Enhancement of CURB65 score with proadrenomedullin (CURB65-A) for outcome prediction in lower respiratory tract infections: derivation of a clinical algorithm

Abstract

Background: Proadrenomedullin (ProADM) confers additional prognostic information to established clinical risk scores in lower respiratory tract infections (LRTI). We aimed to derive a practical algorithm combining the CURB65 score with ProADM-levels in patients with community-acquired pneumonia (CAP) and non-CAP-LRTI.

Methods: We used data of 1359 patients with LRTI enrolled in a multicenter study. We chose two ProADM cut-off values by assessing the association between ProADM levels and the risk of adverse events and mortality. A composite score (CURB65-A) was created combining CURB65 classes with ProADM cut-offs to further risk-stratify patients.

Results: CURB65 and ProADM predicted both adverse events and mortality similarly well in CAP and non-CAP-LRTI. The combined CURB65-A risk score provided better prediction of death and adverse events than the CURB65 score in the entire cohort and in CAP and non-CAP-LRTI patients. Within each CURB65 class, higher ProADM-levels were associated with an increased risk of adverse events and mortality. Overall, risk of adverse events (3.9%) and mortality (0.65%) was low for patients with CURB65 score 0-1 and ProADM ≤0.75 nmol/l (CURB65-A risk class I); intermediate (8.6% and 2.6%, respectively) for patients with CURB65 score of 2 and ProADM ≤1.5 nmol/l or CURB classes 0-1 and ProADM levels between 0.75-1.5 nmol/L (CURB65-A risk class II), and high (21.6% and 9.8%, respectively) for all other patients (CURB65-A risk class III). If outpatient treatment was recommended for CURB65-A risk class I and short hospitalization for CURB65-A risk class II, 17.9% and 40.8% of 1217 hospitalized patients could have received ambulatory treatment or a short hospitalization, respectively.

Conclusions: The new CURB65-A risk score combining CURB65 risk classes with ProADM cut-off values accurately predicts adverse events and mortality in patients with CAP and non-CAP-LRTI. Additional prospective cohort or intervention studies need to validate this score and demonstrate its safety and efficacy for the management of patients with LRTI.

Trial registration: Procalcitonin-guided antibiotic therapy and hospitalisation in patients with lower respiratory tract infections: the prohosp study; isrctn.org Identifier: ISRCTN: ISRCTN95122877.

Figures

Figure 1

Estimated association between initial ProADM values and the risk of adverse events (A) and death (B). Estimated association between initial ProADM values and the risk of adverse events (upper black line) and death (lower blue line). Estimates are based on generalized additive models and shaded gray regions correspond to (point-wise) 95% confidence intervals. The rugs at the bottom of the plots display the distribution of the biomarker. Solid lines (and confidence intervals) based on imputed data, dashed lines based on complete-case analysis.

Figure 2

Observed and expected adverse events (A) and mortality (B) within CURB65 class 0/1 (low risk patients). Observed and expected adverse events and mortality within CURB65 class 0/1 (low risk patients). A. Calibration of initial ProADM deciles and adverse events. B. Calibration of ProADM deciles and mortality.

Figure 3

Observed and expected adverse events (A) and mortality (B) within CURB65 class 3-5 (high risk patients). Observed and expected adverse events and mortality within CURB65 class 3-5 (high risk patients). A. Calibration of ProADM deciles and adverse events. B. Calibration of initial ProADM deciles and mortality.

Figure 4

Combination of CURB65 risk classes and ProADM tertiles in all LRTI patients. Combination of CURB65 risk classes and initial ProADM tertiles in all LRTI patients. A. Adverse events in CURB65 and ProADM categories. B. Mortality in CURB65 and ProADM categories. Risk class I = green, risk class II = orange, risk class III = red.

Figure 5

Combination of CURB65 risk classes and ProADM tertiles in CAP patients. Combination of CURB65 risk classes and initial ProADM tertiles in CAP patients. A. Adverse events in CURB65 and ProADM categories. B. Mortality in CURB65 and ProADM categories. Risk class I = green, risk class II = orange, risk class III = red.

Figure 6

Risk classification stratified for outpatients (A) and inpatients (B). Risk classification stratified for outpatients (A) and inpatients (B) according to CURB65-A. Risk class I = green, risk class II = orange, risk class III = red.