Psychiatric and Substance Use Disorders and Their Association With Clinical Outcomes in Diabetic Ketoacidosis and Hyperosmolar Hyperglycemic Syndrome

Background and Objective

Diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic state (HHS) are life-threatening conditions that send nearly 180,000 patients to the intensive care unit each year, with mortality rates up to 5–10%. Little is known about the impact of concurrent psychiatric disorders on specific DKA/HHS outcomes. Identifying these relationships offers opportunities to improve clinical management, treatment planning, and mitigate associated morbidity and mortality.

Methods

We conducted a retrospective review including adult DKA/HHS admissions within a large Massachusetts hospital system from 2010 to 2019. We identified patients admitted inpatient for DKA or HHS, then filtered by International Classification of Disease-9-CM and International Classification of Disease-10-CM codes for psychiatric diagnoses that were present in patients electronic medical record at any point in this observational period. Outcomes included the number of inpatient admissions for DKA/HHS, age of death, rates of discharging against medical advice (AMA) from any inpatient admission, and end-stage renal disease/dialysis status. Multivariate regression was conducted using R software to control for variables across patients and evaluate relationships between outcomes and concurrent psychiatric disorders. Significance was set at P < 0.05.

Results

Seven thousand seven hundred fifty-six patients were admitted for DKA or HHS, 66.9% of whom had a concurrent psychiatric disorder. Of these patients, 54.5% were male, 70.4% were White, and they had an average age of 61.6 years. This compares with 26.1% with concurrent psychiatric condition within the general diabetes population, 52.1% of whom were male, 72.1% were White, and an average age of 68.2 years. A concurrent psychiatric disorder was associated with increased odds of rehospitalization (adjusted odds ratio [aOR] = 1.62 95% confidence interval [CI] 1.35–1.95, P < 0.001), of being diagnosed with end-stage renal disease and on dialysis (aOR = 1.02 95% CI 1.002–1.035, P = 0.02), and of leaving AMA (aOR = 6.44 95% CI 4.46–9.63, P < 0.001). The average age of death for those with a concurrent psychiatric disorder had an adjusted mean difference in years of −7.5 years (95% CI −9.3 to 5.8) compared to those without a psychiatric disorder.

Conclusions

Of patients with DKA/HHS, 66.9% have a concurrent psychiatric disorder. Patients with a concurrent psychiatric disorder admitted for DKA/HHS were more likely to have multiple admissions, to leave AMA, to be on renal dialysis, and to have a lower age of mortality.