Benefit and risk of early surgery

Early surgery is recommended for patients with complicated IE,3 ,4 ,9 but data supporting early surgery from randomised controlled trials are scarce due to ethical concerns about withholding surgery in control patients.10 Given the absence of randomisation in observational studies, baseline differences in treatment groups make it difficult to assess the impact of surgery on outcomes. Groups of patients treated medically and surgically in observational studies showed significant differences in baseline characteristics11–17 and plausible explanations of discrepant results among studies include difference in the baseline characteristics. As investigators choose the type of treatment in observational studies, selection bias also affects comparisons of outcomes.18 Selection bias can be adjusted by developing a propensity score (estimated possibility that a patient undergoes surgery) and matching on this score.11 ,15 ,18 Several studies used propensity score-matched analysis to evaluate the effect of surgery on survival in patients with IE (table 1). Although propensity analyses have yielded conflicting results due to differences in statistical methods and the duration of follow-up,19 a large multicentre study which adjusted for treatment selection, survivor and hidden biases, confirmed the beneficial effects of surgery.15 In addition, an analysis of 24 published observational studies showed a significant inverse correlation between the percentage of patients undergoing early surgery and in-hospital mortality rate.18

The main indications for early surgery in IE are HF, uncontrolled infection and prevention of embolism3 ,4 (figures 1 and 2). However, the risk–benefit ratio of early surgery may differ according to the type of complications. In analyses of 218 patients matched by propensity scores, valve surgery was associated with reduced mortality, and the greatest reduction in mortality with surgery was shown in patients with moderate to severe HF.11 In the recent analysis of 1359 IE patients complicated with HF, the in-hospital mortality rates in patients who underwent early surgery and those treated medically were 21% and 45%, respectively, and absolute mortality risk reduction of early surgery was greater for patients with moderate to severe HF.16 Moderate to severe HF is the most important predictor of mortality as well as the most frequent and clearest indication for early surgery.7 ,9 The appropriate management strategy for patients with mild HF and good response to initial medical therapy, however, remains controversial. Surgery can be performed at a later stage after healing of the infection with antibiotics, but despite appropriate antibiotic therapy, mild HF may progress insidiously to severe HF due to worsening of valvular regurgitation.9 ,20 In patients with well-tolerated valvular regurgitation and no other indication for surgery, initial medical management with antibiotics is recommended under careful clinical and echocardiographic observation.9

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Indications for early surgery: 2009 European Society of Cardiology (ESC) guidelines. 2009 ESC guidelines: 2009 European Society of Cardiology guidelines; Emergent surgery: surgery performed within 24 h; HF, heart failure; IE, infective endocarditis; Urgent surgery, surgery performed within a few days. Adapted from Habib et al.3

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Indications for early surgery: 2014 American Heart Association/American College of Cardiology (AHA/ACC) guidelines. Early surgery: surgery performed during initial hospitalisation before completion of a full therapeutic course of antibiotics; HF, heart failure; IE, infective endocarditis. Adapted from Nishimura et al.4

Perivalvular extension of IE is common, occurring in 10%–40% of native valve IE and in 56%–100% of prosthetic valve IE and is the most frequent cause of uncontrolled infection.3 ,9 Abscess of perivalvular structures cannot be cured with medical therapy alone and usually requires surgical drainage of abscess cavities.4 ,20 Surgical results of perivalvular complications are related to surgeon's ability to excise infected tissues completely,4 ,18 because total removal of infected tissues are essential for controlling infection and preventing recurrence after surgery. Early recognition and urgent performance of surgery is necessary for successful management of perivalvular complications including abscess, pseudoaneurysm and fistula,3 ,7 because valve replacement and surgical reconstruction are difficult when perivalvular supporting structures are extensively destroyed.

Embolic events can occur before the diagnosis of IE and during antibiotic therapy after its diagnosis. Early diagnosis of IE and prompt start of antibiotic therapy are essential for preventing embolic complications of IE because the incidence of embolic events markedly decreases after appropriate antimicrobial therapy.21 ,22 Embolic events occurring during antibiotic therapy may be prevented by surgical removal of vegetation, but the decision to perform surgery has been a clinical dilemma. Early identification of patients at high risk of embolism, increased experience in complete excision of infected tissue and valve repair, and low operative mortality have raised arguments for early surgery,9 but there have been concerns that such surgery may be more difficult to perform in the presence of active infection and inflammation, which leads to a high operative mortality and a high risk of postoperative valve dysfunction.8 Consensus guidelines for performance of early surgery on the basis of vegetation were different (figures 1 and 2). The 2006 ACC/AHA guidelines recommended early surgery as a class IIa indication only in patients with recurrent emboli and persistent vegetation despite appropriate antibiotic therapy,23 whereas the 2009 ESC guidelines recommend urgent surgery as a class I indication in patients with one or more embolic episodes and large vegetations (>10 mm in length) despite appropriate antibiotic therapy and urgent surgery as a class IIb indication in patients with isolated, very large vegetations (>15 mm).3 The recently revised 2014 AHA/ACC guidelines have added a class IIb indication for early surgery in patients with mobile, large vegetations (>10 mm),4 but are still more conservative than the ESC guidelines because early surgery is recommended only after recurrent embolic episodes occur during antibiotic therapy. Operative risk and severity of valvular lesion should be also considered for clinical decision of surgery. Early surgery is indicated in patients with large vegetations only when embolic risk exceeds operative risk, and the benefit of surgery would be greater if conservative procedure preserving the native valve is likely or severe valvular regurgitation is associated.

Neurological events occur in 20%–40% of patients with IE, mainly as a consequence of vegetation embolism.6 ,22 The clinical spectrum of these complications include ischaemic stroke with or without haemorrhagic transformation, transient ischaemic attack, silent cerebral embolism, mycotic aneurysm, brain abscess, meningitis and encephalopathy.3 Early diagnosis and rapid initiation of antibiotic therapy is effective in preventing neurological complications, with 86% of neurological complications observed before or during the first week of antibiotic therapy and the incidence of neurological complications markedly decreasing after appropriate antimicrobial therapy.22 The International Collaboration on Endocarditis Prospective Cohort Study (ICE-PCS) also showed that the incidence of stroke was 4.82 per 1000 patient-days in the first week of antibiotic therapy and fell to 1.71 per 1000 patient-days in the second week, with further decreases thereafter.21 However, vegetations >1 cm were associated with a less pronounced decrease in risk after antibiotic therapy, and in those with vegetations >3 cm, the incidence of embolic events after the first week of antibiotic therapy remained very high (20%), supporting an embolic indication of surgery for very large vegetations.22

Patients with a neurological complication may have other complications requiring surgery, but early surgery may pose significant risks for neurological deterioration and perioperative cerebral bleeding.9 ,24 Anticoagulation during cardiopulmonary bypass may extend haemorrhagic infarction and increase the risk of cerebral haemorrhage.25 In addition, hypotension during cardiopulmonary bypass may aggravate ischaemic neurological deficit and potentiate cerebral oedema.26 If possible, surgery should be delayed to prevent neurological deterioration in patients with a large cerebral infarction who are at risk of haemorrhagic transformation or intracerebral haemorrhage.24–26 Systematic early MRI of the brain may reveal cerebral abnormalities in up to 80% of patients and help clinical decision on timing of cardiac surgery.27

Prosthetic valve endocarditis (PVE) is the most severe form of IE and more difficult to treat by antibiotic therapy alone.3 ,28 Even with the use of an aggressive surgical strategy, PVE is clearly associated with higher mortality rates.3 ,4 ,28 ,29 In a series of 1025 patients with PVE, mortality was associated with healthcare-associated infection, Staphylococcus aureus, and complications of PVE.30 After adjustment for selection and survivor bias, early valve surgery was not associated with lower mortality compared with medical therapy in the overall cohort, but patients with the strongest indications for surgery had a significantly lower 1-year mortality rate with early surgery.30 These results support the recommendation of current guidelines that a surgical strategy is considered in high-risk subgroups of PVE complicated with HF, abscess or persistent fever.3 ,4 Surgery for PVE follow the general principle outlined for native valve IE and most of the indications and timing of surgery are the same for both types of IE.3 ,4

The rate of early surgery has increased during the last decade,1 with a recent study reporting that surgery was performed in 57% of 1296 patients with left-sided IE and in 76% of patients with a surgical indication.2 Appropriate early surgery can significantly improve in-hospital mortality and long-term prognosis in IE,15 ,19 but nearly one-quarter of patients who have a surgical indication do not undergo surgical intervention.2 Implementation of a standardised strategy based on close collaboration between cardiologists, infectious disease specialists and cardiac surgeons may improve compliance with guidelines for surgical indication and timing.18 In summary, multidisciplinary collaborations are required for appropriate decisions about indication and timing of surgical intervention,4 ,9 and these decisions should be based on individual risk–benefit analysis (table 2). The potential benefits of early surgery need to be weighed against its operative risks and long-term consequences.

Timing of surgery

The ESC guidelines3 have established optimal timing for each surgical indication and classified early surgery into emergency (within 24 h), urgent (within a few days) and elective (after at least 1–2 weeks of antibiotic therapy) (figure 1). However, most previous observational studies defined early surgery as surgery performed during initial hospitalisation for IE,11–17 and there is no consensus on the optimal timing of early surgery due to a lack of evidence. Thus, the 2014 AHA/ACC guidelines vaguely defined early surgery as occurring during initial hospitalisation before completion of a full therapeutic course of antibiotics (figure 2), and recommended that decisions about timing of surgery be made by a Heart Valve Team of cardiologists, cardiac surgeons and infectious disease specialists.4

Prevention of embolism

As the risk of embolism is highest during the first few days after initiation of antibiotic therapy, the ESC guidelines clearly recommend that surgery to prevent embolism be performed very early on an urgent (within a few days) basis.3 By contrast, the 2014 AHA/ACC guidelines have not established the optimal timing of early surgery for embolic indication.4 Since the benefits of surgery to prevent embolism are greatest during the first week of the diagnosis, deferring surgery after 1–2 weeks is of little value.9 ,21 A recent randomised trial compared clinical outcomes of early surgery and a conventional treatment strategy based on current guidelines in left-sided IE patients with high embolic risks.10 Patients were eligible for enrolment if they were diagnosed as definite IE and had both severe mitral or aortic valve disease and maximal length of vegetation >10 mm, and were randomly assigned to early surgery (37 patients) or to conventional treatment (39 patients). All patients in the early surgery group underwent valve surgery within 48 h after randomisation. Of the 39 patients in the conventional treatment group, 30 (77%) patients underwent surgery during initial hospitalisation (n=27) or during follow-up (n=3). The primary end point of in-hospital death and embolic events at 6 weeks occurred in 1 (3%) patient in the early surgery group as compared with 9 (23%) in the conventional treatment group (HR, 0.10; 95% CI 0.01 to 0.82; p=0.03). There was no significant difference in all-cause mortality rate at 6 months (3% vs 5%; HR, 0.51; 95% CI 0.05 to 5.66; p=0.59) (figure 3A). The rate of the composite of death from any cause, embolic events, recurrence of IE, or repeat hospitalisations due to development of congestive HF at 6 months was 3% in the early surgery group as compared with 28% in the conventional management group (HR, 0.08; 95% CI 0.01 to 0.65; p=0.02) (figure 3B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Kaplan–Meier curve for the cumulative probabilities of death (A) and of the composite end point (B), according to treatment group. There was no significant between-group difference in all-cause mortality at 6 months (A). The rate of the composite of death from any cause, embolic events, recurrence of infective endocarditis or repeat hospitalisations due to development of congestive heart failure at 6 months was 3% in the early surgery group versus 28% in the conventional management group (HR, 0.08; 95% CI 0.01 to 0.65; p=0.02) (B). Adapted from Kang et al.10

This randomised trial demonstrated that early surgery performed within 48 h after diagnosis reduced the primary end point (composite of in-hospital death and embolic events) by effectively decreasing systemic embolisms in patients with IE. Moreover, these improvements in clinical outcomes could be achieved without increases in operative mortality or recurrence of IE. Because IE is a highly variable disease and the risk–benefit ratio of early surgery over conventional treatment may differ according to the type of high-risk situation, additional randomised trials will be necessary to evaluate the efficacy and safety of early surgery in patients with complicated IE.

Conclusion

Early surgery is performed in approximately half of patients for the treatment of IE, but the decision about timing of surgery has been difficult. Urgent surgery (within a few days) is indicated for patients who have moderate to severe HF, uncontrolled infection and large vegetations associated with severe valvular disease, but surgery should be delayed for 2–4 weeks in patients with large cerebral infarction and at least 4 weeks in those with intracerebral haemorrhage if possible. The timing of surgery should be based on individual risk–benefit analysis, and early surgery is strongly indicated if its benefits exceed operative risks.