Introduction: Using biomarkers to predict mortality in patients with sepsis is important because these patients frequently have high mortality rates and unsatisfactory outcomes. The performance of N-terminal pro-brain natriuretic peptide (NT-pro-BNP) and C-reactive protein (CRP) to predict clinical outcomes in elderly sepsis patients is unimpressive. We aimed to assess the prognostic value of NT-pro-BNP, CRP and the combination of both in selected medical ICU sepsis patients more than 75 years old.
Methods: In total, 245 consecutive patients were screened for eligibility and followed during their ICU stays. We collected the patients' baseline characteristics, including their Acute Physiology and Chronic Health Evaluation II (APACHE II) scores and NT-pro-BNP and CRP levels. The primary outcome was ICU mortality. Potential predictors were analyzed for their possible associations with the outcome. We also evaluated the ability of NT-pro-BNP and CRP levels combined with the APACHE II score to predict ICU mortality by calculating the C-index and net reclassification improvement (NRI).
Results: Univariate regression revealed that CRP, NT-pro-BNP, APACHE II score, lactic acid level, NEU count(neutrophil count)and HCT level independently predicted ICU mortality (all P < 0.01). The C-index for the prediction of ICU mortality by the APACHE II score (0.847 ± 0.029; P < 0.001) was greater than that for NT-pro-BNP (0.673 ± 0.039; P < 0.01) or CRP (0.626 ± 0.039; P < 0.01) (all P < 0.01). Compared with the APACHE II score (0.847 ± 0.029; P < 0.001), the combination of CRP (0.849 ± 0.029; P < 0.01) or NT-pro-BNP (0.853 ± 0.028; P < 0.01) or both (0.853 ± 0.030; P < 0.01) with the APACHE II score did not significantly increase the C-index for predicting ICU mortality (all P > 0.05). However, the addition of NT-pro-BNP to the APACHE II score gave an NRI of 8.6% (P = 0.000), the addition of CRP to the APACHE II score provided an NRI of 11.34% (P = 0.012), and the addition of both markers to the APACHE II score yielded an NRI of 29.0% (P = 0.000). In the MOF subgroup (N = 118), CRP (OR = 2.62, P < 0.05) but not NT-pro-BNP (OR = 1.73, p > 0.05) independently predicted ICU mortality, and the addition of CRP to the APACHE II score obviously increased its predictive ability (NRI = 13.88%, P = 0.000). In the non-MOF group (N = 127), neither CRP (OR = 5.447, P = 0.078) nor NT-pro-BNP (OR = 2.239, P = 0.016) was an independent predictor of ICU mortality.
Conclusions: In sepsis patients older than 75 years, NT-pro-BNP and CRP can serve as independent predictors of mortality, and the addition of NT-pro-BNP or CRP or both to the APACHE II score significantly improves the ability to predict ICU mortality. CRP appears to be useful for predicting ICU outcomes in elderly sepsis patients with multiple-organ failure.
Keywords: C-reactive protein; NT-pro-BNP; Predictive value; Sepsis.