Advances in surgical therapy and perioperative care have led to improved survival rates for children with single-ventricle CHD.Reference Alsoufi, Mori and Gillespie 1 Reduction of long-term morbidity, improved health-related quality of life, and enhancement in psychosocial performance and neurodevelopmental outcome have become increasingly important in the care of this growing high-risk population.Reference Marino, Lipkin and Newburger 2 Children with single-ventricle CHD are at risk for impaired health-related quality of life in adolescent years,Reference Marino, Lipkin and Newburger 2 – Reference O’Connor, Wray and Tomlinson 6 and into adulthood.Reference Uzark, Jones, Slusher, Limbers, Burwinkle and Varni 7 However, studies on health-related quality of life and its associated factors in pre-school age children are rare.Reference Williams, Gelijns and Moskowitz 8 , Reference Goldberg, Lu and Sleeper 9 , Reference Atz, Zak and Mahony 10
Health-related quality of life was found to be comparable to a healthy reference population after stage I, with a decline for all dimensions after stage II, followed by improvement thereafter.Reference Williams, Gelijns and Moskowitz 8 Others reported on poorer health-related quality of life for single-ventricle CHD patients as early as at the age of 3 years compared with healthy reference.Reference Goldberg, Lu and Sleeper 9 This seems to persist throughout school age and adolescents with mixed CHD.Reference O’Connor, Wray and Tomlinson 6 , Reference Mussatto and Tweddell 11 Patients with single-ventricle CHD were found to be at greatest risk for poor perceived quality of life.Reference Marino, Lipkin and Newburger 2 , Reference Eagleson, Justo, Ware, Johnson and Boyle 5
Both, patient and medical factors, are known predictors for health-related quality of life. Despite complexity of the underlying CHD,Reference O’Connor, Wray and Tomlinson 6 , Reference Mussatto and Tweddell 11 an underlying genetic defect and family income are negative predictors.Reference Marino, Lipkin and Newburger 2 , Reference Werner, Latal, Valsangiacomo Buechel, Beck and Landolt 12 , Reference Cassedy, Drotar and Ittenbach 13 Other risk factors for poor health-related quality of life are predominantly treatment-related, such as longer circulatory arrest time, shorter time since last surgery or last hospitalisation, number of doctor visits, and tube feeding.Reference Eagleson, Justo, Ware, Johnson and Boyle 5 , Reference Williams, Gelijns and Moskowitz 8 , Reference Werner, Latal, Valsangiacomo Buechel, Beck and Landolt 12
The aim of this study was to examine the impact of patient-related and surgical variables on health-related quality of life at a time when disease burden is high due to mandatory surgical procedures.Reference Alsoufi, Mori and Gillespie 1 We hypothesised that children with complex CHD undergoing staged cardiac surgery in early life would have lower health-related quality of life compared to patients with less severe CHD.
Materials and methods
Setting and study population
This study was part of a two-centre study of the University Paediatric Heart Centre Giessen, Germany – Centre A – and the University Children’s Hospital Zurich, Switzerland – Centre B – among preschool-aged patients undergoing Fontan procedure for single-ventricle CHD.Reference Reich, Heye and O’Gorman Tuura 14 The study protocol conforms to the ethical guidelines as reflected in approval by both local ethics committee. Written informed consent was provided by parents or caregivers at study entry.
Using a longitudinal design, the original study population of children with single-ventricle CHD without genetic comorbidities enrolled between August 2012 and July 2015 – details were published previouslyReference Reich, Heye and O’Gorman Tuura 14 – was followed for health-related quality of life assessment until the age of 3–4 years. From the original cohort, two patients were excluded: one moved away and another refused from participation. The follow-up rate was 97.9%, and the participation rate was 95.8%. Demographics, cardiac diagnosis, and surgical treatment variables of the excluded children did not differ from the included children, the data are not shown.
Demographic measures and clinical data
Demographics were derived from patient charts. Growth measures were taken at birth and transformed to World Health Organisation growth standard deviation scores. 15 Small for gestational age was defined as weight or length lower than the 10th percentile of reference data. Family status, defined as married/living together, or divorced, and the socioeconomic status were assessed as proposed by Largo et al.Reference Largo, Pfister, Molinari, Kundu, Lipp and Duc 16 Socioeconomic status was determined based on maternal education and paternal occupation, each scored from one to six, resulting in a score ranging from two to 12, with higher scores indicating higher socioeconomic status. Perioperative variables, that is cumulative length of hospital stay, cumulative length of ICU stay, cumulative duration of mechanical ventilation until stage II, number of cardiac medication at admission for stage I, II, and III, and the occurrence of complications – such as re-operations, catheter re-interventions, need for extra-corporal membrane oxygenation, or re-animation – were derived from patient charts and entered into a pre-defined database as described earlier.Reference Reich, Heye and O’Gorman Tuura 14
Surgical procedures
Standard of care was a three-staged procedure. Palliation at stage I was either a classical Norwood procedure or a shunt palliation, including modified Blalock–Taussig shunt, systemic ventricular to pulmonary artery shunt, pulmonary artery banding, followed by bidirectional cavopulmonary anastomosis, also known as Glenn procedure, at stage II. Other children were followed by the hybrid approach, undergoing catheter-guided stenting of patent arterial duct and surgical bilateral pulmonary banding as stage I, followed by a comprehensive stage II procedure including removal of pulmonary artery bandings and the patent arterial duct stent, aortic arch reconstruction, and bidirectional cavopulmonary anastomosis. Two patients did not undergo the Fontan procedure/stage III (complete separation of passive pulmonary blood flow) at the time of health-related quality of life assessment: one patient was heart transplanted after comprehensive stage II, and one patient was still awaiting the stage III operation.
Neurodevelopmental testing and brain MRI
Patients underwent neurodevelopmental assessment and brain MRI scan at a mean age of 26 months and 1 week (SD 14 weeks) with the third version of the Bayley Scales of Infant and Toddler Development, Bayley-III,Reference Bayley 17 and a neurological examination by trained paediatric neurologists or developmental paediatricians. Severe developmental delay was indicated by Bayley-III scores below −2 SD, and mild/moderate developmental delay was defined as scores below −1 SD relative to the normative mean. The neurological examination focussed on motor functioning, graded as normal, reflex or tone abnormality, reflex and tone abnormality, or cerebral paresis, and on abnormalities in sight and hearing, with the overall result summarised in a single score.Reference Hajnal, Sahebkar-Moghaddam, Barnwell, Barkovich and Ferriero 18 Details on brain MRI and neurodevelopmental outcome were published earlier.Reference Reich, Heye and O’Gorman Tuura 14 , Reference Knirsch, Mayer and Scheer 19
Health-related quality of life
Health-related quality of life was assessed with the German version of the Pre-school Paediatric Cardiac Quality of Life Inventory by parental report.Reference Niemitz, Seitz and Oebels 20 All families were comfortable completing the questionnaire in German. The Paediatric Cardiac Quality of Life Inventory is a 51-item, five-scale generic instrument measuring the frequency of disease-related problems and severity of negative emotions about these by proxy report. The total score is calculated as the average score of the five scales. The physical scale assesses impact of the underlying CHD on physical capacity and functioning, for example, quality and quantity of physical activity, restlessness, and physical appearance, sweating, weight/eating problems, and frequency of pain experience. Disease impact on emotional wellbeing and behaviour is measured by the emotional scale, including anxiety, self-confidence, self-regulation, and reactions towards unfamiliar environment and medical interventions. The social scale assesses integration in the society and asks for the amount of peer contact, the amount of help needed, and the reaction of unrelated persons towards the child. The therapeutic scale assesses treatment burden, for example, implements needed, side effects of medication, how the child handles pain, and the frequency of medical visits. The functional scale estimates impact on everyday life, such as problem-solving approaches, restriction arising from the underlying CHD, apparent developmental delay, sleeping, recognition of performance limits, and the understanding of the heart disease. Higher scores relate to higher health-related quality of life in all scales. We used the computerised analysis tool, calculating transformed values of the raw data, ranging from 0 to 100. Of the reference group,Reference Niemitz, Seitz and Oebels 20 scores from 26 children with biventricular CHD, of which 16 were males, were in the target age range between 3 and 4 years, with mean age of 41 months 3 weeks, SD 13 weeks, and eligible for comparison.
Statistical analysis
Data were analysed with IBM SPSS Statistics for Macintosh, version 24.0, IBM Corp., Armonk, NY, United States of America. Descriptive statistics are presented as mean with SD or median with interquartile range for the respective dispersion of the data. Normality of data was tested with the Shapiro–Wilk test. Internal consistency of various measures was tested with the Cronbach’s α. The Welch’s t-test was applied to compare Preschool Paediatric Cardiac Quality of Life Inventory scores of our patients with the reference data of 26 patients with biventricular CHD. Between-groups comparisons were performed with χ2 tests in case of binary variables. To correct for multiple testing, a Bonferroni correction for group comparison was applied. Linear regression analysis was used to test for statistical relevance of association with health-related quality of life. The natural logarithm was calculated for non-normally distributed data. Collinearity among factors entered in the regression analysis was identified with Pearson’s or Spearman’s correlation for the respective dispersion of the data. Owing to the exploratory nature of the correlational analyses among candidate factors for the regression analyses, we did not adjust for multiple testing. In the first step, univariate linear regression for each factor was assessed. Variables with p-value <0.1 in the univariate analysis were included in the second step in a multi-variate linear regression model for each Preschool Paediatric Cardiac Quality of Life Inventory dimension. A maximum of four variables were allowed in each regression model as given by the sample size. A surgery severity score reflecting the broad spectrum of medical variables was thus calculated as sum of median split-half of each medical factors, such as cumulative length of hospital stay, cumulative length of ICU stay, mechanical ventilation, number of cardiac medications at stages I, II, and III, status post re-animation or need for extra-corporal membrane oxygenation, number of re-operations, and re-interventions, encoding the lower as zero and the higher as one. The score ranged from zero to nine, with higher scores indicating greater complexity. A two-sided significance level of p <0.05 was applied.
Results
Patient characteristics
Detailed patients’ characteristics, cardiac diagnosis, and surgical procedures are presented in Table 1. Mean age at follow-up of the 46 patients included was 38 months and 3 weeks (SD 14 weeks), with a median time after the patient’s last operation of 10 months and 3 weeks; interquartile range 6.67, 29.24 months. Demographic characteristics did not differ among the two centres, data not shown.
Patient characteristics (n = 46).
Data are presented as n (%), mean ± SD or median (interquartile range), SDS, standard deviation score.
* n = 42, Four children (9%) with balanced haemodynamic state because of pulmonary artery stenosis did not undergo stage I procedure and were later palliated with bidirectional cavopulmonary anastomosis, known as Glenn procedure, at stage II
** n = 44, one did not undergo stage III because of balanced haemodynamics; one patient received heart transplantation
Neurodevelopmental outcome and brain MRI findings
Mean Bayley-III cognitive composite score was 97, SD 12, ranging from 65 to 120. Severe and mild/moderate developmental delay was present in 4% (n = 2) and 7% (n = 3), respectively. All Bayley scales were intercorrelated: cognitive composite score with language composite score r = 0.69, p <0.001; cognitive composite score with motor composite score r = 0.75, p <0.001; language composite score with motor composite score r = 0.66, p <0.001. Neurological examination revealed mild abnormalities such as reflex abnormalities or muscular hypotonia in 16 patients (35%). None of the patients were subject to major neurological handicap, that is cerebral palsy, deafness, or blindness.
From 45 patients undergoing a brain MRI scan, 17 patients (38%) had at least one abnormality, including ventriculomegaly or atrophy (n = 10), status post infarction (n = 9), white matter injury (n = 5), suspected hypoxic brain injury (n = 2), with multiple findings in eight patients (18%). Brain MRI scan was not feasible in one patient due to an MRI-incompatible pacemaker.
Health-related quality of life and associated factors
Overall, health-related quality of life of our population was favourable (Table 2). Health-related quality of life of our single-ventricle CHD cohort was comparable to the reference data of patients with biventricular CHD. All differences were not significant both with Bonferroni correction and without.
Comparison of health-related quality of life between the sample and the reference cohort
Group differences were tested with the Welch’s t-test for continuous variables and with the χ2 test for binary variables.
All Preschool Paediatric Cardiac Quality of Life Inventory subscales showed acceptable to excellent internal consistency reflected by following Cronbach’s α values: total score: 0.94; physical: 0.85; emotional: 0.70; social: 0.85; therapy: 0.73; functional: 0.67 and were eligible for the risk factor analyses.Reference DeVellis 21 Given the high collinearity among Bayley-III scales, only the cognitive composite score was included for the regression analysis. Furthermore, the Apgar at 5 minutes showed high collinearity with the risk factor preterm birth (r = 0.50, p <0.001). The latter was included in the subsequent analysis. Results of the regressions analysis are given in Table 3. Perioperative variables were important determinants for all Preschool Paediatric Cardiac Quality of Life Inventory dimensions at 3–4 years of age in the univariate analysis (Supplementary Table S4). Interestingly, the emotional score was associated with head circumference at birth, but not with any perioperative candidate variables. The total score was associated with length of ICU stay, mechanical ventilation, and number of cardiac medication after stage II and stage III. The same variables were significant determinants for the physical score, which was additionally associated with cumulative length of hospital stay. Lower social, therapy and functional health-related quality of life were all associated with longer ICU stay, and greater number of cardiac medication after stage II and stage III, whereas a status of pre-term birth also predicted poor therapeutic health-related quality of life.
Associated variables for lower health-related quality of life
Bold indicates the variables entered in the multi-variate model. The surgery severity score reflects a summary of following surgical variables: cumulative length of hospital stay, cumulative length of ICU stay, mechanical ventilation, number of cardiac medications at stages I, II, and III, status post re-animation or need for extra-corporal membrane oxygenation, number of re-operations and re-interventions
ß unstandardised beta, SE, standard error, R2, adjusted coefficient of determination. Bold numbers represent significant associations. Of following non-normally distributed variables, the natural logarithm was used: socioeconomic status, total hospital stay, total ICU stay, cumulative time of mechanical ventilation, number of cardiac medication, number of re-operations and re-interventions. ECMO, extra-corporal membrane oxygenation
*p <0.05, **p <0.01, ***p <0.001, n = 46
‡ n = 45 (MRI scan not feasible in one patient with MR-incompatible pacemaker)
#n = 44 (one patient had heart transplantation, one patient awaiting stage III)
To adjust for the small sample size in the multi-variate model, the surgery severity score reflecting overall medical case complexity was entered. The score showed acceptable internal consistency (Cronbach’s α: 0.67). Multi-variate linear regression identified the surgery severity score as the only independent associated factor for the total health-related quality of life score, the physical score, and the social score (Table 3). Preterm birth and a higher surgery severity score were independently associated with lower scores in the therapy dimension. Preterm birth was the only remaining factor in the multi-variable analysis for the functional dimension, although there was no collinearity between preterm birth and the surgical severity score.
Health-related quality of life scores were neither associated with CHD-related pathological brain MRI findings, pathological neuroscore, nor with mild/moderate neurodevelopmental delay. Health-related quality of life was independent of socioeconomic status, family status, prenatal diagnose, small for gestational age or surgical approach (all p >0.05).
Discussion
In this study, overall health-related quality of life in young children with single-ventricle CHD was found to be favourable. In contrast to our hypothesis and findings of other studies, health-related quality of life was comparable to children with biventricular forms of CHD.Reference Marino, Lipkin and Newburger 2 , Reference Eagleson, Justo, Ware, Johnson and Boyle 5
Despite overall good health-related quality of life in children with mixed CHD at pre-school age,Reference O’Connor, Wray and Tomlinson 6 , Reference Mussatto and Tweddell 11 , Reference Werner, Latal, Valsangiacomo Buechel, Beck and Landolt 12 children with single-ventricle CHD seem to be at greatest risk for less favourable health-related quality of life.Reference Marino, Lipkin and Newburger 2 , Reference Marino, Tomlinson and Wernovsky 4 – Reference O’Connor, Wray and Tomlinson 6 Goldberg et al reported on lower quality of life accompanied by poor functional status in single-ventricle CHD in comparison to healthy controls at 3 years of age.Reference Goldberg, Lu and Sleeper 9 An important difference to this particular study despite the comparison to healthy children, is the exclusion of children with genetic syndromes in our study, which are known to be at greatest risk for impaired quality of life and the comparison to healthy children.Reference Werner, Latal, Valsangiacomo Buechel, Beck and Landolt 12
We identified treatment-related variables as important determinants of health-related quality of life, reflected in the association of the surgery severity score with the Preschool Paediatric Cardiac Quality of Life total score. This finding is in line with the studies investigating health-related quality of life reports of school-aged children and adolescents with mixed CHD, where the number of doctor visits and cumulative hospital stay were identified as significant associated factors of health-related quality of life,Reference Eagleson, Justo, Ware, Johnson and Boyle 5 , Reference Cassedy, Drotar and Ittenbach 13 whereas initial surgical risk score had no impact on outcome in adolescence.Reference Eagleson, Justo, Ware, Johnson and Boyle 5 These findings suggest that variability in complexity of the treatment courses influence prospective health-related quality of life, rather than the nature of the cardiac diagnose in a subgroup of single-ventricle CHD.
Pre-term birth was associated with lower treatment and functional dimension of quality of life. Despite being at risk for a complicated course, greater parenting stress in families of pre-term infants was previously shown.Reference Gray, Edwards, O’Callaghan, Cuskelly and Gibbons 22 These families are challenged with numerous medical and developmental needs in addition to the palliation of the underlying CHD. These additional stressors may influence parental report, as elevated state of anxiety was shown to alter perception of a child’s health-related quality of life.Reference Lisanti, Allen, Kelly and Medoff-Cooper 23 Interestingly, socioeconomic status and family status, as a surrogate for social integrity, were not associated with health-related quality of life in our cohort. The relatively high socioeconomic status, with only 20% parents with low socioeconomic status, and few number of divorced couples (4%) in our sample might explain this finding.
Neurocognitive functioning at 26 months did not predict health-related quality of life at 3–4 years of age. This is in contrast to the studies in older CHD patients showing deterioration of health-related quality of life at school-age and adolescence.Reference Alsoufi, Mori and Gillespie 1 , Reference Eagleson, Justo, Ware, Johnson and Boyle 5 , Reference Uzark, Jones, Slusher, Limbers, Burwinkle and Varni 7 , Reference Frigiola, Bull and Wray 24 The rather minor neurodevelopmental impairments found in our young population and the disability paradox might account for this finding to some extent.Reference Albrecht and Devlieger 25 However, minor functional disabilities may not be relevant until school-aged children grow into a more and more demanding environment, when executive functions, demanding communication skills, visuo-constructive functions, and fine motor skills become more important for academic achievement and peer interaction.Reference Mussatto and Tweddell 11
The results of this study have several implications for future research and clinical management. Treatment-related variables were associated with health-related quality of life. Thus, it is essential to provide comprehensive care during the vulnerable clinical phase with high treatment burden, and to identify patients with special needs. Standardised long-term follow-up assessments are warranted exploring the patients’ self-reported health-related quality of life of clinical significance and to plan for interventions if needed. These follow-up visits can ideally be combined with routine cardiac examinations at the same centre to minimise time and effort for the accompanying parents. Future research evaluating family emotions towards cardiac diagnosis and treatment will allow further insight in this underlying comorbidity in children with single-ventricle CHD.
The strengths of our study are the prospective study design and the homogeneous sample regarding cardiac diagnosis and age at assessment. However, this study is subject to several limitations. One limitation arises from the small sample size, limiting the statistical power of the study. From the original reference data,Reference Niemitz, Seitz and Oebels 20 26 cases with biventricular CHD aged 3–4 years at parental report were eligible for the analysis. The different group size is a limitation to the statistical analysis. The reference data were tested somewhat earlier, and improved patient care might have biased some of the results. Our multi-variate models explained 6–28% of variance in health-related quality of life. Additional factors that may also influence health-related quality of life such as financial problems or the parents’ well-being were not evaluated in this study. Although the patients’ demographic characteristics did not differ between the two study centres, some disparity in living condition among the two countries cannot be excluded. Another limitation arises from lacking details on parental stress, for example depression and anxiety levels. Parents stress levels have been shown to remain moderate to high in CHD patients hospitalised at the cardiac ICU after cardiac surgery regardless of the illness severity,Reference Franck, Mcquillan, Wray, Grocott and Goldman 26 and may alter their perception of the child’s health-related quality of life. Future research is warranted to further explore impact of these variables on the parents’ rating. The surgical complexity score used in the analysis was developed to identify patients with highest disease burden in this sample with complex CHD. The score did show statistical reliability, however, its application in other population warrants further investigations.
Conclusions
The complex treatment-related and the psychosocial environment in children with single-ventricle CHD and evaluation by parental report make it difficult to draw conclusions on health-related quality of life. However, our finding of comparable health-related quality of life in early childhood in single-ventricle CHD without genetic abnormalities compared to biventricular CHD is encouraging. Treatment-related variables and preterm birth are important determinants of health-related quality of life at 3–4 years of age. It is important to recognise social concerns and provide comprehensive support to the children and their families during the time of high disease burden. To better understand the evolution of health-related quality of life and its associated factors, a follow-up assessment at school age is planned.
Acknowledgements
We kindly thank the patients and their parents for their participation. Special thanks to Lutz Goldbeck and Mandy Niemitz for providing the German version of the Preschool Paediatric Cardiac Quality of Life Inventory.
Financial Support
This study has been funded by Fördergemeinschaft Deutsche Kinderherzzentren e.V., Bonn, Germany (grant number W-GI-005-2010), and the Mäxi Foundation, Zurich, Switzerland. The two sponsors, the Fördergemeinschaft Deutsche Kinderherzzentren e.V., Bonn, Germany, and the Mäxi Foundation, Zurich, Switzerland, had no influence on the study design, the collection, analysis, and interpretation of the data, on writing of the paper, and the decision to submit the paper for publication.
Conflicts of Interest
None.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1047951118001993
