Fig. 6. Regression plots between transcriptional phenotypes and postoperative cardiac function. Multivariate analysis including the explanatory variables peroxisome proliferator activated receptor γ coactivator-1α (PGC-1α) pathway (an index of long-term energy metabolism) and the anesthetic-induced short-term changes in gene expression was used to determine the best predictors of postoperative cardiac function. (  A ) Perioperative changes in N-terminal pro brain natriuretic peptide (NT-proBNP) were best predicted by changes in fatty acid (FA) oxidation pathway activity. (  B ) PGC-1α pathway status best predicted perioperative changes in cardiac index (CI), whereas DNA-damage signaling best predicted postoperative CI (  C ). Granulocyte colony-stimulating factor (G-CSF) survival signaling best predicted postoperative early diastolic (ED) function measured at the septum (  D ), the mitral annulus (  E ), and the lateral wall (  F ). PROP = propofol patients; SEVO = sevoflurane patients.