limitations of NT-proBNP with renal failure

 I don’t mean to perseverate on this topic, but there is a remarkable dearth of information on how to interpret NT-proBNP levels in patients with chronic kidney disease (CKD), an unfortunately common scenario. A recent article found that the ratio of NT-proBNP/cystastin C may have the best association with clinical heart failure (see heart failure ratio NT-proBNP cystat best frontcardmed2021 in dropbox or doi: 10.3389/fcvm.2021.731864) 

  

Details: 

-- 396 patients in this Chinese study were hospitalized with CKD and had both cystatin C (CysC) as well as NT-proBNP levels as well as echocardiography done

-- these patients were divided into heart failure group (n=216) and control group (n=180) 

-- mean age 73, 62% male, BMI 25 

-- the variables assessed that were significantly different between the heart failure and control groups were (or the definition of the echocardiographic abbreviations, see below): LVEF (57% versus 62%), E/e' (19 versus 15), LAVI (38 versus 33), LVMI (111 versus 108), serum creatinine (257 mmoles/L versus 202 mmoles/L; or 2.9 mg/dL vs 2.5 mg/dL), cystatin C (3.2 versus 3.3), hs-cTnT (highly sensitive cardiac troponin: 0.018 versus 0.015). Also,

    -- NT-proBNP (8505 pg/mL versus 4271 pg/mL) 

    -- NT-proBNP/CysC (1283 versus 637) 

 

Results: 

-- NT-proBNP/CysC was an independent risk factor for cardiac dysfunction 

-- when comparing the actual benefits of different tests, the test reaching closest to the upper left corner (1-specificity closest to 0; sensitivity closest to 1) in the ROC curve is the best test: see graph below, where the NT-proBNP /CysC^1.53  had the best association with cardiac dysfunction

-- the area under the ROC curve for the NT-proBNP/CysC^1.53 was 0.815 (0.772-0.858) 

    -- for a cut-off point of 848, there was a sensitivity of 78% and specificity of 69%  

 

 

 

-- echocardiographic abbreviations in the article:

    -- LVDd and LVDs =left ventricular end-diatostolic and end-systolic dimesions

    -- LVEF= left ventricular ejection fraction

    -- LAVI= left atrial volume index

    -- LMVI= left ventricular mass index

    -- GAS= global area strain

    -- GLS= global longitudinal strain

    -- FS= ventricular fractional shortening

    -- SV= stroke volume

    -- E/e'= echocardiographic ratio of early diastolic mitral inflow velocity to early diastolic mitral annulus velocity (a measure for diastolic dysfunction)

-- there was no correlation between the NT-proBNP and NT-proBNP/CysC^1.53 ratio in cardiac dysfunction

 

Commentary: 

-- CKD and heart failure (HF) are intimately linked: 

    -- the main cause of death among patients with CKD is cardiovascular disease, including myocardial infarction and heart failure 

    -- CKD does lead to sodium and fluid retention, a frequent pre-condition for clinical HF, and may well be associated with increased cardiac preload, ventricular pressures, and NT-proBNP levels in the blood 

    -- a recent study found that routinely measuring BNP levels in patients with CKD decreased both kidney replacement therapy as well as heart failure hospitalizations: https://gmodestmedblogs.blogspot.com/2023/12/routine-bnp-assessment-helpful-for.html

-- the problem is that there are factors that affect the results of NT-proBNP (more so than plain BNP measurements), with the European Society of Cardiology (see below) commenting: age is an important factor but the "results should always be interpreted in consideration of renal function and body mass index, the two most powerful confounders of NP concentrations (BNP, NT-proBNP, MR-proANP) "

    -- by the way, I am not so sure that age is necessarily such a clear risk factor for elevated NT-proBNP: it could well turn out that age is at least a partial surrogate marker for CKD, since CKD increases so significantly with age. A well-designed multi-factorial study could help disaggregate the real risk factors involved in interpreting NT-proBNP levels

-- BNP, and especially NT-proBNP levels, are elevated in those with decreased glomerular filtration  (eg: see https://www.sciencedirect.com/science/article/pii/S0735109705023521?via%3Dihub )

-- AND, we do not have a reliable nomogram that takes into account eGFR levels in interpreting NT-proBNP (or BNP) in the setting of CKD

-- unfortunately, the standard HF physical exam of assessing inspiratory rales on chest exam, elevated jugular venous pressure, and leg edema have a sensitivity only 50 to 60%, though the specificity is high. EKGs and chest x-rays don't add much to this 

-- this study found that NT-proBNP was not so useful as a measure of cardiac dysfunction in those with CKD, but that incorporating cystatin C measurements in the calculation as the ratio NT-proBNP/CysC^1.53 had the best correlation with heart failure. And the predictive value of this ratio was completely independent of just assessing the NT-proBNP value by itself

 

-- the European Society of Cardiology had a 2019 position paper on this issue of NT-proBNP measurements (https://onlinelibrary.wiley.com/doi/10.1002/ejhf.1494 )

   --they incorporated the results of J Januzzi's studies on age and NT-proBNP levels, noting that a patient >75yo who likely has heart failure has an NT-proBNP level >1800 pg/mL (with a BNP level >400); "with a grey zone" (indeterminate, requiring ancillary testing) of 300-1800 for NT-proBNP in someone >75yo or BNP 100-400. See table:

 

 

        -- these conclusions were based on the 2018 report by Januzzi ea al (J Am Coll Cardiol 2018; 71: 1191–1200, which is an updated analysis of their 2006 study Eur Heart J 2006;27:330-337). Of note, the BNP concentrations are much less affected by age, leading to the above age-related cut-points only for NT-proBNP levels.

-- a 2023 update by the American College of Cardiology used the same NT-proBNP cutpoints as the European society ones above for determining the likelihood of heart failure (based on the same Januzzi analyses), with those same age cutpoints: J Am Coll Cardiol2018;71:1191–1200, and the 2023 update https://www.acc.org/Latest-in-Cardiology/Articles/2015/02/09/13/00/Cardiac-Biomarkers-and-Heart-Failure

-- also, of note, patients with similar clinical degrees of heart failure with preserved ejection fraction have lowered BNP/NT-proBNP levels than those with reduced ejection fraction [heart failure subtypes were not differentiated in the above analysis by Januzzi]

-- the American guidelines also do note that "Cystatin C and β trace protein (BTP) performed better than traditional renal markers for determining prognosis in HF" 

 

-- i do need to bring up the elephant in the room here (yet again)

-- eGFR is poorly correlated with measured GFR. Some studies have suggested cystatin-based eGFR is better than creatinine-based eGFR, which does make intuitive sinse since there are so many factors affecting creatinine excretion (eg muscle mass, drug interactions, etc), though a recent study did not find much difference in estimated GFR by cystatin vs creatinine in comparison to measured GFR (ie, both were bad): https://gmodestmedblogs.blogspot.com/2022/07/egfr-not-such-great-estimate-of-renal.html

-- however,  cystatin-based eGFR seems to be much more predictive of cardiovascular and renal outcomes than the creatinine based one: https://gmodestmedblogs.blogspot.com/2023/12/cystatin-c-better-predictor-of-bad.html

 

Limitations: 

-- This study had relatively low numbers of patients in both the heart failure and control groups, limiting the potential accuracy of their numbers 

-- this was also a Chinese study with a homogeneous group of patients, limiting generalizability to other areas of the world 

-- as noted in this article "There are no large-scale prospective clinical trials that provide an accurate cutoff value of NT-pro BNP with a diagnosis of heart failure in patients with CKD" 

-- though BNP levels (vs NT-proBNP) are less affected by CKD, there is a real concern about the BNP levels, where there is less standardization across assays

-- in this study there were only 9 patients with stage 1-2 CKD, limiting generalizability of the results to those with less severe CKD; similarly there were only 12 patients with symptoms/signs of HF and with a "grey zone" HFA score of between 2-4 (Heart Failure Assessment score, which applies to those with HF with preserved ejection fraction (HFpEF), where <2=low probability of HFpEF, and >4=diagnostic of HPpEF); all of this limiting overall generalizability

-- the major studies cited for deriving values for both NT-proBNP and BNP values were done by age, with a few concerns:

-- these were all published by the same group (with Januzzi), and one should be more skeptical when only one group has published results (ie no one else verifying them from a non-conflicted vantage point)

-- their age buckets were very large  (<50, 50-75, >75) which makes it pretty imprecise (being 49yo vs 75yo is a much bigger difference than 49yo vs 50yo)

-- there is no evaluation of the role of kidney function, or BMI (perhaps the major confounders)

-- there should likely be different cutpoints for the BNPs if the patient has reduced vs preserved ejection fractions (the former is associated with higher BNP values for the same degree of clinical heart failure)

-- this study did not differentiate between those patients who had heart failure with reduced vs preserved ejection fractions (see https://gmodestmedblogs.blogspot.com/2023/10/heart-failure-preserved-in-obese.html for more info on the pathophysiologic differences), limiting our ability to apply their results to the individual patients we see with heart failure

 

so, what does this small study suggest:

-- it confirms the well-known fact that NT-proBNP values, and their relationship with clinical heart disease, requires the input of the level of CKD for the individual patient

    -- there is likely a real value of incorporating the individual’s cystatin-C into interpreting the NT-proBNP (likely along with BMI and age)

-- BUT, we have almost no studies informing us on how to incorporate these items into interpreting the NT-proBNP test results

    -- the study by Januzzi is a step forward, incorporating age into the NT-proBNP interpretation, though this has some limitations as noted above.

        – in a separate study, Januzzi did find major changes in NT-proBNP values in those with renal transplants depending on the functioning of the allograft: https://link.springer.com/article/10.1007/s00508-009-1248-x ; and another on lesser degrees of CKD https://www.sciencedirect.com/science/article/pii/S0735109705023521?via%3Dihub.  But there was no assessment of CKD cutpoints in determining cardiac function

    -- the big issue here is that we in primary care, as well as specialists in renal and cardiology, do rely on the NP tests to help diagnose as well as monitor patients with heart failure. And there is a huge overlap of heart failure with CKD, increasingly so as the population ages

    -- which really brings up the issue: in this common combination of medical problems often associated with significant morbidity and mortality, we really need good, large studies to inform us of the specifics of how to interpret the NP test results in order to optimize patient care.

 

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs:

 

 go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

 

if you would like to see the articles in this blog, please email me. 

 

please feel free to circulate this to others. also, if you send me their emails (gmodest@bidmc.harvard.edu), i can add them to the list