Inflammation has been suggested to underlie in part the risk of arrhythmias and cardiopulmonary dysfunction during the first weeks after cardiac surgery.Reference Hassanabad, Furman, Parratt and Aughey1 Cardiopulmonary by-pass elicits an intense inflammatory reaction, with subsequent deleterious effects on the vascular function.Reference Verrier and Boyle2 C-reactive protein, an important inflammatory mediator mainly synthesized by the liver in response to circulating cytokines, rises rapidly after surgery, reaching a peak after approximately 24 to 48 hours.Reference Franke, Lante and Fackeldey3 Several studies have suggested an association of C-reactive protein with the risk of developing atrial fibrillation and myocardial dysfunction after such surgery,Reference Bruins, te Velthuis and Yazdanbakhsh4 while increased levels of the C-reactive protein has also been suggested to have deleterious effects on myocardial function even in the absence of cardiac surgery.Reference Wiklund, Ronquist, Roomans, Rubertsson and Waldenstrom5 These observations lend support to the concept that increased levels of the C-reactive protein might be an independent risk factor for myocardial disease.Reference Paul6
To date, however, there is no direct evidence to link the possible association between levels of C-reactive protein and the changes in flow through the coronary microcirculation occurring after cardiopulmonary bypass surgery. Data from both experimental and clinical studies using recombinant C-reactive protein, and leading to a wide range of circulating levels, suggested possible dichotomous effects of the C-reactive protein on the vasomotor tone and reactivity of the conduit arteries.Reference Sternik, Samee and Schaff7 Some possible explanations for these contradictory results include contamination of the C-reactive protein extract with other substances, such as endotoxin and sodium azide.Reference Zhong, Li and Liu8
Our recent transthoracic Doppler echocardiographic studies9 have suggested that basal coronary arterial flow increases in children up to 4 to 6 days after cardiopulmonary bypass surgery. In this same population of children, therefore, we have investigated whether the postsurgical rise in C-reactive protein was correlated with the observed changes in coronary arterial flow.
Subjects and methods
We studied 17 infants referred for cardiopulmonary bypass surgery for closure of ventricular- or atrioventricular septal defects. Their ages ranged from 3 to 18 months. Clinical data was obtained from the hospital records, and written consent had been obtained from the guardians of all the children. The study was approved by the ethics committee for human research at Lund University.
Transthoracic Doppler echocardiography was performed on all children 1 day before, and 5 days plus or minus 1 day postoperatively. The criterions for exclusion were clinical signs of infectious illness, levels of C-reactive protein greater than 1 milligrams per litre prior to surgery, heart failure and preoperative therapy with vasoactive drugs. The Doppler examination was performed using the Sequoia™ C512 machine, manufactured by Acuson in Mountain View, California, United State of America. We used a 7 megaherz transducer, and carried out standard M and B-mode, and Doppler, echocardiographic interrogations to determine the anatomy and function of the heart. Measurements of basal coronary arterial flow, and peak velocity of flow, were performed according to the protocol described in previous studies.Reference Hozumi, Yoshida and Ogata10, Reference Saraste, Koskenvuo and Knuuti11 Pulsed Doppler of 4.5 megaherz, a sweep rate of 100 millimetres per second, and velocity settings of 15 to 60 centimetres per second before, and 30 to 60 centimetres per second after, were used. All measurements were corrected for the angle between the Doppler beam and the direction of coronary arterial flow. True velocity was defined as the measured velocity divided by the cosine of the angle between the Doppler beam and the direction of blood.
The bifurcation of the left coronary artery was imaged from the standard short-axis view of the arterial trunks. The internal dimension of the anterior interventricular artery was measured at end-diastole with callipers applied on the inner borders 2 to 3 millimetres distal to the bifurcation of the left coronary artery. The velocity scale was decreased to the minimum range, and then gradually increased until colour signals were optimized within the lumen of the artery. After obtaining good signals of flow, the sample volume was adjusted to 0.5 to1.0 millimetres. The sample was positioned on the flow signal 2 to 3 millimetres distal to the bifurcation, and the velocity of flow recorded. We chose a sample volume that gave the best possible quality envelope, and pure sound throughout the cardiac cycle.
The images were saved on a magnetic optical disc. They were reviewed offline in slow motion, and analyzed in single frame advance mode. The diameter of the aortic root was measured in M-mode using a long axis view at the level of the proximal hinge points of the valvar leaflets. Left ventricular mass was calculated according to Devereux and colleagues,Reference Devereux, Alonso and Lutas12 and as recommended by the American Society of Echocardiography.13
Left ventricular mass per gram is equal to 0.8 [1.04 (left ventricular internal diameter in diastole plus ventricular septal thickness in diastole plus left ventricular inferior mural thickness in diastole)Reference Franke, Lante and Fackeldey3 (minus left ventricular internal diameter in diastole)Reference Franke, Lante and Fackeldey3] plus 0.6.
We corrected the increase in the velocity time integral for the left anterior interventricular artery for left ventricular mass.Reference Hildick-Smith, Johnson, Wisbey, Winter and Shapiro14
The velocity time integral corrected in centimetres per 100 grams is equal to [(velocity time integral on the fifth day minus velocity time integral pre-operatively) per velocity time integral pre-operatively] multiplied by the left ventricular mass pre-operatively−1 multiplied by 100.
Left ventricular fractional shortening was computed from the standard formula. Arterial blood pressure was measured by an automatic oscillometer cuff sphygmomanometer (Dynamap, Critikon Inc, Tampa, Florida, United States of America).Reference Czernin, Muller and Chan15 The product of rate and pressure was calculated by multiplying heart rate with the systolic blood pressure.Reference Baller, Bretschneider and Hellige16 The analysis package of the ultrasound unit was used for manual tracing of the spectral envelope. The measurements of the velocity of flow across the aortic valve were averaged over three consecutive cardiac cycles, and included in measurements of cardiac output. We measured the peak velocity of flow in diastole, and the velocity time integral in diastole and systole. The velocity time integral per minute is equal to the velocity time integral in systole and diastole multiplied by the heart rate. The flow of blood per minute was calculated as the velocity time integral per minute multiplied by cross-sectional area, where cross-sectional area is equal to pi multiplied by the square of the coronary arterial diameter divided by 2.
Postoperative therapy
All children received opioid derivatives in the form of intravenous ketobemidone hydrochloride, given at 25 micrograms per kilogram during the first 3 postsurgical days, and oxicodon, given at 0.1 milligram per kilogram orally between the third and fifth postoperative days. None of our patients received inhibitors of angiotensin converting enzyme, inotropes, beta-blockers or other vasoactive medications.
Statistics
We used paired Student’s t test for comparison of the data obtained before and after surgery. The effects postsurgically of the levels of C-reactive protein on the variations in coronary arterial flow were assessed by ANOVA. Simple regression was used to assess the correlation, if any, between the variations in coronary arterial flow and the levels of C-reactive protein in the plasma. The levels of C-reactive protein were log-transformed, given their skewed distribution. All statistical analyses were performed using StatView® (SAS Institute, Cary, NC, USA), version 5.0 as a statistical software package. A value of p of less than 0.05 was considered statistically significant. Results are presented as mean values plus or minus the standard deviation.
Results
All children had similar pre- and peri-operative characteristics, and no adverse clinical events occurred during the postoperative period. Prior to surgery, levels of C-reactive protein were below the limits for detection in all the children, in other words less than 0.8 milligrams per litre. Levels of the C-reactive protein increased significantly by the second day, reaching a median of 25, and a range from 4 to 142 milligrams per litre. They remained elevated on the fifth day after surgery, when the median was 11, and the range from 3 to 20 milligrams per litre.
Compared to the preoperative values, blood flow in the left anterior interventricular artery, the velocity time integral in diastole and systole, and peak velocity of flow in diastole, all increased significantly on the fifth day, at 27 plus or minus 10 versus 47 plus or minus 15 millilitres per minute, p equal to 0.0001; 9 plus or minus 3 to 14 plus or minus 5 centimetres, p equal to 0.05; and 36 plus or minus 10 to 49 plus or minus 16 centimetres per second, p equal to 0.05, respectively.
The velocity time integral after surgery, and the peak velocity of flow in diastole, correlated inversely with log-transformed values for C-reactive protein, with r equal to −0.54, p less than 0.02, and r equal to −0.74, p less than 0.01, respectively (Fig. 1, Panels A and B). When patients were further divided in relation to the period of aortic cross-clamping, which had a mean of 59 minutes, and ranged from 33 to 121 minutes, these associations remained significant only in those in whom the period of clamping exceeded 1 hour (Fig. 2). The increase in postsurgical velocity time integral corrected for left ventricular mass was significantly lower among those patients with levels of C-reactive protein greater than 10 milligrams per litre than in the remaining patients, at 5 plus or minus 2 versus 2 plus or minus 1, respectively; p less than 0.05 (Fig. 3).
Figure 1 Panel A shows the velocity time integral for the left anterior interventricular artery in centimetres (LAD VTI, Panel A), while Panel B shows the diastolic peak velocity of flow (LAD PFV, Panel B) both inversely correlated with log C-reactive protein on the fifth day after surgery (r = −0.54, p < 0.02, and r = −0.74, p < 0.01, respectively).
Figure 2 The effect of the duration of aorta cross-clamping on the association between the velocity time integral in diastole and systole (LAD VTI) and log C-reactive protein on the fifth day after surgery.
Figure 3 The effect of C-reactive protein on the velocity time integral for the left anterior interventricular coronary artery (LAD VTI) corrected for left ventricular mass on the fifth day after surgery.
On the fifth day postoperatively, neither the diameter of the left anterior interventricular artery, nor flow through the artery, correlated with the levels of C-reactive protein, p being greater than 0.5. Similarly, levels of C-reactive protein showed no association with left ventricle fractional shortening, cardiac output, the product of rate and pressure, and mean arterial blood pressure, the value of p being greater than 0.5 for all.
Discussion
We have shown that, in children undergoing cardiopulmonary bypass surgery, the rise in the velocity of flow through the coronary arteries after the procedures was attenuated in those with higher levels of C-reactive protein. This inverse association between levels of C-reactive protein and the velocity of coronary arterial flow was strengthened by prolonged aortic cross-clamping during the period of surgery. Although these findings do not prove a causal relationship, they are in keeping with the hypothesis that systemic inflammation might be an important mechanism in the pathogenesis of postsurgical myocardial dysfunction, acting in part by possible vasoconstrictive effects on the coronary circulation.
Cardiopulmonary bypass and surgery are strong stimuluses for systemic inflammation in both children and adults, being characterized by an intense and rapid rise of inflammatory cytokines, followed from the second day by 100 to 200 fold increase in levels of C-reactive protein.Reference Paparella, Galeone and Venneri17 This postoperative rise in levels of the C-reactive protein, particularly in adults, coincides with an increased risk for cardiovascular complications, such as haemodynamic instability, myocardial ischaemia, arrhythmias, and even multi-organ failure in severe cases.Reference Bruins, te Velthuis and Yazdanbakhsh4, Reference Gaspardone, Versaci and Tomai18
After cardiopulmonary bypass surgery, the coronary microcirculatory phenotype is shifted to a highly disturbed state, characterized by decreased myogenic tone, and impaired ability to dilate in response to physiological or pharmacological stimuluses.Reference Wang, Friedman, Franklin and Sellke19, Reference Hiratzka, Eastham and Carter20 The mechanisms underlying the impaired coronary flow reserve appear to be manifold, and may include an imbalance between dilating and constricting compounds synthesized by the arterioles. Importantly, C-reactive protein stimulates release from endothelial cells of endothelin-1, which is a powerful vasoconstrictor.Reference Verma, Li and Badiwala21 Cardiopulmonary bypass surgery is an important source of endothelin-1.Reference Pernow and Wang22 C-reactive protein also poses a down-regulatory action on the constitutive form of nitric oxide synthase via decreased stability of endothelial nitric oxide synthase messenger ribonucleic acid, thereby resulting in decreased amounts of nitric oxide, which is a potent vasodilator. In one previous experimental study,Reference Lafuente, Azcutia and Matesanz23 exposure of coronary arterioles to human recombinant C-reactive protein at levels comparable to those attained in our study, namely 7 versus 11 milligrams per litre, resulted in significant inhibition of the endothelium-dependent nitric oxide-mediated dilation of the arterioles by increasing production of superoxide from nicotinamide adenine dinucleotide phosphate oxidase via activation of the p38 kinase.
Alternatively, the coronary microcirculatory phenotype associated with C-reactive protein in our study might be a consequence of the surgically induced disturbances in the adjacent ventricular myocytes, which closely interact with the arteriolar smooth muscle cells. Particularly during cardiopulmonary bypass surgery, the myocardial cells are highly susceptible to damage via the inflammatory response, ischaemia-reperfusion, and surgery as such, with subsequent infiltration of leucocytes and oedema surrounding the arterioles, and hence, arteriolar compression.Reference Heltne, Koller and Lund24 This hypothesis, however, is less plausible, since neither the haemodynamic parameters relating to the myocardium, such as heart rate, contractile function and blood pressure, nor the surgical parameters, such as clamping, perfusion and overall duration of surgery, correlated with the measured levels of C-reactive protein.
Our study does have its limitations, the main one residing in the lack of cardiovascular complications in our material, which precludes conclusion on the possible impact of the observed association on the clinical outcome. It is also the case that screening for additional markers of inflammation, such as interleukins, might have provided additional mechanistic clues.
We conclude, nonetheless, that the rise we observed in C-reactive protein after cardiopulmonary bypass surgery is inversely associated with the velocity of coronary arterial flow as assessed by transthoracic Doppler echocardiography. Although this finding in itself is not proof of a causative relationship, it is in keeping with the hypothesis that systemic inflammation following cardiac surgery could be an important mechanism in the onset of myocardial complications due to dysfunction of the coronary microvasculature. Additional studies are needed to investigate whether this association could, in part, account for the previously suggested adverse effects of inflammation on myocardial physiology during the immediate postoperative period.
Acknowledgements
Our study was partly supported by grants from the Lund University Hospital and the Medical Faculty of Lund University. We thank our registered research nurse, Ms. Annica Maxedius, for her help in the recruitment of the patients.
