Changes in cardiac biomarkers with bortezomib treatment in patients with advanced cardiac amyloidosis

To the Editor

Therapy of light chain (AL) amyloidosis utilizes anti-plasma cell chemotherapies which target the underlying amyloidogenic plasma cell clone but have little effect on pre-formed/organ amyloid. Organ response thus lags hematologic response; those with advanced stage disease in particular have a high chance of dying early after diagnosis. In recent years, bortezomib has moved rapidly to the frontline of anti-amyloid therapies based on its ability to induce deep, rapid hematologic responses [1, 2]. Because published prospective bortezomib data was restricted to patients with limited cardiac involvement [2], it is unclear how bortezomib impacts those with advanced cardiac involvement. Bortezomib has been associated with heart failure, arrhythmias and sudden cardiac death with pre-clinical studies show that bortezomib impairs cardiac function with ultrastructural cardiomyocytes abnormalities, decreased mitochondrial ATP synthesis and decreased cardiomyocyte contractility [3]. We hypothesized that bortezomib causes fluctuations in cardiac biomarkers, particularly in advanced AL. We assessed cardiac biomarkers before and after bortezomib exposure, changes in relation to disease stage and therapy exposure time and tested if there was a prognostic significance to biomarker trends among patients with AL amyloidosis treated with bortezomib-based therapies.

Sixty five consecutive patients with proven AL amyloidosis who received bortezomib (either with dexamethasone alone, or combined with cyclophosphamide and dexamethasone) during July 2008 - December 2014 were identified. Based on the year of treatment, patients received both intravenous and subcutaneous bortezomib. Patients were staged using the 2012 staging system. A stage could not be calculated in 17 patients (13- baseline BNP and/or troponin I was obtained instead of NT-proBNP or TnT, 4- no baseline biomarkers obtained). Biomarkers were collected every 1-2 cycles of treatment and trended at baseline (denoted with B), the highest values in the first 3 cycles post-bortezomib (C_1-3), between cycles 4-6 post-bortezomib (C_4-6), and at last bortezomib therapy (L). A creatinine level at the same time point as the biomarkers was collected. Statistical analysis was performed using SAS v 9.3 (Cary, NC).

The median age at diagnosis was 67 (39-87) years with 52% females. Stage was I in 12 (18%), II in 20 (31%), III in 9 (14%), IV in 7 (11%) and missing in 17 (26%). Cardiac involvement was seen in 40 (61%) with a median baseline NT-proBNP of 996.5 (range, 42-41,820) pg/ml and TnT of 0.02 (range, <0.01-0.36) ng/ml. Renal involvement was seen in 39 (60%) and liver involvement in 16 (25%); 12 (18%) had ≥4 organs involved. The median difference in involved versus uninvolved free light chains (dFLC) was 18.2 (0.54-6098.3) mg/dl, with a lambda clone in 72%. The median duration of bortezomib therapy was 5 (0.5-12.1) months. Trends during the aforementioned times, available in a subset, are shown in the Table. Patients with stage I/II had no significant fluctuations in cardiac biomarkers, whereas these changes were significantly higher in those with stage III/IV particularly between the time points, B and C_1-3. Renal function was also trended; while median creatinine level was mildly higher in the stage III/IV group compared to stage I/II, 1.3 mg/dl versus 0.9 mg/dl respectively (p 0.03) and increased to 1.6 mg/dl versus 0.8 mg/dl at C_1-3, no significant changes in median creatinine level occurred from baseline to the first 3 months of treatment. With a median follow up of 21(3-142) months, 12 deaths were seen with a 1-year mortality after the start of therapy of 20%. Out of these, 2 had stage I/II, 5 patients had stage III/IV and 5 patients had missing stage. All 5 patients with stage III/IV disease died within 3 months of starting therapy. For the 23 patients for whom the change from B to C_1-3 in NT-proBNP was evaluable, for every 5000 increase in NT-proBNP the relative risk of death was 1.32 (95% CI 1.03-1.70; p = 0.03).

Table.

Trends in cardiac biomarkers at various treatment stages

Stage I/II (N=32)					Stage III/IV (N=16)
NT-proBNP
	N	Median change	Range of change	p-value		N	Median change	Range of change	p-value
B-C1-3	10	305	−1933, 704	0.6	B-C1-3	12	4988	−4077, 43746	0.02
C1-3-C4-6	5	−205	−618, 104	0.2	C1-3-C4-6	4	−8529	−15849, 40071	0.9
C4-6-L	7	0	−704, 519	1.0	C4-6-L	6	−2059	−18152, 0	0.06
B-L	16	66	−1987, 13332	0.7	B-L	11	3194	−14470, 43746	0.3
Troponin T
B-C1-3	7	0	−0.03, 0.06	0.6	B-C1-3	10	0.04	0.005, 0.75	0.002
C1-3-C4-6	4	0	−0.05, 0	0.5	C1-3-C4-6	5	−0.03	−0.06, 0.02	0.2
C4-6-L	6	0	−0.038, 0	1.0	C4-6-L	4	−0.02	−0.1, 0.35	1.0
B-L	15	0	−0.04, 0.127	0.1	B-L	7	0.07	−0.005, 0.75	0.03
Creatinine
B-C1-3	26	−0.03	−7.99, 0.80	0.07	B-C1-3	16	0.03	−0.89, 1.47	0.5
C1-3-C4-6	26	0	−0.37, 1.07	0.5	C1-3-C4-6	10	−0.04	−0.43, 2.89	0.9
C4-6-L	30	0.07	−0.64,1.39	0.01	C4-6-L	10	−0.05	−0.36, 0.33	0.9
B-L	29	−0.01	−7.89, 2.33	0.9	B-L	13	−0.08	−0.63, 2.07	0.8

We make the following observations: 1) There is significant increase in NT-proBNP and TnT in patients with Stage III/IV AL amyloidosis within the first 3 months of therapy; 2) These increases stabilize and improve during 4-6 months after bortezomib and 3) increases in NT-proBNP of >5000 pg/ml are associated with death within first year after diagnosis. There is scant data regarding cardiac safety of bortezomib among patients with advanced AL as prospective bortezomib data restricted patients with advanced cardiac involvement. Dubrey, et al. evaluated NT-proBNP after bortezomib (without dexamethasone) therapy in those with NYHA classification I/II and showed mild early elevations in cardiac biomarkers [6]. Notably this study did not have patients with advanced cardiac involvement; median NT-proBNP was only 384 pg/mL. Similar initial worsening in cardiac biomarkers has also been reported with immunomodulatory drugs such as thalidomide, lenalidomide and pomalidomide sometimes resulting in therapy discontinuation despite a FLC response [5]. We measured creatinine to assess if renal dysfunction would explain the increases in biomarkers, and found that creatinine levels were stable over this period. Although creatinine between the 2 groups was higher in stage III/IV than stage I/II, the levels were not at values that explain the significant increases in cardiac biomarkers between baseline and C_1-3. Further, creatinine changes between treatment phases were minimal, and thus unlikely to explain the biomarker changes in the patients with Stage III/IV AL. Our analysis is limited by our small sample size and our inability to perform adjusted analyses. The increases in biomarkers seen in advanced cardiac amyloidosis may well be a measure of ongoing, worsening organ damage from pre-formed amyloid rather than toxicity from bortezomib therapy. While others have shown that the use of bortezomib does not improve outcomes of patients with advanced cardiac amyloidosis [6], our data provokes us to ask if bortezomib could be contributing to worsening cardiac biomarkers. In conclusion, we show that patients with advanced stage amyloidosis can develop significant increases in NT-proBNP and TnT in the first 3 months of bortezomib treatment despite a hematologic response. Patients with >5000 pg/mL increase in NT-proBNP from baseline after starting bortezomib had an increased risk of death. Thus it is essential to follow these biomarkers closely after starting treatment not only for the purpose of organ response but also to assess any biomarker worsening even as hematologic response is occurring.