Quantification of stress-induced hyperglycaemia associated with key diagnostic categories using the stress hyperglycaemia ratio.

Abstract

Stress-induced hyperglycaemia (SIH) is the acute increase from preadmission glycaemia and is associated with poor outcomes. Early recognition of SIH and subsequent blood glucose (BG) management improves outcomes, but the degree of SIH provoked by distinct diagnostic categories remains unknown. Quantification of SIH is now possible using the stress hyperglycaemia ratio (SHR), which measures the proportional change from preadmission glycaemia, based on haemoglobin A1c (HbA1c ). We identified eligible patients for eight medical (n= 892) and eight surgical (n= 347) categories. Maximum BG from the first 24 h of admission for medical, or postoperatively for surgical patients was used to calculate SHR. Analysis of variance indicated differing SHR and BG within both the medical (p< 0.0001 for both) and surgical cohort (p< 0.0001 for both). Diagnostic categories were associated with signature levels of SHR that varied between groups. Medically, SHR was greatest for ST-elevation myocardial infarction (1.22 ± 0.33) and sepsis (1.37 ± 0.43). Surgically, SHR was greatest for colectomy (1.62 ± 0.48) and cardiac surgeries (coronary artery graft 1.56 ± 0.43, aortic valve replacement 1.71 ± 0.33, and mitral valve replacement 1.75 ± 0.34). SHR values remained independent of HbA1c , with no difference for those with HbA1c above or below 6.5% (p> 0.11 for each). BG however was highly dependent on HbA1c , invariably elevated in those with HbA1c ≥ 6.5% (p < 0.001 for each), and unreliably reflected SIH. The acute stress response associated with various medical and surgical categories is associated with signature levels of SIH. Those with higher expected SHR are more likely to benefit from early SIH management, especially major surgery, which induced SIH typically 40% greater than medical cohorts. SHR equally recognised the acute change in BG from baseline across the full HbA1c spectrum while BG did not and poorly reflected SIH.