Introduction
Amyloid light-chain (AL) amyloidosis is a clonal plasma cell disorder characterized by the deposition of fibrils, which are derived from monoclonal immunoglobulin light chains [1,2]. Amyloid fibrils are deposited in many organs, including kidney, heart, liver, and nerve, destructing organ structures and functions [3]. Especially, kidney involvement occurs in about 70% of patients with AL amyloidosis and results in progressive kidney dysfunction. The incidence rate of progression to end-stage kidney disease (ESKD) in patients with kidney amyloidosis ranged from 10% to 33% in previous studies [4–6]. It is one of the main determinants of morbidity and quality of life and also limits therapeutic options [7].
Over the past two decades, advancements in diagnostic tools and treatment options have significantly improved kidney outcomes in AL amyloidosis. In 2014, Palladini et al. [5] developed the criteria of kidney stage, response, and progression based on proteinuria and estimated glomerular filtration rate (eGFR) and validated these criteria as predictive factors of progression to ESKD. Moreover, several studies highlighted the importance of the early reduction in difference between involved and uninvolved free light chain (dFLC), which is one determinant of hematologic response, as a crucial factor in predicting kidney prognosis [8–10].
In this retrospective cohort study, we evaluated kidney outcomes of 138 patients with kidney amyloidosis who received recent standardized care for AL amyloidosis. The aim of our study was to identify predictors of kidney survival, focusing on hematologic response (both categorical criteria and dFLC changes) and kidney response, as defined by Palladini et al. [5]. Furthermore, we examined the temporal changes in eGFR and proteinuria in relation to the grade of hematologic response, which could provide insights into predicting kidney response after achieving hematologic response.
Methods
Study population
This retrospective study identified 315 patients diagnosed with AL amyloidosis at Samsung Medical Center between March 2011 and December 2019. A total of 165 participants (52.4%) had kidney involvement, defined as either biopsy-proven kidney amyloidosis or proteinuria over 0.5 g per day, predominantly albuminuria in the presence of histologically confirmed amyloidosis of other organs according to the 2005 International Society of Amyloidosis (ISA) criteria [3]. Among them, individuals were excluded based on the following criteria: concurrent malignancy other than multiple myeloma (n = 4), initiation of hemodialysis within 1 month after diagnosis (n = 9), failure to initiate chemotherapy within 3 months after diagnosis or loss to follow-up within 1 month after diagnosis (n = 14). Finally, a total of 138 participants were included (Supplementary Fig. 1, available online). Multiple myeloma was diagnosed based on the 2014 revised diagnostic criteria for multiple myeloma [11]. Out of the 138 participants, 41 (29.7%) were considered as having kidney amyloidosis based on proteinuria over 0.5 g/day with pathologic confirmation in other organs, while in 97 participants (70.3%), kidney biopsy confirmed the diagnosis.
Ethics considerations
Clinical investigations were conducted in accordance with the principles of the Declaration of Helsinki. This study was approved by the Institutional Review Board (IRB) of Samsung Medical Center (No. SMC 2023-07-058). The need for informed patient consent was waived by the IRB due to the retrospective design of the study.
Data collection and definitions
We collected baseline data at the time of diagnosis. Hematologic evaluation of the plasma cell clonality included bone marrow biopsy and determination of monoclonal light chain burden through electrophoresis, immunofixation, and serum free light chain concentrations measurement. The involvement of other organs, such as heart, kidney, liver, gastrointestinal tract, and nerve, was assessed based on the 2005 ISA criteria specific to each organ [3]. For kidney staging, we used the current criteria developed by Palladini et al. [5]. Kidney stage Ⅰ was defined as eGFR equal to or higher than 50 mL/min/1.73 m2 and proteinuria equal or less than 5 g/day, and stage Ⅱ was when eGFR was lower than 50 mL/min/1.73 m2 or when there was proteinuria exceeding 5 g/day. Lastly, stage Ⅲ was assigned for individuals with eGFR lower than 50 mL/min/1.73 m2 and proteinuria greater than 5 g/day. eGFR was calculated using the equation of chronic kidney disease-epidemiology collaboration equation. The 24-hour urine collection, routinely examined at the time of diagnosis, was used to quantify proteinuria. The urine protein-creatinine ratio was substituted for two patients where 24-hour urine collection was not available. For risk stratification for survival outcomes associated with cardiac involvement, revised Mayo staging system was used [10]. Score of 1 was assigned to each of the three prognostic markers for cardiac involvement: N-terminal natriuretic peptide type B, ≥1,800 pg/mL; troponin T, ≥0.025 ng/mL or troponin I, ≥0.1 ng/mL; and dFLC, ≥180 mg/L. As a result of summing up the scores, 0, 1, 2, and 3 were classified as stage I, Ⅱ, Ⅲ, and Ⅳ, respectively. A single neurologist assessed neurologic issues in all patients diagnosed with AL amyloidosis and performed nerve conduction study. Positive results in these studies were defined as nerve involvement. Gastrointestinal involvement was defined by biopsy-confirmed findings in patients with relevant symptoms.
Hematologic response was assessed at 3 and 6 months in accordance with the 2012 ISA criteria [2]. Given the known impact of deep hematologic response, particularly achieving at least a very good partial response (VGPR), on prognosis and organ response, hematologic response was divided into two groups: deep hematologic response (complete response/VGPR) and no deep hematologic response (partial response/no response/progression) [1,4]. The dFLC was calculated by the difference between involved and uninvolved free light chain. The percentage changes in dFLC at 3 and 6 months from baseline were assessed using continuous and categorical variables, with the latter being divided into two groups by 90% reduction. Kidney response and progression were evaluated at 6 months according to the criteria of Palladini et al. [5]. Kidney progression was defined as a decrease in eGFR of 25% or more, and kidney response was defined as a decrease in proteinuria of 30% or more or below 0.5 g/day in the absence of kidney progression. Individuals who did not belong to either kidney response or kidney progression were classified as stable disease. Thereafter, eGFR and proteinuria were evaluated every 6 months for 3 years.
Outcomes
The primary outcome was kidney survival, defined as progression to ESKD and undergoing dialysis for more than 3 months. We did not consider cases where temporary dialysis was performed for acute kidney injury treatment. The participants were followed up until initiation of dialysis, death, or the last clinical visit date. The secondary outcome was changes in eGFR and proteinuria with time according to deep hematologic response and kidney response for 3 years following treatment initiation.
Statistical analysis
Continuous variables were expressed as median with interquartile range (IQR), and categorical variables were expressed as numbers with percentages. Comparison between groups was performed using the Kruskal-Wallis test or Mann-Whitney U test for continuous variables and the chi-square test for categorical variables. Kaplan-Meier curves were plotted for the cumulative incidence, and statistical difference was analyzed using log-rank test. Because a substantial number of patients died before progressing to ESKD during treatment, we accounted for death before requiring dialysis as a competing risk for ESKD using fine and gray substitution hazard model. Cox regression analysis was performed to identify predictive factors of kidney survival. Multivariable analysis included variables with a p-value of <0.10 in univariable analysis, age and sex. The results were expressed by the hazard ratio (HR) with 95% confidence interval (CI). Generalized estimating equation was used to investigate the interaction of the time and deep hematologic response or kidney response in eGFR and proteinuria. Statistical analysis was performed using R Statistical Software version 4.3.1 (R Foundation for Statistical Computing) and IBM SPSS version 22 (IBM Corp.).
Results
Baseline characteristics and overall patients’ survival
The baseline clinical characteristics and laboratory data are presented in Table 1. The median age was 64 years (IQR, 56–70 years) and 72 participants (52.2%) were male. A total of 20 participants (14.5%) had diabetes mellitus (DM), 99 (71.7%) had heart involvement, and 11 (8.0%) had multiple myeloma concurrently. At baseline, the median dFLC was 162.2 mg/L (IQR, 63.1–482.8 mg/L), the median eGFR was 79 mL/min/1.73 m2 (IQR, 52–94 mL/min/1.73 m2), and the median proteinuria was 3.8 g/day (IQR, 2.0–5.9 g/day). According to the revised Mayo staging, 31 (22.5%), 27 (19.6%), 30 (31.2%), and 26 (26.1%) belonged to stage Ⅰ, Ⅱ, Ⅲ, and Ⅳ, respectively. In terms of kidney staging, 70 (50.7%) were in stage Ⅰ, 57 (41.3%) were in stage Ⅱ, and 11 (8.0%) were in stage Ⅲ.
The median follow-up duration was 46.4 months (IQR, 10.0–79.4 months). A total of 53 patients (38.4%) died, of which 47 (34.1%) died before requiring dialysis. The overall survival was different among the revised Mayo stages (p < 0.001). The median overall survival was 80 months in stage III and 21 months in stage IV, while not reached in stage I and II. Post hoc analysis revealed that the revised Mayo stage Ⅰ had significantly longer overall survival than stage Ⅱ (p = 0.003), Ⅲ (p < 0.001), and Ⅳ (p < 0.001) (Supplementary Fig. 2, available online).
Initial kidney staging and kidney survival
Out of the 138 patients, 17 (12.3%) progressed to ESKD. By kidney stage, five (7.1%), 10 (17.5%), and two patients (18.2%) initiated dialysis during the follow-up in stage Ⅰ, Ⅱ, and Ⅲ, respectively. The kidney survival varied across three kidney stages (p = 0.02) (Fig. 1). The cumulative incidence of ESKD at 5 years was 2.1% in kidney stage Ⅰ, 15.5% in stage Ⅱ, and 18.2% in stage Ⅲ, respectively. Alongside kidney staging, DM status, and hemoglobin levels showed an association with kidney survival in univariable analysis (Table 2). In multivariable analysis involving baseline characteristics, only kidney stage Ⅱ exhibited an increased risk of progression to ESKD compared to kidney stage Ⅰ (HR, 3.75; 95% CI, 1.38–10.15; p = 0.01).
Treatment responses within 6 months and probabilities of end-stage kidney disease
At 6 months, 47 out of 109 participants (43.1%) achieved kidney response. In terms of hematologic response, 59 of 122 patients (48.4%) achieved deep hematologic response at 3 months, while 68 of 111 patients (61.3%) achieved deep hematologic response at 6 months. The percentage changes in dFLC from baseline were –78.9% (–97.0% to –45.9%) at 3 months and –85.8% (–96.9% to –65.2%) at 6 months.
We analyzed the proportions of treatment responses at 6 months relative to the baseline kidney stages. The results indicated that the proportions of both kidney responses and deep hematologic responses were similar across the different baseline kidney stages (Supplementary Table 1, available online). Next, to assess the association between kidney response and hematologic responses, we compared hematologic response indicators between the two groups based on kidney response at 6 months (Supplementary Table 2, available online). There was no significant difference in the frequency of hematologic responses between the groups. However, we observed that the percentage reduction in dFLC at 6 months was significantly greater in the group with a kidney response (median, –90.7%; IQR, –98.8% to –80.0%) compared to the group without a kidney response (median, –83.3%; IQR, –94.5% to –59.9%; p = 0.02).
Among treatment response parameters, only kidney response independently predicted kidney survival in multivariable analysis (Table 3). Specifically, both stable disease (HR, 8.42; 95% CI, 1.71–41.35; p = 0.01) and kidney progression (HR, 7.36; 95% CI, 1.25–43.33; p = 0.03) were associated with an increased risk of ESKD compared to kidney response after adjusting for age, sex, DM status, hemoglobin levels, and initial kidney stage. dFLC percentage changes at 6 months were associated with kidney survival in univariable analysis, but this association was marginally significant after adjusting for initial kidney stages. Hematologic response at both 3 and 6 months did not predict progression to ESKD.
We categorized participants into four groups based on their hematologic response and kidney response at 6 months and then compared the probabilities of ESKD, as illustrated in Fig. 2. None of the participants who achieved both deep hematologic response and kidney response progressed to ESKD. Kidney survival was longer in patients with both deep hematologic response and kidney response than those with only hematologic response (p = 0.004) (Fig. 2A). Regarding overall survival, only participants who achieved kidney and deep hematologic response had significantly longer overall survival than those who did not achieve both (p = 0.01) (Fig. 2B).
First-line therapy and kidney outcomes
We divided patients into subgroups according to first-line therapy to evaluate the impact of treatments on kidney outcomes: autologous stem cell transplantation (ASCT), bortezomib-based, and other chemotherapies. A total of 41 participants (29.7%) underwent upfront ASCT with or without preceding induction chemotherapy. Except for those who underwent upfront ASCT, 54 (39.1%) were treated with bortezomib and 43 (31.2%) received other chemotherapies, including immunomodulatory drugs (thalidomide or lenalidomide), daratumumab, and conventional chemotherapy (cyclophosphamide, melphalan, steroid only, etc.), as the first-line therapy. Supplementary Table 3 (available online) shows baseline characteristics of patients according to the first-line therapy. Patients who underwent ASCT were younger than those who received chemotherapy, while those who were treated with bortezomib-based chemotherapy had a higher prevalence of heart involvement and a lower eGFR than other groups.
Four (9.8%), nine (16.7%), and four patients (9.3%) progressed to ESKD in treatment groups of ASCT, bortezomib-based, and other chemotherapies, respectively. In univariable analysis, types of first-line therapy, ASCT (HR, 1.05; 95% CI, 0.29–3.83; p = 0.94) and bortezomib-based chemotherapy (HR, 2.20; 95% CI, 0.72–6.69; p = 0.16) was not associated with ESKD risk when other chemotherapies was used as the reference category (Supplementary Table 4, available online). Then, we conducted subgroup analysis to evaluate the association between hematologic and kidney responses and ESKD risk in each ASCT and chemotherapy treatment group. Initial kidney stage Ⅱ and kidney response remained predictive of progression to ESKD, particularly in those who received chemotherapy, but not in those who underwent ASCT (p for interaction < 0.01) (Supplementary Table 5, available online).
Serial changes of estimated glomerular filtration rate and percentage changes in proteinuria according to treatment responses
We investigated the temporal changes in the eGFR and proteinuria over a period of 3 years based on hematologic response (Fig. 3). In the deep hematologic response group, eGFR exhibited a marginally significant decrease from baseline to 6 months (baseline vs. 6 months: 83.3 mL/min/1.73 m2 vs. 72.3 mL/min/1.73 m2; p = 0.07), followed by a subsequent stabilization (Fig. 3A). The percentage changes in proteinuria from baseline were –43.8% and –60.4% during the initial 6 months and the subsequent 6 months, respectively, showing a marginal difference (p = 0.09) (Fig. 3B). Afterward, there was a slight non-significant decrease in proteinuria levels. Conversely, the no deep hematologic response group showed no significant variations in the values of both eGFR and proteinuria assessed every 6 months. Overall, both eGFR (p = 0.38) and percentage changes in proteinuria (p = 0.56) did not exhibit significant differences between hematologic response groups throughout 3 years.
We also assessed whether the changes in the eGFR and proteinuria differed according to kidney response. There were significant interactions between time and kidney responses in both eGFR (p < 0.01) (Fig. 3C) and percentage changes in proteinuria (p = 0.02) (Fig. 3D). Specifically, eGFR decreased significantly from baseline to 6 months in patients with kidney progression (baseline vs. 6 months: 70.0 mL/min/1.73 m2 vs. 41.9 mL/min/1.73 m2; p < 0.01).
Discussion
In our longitudinal study involving 138 patients with kidney amyloidosis, the initial kidney stage emerged as an independent predictor of progression to ESKD. Notably, in stage Ⅱ, the risk of ESKD increased by 3.75 times compared to stage Ⅰ, emphasizing the critical importance of initiating appropriate hematological treatment before kidney involvement reaches a more severe stage. Our analysis has brought that kidney response, rather than hematologic response following chemotherapy, exhibited a substantial association with progression to ESKD. Of particular concern is the finding that, even among those achieving deep hematologic response at the 6-month mark, subsequent improvements in eGFR and proteinuria after 12 months were insignificant, and especially, a lack of kidney response had an association with an increased risk of ESKD. This observation implies that despite a favorable hematologic response, additional interventions may be required when the kidney response remains suboptimal. This is in line with recent arguments suggesting that patients who have not received sufficient organ responses need new treatment targeting the amyloid deposits [12,13].
The achievement of hematologic response is a crucial goal in the treatment of AL amyloidosis, aiming to impede the further accumulation of amyloid fibrils. Previous studies have demonstrated its association with overall survival [2,14], as well as organ responses, including kidney response [4,6] and kidney survival [4,5]. However, recent research indicates a shifting emphasis towards the importance of evaluating organ response, encompassing the heart, kidneys, liver, and more, as a key prognostic factor compared to hematologic response [15–18]. In line with these insights, our study consistently found that assessing kidney response holds more significance in predicting progression to ESKD than hematologic response alone. Unlike multiple myeloma, where morbidity and mortality primarily stem from plasma cell proliferation, the presence of organ dysfunction adversely affects the patient’s prognosis in AL amyloidosis [3]. Given this intricate pathophysiology, there is a growing need to expand current treatment surrogate endpoints in AL amyloidosis, which are highly dependent on hematologic response, to encompass organ responses as well [19,20].
The current standard for evaluating kidney response in AL amyloidosis involves categorizing it into three stages: progression, stable, and response [5]. Our study found that the HR for progression to ESKD was comparable between stable disease and kidney progression. In post hoc analysis, the risk of ESKD was not different between kidney progression and stable disease (p = 0.80, data not shown). This suggests that stable disease has limited discriminatory value compared to kidney progression, and kidney response serves as a more reliable predictor of kidney survival. There is a need for further exploration and confirmation of criteria related to kidney response, which is widely utilized in AL amyloidosis, particularly those associated with long-term kidney outcomes. Questions also remain regarding the timing of evaluating organ response. Notably, in the group with deep hematologic response, there was no significant reduction in proteinuria after 12 months, indicating that the 12-month mark posttreatment initiation may be an appropriate time point for kidney response evaluation. This observation aligns with a study conducted at the Mayo Clinic, which suggested a median time of around 11 months to reach kidney response [21].
As a determinant of hematologic response, dFLC changes have been identified as an independent predictor of overall patient survival [9,22] and kidney survival in patients with AL amyloidosis [5,8,9]. Especially, dFLC is preferred over levels of involved free light chain or free light chain ratio for assessment of light chain burden in individuals with kidney failure because it is less likely to be confounded by kidney function [23]. While it did not emerge as a predictor of kidney survival in our study, it showed an association with kidney response. Specifically, the percentage changes in dFLC at 3 months (no kidney response vs. kidney response: –72.6 vs. –89.3; p = 0.10) and 6 months (–83.3 vs. –90.7; p = 0.02) were more substantial in patients who achieved kidney response at 6 months compared to those who did not (Supplementary Table 2, available online).
This study has several limitations. Firstly, its retrospective design introduced a notable constraint, with slight variations in the timing of response assessments. Also, kidney responses at 6 months were unavailable for two patients due to missing laboratory results. Secondly, advancements in therapeutic strategies over the study period, such as increased utilization of bortezomib and ASCT since 2014, may have influenced patient outcomes [24]. This enhancement in patient survival allowed for a comprehensive assessment of kidney outcomes in many cases. However, analysis involving first-line therapy was limited by small numbers of patients in each treatment group and selection bias in that individuals who underwent ASCT were younger and had better organ function than those who received other treatments. Third, albumin-dominant proteinuria without pathologic confirmation, one inclusion criterion of the present study could be attributed to kidney disease unrelated to AL amyloidosis, such as diabetic kidney disease or glomerulonephritis. The changes in kidney function in these conditions may occur independently of hematologic response.
In conclusion, our study underscores the importance of kidney response, rather than post-chemotherapy hematologic response, in predicting the progression to ESKD. While a significant reduction in dFLC substantially increased likelihood of kidney response, discrepancies between dFLC changes and improvements in kidney parameters were noted in some cases. Additionally, even among patients exhibiting deep hematologic response at the 6-month mark, subsequent improvements in eGFR and proteinuria after 12 months were found to be insignificant. This highlights the imperative for new treatment strategies to improve the prognosis of patients with insufficient organ responses. Further research is warranted to establish specific criteria and optimal timings for organ responses.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement table 1
Print
