Understanding real‐world treatment patterns and clinical outcomes in AL amyloidosis patients diagnosed in Canada: A population‐based cohort study

Abstract Amyloid light chain (AL) amyloidosis is a rare and chronic bone marrow disorder. Existing claims data can be used to help understand the real‐world treatment patterns and outcomes of this patient population. Various population‐based administrative databases in Alberta, Canada were queried from 2010 to mid‐2019 to identify cases of AL amyloidosis. Baseline patient and disease characteristics, sequencing of pharmacologic therapies, overall survival, and healthcare resource utilization were evaluated. A total of 215 individuals with AL amyloidosis were included. Among patients diagnosed between 2012 and 2019, 149 (85.1%) initiated first‐line, 67 (38.3%) initiated second‐line, 22 (12.6%) initiated third‐line, and 11 (6.3%) initiated fourth‐line systemic therapy. In the first‐line setting, 99/149 (66.4%) received bortezomib, cyclophosphamide, and dexamethasone (CyBorD) and 21/149 (14.1%) received another bortezomib‐based regimen. Survival from time of diagnosis improved over time, with a median overall survival of 25.8 months (95% CI: 9.8, 57.1) for individuals diagnosed in 2010–2011 versus 52.1 months (95% CI: 25.6, NA) for those diagnosed in 2012–2019. Despite this improvement, the proportion of individuals diagnosed in 2012–2019 who survived beyond five‐years remained low (5‐year survival: 48.4%; 95% CI: 40.9, 57.2) which highlights an unmet need for more efficacious therapies.


INTRODUCTION
Amyloidosis results from the inability of certain proteins to retain a stable structure, which leads them to become amyloid fibrils in various tissues, including the heart, kidney, and liver [1,2]. Over 25 different proteins have been described as amyloidogenic precursors [3]. Amyloid light chain (AL) amyloidosis is the most severe and common type of amyloidosis and is characterized by fibrils composed of a monoclonal immunoglobin light chain [4,5]. AL amyloidosis typically manifests as a systemic disease, with local confinement of the disease to a single organ considered rare [1]. AL amyloidosis is related to multiple myeloma, another condition characterized by abnormal production of antibody-producing cells [6]. There is substantial overlap between these two conditions, and patients are often diagnosed with both [6].
The main distinguishing factor between the two conditions is the type of cells involved; in AL amyloidosis, light chains are of primary concern, while in multiple myeloma, growth of abnormal cells in bone marrow are of primary concern [1,2].
Real-world evidence pertaining to AL amyloidosis is limited, particularly in Canada. We previously described the characteristics and outcomes of 34 individuals diagnosed with AL Amyloidosis who were treated with CyBorD at a single center in Alberta, Canada [11,12].
The purpose of this investigation was to build upon this prior work by leveraging population-based administrative data to describe the characteristics, treatment patterns, and clinical outcomes of individuals diagnosed with AL amyloidosis in a Canadian real-world setting.

Study population
Population-based administrative databases from Alberta, Canada were queried to identify individuals who were diagnosed with AL amyloidosis in the province between January 1, 2010 and June 30, 2019 using a modified version of the algorithm implemented in Quock et al. [13,14].

Treatment and outcomes
Lines of pharmacologic therapy were identified using the Pharmaceutical Information Network database, which contains records from community pharmacies within the province. This database was queried for the following AL amyloidosis pharmacologic therapies: cyclophosphamide, bortezomib, bendamustine, carfilzomib, lenalidomide, daratumumab, melphalan, pomalidomide, thalidomide, ixazomib, dexamethasone, prednisone, doxycycline, and methylprednisolone [14,20]. We also examined hospitalization and ambulatory care records

Statistical analysis
Baseline patient and disease characteristics, sequencing of pharmacologic therapies, overall survival (OS), and time to next treatment or death (TTNT) were evaluated. Patients were followed from the date of diagnosis until death, the last interaction with the healthcare system,

Healthcare resource utilization
Healthcare resource utilization of individuals with AL amyloidosis was quantified and compared to that of the general population with respect to hospitalizations, ambulatory care encounters, and health practi- To account for a lack of independence due to matching, 95% confidence intervals were estimated using cluster-robust variance estimation. In addition to the comparison with the general population, we also compared the mean number of healthcare encounters among AL amyloidosis who had concurrent multiple myeloma with those who did not have concurrent multiple myeloma. In these analyses, healthcare resource utilization was examined from the time of initial diagnosis with AL amyloidosis until the end of follow-up.

Ethics statement
The Health Research Ethics Board of Alberta Cancer Committee approved this study (HREBA.CC-20-0481).

Baseline characteristics
A total of 6,238,529 unique individuals were queried, of which 622 were initially flagged as having a potential AL amyloidosis diagnosis.
After chart review, 215 individuals were confirmed to have AL amyloi-

Patient outcomes
Median OS was 39.

Healthcare resource utilization
The average number of healthcare encounters within the observed follow-up period for AL amyloidosis patients was significantly higher

DISCUSSION
Herein, we describe the treatment patterns and outcomes of a large, real-world, Canadian cohort of AL amyloidosis patients. Between 2012 and 2019, 85% of individuals initiated some form of pharmacologic therapy, which was most commonly CyBorD or another bortezomibbased regimen. Considerable attrition between lines was observed.
As highlighted in previous real-world studies, survival of AL amyloidosis has improved overtime, which is likely attributable to improved response rates achieved by CyBorD [9,11,17,21]. Despite this improvement, our study suggests that long-term OS remains poor, with only one in two patients surviving beyond 5-years in the modern treatment era. These findings emphasize the unmet need for additional, more efficacious therapeutic options in this patient population.
In addition, the healthcare resource utilization of individuals with AL amyloidosis was almost twice that of age-sex matched members of the general population, which suggests considerable disease burden.

TA B L E 5
Comparison of healthcare resource utilization (mean events per patient) among individuals diagnosed with AL amyloidosis in Alberta, Canada between 2010 and mid-2019 with age-sex matched members of the general population and between individuals with and without concurrent multiple myeloma Abbreviations: CI, confidence interval; MM, multiple myeloma; SMD, standardized mean difference. a Cluster robust variance estimation is used to account for bias in the estimation of the standard error due to clustering within the data caused by matching.
Ignoring this lack of independence tends to result in the underestimation of the standard error (i.e., the 95% CIs tend to be too narrow).
Despite having a much shorter lifespan, the healthcare resource utilization of individuals with AL amyloidosis who had concurrent multiple myeloma was comparable to that of those who did not. These findings suggest that the rate of healthcare resource utilization may be particularly high for those with concurrent multiple myeloma since they  [24]. With respect to healthcare resource utilization, other investigations have similarly observed a high burden of disease with AL amyloidosis [22,23]. For example, one investigation found that in the first year following relapse, average resource utilization per patient per month was 0.14 emergency room visits, 0.16 inpatient admissions, and 8.2 days per stay for inpatient admissions [23].
There are several strengths of this investigation. no diagnosis are prone to immortal-time bias due to the definition of concurrent multiple myeloma. However, such bias would attenuate the association by giving an artificial survival advantage to the concurrent multiple myeloma cohort. Therefore, the estimates generated within our study are likely conservative in that they likely underestimate the true burden of concurrent multiple myeloma in this patient population.
In conclusion, this investigation provides population-based realworld evidence pertaining to the management and outcomes of AL amyloidosis patients in Canada. CyBorD and other bortezomib-based regimens were found to be the primary frontline treatment modality used since 2012. High rates of attrition were observed between lines, with less than half of patients receiving two lines of therapy. Despite meaningful improvements in mortality over time, long-term survival remained poor, and the healthcare resource utilization was high, which highlights an unmet need and a high disease burden in this patient population.

ACKNOWLEDGMENTS
This study was funded by Janssen Inc.

DATA AVAILABILITY STATEMENT
Data from this investigation are not available for sharing due to data privacy legislation.