Does preoperative exposure to SGLT2 inhibitors reduce postoperative cardiovascular complications in adult patients with heart failure undergoing cardiac surgery with cardiopulmonary bypass?
Sodium-glucose cotransporter-2 inhibitors (SGLT2i) are now recommended in the management of chronic heart failure. 1, 2 Several randomized trials have reported a reduction in morbidity and mortality in heart failure patients treated with SGLT2i. 1 As a result, these agents are widely prescribed, presenting new challenges for the perioperative management of these treatments. 3 In particular, the impact of this therapeutic class on the postoperative outcomes of heart failure patients is unknown. 3 We aimed to describe the impact of SGLT2i on the cardiovascular outcomes of patients undergoing scheduled cardiac surgery. METHODS We performed a single-institution retrospective study in adult patients with heart failure and reduced ejection fraction, undergoing cardiac surgery with cardiopulmonary bypass (CPB). After institutional review board approval (RnIPH 2024-101, 29 September 2024), patients scheduled for cardiac surgery between January 2020 and December 2022 were screened. Obtaining informed consent was not necessary. We included patients over 18 years of age who underwent scheduled or semiurgent cardiac surgery requiring the use of CPB and had chronic heart failure with moderate to severe reduced left ventricular ejection fraction (LVEF ≤50%). We excluded patients who underwent emergency surgery, cardiac procedures performed without CPB, cardiac transplantation, or left ventricular assist device implantation, as well as those with existing ventricular assist devices or preoperative extracorporeal life support. To describe the postoperative outcomes, the cohort was divided into 2 groups based on preoperative exposure to SGLT2i: a group preoperatively treated with SGLT2i and a control group who was not exposed to SGLT2i. In the SGLT2i group, treatment was not discontinued before surgery and was not administered on the morning of surgery. During the inclusion period, treatments were not discontinued 72 hours before surgery, as the guidelines on the perioperative management of cardiovascular treatments had not yet been published. 4 The primary end point was a composite criterion aimed at describing the main postoperative cardiovascular complications: postoperative myocardial injury (PMI) defined by an increase in high-sensitivity troponin T to more than 70 times the upper reference limit (14 ng/L), as previously describe5; low cardiac output syndrome (LCOS) defined as the administration of inotropes for more than 48 hours or the need for mechanical circulatory support; and postoperative mortality 30 days after surgery. Cohorts were propensity score-matched in a 1: 1 ratio using predetermined variables that included: patients with LVEF <40%, chronic kidney disease, diabetic patients, and those undergoing preoperative levosimendan treatment. After matching, quantitative data were compared using the Mann-Whitney test, while qualitative data were analyzed using the χ2 test. All P-values were 2-tailed and considered statistically significant if <. 05. RESULTS Of the 247 eligible patients, 241 patients were included (Supplemental Digital Content, Supplemental Figure 1, https: //links. lww. com/AA/F429). After propensity score matching, all preoperative covariates were well-balanced between cohorts (Table 1; Supplemental Digital Content, Supplemental Figure 2, https: //links. lww. com/AA/F429 and Supplemental Digital Content, Supplemental Tables 1, 2 and 3, https: //links. lww. com/AA/F429). The cardiovascular complications described by our primary outcome measure was significantly lower in the SGLT2i group (36/86 (41. 9%) control group vs 23/86 (26. 7%) SGLT2i group; OR 0. 51 95% confidence interval CI, 0. 27–0. 96, P =. 04) (Table 2, Supplemental Digital Content, Supplemental Table 4, https: //links. lww. com/AA/F429). This reduction is primarily associated with the decrease in the incidence of PMI (36/86 (41. 9%) control group vs 13/86 (15. 1%) SGLT2i group; OR 0. 46 95% CI, 0. 22–0. 97, P =. 04). We did not observe any difference in either the incidence or the duration of administration of inotropes or vasopressors (Table 2). We observed a mean reduction in lactate levels on day 1 (1. 0 2. 3, P <. 0001) and day 2 (0. 2 0. 2, P =. 0009) in the group of patients receiving SGLT2 inhibitors. Table 1. - Baseline Patients Characteristics, After Propensity Score Matching Variables Total N = 172 Control group N = 86 SGLT2i group N = 86 P Value Preoperative data Age, y 64. 3 (9. 6) 64. 2 (9. 4) 64. 5 (9. 8). 74 Men, n (%) 149 (86. 6) 75 (87. 2) 74 (86. 0). 82 Weigh (kg) 81. 0 (14. 8) 82. 8 (14. 0) 79. 0 (14. 0). 08 Height (m) 1. 72 (0. 09) 1. 72 (0. 08) 1. 71 (0. 09). 32 BMI (kg/m2) 27. 5 (5. 0) 28. 0 (5. 2) 27. 1 (4. 7). 23 Euroscore II % 3. 1 (2. 3) 3. 0 (1. 5) 3. 2 (2. 8). 19 Arterial hypertension, n (%) 95 (55. 2) 52 (60. 5) 43 (50. 0). 17 Smoking current, n (%) 45 (43. 6) 27 (31. 4) 28 (32. 5). 87 Chronic occlusive arteriopathy, % (n) 25 (14. 5) 12 (14. 0) 13 (15. 1). 83 Diabetes (type 1 or 2), n (%) 76 (44. 2) 38 (44. 2) 38 (44. 2) 1. 00 COPD, n (%) 33 (19. 2) 19 (22. 1) 14 (16. 3). 33 Stroke, n (%) 10 (5. 8) 5 (5. 8) 5 (5. 8) 1. 00 LVEF % 41. 8 (10. 5) 41. 3 (8. 4) 42. 2 (12. 3). 92 LVEF <40%, n (%) 67 (39. 0) 31 (46. 4) 36 (41. 9). 43 Chronic renal failure, n (%) 88 (51. 2) 50 (58. 1) 38 (44. 2). 07 eGFR, mL/min 72 (23) 75 (23) 70 (22). 09 Preoperative hemoglobin (g/dL) 13. 3 (1. 7) 13. 3 (1. 6) 13. 2 (1. 7). 54 Preoperative iron administration, n (%) 6 (3. 5) 3 (3. 5) 3 (3. 5) 1. 00 Preoperative levosimendan administration, n (%) 18 (10. 5) 6 (7. 0) 12 (14). 13 Operative outcomes Type of surgery. 99 CABP, n (%) 85 (49. 4) 42 (48. 8) 43 (50. 0) Aortic valve replacement, n (%) 22 (12. 8) 11 (12. 8) 11 (12. 8) Mitral valve surgery, n (%) 15 (8. 7) 8 (9. 7) 7 (8. 1) CABP + aortic valve surgery, n (%) 22 (12. 8) 12 (14. 0) 10 (11. 6) Ascendant aortic surgery, n (%) 25 (14. 5) 12 (14. 0) 13 (15. 1) Other, n (%) 3 (1. 7) 1 (1. 2) 2 (2. 3) CPB time, min 96 (38) 101 (42) 91 (34). 12 Aortic cross clam time, min 67 (31) 70 (31) 64 (26). 32 CPB weaning time, min 22 (20) 25 (20) 20 (21). 005 Fluid resuscitation, mL 2444 (928) 2371 (1007) 2519 (841). 14 Quantitative data: mean (standard deviation). Comparison by Mann-Whitney test. Qualitative data: n = number of subjects (frequency in %). Comparison by χ2 test. Abbreviations: BMI, body mass index; COPD, chronic obstructive pulmonary disease; chronic renal failure (eGFR <60 mL/min/1. 73 m2) ; eGFR, estimated glomerular filtration rate; LVEF, left ventricular ejection fraction. Table 2. - Postoperative Adjusted Outcomes Control groupN = 86 SGLT2i groupN = 86 Odds ratio (95% CI) P Value Primary outcome Poor postoperative cardiovascular outcome 36 (41. 9) 23 (26. 7) 0. 51 (0. 27–0. 96). 04 LCOS, n (%) 21 (24. 1) 17 (19. 8) 0. 76 (0. 37–1. 56). 46 Inotrope 21 (24. 1) 17 (19. 8) 0. 76 (0. 37–1. 56). 46 ECLS 3 (3. 5) 4 (4. 6) 1. 35 (0. 32–5. 63). 70 Perioperative myocardial injury, n (%) 36 (41. 9) 13 (15. 1) 0. 46 (0. 22–0. 97). 04 Hospital death 4 (4. 5) 1 (1. 2) 0. 24 (0. 04–1. 57). 16 Secondary outcomes Vasopressor administration, n (%) 78 (89. 5) 76 (88. 3) 1. 14 (0. 43–3. 03). 80 Vasopressor duration, d 2. 3 (2. 4) 2. 7 (3. 3) -. 94 Inotrope administration, n (%) 27 (31. 4) 21 (24. 4) 0. 71 (0. 36–1. 37). 31 Inotrope duration, d 1. 3 (2. 6) 1. 0 (2. 4) -. 27 Total fluid resuscitation, mL 1683 (901) 1044 (575) - <. 0001 Troponin T hs, ng/L 1334 (3118) 847 (1419) -. 03 Lactatemia D0, mmol/L 1. 7 (0. 7) 1. 7 (1. 3) -. 12 Lactatemia D1, mmol/L 2. 7 (2. 3) 1. 7 (0. 9) - <. 0001 Lactatemia D2, mmol/L 1. 4 (0. 5) 1. 2 (0. 7) -. 0009 pH D0 7. 38 (0. 05) 7. 35 (0. 07) -. 004 pH D1 7. 36 (0. 07) 7. 38 (0. 04) -. 02 pH D2 7. 41 (0. 05) 7. 42 (0. 05) -. 19 Hemoglobin D1, g/dL 11. 2 (1. 9) 12. 1 (1. 8) -. 004 Transfusion, n (%) 25 (29. 0) 27 (31. 4) 1. 11 (0. 58–2. 13). 74 Pneumoniae, n (%) 12 (14. 0) 22 (25. 6) 2. 12 (0. 98–4. 57). 06 Atrial fibrillation, n (%) 28 (32. 6) 24 (25. 9) 0. 80 (0. 42–1. 53). 51 Acute renal failure, n (%) 43 (50. 0) 49 (57. 0) 1. 32 (0. 73–2. 40). 36 Dialysis, n (%) 3 (3. 5) 3 (3. 5) 1. 00 (0. 22–4. 54). 99 Delirium 10 (11. 6) 3 (3. 5) 0. 27 (0. 08–0. 96). 04 ICU length of stay, d 5. 9 (4. 7) 6. 5 (7. 1) -. 95 Quantitative data: mean (standard deviation). Comparison by Mann-Whitney test. Qualitative data: n = number of subjects (frequency in %). Comparison by χ2 test. Abbreviations: CPB, cardiopulmonary bypass; ECLS, extracorporeal life support; ICU, intensive care unit; LCOS, low cardiac output syndrome; pH, potential of hydrogen. DISCUSSION In this cohort study, the use of SGLT2i was associated with a lower incidence of several postoperative complications. The observed effect can be attributed to the reduction in PMI, since no significant changes were noted in the other elements of the composite end point. We find this result particularly relevant, as PMI are associated with increased 30-day postoperative morbidity and mortality. 5 The postoperative troponin peak reflects the extent of myocardial injury, primarily due to ischemia induced during CPB. 5 SGLT2i could therefore play a cardioprotective role in this context. We can hypothesize that myocardial damage would be reduced due to a shift in energy sources induced by SGLT2i during ischemia-reperfusion, favoring beta-hydroxybutyrate over glucose. 6, 7 This change, linked to increased glucagon and decreased insulin levels, enhances lipolysis and ketogenesis, leading to lower oxygen consumption and a decreased peak of troponin. 8 However, although a positive trend was observed, the reduction in PMI did not result in a statistically significant decrease in LOCS or mortality, which may be explained by the limited statistical power of our study to detect such outcomes. Nevertheless, our findings may help inform sample size estimations for future studies aimed at validating these hypotheses. Furthermore, our findings challenge the validity of the recommendations that advocate for the discontinuation of SGLT2i 72 hours before scheduled surgery due to the risk of euglycemic ketoacidosis (EDKA). 4, 9 These guidelines are based on the publication of several case reports that indicated that EDKA may occur after surgery in patients on SGLT2i. 3 However, the benefit-risk balance is poorly documented in the literature, and recent data question the need for preoperative discontinuation. 3 Indeed, recent evidence suggests that, in a medical context, discontinuation of SGLT2i is associated with an increased risk of cardiovascular death and hospitalization for heart failure within 90 days. 10 These findings are in line with our results and support the hypothesis that continuing SGLT2i may benefit patient outcomes. However, due to the retrospective nature of our study, we did not perform systematic perioperative ketone level measurements and are therefore unable to report the incidence of EDKA (including mild or asymptomatic forms). This limitation underscores the need for routine ketone monitoring. Future research should evaluate the risk of EDKA in relation to the risk of postoperative cardiovascular decompensation, as well as the potential cardiovascular benefits of SGLT2i continuation. In conclusion, our study highlights the positive impact of SGLT2i on postoperative outcomes after cardiac surgery, primarily driven by a reduction in PMI. Our results raise the question of whether SGLT2i should be continued or discontinued preoperatively in accordance with current guidelines and reinforce the need for randomized studies to evaluate the benefit-risk balance of these strategies. DISCLOSURES Conflicts of Interest: None. Funding: None. This manuscript was handled by: Girish P Joshi, MBBS, MD, FFARCSI.
Labaste et al. (Fri,) studied this question.