Long-term Outcomes of Tetralogy of Fallot in the Kingdom of Bahrain.

Introduction: Tetralogy of Fallot (TOF) is the most common cyanotic congenital heart disease. Surgical correction has improved survival but re-intervention is often required.
Objectives: The objective is to assess outcomes after surgical repair of TOF, long-term follow-up, and factors that influence these results. Materials and
Methods: This is a retrospective study conducted in a tertiary care center. Records of patients diagnosed with TOF from 1992 to 2019 (37 years) were retrieved from a detailed database. Patients who underwent complete correction were grouped according to diagnosis, the technique utilized in surgical repair, need for staged repair, and syndromic association. Univariate actuarial and event-free survival analysis was performed. The endpoint for an event was death or re-intervention.
Results: A total of 230 patients were diagnosed with TOF and 174 patients underwent complete surgical repair. At 40 years postoperatively, survival was 96%. Actuarial survival was independent of syndromic associations, anatomical diagnosis, type of surgery, or previous shunt. Event-free survival (EFS) survival was 8.12%. EFS was significantly worse for patients with pulmonary atresia (PA) (Hazard ratio, 4.1125; 95% confidence interval [CI], 1.2654-13.3657; P < 0.0001) and for those that required homograft/conduit. The median duration for EFS was 22.73 years, 19.58 years, and 9.12 years for transannular patch (TAP), pulmonary valve-sparing (PVS), and homograft group, respectively. The survival curve for the PVS group merged with that of TAP 20 years postoperatively. Similarly, it merged at 22 years for staged versus primary repair and at 22.73 years for syndromic versus nonsyndromic patients. A weak correlation was found between age at surgery and event-free duration (cc, 0.309; P < 0.0001). The need for TAP was not influenced by the previous palliation, χ2(1, n = 154) = 3.36, P = 0.0667, or with interval to complete correction after the shunt procedure (P = 0.9672). Conclusions: Total correction of TOF has low perioperative mortality and good long-term survival, but the need for re-interventions is high. This study demonstrated that patients requiring homograft/conduit and those with a diagnosis of PA had worse outcomes. Comparison between different surgical groups showed merging of survival curves in follow-up that signifies gradual loss of survival advantage over time. Copyright:

INTRODUCTION

Tetralogy of Fallot (TOF) is the most common cyanotic congenital heart disease with an incidence of 4 per 10,000 live births.[1] Before corrective surgery, survival was only 50% beyond the first few years of life and very few patients reached adulthood.[2] The first corrective surgery of TOF was attempted by Lillehei in 1954.[3] From the earliest era of total correction, results were good with most of the patients reaching adulthood, however, long-term survival is still lower than that of the general population.[4] After complete correction, there is often a need for interventions for residual lesions.[567] This largely depends on whether the pulmonary valve was spared (PVS) during a complete correction or whether a transannular patch (TAP) or homograft/conduit is implanted.[56]

The incidence of TOF in Bahrain is higher than in international data (5.9/10,000 live births).[8] Here, the surgical treatment of TOF started with the insertion of the Blalock–Thomas–Taussig shunt (BTS) in 1986. The first total correction of TOF was done in 1995. Before that, patients were referred abroad. This study aimed to review our results of surgical repair of TOF and their long-term follow-up.

MATERIALS AND METHODS

In this retrospective study, the records of all patients diagnosed with TOF from 1992 to 2019 (37 years) were retrieved from a detailed database. Patients operated outside Bahrain but being followed up here were also included in the follow-up analysis. Patients with additional complex cardiac lesions were excluded from the study, for example, those with TOF and atrioventricular septal defects. Patients who underwent complete correction were compared based on diagnosis, i.e., TOF without pulmonary atresia (PA) versus TOF with PA, the type of surgery required, i.e., PVS, TAP, or homograft/conduit placement, and between patients with and without syndromic association. Survival analyses also compared patients who underwent palliative shunt procedures (categorized as staged repair) before complete correction to those with primary correction.

Statistical methods

Shapiro–Wilk test was applied to analyze the normality of quantitative data. Normal data are described as mean with standard deviations and analyzed with Student's t-test or one-way ANOVA. Non-normal data are represented as median and range and analyzed with the Kruskal–Wallis test or a Mann–Whitney U test as and when appropriate. Spearman rank correlation was utilized to assess the correlation between nonnormal quantitative data. Categorical variables were summarized as frequencies and percentages and analyzed with Chi-square or Fisher's exact test as and when appropriate. The probability of long-term survival was estimated using the Kaplan–Meier survival analysis method (KM analysis) and represented in graphs. Differences between groups were calculated by the log-rank test. Log-rank analysis was not utilized when survival curves crossed and the proportional hazards assumption did not hold. For analyzing long-term actuarial and event-free survival (EFS), steep, nonlinear, the first phase of survival curves was excluded which is due to mortalities experienced in the immediate perioperative period. The event was defined as death or re-intervention due to any cause. Statistical analyses were facilitated with the help of Data was analyzed by using SPSS Statistics for Windows, Version 23.0 (IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY: IBM Corp).

RESULTS

A total of 230 patients were diagnosed with TOF. 174 patients underwent complete surgical repair. Their demographic and surgical details are summarized in Tables 1 and 2. 39 patients had incomplete long-term follow-up data and were excluded from the long-term analysis. 14 patients are awaiting surgery and 3 patients died before surgical correction.

Table 1

Patient characteristics

Patient Demographics	n(%)
Number of patients diagnosed as TOF	230
Sex (male:female)	1.52 (139/91)
Median age at diagnosis (years)	0.44 (0-48)
Diagnosis
TOF with pulmonary stenosis	186 (80.9)
TOF with pulmonary atresia	38 (16.5)
TOF with absent pulmonary valve	6 (2.6)
Patients with syndromes (total)	56 (24.3)
22 Q deletion	30 (13.0)
Down’s syndrome	11 (4.8)
VACTREL anomaly	2 (0.8)
Trisomy 13	2 (0.8)
Multiple congenital anomalies	4 (1.7)
Kayler syndrome	3 (1.3)
Del 2p13	1 (0.4)
46XXX syndrome	1 (0.4)
Deletion/TX 12/14	1 (0.4)
Kabuki syndrome	1 (0.4)
Lost to follow up	39 (16.9)
Awaiting surgery	14 (6.09)
Expired before surgical correction	3 (1.3)
Complete repair	174 (75.6)
Staged repair	52 (29.9)
Primary repair	122 (70.1)
Median age at the time of palliation (years) (range)	0.17 (0.02-3.46)
Median interval to definitive correction (years) (range)	1.52 (0.52-11.86)
Median age at complete correction (years) (range)	1.43 (0.01-31.2)
Median follow up duration (years) (range)	5.42 (0-43.9)
Number of patients with events after complete correction	38 (23)
Median event free survival (years)	21.85

TOF: Tetralogy of fallot

Table 2

Surgical details of tetralogy of fallot patients categorized according to type of surgery

	Transannular patch (n=96), n (%)	Pulmonary valve sparing (n=57), n (%)	Homograft/conduit (n=21), n (%)	P
Sex ratio (male: female)	1.28	1.47	2.5	0.43
Syndromic association	25 (26.0)	13 (22.8)	8 (38.1)	0.39
Staged repair	30 (31.2)	9 (15.8)	13 (61.9)	<0.001
Diagnosis
Pulmonary stenosis	88 (91.7)	57 (100)	9 (42.8)
Pulmonary atresia	8 (8.3)	0	9 (42.8)
Absent pulmonary valve	0	0	3 (14.3)
Perioperative mortality	6 (6.3)	2 (3.5)	0
Staged repair	3 (3.1)	0	0	0.71
Primary repair	3 (3.1)	2 (3.5)	0
Median age at the time of palliation (years) (range)	0.24 (0.02-1.07)	0.08 (0.04-0.49)	0.08 (0.02-3.46)	0.20
Median interval to definitive correction (years) (range)	1.25 (0.52-8.7)	1.55 (0.97-7.28)	2.34 (0.87-3.07)	0.02
Median age at complete correction (years) (range)	1.35 (0.01-13.6)	1.45 (0.28-31.2)	2.45 (0.08-17.7)	0.24
Median follow up duration (years) (range)	5.9 (0-43.9)	2.3 (0-32.9)	2.5 (0-20.4)	0.02
Number of patients with events	23 (23.9)	7 (12.3)	8 (38)	0.03
Number of events	39	9	12
Median event free survival (years)	22.73	19.58	9.12

Age at surgery

A gradual decline in surgical age was observed over the study period, except for a few patients who were referred late [Figure 1].

Figure 1

Trend for age of complete correction over years

Survival

Details of patients who underwent complete surgical correction are described in Table 2. 24 patients who were operated on before 1995 were referred outside Bahrain for surgical correction and are also included in the follow-up analysis. Total perioperative mortality was 4.6%.

Long-term survival

The median duration of follow-up was 5.42 years (range from 0 to 43.9). Actuarial 5-, 10-, 20-, 30-, and 40-year survival rates in 166 patients who survived the perioperative period were 97.4%, 96%, 96%, 96%, and 96%, respectively [Figure 2]. During follow-up, 3 mortalities were experienced in the TAP group (1 due to sepsis, 1 during re-intervention, and 1 sudden cardiac death) and 1 sudden cardiac death in the homograft group.

Figure 2

Kaplan–Meier graph actuarial and event-free survival (years) probability (event = re-intervention or death) in patients who survived the perioperative period after complete correction. (Vertical lines represent censored patients, number of patients at risk are displayed in table below the graph)

Survival according to diagnosis

Among PS patients, the need for TAP was influenced neither by previous palliation, χ2(1, n = 154) = 3.36, P = 0.0667 nor with interval to complete correction after the shunt procedure (P = 0.9672). Of 166 TOF patients who survived the perioperative period following complete correction 17 were PA and 3 were absent pulmonary valve (APV). All the patients had excellent long-term actuarial survival irrespective of diagnosis. Patients with APV were insufficient for detailed long-term survival statistical analysis.

Survival according to the type of surgical procedure

Patients were grouped according to the type of surgery (TAP, PVS, or homograft) for long-term survival comparison. All three groups had excellent actuarial survival with insignificant variation [Figure 3].

Figure 3

Survival following different types of TOF repair. Perioperative mortality (years) after TOF repair is excluded. (Vertical lines represent censored patients, number of patients at risk are displayed in table below the graph). TOF: Tetralogy of Fallot

Survival according to previous palliative operations

Around one-third of the patients (n = 52 [29.9%]) underwent primary palliation (peripheral shunt or central shunt) before complete correction [Table 2]. 37 (24%) patients with PS and 15 (88.2%) patients with PA underwent staged repair. The median age at the time of palliation was around 2 months (range, 0.02–3.46 years) and the median interval to definitive correction was 1.52 years (range, 0.52–11.86 years). Five patients required a second shunt surgery. Actuarial survival was indifferent in comparison between primary repair and staged repair groups.

Survival according to the syndromic association

Fifty-six patients (24.3%) had syndromic associations described in Table 1. Among them, 46 (82.1%) underwent complete correction, of which 8 (17.4%) had homograft, 13 (28.3%) had valve-sparing surgery, and remaining 25 (54.3%) underwent TAP. Both syndromic and nonsyndromic patients had similar long-term actuarial survival of > 96%.

Event-free survival

Details of all the events are described in Table 3. EFS at 5-, 10-, 20-, 30-, and 40-years was 86.2%, 77.7%, 56.3%, 24.4%, and 8.12%, respectively. Median EFS was 21.85 years (95% confidence interval [CI], 18.930–43.960 years) [Figure 2]. A significant but weak correlation was found between age at surgery and event-free duration (cc, 0.309; P < 0.0001).

Table 3

Detail of events (re-intervention or death) after complete correction for patients who survived perioperative period

Events	Total patients with complete correction (n=166)	Transannular patch (n=90)	Pulmonary valve sparing (n=55)	Homograft (n=21)
Total events	60	39	9	12
Number of patients with events (%)	38 (22.8)	23 (25.5)	7 (12.7)	8 (38.1)
Pulmonary valve replacement	27	20	4	3
Conduit/homograft replacement	13	5	2	6
Permanent pacemaker implantation	4	3	1	-
Stenting/balloon dilatation of pulmonary arteries	4	1	1	2
Radiofrequency ablation	3	3	-	-
Relief of right ventricular outflow tract obstruction	2	1	1	-
Coiling of major aorto-pulmonary collaterals	1	1	-	-
Balloon dilatation of pulmonary stenosis	1	1	-	-
Cardioversion	1	1	-	-
Death	4	3	-	1

Event-free survival according to diagnosis

EFS for PA patients after the complete correction was worse than for patients with classical TOF [Figure 4](Hazard ratio, 4.1125; 95% CI, 1.2654–13.3657; P < 0.0001). Median survival was only 8.9 years for PA patients compared to 22.73 years for pulmonary stenosis (PS) patients.

Figure 4

Event-free survival (years) for patients according to diagnosis. Perioperative mortality after TOF repair is excluded. (Vertical lines represent censored patients, number of patients at risk are displayed in table below the graph). TOF: Tetralogy of Fallot

Event-free survival according to the type of surgical procedure

The median duration for EFS was 22.73 years, 19.58 years, and 9.12 years for TAP, PVS, and homograft groups, respectively. Although there was an early survival benefit of preserving the pulmonary valve during surgery over other methods [Figure 5]. At a postoperative duration of 19.58 years, this survival benefit is lost and survival curves crossed. All the patients needed some re-intervention by, 32 years in the PVS group and by 43 years in the TAP group. 16 (17.8%) patients in TAP group, 2 (3.6%) in PVS and 4 (19.0%) in homograft group required multiple re-interventions.

Figure 5

Event-free survival following different types of TOF repair (years). (Perioperative mortality after TOF repair is excluded. Vertical lines represent censored patients, number of patients at risk are displayed in table below the graph). TOF: Tetralogy of Fallot

Event-free survival according to previous palliative operations

Median EFS for patients with staged repair (15.12 years) was lower than primary repair (21.85 years) and crossing of the curve was observed at 21.85 years [Figure 6].

Figure 6

Event-free survival (years) for patients with or without primary palliation. Perioperative mortality after TOF repair is excluded. (Vertical lines represent censored patients, number of patients at risk are displayed in table below the graph). TOF: Tetralogy of Fallot

Event-free survival according to the syndromic association

Median survival was 22.730 years for patients without syndromic associations, whereas it was only 9.12 years for patients with syndromes with an intersection at 22.73 years [Figure 7].

Figure 7

Event-free survival (years) for patients with or without syndromic association. Perioperative mortality after TOF repair is excluded (Vertical lines represent censored patients, number of patients at risk are displayed in table below the graph). TOF: Tetralogy of Fallot

DISCUSSION

TOF is the most common cyanotic congenital heart disease and its successful surgical correction started as early as the 1950s.[3] Studies describing the long-term follow-up in Middle Eastern countries are scarce. This study presents our long-term experience after the correction of TOF in Bahrain.

We had perioperative mortality of 4.6% which is equivalent to others that ranges from 3% to 7.8%.[910111213141516] The gradual reduction in the mean age of surgery over the study period was seen and is similarly described by others.[1617] It is the direct result of increasing experience and confidence of the surgical team over time, in the management of pre- and post-operative cases.

Patients who survive the perioperative period have excellent long-term actuarial survival.[61516] In this study, a 96% actuarial survival probability at 40 years reflects similar findings. This is irrespective of syndromic associations, diagnosis, type of surgery, or previous shunt. Actuarial survival with EFS showed a gradual decline with time and a 96% actuarial survival at 40 years postoperatively correlates with only 8.12% of EFS [Figure 2]. This indicated that although almost all of the patients are alive at long-term follow-up, few have an uneventful course. These results are comparable to previous studies describing a high rate of re-interventions in follow-up after TOF correction.[714171819] There was also a weak correlation between the age of surgery and event-free duration.

Studies have contradicting results over survival analysis between different surgical options. Some conclude that there are no significant differences.[4141516202122232425] Our study corroborates with Murphy et al.,[4] and others as long-term actuarial survival was similar in all surgical groups irrespective of technique utilized in repair.

Similar results cannot be described for EFS as most of the patients in our study, irrespective of the type of surgery needed a second procedure after long-term follow-up. Our data shows a non-sustained early EFS benefit of PVS during surgery over TAP which eventually equalized at around 20 years in follow-up [Figure 5]. All the patients in the PVS or TAP groups required re-intervention in long term. Other studies have also described no difference in long-term follow-up among these groups.[7162326] PVS is not as advantageous as previously thought and the loss of survival advantage of PVS over TAP can be attributed to the progressive nature of pulmonary regurgitation (PR).[2427] It has also been reported earlier that as much as 20% of patients in the PVS group can develop moderate PR within 1st year after surgery.[72426]

Homograft/conduit placement has a worse prognosis compared to PVS and TAP [Table 2]. This finding shows that the homograft and TAP group require re-interventions earlier than the PVS group [Figure 4] and also that the re-intervention rate and need for multiple re-interventions are higher in these groups [Tables 2 and 3]. Similar findings have also been reported by others who have demonstrated a significantly higher risk of reoperation in the TAP group, as early as 5 years after complete correction.[14151720212228]

A mild right ventricular outflow tract (RVOT) gradient is shown to be advantageous in reducing the need for future pulmonary valve replacements (PVR),[29] thus inclination towards preservation of pulmonary valve and intentionally leaving a small gradient in RVOT, has led to a higher incidence of residual RVOT obstruction and lower EFS in PVS group.[71725] However, this policy is still not utilized at our center.

This study is unique in that it shows the crossing of survival curves in long-term, which signifies that the effect of the type of surgery is not constant and the survival benefit is lost over time. This can explain the contradicting results of earlier studies. We also found a significant difference in follow-up duration among these three categories which is mainly driven by the TAP group [Table 2]. The fact responsible for this observation is that in the early surgical era surgeons had a preference for TAP and utilization of PVS or homograft strategy was implemented later.

Evidence favoring primary palliation versus primary repair is contradictory with studies supporting either strategy.[14153031323334] Even though earlier studies have shown that BTS may help in annular growth and reduce the requirement of TAP during complete repair,[313334] this finding was not true in our study. In this study, a significant difference was observed in the need for staged surgery between different surgical groups (P < 0.001) which was mainly driven by the PA-homograft group [Table 2]. Also as there is a preference for delayed repair if homograft is utilized, it resulted in a significant increase in the interval to total correction [Table 2]. Earlier studies have described a higher requirement of PVR in patients with staged repair during follow-up.[14172228] Our follow-up analysis did show some short-term benefits after complete correction but it got negated with age and EFS probability became similar by 22 years' postoperative. This initial difference can be the result of a biased selection of cases for staged repair among whom the effect of poor anatomy decreases as the child grows and thus in the long-term outcome is equivalent with either strategy [Figure 5].

In our study, the need for PVS or conduit in syndromic patients was comparable to earlier studies.[3536] Michielon et al. and others had described lower survival rates and lower freedom from re-intervention in syndromic patients.[61635] In this study, after an initial decline in EFS for syndromic patients, the survival curves crossed 22.73 years. As only 1 syndromic patient was in follow-up at that time, this result might not represent the true picture and needs to collaborate with further studies for a definitive outcome [Figure 6].

In comparing PA with classical TOF, Hickey et al. described that patients with PA have 3 times more risk for late death and only half of the patients were alive 40 years postoperatively.[16] Other studies have shown that there is a higher risk for reoperation and PVR in PA patients.[1516183738] In this study, actuarial survival was not as poor as described by Hickey et al., but we did find significantly poor EFS [Figure 4]. Median EFS in patients with PA was way worse than classical TOF. As these patients are at the poorer end of the TOF spectrum to begin with and most of them require repair with homograft/conduit [Table 3], it results in high re-intervention rates.[39]

To summarize, TOF patients have an excellent actuarial survival after complete correction but the need for repeated re-interventions is common. In all the categories except for homograft and PA, the univariate analysis showed no difference in EFS in long term. This group had significantly worse EFS and an increased need for re-interventions.

Limitations

The retrospective study design and long duration of the study period are one of the major limitations of this study. Diagnostic methods, technology, indications, and timing for (re) interventions all have changed over the study period, which could have influenced our results. Retrospective nature further contributes to the difficulty in gathering detailed medical records with technical details, intra-operative information, and other confounding variables such as preoperative and postoperative outflow gradients, length of patching, and cardiopulmonary bypass timings which can influence the outcome. Further, a decision regarding the type of surgery is determined by the surgeon's preference which could have introduced potential selection bias, as children treated by a PVS procedure have friendlier RVOT anatomy, resulting in better functional outcome and prognosis. Finally, most of the patients were operated on recently and the number of patients in follow-up for more than 20 years is less which may further add to inconsistencies in analysis and difficulty in generalization of the outcome.

CONCLUSIONS

The present study adds to the knowledge that total correction of TOF can be performed with low perioperative mortality and good long-term survival, but they still have high re-intervention rates in follow-up. Our study provides important information about long-term outcomes in different surgical groups. The study is unique in the way that it showed a gradual loss of survival advantage of sparing pulmonary valve or of doing a primary complete repair. It also showed that patients with homograft or PA patients have a significantly worse outcome. Despite medical advances, TOF remains a life-long disease with long-term morbidity. It is not cured by complete repair and continuous surveillance of this cohort is important.

Financial support and sponsorship

Nil.

Conflicts of interest

Abhinav Agarwal (none), Suad R Al Amer (none), Habib Al Tarif (none), Aieshah Ahmed Ismael (none) Neale Nicola Kalis (none).