NT-proBNP predicts future heart failure and mortality in asymptomatic

 A recent article tracked changes over time of N-terminal B-type natriuretic peptide (NT-proBNP) in those without known heart failure (ie primary prevention) finding that changes in NT-proBNP were associated with both incident heart failure and death (see chf bnp predicts HH development JAMAcardiol2023 in dropbox or  doi:10.1001/jamacardio.2022.5309) 

 

Details:

-- 9776 individuals in the Atherosclerosis Risk in Community (ARIC) study from 4 US communities (in North Carolina, Mississippi, Minnesota, and Maryland)

-- data were collected from patient at visits 2 and 4, approximately 6 years apart, both with measurements of NT-proBNP

-- patients with prevalent heart failure were excluded

-- mean age 63, 57% women, 79% white/21% Black

-- hypertension 46%/on hypertension medications 34%/mean blood pressure 128/70, LDL 123, triglyceride 62, on cholesterol-lowering medications 14%, diabetes 15%, BMI 29, eGFR 86, current smoking 14%, prevalent coronary heart disease 7%

-- primary outcome: incident heart failure (HF) hospitalizations and all-cause death, with NT-proBNP grouped as either <125 pg/ml vs >125 pg/ml, as well the relationship with the percent change in NT-proBNP

-- also assessed was the association between changes in cardiovascular risk factors with changes in NT-proBNP

-- the fully-adjusted model included systolic and diastolic blood pressures, being on hypertensive medication use, diabetes, fasting glucose, LDL, triglyceride, being on cholesterol-lowering medication, cigarette smoking, eGFR, BMI, and prevalent heart disease.

 

Results:

--  as compared to those with NT-proBNP <125pg/ml at both visits (reference), all with adjusted hazard ratios:

    -- NT-proBNP <125pg/ml at visit 2 and  >125 pg/ml at visit 4

        -- incident HF: HR 1.86 (1.60-2.16)

        -- mortality risk: HR 1.32 (1.19-1.47)

    -- NT-proBNP >125pg/ml at visit 2 and  <125 pg/ml at visit 4

        -- incident HF: HR 1.01 (0.71-1.43)

        -- mortality risk: HR 0.79 (0.61-1.01)

    -- NT-proBNP >125pg/ml at visit 2 and  >125 pg/ml at visit 4

        -- incident HF: HR 2.40 (2.00-2.88)

        -- mortality risk: HR 1.68 (1.47-1.91)

    -- ie, as compared to those who remained with a low NT-proBNP, those whose NT-proBNP increased on visit 4 had a significant rise in HF, which was even more significant in those who started with a higher NT-proBNP and remained high. But those who started high and then had a lower level of NT-proBNP had no worse outcomes than those who began low and remained low

    -- there was no interaction between NT-proBNP’s relationship between HF incidents vs death (i.e. these were independent phenomena)

-- Stratification by percent change in NT-proBNP:

    -- visit 2 NT-proBNP <125pg/ml, as compared to no change by visit 4 (reference), all with adjusted HRs:

        -- >25% decrease:

            -- incident HF: HR 1.15 (0.86-1.53)

            -- mortality: HR 1.14 (0.94-1.38)

        -- >25% increase:

            -- incident HF: HR 1.93 (1.39-2.67)

            -- mortality: HR 1.56 (1.26-1.94)

    -- visit 2 NT-proBNP >125pg/ml, as compared to no change by visit 4:

        -- >25% decrease:

            -- incident HF: HR 0.86 (0.57-1.29)

            -- mortality: HR 0.59 (0.44-0.79)

        -- > 25% increase:

            -- incident HF: HR 6.56 (2.27-18.99)

            -- mortality: HR 5.47 (2.58-11.60)

-- Changes in cardiovascular risk factors:

    -- levels of NT-proBNP were positively associated with age, systolic blood pressure, being on hypertension medication use, smoking, and prevalent heart disease

    -- but there was an inverse association of NT-proBNP levels in men, diastolic blood pressure, triglyceride level, cholesterol medication use, eGFR, BMI, and those of Black race

        -- absolute change in NT-proBNP did correlate positively with systolic blood pressure change, particularly in those with an increase in NT-proBNP by at least 132 pg/mL

 

Commentary:

-- NT-proBNP has been documented to be an effective biomarker for both the diagnosis and prognosis of heart failure. In those without known cardiovascular disease, it is associated with subsequent HF, cardiovascular disease, stroke, and all-cause mortality. The current AHA/ACC guidelines on heart failure include NT-proBNP as part of their HF classification, with those having levels greater than 125 pg/mL being considered to be in stage B or pre-HF category (stage B includes either high BNP determinations as well as persistently elevated cardiac troponin levels: 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines | Circulation (ahajournals.org)

-- this all brings up the question of the differences between the NT-proBNP and BNP markers

    -- whereas BNP is the biologically active hormone and is degraded, NT-proBNP is its nonfunctional precursor, is cleared passively, and has a much longer half-life

    -- levels of both NT-proBNP and BNP do have a significant genetic component, explaining about 40% of their variation

    -- BNP levels tend to increase with age and be higher in women; NT-proBNP levels tend to be higher in those with left ventricular dysfunction; and the measured levels tend to be more standardized for NT-proBNP than for BNP. Levels of both hormones tend to be lower in obese individuals

    -- both levels are higher in patients with renal failure, though much more significantly so with NT-proBNP (so these NT-proBNP levels are not so helpful in diagnosing and managing heart failure, but a low level of BNP would reasonably effectively rule out left ventricular dysfunction. NT-proBNP is renally excreted, which might explain its higher levels in patients with renal failure (see https://pubmed.ncbi.nlm.nih.gov/24372567/ and https://bmcnephrol.biomedcentral.com/articles/10.1186/1471-2369-14-117) 

        -- one study (J Am Coll Cardiol. 2006;47(1):91) found “NT-proBNP values of >450 pg/ml for patients ages <50 years and >900 pg/ml for patients ≥50 years had a sensitivity of 85% and a specificity of 88% for diagnosing acute CHF among subjects with GFR ≥60 ml/min/1.73 m². Using a cut point of 1,200 pg/ml for subjects with GFR< 60 ml/min/1.73 m², we found sensitivity and specificity to be 89% and 72%, respectively” : [ie pretty large variations in NT-proBNP in different types of patients, and much larger numbers than the small 125 pg/mL cutpoint in this study, or the 25% changes]

    -- of note, the Framingham Study found that higher BNP levels predated higher blood pressure (as was found in the ARIC trial with NT-proBNP): see https://www.ahajournals.org/doi/10.1161/01.hyp.0000061116.20490.8d#:~:text=Increased%20brain%20natriuretic%20peptide%20(BNP,BP%20increase%20and%2For%20hypertension )

    -- NT-proBNP levels are more accurate in those on sacubitril-valsartan than BNP levels, since the latter rises with neprilysin inhibition as found in the combo sacubitril-valsartan:  https://pubmed.ncbi.nlm.nih.gov/30846338/ 

    -- and, unfortunately, there was a poor correlation between BNP and NT-proBNP in a study of 3,029 patients, both inpatient and outpatient: https://www.aacc.org/science-and-research/scientific-shorts/2019/bnp-or-nt-probnp . so makes sense to stick with one of BNP/NT-proBNP to follow for changes

 

-- this ARIC study had several important findings:

    -- there was a graded response of both increased risk of heart failure and mortality (though these were found to be independently associated) as patients were stratified from low levels of NT-proBNP in both the initial and subsequent visits, and increasing lots to those with higher levels at both visits

    -- those with higher NT-proBNP initially (visit 2), but had at least a 25% decrease 6 years later by visit 4, actually had a lower risk of death than those who had low levels at both visits (ie a decrease in NT-proBNP from high to low seemed to be unexpectedly better than just remaining low)

        -- and, this relationship was found with as little as a NT-proBNP being 9 pg/mL higher at visit 2 than visit 4

-- some other studies have found a relationship between increased NT-proBNP and cardiovascular disease, including the MESA study, but these other studies have included people both with and without baseline heart failure

    -- and, this ARIC study had younger patients than the other studies, had longer follow-up than others, and involved patients without baseline heart failure (ie, likely a real "primary prevention" study)

 

Limitations:

-- this was an observational study, and as a result one cannot draw causal conclusions (i.e. one cannot clearly conclude that it is specifically the NT-proBNP levels that are causing incident heart failure and mortality, since there could be known or unknown confounders that mediate this relationship hence the need for RCTs); for example several cardiovascular risk factors were increased or decreased with NT-proBNP levels which could potentially mediate NT-proBNP's association with these clinical outcomes)

-- in particular, as an observational study, we cannot draw clear conclusions that in a cohort of asymptomatic patients without known history of heart failure such as those included in ARIC, that interventions to lower NT-proBNP is beneficial

-- this study also had only two measurements of NT-proBNP 6 years apart. We do not have information on any intermediate levels to get a more informed understanding of the clinical effect of fluctuations in NT-proBNP or when to consider repeated measuring NT-proBNP (eg, the unexpected finding that going from a higher to a lower NT-proBNP might be explained by unmeasured changes in the 6 year period between ARIC visits)

-- there was no echocardiographic confirmation of the heart failure, or what classification of heart failure was present, though these patients were all admitted to the hospital and presumably this information was known

-- they limited the outcomes to clinical heart failure leading to admission to the hospital, limiting our understanding of those with lesser degrees of heart failure who were treated as outpatients

-- there was also no granular information about the causes of the heart failure in those admitted to the hospital: was it related to atrial fibrillation, valvular heart disease (both of which can impact NT-proBNP levels), thyroid disease, tachycardia, etc. So, not sure we can generalize the aggregate outcome (hospital admission for HF) to all subtypes or causes of HF

 

So, a few comments:

-- it seems to may make sense to check NT-proBNP levels as a predictor of higher risk of heart failure and mortality (though we do not have information on them appropriate interval for repeated tests, and it seems that a single high value that subsequently normalizes at 6 years is actually beneficial vs the level always being low...)

-- given the strong push for prescribing SGLT-2 inhibitors as a means to decrease heart failure, perhaps it makes sense to check NT-proBNP levels and selectively consider prescribing SGLT-2’s, maybe especially in those with high levels (eg >125 pg/mL) who have a subsequent one also high (especially since SGLT-2’s are not such benign drugs: see https://gmodestmedblogs.blogspot.com/2019/05/sglt-2-inhibitors-and-fourniers-gangrene.html  and https://gmodestmedblogs.blogspot.com/2018/11/sglt-2-inhibitors-twice-risk-of.html , as well as being assoicated with severe urinary tract infections (and urosepsis), mycotic infections, DKA at even low blood sugar levels, etc

-- it remains likely but unproven that interventions to lower NT-proBNP levels would be clinically helpful in asymptomatic patients with higher levels.

-- it does seem that the conclusions of this current ARIC study should lead to more aggressive general cardiovascular risk factor reductions (lipids, hypertension, diabetes, smoking, overweight, etc) in those with high NT-proBNP levels, since this seems to be a predictor of cardiovascular disease

-- given the clinical differences between NT-proBNP and BNP, it would be useful to have a study assessing the relationship between changes in BNP levels and incident heart failure and mortality (notably, per the AHA/ACC guidelines, they do not differentiate between these two entities in terms of defining pre-HF)

-- and it is important that the laboratories we use have access to each of these tests 

geoff