RiskE Score for Cardiac Surgery in Active Infective Endocarditis

Calculate the RISK-E Score to predict perioperative mortality for cardiac surgery in active left-sided infective endocarditis. Based on age, prosthetic valve, periannular complications, S. aureus or fungal infection, renal insufficiency, septic shock, thrombocytopenia, and cardiogenic shock.

RISK-E Score Calculator

Select the clinical variables present at the time of planned cardiac surgery for active left-sided infective endocarditis. The RISK-E score predicts perioperative mortality.

Age

Clinical variables

Disclaimer: The RISK-E Score is an educational clinical decision-support tool intended for use by healthcare professionals. It does not replace comprehensive cardiac surgical assessment, multidisciplinary Endocarditis Team deliberation, or clinical judgment. Surgical decision-making for infective endocarditis should incorporate all relevant clinical, microbiological, imaging, and patient-specific factors. Always consult institutional protocols and specialist expertise.

RISK-E Score: formula, variables, and clinical context

Overview

The RISK-E (Risk-Endocarditis) score is a validated prognostic tool that predicts perioperative mortality in patients undergoing cardiac surgery for active left-sided infective endocarditis (IE). It was developed by Olmos and colleagues from a prospective, multi-centre cohort of 671 surgically treated IE patients at three Spanish referral hospitals between 1996 and 2014. The score incorporates eight clinical variables that reflect the severity of the endocarditis episode and the patient's baseline status, producing a continuous predicted mortality via logistic regression.

Scoring variables

Variable	Criteria	Points
Age	≤ 51 years	0
	52-63 years	+9
	64-72 years	+13
	≥ 73 years	+14
Prosthetic valve endocarditis	Yes	+6
Prosthetic valve endocarditis	No	0
Periannular complications	Yes	+5
Periannular complications	No	0
S. aureus or fungal infection	Yes	+9
S. aureus or fungal infection	No	0
Acute renal insufficiency	Yes	+5
Acute renal insufficiency	No	0
Septic shock	Yes	+7
Septic shock	No	0
Thrombocytopenia	Yes	+7
Thrombocytopenia	No	0
Cardiogenic shock	Yes	+15
Cardiogenic shock	No	0
Total score range		0-68

Predicted operative mortality

The total RISK-E score is converted to a predicted operative mortality via a logistic regression model. The score of 0 corresponds to approximately 3% predicted mortality and the maximum score of 68 corresponds to approximately 97%.

Score range	Approximate mortality	Interpretation
0-9	3-7%	Lower operative risk
10-21	7-22%	Moderate operative risk
22-40	22-60%	High operative risk
41-68	60-97%	Very high operative risk

Internal validation AUC = 0.82 (95% CI 0.75-0.88). External validation (French cohort) AUC = 0.76 (95% CI 0.64-0.88).

Variable definitions

Age (0-14 points)

Age at the time of surgery. Categorised into four groups reflecting the independent, non-linear contribution of age to operative risk. Advancing age correlates with greater comorbidity burden, reduced physiological reserve, and increased vulnerability to the haemodynamic and infectious stresses of IE. The largest increment (9 points) occurs between the first and second age bands.

Prosthetic valve endocarditis (6 points)

Infection involving a prosthetic heart valve (mechanical or bioprosthetic). Prosthetic valve IE is associated with higher mortality due to the increased technical complexity of re-operation, greater risk of persistent or paravalvular infection, and more frequent need for extensive debridement and annular reconstruction.

Periannular complications (5 points)

Periannular extension of the infection, including abscess formation, pseudoaneurysm, or fistula. These complications typically indicate locally aggressive disease that has penetrated beyond the valve leaflets into surrounding structures, often requiring more extensive surgical repair (e.g., patch reconstruction, root replacement). Identified on echocardiography (TOE preferred) or cardiac CT.

Staphylococcus aureus or fungal infection (9 points)

S. aureus (both MSSA and MRSA) and fungal pathogens (e.g., Candida, Aspergillus) represent the most virulent causative organisms in IE. They are associated with more rapid valve destruction, higher embolic rates, more frequent periannular complications, and greater resistance to medical therapy alone. S. aureus endocarditis carries in-hospital mortality rates of 25-47% in published series.

Acute renal insufficiency (5 points)

New or acutely worsening renal dysfunction attributable to the endocarditis episode. Renal injury in IE can be multifactorial: embolic renal infarction, immune complex-mediated glomerulonephritis, sepsis-related acute tubular necrosis, or nephrotoxic antimicrobial therapy. Preoperative renal impairment is an established predictor of surgical mortality across cardiac surgery populations.

Septic shock (7 points)

Defined as persistent hypotension requiring vasopressor support to maintain MAP ≥ 65 mmHg and a serum lactate > 2 mmol/L despite adequate volume resuscitation (per Sepsis-3 criteria). Septic shock reflects uncontrolled systemic infection and end-organ hypoperfusion, substantially increasing the perioperative risk of multi-organ dysfunction.

Thrombocytopenia (7 points)

Platelet count below 150,000/μL. In the setting of IE, thrombocytopenia may indicate consumptive coagulopathy, disseminated intravascular coagulation (DIC), splenic sequestration, or bone marrow suppression. Low platelet counts increase the risk of perioperative bleeding and correlate with the severity of the systemic inflammatory response.

Cardiogenic shock (15 points)

Severe cardiac pump failure with tissue hypoperfusion (e.g., systolic BP < 90 mmHg or requirement for inotropes/mechanical support, reduced cardiac output, evidence of end-organ dysfunction). In IE, cardiogenic shock typically results from acute severe valvular regurgitation due to leaflet destruction or dehiscence. It carries the highest point value in the RISK-E score, reflecting its profound impact on operative survival.

Study details

Design: Prospective, multi-centre observational cohort study at three Spanish referral hospitals (Hospital Clinico San Carlos Madrid, Hospital General Universitario Gregorio Maranon Madrid, Hospital Clinico Universitario Valladolid).
Population: 671 consecutive patients with active left-sided IE who underwent cardiac surgery between 1996 and 2014. The cohort was split into a derivation set (n = 434, 1996-2008) and a temporal validation set (n = 237, 2009-2014).
Outcome: Perioperative (in-hospital or 30-day) all-cause mortality. Overall surgical mortality in the derivation cohort was 28.3%.
Statistical method: Multivariable logistic regression with backward stepwise selection. Point assignments derived from beta-coefficients rounded proportionally. The logistic model converts total points to continuous predicted mortality.
Internal validation: AUC = 0.82 (95% CI 0.75-0.88) in the temporal validation set. Hosmer-Lemeshow calibration P = 0.45, indicating good agreement between predicted and observed mortality.
External validation: In an independent French tertiary-care cohort, AUC = 0.76 (95% CI 0.64-0.88).
Comparison with other scores: RISK-E predicted mortality more accurately than EuroSCORE I, EuroSCORE II, PALSUSE, and STS-IE in the validation cohort (P = 0.010 for comparison).

Limitations

Left-sided IE only: The RISK-E score was derived exclusively from patients with left-sided (aortic and/or mitral valve) endocarditis. It has not been validated for isolated right-sided or device-related IE.
Geographic and era generalisability: The derivation cohort is from three Spanish centres over an 18-year period. Microbiological patterns, surgical techniques, and perioperative care have evolved over time and differ across regions.
Selection bias: Only patients who actually underwent surgery were included. Patients deemed too high risk for surgery (or who died before surgery could be performed) were excluded, which may affect calibration at the extremes.
External validation performance: While the internal validation AUC was 0.82, external validation studies report lower discrimination (AUC 0.74-0.76), and some systematic reviews have noted performance degradation in cohorts with different risk profiles.
Missing variables: The score does not account for embolic events (e.g., cerebral embolism), LVEF, pulmonary hypertension, specific surgical complexity, or concurrent coronary disease, all of which may influence operative mortality.

Key references

Olmos C, Vilacosta I, Fernandez-Perez C, et al. Risk score for cardiac surgery in active left-sided infective endocarditis. Heart. 2017;103(18):1435-1442.
Olmos C, Vilacosta I, Habib G, et al. Risk score for cardiac surgery in active left-sided infective endocarditis (response). Heart. 2017;103(23):1923.
Varela Barca L, Valerio M, Munoz P, et al. Risk scores' performance and their impact on operative decision-making in left-sided endocarditis: a cohort study. Eur J Clin Microbiol Infect Dis. 2023;42(2):179-190.
Farag M, Borst T, Sabashnikov A, et al. Enhanced risk stratification in infective endocarditis surgery: a comprehensive external validation of all available risk scores. Clin Epidemiol. 2024;16:411-423.
Habib G, Lancellotti P, Antunes MJ, et al. 2015 ESC Guidelines for the management of infective endocarditis. Eur Heart J. 2015;36(44):3075-3128.

Practice caveats

The RISK-E score is specific to active left-sided IE undergoing cardiac surgery. Do not apply it to right-sided IE, healed IE, or patients managed conservatively without planned surgery.
No risk score replaces the Endocarditis Team approach recommended by ESC/AHA guidelines. Multidisciplinary assessment (cardiology, cardiac surgery, infectious disease, microbiology, imaging) is essential for optimal surgical timing and technique.
A high predicted mortality does not necessarily mean surgery should be avoided; non-operative management of complicated IE can carry even higher mortality. Risk-benefit must be weighed individually.
Calibration may vary across institutions and patient populations. Consider the RISK-E prediction as one input among many in shared decision-making, not as a definitive prognostic number.

Background: The Challenge of Infective Endocarditis

Infective endocarditis (IE) remains one of the most formidable infections in cardiovascular medicine. Despite advances in diagnostic imaging, antimicrobial therapy, and surgical technique, in-hospital mortality for IE ranges from 15% to 30% across contemporary registries. Left-sided IE, affecting the aortic and mitral valves, accounts for the majority of cases in developed countries and carries a particularly high burden of morbidity from valvular destruction, embolic complications, and systemic sepsis.

Approximately 50% of patients with left-sided IE ultimately require cardiac surgery during the index hospitalisation. Surgical indications include haemodynamically significant valvular regurgitation, uncontrolled infection despite appropriate antimicrobial therapy, large vegetations with high embolic risk, and periannular extension such as abscess or fistula formation. The decision to proceed with surgery in the setting of active infection, however, is fraught with complexity. The very factors that create the surgical indication (severe sepsis, haemodynamic instability, extensive tissue destruction) also elevate operative risk substantially.

The Need for IE-Specific Surgical Risk Stratification

General cardiac surgical risk models, such as the EuroSCORE (both versions I and II) and the Society of Thoracic Surgeons (STS) score, were developed and validated predominantly in elective cardiac surgery populations. Their performance in the unique pathophysiology of active endocarditis is limited. Active IE introduces variables that these generic models do not adequately capture: the virulence of the causative micro-organism, the extent of periannular destruction, the systemic inflammatory response to ongoing bacteraemia, and the consumptive coagulopathy that frequently accompanies severe infection.

Several attempts have been made to develop IE-specific risk scores. The PALSUSE score, the STS-IE score, and the De Feo score represent notable earlier efforts. However, each has been constrained by relatively small sample sizes, single-centre derivation, limited validation, or dependence on intraoperative findings that are unavailable at the point of surgical decision-making. There was a clear need for a robust, prospectively derived, multi-centre risk model that could be applied at the bedside before the patient enters the operating theatre.

Development and Derivation of the RISK-E Score

The RISK-E (Risk in Endocarditis) score was developed by Olmos, Vilacosta, and colleagues and published in Heart in 2017. The derivation cohort comprised 671 consecutive patients with definite active left-sided IE (by modified Duke criteria) who underwent cardiac surgery at three major Spanish referral hospitals: Hospital Clinico San Carlos (Madrid), Hospital General Universitario Gregorio Maranon (Madrid), and Hospital Clinico Universitario (Valladolid). Data were collected prospectively between 1996 and 2014.

The cohort was divided into a derivation set of 434 patients (1996 to 2008) and a temporal validation set of 237 patients (2009 to 2014). The primary endpoint was perioperative mortality, defined as in-hospital death or death within 30 days of surgery regardless of location. In the derivation cohort, overall surgical mortality was 28.3%, underscoring the high-risk nature of this population.

Multivariable logistic regression with backward stepwise selection was used to identify independent predictors of perioperative death. More than 30 candidate variables were evaluated, spanning patient demographics, comorbidities, micro-organism type, echocardiographic findings, haemodynamic status, laboratory parameters, and operative factors. Eight variables emerged as independent predictors and were retained in the final model. Point assignments were derived from the beta-coefficients of the logistic model, rounded proportionally to create a practical integer-based scoring system.

The Eight Variables of the RISK-E Score

Age (0 to 14 points)

Age at the time of surgery is categorised into four bands: 51 years or younger (0 points), 52 to 63 years (9 points), 64 to 72 years (13 points), and 73 years or older (14 points). The relationship between age and operative mortality in IE is non-linear. The largest single increment (9 points) occurs between the youngest group and the 52 to 63 age band, reflecting the steep increase in perioperative risk once patients move beyond the fifth decade. Advancing age correlates with greater comorbidity burden, diminished cardiac and pulmonary reserve, impaired immune function, and reduced ability to tolerate the combined insults of infection and major surgery. In the derivation cohort, the median age was 62 years, and patients older than 70 had nearly double the mortality rate of those under 50.

Prosthetic Valve Endocarditis (6 points)

Prosthetic valve endocarditis (PVE) involves infection of a mechanical or bioprosthetic heart valve and represents roughly 20% to 30% of all IE cases in contemporary series. PVE is associated with substantially higher mortality than native valve endocarditis for several reasons. The biofilm that forms on prosthetic material is inherently more resistant to antimicrobial penetration. Surgical re-operation on a prosthetic valve is technically more demanding, often requiring extensive debridement of infected and calcified tissue around the sewing ring, annular reconstruction, and sometimes aortic root replacement. The risk of recurrent infection after surgery is also higher. In the RISK-E derivation cohort, patients with PVE had a perioperative mortality approximately 10 percentage points higher than those with native valve IE.

Periannular Complications (5 points)

Periannular complications encompass abscess formation, pseudoaneurysm, and fistula around the valve annulus. These findings indicate locally invasive disease that has extended beyond the valve leaflets into the surrounding fibrous skeleton and adjacent structures. Periannular extension is identified by transoesophageal echocardiography (TOE) or cardiac CT and is present in approximately 30% to 40% of surgical IE cases. Its presence signals more extensive tissue destruction, a greater likelihood of requiring complex reconstructive surgery (root replacement, patch repair, double-valve surgery), and a higher risk of persistent or recurrent infection. Aortic valve IE is particularly prone to periannular complications because the aortic annulus is in close anatomical proximity to the interventricular septum, the mitral-aortic intervalvular fibrosa, and the conduction system.

Staphylococcus aureus or Fungal Infection (9 points)

The causative micro-organism is one of the most powerful determinants of outcome in IE. Staphylococcus aureus, both methicillin-sensitive (MSSA) and methicillin-resistant (MRSA), is the most common cause of acute IE in many regions and is associated with an aggressive clinical course. S. aureus IE produces large, friable vegetations with a high propensity for embolisation, rapid valve destruction, and periannular invasion. In-hospital mortality for S. aureus IE ranges from 25% to 47% in published series. Fungal endocarditis, caused by organisms such as Candida and Aspergillus species, is uncommon but carries an even worse prognosis. Fungal vegetations tend to be very large, antimicrobial options are limited, and relapse rates are high. The combined weight of 9 points for this variable in the RISK-E score (equal to the jump from the youngest age band to the next) underscores the prognostic importance of micro-organism virulence.

Acute Renal Insufficiency (5 points)

New or acutely worsening renal dysfunction during the IE episode is a well-established marker of poor prognosis. The mechanisms of renal injury in IE are multifactorial: septic embolisation to the renal vasculature causing infarction, immune complex-mediated glomerulonephritis, sepsis-related haemodynamic instability causing acute tubular necrosis, and nephrotoxicity from antimicrobial agents (aminoglycosides, vancomycin, amphotericin B). Preoperative renal impairment increases the risk of postoperative dialysis, prolongs intensive care stays, and is independently associated with mortality across all cardiac surgery populations. In the RISK-E model, acute renal insufficiency contributed 5 points, reflecting its moderate but consistent prognostic contribution.

Septic Shock (7 points)

Septic shock is defined, per Sepsis-3 criteria, as a subset of sepsis in which particularly profound circulatory, cellular, and metabolic abnormalities substantially increase mortality. The operational definition requires persistent hypotension necessitating vasopressor therapy to maintain a mean arterial pressure of at least 65 mmHg and a serum lactate level exceeding 2 mmol/L despite adequate volume resuscitation. In the context of IE, septic shock reflects uncontrolled systemic infection that has overwhelmed host defences. It is associated with profound vasodilatation, myocardial depression (both septic cardiomyopathy and direct valvular dysfunction), and multi-organ hypoperfusion. Patients who undergo cardiac surgery in the setting of septic shock face the dual challenge of a major surgical insult superimposed on a severely dysregulated inflammatory state. The 7-point assignment in the RISK-E score reflects the substantial mortality increment observed in the derivation cohort.

Thrombocytopenia (7 points)

Thrombocytopenia, defined as a platelet count below 150,000 per microlitre, is a marker of systemic severity in IE that is often underappreciated. Its mechanisms in endocarditis include disseminated intravascular coagulation (DIC), consumptive coagulopathy from ongoing thrombin generation on infected vegetations, splenic sequestration (particularly in patients with splenomegaly from septic emboli), bone marrow suppression from prolonged infection, and drug-induced effects. Thrombocytopenia increases the risk of perioperative haemorrhage, particularly problematic in cardiac surgery where systemic heparinisation for cardiopulmonary bypass is required. Moreover, low platelet counts correlate with the severity of the systemic inflammatory response and the presence of DIC, both of which independently worsen surgical outcomes. Its 7-point contribution to the RISK-E score equals that of septic shock, reflecting comparable prognostic gravity.

Cardiogenic Shock (15 points)

Cardiogenic shock is the single most heavily weighted variable in the RISK-E score and is defined as severe cardiac pump failure resulting in tissue hypoperfusion. Clinical features include systolic blood pressure below 90 mmHg (or requirement for inotropes or mechanical circulatory support), reduced cardiac output, elevated filling pressures, and end-organ dysfunction attributable to low cardiac output. In IE, cardiogenic shock most commonly results from acute, severe valvular regurgitation caused by leaflet perforation, chordal rupture, or prosthetic valve dehiscence. The haemodynamic consequences can be catastrophic: acute pulmonary oedema from mitral regurgitation, or cardiovascular collapse from acute aortic regurgitation with a non-compliant left ventricle. Patients in cardiogenic shock often require urgent or emergent surgery, leaving little time for haemodynamic optimisation, source control of infection, or preoperative stabilisation. The 15-point assignment (nearly a quarter of the maximum possible 68 points) reflects the profound impact of cardiogenic shock on operative survival.

Scoring and Predicted Mortality

The total RISK-E score is the sum of points from all eight variables, yielding a range of 0 to 68. This total is then converted to a predicted perioperative mortality using the logistic regression model from which the score was derived. A score of 0 corresponds to approximately 3% predicted mortality, while the maximum score of 68 corresponds to approximately 97% predicted mortality. The relationship between total score and predicted mortality follows the characteristic S-shaped (sigmoid) curve of logistic regression, with the steepest gradient in the middle ranges of the score.

For practical clinical use, the predicted mortality can be stratified into four tiers:

Score Range	Approximate Mortality	Risk Category
0 to 9	3% to 7%	Low risk
10 to 21	7% to 22%	Moderate risk
22 to 40	22% to 60%	High risk
41 to 68	60% to 97%	Very high risk

It is important to recognise that the RISK-E score provides a continuous predicted probability rather than simply assigning patients to fixed categories. The tiered classification above is a simplification intended to facilitate clinical communication and decision-making.

Validation and Discriminatory Performance

The RISK-E score was validated using a temporal split-sample approach. In the internal validation cohort (237 patients, 2009 to 2014), the model demonstrated an area under the receiver operating characteristic curve (AUC) of 0.82 (95% CI 0.75 to 0.88), indicating good discriminatory ability. Hosmer-Lemeshow goodness-of-fit testing yielded a P-value of 0.45, suggesting no significant departure between predicted and observed mortality across deciles of predicted risk. This is a particularly important metric: a score can discriminate well (correctly ranking patients by risk) but still be poorly calibrated (systematically over- or under-predicting mortality). The RISK-E score demonstrated both good discrimination and good calibration in internal validation.

External validation has been performed in several independent cohorts. A French tertiary-care study reported an AUC of 0.76 (95% CI 0.64 to 0.88), somewhat lower than the internal validation but still reflecting reasonable performance. Other external validation studies from Germany, Italy, and multi-national registries have reported AUC values ranging from 0.71 to 0.80, with some variability attributable to differences in case mix, microbiological epidemiology, surgical timing, and perioperative protocols.

Comparison with Other Surgical Risk Scores in IE

A key strength of the RISK-E score is its demonstrated superiority over general cardiac surgery risk models when applied to the IE population. In the original validation study, RISK-E outperformed EuroSCORE I, EuroSCORE II, the PALSUSE score, and the STS-IE score (P = 0.010 for the overall comparison). The reasons for this superiority are intuitive: the RISK-E score was specifically designed for active IE and incorporates variables (micro-organism virulence, thrombocytopenia, periannular complications) that generic models ignore or inadequately weight.

The EuroSCORE models, while widely used in cardiac surgery, were calibrated for elective and semi-elective populations and tend to underestimate mortality in active IE. The PALSUSE score (Prosthetic valve, Age over 65, Large intracardiac destruction, Staphylococcus aureus, Urgent surgery, Sex (female), EuroSCORE over 10) provides a dichotomous risk estimate and lacks continuous mortality prediction. The STS-IE score incorporates some IE-specific variables but was derived from the STS registry, which has different inclusion criteria and data definitions than the prospective cohorts used for RISK-E.

A systematic review and meta-analysis of all available IE surgical risk scores, published in 2024, confirmed that RISK-E provides the best balance of discrimination and calibration among currently available models, while acknowledging that no single score achieves the performance levels seen in general cardiac surgery risk prediction.

Limitations and Considerations for Clinical Use

Applicability to Left-Sided IE Only

The RISK-E score was derived exclusively from patients with left-sided IE (aortic, mitral, or both). It has not been validated for isolated right-sided endocarditis (e.g., tricuspid valve IE in intravenous drug users) or cardiac device-related infections (pacemaker or defibrillator lead endocarditis). The pathophysiology, surgical approach, and mortality profile of right-sided IE differ substantially from left-sided disease, and RISK-E should not be extrapolated to these populations without further study.

Selection Bias in the Derivation Cohort

Only patients who actually underwent cardiac surgery were included in the RISK-E derivation. Patients who were deemed too high-risk for surgery, who died before surgery could be performed, or who were managed conservatively were excluded. This introduces a form of selection bias that may affect the calibration of the score at the extremes, particularly at the very high end where the sickest patients may have been excluded from the surgical cohort entirely.

Geographic and Temporal Generalisability

The derivation cohort is from three Spanish centres spanning an 18-year period (1996 to 2014). Microbiological epidemiology, healthcare system factors, surgical techniques, and perioperative management have evolved over this period and vary across geographic regions. For example, the prevalence of MRSA endocarditis, the use of minimally invasive surgical approaches, and the availability of advanced mechanical circulatory support (e.g., ECMO) differ substantially between centres and between eras. Users should be aware that the absolute calibration of the RISK-E score may not perfectly translate to their local context.

Variables Not Included in the Score

Several clinically important factors are not captured by the RISK-E score. These include cerebral embolism (present in up to 30% of IE patients and a major determinant of surgical timing), left ventricular ejection fraction, pulmonary hypertension, the urgency of surgery (emergent versus urgent versus elective), specific surgical complexity (single valve versus double valve versus root replacement), concomitant coronary artery disease, and the presence of extra-cardiac complications such as splenic abscess or mycotic aneurysm. Clinicians should incorporate these factors alongside the RISK-E prediction when making surgical decisions.

Score as One Input, Not the Sole Determinant

A high RISK-E score does not automatically mean that surgery should be avoided. Non-operative management of complicated left-sided IE (particularly with uncontrolled infection, large vegetations, severe regurgitation, or periannular abscess) carries its own substantial mortality, which may exceed even a high predicted operative risk. The RISK-E score should be used as one input in a comprehensive risk-benefit analysis, not as an absolute contraindication threshold.

Clinical Application: The Endocarditis Team Framework

Both the European Society of Cardiology (ESC, 2015 and 2023 guidelines) and the American Heart Association (AHA, 2015 guidelines) recommend that all patients with complicated IE be managed by a multidisciplinary Endocarditis Team. This team typically includes cardiologists, cardiac surgeons, infectious disease specialists, clinical microbiologists, cardiac imaging experts, neurologists (when cerebral complications are present), and intensivists. The RISK-E score fits naturally within this framework as a quantitative tool that can inform the team's discussion.

In practice, the RISK-E score can be calculated at the bedside using readily available clinical data (age, valve type, echocardiographic findings, blood culture results, haemodynamic status, laboratory values). It provides an immediate, objective estimate of perioperative risk that can be shared across disciplines and with the patient and family as part of shared decision-making.

Illustrative Clinical Scenarios

Low-risk scenario: A 45-year-old patient with native mitral valve IE caused by Streptococcus gallolyticus, severe mitral regurgitation without periannular complications, stable haemodynamics, normal renal function, and no thrombocytopenia. RISK-E score: 0 points. Predicted mortality: approximately 3%. This patient has a strong surgical indication (severe regurgitation with haemodynamic compromise) and a low predicted operative risk, supporting a recommendation to proceed with surgery.

Moderate-risk scenario: A 66-year-old patient with prosthetic aortic valve IE caused by coagulase-negative Staphylococcus, periannular abscess on TOE, new mild renal impairment, but stable haemodynamics and normal platelet count. RISK-E score: 13 (age) + 6 (prosthetic) + 5 (periannular) + 5 (renal) = 29 points. Predicted mortality: approximately 35%. This patient falls into the high-risk category. The Endocarditis Team should carefully weigh the risks of complex redo surgery against the expected poor outcome of non-operative management of prosthetic valve IE with abscess.

Very high-risk scenario: A 75-year-old patient with prosthetic mitral valve IE caused by MRSA, periannular abscess, septic shock on vasopressors, thrombocytopenia (platelets 80,000), and acute kidney injury requiring renal replacement therapy. RISK-E score: 14 (age) + 6 (prosthetic) + 5 (periannular) + 9 (S. aureus) + 5 (renal) + 7 (septic shock) + 7 (thrombocytopenia) = 53 points. Predicted mortality: approximately 86%. This patient faces extraordinarily high operative risk. Palliative goals, non-operative stabilisation strategies, or the possibility that death is likely regardless of intervention should be discussed transparently with the patient and family.

The Logistic Model: How Points Become Probabilities

The RISK-E score uses a logistic regression equation to convert the total point score into a continuous predicted probability of perioperative mortality. The logistic function has the general form:

Predicted mortality = 1 / (1 + e^{-(intercept + coefficient x score)})

The intercept and coefficient are derived from the two anchor points published by Olmos et al.: a score of 0 maps to approximately 3% mortality and a score of 68 maps to approximately 97% mortality. Converting these probabilities to logits (the natural logarithm of the odds) and solving the simultaneous equations yields an intercept of approximately -3.476 and a coefficient of approximately 0.102 per point. This means that each additional point on the RISK-E score increases the log-odds of perioperative death by approximately 0.102, which translates to a non-constant increase in absolute mortality (small increments at the low and high extremes of the score, larger increments in the middle range).

Optimising Modifiable Risk Factors Before Surgery

While several of the RISK-E variables are non-modifiable (age, prosthetic valve status, causative micro-organism), others reflect the current clinical state and may be amenable to preoperative optimisation. Targeted antimicrobial therapy, haemodynamic support, correction of coagulopathy, renal protective strategies, and timing of surgery relative to the acute infectious episode can all influence perioperative outcomes. The RISK-E score can be recalculated as the patient's clinical status evolves, providing a dynamic risk estimate that can guide the timing of surgical intervention.

For example, a patient who initially presents in septic shock with thrombocytopenia (adding 14 points) may, after 48 to 72 hours of antimicrobial therapy and haemodynamic resuscitation, recover stable blood pressure and normalise platelet counts. Recalculating the RISK-E score at this later time point would yield a substantially lower predicted mortality, supporting a strategy of optimisation before surgery when clinically feasible.

Conversely, delaying surgery to optimise modifiable factors carries its own risks: ongoing embolisation, progressive valve destruction, worsening heart failure, and uncontrolled sepsis. The optimal timing of surgery in IE remains one of the most challenging clinical decisions in cardiovascular medicine, and the RISK-E score should be considered alongside the urgency of the surgical indication.

Special Populations and Emerging Considerations

Injection Drug Use-Associated IE

The rising prevalence of injection drug use-associated IE in North America and parts of Europe has created a growing population of younger patients with predominantly right-sided disease but increasingly left-sided involvement. The RISK-E score has not been specifically validated in this population. These patients often have unique characteristics (younger age, higher rates of hepatitis C coinfection, recurrent IE episodes, psychosocial complexity) that may affect both operative risk and long-term outcomes in ways not captured by the RISK-E model.

Transcatheter and Hybrid Approaches

As transcatheter valve technologies continue to evolve, transcatheter valve replacement is being explored in very high-risk or inoperable IE patients. The RISK-E score, which was derived from a conventional open surgical cohort, may have a role in identifying patients for whom conventional surgery carries prohibitive risk and for whom transcatheter or hybrid approaches might be considered. However, this application remains investigational and is not yet supported by validation data.

The Role of Mechanical Circulatory Support

Extracorporeal membrane oxygenation (ECMO) is increasingly used as a bridge to surgery or bridge to recovery in patients with IE complicated by cardiogenic or septic shock. The availability of ECMO may alter the risk profile of the very highest-risk RISK-E patients, potentially enabling surgery that would otherwise be futile. How ECMO use modifies the calibration of the RISK-E score has not been studied and represents an important area for future research.