Aims of this study were to evaluate
1. the long-term survival of patients operated on for a congenital heart defect during childhood
2. the morbidity of CHD patients
3. the causes of late deaths of CHD patients
4. the life situation of CHD patients late after surgery
METHODS
Number of patients and operations
The Research Registry of Paediatric Cardiac Surgery was established with permission of the Finnish Ministry of Social Affairs and Health in 1995. The long-term data collection began that year with the aim of producing a to database including all patients operated on during childhood for a congenital heart defect in Finland. Child was defined as a person less than 15 years of age at the time of the first cardiac operation.
The patients and procedures were collected from various data sets. The data of all operations conducted at five university hospitals (Helsinki, Turku, Tampere, Oulu and Kuopio) and in one district hospital (Aurora Hospital, Helsinki) were obtained. The diaries of the operation theatre, the diaries of the children’s ward and surgeons’ own records were used to find the operations. Once the procedures were found, the patient records were read to check the data of the diaries.
The procedural data of years 1953–1989 covered the whole of Finland. The numbers of the 1990s and 2000–2007 included only patients operated on at Helsinki University Central Hospital. In the 1990s, about 1400 patients were operated on at other University Hospitals. Since 1998, all paediatric open-heart surgeries and all heart surgeries for neonates were centralized to Helsinki. Thus, the figures of the 21st century covered practically all of Finland. The only procedures missing from 2000–2007 were the closures of PDA in premature babies (about 400) and about 100 PDA closures and COA repairs for older children.
Patients were given one primary diagnosis to facilitate comparison of survival between defects. Heart defects were sorted hierarchically (Table 3); patients with more than one defect were given the diagnosis highest on the list. For instance, a patient with COA, VSD and stenotic mitral valve was classified as VSD. The hierarchy was based on hierarchies published previously (Fyler 1980, Carlgren et al. 1987).
Patients’ current status and address were obtained from the Finnish Population Registry. The date of death or emigration was also provided to us. The address was needed, since we aimed to find the patients last contact with secondary health care, which gave us the information about reoperations.
Univentricular heart
Hypoplastic left heart syndrome Truncus arteriosus
Interrupted aortic arch
Transposition of great arteries Atrioventricular septal defect
Total anomalous pulmonary venous drainage Pulmonary atresia with intact ventricular septum Tetralogy of Fallot
Ventricular septal defect Coarctation of the aorta Aortic stenosis
Pulmonary stenosis Mitral valve disease
Partial anomalous pulmonary venous drainage Atrial septal defect, secundum
Patent ductus arteriosus Other
Table 3. Hierarchy of diagnosis.
All data were stored in a computer running the custom-made research registry program ProCardio (Melba Group, Helsinki, Finland), based on FileMaker™ Pro 3.0 (Claris Corp, CA, USA). The current version of ProCardio is 8 and it runs under FileMaker™
Pro 8.5 (FileMaker Inc, CA, USA).
Mortality and survival (I, III)
The early mortalities (death within 30 days of first operation) of all patients and patients with PDA, ASD, COA, VSD, TOF, TGA and UVH were presented as percentages (I).
Late mortality was assessed with survival analysis, which was done twice, excluding and including early deaths (I). Survival was calculated for all patients and separately for gender and diagnostic groups. The survival rate of patients operated on in different decades was also calculated separately to evaluate the progress of treatment (III). We compared the survival rates of different gender and decade groups, and analysed the survival with and without non-CHD-related causes of death.
We compared each patient’s individual survival with that of the age-, sex- and time-matched general population (I). For every patient and for every year of follow-up, we obtained the mortality of the general population. The expected probability of surviving a year was obtained for all individuals who, regardless of their survival status, were not censored prior to the beginning of the interval. We used the Hakulinen method (Hakulinen 1982) for combining the individual information with an estimate for the expected survival of the group.
Cardiac morbidity (I, II) and causes of late deaths (III)
Operations per patients (I) and use of cardiac medicines (II)
Cardiac morbidity was assessed by the number of cardiac operations needed per patient and by usage of cardiac medications. The cardiac operations performed on patients after childhood were searched for from the hospital records. The usage of medications was investigated by linking the patients’ data from the Research Registry of Paediatric Cardiac Surgery to the data of the Social Insurance Institution’s registry of special reimbursement entitlements and medicine purchases.
The prevalence of chronic diseases was determined from the registry provided by the Finnish Social Insurance Institution. Individuals with certain specific chronic diseases (about 80 diseases) are entitled to special reimbursement for the drug treatment costs of that disease. To get the entitlement for special reimbursement (ESR), a patient applies for it from the Social Insurance Institution, and the application requires a certificate from a physician giving the details of diagnosis and disease treatment. The Social Insurance Institution centrally records the data of the ESRs. The four chronic cardiac diseases that entitle the patient to receive special reimbursement are heart failure, hypertension, arrhythmia and coronary heart disease.
The actual usage of medicines was determined from the medicine registry of the Social Insurance Institution. The medicine registry contains information on all details and costs of purchases of prescribed medicines. The records are based on on-line computerized information provided to the Social Insurance Institution by all Finnish pharmacies. In the registry, the medications are categorized by Anatomical Therapeutic Chemical (ATC) classification (WHO Collaborating Centre for Drug Statistics Methodology 2004). Use of a medication was defined as at least one purchase during 2004.
For each patient in our study population, two matched control subjects were chosen from the population registry. Matching criteria were sex, age (± 2 years) and hospital district in which the patient lived.
Causes of deaths (III)
The causes of late death were identified from death certificates obtained from Statistics Finland. If something unclear appeared on the death certificate, we examined the patient’s medical records.
The causes were divided into CHD-related and non-CHD-related causes. CHD-related causes included deaths caused by diseases with ICD-10 (International Classification of Diseases) diagnosis numbers Q20 to Q28 (in older cases ICD-9 numbers 745 to 747). All other causes were classified as non–CHD-related, also the rare cases of acquired heart disease.
The causes of CHD-related deaths were evaluated further by dividing them according to an existing classification of the cardiac cause of deaths (Oechslin et al. 2000).
The four classes of cardiac death were heart failure and sudden, perioperative and cardiovascular death. Sudden death was defined as death due to cardiovascular causes within 1 hour of onset (or significant worsening) of symptoms or unwitnessed death during sleep (Myerburg et al. 1992). The perioperative deaths included all early postoperative (within 30 days) deaths due to the patient’s second, third or fourth operation. The cardiovascular deaths included all CHD-related deaths that could not be classified into the other categories.
Comorbidity and causes of non-CHD-related deaths (II, III)
Comorbidity was assessed by the usage of medications. The methods are described in detail under the heading Cardiac morbidity. Data of the causes of deaths were gathered as mentioned above (see Causes of deaths section).
The non-CHD-related causes of death of the patients were compared with the cause-specific mortality of the general population. The results appear as numbers of deaths observed and expected and as standardized mortality ratios (SMRs). SMR expresses the ratio of observed causes of death to that expected on the basis of overall mortality rates in the general Finnish population. The expected numbers of deaths were standardized for age, sex and time. The Poisson distribution was used to calculate the 95% confidence intervals. The significance of SMR was calculated using the approximation of Byar for the exact Poisson test (Breslow et al. 1987).
Life after surgery (IV)
Life situation of the patients was studied with a mailed questionnaire designed to elicit information about the patient’s general state of health, exercise tolerance, education, occupation, marital status, offspring and cardiac follow-up. The questionnaire was designed to be easy and quick to complete and unambiguous to interpret.
Each operating hospital obtained permission to conduct this part of the study. They sent the inquiry to their own patients, and the results were combined.
The educational, employment and family statuses were compared with these of the general population, using data from the national statistical centre, Statistics Finland.
The expected values with 95% confidence intervals were calculated as weighted averages of published age- and gender-specific rates. The calculations were carried out separately for each patient group. In the education section, patients under 20 years of age were excluded.
Statistical methods (I–IV)
The survival statistics were calculated with SURV3, the latest version of the survival analysis package developed at the Finnish Cancer Registry (Hakulinen et al. 1985).
Relative risk (RR with 95% confidence intervals) analysis was used in evaluating the differences between the usage of medicines among patients and controls (II), and between observed and expected numbers of deaths (III). The differences were significant if the confidence intervals of RR did not include 1.0.
Student’s unpaired t-test was used to compare the ages of patients and controls, and the ages of patients in different groups and statuses. The Chi-square test was used to compare frequencies between patients and the general population (II, IV).
RESULTS
Number of patients and operations (I)
A total of 7240 cardiac operations were performed on 6461 children during the first 37 years of cardiac surgery (1953–1989). The number of operations increased regardless of a concurrent decrease in the Finnish child population (Figure 2).
The patients’ age at operation averaged 5.1 years (median 4.5; range 0–15). The mean age decreased during the study period from 8.9 to 3.4 years and the median even more, from 9.0 to 1.7 (Table 4). Since 1990, the age at operation declined further.
In the Hospital for Children and Adolescents, the median age of operated children was less than one year (Table 4).
400
Number of children x 1 000 Operations
1953 1957 1961 1965 1969 1973 1977 1981 1985 1989
0
1-15 years
< 1 month 1-12 months
Figure 2. Number of cardiac operations in Finland in 1953–1989. Bars are divided by patients’ age at operation. The continuous line represents the number of children (<15 years) of the whole population during the study years.
The patient material varied over time, as more defects became surgically treatable. In the 1950s, most patients were operated on for PDA. During the subsequent decade, the number of operations and the variation in defects increased rapidly as open-heart surgery began. The number of patients with different defects operated on in different decades is presented in Table 5. The development continued. During the 21st century the proportions of patients with PDA and ASD diminished, while the proportion of miscellaneous defects rose (Table 5). The most common miscellaneous defects were atrioventricular septal defect (AVSD) and aortic stenosis during 1953–1989 and AVSD and hypoplastic left heart syndrome (HLHS) during 2000–2007.
Age, years Operative mortality No. of
operations mean median n %
1953–1959 329 8.9 9.0 11 3.3
1960–1969 1505 7.1 7.3 87 5.8
1970–1979 2337 5.5 5.3 193 8.3
1980–1989 3069 3.4 1.7 208 6.8
1953–1989 7240 5.1 4.5 499 6.9
1990–1999* 3131 2.3 0.7 218 7.0
2000–2003* 1138 2.0 0.4 50 4.4
2004–2007* 1173 2.1 0.5 27 2.3
* Procedures conducted at Helsinki University Hospital Table 4. Age at operation and operative mortality.
Mortality and survival (I, III)
Early mortality (I)
During the decades of the earlier era (1953–1989) the operative mortality varied from 3.3% to 8.3%, with an overall average of 6.9% (Table 4). In the 1990s a slight rise occurred in early mortality, as increasingly complicated patients came into the reach of operative treatment. During 2000–2003 practically no defects were beyond the operative range. Thus, the operative mortality of 4.4% was excellent. In recent years, the operative mortality has decreased even more (Table 4).
When focusing on patients, the results looked slightly different. A total of 433 patients (7%) died within 30 days of their first operation (Table 5). The operative mortality was lowest (0–1%) in the ASD group throughout the years. The mortality increased as the defects became more complicated; in the UVH group, the operative mortality was 26%
during the earlier era, and 8% (5/60) during 2000–2007 (Table 5).
The strides in development of treatment were also seen in the operative mortality of HLHS: 1953–1989, it was an inoperable defect. When new techniques (e.g. Norwood operation) were taken into practice in the 1990s the operative mortality at Helsinki University Hospital was 60% (25/42). During 2000–2003 the early mortality was 23%
(11/47), decreasing further in 2004–2007 to 5% (2/40).
Late mortality (I, III)
A total of 592 patients out of the 6028 early survivors operated on during 1953–1989 died late (>30 days) after surgery. The mortality was highest shortly after the first operation. During the first year 164 (28%) deaths occurred, almost half of them (n=79, 48%) during the first three months after surgery. The patients who died were
1953–1989Operative1950s1960s1970s1980s n(%)mortality, n (%)n(%)n(%)n(%)n(%) Patent ductus arteriosus2050(32)49(2)250(77)701(48)498(24)601(23) Coarctation of aorta926(14)33(4)19(6)196(13)369(18)342(13) Atrial septal defect812(13)8(1)2(1)151(10)324(16)335(13) Ventricular septal defect767(12)61(8)--115(8)260(13)392(15) Tetralogy of Fallot453(7)40(9)33(10)114(8)131(6)175(7) Transposion of the great arteries298(5)40(13)--25(2)114(5)159(6) Univentricular heart137(2)35(26)1(0)20(1)42(2)74(3) Miscellaneous defects1018(16)167(17)19(6)139(10)336(16)524(20) All6461(100)433(7)324(100)1461(100)2074(100)2602(100) 2000–2003*Operative2004–2007*Operative n(%)mortality, n (%)n(%)mortality, n (%) Patent ductus arteriosus82(11)8(10)107(13)3(3) Coarctation of aorta102(13)1(1)93(11)0(0) Atrial septal defect54(7)0(0)91(11)0(0) Ventricular septal defect128(17)2(2)152(18)1(1) Tetralogy of Fallot57(7)2(4)64(7)0(0) Transposion of the great arteries51(7)2(4)57(7)1(2) Univentricular heart35(5)2(6)25(3)3(12) Miscellaneous defects253(33)21(8)250(30)9(4) All762(100)38(5)839(100)17(2) * Procedures conducted at Helsinki University Hospital
Table 5. Number of patients and proportion of defects divided by the time of first operation. Operative mortality in different defects and eras.
significantly younger at the time of their first operation than the patients who survived, 3.9 (median 1.3, range 0–15) vs. 5.4 (median 5.1, range 0–15) years (p<0.001). The mean age of late death was 13.3 years (median 10.2, range 0.1–53). The patients died on average 9.4 years (median 5.8, range 0.08–39) after their first operation.
The cause of death was identified in 582 of the 592 patients (98%), and was confirmed with post-mortem examination in 474 cases (81%). Ten death certificates (2%) were unavailable, and thus, the cause of death was not determined. A total of 397 deaths (67%) were related, and 185 (31%) were unrelated to congenital heart defect. In most diagnostic groups, CHD-related causes of death dominated. Only patients with PDA and ASD died more often due to non-CHD-related causes (Figure 3).
1 numbers of deaths in each diagnostic group. PDA=patent ductus arteriosus, COA=coarctation of aorta, ASD=atrial septal defect, VSD=ventricular septal defect, TOF=tetralogy of Fallot, TGA=transposition of the great arteries, UVH=univentricular heart, Misc=miscellaneous defects.
Late survival (I, III)
Compared with the general population (I)
The overall survival for 45 years was 78% for the patients and 93% for the general population. The average longevity of the patients was ten years shorter than that of the general population, 58 and 68 years, respectively. After excluding operative mortality patients’ survival increased to 84%, and the average longevity to 62 years.
The negative effect of the disease gradually disappeared over the years and patients’
risk of death became similar to that of the control population. This was seen as planar or slightly rising curves of relative survival (Figure 4). With severe defects, the relative survival continued to decrease. The sharp drops of these curves were caused by the small numbers of living patients. population is 100%. Numbers beside the curves indicate the number of patients alive at 25, 36 (COA) and 38 (TOF) years post-surgery. Follow-up ended in October 1998.
PDA=patent ductus arteriosus, COA=coarctation of aorta, ASD=atrial septal defect, VSD=ventricular septal defect, TOF=tetralogy of Fallot, TGA=transposition of the great arteries, UVH=univentricular heart.
Late survival in different diagnostic groups in different decades (III)
The late survival rates of all patients operated on in 1953–1989 decreased decade after decade (Figure 5). This reflects that more children with severe congenital heart defects were treated during the later decades. In the 1960s, 1347 patients were operated on and most exhibited simpler defects than the 2418 patients operated on in the 1980s.
The late survival of those patients with PDA who underwent surgery in the 1980s was significantly poorer than that during earlier decades (p<0.001). This result was due to the difference in the number of neonates undergoing surgery: 31% (186/601) in the 1980s, and only 4% (22/498) in the 1970s. The difference in survival rates between decades in PDA patients disappeared when non-CHD-related causes were excluded from analysis (Figure 5), indicating that most of the neonates were small, premature babies dying from non-cardiac causes. The same effect of neonates was seen also in the analysis of all patients, because PDA was the biggest diagnostic group (23%, 601/2602).
As expected, the survival of patients with COA, ASD and VSD improved over the first decades of cardiac surgery in Finland (Figure 5). The late survival of patients with TOF who underwent surgery in the 1980s was worse than of those operated on in the 1970s. The increase in mortality was mainly a result of patients in the 1980s being younger at the time of their operation (mean age 2.8 vs. 6.3 years, p<0.001). Also patients operated on in the 1980s the impact of age was significant. The mean age at operation for survivors was 3.0 years, and for those who died late after surgery 1.1 years, p=0.015. The results of TOF patients seemed to get better during 1990–2003, even though the mean age at operation (0.9 years) was lower than during earlier decades. The 16-year survival was 96%, while during 1970s and 1980s it was 94%
and 90%, respectively. However, these differences were not statistically significant.
50
Coarctation of aorta Atrial septal defect
Ventricular septal defect Tetralogy of Fallot Survival (%)Survival (%)Survival (%)
Years after surgery Years after surgery
*
*=the patient died 15 years after closure of PDA, VSD was not surgically treated
p=0.023
Figure 5. Survival of patients after their first operation excluding operative mortality.
Division is by the decade in which the patient underwent the first surgery. The dotted green line in the figures for all patients and for patients with patent ductus arteriosus represents survival in the 1980s, when only CHD-related deaths were included in the analysis. Follow-up started from the patients’ first operation and ended in October 1998.
The survival of patients with TGA improved within time. The 15 year survival in 1970s was 72% and in 1980s 87%. The positive development of treatment continued during the later eras. The mid-term survival of the patients operated on for TGA during the 1990s and at the beginning of the 21st century was better than earlier, as seen in Figure 6.
The late survival of patients with UVH was quite poor. The 19-year survival for those operated on in the 1970s was better than for those operated on in the 1980s, 68% vs.
55%. The higher mortality in the 1980s is explained by the complexity of the defects operated on. For example, during the 1980s, eight (14%) of the 56 operated UVH patients had a combination of UVH, TGA, and COA, whereas in the 1970s only one (3%) of the 33 patients had this complex defect. During the 1990s and in 2000–2003 a trend towards better late survival of UVH patients emerged (Figure 6). However, the difference was not significant. Transposition of the great arteries
1970s 1980s
Years after first operation Years after first operation
Figure 6. Survival of patients with transposition of the great arteries and univentricular heart after their first operation excluding operative mortality. Division is by the decade in which the patient underwent the first surgery. Follow-up started from the patient’s first operation and ended in February 2007.
Cardiac morbidity (I, II) and causes of late deaths (III)
Operations per patients (I) and use of cardiac medicines (II)
The studies of cardiac morbidity and causes of late deaths included all patients operated on during 1953–1989.
Many patients needed more than one surgical procedure to survive. Altogether 949 patients were operated on several times; these comprised 16% of those 5774 who survived their first operation and had a complete follow-up. The number of operations needed varied according to the defect. PDA or ASD rarely required reoperation, 16%
of patients with COA, every third VSD, TOF and TGA patient, and 80% of patients with UVH were operated on at least twice.
A total of 17% (855/5116) of patients and 8% (791/10232) of control subjects used medication for cardiac diseases, RR 2.16 (Cl 1.97–2.37). The patients using cardiac medicines were significantly younger than controls (mean age 38.2±0.4 vs. 44.9±0.4 years, p<0.001). Patients using medication for cardiovascular diseases were older and had been operated on at an older age than the other CHD patients (mean age 38.9±0.4 vs. 32.3±0.2 years, p<0.001, and mean age at first operation 6.4±0.2 vs.
A total of 17% (855/5116) of patients and 8% (791/10232) of control subjects used medication for cardiac diseases, RR 2.16 (Cl 1.97–2.37). The patients using cardiac medicines were significantly younger than controls (mean age 38.2±0.4 vs. 44.9±0.4 years, p<0.001). Patients using medication for cardiovascular diseases were older and had been operated on at an older age than the other CHD patients (mean age 38.9±0.4 vs. 32.3±0.2 years, p<0.001, and mean age at first operation 6.4±0.2 vs.