AL amyloidosis cardiac staging updated using BNP

The inability to perform gold-standard testing for cardiac involvement in immunoglobulin light chain amyloidosis (AL amyloidosis) remains a barrier in many centers as they try to rapidly identify and prognosticate the highest-risk patients. In this issue of Blood, Lilleness and colleagues sought to update the Mayo 2004 cardiac staging system by examining the utility of substituting the established N-terminal pro-brain natriuretic peptide (NT-proBNP) with the more widely available brain natriuretic peptide (BNP).¹  Importantly, the Boston University (BU) staging system defined in their article was built on a more contemporary cohort treated in an era in which proteasome inhibitors are now a mainstay of first-line therapy, and more stringent patient selection has optimized the use of high-dose chemotherapy and autologous stem cell transplantation.

The severity of cardiac involvement remains the most important prognostic factor in AL amyloidosis. Despite advances in therapy, many patients never live long enough to benefit from these improvements because of early cardiac death. More rapid determination of cardiac involvement helps identify those at highest risk, thus providing an opportunity to begin more aggressive supportive care strategies and even modify a given chemotherapeutic approach. Early work by the group at Mayo established the first broadly applicable tool to risk stratify patients. However, its dependence on NT-proBNP does limit its usefulness because many sites cannot run the test. Although BNP is more widely available, it cannot act directly as a substitute, hence the importance of the work presented by Lilleness et al.² 

In the Lilleness et al study, the first step was to use a derivation cohort to identify the BNP level that best correlated with the NT-proBNP level of 332 pg/mL established in the original Mayo criteria.³  Second, the cohort was used to determine a BNP cutoff indicative of cardiac involvement. Importantly, the authors defined that cutoff by incorporating modern diagnostics, including cardiac magnetic resonance imaging (MRI) and more detailed echocardiographic criteria now commonly used in practice but not captured in the 2005 consensus criteria.⁴  It reflects a broader and more clinically relevant spectrum of cardiac involvement known to occur in AL amyloidosis.

Acknowledging the relative interchangeability of cardiac troponin I (TnI) and troponin T (TnT) established in the original Mayo publication, the authors chose the TnI threshold of 0.1 ng/mL as a partner for the BNP vs NT-proBNP analysis. Examining a cohort of 250 consecutively evaluated patients that had both NT-proBNP and BNP levels drawn at diagnosis, the BNP threshold of 81 pg/mL was identified as a match for the NT-proBNP level of 332 pg/mL noted in the original Mayo 2004 criteria. Using the same derivation cohort, the authors then went on to define both the BNP and NT-proBNP thresholds that correlate with cardiac involvement. Interestingly, the same BNP cutoff of 81 pg/mL was predictive of light chain cardiac deposition. In contrast, an NT-proBNP cutoff of 288 pg/mL was determined to be indicative of cardiac involvement, which was slightly lower than the level used for stratification in the Mayo staging system. This may be related to the initial study being designed primarily around survival rather than cardiac involvement as well as being done in an era without widespread availability of cardiac MRI.⁴  Thus, both of these values may be useful as an initial screen at diagnosis to predict whether the heart might be involved and prompt more detailed workup when required.

The authors do acknowledge the variable renal metabolism of both of these biomarkers, with NT-proBNP being more dependent on the kidney for clearance.⁵  They hypothesized that BNP may in fact be superior in the setting of chronic kidney disease (CKD) stage III or greater at predicting cardiac involvement and that NT-proBNP may be more advantageous in CKD stage II or lower. In their analysis, the authors concluded that the latter was in fact true, with NT-proBNP being more sensitive in patients with minimal or no renal impairment. However, in those with impaired renal function, both biomarkers were equally valid albeit at different thresholds: 427 pg/mL for BNP and 2930 pg/mL for NT-proBNP. Although both tests seem to be reasonable screening strategies for cardiac involvement (on the basis of a patient’s renal function), one must consider this important nuance when applying the data in the clinic.

The derivation cohort was established by using consecutively diagnosed patients over a period of 6 months with a maximum possible follow-up of only 2 years. Thus, to establish the prognostic utility of the BU staging system, the authors examined a second complementary cohort with 10 years of follow-up. By using this larger data set of more than 1000 patients, the authors were able to demonstrate that the BU staging system nicely stratified patients on the basis of overall survival. Furthermore, consistent with recent publications that identify ultra-high-risk patients on the basis of markedly elevated NT-proBNP levels (>8500 pg/mL),⁶  an additional BNP threshold of 700 pg/mL was established in the BU staging system to define patients with advanced stage IIIb disease. It is remarkable that, although the BU staging system nicely differentiates the groups in a way similar to the Mayo staging system, the estimated survivals for each risk group are much longer (see table). Stage I patients would be expected to live for more than 10 years. Even in patients with advanced stage III disease, approximately 20% of them may also reach this 10-year milestone. Because the complementary cohort was more contemporary, the outcomes reflect the advances we are experiencing with modern methods of care for patients with AL amyloidosis. Although therapy for these patients continues to evolve, there is hope that if enough therapy is given and deep enough clonal responses are attained, durable remissions with organ response can be achieved, even in patients with the most advanced-stage disease.

The authors do acknowledge some important limitations of the study. The main limitation is the lack of data using BNP as a marker of cardiac response after treatment. The NT-proBNP criteria for cardiac response remain the best described and validated staging system.⁷  Thus, that system will continue to have an important role in management and follow-up, especially in the clinical trial setting. That being said, the data presented in the Lilleness et al article warrants future study to validate the proposed BU system prospectively and to examine the utility of BNP in response assessment.

Overall, the study by Lilleness et al is important because it addresses a very practical question that many laboratories face when they do not have the gold-standard test to identify high-risk patients with AL amyloidosis. It is a step toward validating the use of BNP in staging this disease. Moreover, it re-establishes the utility of these 2 biomarkers in a more contemporary cohort treated in the era of novel agents. Ultimately, it may be a pathway to more accessible testing modalities that help clinician with earlier diagnosis and better characterization of their patients.