Our search strings identified a total of 1,503 citations. Of these, nine studies were included. Among the selected studies were seven retrospective studies, one prospective study, and one RCT. Table 1 summarises the 2015 and 2023 ESC guidelines’ recommendations for surgery to prevent emboli.

The total number of patients from the nine studies was 3,565 and the median number of patients was 355 (range: 71–1,006). The median age was 66 years (range: 17–80) and the median percentage of male patients was 65.6% (range: 61.8 − 71.4%). The causative microorganisms were identified in 2,416 patients (67.8%); the most frequently found were: Viridans group streptococci (33.2%), Staphylococcus aureus (28.5%), Enterococci (16.8%).

Three studies were multinational, three were national ones collecting data from multiple tertiary centres, and the remaining three were single centre studies. All nine studies focused on left-sided IE only. A single study examined IE in patients with prosthetic valves, the remaining eight analysed patients with native valve IE.

All nine studies divided their populations into patients treated with surgery + antibiotic therapy and patients managed with antibiotic therapy alone. The eight cohort studies retrieved patient data from either international databases, or from electronic medical records. To even out differences in baseline characteristics and compare clinical outcomes, propensity score matching was performed in five studies [15,16,17,18,19]. Five studies subdivided patients based on vegetation size (> 10 mm vs. <10 mm) [15, 16, 18, 20, 21]. Three studies identified patients who underwent surgery with embolic risk reduction as the only surgical indication [19, 22]. The RCT randomised patients with vegetations > 10 mm to surgery + antibiotic therapy or antibiotic therapy alone. The one study investigating prosthetic valve IE [17] compared surgery + antibiotic therapy to antibiotic therapy alone, regardless of vegetation size.

Five studies [16, 18, 21,22,23] evaluated long-term (+ 5 years) survival outcomes between the surgery + antibiotic therapy groups and the antibiotic therapy alone groups, four also included the occurrence of embolism after being admitted to a tertiary centre as well as recurrence of IE. Two studies [15, 20] analysed 6-month mortality, one study reported 90-day mortality [19], and one on in-hospital mortality [15]. Table 2 summarises the results of the included articles.

Three studies [2, 16, 21] compared the rate of embolism after admission between the surgery + antibiotic therapy groups and the antibiotic therapy alone groups. In the Kang et al. study, eight patients (21%) in the antibiotic therapy alone group suffered systemic emboli within six weeks compared to zero in the surgery + antibiotic therapy group. Further, eight patients (21%) form the antibiotic therapy alone group suffered a systemic embolism within six months. Again, no patients from the surgery + antibiotic therapy group suffered a systemic embolism within six months. Kim et al. conducted a similar evaluation where 14 patients in the antibiotic therapy alone group suffered a systemic embolism during initial hospitalization, compared to zero in the surgery + antibiotic therapy group. During follow-up, two patients from each group had suffered a systemic embolism. Song et al. evaluated the rate of systemic embolism after initial hospitalization. 49 (33.6%) patients from the antibiotic therapy alone group suffered a systemic embolism, compared to 77 (28.2%) in the surgery + antibiotic therapy group. None of the studies statistically tested if surgery was associated with a significant risk reduction of systemic emboli.

Two studies [2, 16] evaluated the recurrence of IE. In both studies, one patient from the antibiotic therapy alone group was readmitted with IE, and none from the surgery + antibiotic therapy group.

Of the nine studies included in this review, seven indicate that surgery is beneficial to short- and long-term mortality. This includes the RCT, where in-hospital mortality or embolism was significantly lower in the surgery + antibiotic therapy group, compared to the antibiotic therapy alone group [hazard ratio (HR) = 0.10; 95% confidence interval (CI) 0.01–0.82.]. Further, the long-term survival rates at 7-year follow-up were also better in the surgery + antibiotic therapy group (87% vs. 83%) as was the event-free survival rate (87% vs. 59%) [23]. The largest study in this review [15] (1,006 patients) found larger vegetation sizes to be associated with higher mortality rates, both in-hospital [Odds ratio (OR) 1.45 (95% CI 1.04–2.03)] and 6-months after admission [OR 1.37 (CI 95% 1.01–1.87)]. However, in the surgery + antibiotic therapy group, vegetation size was not associated with a higher mortality rate at either endpoint [OR 1.01 (95% CI 0.69–1.49)]. This study also propensity matched patients using Cox regression to eliminate confounders. When comparing patients with vegetations > 10 mm, the surgically treated patients had significantly higher survival probability compared to the patients treated with antibiotic therapy alone. Another study [20] examining vegetation sizes provided similar results. Vegetation area > 50 mm2 was only associated with mortality in the antibiotic therapy alone group. A fourth study [18] propensity matched 147 patients (49 receiving surgery + antibiotic therapy vs. 98 receiving antibiotic therapy alone) who had prevention of embolism as the only surgical indication. The surgery + antibiotic therapy group had lower mortality rates [HR 0.22 (95% CI 0.08–0.66)] and better estimated 5-year Kaplan-Meier survival rates (95% CI): 82.4% (71.9 − 94.4%) for surgically treated patients vs. 66.5% (57–77.6%) in patients treated with antibiotic therapy alone. Three retrospective database studies [16, 17, 21] also found that surgery + antibiotic therapy, in patients with vegetations > 10 mm, improved either long-term survival rates, in-hospital mortality rates, or both.

The remaining two studies found that surgery + antibiotic therapy, in patients with embolic risk reduction as the only surgical indication, is not beneficial when compared to antibiotic therapy alone (210 patients). The single-centre study identified 71 patients with no surgical indications other than prevention of embolism where 59 patients underwent surgery. Using Cox regression, the study found that surgery was independently associated with increased mortality rates [HR 3.9 (95% CI 0.9–16.6)]. The second study was able to identify 70 patients who underwent surgery with no other indication, and 69 who only received antibiotic therapy alone. The surgery + antibiotic therapy group had a 90-day survival probability of 81.4% vs. 88.4% in the antibiotic therapy alone group.