Development and validation of a nomogram for predicting long-term survival in patients with infective endocarditis: a single-center retrospective study.
Song T, Hai B, et al. • Scandinavian cardiovascular journal : SCJ • 2026
A nomogram integrating clinical, microbiological, and treatment-related variables to predict long-term survival in IE patients demonstrated good discrimination with AUCs of 0.872, 0.840, and 0.846 for 1-, 3-, and 5-year survival prediction, respectively, and may support individualized risk assessment and prognostic stratification.
Key Findings
Results
Seven independent prognostic factors were identified and incorporated into the nomogram for predicting long-term survival in infective endocarditis patients.
LASSO regression was used for variable selection, identifying 11 candidates, from which 7 independent prognostic factors were confirmed via multivariable Cox regression.
The study retrospectively analyzed 210 patients with IE, of whom 67.1% were male with a median age of 55 years.
125 patients (59.5%) underwent surgical treatment during the study period.
Variables incorporated spanned clinical, microbiological, and treatment-related domains.
Results
The nomogram demonstrated good discrimination for predicting survival at 1, 3, and 5 years.
Time-dependent ROC curve analysis yielded AUCs of 0.872, 0.840, and 0.846 for 1-, 3-, and 5-year survival prediction, respectively.
The concordance index was 0.826, indicating strong overall model discrimination.
Internal validation was performed using bootstrap resampling.
Calibration and decision curve analyses demonstrated good predictive accuracy and clinical utility.
Results
Blood culture-negative infective endocarditis was associated with poorer long-term survival.
Blood culture-negative IE was identified as an independent prognostic factor in the multivariable Cox regression model.
This finding highlights the prognostic relevance of microbiological characteristics in IE outcomes.
Blood culture-negative status was among the 7 factors retained in the final nomogram.
Results
Gram-negative bacterial infection was associated with poorer survival in patients with infective endocarditis.
Gram-negative bacterial infection was identified as an independent predictor of worse long-term outcomes.
This microbiological characteristic was included among the 7 independent prognostic factors in the nomogram.
The finding underscores the importance of pathogen type in prognostic stratification of IE patients.
Results
Surgical treatment was associated with improved long-term survival in infective endocarditis patients.
Surgical treatment was identified as an independent prognostic factor predicting improved outcomes in multivariable Cox regression.
Among the 210 patients analyzed, 125 (59.5%) underwent surgery.
Surgical treatment was incorporated as one of the 7 variables in the final nomogram.
Methods
The study used a retrospective single-center design with LASSO and multivariable Cox regression methodology to build and internally validate the prognostic nomogram.
Data collected included clinical characteristics, laboratory data, echocardiographic findings, microbiological profiles, and treatment information.
LASSO regression was applied first for variable selection to reduce overfitting.
Model performance was evaluated using time-dependent ROC curves, concordance index, calibration curves, and decision curve analysis.
Internal validation was performed using bootstrap resampling to assess model generalizability.
What This Means
This research suggests that a prediction tool called a nomogram, built using data from 210 patients with infective endocarditis (a serious infection of the heart's inner lining), can reliably estimate how likely a patient is to survive over the long term. The researchers identified seven key factors that independently influenced survival, including whether bacteria could be detected in blood cultures, what type of bacteria caused the infection, and whether the patient had surgery. The model performed well, correctly ranking survival risk in roughly 83% of cases and showing strong accuracy at predicting 1-, 3-, and 5-year survival.
Two microbiological findings stood out: patients whose blood cultures came back negative (meaning the causative bacteria could not be identified) and patients infected with Gram-negative bacteria both had worse survival outcomes. On the other hand, patients who underwent surgical treatment had better long-term survival. These findings highlight how the type of infecting organism and treatment approach are not just clinical concerns but also critical factors in predicting a patient's prognosis.
This research suggests that clinicians may be able to use this nomogram as a practical tool to estimate individual patients' long-term survival risk at the time of diagnosis, potentially helping guide decisions about treatment intensity and surgical intervention. However, since this was a single-center retrospective study, external validation in larger and more diverse patient populations would be needed before the tool could be broadly adopted in clinical practice.
Song T, Hai B, Yang W, Wu S, Wu N. (2026). Development and validation of a nomogram for predicting long-term survival in patients with infective endocarditis: a single-center retrospective study.. Scandinavian cardiovascular journal : SCJ. https://doi.org/10.1080/14017431.2026.2726653