Factors associated with health-related quality of life among adults with tetralogy of Fallot

SWEDCON register

The register was started in 1992 and since 1998 it has covered all seven healthcare regions in Sweden. Since 2009, the register also includes paediatric cardiology and congenital heart surgery. In the end of 2012, the registry contained almost 10 000 adults. Data are collected by each centre and contains information about diagnosis, interventions, demographics, medication, symptoms, New York Heart Association (NYHA) functional class, ECG, exercise tests, self-reported level of physical activity, echocardiography pacemaker/implantable cardioverter defibrillators and QoL assessed by EQ-5D.

Symptoms refer to cardiovascular symptoms (palpitations, syncope, dyspnoea, fatigue, oedema and chest pain). NYHA functional class is reported by the staff.

The register does not contain information about hospitalisations other than for surgical interventions.

Patient selection

This registry study obtained data from SWEDCON. Inclusion criteria were patients with ToF ≥18 years and valid EQ-5D questionnaire. All data available until 19 February 2013 were searched. The data from the last available clinical visit/test were used.

Data on patients with ToF were extracted and the file was manually validated case by case to verify that only patients with complete repair of ToF were included. Initially, 558 patients were identified. Based on intervention codes and concomitant diagnoses, 28 patients were excluded considered having more complex lesions or other diagnosis than ToF (figure 1).

Request permission

In the remaining 530 patients, EQ-5D data were available in 288 who met the inclusion criteria and were included in the analysis.

The patients excluded due to lack of EQ-5D data did not differ regarding age, sex, body mass index (BMI), NYHA class, symptoms or level of physical activity (table 1).

EQ-5D

The EQ-5D questionnaire was developed by the EuroQoL Group and enables the respondent to classify their health according to five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression.

Each dimension has three levels describing no problems, some problems or severe problems.

Thus, 243 different health states can be described. The information derived from the questionnaire can be converted in to a single summary index (EQ-5D_index) by applying a formula weighting all levels in each dimension using an index tariff. Since there is no Swedish EQ-5D_index value set, the British index was adopted as reference.17 EQ-5D_index equal to 1 represents perfect health and 0 worst possible health status or even death.18 Using these calculations, it is possible to generate values below zero, often replaced by zero, but this does not apply in the present study since minimum score here was zero.

Self-reported level of physical activity

In the SWEDCON registry, self-reported level of physical activity is defined as sports/exercise performed more or less for its own sake. There are three levels: none, <3 h/week or >3 h/week.

Statistical analysis

The data were assessed for normality. Data are presented as ratios (percentage), means with SD and ORs with 95% CIs. For comparison between groups, χ² test was used for categorical variables and independent samples t-test for continuous variables. The outcome variable EQ-5D_index was dichotomised into EQ-5D_index=1 or EQ-5D_index<1.

Independent variables were assessed using logistic regression. The independent variables were reviewed and considered for associations. Univariate logistic regression was used to select variables for multivariate regression. Variables with a significant p value in univariate analysis were entered in multivariate analysis. In addition, smoking was entered in the model due to clinical relevance and borderline significance in univariate mode. The statistical analysis was performed using IBM SPSS statistics V.23.0. A p value <0.05 was considered significant.

Discussion

Here, we report that HRQoL was associated with NYHA class, physical activity level, presence of symptoms and cardiovascular medication in a large, register-based population of patients with ToF. This stresses the importance to carefully assess these variables during periodic outpatient follow-up. Furthermore, this is the first time an association between a high physical activity level and best possible HRQoL is reported in adult patients with ToF.

Physical activity level was associated with HRQoL, which is in line with our previous report on adult patients with aortic valve disease.16 Associations between exercise test performance and general health has also been reported, and it may be that physical activity level is a marker of high HRQoL.19 Earlier studies20 21 have reported improvement in QoL after exercise intervention, and it is tempting to speculate that interventions targeting physical activity level would improve HRQoL in patients with ToF as well. Randomised clinical trials to test such a hypothesis are therefore warranted.

The NYHA class was strongly associated with HRQoL. In this study, we analysed staff-reported NYHA. Earlier findings suggest that there is a strong association between self-reported NYHA and HRQoL.8 22 This underscores the importance of correctly assessing this variable during follow-ups and that staff-reported NYHA here seems to be as valid as self-reported NYHA in that aspect.

Presence of symptoms were associated with HRQoL but not as strongly as NYHA. There are multiple reasons for this: first of all, many of the patients (60%) with worse self-reported health status (EQ-5D<1) were reported as asymptomatic. Whether this reflects the truth or is a consequence of deficient evaluation of symptoms during follow-up is not possible to tell.

Second, NYHA relates to limitations in physical activity whereas symptoms may occur unrelated to physical exercise.

Ongoing cardiovascular medication was associated with HRQoL in one of the multivariate models and borderline in the other, but with similar point estimates. Medication is probably an indicator of more severe disease but may also itself affect HRQoL through side effects, both physical and psychological. Furthermore, daily intake of drugs reminds the person of the chronic cardiac disease that not necessarily causes daily symptoms or experience of limitations.23 24

We found no association between age and HRQoL in the multivariate analysis. Since it is well known that morbidity increases with age in the ToF population,7 one could have expected an association. The lack of association in multivariate mode may be due to the strong associations between NYHA, physical activity and HRQoL outweighing the weaker variable age.

Study limitations

In registry studies, there are certain factors that can influence the results. Data are delivered from several sites and by many individuals, which potentially may reduce the validity of data. On the other hand, the large population likely compensates for such possible variations.25 Missing data pose a potential problem. In the current study, 242 patients were excluded due to missing data on EQ-5D. This is, however, most likely explained by changes in routines regarding registration of this variable. Comparison between the groups with and without available EQ-5D data did not show any differences with regard to demographics or medical factors. However, NYHA class and presence of symptoms was borderline significant with a slightly higher proportion of symptoms and a lower proportion of NYHA class I in those lacking EQ-5D data (table 1). Therefore, inclusion of the 242 patients without EQ-5D data would likely have strengthened the OR for these variables. We find it unlikely that this effect would have been strong enough to expel the physical activity variable from the model. The register does not contain information on hospitalisations other than for cardiac interventions. Therefore, the present study potentially underestimates factors that may influence on HRQoL.