Systematic literature review of evidence in amyloid light-chain amyloidosis

Four literature searches (clinical evidence, HRQoL evidence, economic evidence and resource use/indirect costs evidence) were conducted by an information specialist using a peer-reviewed search strategy. The Embase and MEDLINE databases and Cochrane Controlled Register of Trials (CENTRAL) were searched using the Ovid platform. The search strategies included a combination of controlled vocabulary (e.g., ‘amyloidosis’) and keywords (e.g., ‘light chain, amyloid’). Language and publication year were not restricted at the search stage. For the HRQoL, economic and resource utilization searches, study selection was not limited by line of therapy, interventions or comparators (i.e., studies including newly diagnosed or relapsed/refractory AL amyloidosis patients were included). The reference lists from all published, on-topic SLRs identified during the screening process were reviewed as a supplementary search for relevant articles. Gray literature searches were also conducted to supplement the evidence searches.

The clinical evidence search of this SLR is registered with PROSPERO international prospective register of systematic reviews (protocol #CRD42020175412) and is available at: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42020175412.

The clinical evidence literature search focused on randomized controlled trials (RCTs) and observational studies of pharmacotherapies used as first-line treatment for adults with newly diagnosed AL amyloidosis (Supplementary Table 1). Observational studies were included because a paucity of RCTs in this rare disease space was anticipated.

Studies published prior to the publication date of the consensus opinion for organ involvement and response by the 10th International Symposium in Amyloid and Amyloidosis [27] and studies of nonpharmacological treatments (e.g., ASCT) were excluded. Of note, the clinical evidence searches permitted studies of patients at any stage of the disease.

The HRQoL evidence literature search focused on assessments of quality of life or algorithms used to derive utility values for adult patients with AL amyloidosis (Supplementary Table 2). During title and abstract screening, studies reporting any type of HRQoL data were included.

Studies were excluded at the full-text screening phase if they did not report any of the following commonly used HRQoL assessment instruments: European Quality of Life Five Dimension (EQ-5D), European Organization for the Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30), Short Form 36-Item Health Survey (SF-36), Short Form Six-Dimension Health Survey (SF-6D) and the Health Utilities Index.

The economic evidence literature search captured cost–effectiveness analyses (CEAs) with a specific focus on cost–utility analyses (CUAs; i.e., model outcomes reported as cost per quality-adjusted life year [QALY]) for adult patients with AL amyloidosis (Supplementary Table 3).

The resource use/indirect costs evidence literature search aimed to capture direct and indirect burden of illness costs for adult patients undergoing treatment for AL amyloidosis (Supplementary Table 4).

To reflect more relevant and current costs for treatment of AL amyloidosis, publications prior to 2010 (i.e., older than the last ∼10 years) were excluded.

Full search strategies for the four types of evidence are presented in Supplementary Tables 5–9. After the removal of duplicate citations, titles and abstracts were reviewed by two independent reviewers for study eligibility according to prespecified inclusion and exclusion criteria established using the population, intervention, comparators, outcomes and study design framework. Non-English-language publications were excluded during title and abstract screening, provided that the publication language was identified.

Study screening was performed using the systematic review software DistillerSR (Evidence Partners, Ottawa, Canada). Studies that met the inclusion criteria and those that did not provide enough information for exclusion were further reviewed at the full-text screening phase. Full-text articles were reviewed by two independent reviewers. During both screening phases, any discrepancies between the two reviewers were resolved by consensus or by a third reviewer.

Data extraction was performed for studies meeting all inclusion criteria using a standardized spreadsheet to capture all relevant information. Information was extracted by one reviewer and validated by a second reviewer; a third reviewer was consulted to resolve discrepancies as necessary.

Quality assessments of studies from the clinical evidence, HRQoL evidence and economic evidence literature searches were performed independently by two reviewers and conducted using the appropriate questionnaires [28–32]; quality assessments for included studies from the resource use/indirect costs literature search were not conducted. The evaluations were compared in order to ensure consensus.

Given the limited information available in abstracts, quality assessments were only performed for studies with an associated full-text publication.

The clinical evidence literature search identified five unique RCTs (Table 1) and 54 unique observational studies (Table 2) evaluating the efficacy and safety of several first-line pharmacological treatments for adults with AL amyloidosis (Figure 1A).

Four of the five included RCTs were phase III trials [12,14,15,34–40]; none of the included RCTs were placebo-controlled trials. One RCT (NCT01998503) investigated the addition of bortezomib and dexamethasone (Vd) induction treatment prior to high-dose melphalan and ASCT [12], whereas the remaining four RCTs compared chemotherapy regimens (VCd vs VCd with doxycycline [13]; melphalan and dexamethasone [MDex] vs bortezomib, melphalan and dexamethasone [VMDex] [14,34]; VCd vs D-VCd [15,35–40]). Across the studies, where reported, the first patient was enrolled between 2009 and 2018, and the last patient between 2012 and 2019. Median follow-up at the time of publication ranged from 6.1 to 50 months (Table 1). Four studies were multicenter (two were international) and one was a single-center study.

Quality assessment could be completed for only two trials, as they had associated full-text publications [12,14]. Details of random sequence generation ranged from unclear to having a low risk of bias, based on the available detail provided. Both trials had a high risk of bias related to allocation concealment and blinding of patients and care providers, either due to an open-label trial design [14] or as a result of patients in one treatment arm receiving additional concomitant medications compared with the other treatment arm [12]. Of note, blinding of outcome assessors among these trials was deemed unclear or had a low risk of bias. Selective reporting and incomplete outcomes were deemed to have a low risk of bias for both trials. Overall, the quality of reporting varied between low, inconclusive and high, depending on the factors that were assessed. A summary of the assessment of risk of bias of the two RCTs is presented in Supplementary Table 10.

Among the included observational studies, there were 12 prospective studies, 39 retrospective studies and three studies that did not explicitly report the study design [41–43]. 20 studies were described as comparative, whereas 34 studies were described as single arm. However, three comparative studies each contained only one arm that was relevant per the study selection criteria; thus these were treated as single-arm studies in this SLR. Bortezomib-based therapy (e.g., Vd and VCd) and MDex were the most commonly investigated chemotherapies. Nine studies investigated chemotherapy regimens as an induction treatment prior to ASCT [42–51]; three of these had patient populations that were initially ASCT ineligible and investigated the feasibility of ASCT after pharmacological treatment [42,45,49]. The majority of observational studies reported the number of patients who received subsequent therapies; most of these therapies were chemotherapy regimens, although patients undergoing subsequent ASCT and heart transplant were also reported.

Across the observational studies, were reported, patient enrollment began between 2001 and 2018 and ended between 2006 and 2019 (Table 2); 43 studies were single-center and eight multicenter (four were international), while the remaining three did not include this information.

Quality assessment was completed for each comparative observational study with an associated full-text publication (16 of the 19 comparative observational studies). In total, one study was assessed using the criteria for case–control studies and 15 were assessed using the criteria for cohort studies. Overall, 15 of the 16 assessed comparative observational studies were deemed to be of high quality. The design of the remaining study precluded it from proper quality assessment [47]. Because the study involved induction treatment prior to ASCT, only the data pre-ASCT underwent quality assessment. Follow-up data were excluded from the quality assessment, thereby resulting in a lower quality score. All assessed studies reported high-quality patient selection, and the design of most of the studies allowed for suitable treatment comparison. Some variation was noted in the quality of outcome reporting due to differences in the length of follow-up and the number of patients lost to follow-up. Summaries of the risk of bias assessments for qualifying comparative observational studies are presented in Supplementary Tables 11 (case-control studies) & 12 (cohort studies).

The key clinical outcomes that were identified included hematologic response, organ response, time to hematologic response, overall survival, PFS and safety, which comprised dose reductions due to adverse events (AEs), withdrawal due to AEs, and incidence of the following AEs: infection, neuropathy, cytopenia, organ failure, syncope, hypokalemia, peripheral edema and diarrhea (Figure 2).

The HRQoL evidence literature search identified 12 relevant studies for which data extraction was performed (Figure 1B); all were published between 2004 and 2020 and were mostly from the USA and Italy (Table 3). The studies reporting HRQoL consisted of three RCTs [14,38,52], two of which had also been captured through the clinical evidence search [14,38], and nine observational studies. Five studies reported on newly diagnosed or treatment-naive patient populations [14,38,52–54]. Only five studies reported HRQoL outcomes as a result of an intervention, including high-dose melphalan with stem cell transplant [55,56], nutritional counseling [52] and bortezomib-based chemotherapy (VCd vs D-VCd [38] and VMDex vs MDex [14]).

Relevant HRQoL outcomes included direct utility measures and HRQoL survey scores with potential to be mapped to utility values (Figure 3). None of the included studies reported utility values or mapping algorithms. However, all studies reported HRQoL survey scores with potential for mapping to utility values. The vast majority of these studies reported scores obtained from the SF-36, and three studies reported scores from the EORTC QLQ-C30 [14,38,54]. One study reported scores from the EQ-5D-5L visual analog scale [38]. No studies reported Health Utilities Index outcomes. Most studies reported HRQoL survey scores at a single time point. In addition to the outcomes of interest, one study reported the use of a non-cancer-specific global assessment of functioning (on a scale of 0 to 100, where a higher score indicated better functioning) and the Patient Global Impression of Severity scale [11].

Quality assessments demonstrated that patient populations were representative of the general AL amyloidosis patient population. Although the number of patients lost to follow-up varied, most studies had high response rates (100% in seven of the included studies and 77.6–95% in four studies) to the particular HRQoL instrument described in the study. All studies reported: a patient population likely to match the population of interest; the use of a preference-based instrument; and change in HRQoL measures reported directly by the patient. Seven of the 12 studies included satisfied the criterion of valuation of changes in patients’ HRQoL measures being assessed against that of the general population. As all studies reported values for the subscales of the HRQoL instrument, they could not be assessed for the technique used to value health states. A summary of the risk of bias assessments for the HRQoL studies is presented in Supplementary Table 13.

Overall, 39 studies reporting economic evidence were identified for full-text review. No CEAs, CUAs or CUA assessments relevant to AL amyloidosis were identified (Figure 1C). One conference abstract discussed a targeted literature search for drug prices and incremental QALYs of newly approved orphan drugs from health technology assessment publications [57]. Although AL amyloidosis was among the rare diseases identified in the search, the abstract was ultimately excluded due to a lack of sufficient information. In total, 21 studies eligible for full-text review were excluded due to their patient populations. Of these 21 studies, 14 did not discuss AL amyloidosis or failed to report economic outcomes specific to AL amyloidosis, while the remaining seven studies described other forms of amyloidosis such as hereditary and transthyretin amyloidosis. One study discussed economic outcomes among patients with AL and transthyretin amyloidosis but was excluded as disaggregated outcomes were not reported specifically for AL amyloidosis [58]. Similarly, a briefing on a study evaluating ixazomib for the treatment of relapsed and/or refractory AL amyloidosis was excluded as it did not report any outcomes [59]. Two studies were excluded based on the reported outcome, as they did not report costs [60,61]. 15 studies were excluded due to study design (i.e., they were not CEAs or CUAs) [62–76]. Of these, 14 were either cost-of-illness studies or resource utilization studies and reported costs and resource utilization related to AL amyloidosis [62–66,68–76]; the remaining citation was a letter [67]. Of the 14 studies, four focused on patients with AL amyloidosis and reported outcomes such as the number of hospitalizations and length of stay, as well as costs related to treatment, inpatient care and outpatient care [63,68–70].

After full-text review, none of the publications identified by this search fulfilled the criteria for formal inclusion in the final review and thus, data extraction and quality assessment were not performed.

The resource use/indirect costs evidence literature search identified 16 unique studies (one relevant SLR and 15 observational studies) published between 2017 and 2020 (Figure 1D & Table 4). Ten studies reported outcomes among patients with AL amyloidosis, including the SLR conducted by Lin et al. [77]. Six studies reported outcomes among a cardiac amyloidosis population [71–73,75,76,78–81]. Of note, these studies were included as it was implied within the publications that their patient populations could include individuals with AL amyloidosis. Most studies were of USA origin and reported costs in US dollars.

Relevant outcomes for the resource use evidence literature search included usage and cost of various healthcare resources, including but not limited to hospitalization, admission to the intensive care unit (ICU) or emergency department, inpatient care, outpatient care, readmission and treatment (Figure 4). Outcomes related to hospitalization and ICU stay were most commonly reported among the included studies and were reported as the number of stays and length of stay. Outcomes related to emergency department use, outpatient care and readmission were less commonly reported across the included studies.

Descriptions of costs reported across the studies varied. Costs related to hospitalization and inpatient care were reported in eight studies [68–73,75–77], and three studies reported costs related to both emergency and outpatient care [68–70]. Three studies reported various costs related to the treatment of amyloidosis, including chemotherapy, ASCT, organ transplant and various other procedures [68,71,77]. One study reported costs related to hospital readmission [71]. The 2017 study by Lin et al., an SLR of evidence related to AL amyloidosis, reported the costs associated with treating anemia, neutropenia and thrombocytopenia, all of which are common AEs resulting from chemotherapy treatment for AL amyloidosis [77]. These AE costs were reported for the general cancer population, but they were deemed applicable to the AL amyloidosis patient population as well [77]. No articles reporting indirect costs were identified in this search.