Increased Vascular Endothelial Growth Factor in Patients With Cyanotic Congenital Heart Diseases May Not Be Normalized After a Fontan Type Operation

Transcription

1 Increased Vascular Endothelial Growth Factor in Patients With Cyanotic Congenital Heart Diseases May Not Be Normalized After a Fontan Type Operation Kenji Suda, MD, Masahiko Matsumura, MD, Setsuko Miyanish, MT, Kyokun Uehara, MD, Takaaki Sugita, MD, and Masahiko Matsumoto, MD Department of Pediatrics, Division of Pediatric Cardiology, and Department of Cardiovascular Surgery, Tenri Hospital, Department of Clinical Pathology, Tenri Institute of Medical Research, Tenri City, Japan Background. To determine the change of serum concentration of vascular endothelial growth factor (VEGF) in patients with cyanotic congenital heart disease (C-CHD). Methods. Patients conmprised four groups: group A, 19 patients without cyanosis; group B, 24 patients with C-CHD; group C, 17 patients who had C-CHD and underwent biventricular repair; and group D, 15 patients who had single ventricle and underwent a Fontan type operation. Blood samples were obtained from upper arm veins and serum VEGF was determined. We determined correlation between serum VEGF and arterial oxygen saturation and compared levels of serum VEGF among groups. In addition, age and hemodynamic variables derived from cardiac catheterization were analyzed in terms of correlation with serum VEGF. Results. Serum VEGF significantly negatively correlated with arterial oxygen saturation (r 0.62, p < Patients with cyanotic congenital heart disease (C- CHD) often develop various types of collateral vessels. These collateral vessels have a wide range of variety and present either before or after surgical correction of C-CHD [1 4]. Among these surgical corrections, a Fontan-type operation including intra- or extra-cardiac total cavopulmonary connection (TCPC) is offered for the patients with single ventricle physiology, and these collateral vessels have significantly affected the prognosis of patients with this operation. Among these collateral vessels, systemic to pulmonary collateral vessels can increase volume overload to the ventricle and pulmonary arteriovenous collateral vessels or systemic venovenous or venoatrial collateral vessels can bring desaturated blood back to the heart and increase cyanosis deteriorating Fontan physiology. In the formation of these collateral vessels, several angiogenic factors are thought to be involved. Among these angiogenic factors, vascular endothelial growth factor (VEGF) has been known to promote the formation Accepted for publication March 16, Address reprint requests to Dr Suda, Division of Pediatric Cardiology, Department of Pediatrics, Tenri Hospital, 200 Mishima-cho, Tenri City , Japan; kensuda@tenriyorozu-hp.or.jp ). Serum VEGF in B and D were significantly higher than those in A and C (median 443 pg/ml in B and 302 pg/ml in D vs 180 pg/ml in A and 216 pg/ml in C, respectively). Serum VEGF was significantly negatively correlated with patient s age (p 0.54, p < 0.04) in D, but significantly positively correlated with systemic venous pressure (p 0.45, p < 0.01) in pooled data. Conclusions. Patients with C-CHD have increased serum VEGF in parallel with the degree of cyanosis. With biventricular repair, cyanosis and serum VEGF are normalized. However, with a Fontan type operation, cyanosis disappears but serum VEGF may not be normalized because of elevated venous pressure in association with younger age. (Ann Thorac Surg 2004;78:942 7) 2004 by The Society of Thoracic Surgeons of cardiac collateral vessels in ischemic heart disease and was reported to be increased in patients with C-CHD [5, 6]. In addition, lung biopsy specimens from children after cavopulmonary anastomosis demonstrated increased expression of VEGF and its receptor [7]. However, there are still little data available concerning the change of serum VEGF in patients with C-CHD corrected by a Fontan type operation, such as TCPC. Therefore, the aim of this study was to determine serum concentrations of VEGF in patients with C-CHD, to determine the effect of operative correction of C-CHD, such as biventricular repair or TCPC, on serum VEGF and to determine the effect of hemodynamic derangements seen in patients with TCPC, such as high venous pressure, on serum VEGF. Material and Methods All patients were taken care at Tenri Hospital, a regional tertiary care teaching hospital between January 1, 1998 and December 31, For this study, 75 patients who were more than 2 but less than 20 years old were chosen, adjusting patients age. To minimize the possible effect of surgical intervention on serum VEGF level, these patients were enrolled more than 1 year after the surgical 2004 by The Society of Thoracic Surgeons /04/$30.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg SUDA ET AL 2004;78:942 7 VEGF IN CYANOTIC HEART DISEASES Table 1. Diagnoses of the Study Patients Group Number of Patients A(n 19) Healthy control 6 Arrhythmias 4 History of Kawasaki disease 9 without coronary stenosis B(n 24) Tetralogy of Fallot or double outlet 7 right ventricle or corrected transposition of the great arteries and pulmonary stenosis Single ventricle (tricuspid atresia, 14 mitral astresia, pulmonary atresia with hypoplastic right ventricle, and double inlet left ventricle, etc) Miscellaneous 3 C(n 17) Tetralogy of Fallot or double outlet 12 right ventricle and pulmonary stenosis Trasposition of the great arteries 3 or Tasussing-Bing Anomaly Miscellaneous 2 D(n 15) Single ventricle and common 5 atrioventricular valve Tricuspid atresia 6 Miscellaneous 4 intervention if they had surgical intervention before this study. All patients underwent physical examination, electrocardiogram, echocardiogram, and cardiac catheterization if indicated. Arterial oxygen saturation was measured using a pulse oximeter. These patients were divided into four groups based on arterial oxygen saturation and mode of operation (Tables 1 and 2): group A, 19 patients with normal hemodynamics without cyanosis; group B, 24 patients with different types of uncorrected C-CHD with variable cyanosis; group C, 17 patients who had C-CHD and underwent biventricular repair without significant residual lesion; and group D, 15 patients who had single ventricle and underwent TCPC. Group B included 7 patients with anatomy consistent with a biventricular repair, 14 patients with single ventricle physiology, and 3 patients with miscellaneous cardiac defects and there was no significant difference in arterial oxygen saturation among these patients (74% 9%, 81% 6%, and 81% 6%, respectively). Of these patients in 943 group B, 11 patients underwent palliative operation including modified Blalock-Taussig shunt placement (n 6), pulmonary artery banding (n 2), bidirectional Glenn procedure (n 2), and attempted closure of ventricular septal defect (n 1) as the final surgical intervention before this study. In group C, no patient left with ventricular septal defect, more than 30 mm Hg of pulmonary artery stenosis, nor more than grade 2 of pulmonary valve insufficiency evaluated by either echocardiography or cardiac catheterization and all were in class 1 functional status of New York Heart Association Classification. In group D all patients underwent either intracardiac (n 2) or extra-cardiac (n 13) nonfenestrated TCPC using a prosthetic tube with 16 to 20 mm of diameter and were in class 1 functional status of New York Heart Association Classification. Of these patients, 2 patients required additional atrioventricular valve plasty or replacement at 3 years and 7 months or 7 months after the TCPC, respectively. Two other patients required successful catheter intervention to dilate stenosis of the left pulmonary artery at 2 or 3 months after the TCPC, respectively. Patients in group D underwent TCPC at significantly older age than patients in groups B and C, and patients in group C participated in this study at significantly later after operation than groups B and D (Table 2). Blood samples were obtained from upper arm veins in all patients. These blood samples were centrifuged and stored at 20 C until measurement. Serum VEGF was determined by commercially available ELISA kits (R&D Systems Inc, Minneapolis, MN) [5, 8]. In pooled data of groups A, B, and C the correlation between serum VEGF and arterial oxygen saturation was determined. Serum VEGF was compared among four groups. In pooled data of groups A, C, and D as well as in each group the correlation between serum VEGF and age at study was determined separately. In pooled data of groups C and D as well as in each group the correlation between serum VEGF and time after operation was determined separately. Twenty-nine patients, 9 in group A, 10 in group C, and 10 in group D, underwent diagnostic cardiac catheterization within 1 month of blood sampling for determination of serum VEGF. Based on these cardiac catheterizations, hemodynamic variables such as, cardiac index, systolic blood pressure, mean pressure of the superior vena cava CARDIOVASCULAR Table 2. Demographic Data of the Study Patients Group A Group B Group C Group D Age at study (years) Arterial oxygen saturation (%) * Age at operation (years) Time after operation (years) Data are mean 1 standard deviation. There was no significant difference in age among groups. * Significantly higher than Group B, C and D, p 0.01; Significantly higher than Group B and D, p 0.01; Significantly higher than Group B, p 0.01; Significantly higher than Group B and C, p 0.01.

3 944 SUDA ET AL Ann Thorac Surg VEGF IN CYANOTIC HEART DISEASES 2004;78:942 7 as a representative of systemic venous pressure, and pulmonary capillary wedge pressure were determined. These hemodynamic variables and age were analyzed in terms of correlation with serum VEGF. Our institutional review board approved this clinical study and all patients or patients guardians gave written informed consents to participate in this study. Data Analyses Data are expressed as mean with standard deviation, median with range and bar graph as appropriate. We used unpaired t test to compare age, arterial oxygen saturation, age at operation, and time after operation and Kruskal-Wallis s H test with post-hoc analysis using Dunn s method to compare serum VEGF among groups. To determine correlation between serum VEGF and arterial saturation, age at study, time after operation or hemodynamic variables, we determined Pearson s correlation coefficient. These statistical analyses were made by Stat View version 5.0 (SAS Institute, Cary, NC) or Stat- MateIII (ATMS, Tokyo, Japan). Significance level was set for p less than Results There was a significant negative correlation between serum VEGF and arterial oxygen saturation, the lower the arterial saturation, the higher the serum VEGF (Fig 1, VEGF arterial oxygen saturation, r 0.62, p ). In comparison of serum VEGF among groups, serum VEGF in groups B and D were significantly higher than those in groups A and C (median 443 [range 128 to 1077] pg/ml in group B and 302 [ ] pg/ml in group D versus 180 [26 527] pg/ml in group A and 216 [75 355] pg/ml in group C, respectively; Fig 2). Serum VEGF in group D was as high as that in group B and there was no significant difference. Patient s age at study significantly negatively correlated with serum VEGF only in group D (Fig 3, r 0.54, p 0.04), but there was no correlation between them in the each remaining group as well as in pooled data of groups A, C, and D. On the other hand, there was no correlation between time after operation and serum VEGF in group C as well as in group D, but there was only a weak negative correlation between time after operation and serum VEGF in pooled data of groups C andd(r 0.35, p 0.05). In terms of correlation with serum VEGF in pooled data of 29 patients in groups A, C and D that underwent corrective surgery and did not have cyanosis afterward, age and other hemodynamic variables did not significantly correlated with serum VEGF, but only systemic venous pressure represented by pressure of superior vena cava significantly positively correlated with serum VEGF (Fig 4; r 0.45, p 0.01). Comment This study indicates that patients with C-CHD have increased serum VEGF in parallel with the degree of Fig 1. Arterial saturation versus serum VEGF. Arterial saturation significantly negatively correlated with serum VEGF in pooled data of groups A, B, and C (r 0.62, p ). The solid line represents the regression line. The dotted lines represent the 95% confidence interval for slopes and means. (VEGF vascular endothelial growth factor; SpO 2 arterial oxygen saturation measured by a pulse oximeter.) cyanosis. With biventricular repair, cyanosis disappears and serum VEGF may be normalized. In contrast, with a Fontan type operation, cyanosis disappears but serum VEGF may not be normalized. We demonstrated a significant negative correlation Fig 2. Comparison of serum VEGF among four groups expressed by box plot. Serum VEGF in groups B and D were significantly higher than those in groups A and C. In addition, serum VEGF in D was as high as that in B and there was no significant difference; *p 0.01 versus group A, #p 0.01 versus group C. (VEGF vascular endothelial growth factor.)

4 Ann Thorac Surg SUDA ET AL 2004;78:942 7 VEGF IN CYANOTIC HEART DISEASES Fig 3. Age versus serum VEGF in patients with a Fontan type operation. Age at study significantly negatively correlated with serum VEGF in patients with a Fontan type operation. The solid line represents the regression line. The dotted lines represent the 95% confidence interval for slopes and means. (VEGF vascular endothelial growth factor.) between serum VEGF and arterial oxygen saturation and this is in agreement with the reports of Starnes [5], Himeno [6], and Ootaki [9]. Starnes [5] and Ootaki [9] separately demonstrated significantly higher serum VEGF in patients with cyanotic C-CHD than in patients with acyanotic CHD. Also Himeno [6] revealed that plasma VEGF negatively correlated with oxygen saturation and positively correlated with hemoglobin in patients with C-CHD aged between 3 months old and 10 Fig 4. Pressure of superior vena cava versus serum VEGF. Systemic venous pressure represented by pressure of superior vena cava significantly positively correlated with serum VEGF in pooled data of 29 patients who did not have cyanosis. The solid line represents the regression line. The dotted lines represent the 95% confidence interval for slopes and means. (SVC pressure of superior vena cava; VEGF vascular endothelial growth factor.) 945 years old. Our study, in addition to these studies, indicates that systemic hypoxemia can be one of the strong stimuli for the production of VEGF. Concordant with this finding, reversal of hypoxemia could normalize serum VEGF. In fact, patients in group C, patients who had C-CHD and underwent biventricular repair without significant residual lesions, manifested serum VEGF as low as in the group A, acyanotic patients. This finding is compatible with our clinical observation that we have seldom seen patients with collateral vessels after definitive biventricular repair without significant residual lesion. Obviously normalization of serum VEGF might be a combined effect of change in oxygenation and hemodynamics, because these operations normalize not only oxygenation, but also the hemodynamic abnormality seen in patients with C-CHD. Conversely, after a Fontan type operation, serum VEGF may not be normalized, but may stay high despite disappearance of cyanosis. Younger age can be one of the factors that promote production of VEGF after a Fontan type operation. We have indicated significant negative correlation between age at study and serum VEGF in group D. However, this relationship was not seen in the remaining groups and it has been reported that there is no correlation between age and plasma VEGF after 3 months of age [6]. Therefore, age alone should not be the factor that promotes production of VEGF, but be a facilitating factor that enhances production of VEGF in conjunction with other factors. In addition to younger age, earlier time point after surgical intervention may be a factor for increased VEGF. However, in this study, we intentionally excluded patients less than 1 year after operation to minimize the possible effect of surgical intervention itself. Therefore, this is out of scope of this study and another study to determine how long these surgical interventions have influence on the production of VEGF is required. As another factor that promotes production of VEGF, elevated systemic venous pressure in patients after a Fontan operation should be important. We have demonstrated significant positive correlation between mean pressure of superior vena cava and serum VEGF, though we could not show this relationship only in patients with a Fontan operation because of the small sample size. In vitro, mechanical forces on endothelial cells are reported to modulate the production of VEGF. Suzuma and colleagues [10] showed that cyclic stretch in bovine retinal endothelial cells can increase the expression of VEGF and VEGF receptor. In addition, Hollinsworth and coworkers [11] demonstrated that venous stasis experimentally induced by cuff inflation increased plasma VEGF in humans. Therefore, it is possible that increased venous pressure and stretch of venous endothelial cells promote the production of VEGF leading to formation of collateral vessels. In fact, significant collateral formation after a Fontan type operation was frequently observed in patients with increased trans-pulmonic pressure, ie, high systemic venous pressure [12]. Although none of our 15 patients with TCPC has CARDIOVASCULAR

5 946 SUDA ET AL Ann Thorac Surg VEGF IN CYANOTIC HEART DISEASES 2004;78:942 7 suffered from significant collateral formation so far, we need to watch carefully patients with high serum VEGF. Study Limitations There are several limitations in this study. One of the limitations was that we did not determine locally produced VEGF but determined serum level of VEGF. However, in study of cancer patients, serum level of VEGF has shown to be a significant prognostic factor and presumably thought to reflect local anigiogenic activity of various tumors [13 15]. Therefore, we believe that serum level of VEGF can be one of the significant markers of angiogenic activity of these patients. Obviously the small sample size of this study did not allow us to determine hemodynamic factors that have influence on the production of VEGF in cyanotic patients. Larger-scaled and more detailed studies concerning hemodynamic factors and anatomy in cyanotic patients are required. Another limitation was that we did not determine serum VEGF before and after the operation in each patient. To clarify the significance of age and time after surgical intervention on the production of VEGF, further studies including more controlled longitudinal study needs to be performed and our study provides the basis for these studies. Conclusions Patients with C-CHD have increased serum VEGF in parallel with the degree of cyanosis. With bi-ventricular repair, cyanosis disappears and serum VEGF is normalized. In contrast, with a Fontan type operation, cyanosis disappears but serum VEGF may not be normalized. As factors contributing this increased VEGF, elevated systemic venous pressure in association with younger age is important. We thank Dr Julien I. E. Hoffman, Professor of Pediatrics, University of California, San Francisco, for his kind assistance with the manuscript. References 1. Moore JW, Kirby WC, Madden WA, Gaither NS. Development of pulmonary arteriovenous malformation after modified Fontan operations. J Thorac Cardiovasc Surg 1989;98: Clapp S, Morrow WR. Development of superior vena cava to pulmonary vein fistulae following modified Fontan operation: case report of a rare anomaly and embolization therapy. Pediatr Cardiol 1998;19: Heinemann M, Breuer J, Steger V, Steil E, Sieverding L, Ziemer G. Incidence and impact of systemic venous collaterals development after Glenn and Fontan procedures. Thorac Cardiovasc Surg 2001;491: Kaulitz R, Ziemer G, Paul T, Peuster M, Bertram H, Hausdorf G. Fontan-type procedures: residual lesions and late interventions. Ann Thorac Surg 2002;74: Starnes SL, Duncan BW, Kneebone JM, Rosenthal GL, Jones TK, Grifka RG, Cecchin F, Owens DJ, Fearneyhough C, Lupinetti FM. Vascular endothelial growth factor and basic fibroblast growth factor in children with cyanotic congenital heart disease. J Thorac Cardiovasc Surg 2000;119: Himeno W, Akagi T, Furui J, Maeno Y, Ishii M, Kosai K, Murohara T, Kato H. Increased angiogenic growth factor in cyanotic congenital heart disease. Pediatr Cardiol 2003;24: Starnes SL, Duncan BW, Kneebone JM, Rosenthal GL, Patterson K, Fraga CH, Kilian KM, Mathur SK, Lupinetti FM. Angiogenic proteins in the lungs of children after cavopulmonary anastomosis. J Thorac Cardiovasc Surg 2001;122: Obermair A, Tempfer C, Hefler L, Preyer O, Kaider A, Zeillinger R, Leodolter S, Kainz C. Concentration of vascular endothelial growth factor (VEGF) in the serum of patients with suspected ovarian cancer. Br J Cancer 1998;77: Ootaki Y, Yamaguchi M, Yoshimura N, Oka S, Yoshida M, Hasegawa T. Vascular endothelial growth factor in children with congenital heart disease. Ann Thorac Surg 2003;75: Suzuma I, Hata Y, Clermont A, Pokras F, Rook SL, Suzuma K, Feener EP, Aiello LP. Cyclic stretch and hypertension induce retinal expression of vascular endothelial growth factor and vascular endothelial growth factor receptor-2: potential mechanisms for exacerbation of diabetic retinopathy by hypertension. Diabetes 2001;50: Hollingsworth SJ, Tang CB, Dialynas M, Barker SG. Varicose veins: loss of release of vascular endothelial growth factor and reduced plasma nitric oxide. Eur J Vasc Endovasc Surg 2001;22: Weber HS. Incidence and predictors for the development of significant supradiaphragmatic decompressing venous collateral channels following creation of Fontan physiology. Cardiol Young 2001;11: Dirix LY, Vermeulen PB, Pawinski A, et al. Elevated levels of the angiogenic cytokines basic fibroblast growth factor and vascular endothelial growth factor in sera of cancer patients. Br J Cancer 1997;76(2): Shimada H, Takeda A, Nabeya Y, et al. Clinical significance of serum vascular endothelial growth factor in esophageal squamous cell carcinoma. Cancer 2001;92: Coskun U, Gunel N, Sancak B, et al. Significance of serum vascular endothelial growth factor, insulin-like growth factor-i levels and nitric oxide activity in breast cancer patients. Breast 2003;12: INVITED COMMENTARY Suda and colleagues have produced an intriguing investigation of the relationship of vascular endothelial growth factor (VEGF) in patients with cyanosis and the palliative or reparative operations that these patients receive. The topic is an important one, in that VEGF has been hypothesized to play at least a contributory role in the pathophysiology of pulmonary arteriovenous malformations (PAVM). Because PAVM are known to regress in many, if not all, patients who undergo Fontan procedures, the failure of VEGF levels to normalize after the Fontan may at first be taken to undermine the theory that VEGF is a causative factor in PAVM formation and persistence. Although that may indeed be the case, it would be premature to exclude VEGF as an actor in this process. It is known that PAVM occur in a significant proportion of patients with unidirectional or bidirectional cavopulmonary connections over a sufficient 2004 by The Society of Thoracic Surgeons /04/$30.00 Published by Elsevier Inc doi: /j.athoracsur