Recurrent diabetic ketoacidosis burden and precipitant patterns in established type 1 diabetes using continuous...

Diabetic ketoacidosis (DKA) remains a major acute complication in people with established type 1 diabetes, but whether recurrent DKA burden and precipitant patterns differ by insulin-delivery modality is not well defined. We compared recurrent DKA burden and precipitant patterns between continuous subcutaneous insulin infusion (CSII) and multiple daily injections (MDI), and evaluated severity and short-term outcomes of index eligible admissions.

We retrospectively screened DKA hospitalizations in patients with type 1 diabetes between 2014 and 2024. Admissions at initial type 1 diabetes diagnosis, those involving preadmission insulin regimens other than CSII or basal-bolus MDI, and records insufficient to assess key variables or outcomes were excluded. Index-admission analyses used treatment modality at the earliest eligible hospitalization (CSII, n = 33; MDI, n = 95). Patient-level recurrence analyses used predominant modality across eligible hospitalizations (CSII, n = 35; MDI, n = 93). Recurrent DKA was defined as at least two eligible admissions. Logistic regression was used for adjusted analyses.

The cohort comprised 180 eligible DKA hospitalizations in 128 patients. Recurrent admissions accounted for 52/180 hospitalizations (28.9%), and 30/128 patients (23.4%) experienced recurrent DKA. In the primary patient-level analysis, recurrent DKA was more frequent in CSII than in MDI (13/35 [37.1%] vs. 17/93 [18.3%]); the adjusted odds ratio was 2.64 (95% confidence interval 1.07-6.55; P = 0.036). In index admissions, hospital stay was shorter in CSII than in MDI (7.0 [5.0, 9.0] vs. 9.0 [7.0, 11.0] days; P = 0.006). Among adults, CSII was independently associated with lower odds of severe DKA (adjusted odds ratio 0.30, 95% confidence interval 0.11-0.77; P = 0.012). Recorded precipitating factors were summarized descriptively. Among recurrent CSII admissions, device/infusion-related failure was recorded in 7/30 admissions (23.3%); among recurrent MDI admissions, infection/intercurrent illness and insulin omission/insufficiency were recorded in 10/22 (45.5%) and 6/22 (27.3%) admissions, respectively.

Among patients with established type 1 diabetes who had experienced at least one DKA hospitalization, recurrent DKA was common. Recurrence was more frequent among patients classified as CSII users in the primary analysis, although the strength and statistical significance of this association varied across sensitivity analyses. Recorded precipitating factors suggested treatment-specific clinical contexts and should be interpreted descriptively.

Not applicable.