TO THE EDITOR:
Systemic amyloidoses are disorders caused by amyloid deposits in various tissues, leading to progressive organ failure and death. For any amyloidosis type, heart involvement is fundamental for prognostication. In light chain (AL) amyloidosis, the heart is involved in 70% to 80% of patients.1-4 For age-related transthyretin amyloidosis (wild-type ATTR [ATTRwt]), virtually all patients have cardiac involvement,5 whereas the majority of hereditary ATTR (also mutant ATTR [ATTRmt]) have cardiac involvement.6
Soluble cardiac biomarkers are invaluable for decision making in AL amyloidosis, and data are emerging about their importance in ATTR amyloidosis.5,7 With a changing landscape of cardiac biomarker assays, risk models may appear inaccessible. However, there are opportunities for harmonization. We briefly discuss the evolution of cardiac biomarker use in systemic amyloidoses and provide a conversion tool between cardiac troponin T (cTnT) and high-sensitivity cardiac troponin T (hs-cTnT) to allow a bridging pathway between past data and future developments.
Cardiac troponin I (cTnI) and cTnT are sensitive and specific for myocardial injury. Both cardiac troponin assays were initially assessed 15 years ago for their prognostic role in AL amyloidosis.8 In a pioneering work, cTnT had superior prognostic discrimination over cTnI in patients with AL amyloidosis using conventional, non–high-sensitivity assays. Moreover, multiple cTnI assays existed, whereas there was only 1 cTnT assay, allowing for greater standardization and reproducibility using cTnT. As a result, cTnT had been the main assay reported in the amyloidosis literature. hs-cTnT assay (also known as fifth-generation cTnT assay) was subsequently developed and has replaced cTnT in many laboratories because it has a fourfold to fivefold increased sensitivity over the fourth-generation cTnT assay. Correlation between the hs-cTnT assay and the fourth-generation cTnT assay at low concentrations is poor.9,10 In AL amyloidosis, the correlation between cTnT and hs-cTnT assays was excellent (R = 0.97) only for hs-cTnT >77 ng/L.10 When the hscTnT assay became available, it was shown by 2 separate groups to be positively correlated with severity of cardiac involvement and serves as an independent prognostic marker. hs-cTnT at 50 ng/L cutoff best predicted for overall survival in 1 study,11 whereas a cutoff of 77 ng/L was suggested by another study.12
Natriuretic peptides are secreted from myocardial cells in response to volume expansion and increased wall strain. The prohormone pro–brain natriuretic peptide (proBNP) is cleaved into the active hormone BNP and biologically inactive N-terminal proBNP (NT-proBNP) fragment. NT-proBNP was initially explored in AL amyloidosis and was proven to be a powerful biomarker for detection of cardiac involvement and survival prediction.13 NT-proBNP is important in assessing response to therapy.14,15 BNP was less investigated in AL amyloidosis, in part due to the multitude of available assays and lack of stability on stored samples.
In old series of AL amyloidosis, overt heart failure was an ominous sign, with a median survival of <8 months.16 With a need to better risk-stratify heart involvement, soluble cardiac biomarkers have emerged as the leading tool to predict outcome and aid in treatment choice. There are several risk models,2,17-19 all based on soluble cTnT and NT-proBNP (Table 1).
The first universally accepted model was the Mayo 2004, where patients were separated into 3 prognostic groups (Table 1).19,20 As part of this original study, investigators ran cTnI simultaneously, allowing for prognostication if cTnI was the troponin assay available, although it was less predictive. On parallel measure of hs-cTnT and cTnT, it was determined that a threshold of hs-cTnT of 54 ng/L was comparable to the 0.035 μg/L cTnT used in the Mayo 2004 model,10 enabling the use of the original model with the hs-cTnT assay. We reanalyzed the same data set using a quartic formula (provided in supplemental Data, available on the Blood Web site; JMP software, SAS, Cary, NC) to derive more accurate conversion of hs-cTnT for each of the existing thresholds, as shown in Table 2. This study was approved by the institutional review board. Using this extrapolation, 50 ng/L is found to be comparable to the 0.035 μg/L threshold of the Mayo 2004 model. A recent work by the Boston group has provided a conversion between NT-proBNP and BNP using a cohort of patients with simultaneous measurements of BNP, NT-proBNP, and cTnI. A BNP threshold of >81 ng/L best correlated with the Mayo 2004 model 332 ng/L NT-proBNP threshold.21
The European 2015 model was built on the Mayo 2004 model and subclassifies stage III patients into higher- and lower-risk categories by incorporating a second NT-proBNP threshold of 8500 ng/L2 (Table 1).The Mayo 2012 system uses different thresholds for NT-proBNP and cTnT, and it incorporates serum free light chain level as additional risk factor.18 A BNP of 400 ng/L was offered as an equivalent measure to NT-proBNP at an 1800-ng/L cut point. Table 2 provides the extrapolated hs-cTnT threshold value of 41 ng/L (rounded value 40 ng/L), comparable to the cTnT value of 0.025 μg/L in that model. Survival curves based on the extrapolated hs-cTnT cutoff for the Mayo 2004 and the European 2015 models are provided in the supplemental Data, whereas for Mayo 2012 was reported separately.22
Autologous stem cell transplant (ASCT) is an important therapy for AL amyloidosis. Given its toxicities, it can be offered only to selected patients. In our practice, approximately one-third of the referral AL population undergoes this procedure.3 We have demonstrated that baseline cardiac biomarkers are predictive for early death following ASCT. Proposed cardiac biomarker-based exclusion criteria for ASCT are ≥0.06 μg/L for cTnT and >5000 ng/L for NT-proBNP.23,24 The cTnT has withstood the test of time, but the NT-proBNP has less predictive power and is less often used as exclusion criterion for ASCT. A threshold hs-cTnT of 73 ng/L (rounded value 75 ng/L) can substitute for the cTnT 0.06 μg/L (Table 2).
The literature employing cardiac biomarkers for staging in ATTR amyloidosis is less mature. Two models have been described for ATTR, 1 from the Mayo group, which was performed in ATTRwt only,5 and the other from the National Amyloidosis Centre (NAC), which included ATTRwt and ATTRmt patients (Table 1).7 The Mayo model used cut points of NT-proBNP 3000 ng/L and cTnT of 0.05 μg/L to derive a 3-stage model. The NAC system utilizes NT-proBNP 3000 ng/L and estimated glomerular filtration rate (eGFR) <45 mL/min per body surface area to define 3 alternative risk groups. The NAC authors tested the Mayo ATTR model using their patient cohort and found comparable performance to their model. However, in various subgroups (ATTRwt, V122I, non-V122I), there was a trend for better discrimination with the eGFR/NT-proBNP model than the pure cardiac biomarker model. A drawback of their analysis, however, is the hs-cTnT threshold used to test the Mayo model (which uses cTnT) was not specified. As shown in Table 2, a cTnT of 0.05 μg/L is comparable to an hs-cTnT of 65 pg/L.
We conclude that using conversion method, hs-cTnT levels can be harmonized with cTnT values. The mostly commonly used cTnT thresholds of 0.025, 0.035, 0.05, and 0.06 μg/L can be extrapolated into hs-cTnT thresholds of 40, 50, 65, and 75 pg/L, respectively. Current models should not be discarded or made inaccessible purely based on lack of access to a desired assay. There are multiple hs-cTnI assays, and it may be time to readdress the value of cTnI in this arena. There also are several alternative natriuretic peptide assays that may be helpful. Thus, it is likely time to reexplore this area more comprehensively, attempting to make approaches with soluble biomarkers more universally and easily accessible.
The online version of this article contains a data supplement.
Contribution: E.M. and A.D. conceived of and designed the study and wrote the first draft of the paper; A.D. and S.K.K. performed the statistical analysis; and M.A.G., M.G., O.F.A., and A.S.J. participated in the analysis and interpretation of data and critical revision of the manuscript.
Conflict-of-interest disclosure: S.K.K. received consultancy fees from Celgene, Millennium, Onyx, Janssen, and BMS and research funding from Celgene, Millennium, Novartis, Onyx AbbVie, Janssen, and BMS. M.A.G. received consultancy fees from Millenium and honoraria from Celgene, Millenium, Onyx, Novartis, Smith Kline, Prothena, Ionis. A.S.J. presently or has in the past consulted for most of the major cardiac biomarker diagnostic companies. A.D. received research funding from Celgene, Millennium, Pfizer, and Janssen and a travel grant from Pfizer. The remaining authors declare no competing financial interests.
Correspondence: Angela Dispenzieri, Division of Hematology, Mayo Clinic, 200 First St, SW, Rochester, MN 55905; e-mail: firstname.lastname@example.org.