Evidence has consistently demonstrated the lack of clinical utility provided by natriuretic peptides. Further studies documenting their incompetencies are no longer necessary, but it is rather enjoyable to highlight their dependable failures. The most recent illustration of natriuretic peptides clinical ineptitude was published this week in CIRCULATION by Stienen et al (1).
The authors randomized patients admitted to the hospital with an acute exacerbation of heart failure after their initial stabilization phase. Discharge decision was guided by NT-proBNP or by their treating physician.
Patients were enrolled shortly after admission, but did not undergo randomization until they had obtained clinical stability (defined as at least three of the following: target weight reached, stable levels of serum creatinine, improved symptoms and mobilization was possible, and/or adequate discharge medication were prescribed). All patients underwent clinician guided therapy at admission with an initial NT-proBNP level drawn. A second value was obtained following clinical stabilization. Clinicians were blinded to these values. Only after randomization patients in the NT-proBNP-guided strategy had their natriuretic peptide values made available to the treating physicians.
For patients with a 30% reduction in admission NT-proBNP at the point of clinical stability, discharge was recommended. Patients who did not achieve a 30% reduction, entered a predefined algorithm consisting of several steps including additional NT-proBNP measurements and suggested interventions in the hopes of attaining the desired >30% reduction. In the conventional therapy arm, discharge and follow-up were determined by the treating physician.
From 2011 to 2015, 411 patients were included in the primary analysis. The blinded NT-proBNP levels at admission and following randomization were similar between the NT-proBNP and conventional therapy groups. 64% vs. 63% achieved the desired 30% reduction in NT-proBNP levels at the time of randomization. Not surprisingly, once the physicians in the NT-proBNP group were made aware of the NT-proBNP levels and given a protocol to reduce these levels, they far more effectively achieved the desired 30% than the physicians treating patients in the conventional care arm (80% vs. 64%). These differences translated into significant changes in clinical care. A total of 71 NT-proBNP-guided patients failed to achieve the desired 30% reduction in NT-proBNP at randomization and were subjected to the study algorithm, and received significantly more aggressive care than the peptide-free control group.
These natriuretic induced treatment modifications failed to translate into any benefit in meaningful clinical outcomes. There was no difference between the number of days from admission to randomization (both median 5 days) nor from randomization to discharge (median 3 days). All-cause mortality or readmission within 180-days of randomization was 36% in both groups, with readmission rates of 24% and 26% of the NT-proBNP-guided and conventional therapy groups, respectively. Likewise all-cause mortality occurred in 19% and 17% of patients, respectively.
Even more interesting was how the NT-proBNP affected clinical care. Patients in the NT-proBNP-guided therapy group who failed to attain 30% NT-proBNP reduction at randomization had a significantly longer duration of admission when compared to NT-proBNP-guided patients who attained that target at randomization (11 days vs 8 days). They also experienced a longer stay than patients in the conventional therapy arm. Conversely, there was no significant difference in duration of admission in conventionally treated patients regardless of changes in NT-proBNP levels.
This study is a wonderful glimpse into one small microcosm of natriuretic peptides larger scale incompetency. We have discussed the multitude of studies which have discredited their diagnostic prowess in the Emergency Department, their lack of utility in the inpatient setting and even their inability to assist in the outpatient management of CHF. This study addresses one single question, does % NT-proBNP clearance assist in the discharge decision in patients admitted for a CHF exacerbation when added to unstructured clinical judgment? The short answer is, no. In approximately 65% of patients, desired NT-proBNP levels were achieved with conventional therapy alone. Natriuretic peptides failed to provide additional clinical guidance. The remaining minority, when treated using NT-proBNP-guided algorithms, were exposed to more aggressive medical therapies and needless delays in hospital discharge without any gains in clinically important outcomes.
This publication serves as a reminder of the dangers of basing clinical decisions off surrogate endpoints. Without a control group, the results of the PRIMA II trial would have been considered a success. The authors noted a significant portion of the cohort had their treatment plan altered as a result of the NT-proBNP level. But the addition of the control arm reveals that while the use of natriuretic peptides may change care, they do so in a self-serving manner. Stienen et al have demonstrated that the use of a surrogate-guided treatment strategy can effectively modify this same surrogate in a positive manner. But when such surrogates fail to translate into clinically meaningful endpoints, their use is nothing more than a costly extraneous diagnostic redundancy.
- Stienen S, Salah K, Moons AH, et al. NT-proBNP-Guided Therapy in Acute Decompensated Heart Failure: The PRIMA II Randomized Controlled Trial. Circulation. 2017;