Surgical outcomes of infective endocarditis in children: should we delay surgery for infective endocarditis?

Abstract

 

Open in new tabDownload slide

OBJECTIVES

We compared the surgical outcomes of infective endocarditis (IE) between early surgery and non-early surgery groups in children.

METHODS

From January 2000 to April 2020, we retrospectively reviewed 50 patients <18years of age who underwent first surgery for IE. Early surgery was defined as that performed within 2 days for left-sided IE and 7 days for right-sided IE after diagnosis.

RESULTS

The median age and body weight at operation were 7.7 years [interquartile range (IQR), 2.3–13.2] and 23.7 kg (IQR, 10.3–40.7), respectively. The median follow-up duration was 9.5 years (IQR, 4.0–14.5). In 28 patients with native valve endocarditis, the native valve was preserved in 23 (82.1%). The most common causative microorganism was Streptococcus viridans (32.0%). The operative mortality was 2.0%, and 13 (26.0%) patients required reoperation most commonly for prosthesis failure (n = 7). There were no significant differences in patient characteristics and perioperative data between early surgery (n = 9) and non-early surgery (n = 36) groups, except for the interval between diagnosis and surgery (early surgery < non-early surgery, P < 0.001) and preoperative negative blood culture conversion (early surgery < non-early surgery, P = 0.025). There were no significant differences in overall survival, recurrent IE, and reoperation rate between the groups. Early surgery and preoperative negative blood culture conversion were not found as significant factors for surgical adverse outcomes.

CONCLUSIONS

Surgical outcomes for IE in children were acceptable irrespective of the time of surgery. Our results suggest that it may not be required to delay surgery for IE and the potential benefit of early surgery could be expected in children.

INTRODUCTION

Infective endocarditis (IE) is a potentially serious condition that can lead to significant morbidity and mortality [1, 2]. Although IE is relatively rare in children, its incidence appears to be increasing [1, 2]. The increased incidence of IE is related to improved survival rates among children at risk for IE, including those with congenital heart disease and hospitalized newborn infants [1, 2]. The complexities of patient management in neonatal and paediatric intensive care units also have increased the risk of IE in children with normal structural hearts with prolonged central indwelling venous catheters [1, 2].

However, in children, no randomized controlled trials have been performed to evaluate the treatment of IE. Therefore, many indications are based on consensus [2]. In addition, since the majority of children with IE are expected to survive solely with conservative medical treatment, candidates for surgery should be selected cautiously [3]. Therefore, surgery for IE is more likely to be avoided in children compared to adults, even in cases of children at risk of haemodynamic compromise, embolism, failure of antibiotic therapy [3, 4]. Furthermore, hesitations in the liberal use of the prosthetic valves in children may affect the above tendency.

Early surgery during the acute phase of infection has been reported with good outcomes in adults. It is also likely that optimal surgical timing is crucial for improved outcomes in children [5]. However, optimal timing of surgery is still contentious in children. Delaying surgery allows for the longer use of antibiotics that may eradicate microorganisms in the bloodstream and lead to haemodynamic stabilization. However, it incurs the risk of disease progression with valve destruction, abscess formation, heart block, embolic complications and even death [6].

The surgical outcomes of IE in children have rarely been reported. Moreover, clinical data on the impact of surgical timing on the outcomes of IE in children are lacking. We reviewed and compared the surgical outcomes of IE between early surgery and non-early surgery in children.

MATERIALS AND METHODS

Patient selection

This retrospective study was approved by the institutional review board (No.: H-1909-025-1061). The requirement for informed consent was waived.

From January 2000 to April 2020, 67 patients underwent their first surgery for IE in our institution. Five patients, although not preoperatively diagnosed with IE, were included due to clear evidence of IE found intraoperatively. There were 73 consecutive cases of IE, of whom, 4 required surgery more than once due to recurrent IE. Patients >18 years of age (n = 14), those with severe neurological damage (n = 1) and those with lost medical records (n = 2) were excluded. Thus, 50 patients were enrolled. Patients were identified using the modified Duke criteria for definite IE [7].

Treatment strategy and definitions

IE was managed by a multidisciplinary team as per previous existing guidelines [2, 8, 9]. Except for the 5 patients diagnosed intraoperatively, all patients received preoperative antibiotics for a median of 18 days [Interquartile range (IQR), 5.5–32.5] after clinical diagnosis. Perioperatively, adequate intravenous antibiotics were administered based on causative microorganisms and guidelines [2, 8, 9]. Postoperatively, the median duration of antibiotic therapy was 34.5 days (IQR, 15.3–42.0). In several patients with shorter hospitalization, antibiotics were fully administered even after discharge.

The surgery for IE consisted of complete debridement and the repair or replacement of infected structures, if needed. Preserving intra-cardiac structures and native valves rather than replacing with prosthetic valves were preferred in native valve endocarditis (NVE). Prosthetic valve endocarditis (PVE) was defined as infection on parts of the valve prosthesis or reconstructed native heart valves [10]. One case of native valve infection associated with adjacent prosthetic material infection was included in PVE.

Early surgery was defined as surgery within 2 days for left-sided IE (LSIE) with the involvement of systemic circulation and within 7 days for isolated right-sided IE (RSIE) after clinical confirmation [11, 12]. IE that occurred in the Fontan conduit was regarded as RSIE because it was directly related to pulmonary circulation. The date of diagnosis of IE was determined based on the date on which the patient met the modified Duke criteria [7]. If the indications for surgery were met, delays in surgical treatment were avoided to minimize complications associated with IE.

The surgical indications of IE [2, 8, 9] were as follows: (i) acute decompensated heart failure, (ii) uncontrolled infection and (iii) high embolic risk, including mobile vegetation.

Evaluation of clinical outcomes and end points

Operative mortality was defined as death within 30 days of operation or during the same hospitalization. Late mortality was defined as death after discharge. IE was defined as active if patients presented a fever or positive blood cultures persisted regardless of the use of adequate intravenous antibiotics before surgery. Preoperative complications on the echocardiographic findings were diagnosed based on the guidelines [2, 8, 9].

The end points were all-cause mortality after surgery, recurrent IE, and reoperation. Other surgical outcomes, including valve replacement, were also evaluated and compared between the early surgery (E) and non-early surgery (N) groups. For between-group comparison, the 5 patients intraoperatively diagnosed with IE were excluded to minimize bias. Regular postoperative follow-up was performed at 3–6-month intervals.

Statistical analyses

Continuous variables were expressed as the mean ± standard deviation or the median (IQR). Categorical variables were presented as frequency (%). The differences in baseline characteristics and clinical factors between the groups were evaluated using Student’s t-test or Wilcoxon rank sum test for continuous variables according to normality assumption by the Shapiro–Wilk test and using the chi-square test or Fisher’s exact test for categorical variables. Fisher’s exact test was used if >20% of cells with an expected value of <5 were in a contingency table.

The logistic regression model and Cox proportional-hazards model were applied to identify risk factors for adverse outcome-related events. The univariable analysis was used to select an initial screening of candidate risk factors. Significant variables of P-value <0.2 in the univariable analysis were included in the multivariable analysis. A stepwise selection method was performed to evaluate clinical outcome risk factors. Variables with P-value <0.05 were regarded eligible to enter the test, and those with P-value >0.10 were removed. Survival and event-free survival rates were estimated using the Kaplan–Meier method. The log-rank test was performed to compare the survival curves between groups.

Statistical analyses were performed using IBM SPSS statistical software (version 25.0, IBM Inc., Armonk, NY) and SAS statistical software (SAS system for Windows, version 9.4; SAS institute, Cary, NC). All P-values were two-tailed, and P < 0.05 was considered statistically significant.

RESULTS

Baseline characteristics and perioperative data

Baseline characteristics and perioperative data are summarized in Table 1. The IE-involved side is presented in Supplementary Material, Table S1.

The median age and body weight at operation were 7.7 years (IQR, 2.3–13.2) and 23.7 kg (IQR, 10.3–40.7), respectively. The tricuspid valve was mostly involved (24/50, 48.0%). No patients had a history drug abuse or percutaneous valve replacement. Surgery for IE was performed earlier in patients with LSIE than in those with RSIE (median, 5 vs 20.5 days). The median follow-up duration was 9.5 years (IQR, 4.0–14.5).

Preoperative blood culture was performed in 46 (92.0%) patients. Of 5 patients diagnosed with IE intraoperatively, 4 patients did not present clinical symptoms or signs suspicious for IE preoperatively. Negative blood culture conversion was confirmed preoperatively in 30 (60.0%) patients. Residual vegetation after embolic episode despite appropriate antibiotic therapy or persistent vegetation with high embolic risk was the most common indications for surgery (39/50, 78.0%).

Of 28 patients with NVE, the native valve was preserved in 23 (82.1%) patients (atrioventricular valves in 18, semilunar valves in 9, and both in 4); 54.5% for LSIE and 100.0% for RSIE, respectively.

The causative microorganisms and preoperative blood culture results according to IE-involved side or early surgery are summarized in Supplementary Material, Tables S2 and S3. The most common causative microorganism was Streptococcus viridans (16/50, 32.0%).

Clinical results

Overall, all-cause mortality occurred in 2 (4.0%) patients, and operative mortality was 2.0% (1/50). The cause of death was irreversible septic shock, even though the surgery was performed 3 days after the clinical diagnosis with RSIE. The patient underwent corrective surgery for a double-outlet right ventricle with a subaortic ventricular septal defect and was diagnosed with RSIE on postoperative day 10. Adequate antibiotics were administered for the causative microorganism (Staphylococcus aureus). We performed early surgery due to haemodynamic instability with huge vegetation near-completely occluding the tricuspid valve. However, the patient died due to septic shock after 2 days of surgery. Of 49 early survivors, there was late mortality in 1 case. The patient underwent surgery 58 days after the diagnosis with RSIE. The cause of death was accidental respiratory arrest due to the aspiration of food material during follow-up at postoperative 48.8 months. The overall survival rates at 1, 5 and 10 years were 97.8%, 94.9% and 94.9%, respectively.

There were recurrent IE in 4 (8.0%) patients; 3 had previous surgery for fungal IE and 1 had surgery for IE by S. aureus. All recurrent IE occurred >6 months after surgery (at postoperative 6, 19.1, 21.2 and 67 months). The causative microorganisms isolated in recurrent IE were the same as previous causative microorganisms in 2 patients. Of these, 1 patient underwent reoperation due to IE combined with conduit failure at postoperative 21.2 months, and the others were medically treated. The probability of freedom from recurrent IE at 1, 5 and 10 years was 97.6%, 92.5% and 88.6%, respectively.

Thirteen (26.0%) patients underwent reoperation during follow-up. The causes of first reoperation after surgery for IE were as follows: prosthesis failure associated with stenosis (n = 7), native valve failure (n = 5) and recurrent IE (n = 1). The probability of freedom from reoperation at 1, 5 and 10 years was 95.5%, 77.8% and 72.9%, respectively.

In NVE, 8 patients ultimately underwent valve replacement during follow-up: 5 at the time of operation for IE (4 valve replacement, 1 Ross procedure) and 3 during follow-up due to progressive regurgitation of repaired native valve. The probability of freedom from overall valve replacement at 1, 5 and 10 years was 80.7%, 76.9% and 71.4%, respectively.

Early versus non-early surgery

Baseline characteristics and perioperative data

At the timing of surgery, there were no significant differences predominantly in baseline characteristics and perioperative data between groups E (n = 9) and N (n = 36); however, group E underwent surgery earlier than did group N (median, 3 vs 20.5 days, P < 0.001) and had a lower rate of negative blood culture conversion (33.3% vs 75.0%, P = 0.025) (Table 1).

Early and late clinical outcomes

Early results were comparable between groups E and N (Table 2), and no significant differences existed between the groups when divided by the IE-involved side (Supplementary Material, Table S4). Among 45 patients, 1 late mortality (group E 0 versus group N 1) and 4 recurrent IE (group E 1 versus group N 3) occurred. Eleven patients (group E 3 versus group N 8) required reoperation. In the Kaplan–Meier curve, no significant differences in overall survival, recurrent IE, reoperation due to a cardiac cause or overall valve replacement in NVE were found between the groups (Figs. 1 and 2).

Of 45 patients (group E 9 versus group N 36), 1 late mortality (group E 0 versus group N 1) and 4 recurrent IE (group E 1 versus group N 3) occurred. Eleven patients (group E 3 versus group N 8) required reoperation.

Risk factor for adverse outcomes

LSIE (P = 0.030) was a risk factor for native valve-sparing failure in NVE. In addition, preoperative cerebral infarction (CI) (P = 0.010) was a risk factor for postoperative CI. Younger age (P = 0.002) and perivalvular extension of infection (PVEI) (P = 0.003) were significant risk factors for reoperation. Fungal infection (P = 0.029) was a significant risk factor for recurrent IE. The significant risk factor for overall valve replacement in NVE was PVEI (P = 0.023). However, early surgery and preoperative negative blood culture conversion did not affect surgical adverse outcomes significantly (Table 3 and Supplementary Material, Table S5).

DISCUSSION

Reports on surgical outcomes of IE in children have been limited to small series of patients. Hickey et al. [13] reported surgical outcomes of IE in 30 children and suggested that the survival and functional outcomes were favourable and native valve preservation was frequently possible. Russell et al. [14] reported clinical outcomes of 34 patients who underwent surgery for IE; the overall mortality was 15%, and 28% (10/34) of the patients underwent reoperation mainly due to somatic growth. Khoo et al. [4] reported good surgical outcomes of IE showing high survival, low recurrent IE, and low reoperation rates in children.

Despite several reports on improved surgical outcomes over time, how and when to treat IE in children is still controversial. Hence, surgery for IE in children tended to be delayed and reserved for children presenting medically intractable heart failure with haemodynamic instability, a high risk of an embolic event, or failure of antibiotic therapy [3].

The role of early surgery in adults with IE has been studied. A recent randomized trial demonstrated that early surgery performed within 48 h after the diagnosis of IE effectively reduced systemic embolisms without increasing operative mortality or recurrent IE [11, 12]. Kang et al. [15] demonstrated that the beneficial effect of early surgery was maintained during long-term follow-up, and late clinical outcome after surgery was excellent. The European Society of Cardiology guidelines (2015) clearly recommend that early surgery should be performed as soon as possible if indicated, in case of complicated IE (heart failure, uncontrolled infection and high embolic risk) [9].

In children, Nomura et al. [16] suggested that early surgical intervention would be required in children with active or acute infection despite advancement in antibiotic therapy. Khoo et al. demonstrated that surgery performed during the active phase of IE did not increase the risk of mortality or reoperation. The median interval between the diagnosis and the surgery was 10 days (LSIE versus RSIE, 5 vs 28 days) [4]. Shamszad et al. suggested that early surgery can be successfully performed in children with high rates of native valve repair and low rates of recurrence and mortality. The median interval to surgery was 3 days, and patients who underwent surgery within 3 days and >7 days after diagnosis were compared [5].

Nevertheless, there is no consensus on the optimal timing of surgery and the impact of early surgery on the clinical outcomes of IE in children. Surgical therapy during the active phase of the disease could be associated with the significant risk. However, reasons for considering early surgery in the active phase include avoiding progressive heart failure, preventing from systemic embolism and minimizing irreversible damages of valvular structures that might be an important issue for young children because of the limitations on the use of prosthetic valves replacing the destroyed native valve [9]. Given the above considerations, the potential benefit of the early surgery in children should be weighed against its operative risks and long-term consequences.

Hence, we sought to evaluate the risk–benefit of early surgery for IE in children. Currently, there is no clear definition or consensus of early surgery for IE worldwide and there are differences between the European Society of Cardiology and the American Heart Association guideline. Facing an issue of the definition of early surgery, we concentrated on the expectable benefits and basic surgical indications of early surgery. First, we focused on the high systemic embolic risk in LSIE, considering that embolic events are important prognostic indicators for IE outcomes [8, 9, 15]. In addition, the concept of uncontrolled infection including persistent infection was considered for maximizing the potential benefit of the early surgery even in RSIE [8, 9]. Since patients with LSIE are exposed to high systemic embolic risk under a high-pressure system, early surgery in LSIE was defined as surgery performed within 2 days, based on previous studies [11, 12]. LSIE included isolated left-sided IE and both-sided IE, considering the involvement of systemic circulation and worse outcomes of both-sided IE than that of RSIE [4, 17]. Early surgery in RSIE was defined as surgery performed within 7 days because less-resistant or less-virulent microorganisms could not persist within 7 days generally under appropriate antibiotic treatment, based on the guideline [9]. RSIE is considered more favourable to medical treatment, with a relatively benign prognosis, because it is involved in pulmonary circulation with a relatively low-pressure system [18].

In our study, S. aureus, active IE and PVE were not significant risk factors for time-related events, including death, reoperation, and IE relapse or reinfection. This finding can be explained by a more proactive approach with timely surgery according to the involved side than other pathogens. In addition, as aforementioned, we included 1 case of native valve infection that occurred directly adjacent to the prosthesis as PVE. Even when we put this case in others (non-NVE and non-PVE), it did not result in significant differences.

Mortality

Russell et al. [14] showed that infancy, prematurity, and fungal infection were risk factors of mortality. Khoo et al. reported several risk factors for mortality such as younger age, PVE, coagulase-negative staphylococci and increased duration of preoperative antibiotic therapy. Notably, heart failure and a high vegetation size to body surface area ratio were risk factors of mortality in patients with LSIE [4]. However, there was no significant risk factor for overall survival, because only 2 died in our study. In adults, several recent studies reported scoring systems predicting mortality with variables related to high-risk IE [19, 20]. In this study, operative mortality occurred in a patient at high risk, including active PVE and S. aureus infection and the cause of late death was non-cardiac.

Recurrent IE

Renzulli et al. [21] reported that PVE, positive valve culture and persistent fever were significant predictive factors for recurrent IE and that it could be prevented by complete surgical debridement. Pettersson et al. [22] reported incomplete debridement of the valve was the most important risk factor for relapse of IE; removing the source of bacteraemia is essential. In our study, fungal infection was a significant risk factor for recurrent IE. Considering the follow-up duration and isolated microorganisms in patients with recurrent IE, all cases of recurrent IE in our study were likely due to reinfection rather than relapse. Three of 4 patients underwent surgery for IE caused by fungal infections and had prostheses or central indwelling venous catheter for at least 6 months after surgery. These findings may be associated with statistical confounders or interactions and may influence the risk factor analysis.

Reoperation

Previous studies showed the results suggesting the effect of somatic growth, prior cardiac surgery and PVEI on reoperation [4, 14]. Younger age and PVEI were revealed as significant risk factors for reoperation in our study. More than half of reoperations (7/13, 53.8%) were performed due to the stenosis of prosthesis from previous cardiac surgery and required a larger prosthesis mainly due to somatic growth. Younger age as a significant risk factor for reoperation is expected to reflect these results. Furthermore, PVEI is the most common cause of uncontrolled infection that has been consistently regarded as a risk factor associated with poor prognosis [2, 4, 8, 9, 12].

Native valve preservation

Several studies indicated that native valve preservation was preferred, whenever possible [4, 5, 13]. We preferred the repair and reconstruction of the native valve to preserve and maintain valve function, and our results showed good outcomes in NVE (native valve-sparing 82.1% and freedom from overall valve replacement rate 73.7% at 10 years). LSIE was a risk factor for native valve-sparing failure. These findings reflect the relatively benign course of RSIE and the difference in preoperative manifestations between LSIE versus RSIE in our study (higher rate of valvular dysfunction, PVEI and vegetation in LSIE) (Supplementary Material, Table S1). A significant risk factor for overall valve replacement was PVEI. Despite aggressive debridement of the infected site and appropriate repair, reconstruction or replacement of the intra-cardiac structure, PVEI may affect the outcomes.

Postoperative CI

Preoperative CI was a risk factor for postoperative CI. However, recently reported studies suggested that early surgery can be safe, even in selected patients with active IE complicated with CI [23–25]. In our study, early surgery did not increase the risk of postoperative CI. Even in a patient who underwent early surgery for active IE with CI 1 day after diagnosis, preoperative stupor status improved to an alert status before discharge.

Early surgery for IE in children

Above all, considering that surgical treatment tends to be delayed in patients with stable IE, it should be taken into account that patients who underwent early surgery may have surgical treatment in a more critical and worse situation with an abrupt progression of the disease. Nonetheless, early surgery and preoperative negative blood culture conversion were not significant risk factors for surgical adverse outcomes of IE. These findings suggest that early surgery can be performed in children safely if patients have indications for early surgery. Therefore, just as surgery for IE in adults is not delayed in current practice, there is no need to hesitate and delay surgery for IE in children and the potential benefit of the early surgery might be expected and pursued.

Limitations

This study has several limitations. Basically, this study was a non-randomized, retrospective study with a small sample size of cases at a single centre. Therefore, given the fact that baseline characteristics in patients who underwent surgery for IE might be different between the groups, the results might be biased. Considering the preoperative status of patients and the course of the disease progression, baseline characteristics and perioperative data between the groups could not be perfectly controlled and could not be identical. Finally, although we referred to previously reported studies and guidelines, there has been no clear consensus on a couple of definitions that were used in our study, for example, ‘early surgery for IE’ or ‘uncontrolled infection’.

CONCLUSION

Surgical outcomes for IE in children were acceptable irrespective of surgical timing. Early surgery for IE could be performed safely and indicated without delaying if corresponding to the surgical indication in children. We expect the potential benefit of the early surgery, and believe that earlier surgery in paediatric patients diagnosed with IE can limit disease progression and increase the possibility of repair of valvular structures compared to delayed surgery, with other comparable clinical results. Of course, the decision for surgical timing should be based on an individual risk–benefit analysis.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Conflict of interest: none declared.

Presented at the 34th Annual Meeting of the European Association for Cardio-Thoracic Surgery, Barcelona, Spain, 8–10 October 2020.

Author contributions

Jae Hong Lee: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Visualization; Writing—original draft. Jae Gun Kwak: Conceptualization; Project administration; Resources; Supervision; Validation; Writing—review & editing. Sungkyu Cho: Resources; Supervision; Validation. Woong-Han Kim: Resources; Supervision; Validation. Jeong Ryul Lee: Resources. Hye Won Kwon: Resources. Mi Kyoung Song: Resources. Sang-Yun Lee: Resources. Gi Beom Kim: Resources. Eun Jung Bae: Resources.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Antonios Kallikourdis, Hani Najm and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

 
• CI

  Cerebral infarction

 
• IE

  Infective endocarditis

 
• IQR

  Interquartile range

 
• LSIE

  Left-sided infective endocarditis

 
• NVE

  Native valve endocarditis

 
• PVE

  Prosthetic valve endocarditis

 
• PVEI

  Perivalvular extension of infection

 
• RSIE

  Isolated right-sided infective endocarditis