Prognostic factors and etiological insights in constrictive pericarditis

Abstract

Abstract Constrictive pericarditis (CP) is an infrequent disease that poses a diagnostic and therapeutic challenge, with mortality often linked to underlying etiological factors. Our unicentric retrospective research aims to elucidate prognostic indicators and unravel the diverse etiological spectrum of this condition. The study is accepted by the Ethical Committee of our Center. Between January 2010 and December 2023, we have identified a total of 112 patients diagnosed with CP, with an average follow-up time of 6.42 years (IQR 1.98 - 11.45). We present the baseline variables of the patients, etiology and prescribed treatments in the table 1, comparing patients who survive with those who do not survive. 66.1% of patients were males and median age was 60.8 years old. We present two Kaplan-Meier curves (Figures 1 and 2) comparing the survival rates between CP etiologies. In Figure 1, all causes are represented separately. In Figure 2, CP related to cardiac surgery, radiotherapy, purulent pericarditis, tuberculous pericarditis, mixed and other causes are combined. Our results suggest a better prognosis in patients with idiopathic CP and worse prognosis in patients with oncological related CP. The significant predictors in the univariate Cox analysis were age, cachexia, ascites, oncology etiology, hemoglobin, sodium, Euroscore II and surgical pericardiectomy. The best predictive model (selected with a lower AIC) to predict mortality throughout follow-up, with a Harrell's C of 0.835, included the variables age, cachexia, oncological cause and hyponatremia as risk factors for mortality and surgical treatment by pericardiectomy as a protective factor (Figure 3). In conclusion, our study suggests that CP etiology has an impact on survival with oncological cause having the worst prognosis. Furthermore, other factors as hyponatremia, cachexia, shock and age at diagnosis have also resulted in indicators of poor outcome.Figure 1.Kaplan meierFigure 3.Predictive model