Closing the implementation Gap: laboratory-integrated decision support for NT-proBNP-based heart failure diagnosis and risk stratification
L M Laura Martin, R De La Espriella, A P Adela Pozo, R M Rosa Murria, E S Enrique Santas, M L Miguel Lorenzo, J M Jorge Montiel, A G Andrea Gasull, G N Gonzalo Nunez, S V Sandra Villar, J S Juan Sanchis, E R Enrique Rodriguez, J N Julio NunezAbstract
Background/Introduction
Despite guideline endorsement, implementation gaps persist in natriuretic peptide testing due to multiple cut-points, dependence on pre-test probability, frequent modifiers (age, renal function, obesity), and limited applicability of conventional thresholds for detecting decompensation in established HF. These factors contribute to delays in diagnosis and treatment. A laboratory-integrated clinical decision support (CDS) system could standardize NT-proBNP interpretation and support earlier risk stratification.
Purpose
To evaluate the diagnostic and prognostic utility of a laboratory-integrated CDS pathway for NT-proBNP-based evaluation of suspected de novo HF and HF decompensation/progression in primary care and the emergency department.
Methods
Pragmatic prospective cohort study embedded within our Institute, leveraging bidirectional integration between laboratory information system and electronic health records. Two CDS-enabled profiles were implemented. (i) De novo HF pathway: eligibility required ≥2 clinical signs/symptoms; NT-proBNP stratified by age/obesity in primary care (<50 years or obesity: <149 pg/mL=rule-out, 150-2000=rule-in, >2000=high-risk; ≥50 years without obesity: <300=rule-out, 300-2000=rule-in, >2000=high-risk). In ED, age-stratified thresholds of 450/900/1800 pg/mL for <50/50–75/>75 years were applied. (ii) HF progression pathway: individualized "dry" NT-proBNP estimated as the median of prior stable outpatient values; alerts generated when NT-proBNP exceeds >100% of dry value. Primary endpoint: time to all-cause death or worsening HF (WHF) at 180 days. Secondary endpoints: all-cause mortality and WHF (competing-risks analysis, death as competing event).
Results
Between February 2024 and January 2025, 2508 patients were included (mean age 77.0±11.6 years; 55.7% women; 55.3% suspected de novo HF; 60.2% ED activations). After applying CDS rules, 1420 (56.6%), 247 (9.9%), and 841 (33.5%) patients were classified as rule-out, grey-zone, and rule-in, respectively. Over 180 days, 398 primary composite events occurred. Cumulative incidence across CDS categories is shown in the Table. In multivariable Cox regression, rule-in was independently associated with the primary endpoint (HR 2.32; 95% CI 1.82–2.96; P<0.001) and mortality (HR 2.32; 95% CI 1.82–2.96; P<0.001). In adjusted Fine–Gray models, rule-in had a higher WHF risk (SHR 2.55; 95% CI 1.75–3.72; P<0.001). Grey-zone showed no significant difference versus rule-out for any endpoint.
Conclusions
Laboratory-integrated CDS for NT-proBNP interpretation effectively stratifies short-term risk in patients evaluated for suspected new-onset or decompensated HF. Rule-in classification identifies individuals at substantially higher risk, supporting its role in earlier intervention and specialist referral.180-day cumulative incidence riskFor image description, please refer to the figure legend and surrounding text.180-days outcomesFor image description, please refer to the figure legend and surrounding text.