with the standard insulin infusion protocol and transitioned to subcutaneous basal-bolus insulin as he stabilized. He had no prior diagnosis of diabetes, a BMI of 34, and no family history he's aware of on either side.
What complicates the simple "he's now insulin-dependent" read of a DKA presentation is his full picture on closer review: his GAD antibody testing came back negative, and his phenotype — middle-aged, obese, no prior symptoms until an acute, severe presentation — matches a recognized pattern distinct from classic autoimmune type 1 diabetes, sometimes called ketosis-prone or "Flatbush" diabetes. In that pattern, the DKA at presentation is often driven by severe glucotoxicity temporarily suppressing beta-cell insulin secretion, rather than permanent autoimmune beta-cell destruction — and a meaningful share of these patients regain enough functional beta-cell capacity, once glucose control is achieved, to come off insulin entirely and be managed like typical type 2 diabetes going forward.
A DKA presentation that may not mean what it usually means
I'd keep him on insulin indefinitely and not chase the discontinuation question yet. He presented in real, severe DKA — that's a marker of significant insulin deficiency, whatever the long-term etiology turns out to be. A recurrent DKA episode from stopping insulin too soon is a serious, avoidable harm against a regimen that's already working.
His full picture is genuinely more specific than "presented in DKA." Negative GAD antibodies, middle-aged onset, obesity, no prior symptoms until an acute severe presentation — that's a recognized pattern, ketosis-prone or Flatbush diabetes, distinct from autoimmune type 1. Umpierrez and colleagues, who characterized this Flatbush pattern in the first place, support reassessing beta-cell function once glucotoxicity has resolved, because a real share of these patients recover enough function to come off insulin.
I'm not proposing we stop insulin today, or without a real reassessment — I'm proposing we don't default to "insulin forever" just because that's the conservative-sounding answer, when his actual phenotype points somewhere more specific and more hopeful.
I don't think the two of you actually disagree on what should happen, just on how it's framed. I'd support the reassessment and, if his C-peptide recovers, a real discontinuation trial — but I want explicit safety-netting built in from day one, not assumed.
That means direct, concrete DKA warning-sign education before discharge, close early follow-up rather than a routine interval, and a clear, written plan for resuming insulin immediately if anything looks off. Whether he turns out to have KPD or not, a discontinuation attempt should never quietly become an unmonitored gamble.
Insulin continued at discharge, C-peptide reassessment scheduled for three months once good control is established, and a written DKA safety-net plan reviewed with Terrence before he left the hospital, including exactly when to call and when to go to the emergency department directly.
Not agreed: whether the three-month timeline is long enough to let glucotoxicity fully resolve before testing, or whether it should be pushed to four or six months for a more reliable result. The endocrinologist felt three months was consistent with the KPD literature's own timing; the hospitalist would have preferred a longer, more conservative interval. Left for the outpatient follow-up to finalize.