CLINICAL CONSEQUENCES OF INSULIN DEPRESCRIBING IN TYPE 2 DIABETES

INSIGHTS FROM A DISTRICT HOSPITAL IN MALAYSIA

Abstract

INTRODUCTION
Type 2 diabetes (T2D) is characterised by insulin resistance and progressive beta-cell dysfunction, leading to failure on oral glucose-lowering drugs (OGLDs). Insulin deprescribing requires individualisation, considering factors like residual beta-cell function, disease duration and insulin dosage. This study explores the characteristics and outcomes of patients with T2D who had insulin deprescribed at a local health clinic before hospitalisation.

METHODOLOGY
We performed a retrospective medical records review involving adult patients with T2D admitted to Hospital Jempol from September 2024 to April 2025. These patients had their insulin deprescribed within 12 months prior to admission. Basic demographics, Charlson Comorbidity Index (CCI), baseline HbA1c, admission random blood glucose (RBS), interval from insulin discontinuation to admission, length of stay (LOS) in the ward, insulin deprescribing success (OGLDs maintained at discharge) and factors favoring deprescribing were explored.

RESULT
Among 14 patients with a median age of 62.6 years (range: 41-78), 85.7% had a CCI ≥3. All had OGLDs with insulin (basal bolus 71.4% or premixed 28.6%), with median total daily dose (TDD) of 57 units (range: 28-98) pre-admission. Duration of insulin discontinuation to admission was 3-31 weeks. Median baseline HbA1c was 11.5% (range: 6.4-14%), while median admission RBS was 19.6 mmol/L (range: 6.8 – 31 mmol/L). Infections accounted for 57% of them, in which 5 were complicated (2 DKA, 1 respiratory failure, 1 gram-negative bacteremia, 1 septic shock and intubated). Other indications for admission included symptomatic hyperglycemia (28.6%), decompensated heart failure (7.1%), and hypertensive emergency (7.1%). The median LOS was 4 days (range: 1–11). 28.6% transitioned successfully to OGLDs while 71.4% resumed de-intensified insulin regime. Successful deprescribing was noted in older patients (median: 68 vs 64; p-value 0.178), patients with lower baseline HbA1c (median: 8.7 vs 12; p-value 0.288) and higher RBS (median: 20.4 vs 18.2 mmol/L, p-value=1.00).

CONCLUSION
Although statistically insignificant, lower HbA1c may favour deprescribing success. These preliminary trends may inform future studies on safer deprescribing practices to prevent adverse outcomes and hospitalisations.