Less stringent perioperative glucose control improves outcomes after coronary artery bypass graft in patients with insulin dependent diabetes mellitus

Abstract

Abstract Objective After coronary artery bypass grafting (CABG), studies testing strict glycemic control did not took into account previous insulin use. Our aim was to evaluate the effect of perioperative (PO) control in patients with NonDM (ND), noninsulin- (NIDDM) and insulin-dependent (IDDM) diabetes submitted to CABG. Methods We included CABG patients (2010–2017), stratified by DM presence and type. PO glucose was managed with a standardized protocol (IV insulin targeting a glucose <180 mg/dL. Endpoints included in-hospital death, major adverse cardiac and cerebrovascular (MACCE) events, deep sternal wound infection [DSWI], and major infections. PO glucose control was stratified in 4 quartiles using the highest glucose level: 1st, <140mg/dL; 2nd, 140–179mg/dL; 3rd, 180–239mg/dL; and 4th, >240mg/dL). Multivariable analysis was used to model the association of PO glucose control with outcomes, taking the presence and type of DM into account. All analysis on SPSS 23 (SPSS, IBM, 2018), with p<0.05 as statistically significant. Results 2020 CABG pts, age 64±10 y, 31% female, DM 37,5% (66% NIDDM, 34% IDDM). Compared with DM, neither NIDDM nor IDDM were associated with mortality (OR=0.99 [0.61–1.61], p=0.985) or MACCE (OR=0.12 [0.81–1.54], p=0.477). In multivariable analysis, DM (OR=1.58 [1.01–2.49], p=0.049) and IDDM (OR=2.55 [1.57–4.16], p<0.001) were independently associated with DSWI. In a mediation analysis, the inclusion of IDDM in the model, weakened the association of DM with DSWI into nonsignificance (OR=1.20 [0.73–1.98], p=0.456), while IDDM retained significant association with DSWI (OR=2.37 [1.39–4.01], p=0.001), indicating that the impact of DM on the incidence of DSWI is fully explained by the IDDM patients. Poor PO glycemic control was an independent predictor of death (OR=1.09, 95% CI 1.04–1.13, p<0.001; 9% higher mortality to each 10mg/dL increase). Accordingly, patients from ND group in the highest stratum of glucose control (4th quartile), compared with the other 3 quartiles, showed increased mortality (3.8%, 2.9%, 4.4%, and 17.9%, p<0.001) and MACCE (8.7%, 8.5%, 14.3%, and 39.3%, p<0.001). However, in IDDM patients, poor PO glycemic control was not an independent predictor of mortality (OR=3.34, 95% CI 0.34–38.23, p=0.347). In fact, strict PO glycemic control in IDDM patients (but not in ND and NIDDM) increased the incidence of DSWI (<140mg/dL, 19.8%, vs >240mg/dL, 8.2%, p=0.015) and other major infections (<140mg/dL, 26.2%, vs >240mg/dL, 12.2%, p=0.048). Conclusion After CABG, poor glycemic control is an independent predictor of mortality. However, in IDDM patients, PO hyperglycemia is not associated with mortality, and strict glycemic control increased DSWI, major infections and LOS in these patients. These results suggest that IDDM patients submitted to CABG may benefit from less stringent glycemic targets in PO management. Funding Acknowledgement Type of funding source: None