Introduction

Multiple myeloma (MM) is a neoplastic disease associated with several symptoms. These symptoms may include bone pain, fatigue, impaired renal function, and anemia, which cause an impairment of health-related quality of life (HRQOL) [1]. This term refers to a multidimensional construct which represents the patient’s general perception of the effect of illness and treatment on physical, psychological, and social aspects of life [2]. However, HRQOL is only one out of a variety of other patient-reported outcomes (PRO) that are now frequently used in clinical research. PRO can be defined as any measure of patient’s health, coming directly from the patient without interpretation by a physician or anyone else [3].

Achieving a durable complete response while, at the same time, an adequate level of toxicity is the primary aim of therapies for MM [4]. Development of novel treatments in recent years has contributed to major advances in clinical outcomes. For example, in the last decade, overall survival of newly diagnosed patients has increased [5], and this was more evident in younger patients [6,7,8]. Recent studies have demonstrated the role of proteasome inhibitors (PIs) and immunomodulatory drugs (IMiDs), such as bortezomib, lenalidomide, and thalidomide, in improving time to progression (TTP) and progression-free survival (PFS) on MM patients [9,10,11]. Despite these major treatment advances, the disease remains broadly incurable. The only potentially curative option is allogeneic hematopoietic stem cell transplantation (allo HSCT), but this treatment is often limited to a small proportion of patients. Nevertheless, the natural history of the disease comprises sequential phases of responses, remissions and relapses, and the depth and duration of responses following each relapse usually decrease [4].

Relapsed/refractory multiple myeloma (RRMM) patients are a heterogeneous population, whose characteristics depend on the number and type of treatment used, and the type of relapse (early relapse, late relapse or multiple relapses). Basically, the following groups are considered: primary refractory, relapsed, and relapsed and refractory. Relapsed MM, that is the recurrence of the disease after a previous response, is defined based on laboratory criteria, including a ≥ 25% increase of the serum or urine monoclonal protein (M-protein) [12, 13]. Relapsed and refractory myeloma is defined as disease that is nonresponsive while on salvage therapy, or progresses within 60 days of last therapy in patients who have achieved minimal response (MR) or better at some previous point [14]. Primary refractory, instead, refers to disease which fails to achieve a MR with any therapy [12]. Selection of treatment strategies for these patients is challenging for a number of reasons, including the increased age and the often associated comorbid conditions [15]. Relapsed/refractory myeloma patients are also more symptomatic, more vulnerable to adverse events (AEs) and, therefore, more likely to incur a dose reduction or early discontinuation of therapy. Indeed, AEs and treatment discontinuation are associated with poorer outcomes and greater risk of death [16]. For example, patients who become refractory to PIs and IMiDs have poor survival, reporting an average life expectancy of 9 months from the time of failing first line PIs and IMiDs [17]. Balancing expected efficacy of therapy against potential toxicities and impact on patients’ HRQOL is therefore critical in this patient population. Also, the benefits of early palliative care should not be underestimated [18].

The majority of studies that have assessed HRQOL in MM patients have mainly focused on newly diagnosed patients enrolled in randomized controlled trials (RCT) [19] or have addressed specific methodological issues in the design and the reporting of HRQOL outcomes [20]. For example, Nielsen et al. have recently published a review looking into changes in EORTC QLQ-C30 [21] scores, in MM longitudinal studies [22].

Therefore, we performed a systematic review with the objective of quantifying the amount of evidence-based data available on HRQOL in RRMM patients published since 1990. In addition, we also examined the possible added value of RRMM patients’ perspectives in supporting clinical decision-making.

Materials and methods

Search strategy for identification of studies

We conducted a systematic literature search in PubMed/Medline and in Cochrane Library to identify studies published between January 1990 and May 2017. Additional publications were identified by hand-searching reference lists of relevant articles. We have also consulted with content experts in MM research to possibly identify additional eligible articles that we might have missed through electronic search. The starting point of the research was 1990 in order to best capture the studies focusing on new drug therapies for RRMM. The key search strategy used to retrieve eligible articles is reported in Appendix 1. The search strategy included all varieties of studies, excluding reviews and case reports. Only English language articles were considered and abstracts of identified articles were screened for inclusion. If a selected study had multiple publications, we considered them all in the data extraction process in order to maximize quality of information in our review.

Selection criteria

Types of participants

Studies with adult patients diagnosed with primary refractory, relapsed, and relapsed and refractory multiple myeloma were included. Studies focusing on newly diagnosed patients were excluded. No restriction was applied with regard to the number of patients enrolled.

Types of intervention

All studies conducted in RRMM patients receiving any kind of treatment were eligible for inclusion.

Types of studies

We included all type of studies (regardless of the design) conducted in RRMM patients incorporating a PRO measure. Studies including a heterogeneous sample of cancer sites were excluded. We excluded studies including both RRMM and newly diagnosed MM patients, in which the results were analyzed together [23, 24] and studies where it was not specified that the population was composed by RRMM [25, 26], despite the patients receiving different lines of treatment, due to the difficulty to attribute the findings to RRMM population. Conference abstracts and case reports were excluded due to their lack of necessary information to assess quality of PRO reporting.

Data extraction and type of information considered

Before beginning the literature search, a predefined coding schema for data extraction was developed. The data extraction form included various information on (1) basic trial demographics (i.e., year, journal, age of patients, overall study sample size, study location); (2) clinical characteristics (i.e., type of therapy, prior treatments, prior lines of therapy); (3) PRO design characteristics (i.e., patients included in the PRO analysis, PRO instrument used and timing of HRQOL assessment); (4) summary of findings (i.e., statistically and clinically meaningful PRO outcomes, clinical endpoints) and treatment recommendations by authors. This latter was based on how authors themselves summarized their findings and what treatments they recommended, based on traditional clinical and PRO outcomes. PRO information extracted was grouped into predefined categories depending on whether or not statistically or clinically meaningful differences between treatment arms (if more than one) or from baseline (if single arm trial) at any time point were found. If the difference was found only in a single scale, PRO outcomes were categorized as “symptoms only,” “functional scale only,” or “Global Health Status/QoL scale only,” depending on the single item or scale where the difference was registered. If the difference was found in more than a scale, the outcomes were categorized as “mixed outcome,” otherwise, if no difference were found, we categorized it as “study equivalence.” Each study selected was evaluated independently by two reviewers (FS and FE), and, when disagreements occurred with regard to the extracted information, the reviewers revisited the paper/s to reconcile any differences until consensus was achieved.

Results

The literature search yielded 1680 records, published until May 2017, that were screened for eligibility. We retrieved a total of 18 papers, published from June 2003 to March 2017, comprising of 11 studies. Figure 1 details the search strategy and the selection process of the articles included in this review, which were complied with Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines [27].

Fig. 1
figure 1

Schematic breakdown of literature search results of RRMM studies

Full size image

Overview of patient characteristics and study design

The overall number of patients enrolled was 4035, ranging from a sample size of 32 to 792 (63.6% of the studies enrolled more than 200 patients). Six studies were RCT and the median age of patients enrolled in all studies identified ranged from 30 to 91 years. All the studies included were multicenter and ten of them (90.9%) were international studies. Two of the most recent studies investigated the role of carfilzomib [28, 29] and pomalidomide [30,31,32], that is the most recently approved IMiDs for the treatment of myeloma patients [33]. Two studies (18.2%) focused on bortezomib therapy [34,35,36,37,38,39] two on thalidomide [40, 41], one (9.1%) on the most recently Food and Drug Administration (FDA) approved PIs, ixazomib [42], one on elotuzumab [43], one on lenalidomide [44], one both on lenalidomide and bortezomib [45], and one on ALCAR [46]. The number of prior lines of therapy ranged from one to 17. Full details are reported in Table 1.

Table 1 General study characteristics
Full size table

Overview of PRO instruments used

In all studies, except one [45], PRO were included as secondary endpoints. The EORTC QLQ-C30, used alone or in conjunction with its specific myeloma module [47], was the most frequently used instrument, being employed in ten studies (90.9%). The FACT/GOG-Ntx [48] was used in three studies (27.3%) and the FACIT-Fatigue scale [49] was used in two studies (18.2%).

In one study [30,31,32], authors employed five of the 15 EORTC QLQ-C30 scales (Global Health Status/QoL scale, Physical Functioning, Emotional Functioning, Fatigue and Pain), two of the four EORTC QLQ-MY20 scales (Disease Symptoms and Side effects of treatment) and EQ-5D Health Utility Index. Stewart AK et al. [28, 29] used six EORTC QLQ-C30 scales (Global Health Status scale/QoL scale, fatigue, pain, nausea/vomiting, physical functioning, and role functioning) and two EORTC QLQ-MY20 scales (disease symptoms and side effects of treatment), whereas remaining studies used all the domains of the instruments employed. Details are reported in Table 2.

Table 2 PRO characteristics
Full size table

Proteasome inhibitors and PRO

Six studies (54.5%) investigated the use of PIs (ixazomib, carfilzomib, and bortezomib) and their impact on HRQOL and overall survival (OS). In one study [37,38,39], bortezomib was associated with a median OS of 16 months (data on statistical significance are not reported), while it induced clinically meaningful improvements in 10 PRO scales. In another study, bortezomib was compared with dexamethasone [34,35,36] and it demonstrated improved survival, as OS for this group was 29.8 compared with 23.7 of dexamethasone. Also, this study showed that bortezomib was associated with higher AEs but better HRQOL, since ten scales resulted in better outcomes with respect to the dexamethasone group. Leleu et al. [45] observed that patients who completed the 6 month treatment with bortezomib worsened only in the Global Health Status/QoL scale, compared with those who discontinued the therapy prior to 6 months, who declined on eight scales. In Callander et al. [46], the addition of ALCAR to bortezomib, doxorubicin and dexamethasone did not demonstrate statistical significant differences in OS and PRO outcomes. In Moreau et al. [42], ixazomib was associated with longer progression-free survival (PFS), but AEs (any grades), including constipation, nausea and vomiting were generally reported at higher rates with this drug. PRO outcomes were similar between groups, with a trend for less fatigue and better physical and emotional functioning in the ixazomib regimen. In Stewart et al. [28, 29], median overall survival was not reached in either group, but resulted a trend in favor of carfilzomib group. Any grade non-hematologic AEs were reported at higher rates in the experimental arm. The same group reported better clinically meaningful outcomes in the EORTC QLQ-C30 Global Health Status/QoL scale, reaching minimal important differences (MID) at cycle 12 and approaching it at cycle 18 of treatment.

Immunomodulatory drugs and PRO

Five studies (45.5%) investigated the impact of IMiDs (Lenalidomide, Thalidomide and Pomalidomide) on HRQOL and OS. Thalidomide, in Waage et al. [41], was associated with a median overall survival of 12 months (statistical significance is not reported) and with stable HRQOL scores, except for clinically significant improvement of pain and increase of constipation. The same therapy, assessed in a RCT by Hjort et al. [40], was associated with an OS of 22.8 months, compared with 19 months of the control group (bortezomib), but no data about statistical significance was reported. In the PRO analysis, no clinically meaningful differences in favor of thalidomide were found. Lenalidomide was investigated in Alegre et al. [44], where median overall survival was not reached, while clinically meaningful improvements were only found in the EORTC QLQ-MY20 “Future perspective” scale. In Leleu et al. [45], patients in lenalidomide group maintained substantially their HRQOL, with only a clinically meaningful deterioration on diarrhea. Better results were reached with pomalidomide [30,31,32]: OS was 12.7 months comparing with 8.1 months of the control group; statistical and clinically significant differences were found in most scales of the PRO instruments used.

Monoclonal antibodies and PRO

One study (9.1%) evaluated the efficacy and safety of elotuzumab, a humanized monoclonal antibody anti-SLAMF7, on RRMM patients [43]. This agent, in combination with lenalidomide and dexamethasone, improved PFS and overall response rate (ORR), compared with lenalidomide and dexamethasone alone (19.4 vs. 14.9 months and 79 vs. 66%, respectively). However, AEs (any grades), including fatigue, diarrhea and constipation were generally reported at higher rates with this drug. In the PRO analysis, there were no significant differences between groups in pain and in health-related quality of life measures.

Overview of treatment recommendations

Recommendations by authors are summarized in Table 2. Authors of TOURMALINE-MM1 study [42], on the basis of benefits reported on PFS and a maintained HRQOL, recommended the addition of ixazomib to the doublet regimen of lenalidomide-dexamethasone, for RRMM patients. Authors of SUMMIT study [37,38,39] reported that bortezomib induced, in RRMM patients, clinically significant responses, and that the most common AEs (e.g., gastrointestinal symptoms, fatigue, thrombocytopenia) were manageables. In addition, the study demonstrated the complementary value of PRO assessments. Authors of APEX study [34,35,36] demonstrated that bortezomib therapy was superior to high-dose of dexamethasone, for RRMM patients who received up to three previous therapies other than bortezomib. Furthermore, consistent with clinical outcomes, this therapy was associated with better HRQOL outcomes. Callander et al. [46] did not recommend the addition of ALCAR to bortezomib, doxorubicin, and dexamethasone, as it did not eliminate treatment-related peripheral neuropathy. Stewart et al. [28, 29] stated that the addition of carfilzomib to lenalidomide and dexamethasone had a favorable risk-benefit profile, leading to improvements in PFS, HRQOL, and response rates, despite reporting higher numbers of common AEs. The authors highlighted the importance that the improvement of the efficacy was not at the cost of an impairment of HRQOL.

Regarding IMiDs studies, Waage et al. [41] affirmed that no relationship between thalidomide and effects after 12 weeks was found, although they did not exclude that such a relationship exists. According to Hjorth et al. [40], thalidomide, in combination with dexamethasone, seemed to have an efficacy compared with bortezomib plus dexamethasone in melphalan refractory myeloma patients; nonetheless, the statistical strength of the study was not optimal. However, authors recommended bortezomib for patients with advanced disease and in need of a rapid response. According to Alegre et al. [44], the study demonstrated that lenalidomide was effective and generally tolerated by RRMM patients. Regarding PRO analysis, no changes in median HRQOL scores were seen, except in “Future perspective” scale. For the authors, this suggested that, overall, the therapy did not negatively impact patient’s HRQOL and that, for an appreciable percentage of them, it produced clinically meaningful improvements. Finally, authors of MM-003 trial [30,31,32] demonstrated that pomalidomide could be considered a new treatment option for patients with RRMM who failed with bortezomib and lenalidomide therapy.

Leleu et al. [45] stated that continuous treatment with lenalidome or bortezomib, despite being associated with adverse events, did not deteriorate patients’ HRQOL.

Authors of ELOQUENT-2 study [43] elaborated their conclusions only on the base of primary clinical endpoints, highlighting the role of elotuzumab on improving PFS and ORR, but no mention was made to PRO.

Discussion

In this systematic review, we have quantified the amount of evidence-based data on the impact of salvage therapies on HRQOL and other type of PRO in patients with RRMM. Although we searched for studies published up to May 2017, we found only eleven studies. This dearth of information is striking considering the potential value that PRO information could provide to facilitate clinical decision-making in this vulnerable population.

Past reviews have focused on methodological quality of HRQOL assessments [20] by investigating methods of data collection, analysis, and reporting [19] or included studies both on newly diagnosed patients and RRMM [22]. In this work, we focused on the added value of PRO assessment in supporting clinical decision-making for RRMM therapies. We found that in most cases PRO assessment was an important factor, as PRO added novel information compared to physician-reported AEs. In this review, the studies focusing on PIs were associated with better OS and with clinical meaningful differences in several PRO scales. Carfilzomib reported improved results on Global Health Status/QoL scale [28, 29]; the use of bortezomib was associated with improvements both on Global Health Status/QoL scale and on symptoms and functional scales, in particular physical, role, emotional, and social functioning [34,35,36,37,38,39]. The HRQOL was maintained by the use of ixazomib [42] and a trend for less fatigue, emotional and physical functioning was found. In the studies about IMiDs, fewer PRO scales reached clinical meaningful differences and data about the statistical significance were frequently not reported. Pomalidomide was associated with better HRQOL on symptoms and functional scales, a result that was concordant with better clinical outcomes reached from patients treated with this therapy [30,31,32]. Lenalidomide was associated with meaningful improvements in the EORTC QLQ-MY20 “Future perspective scale” [44], while thalidomide was associated with an improvement and a worsening in, respectively, EORTC QLQ-C30 pain and constipation scales [41]. In the only study about monoclonal antibodies [43], elotuzumab did not improve HRQOL, but better results were found for PFS. Inspection of the reported AEs showed that, in most cases, all grades AEs reported in the experimental group were higher than those reported in the control group. Based on toxicity data, we may have expected carfilzomib, ixazomib, pomalidomide, bortezomib, and elotuzumab to cause an impairment on patient’s HRQOL. Instead, these therapies were associated, in the same RCT, with improved PRO outcomes or, in the case of elotuzumab and ixazomib, with no differences in PRO. For example, pomalidomide was associated with higher percentages of fatigue measured by CTCAE but, from the patient’s perspective, the same therapy was associated with improved outcomes. In the APEX trial, patients in the bortezomib group reported better outcomes in nausea, diarrhea, and neurotoxicity but this was not reflected in toxicity results. In TOURMALINE-MM1 study [42], physicians reported higher percentages of nausea and vomiting in the ixazomib group, while the same symptoms were similar between treatment arms when reported by patients themselves. This emphasizes the added value of taking into consideration the patients’ perspectives and the importance of PRO in the evaluation of treatment effectiveness. Focusing attention only on measures such as toxicity, which are typically based on laboratory information or physician’s judgements, may provide a partial view of overall treatment effectiveness. As RRMM patients are vulnerable population, often characterized by low survival rates, a more systematic implementation of PRO assessment may provide novel additional information to more robustly inform patient care.

This study has limitations. Although we used a comprehensive key search strategy, it is possible that we have missed some studies. Furthermore, the exclusion of non-English-language published papers from this review may have decreased the number of studies examined, but is unlikely that the quality of the review has been reduced [50]. The strength of our study was the analysis and inclusion of most recent studies that take into consideration novel drugs approved for RRMM patients.

In conclusion, there is paucity of evidence-based data on PRO in RRMM patients. The management of patients who have already received prior treatments represents a major challenge, as treatment options are continuously increasing. There is urgent need to implement PRO assessment in forthcoming studies in this area to more comprehensively evaluate treatment effectiveness.