Metabolomic Profiling Reveals Distinct Patterns of Myocardial Substrate Use in Humans With Coronary Artery Disease or Left Ventricular Dysfunction During Surgical Ischemia/Reperfusion
Background— Human myocardial metabolism has been incompletely characterized in the setting of surgical cardioplegic arrest and ischemia/reperfusion. Furthermore, the effect of preexisting ventricular state on ischemia-induced metabolic derangements has not been established.
Methods and Results— We applied a mass spectrometry–based platform to profile 63 intermediary metabolites in serial paired peripheral arterial and coronary sinus blood effluents obtained from 37 patients undergoing cardiac surgery, stratified by presence of coronary artery disease and left ventricular dysfunction. The myocardium was a net user of a number of fuel substrates before ischemia, with significant differences between patients with and without coronary artery disease. After reperfusion, significantly lower extraction ratios of most substrates were found, as well as significant release of 2 specific acylcarnitine species, acetylcarnitine and 3-hydroxybutyryl-carnitine. These changes were especially evident in patients with impaired ventricular function, who exhibited profound limitations in extraction of all forms of metabolic fuels. Principal component analysis highlighted several metabolic groupings as potentially important in the postoperative clinical course.
Conclusions— The preexisting ventricular state is associated with significant differences in myocardial fuel uptake at baseline and after ischemia/reperfusion. The dysfunctional ventricle is characterized by global suppression of metabolic fuel uptake and limited myocardial metabolic reserve and flexibility after global ischemia/reperfusion stress in the setting of cardiac surgery. Altered metabolic profiles after ischemia/reperfusion are associated with postoperative hemodynamic course and suggest a role for perioperative metabolic monitoring and targeted optimization in cardiac surgical patients.
Received August 25, 2008; accepted January 9, 2009.
Ischemia/reperfusion (I/R) is a well-recognized mechanism of injury after restoration of myocardial blood flow, either regionally in acute coronary syndromes or globally in the setting of cardiac surgery with cardioplegic arrest. Although I/R has characteristic thrombotic and inflammatory components, elucidating its poorly understood metabolic consequences is becoming feasible through the application of novel metabolomic technologies. Such high-throughput approaches have already identified metabolic status alterations in common cardiac disorders, including myocardial ischemia,1 atrial fibrillation,2 and heart failure.3 Furthermore, because current intraoperative cardioprotective strategies do not invariably prevent contractile dysfunction, myocardial apoptosis, or myocardial necrosis, perioperative metabolic management and substrate optimization for patients undergoing cardiac surgery continue to be important.
Editorial p 1700
Clinical Perspective p 1746
At baseline, the heart is known to use a diverse set of fuel substrates, including lactate, glucose, amino acids, ketones, and particularly free fatty acids (FFA). Studies in animals and humans have demonstrated significant alterations in myocardial substrate extraction and oxidation with I/R. These changes are potential targets of therapeutic modulation in the failing heart and acute ventricular dysfunction commonly seen after cardiac surgery.
Significant gaps remain in our understanding of ischemia-induced myocardial metabolic derangements. Little is known about the metabolic responses of the dysfunctional ventricle to global I/R in humans because previous studies have included only low-risk patients with normal ventricular function. Existing reports of post-I/R changes in myocardial substrate use have been limited to only a few metabolites (eg, glucose, FFA, and lactate). Recently, mass spectrometry–based methods for targeted metabolic profiling have enabled comprehensive analyses of changes in metabolic fuel selection in a variety of cellular and animal models.4–6 We hypothesized that among patients undergoing cardiac surgery with cardioplegic arrest, the underlying myocardial pathology, in terms of coexisting coronary artery disease (CAD) or left ventricular (LV) dysfunction (LVD), would be associated with different myocardial metabolic phenotypes at baseline and early after global I/R. This hypothesis was tested using targeted metabolic analysis to measure transmyocardial arteriovenous gradients in a wide array of metabolites before and after I/R.
A total of 37 consecutive patients undergoing cardiac surgery with cardioplegic arrest were included in this study. The Institutional Review Board approved the protocol, and all patients provided written informed consent. Eligibility criteria were age ≥18 years and admission for an elective cardiac surgical procedure with planned coronary sinus (CS) catheter placement for administration of retrograde cardioplegia. Patients with end-stage renal or liver disease, ongoing infection, and long-term immunosuppression were excluded. On the basis of preoperative cardiac catheterization and echocardiographic assessment of LV function, the study population was divided into 3 groups as follows: (1) CAD, defined as luminal stenosis of >50% in any major epicardial coronary artery, history of coronary revascularization, or myocardial infarction (n=12); (2) LVD, defined as an LV ejection fraction <45% (n=10); and (3) control subjects, patients with angiographically normal coronary arteries and normal LV systolic function (n=17). Two patients had both CAD and LVD and are included in both groups.
Study Protocol and Sample Collection
All patients underwent an overnight fast and standardized anesthetic and perfusion management. Surgical approach and cardioprotective management were at the discretion of the attending surgeons, with intermittent antegrade and retrograde cold blood cardioplegia used in all cases. The composition of cardioplegia solution at our institution is presented in Table I of the online-only Data Supplement. Glucose-insulin-potassium infusions were not used in any of the study patients. However, all patients received insulin infusions titrated as required to maintain intraoperative normoglycemia and continued into the intensive care unit per institutional protocol.
Immediately after cardiopulmonary bypass (CPB) was established, the CS catheter was inserted and its correct position was verified echocardiographically. Paired arterial and CS blood samples were collected simultaneously at baseline before aortic cross-clamp (AXC) placement (pre-I/R) and 10 minutes after removal of AXC and myocardial reperfusion (post-I/R). Samples were drawn slowly to avoid hemolysis and contamination with right atrial blood into EDTA- and protease inhibitor–treated tubes (BD P100, BD Diagnostics, Franklin Lakes, NJ). The first 10 mL from the CS catheter was discarded to avoid contamination from saline, residual cardioplegia, or stagnant blood. All samples were immediately centrifuged at 4°C and 5000 rpm for 5 minutes to remove cellular elements, and plasma aliquots were stored at −80°C until analysis.
Plasma Metabolite Analyses
Plasma concentrations of 63 metabolites were measured, including 6 conventional metabolic substrates (glucose, lactate, pyruvate, total FFA, total ketones, and 3-hydroxybutyrate), 12 amino acids, and 45 acylcarnitine derivatives. Specific metabolites are listed in online-only Data Supplement Table II.
Plasma glucose and lactate concentrations were determined with kits from Roche Diagnostics (Indianapolis, Ind); total FFA (nonesterified) and ketones (total and 3-hydroxybutyrate) were determined with kits from Wako (Richmond, Va). Plasma pyruvate was measured with a commercially available assay from Biovision (Mountain View, Calif).
Plasma acylcarnitines and amino acids were measured with flow injection tandem mass spectometry (Micromass Quattro Micro TM instrument, Waters Corp, Milford, Mass). Sample preparation and quantification by stable isotope dilution were performed as previously described.4,5,7,8 Additional methods relative to amino acid measurements are described in the online-only Data Supplement, Methods 1.
Transmyocardial extractions were determined by subtracting CS concentrations of each analyte from the arterial concentrations. Transmyocardial extraction ratios were calculated by the following formula: extraction ratio (%)=(CS concentration−arterial concentration)/(arterial concentration)×100.
Perioperative Assessment of Ventricular Function
Standard transesophageal echocardiographic images were obtained before and after CPB with a Philips iE33 system and analyzed posthoc by an investigator blinded to the metabolomic data. LV end-diastolic area (LVEDA) and end-systolic area (LVESA) were measured in the transgastric short-axis midpapillary view, and LV fractional area change (FAC; %) was calculated as follows: FAC=(LVEDA−LVESA)/LVEDA×100%.
Serial cardiac outputs measurements were obtained intraoperatively after separation from CPB and in the intensive care unit (ICU) with pulmonary artery thermodilution. The average cardiac output over the first 4 postoperative hours after admission to the ICU was recorded to avoid potential biases from single measurements.
Duration and cumulative dose of perioperative inotropic use (within the first hour after separation from CPB and total postoperative course) were collected. Inotropic support, defined as epinephrine infusions at any dose or dopamine infusions >4 μg · kg−1 · min−1, was initiated to facilitate weaning from CPB or to maintain a cardiac index >2 L · min−1 · m−2 and a mean arterial pressure ≥50 mm Hg per institutional protocol.
Clinical, procedural, and functional data are presented as mean±SD or percentages as appropriate. Differences between patient groups were evaluated by χ2 test for discrete clinical variables and by t test for continuous variables. Values for all plasma metabolites are reported as median (quartiles 1 and 3). Differences in paired arterial and CS concentrations of each analyte and differences in paired extraction ratios before versus after I/R within patient groups were evaluated by Wilcoxon signed-rank test. To assess the differences in myocardial metabolic substrate extraction ratios between study groups, the Mann–Whitney U test was used. Tests of correlation were performed with the Pearson correlation (r) or Spearman rank-order test (rs), depending on the satisfaction of the normality assumption. The 2-sided Fisher exact test was used to determine differences in frequency distributions. Statistical significance was defined as a value of P<0.05.
Exploratory Principal Component Analysis of Metabolomic Data
To reduce the dimensionality of the data set while retaining as much of the variance as possible, an exploratory principal component (PC) analysis (PCA) of the correlation matrix of fuel substrate transmyocardial extraction ratios was performed. The metabolic PCs, defined as linear, uncorrelated (orthogonal) combinations of the original metabolites, were ordered according to their decreasing ability to explain variance in the original metabolic panel. Additional details on the PCA are described in the online-only Data Supplement, Methods 2. All analyses were performed with SAS 9.1.3 (SAS Institute Inc, Cary, NC).
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Patient characteristics, LV functional data, and procedural variables are given in Table 1. Control patients were younger than those with CAD or LVD and had fewer cardiovascular risk factors. As expected, patients with LVD had lower FAC measures and greater LV end-diastolic diameter than the other groups. The mean duration of AXC time did not vary significantly between groups (P=0.98 for LVD versus controls, P=0.72 for CAD versus controls).
Preischemia Myocardial Metabolic Substrate Use
Analysis of the entire cohort of 37 subjects at baseline, before AXC placement, revealed that glucose, lactate, FFA, total ketones, 3-hydroxybutyrate, pyruvate, leucine/isoleucine, and glutamate/glutamine were all present at lower concentrations in CS than in arterial samples, indicative of net myocardial uptake (Table 2). Conversely, alanine concentrations were higher in CS than in arterial samples, suggesting net release of this intermediate. No significant extraction of the other amino acids was observed.
No significant differences in extraction ratios of any of the measured metabolites were found at baseline between LVD and control patients. In contrast, compared with control patients, CAD patients had lower extraction and extraction ratios of FFA (60 versus 130 μmol/L and 11.7% versus 25.5%, respectively), lactate (65 versus 240 mmol/L and 7.7% versus 23.2%), and glutamate/glutamine (29 versus 24 μmol/L and 32.9% versus 46.7%) and a greater release of alanine (−34 versus −9 μmol/L and −10.0% versus −4.0%; Table 3).
Postischemia Myocardial Metabolic Substrate Use
In the entire cohort of 37 subjects after reperfusion, the myocardium on average continued to extract FFA, total ketones, 3-hydroxybutyrate, and glutamate/glutamine and to release alanine but now exhibited little to no net extraction of glucose, lactate, or leucine/isoleucine (Table 2). No significant correlations between post-I/R uptake of any of the measured metabolites and either the duration of AXC or the indexed cardioplegia volume (total cardioplegia volume/total AXC time) were found. Post-I/R extraction gradients stratified by CAD, LVD, and control subgroups are presented in Table 4.
Table 5 summarizes median changes in substrate extraction after I/R compared with the preischemic baseline in the CAD, LVD, and control subgroups. Control subjects exhibited a decrease in extraction of almost all fuels after I/R compared with baseline. In the postischemic phase, CAD subjects had a trend toward decreased extraction of all substrates and a rise in alanine compared with control subjects, whereas LVD patients demonstrated even greater reductions in extraction of a number of fuel substrates than either of the other 2 groups.
A reliance on anaerobic myocardial metabolism, defined by net lactate elution rather than uptake, was found after AXC release in 49% of patients. Among LVD patients, 90% had elevated anaerobic metabolism after reperfusion compared with only 36% in those with normal LV function (P=0.008). Postreperfusion lactate extraction was significantly correlated with the ability to extract other fuel substrates, including FFA (r=0.54, P=0.0009), glucose (r=0.60, P=0.0001), pyruvate (r=0.61, P<0.0001), total ketones (r=0.71, P<0.0001), and 3-hydroxybutyrate (r=0.70, P<0.0001), and the release of alanine (r=0.49, P=0.003) but not glutamate/glutamine (r=0.22, P=0.25).
Plasma Acylcarnitine Measurements
At baseline, arterial and CS acylcarnitine levels were not different (online-only Data Supplement Table III). In contrast, analysis of the entire cohort of 37 subjects after I/R revealed a significant net elution of acetylcarnitine (median arterial-CS difference, −0.29 μmol/L [quartiles 1 and 3, −1.21 and 0.15 μmol/L]; P<0.02) and β-hydroxybutyryl-carnitine (median arterial-CS difference, −0.02 μmol/L [quartiles 1 and 3, −0.05 and −0.01 μmol/L]; P<0.0001), both significantly different from baseline (P=0.04 for acetylcarnitine, P<0.0001 for β-hydroxybutyryl-carnitine). Transmyocardial elution of acetylcarnitine after ischemia was 4-fold higher in patients with CAD (median elution, 1.00 μmol) and LVD (median elution, 0.99 μmol) than in control patients (median elution, 0.26 μmol), although these differences did not achieve statistical significance.
Clinical Correlates of Postreperfusion Metabolism
Postprocedure cardiac outputs were recorded in 27 of 37 patients. No significant differences were found in mean cardiac output after termination of CPB or over the first 4 ICU hours between patients displaying postreperfusion aerobic versus anaerobic myocardial metabolism (6.1±1.7 versus 6.0±1.8 L/min, P=0.92; and 5.6±1.5 versus 6.5±2.1 L/min, P=0.21, respectively). However, patients with postreperfusion anaerobic metabolism more often required inotropic support immediately after CPB (78% versus 41%; P<0.05) and in the ICU (72% versus 35%; P<0.05) than patients with continued lactate extraction. Furthermore, the median duration of inotrope infusions was longer among patients with post-I/R net anaerobic metabolism (9 hours [quartiles 1 and 3, 2 and 10 hours] versus 0 hours [quartiles 1 and 3, 0 and 4 hours); P=0.02; Figure 1).
Factor Analysis of Postreperfusion Metabolism
PCA of postreperfusion transmyocardial metabolite gradients included the amino acids, fuel substrates, and acylcarnitines listed in online-only Data Supplement Table II. Nine factors were identified, explaining 74% of the total variance in these concentration gradients. Features of the top 3 PCs are shown in Table 6.
The first PC (PC1; substrate uptake) was composed mainly of metabolic fuel substrate (glucose, lactate, FFA, pyruvate, and ketone) extraction ratios, with an additional contribution by acetylcarnitine. PC2 (short- versus medium-chain acylcarnitines) featured an inverse relationship between the 8- and 10-carbon medium-chain acylcarnitines and the 5-carbon isobars (2- and 3-methylbutyryl carnitine) and the 4-carbon acylcarnitines (butyryl/isobutyryl carnitine and β-hydroxybutyryl-carnitine). PC3 (amino acid catabolism) was made up of the isobaric methylmalonyl- and succinyl-carnitines, adiponyl-carnitine, lactate, pyruvate, alanine, and glutamate/glutamine.
Patient-derived PC1 scores were modestly correlated with preoperative, but not postoperative, fractional shortening (r=0.36, P<0.04) and FAC measurements (r=0.41, P<0.02) (Figure 2). Similarly, mean factor scores were significantly different between the control (0.50±0.64) and the LVD (−0.88±1.19; P=0.0009) and CAD (−0.21±0.86; P<0.03) cohorts. Lower PC1 scores correlated with higher cumulative inotrope dose in the first hour after separation from CPB (rs=−0.37, P<0.04). Furthermore, patients requiring inotropes in the ICU tended to have lower PC1 scores (0.19 [interquartile range, −0.98 to 0.42] versus 0.41 [interquartile range, −0.15 to 0.84]; P=0.09).
Higher PC2 scores correlated with increased cardiac outputs after CPB (r=0.54, P=0.007) and over the first 4 ICU hours (r=0.51, P=0.008) (online-only Data Supplement Figure I). In quartile analysis, the highest cardiac outputs were found in the highest group of PC2 scores (Figure 3). PC3 scores were not associated with either inotrope use or cardiac output measures.
This study reports on myocardial metabolic substrate use before and after planned global I/R. Our findings advance our knowledge of human myocardial metabolism in several substantive ways. First, previous smaller studies on the effects of iatrogenic I/R on cardiac metabolism in humans have been limited to low-risk patients with normal LV function and across single procedure types (usually isolated coronary artery bypass graft surgery).9–12 Second, we were able to measure a significantly wider array of intermediary metabolites than previously attempted in any human study of this kind. These novel study features led to new insights into cardiac responses to I/R in general, as well as the influence of 2 major forms of cardiovascular disease, CAD and LVD.
Analysis of the complete cohort of 37 subjects after I/R revealed a clear decrease in glucose uptake, a trend to decreased fatty acid extraction, a clear switch from net lactate extraction to lactate release, a fall in ketone extraction (both total ketones and β-hydroxybutyrate), a decrease in glutamate/glutamine extraction, and increased alanine release. We also observed an increase in elution of 2 acylcarnitine species, acetylcarnitine and β-hydroxybutyryl-carnitine. These findings are consistent with a general decrease in oxidative fuel metabolism and a greater reliance on anaerobic metabolism of glucose for energy after I/R. The increase in lactate suggests that pyruvate derived from glycolysis is diverted away from the pyruvate dehydrogenase reaction and toward the lactate dehydrogenase reaction. Inhibition of pyruvate dehydrogenase may be facilitated by a rise in acetyl-CoA, which continues to be derived from fatty acid oxidation, but accumulates as a result of lowered trichloroacetic acid (TCA) cycle flux. Consistent with this idea is the nearly 8-fold increase in myocardial acetylcarnitine production across the entire study population. Acylcarnitines equilibrate rapidly with their cognate acyl-CoA intermediates, but unlike acyl-CoAs, acylcarnitines can be transported across both the mitochondrial and plasma membranes to appear in the circulation. Thus, transmyocardial plasma acylcarnitine gradients serve as a surrogate measure of myocardial mitochondrial acyl-CoA levels.
Along with acetylcarnitine, a significant myocardial elution of β-hydroxybutyryl-carnitine with reperfusion was also found. This metabolite reports on mitochondrial levels of β-hydroxybutyryl-CoA, an intermediate in ketone oxidation and product of fatty acid β-oxidation. Along with decreased ketone extraction after I/R, elution of β-hydroxybutyryl-carnitine suggests impaired oxidation of this substrate after reperfusion. Recent data from Mayr et al2 have highlighted an association between greater atrial tissue concentrations of ketogenic amino acids and β-hydroxybutyrate and persistent atrial fibrillation, suggesting that ketone metabolism may have an important role in cardiac metabolic pathology.
Impaired TCA cycle flux appears to occur after I/R in part through limiting levels of anaplerotic substrates, which are derived from catabolism of glucose and amino acids.13 The fall in GLX uptake after I/R observed in the present study is consistent with this idea because GLX would normally be transaminated to form the anaplerotic substrate α-ketoglutarate. However, we also noted a postreperfusion increase in alanine release. Net production of alanine likely occurs via transamination of pyruvate, with glutamate as the nitrogen donor in the alanine transaminase reaction. Taken together, this metabolic signature is consistent with impaired glucose oxidation in the postischemic period and resultant diversion of pyruvate away from the TCA cycle and into the LDH and alanine transaminase reactions. Although GLX is used for pyruvate-to-alanine transamination in the postischemic period, its net consumption falls, probably because less of the GLX pool is being oxidized via the glutamate dehydrogenase reaction or used for other transamination reactions. Consistent with this idea, leucine/isoleucine is also extracted less efficiently (catabolized) after I/R (Table 2).
Although results from our PCA should be taken as exploratory, we find that PC1, composed primarily of fuel substrate uptake and acetylcarnitine, explained more of the sample variance than any other factor and was associated, albeit modestly, with preoperative LV function and the need for postoperative inotropic support. This suggests that myocardial fuel extraction has a role in pre- and post-I/R myocardial function. PC2, comprised of small- and medium-chain acylcarnitines, was associated with higher postprocedural cardiac outputs, suggesting the importance of retaining the ability to continue β-oxidation of medium-chain fatty acids after I/R, seemingly at the expense of ketogenic substrates. A proposed schema of metabolic changes associated with I/R in the human heart is summarized in Figure 4.
Although our description of general metabolic changes occurring in the human heart in response to I/R is more detailed than prior reports, the more novel findings of our study are the significant differences in myocardial metabolism that are dependent on preexisting myocardial state. CAD patients exhibited reduced baseline fatty acid and ketone extraction compared with control subjects, whereas LVD patients, although similar to control subjects at baseline, exhibited decreased fatty acid and ketone extraction and increased lactate release after I/R. Overall, the myocardium of both CAD and LVD subjects exhibits deficits in metabolic fuel extraction relative to subjects without CAD or LVD but with more pronounced changes in LVD subjects.
In the preischemic state, we found FFA extraction ratios to be lower in patients with CAD than in control subjects, similar to studies from animal models14,15 and human radionucleotide studies.16 The greater dependency on carbohydrate fuels in the setting of CAD appears to be due to a hypoxia-driven increased glycolytic conversion of glucose to lactate at the expense of fatty acid β-oxidation and oxidation of acetyl-CoA in the TCA cycle.
Unlike the situation for CAD patients, we did not find significant differences in baseline myocardial substrate extraction between patients with and those without LVD. The literature is unclear on this subject; some groups have found greater glucose and lower FFA metabolism associated with LVD,17,18 whereas others have reached the opposite conclusion.19,20 The small sample size of LVD patients precluded stratification by ischemic or nonischemic cardiomyopathy, although it is possible that these 2 types may have divergent metabolic signatures that are masked by grouping them together.
The striking picture that emerges for patients with LVD after I/R is one of profound impairment in extraction of a wide range of metabolic fuels, likely reflecting a severe energy deficit that contributes to impaired cardiac function in these patients. In fact, the LVD myocardium actually elutes glucose, suggesting a deficit in glucose metabolism so profound that it limits glucose uptake through the GLUT-1 and GLUT-4 glucose transporters. We do not think that this observation is confounded by the high concentrations of glucose in cardioplegia solution because glucose elution was generally coupled with lactate and ketone release (which are not contained in cardioplegia solution) and not found in control subjects who also received cardioplegia. Studies in failing human hearts and in dog models of end-stage heart failure suggest a substantive downregulation of enzymes of fatty acid oxidation and glucose metabolism, including the GLUT-1 and GLUT-4 glucose transporters.21–24
In addition to differences in myocardial extraction, the net lactate release associated with LVD suggests a dependence on anaerobic glycolysis to generate ATP. This is in keeping with prior studies showing that mitochondrial oxidative capacity is adversely affected by the presence of LVD.25,26 Along these lines, we also noted decreased FFA uptake in the reperfusion phase. Although a number of ex vivo animal studies have suggested that the reperfused heart takes up and oxidizes FFA avidly,27,28 others have reached an opposite conclusion consistent with our data.29,30 Furthermore, previous studies examining FFA extraction in the postreperfusion phase after global myocardial ischemia in humans also have found decreased FFA extraction and oxidation.9–11 This decrease in extraction was most striking in the LVD patients. Taken together, these findings illustrate that subjects with LVD have limited myocardial metabolic reserve and flexibility after global I/R stress.
With regard to implications of this study for clinical practice, strategies that have been investigated for enhancing metabolic function in heart failure include administration of the pyruvate dehydrogenase activator dichloroacetate31,32 or carnitine,33,34 which will facilitate conversion of CoA esters to carnitine esters in the mitochondria, thereby decreasing the acetyl-CoA pool and enhancing oxidation while relieving inhibition of pyruvate dehydrogenase.22 Studies in LVD with these agents to date have been limited, and further work is required to fully understand their utility for treating these subjects.
Several practical limitations must be acknowledged. Although our study is the largest series of its type published to date, sample sizes are still relatively small and involve comparisons between multiple groups. A second important limitation is the lack of a completely normal control population. Control patients generally underwent cardiac surgery for valvular lesions potentially resulting in LV pressure or volume overload but were selected on the basis of normal coronary angiography and LV systolic function.
The CS catheters were placed shortly after initiation of CPB but before AXC placement and the onset of ischemia. Although the heart is still beating and CPB has been ongoing for only a short time before the first arteriovenous sample is obtained, during this brief period, preload and afterload conditions are considerably different from baseline. Similarly, after AXC removal and myocardial reperfusion, the patient remains supported by CPB, and myocardial load conditions again are clearly different from baseline. This, however, serves to standardize both the postreperfusion cardiac output and load conditions. This would imply that the observed metabolic impairment seen with LVD is not due to poor coronary reperfusion or excessive wall stress but rather to an intrinsic property of the myocardium itself.
True measures of myocardial substrate use and metabolic fluxes would require coronary blood flow measurements in addition to extraction data. This limitation notwithstanding, we believe that the major metabolic derangements observed, particularly in the postischemic phase, are unlikely to be due simply to changes in coronary flow, which have been reported to be relatively small compared with the magnitude of changes in extraction noted here.9,35 Additionally, we did not measure rates of oxygen extraction over the coronary circulation before and after ischemia. We therefore cannot exclude that the lower substrate extraction seen after ischemia reflects, at least in part, preferential use of stored myocardial substrates (ie, triglycerides or glycogen) rather than an exclusive defect in oxidative metabolism.
At baseline, the myocardium uses a diverse array of substrates for metabolism. After global I/R, profound changes in substrate uptake occur across all patients. These changes were especially marked in patients with impaired ventricular function. Clinically, this highlights the need to understand and develop novel means of myocardial metabolic optimization before and during periods of ischemia, which extends beyond simple substrate provision. Furthermore, our data suggest that cardiac surgical patients with LVD are at particular risk for perioperative metabolic derangements and therefore deserve strict attention to multimodal cardioprotective strategies.
Sources of Funding
This work is supported by National Institute of Health grants T32 GM08600-12 (Dr Turer), R01 DK58398 (Dr Newgard), and R01 HL092071 (Dr Podgoreanu).
Sabatine MS, Liu E, Morrow DA, Heller E, McCarroll R, Wiegand R, Berriz GF, Roth FP, Gerszten RE. Metabolomic identification of novel biomarkers of myocardial ischemia. Circulation. 2005; 112: 3868–3875.
Ferrara CT, Wang P, Neto EC, Stevens RD, Bain JR, Wenner BR, Ilkayeva OR, Keller MP, Blasiole DA, Kendziorski C, Yandell BS, Newgard CB, Attie AD. Genetic networks of liver metabolism revealed by integration of metabolic and transcriptional profiling. PLoS Genet. 2008; 4: e1000034.
Chavez PN, Stanley WC, McElfresh TA, Huang H, Sterk JP, Chandler MP. Effect of hyperglycemia and fatty acid oxidation inhibition during aerobic conditions and demand-induced ischemia. Am J Physiol. 2003; 284: H1521–H1527.
Recchia FA, McConnell PI, Bernstein RD, Vogel TR, Xu X, Hintze TH. Reduced nitric oxide production and altered myocardial metabolism during the decompensation of pacing-induced heart failure in the conscious dog. Circ Res. 1998; 83: 969–979.
Taylor M, Wallhaus TR, Degrado TR, Russell DC, Stanko P, Nickles RJ, Stone CK. An evaluation of myocardial fatty acid and glucose uptake using PET with [18F]fluoro-6-thia-heptadecanoic acid and [18F]FDG in patients with congestive heart failure. J Nucl Med. 2001; 42: 55–62.
Stanley WC, Recchia FA, Lopaschuk GD. Myocardial substrate metabolism in the normal and failing heart. Physiol Rev. 2005; 85: 1093–1129.
Sebastiani M, Giordano C, Nediani C, Travaglini C, Borchi E, Zani M, Feccia M, Mancini M, Petrozza V, Cossarizza A, Gallo P, Taylor RW, d'Amati G. Induction of mitochondrial biogenesis is a maladaptive mechanism in mitochondrial cardiomyopathies. J Am Coll Cardiol. 2007; 50: 1362–1369.
Ide T, Tsutsui H, Hayashidani S, Kang D, Suematsu N, Nakamura K, Utsumi H, Hamasaki N, Takeshita A. Mitochondrial DNA damage and dysfunction associated with oxidative stress in failing hearts after myocardial infarction. Circ Res. 2001; 88: 529–535.
Liedtke AJ, DeMaison L, Eggleston AM, Cohen LM, Nellis SH. Changes in substrate metabolism and effects of excess fatty acids in reperfused myocardium. Circ Res. 1988; 62: 535–542.
Gorge G, Chatelain P, Schaper J, Lerch R. Effect of increasing degrees of ischemic injury on myocardial oxidative metabolism early after reperfusion in isolated rat hearts. Circ Res. 1991; 68: 1681–1692.
Iliceto S, Scrutinio D, Bruzzi P, D'Ambrosio G, Boni L, Di Biase M, Biasco G, Hugenholtz PG, Rizzon P. Effects of l-carnitine administration on left ventricular remodeling after acute anterior myocardial infarction: the l-Carnitine Ecocardiografia Digitalizzata Infarto Miocardico (CEDIM) Trial. J Am Coll Cardiol. 1995; 26: 380–387.
Alamanni F, Parolari A, Repossini A, Doria E, Bortone F, Campolo J, Pepi M, Sisillo E, Naliato M, Bigi R, Biglioli P, Parodi O. Coronary blood flow, metabolism, and function in dysfunctional viable myocardium before and early after surgical revascularisation. Heart. 2004; 90: 1291–1298.
Metabolic responses to surgical cardioplegic arrest have been incompletely characterized in humans. We applied targeted mass spectrometry–based “metabolomics” analysis to arterial and venous blood samples taken before and after ischemia/reperfusion in patients undergoing cardiac surgery. We observed significant alterations in myocardial fuel substrate uptake at baseline among patients with coronary artery disease and after ischemia/reperfusion among those with preexisting left ventricular dysfunction relative to subjects presenting with neither left ventricular dysfunction nor coronary artery disease. Additionally, net release of acylcarnitine species and lactate by the myocardium after ischemia/reperfusion is suggestive of severe impairment of oxidative metabolism. The ability to maintain oxidative metabolism in the immediate postischemic phase was diverse in the study cohort and associated with improved periprocedural hemodynamics. We conclude that myocardial metabolism is severely altered after surgically induced global ischemia, with the dysfunctional ventricle being particularly susceptible to this insult. Furthermore, metabolic profiling in the perioperative period has the potential to predict the postoperative hemodynamic course, highlight pathways most severely dysregulated by ischemia/reperfusion, and target future cardioprotective strategies.
↵*Drs Newgard and Podgoreanu contributed equally as senior authors on this work.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.108.816116/DC1.