Nomogram Models to Predict Postoperative Hyperlactatemia in Patients Undergoing Elective Cardiac Surgery

Figures

• 

  Figure 1. Flow chart of the study. POHL, postoperative hyperlactatemia.

• 

  Figure 2. Nomogram for the prediction of postoperative hyperlactatemia after elective cardiac surgery. A true patient is shown to illustrate how to use the nomogram. This is a 65-year-old patient who has experienced his first-time cardiac surgery, with a history of renal insufficiency and an LVEF of 58%. The CPB time is 103 min and the volume of intraoperative RBC transfusion is one unit. The individual item score corresponding to each factor is presented at the top, and the total points are obtained from the sum of the scores corresponding to each factor by a red dot. Based on the given values of the six predictors, the patient can be intuitively mapped onto the nomogram. It can be clearly seen from the nomogram that the total points of this patient are 215 points and the corresponding probability of POHL is 0.153. CPB, cardiopulmonary bypass; LVEF, left ventricular ejection fraction; RBC, red blood cell.

• 

  Figure 3. Assessment and validation of the full nomogram model. (A) calibration plot in the training set; (B) calibration plot in the validation set; (C) ROC curves in both the training and validation sets; (D) decision curves in both the training and validation sets; (E) clinical impact curve in the training set; and (F) clinical impact curve in the validation set. AUC, area under the receiver operating characteristic curve; CI, confidence interval; ROC curve, receiver operating characteristic curve.

• 

  Figure 4. Development, assessment, and validation of the preoperative nomogram model. (A) Preoperative nomogram predicting the development of postoperative hyperlactatemia in patients undergoing elective cardiac surgery; (B) calibration plot in the training set; (C) calibration plot in the validation set; (D) ROC curves in both the training and validation sets; (E) decision curves in both the training and validation sets; (F) clinical impact curve in the training set; (G) clinical impact curve in the validation set. AUC, area under the receiver operating characteristic curve; CI, confidence interval; LVEF, left ventricular ejection fraction; ROC curve, receiver operating characteristic curve.

• 

  Figure 5. Observed and predicted probabilities of postoperative hyperlactatemia by risk intervals. No significant difference was found among the predicted probabilities and the observed probabilities in the training and validation sets in the same risk interval (p > 0.05), indicating good agreement. However, significant differences were found among different risk intervals (p <0.001), indicating good classification.

• 

  Figure 6. Kaplan–Meier curves showing the cumulative mortality of all the patients who underwent elective cardiac surgery (A), and the comparison of cumulative mortality in patients with and without POHL (B). POHL, postoperative hyperlactatemia.