Author + information
- Received June 18, 2007
- Revision received July 24, 2007
- Accepted July 31, 2007
- Published online November 6, 2007.
- Kelly N. Vogt, BSc⁎,
- Cedric Manlhiot, BSc⁎,
- Glen Van Arsdell, MD†,
- Jennifer L. Russell, MD⁎,
- Seema Mital, MD⁎ and
- Brian W. McCrindle, MD, MPH⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Brian W. McCrindle, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada M5G 1X8.
Objectives We sought to define somatic growth patterns for patients with single ventricle (SV) physiology and associated factors.
Background Infants with SV physiology might have somatic growth retardation associated with volume overload and hypoxemia, which might improve after surgical palliation.
Methods We reviewed 126 patients (35% male) who underwent the Fontan procedure from 1994 to 2004. Demographic data, hemodynamic variables, and surgical procedures were recorded. Serial weights and heights were converted to z-scores. Linear regression analysis adjusted for repeated measures was used to model growth trends.
Results Median z-score for weight was −0.7 at birth, −1.6 before bidirectional cavopulmonary shunt (BCPS), −0.7 before Fontan procedure, and −0.7 after Fontan procedure. A significant decline in z-scores for weight was seen before BCPS, which was reversed after the hemi-Fontan and stabilized after Fontan procedure. The z-scores for weight before the BCPS were lower in patients with lower birth weight (p < 0.01), nutritional difficulties (p = 0.01), and higher right atrial pressures (p = 0.02). After the BCPS, impaired growth was seen in patients who had systemic venous collaterals (p < 0.01). Patients who had collaterals embolized had the same growth trends as patients with no collaterals (p = 0.29).
Conclusions Infants with SV physiology show impaired somatic growth before BCPS. Although catch-up growth occurs after BCPS, effective interventions such as more intensive nutritional strategies before BCPS might be targeted at this high-risk population. The presence of systemic venous collaterals might impede growth secondary to hemodynamic impairment. Embolization of collaterals might allow for maximum growth potential.
For patients with congenital heart disease, somatic growth retardation has been well described (1–5). The cause of growth failure is multifactorial and might include poor nutrition, increased metabolic requirements, endocrine factors, impaired hemodynamic status, and hypoxemia (6). For many congenital heart defects, early surgical repair has been shown to lead to improved growth (2,5–8).
Children with single ventricle (SV) physiology represent a heterogeneous group of patients with varying underlying diagnoses. Single ventricle physiology is associated with congestive heart failure and hypoxemia. Surgical intervention aims to optimize systemic and pulmonary blood flow through staged palliation including a bidirectional cavopulmonary shunt (BCPS) followed by Fontan procedure (9). It is thought that early surgical intervention and therefore early removal of volume overload and cyanosis leads to improved clinical outcomes and growth (8,10).
Studies conducted to date examining somatic growth in patients with SV physiology have demonstrated inconsistent results (11–15). Although many patients show growth retardation preoperatively, studies have shown variable periods of catch-up growth after surgical intervention. The demographic, surgical, and hemodynamic factors associated with somatic growth failure in these patients are not well described. Furthermore, the relationship between arterial and venous collateral vessels and somatic growth has not previously been described.
We sought to characterize somatic growth for patients with SV physiology having undergone staged surgical palliation, including BCPS and the Fontan procedure, and to define associated factors, particularly the effects of hemodynamic variables and collateral vessels.
The study population consisted of patients with SV physiology who underwent the Fontan procedure at the Hospital for Sick Children, Toronto, between January 1994 and June 2004. This study was approved by the institutional Research Ethics Board. Individual consent was waived for chart review. Patients were identified from the cardiac surgery and cardiology databases. To have a homogeneous population, inclusion in the study required that both the BCPS and the Fontan procedure had been performed at our institution. Furthermore, all patients must have had cardiac catheterization at our institution during 3 specific time periods: before BCPS, between BCPS and the Fontan, and after the Fontan procedure, to assess hemodynamic status and the presence of collaterals at each stage. During the study period, these catheterizations were performed routinely in almost all patients at our institution to assess hemodynamic status and to perform interventions to address pathway obstruction, occlusion of venous and arterial collateral vessels, and closure of Fontan pathway fenestration. Patients for whom complete surgical or catheterization data were not available, those with genetic syndromes associated with growth failure, or patients who went on to transplantation were excluded. We did not exclude patients with low birth weight or pre-term gestation (<36 weeks), because these can be associated with severe congenital heart disease and are a characteristic of the study population. Furthermore, we wished to examine any impact low birth weight or pre-term gestation might have on subsequent growth trends.
Medical records data collected included gender, primary cardiac diagnosis, morphology of the dominant ventricle, and the presence of other noncardiac diagnoses. The date and details of the procedure performed were recorded for all surgeries and interventional catheterizations. For the BCPS and Fontan procedures, specific data on the type of anastamotic procedures performed, time spent on cardiopulmonary bypass, and whether or not the Fontan pathway was fenestrated were noted. Any documented surgical complications were also recorded.
For each hospital visit, the patient’s height, weight, and current medications were recorded. Height and weight values were converted to z-scores for analysis with the Center for Disease Control normative population data (16) Patients with nutritional difficulties were identified and defined as failure to tolerate or achieve sufficient caloric intake via oral feeding, necessitating nasogastric or gastrostomy tube placement and an increased-calorie feeding regimen. This strategy was not used prophylactically during the study period at our institution.
For each cardiac catheterization, hemodynamic variables recorded included systemic arterial oxygen saturation, pulmonary artery saturation and mean pressure, right atrial pressure, and the end diastolic pressure of the dominant ventricle. Interventional procedures preformed during the catheterization were noted. Data were also recorded regarding the presence of any venous or arterial collaterals identified at the time of catheterization and whether or not these collaterals were embolized. An active arterial or venous collateral was defined as being any collateral vessel that was not embolized during catheterization or surgically ligated.
Data are presented as means with SDs, medians with minimum and maximum values, and frequencies as appropriate. Measurements were grouped into 3 time periods corresponding to the different stages of the BCPS-Fontan surgical pathway. The pre-BCPS period was considered as the time from birth until the BCPS. The pre-Fontan period was considered as the time from BCPS to the Fontan procedure. The post-Fontan period was considered as the time from Fontan to last follow-up.
Demographic, physiologic, surgical, and medical factors potentially affecting weight and/or growth were initially tested individually in a univariate mixed linear regression model adjusted for repeated measures over time. Variables found to significantly affect weight and/or growth in any of the 3 periods were then included in a multivariable mixed linear regression model for each time period. Backward selection was used to obtain a final model for each period. Variable estimates from regression equations were used to illustrate growth over time and the effect of collaterals on growth. Statistical analysis was performed with SAS statistical software (version 9.1, SAS Institute, Cary, North Carolina). Default settings were used.
From a database of all cardiac surgical patients, a total of 231 patients who had undergone BCPS and Fontan procedures at our institution and who had undergone at least 3 cardiac catheterizations were identified. Of these, 91 had not undergone catheterization during the defined time periods at our institution and were therefore excluded from analysis, because hemodynamic status and collaterals status were not available. A further 5 patients who, upon review, did not meet inclusion criteria were excluded from the analysis. For an additional 9 patients, complete hospital records were unavailable at the time of study. Therefore, a total of 126 patients were included, 44 (35%) of whom were female. At the time of data collection, patients had been followed for a median of 7.9 years (range 1.8 to 16.3 years). The median number of weight and height measurements for each patient was 16, ranging up to 58. Other characteristics of the study subjects are given in Table 1.
Of the patients studied, 97 (77%) had a surgical procedure performed before the BCPS. All patients underwent BCPS and Fontan during the study period. Details of the surgical procedures performed are given in Table 2.Three patients had documentation of ligation of collateral vessels at the time of other surgical procedures. The initial Fontan procedure required revision in 2 patients.
During the pre-BCPS period, 87 (69%) patients were receiving cardiac medications and 20 (16%) were receiving noncardiac medications. Cardiac medications included beta-blockers, angiotensin-converting enzyme inhibitors, diuretics, antiarrhythmics, and anticoagulants. During the pre-Fontan period, 105 (83%) patients were receiving cardiac medications and 31 (25%) were receiving noncardiac medications. After the Fontan procedure, 117 (93%) patients were receiving cardiac medications and 32 (25%) were receiving noncardiac medications.
Nutritional difficulties as defined were noted in 42 (33%) patients in the pre-BCPS period, 22 (18%) patients in the pre-Fontan period, and 6 (5%) patients in the post-Fontan period.
A total of 561 cardiac catheterizations were available for analysis. Of these, 312 (56%) were routine diagnostic or preoperative studies, 51 (9%) were performed for investigation of symptoms, and 198 (35%) were performed for interventional purposes, including collateral vessel occlusion and fenestration closure. Hemodynamic variables from the pre-BCPS, pre-Fontan, and post-Fontan catheterizations are shown in Table 3.
In the pre-BCPS period, venous collateral vessels were identified in 5 (4%) patients and arterial collaterals were identified in 7 (6%) patients. In the pre-Fontan period, venous collaterals were identified in 38 (30%) patients and arterial collaterals were identified in 52 (41%) patients. In the post-Fontan period, venous collaterals were identified in 49 (39%) patients and arterial collaterals were identified in 19 (15%) patients.
In the pre-BCPS period, venous collaterals were embolized in 1 patient; no arterial collaterals were embolized. In the pre-Fontan period, venous collaterals were embolized in 15 (12%) patients and arterial collaterals were embolized in 32 (25%) patients. In the post-Fontan period, venous collaterals were embolized in 41 (33%) patients and arterial collaterals were embolized in 11 (9%) patients.
The mean birth weight of studied patients (n = 109) was 3.1 kg (± 0.7), with a median z-score of −0.7 (−3.8 to 1.8); 8 of 106 (8%) were born before 36 weeks of gestational age. The z-scores for weight, height, and weight-for-height at the last clinic visit before the BCPS (median age 0.5 years, range 0.1 to 8.5 years) were −1.6 (−5.5 to 0.8), −0.6 (−5.4 to 2.3), and −1.3 (−10.0 to 2.6), respectively. The z-scores for weight, height, and weight-for-height at the last clinic visit before the Fontan (median age 2.5 years, range 1.3 to 11 years) were −0.7 (−5.2 to 2.1), −0.6 (−4.9 to 1.9), and −0.4 (−3.1 to 2.2), respectively. The z-scores for weight, height, and weight-for-height at the first clinic visit after the Fontan (median age 3.1 years, range 1.5 to 11.5 years) were −0.7 (−5.2 to 2.0), −0.7 (−4.6 to 1.5), and −0.2 (−3.3 to 2.4), respectively. Trends in z-scores for weight over these 3 time periods are shown in Figure 1.A significant decline (p < 0.01) in z-scores for weight was seen before the BCPS. After the BCPS, there was significant catch-up growth (p < 0.01), which leveled off after completion of the Fontan. Trends in z-scores for height over these 3 time periods are shown in Figure 2.There was a small increase in z-scores for height in the pre-Fontan period (p = 0.02). After the Fontan, there was no significant increase in z-scores for height (p = 0.29). Trends in z-scores for weight-for-height are shown in Figure 3.There was no change in z-scores for weight-for-height before BCPS (p = 0.93). In the pre-Fontan period, there was a significant increase in z-scores for weight-for-height (p = 0.01), which leveled off after completion of the Fontan (p < 0.01). In the post-Fontan period, z-scores for weight-for-height for these patients trended to normal.
Factors associated with somatic growth
Factors found to be related to both z-scores for weight and trends in z-scores for weight in our series are shown in Table 4.Lower z-scores for birth weight were associated with lower z-scores for weight throughout all 3 study periods (the results were the same whether or not missing values for birth weight were replaced with imputed mean values). The requirement for surgical procedures before the BCPS was associated with lower z-scores for weight in both the pre-BCPS and the pre-Fontan periods. Furthermore, the requirement of pre-BCPS surgical procedures was correlated with later age at BCPS, which was also associated with impaired growth (p < 0.01). In the pre-BCPS period, the presence of defined feeding difficulties was associated with lower z-scores for weight (p < 0.01). Higher right atrial pressures noted at pre-BCPS catheterization were associated with both lower z-scores for weight and impaired growth before the BCPS (p < 0.01). The association between of right atrial pressure on the trend for z-scores for weight in the pre-BCPS period is shown in Figure 4.In the time between BCPS and Fontan, lower z-scores for weight are seen for patients who had the Fontan at a later age (p < 0.01) as well as those who had 2 or more active venous collaterals during this time period (p < 0.01). Patients taking cardiac medications during this period were also found to have lower z-scores for weight (p = 0.03).
The relationship between active venous collaterals and trend in z-scores for weight can be seen in Figure 5.In both the pre- and post-Fontan periods the presence of 2 or more active venous collaterals was associated with impaired growth (p < 0.01) independently from arterial saturation and pulmonary artery pressure. Those who had venous collaterals embolized had no significant difference in trend for z-scores for weight after embolization from those who had no collaterals or only 1 active collateral (p = 0.89 pre-Fontan; p = 0.19 post-Fontan).
Presence of venous collateral after the BCPS was found to be associated with lower arterial oxygen saturation (p < 0.01), higher mean pulmonary artery pressure (p = 0.01), higher right atrial pressure (p < 0.01), and higher end diastolic pressure in the dominant ventricle (p = 0.02). At the univariate level all impaired hemodynamic variables were found to be significantly associated with impeded growth.
The analysis was redone excluding pre-term patients (n = 8, <36 weeks gestational age), patients with non-cardiac diagnoses (n = 15), and patients who were transferred to our institution after the age of 2 years (n = 2) to exclude the possibility that the associations and trends observed were overtly influenced by patient subsets expected to have the worse growth. Although some variables showed marginally weaker associations than with the entire cohort, both analyses had similar results.
Trends for weight
Patients with SV physiology in our study were underweight at birth and had a steep decline in z-scores for weight before the BCPS (p < 0.01). Catch-up growth was seen after the BCPS (p < 0.01), with leveling off of growth after Fontan. These results are similar to those of Cohen et al. (11), who examined patients who had hemi-Fontan and Fontan. Although their cohort had normal birth weight, a significant decrease in weight z-scores was seen before hemi-Fontan (mean z-score −1.5). Significant increase in weight z-scores was seen before Fontan (mean z-score −0.9), and there were some gains after Fontan, with a z-score at last follow-up of −0.5. Stenbog et al. (14) studied patients only after the BCPS with a pre-operative mean weight z-score of −1.1. Post-operatively they report an increase in z-scores for a weight of 1.1. Day et al. (12) studied patients who underwent either BCPS or Fontan and concluded that the greatest potential for catch-up growth occurs after the BCPS, similar to our findings. Ovroutski et al. (15) studied patients who underwent BCPS and Fontan and found that the greatest growth occurred after Fontan. Their patients, however, had higher birth weights (median in the 25th percentile) than those in our series and therefore, given our finding that lower birth weight is associated with lower z-scores for weight throughout management, are likely not representative of our population. In patients with SV physiology, it seems likely that the most significant impairment in growth might occur in the period before BCPS and that the greatest potential for improvement in growth trends occurs in the period between BCPS and Fontan.
Impact of venous collateral vessels
Venous collaterals are known to develop in patients after BCPS and Fontan and represent either reversal of flow within an existing vessel or the opening of an entirely new vessel between the high-pressure superior vena caval system and the low-pressure inferior vena caval system or pulmonary venous atrium (17,18). Venous collaterals might be associated with increased cyanosis, which might be a reflection of a reduction in effective pulmonary blood flow (17,19). The presence of venous collaterals increases both morbidity and mortality in patients with SV physiology, and their occlusion improves arterial oxygen saturation (20,21). Collateral vessels thought to be of clinical significance are either embolized or surgically ligated. Our findings show that the presence of 2 or more venous collaterals that were not embolized is associated with a decline in z-scores for weight independently of hemodynamic values. For those patients in our series who had venous collaterals embolized or surgically ligated, trends in z-scores for weight were the same as those for patients without active collaterals or with 1 active collateral. It is possible that the presence of clinically detrimental venous collaterals might limit growth secondary to underlying impaired hemodynamic variables, specifically higher pulmonary resistance, and pressure, thereby increasing cyanosis and lowering potential for growth. This might be a reversible effect, and if collaterals are embolized or ligated, these patients might return to a growth trend similar to that which they were following before the development of venous collaterals.
Impact of nutritional difficulties
Our finding that patients with nutritional difficulties have lower z-scores for weight in the pre-BCPS period suggests that these patients likely require early and aggressive nutritional intervention. Results from a cohort of infants undergoing cardiac surgery who were randomized to receive either rapid advancement of a more concentrated formula or usual care demonstrated that infants who received a feeding intervention had significantly greater weight gain and improved clinical outcomes (22). Infants with hypoplastic left heart syndrome have shown improved nutritional outcomes, including higher weight-for-age z-scores at the time of hospital discharge, if they received aggressive nutritional support during their early hospital stays (23). It seems that in the vulnerable period before the BCPS, aggressive interventions including early or prophylactic nutritional support might be indicated to improve outcomes in patients with SV physiology. Because of improved hemodynamic variables and decreased caloric requirements after the BCPS, nutritional interventions should be targeted at the pre-BCPS period.
Impact of other factors
Our results suggest that impaired growth is seen in patients who required surgical intervention before the BCPS and who had the BCPS at a later age. Furthermore, impaired growth was seen in patients who had the Fontan at a later age. These results are supported by both Stenbog et al. (14) and Ovroutski et al. (15), who conclude that early surgical intervention is a requirement for adequate catch-up growth. It is thought that the preservation of heart function early in the management course allows for adequate catch-up growth to occur later in management (5). Elevated right atrial pressure in patients with SV physiology is likely associated with ventricular volume overload and diastolic dysfunction during the period before the BCPS. We note that higher right atrial pressure before the BCPS is associated with both lower z-scores for weight and impaired growth.
Trends for height
The trends for height in our series differed from that for weight, with patients starting out with below normal z-scores for height, which persisted throughout the study period. This is similar to the findings of Day et al. (12), who found that no significant change in z-score for height occurred after completion of the Fontan. In our series, there was a small increase in z-scores for weight seen after completion of the BCPS, but this did not seem to continue after Fontan. A similar small increase was found by Stenbog et al. (14), who found an increase in z-scores for height of 0.8, from a pre-operative value of −0.5. These patients were not followed after further surgical management. These findings suggest that patients with SV physiology have impaired growth with respect to height in the long term. Studies have shown that patients with congenital heart disease might have delayed bone age, which is affected by chronic hypoxemia in childhood (24). This delay in bone age is directly correlated with lower height percentiles. This is further supported by Witzel et al. (25), who report that patients who required a Fontan remained significantly shorter than the normal population even into adolescence. Furthermore, they demonstrate abnormal bone density in these patients and suggest this might play a role in the stunting of growth. It is likely that the chronic hypoxemia experienced by children with SV physiology might impair initial bone development at a crucial stage and impact growth even after surgical palliation.
Trend for weight-for-height
The trend observed in our series in z-score for weight-for-height is as expected, given the large change in z-scores for weight and the lack of change in z-scores for height throughout the early management of these patients. The flat-line in the BCPS period shows that losses in z-scores for weight are proportional to losses in z-scores for height. After BCPS, however, differential gains made in z-scores for weight are likely responsible for the increase seen in weight-for-height z-scores. It should be noted that after completion of the Fontan, the z-scores for weight-for-height are above normal, which might reflect increased adiposity in the setting of short stature and might also be a result of the sedentary lifestyle recently described in this population (26).
The present study is limited by its retrospective nature and the nonstandardized measurements of weight and height. It should be noted, however, that measurements were taken at our institution on standard equipment at each visit. The inclusion of only those patients who survived to the time of post-Fontan catheterization might be a potential source of bias; however, it was necessary to obtain data on hemodynamic variables and collaterals and does create a more homogenous population. Furthermore, although the general standard at our institution is to perform routine cardiac catheterization approximately 6 to 12 months after completion of the Fontan, this was left to the discretion of each patient’s responsible cardiologist, particularly in the early part of the reported experience. Before routine post-Fontan catheterization, patients only underwent catheterization for symptoms or for procedural reasons, and therefore our early experience might be impacted by a lack of data on asymptomatic patients who required no procedural intervention. Finally, the data collected on collateral vessels did not account for the clinical importance of these collaterals, and therefore all noted collaterals were included in the analysis. It is possible that the relationships regarding the presence of collateral vessels might be further magnified by analysis including only clinically important vessels.
This study shows that patients with SV physiology have impaired growth before BCPS. Although catch-up growth occurs after BCPS, early intervention such as implementation of more aggressive nutritional strategies before BCPS surgery might be directed toward this at-risk population. Our results suggest that the presence of venous collaterals is likely associated with impaired growth secondary to hemodynamic impairment and hypoxemia. In patients that have venous collaterals embolized, growth potential might be restored. After completion of the BCPS, it might be necessary to assess and embolize venous collaterals, particularly vessels draining into the inferior vena caval system or the pulmonary venous atrium, to allow for maximum growth potential to be reached. Further research is required to clarify the relationship between venous collaterals and trends in growth and to determine effective nutritional strategies that might improve growth, particularly in the pre-BCPS period.
- Abbreviations and Acronyms
- bidirectional cavopulmonary shunt
- single ventricle
- Received June 18, 2007.
- Revision received July 24, 2007.
- Accepted July 31, 2007.
- American College of Cardiology Foundation
- Cheung M.M.,
- Davis A.M.,
- Wilkinson J.L.,
- Weintraub R.G.
- Kaulitz R.,
- Hofbeck M.
- Ovroutski S.,
- Ewert P.,
- Alexi-Meskishvili V.,
- et al.
- Department of Health and Human Services
- Magee A.G.,
- McCrindle B.W.,
- Mawson J.,
- Benson L.N.,
- Williams W.G.,
- Freedom R.M.
- Gatzoulis M.A.,
- Shinebourne E.A.,
- Redington A.N.,
- Rigby M.L.,
- Ho S.Y.,
- Shore D.F.
- McElhinney D.B.,
- Reddy V.M.,
- Hanley F.L.,
- Moore P.
- McCrindle B.W.,
- Williams R.V.,
- Mital S.,
- et al.