Effect of postoperative haemoglobin variation on major cardiopulmonary complications in high cardiac risk patients undergoing anatomical lung resections

OBJECTIVES: Recent evidence shows that permissive anaemia strategies are safe in different surgical settings. However, effects of variations in haemoglobin (Hb) levels could have a negative impact in high-risk patients. We investigated the combined effect of postoperative Hb concentration and cardiac risk status on major cardiopulmonary complications after anatomical lung resections. METHODS: We retrospectively analysed the records, collected in a prospective clinical database, of 154 consecutive patients undergoing anatomical lung resections at our institution (February 2017-February 2019). Hb levels were displayed as preoperative concentration, nadir Hb level before onset of complications and delta Hb (ΔHb). Cardiac risk was stratified according to the Thoracic Revised Cardiac Risk Index (ThRCRI). Univariable and multivariable logistic regression analyses were used to test the associations between patients, surgical variables and cardiopulmonary complications according to the European Society of Thoracic Surgeons definitions. RESULTS: Cardiopulmonary complications occurred in 63 patients (17%). In the fully adjusted multivariable model, higher values of ΔHb were associated with increased risk of complications [odds ratio (OR) 1.07; P < 0.001], along with higher ThRCRI classes (classes A-B versus C-D: OR 0.09; P < 0.001). Interaction terms with transfusion were not statistically significant, indicating that the harmful effect of ΔHb was independent. According to receiver operating characteristic curve analysis, a ΔHb of 29 g/l was found to be the best cut-off value for predicting complications. CONCLUSIONS: In our series, ΔHb, rather than nadir Hb, was associated with an increased risk of complications, particularly in patients with higher cardiac risk. Restrictive transfusion strategies should be carefully applied in patients undergoing lung resections and balanced according to individual clinical status