Summary of principal findings

The principal finding of this study was that 30-day mortality following primary surgical repair of TOF at our institution during the 10 years was 8%.

The patients who died had a longer duration of cardiopulmonary bypass time, aortic cross-clamp time, ventilation time post-procedure, low preoperative oxygen saturations, RV dysfunction, and LV dysfunction and were more likely to have been operated by the local team alone.

Mortality incidence between high and LICs

Data on outcomes following TOF repair in LMICs remain scarce with the total number of included patients being few compared to HICs. A systematic review and meta-analysis on outcomes after surgical repair of TOF reported an overall 30-day mortality of 4.2% in Africa and the Middle East combined, but only included data from Egypt and Ivory Coast [12]. The few other studies from Africa reporting on perioperative mortality following TOF repair report much higher mortality ranging from 9.0 to 12.9% [8,9,10]. The overall reported mortality in our patient cohort appears to be slightly lower, but it is significantly higher than data from HICs. Some of the studies from HICs reported no mortality in their cohorts [7, 13].

Differences in patient characteristics between LMICs and HICs

The patient demographics differ between LICs and HICs. Patients from LICs tend to be older with lower oxygen saturation and higher haematocrit. This is important to consider because of the impact of chronic cyanosis as a known risk factor for poor postoperative outcomes.

A study from China, an upper middle-income country, observed that lower pre-operative saturations (median 82%, IQR 72–88 versus median 89%, IQR 81–95) were associated with poor outcomes [14]. In the Ethiopian study by Tefera et al., the median pre-operative saturation of patients who died was 76%, compared to 86% in survivors; however, the difference was not statistically significant [9]. These saturations are still much higher than the preoperative saturations in this study cohort. In another study conducted in Cameroon, a lower-middle-income country, the mean oxygen saturation before surgery was as low as 67 ± 5% [8].

In low-middle-income countries, there are few modified Blalock-Taussig shunts (MBTS) recommended to improve saturations, and when done, they tend to perform poorly.

Patients with chronic hypoxia are at increased risk of arrhythmias and other complications due to polycythaemia, such as strokes.

Patients with TOF operated from LMICs tend to be older than those from HIC. Most of the LMICs operate on TOF patients after one 1year of age [8,9,10]. Older age at surgery has been linked to higher mortality due to chronic hypoxia and RV failure [5].

Factors associated with mortality

Our study links mortality to several factors, including time on cardiopulmonary bypass, aortic cross-clamp duration, mechanical ventilation time, low pre-op oxygen saturation, RV and LV dysfunction, and the operating team.Due to delayed surgeries, these children may require significant muscle bundle resection. This can result in extended periods of being on cardiopulmonary and aortic cross-clamp, making them more susceptible to health complications.

Late surgeries can increase the risk of RV failure due to chronic cyanosis. In our study, all patients who passed away experienced RV failure, and in two cases, patients had to be kept on the pump to support the RV due to lack of access to ECMO. These issues contributed to longer cardiopulmonary time.

Patients who died were sicker, required more ventilatory support, and had prolonged ventilation time. Our findings are comparable with studies in LMICs. In the Ethiopian study, prolonged cardiopulmonary bypass time/aortic cross-clamp time and pulmonary valve annulus diameter less than three standard deviations (SD) were independently associated with perioperative mortality [9]. A study by Amirghofran et al. in Shiraz, Iran found a correlation between mortality with pump and ventilation time post-procedure [15]. In Berne Switzerland, found age, cardiopulmonary bypass time, aortic cross-clamp time and higher maximum post-operative troponin levels [16].

One of the common denominators among patients who died in our study was acute kidney injury (AKI) requiring dialysis. This could be attributed to limited access to haemodialysis at UHI, with only peritoneal dialysis available at that time. AKI can also be linked to RV failure. Factors may be due to delayed surgery, but age at surgery was not directly linked to mortality probably due to the small sample size.

Patient selection

The provision of paediatric cardiac surgical services in LMICs has historically been limited to short-term surgical visits from HICs and published data on outcomes from these initiatives are also limited. While the local team operated on all the patients who died in our study, this is unsurprising given the deliberate strategy of selecting patients who are expected to have favourable outcomes to be operated on by the visiting team for teaching /training, while the local team tends to operate sicker patients in comparison.

Patient selection plays a vital role in determining operative outcomes for paediatric cardiac surgery in LMICs like Uganda. With limited resources, prioritizing patients who are likely to benefit most, such as those with favourable anatomy is crucial. This likely influenced more favourable outcomes than if we were operating on all patients, including those with more complicated anatomy. The fact that our patients are older, with a median age of 4 years, suggests that their anatomy was not as complex, enabling them to reach this age. However, older age is associated with more complications, particularly RV dysfunction, following surgery. In our setting, ECHMO is not available to support those patients who may require it.

It has been noted that in low-income countries, TOF surgery is performed on older children as compared to high-income countries. In a study from Ethiopia by Tefera et al., the median age was 7 years with an age range of 1–22 years, while in a study from Cameroon by Tchoumi et al., the mean age was 9.18 +/- 6.5 years with a range of 13.5 months to 26 years [8, 9]. This was not different from this cohort.

In some surgical procedures for TOF repair, patients may experience medium to long-term outcomes. For example, TAP repair relieves RVOTO but may lead to severe pulmonary regurgitation in the long run which may require re-operation. The Finnish research database looked at 600 patients who underwent TOF repair before 15 years of age and in their findings TAP repair carried a high risk of re-operation but had no impact on late survival [17]. Patients with adequate pulmonary valve annulus i.e. PV annular z-sore >-2 are usually preferred during patient selection to avoid the need for a TAP. In our study TAP repairs were few because these children were older with favourable anatomy i.e. many had good pulmonary valve annulus, MPA and branch PAs.

In comparison, a study by Tefera et al. at the children’s Heart fund cardiac centre run in collaboration with international charities and philanthropists, reported that 22/57, 38.6% of patients had a trans annular patch and 2 patients had RV-PA homografts inserted [9].

Similar to our context, this program was initially dependent on international surgeons, but now local teams operate. Data from small cohort of patients undergoing surgical repair in Cameroon, also reported a high incidence of repair with TAP (12/22, 54.5%). However, these surgeries were supported by international surgeons [8]. Patients with adequate pulmonary valve annulus i.e. PV annular z-score >-2 are usually preferred during patient selection at our institution to avoid the need for TAP. In our study, TAP repairs were few because these children were older with favourable anatomy i.e. many had good pulmonary valve annulus, MPA, and branch PAs.

Morbidity

The most frequent causes of morbidity in our study were pleural effusions, arrhythmias, and postoperative infections. This was comparable with what was found in other LMICs in Africa.

In a study conducted in Nigeria by Olukemi et al., pleural effusion was found to be the leading cause of morbidity, followed by pericardial effusions and cardiac dysfunction [10]. In the study conducted in Cameroon, only a small number of patients (22) were involved, and the acute post-surgical complications that were recorded included pericardial effusion and pleural effusion in 4 and 3 patients respectively [8]. It’s possible that the low incidence of complications was due to the small sample size.

Arrhythmias are common in TOF patients, typically within 24 h [18]. In most of our patients, we were able to successfully correct the rhythm disturbances.

The causes of morbidity in HIC are quite like those found in the LMICs. A study conducted by Sameh Ismail et al. at King Saudi University also identified pleural effusions and arrhythmias as the primary causes of morbidity [13].

Suggestions for improvement

Improving access to cardiac surgery can help reduce wait times and enable children to undergo surgery at an earlier stage. Early diagnosis plays a critical role in identifying patients who require surgery. To achieve this, hospitals should encourage routine screening of newborns for oxygen saturation levels after birth, a practice that is not currently widespread.

The success of any surgical team depends on the expertise and skills of its members. To ensure exceptional performance, it is crucial to train more personnel, including cardiologists and cardiothoracic surgeons. These additional team members will not only enhance the surgical team’s capabilities but also improve the quality of care provided to patients. Investing in training now will pay dividends in the future, resulting in better patient outcomes and a more efficient healthcare system overall.

It is imperative to increase the number of surgeries performed. Doing so will significantly improve system efficiency, leading to a considerable reduction in operating time, including CPB and aortic cross-clamp times. Moreover, surgeries must be made accessible to both smaller babies and those with complex anatomy without any delay.

ECMO is highly recommended in our setting to reduce mortality due to RV failure. However, it requires personnel training and sustainable costs.

Study strengths and limitations

This study provides the first documented report of early surgical outcomes of tetralogy of Fallot patients following primary intracardiac repair in Uganda.

The reported mortality rate cannot be considered entirely accurate due to the absence of data for 16 patients.

The fact that this was a single-centre study/ only centre where open-heart surgery is done in the country limits the generalizability of our results.

There was bias during patient selection, only those who qualified for primary intracardiac repair were taken.

Echocardiography timing varied based on the patient’s condition. However, at least one echocardiogram was done before discharge, suggesting no consideration for postoperative remodelling.

Some key statistics could not be measured given the retrospective design.

Medium- and long-term outcomes were not included in this study. A larger sample size was needed to be able to describe rare outcomes.