An automated high-threshold NT-proBNP referral pathway (>5000 pg/mL) is feasible and identifies very high-risk patients, with enrolled patients experiencing rapid therapy optimisation and observed improvements in symptoms and cardiac function.
Key Findings
Results
Over 12 months, 889 referrals were generated by the automated NT-proBNP pathway, distributed across multiple patient categories.
417 referrals (47%) were inpatients
370 referrals (42%) were known outpatients with HF
59 referrals (7%) were deemed not appropriate for follow-up
43 referrals (5%) were new outpatients without a prior HF diagnosis who were enrolled in the pathway
Results
The median time to review for enrolled patients was 2 days, with 70% meeting the 48-hour target.
The pathway aimed for echocardiography and specialist review within 48 hours
Median time to review was 2 days
70% of enrolled patients met the 48-hour review target
Results
Enrolled patients were predominantly elderly with markedly elevated NT-proBNP levels at presentation.
Mean age of enrolled patients was 81 years
56% of enrolled patients were male
Median NT-proBNP was 7480 pg/mL among enrolled patients
Total enrolled cohort consisted of 43 patients
Results
HF with reduced ejection fraction (HFrEF) was the most common diagnosis among newly enrolled patients.
HFrEF was diagnosed in 21 patients (49%)
HF with mildly reduced ejection fraction was diagnosed in 7 patients (16%)
HF with preserved ejection fraction was diagnosed in 10 patients (23%)
Severe aortic stenosis or new atrial fibrillation was found in 4 patients each (10%)
Results
At 12 weeks, 62% of enrolled patients had improved by at least one New York Heart Association (NYHA) functional class.
62% of patients improved by at least one NYHA class at 12-week follow-up
This reflects symptomatic improvement following rapid therapy optimisation
Follow-up was conducted prospectively among the 43 enrolled patients
Results
Mean left ventricular ejection fraction (LVEF) increased substantially from 27% to 45% at 12 weeks in patients with HFrEF.
Baseline mean LVEF in HFrEF patients was 27%
Mean LVEF increased to 45% at 12-week follow-up
This represents an 18 percentage point absolute improvement in LVEF
HFrEF was diagnosed in 21 of 43 enrolled patients (49%)
Results
At 12 weeks, admission and mortality rates in enrolled patients were 36% and 8%, respectively, compared with 35% and 12% in the retrospective comparator cohort.
The retrospective comparator cohort included n=60 outpatients with NT-proBNP >5000 pg/mL not known to the HF service in the 12 months before pathway implementation
Mortality was numerically lower in the pathway group, though the study was not powered for formal statistical comparison
Methods
The automated pathway triggered referrals for all outpatient NT-proBNP results exceeding 5000 pg/mL, embedded within electronic patient records.
The threshold for automatic HF team referral was NT-proBNP >5000 pg/mL
The pathway was embedded in electronic patient records to automate referral generation
The study period spanned 12 months
The pathway operated across inpatient and outpatient settings
What This Means
This research describes a new automated system designed to quickly identify and treat patients with very high levels of NT-proBNP, a blood marker that signals stress on the heart. When a patient's NT-proBNP result exceeded 5,000 pg/mL, the electronic health record system automatically sent a referral to the heart failure (HF) specialist team. Over 12 months, this generated 889 referrals. Of these, 43 patients were newly diagnosed outpatients without a prior HF diagnosis who were enrolled in the rapid-assessment pathway, receiving an echocardiogram (heart ultrasound) and specialist review within 48 hours in most cases.
Among these newly enrolled patients, who were mostly in their 80s with very high biomarker levels, nearly half were found to have HF with severely reduced heart pumping function (HFrEF). After rapid treatment, the results were notable: 62% of patients felt significantly better at 12 weeks (improving by at least one NYHA functional class), and in those with HFrEF, average heart pumping function nearly doubled — from 27% to 45%. Compared to a historical group of similar patients seen before the pathway existed, mortality at 12 weeks was numerically lower (8% vs 12%), although the study was not large enough to confirm this statistically.
This research suggests that embedding an automated high-threshold NT-proBNP alert directly into electronic patient records is a practical and potentially impactful way to fast-track vulnerable heart failure patients to specialist care. The findings indicate that rapid diagnosis and treatment can lead to meaningful improvements in both symptoms and heart function in a very high-risk elderly population, and may help reduce delays that could contribute to worse outcomes.
Henry J, Brown C, Almeida C, Devin J, Jones T, Vidal M, et al.. (2026). Feasibility and outcomes of an automated high-alert NT-proBNP pathway.. Open heart. https://doi.org/10.1136/openhrt-2026-004325