Longitudinal health‐related quality of life in first‐line treated patients with chronic lymphocytic leukemia: Results from the Connect® CLL Registry

Abstract Health‐related quality of life (HRQoL) in patients with chronic lymphocytic leukemia (CLL) is important in guiding treatment decisions. However, the impact of CLL treatment initiation on HRQoL is unclear. We assessed HRQoL using the FACT‐Leu and EQ‐5D‐3L questionnaires in the Connect ® CLL Registry, a large, US‐based, multicenter, prospective observational study of CLL patients enrolled between 2010 and 2014, prior to the introduction of novel therapies. Among 889 patients initiating first‐line therapy with chemoimmunotherapy or rituximab monotherapy, questionnaire completion rates were 95.7% and 95.8% at enrollment, and 70.8% and 69.4% at 12 months, for FACT‐Leu Total and EQ‐5D‐3L, respectively. For 849 patients completing all five FACT‐Leu components, average total scores were 135.7 at enrollment and 141.6 at 12 months. Among 526 patients with FACT‐Leu Total scores at enrollment and 12 months, clinically meaningful (≥11‐point) improvements or reductions were observed in 179 (34.0%) and 88 (16.7%) patients, respectively. Mean EQ‐5D‐3L index scores were 0.87 at enrollment and 12 months. Among 513 patients completing EQ‐5D‐3L at enrollment and 12 months, clinically meaningful (≥0.06‐point) improvements or reductions were observed in 125 (24.4%) and 116 (22.6%) patients, respectively. In the Connect® CLL Registry, HRQoL remained stable or slightly improved after 12 months of follow‐up.


INTRODUCTION
Chronic lymphocytic leukemia (CLL) is generally considered incurable, even with the increasing availability of novel treatments [1]. Most patients experience a chronic disease course with periods of relapse and remission, for which they usually receive multiple lines of therapy. Current guidelines recommend treatment initiation only in symptomatic patients, for both first-line therapy (LOT1) and subsequent lines of therapy [1,2] Patients' health-related quality of life (HRQoL) plays an important role in guiding treatment decisions and assessing the impact of cancer treatment [3]. Previous studies on HRQoL in patients with CLL treated with chemoimmunotherapy (CIT) have mostly focused on global health status and fatigue, which have often been measured using cancer-generic HRQoL instruments, including the European Organization for Research and Treatment of Cancer (EORTC) 30-item quality of life questionnaire (QLQ-C30) [4][5][6][7][8], rather than more sensitive leukemia-specific instruments such as the Functional Assessment of Cancer Therapy-Leukemia (FACT-Leu), which may more accurately represent the symptoms experienced by patients with CLL [9][10][11].
Longitudinal data on HRQoL in CLL are also sparse [12] and have mostly been reported in observational studies with small sample sizes [13,14].
In this study from the Connect ® CLL Registry, longitudinal analyses of HRQoL were performed in a predominantly community-based cohort of patients with CLL, focusing on those patients undergoing LOT1 with CIT or rituximab (R) monotherapy at enrollment into the Registry. As the Registry was initiated prior to the approval of novel agents such as the Bruton tyrosine kinase inhibitors, phosphoinositide 3-kinase inhibitors, and venetoclax, no patients received these agents during LOT1. Patients receiving LOT1 were included in the analysis to avoid introducing additional variability caused by prior therapies and disease progression that may confound the analyses. HRQoL was assessed using the FACT-Leu and the EuroQol fivedimensional three-level (EQ-5D-3L) questionnaires, instruments that are easy to complete, widely used, and validated in CLL and other cancers [9][10][11][12][15][16][17]. To understand whether changes in HRQoL scores were clinically meaningful, the minimally important difference (MID), that is the smallest change in an outcome that a patient would identify as important or that would result in a change in treatment, was utilized [18].

The Connect ® CLL registry
The Connect ® CLL Registry (NCT01081015) is a large, US-based, multicenter, prospective observational cohort study of adult patients with CLL. Full details on the Registry design have been described previously [19]. Further details of the Registry are provided in the Supporting Information Methods.

HRQoL assessments
HRQoL was assessed using the FACT-Leu and EQ-5D-3L questionnaires (see Supporting Information Methods for further details).
Briefly, the FACT-Leu consists of five components: Physical, Social, Emotional, and Functional Well-Being, plus leukemia-specific Additional Concerns [9,10]. FACT-Leu Total scores are presented for patients who completed ≥36 of 44 items overall, regardless of completion of all the individual components [9,20]. Component scores are presented for patients who answered ≥50% of items for each of the five FACT-Leu components. Higher scores across all domains denote better HRQoL.
The EQ-5D-3L consists of five domains covering mobility, self-care, usual activities, pain/discomfort, and anxiety/depression [21]. Results are reported as individual domain scores and as a summary index score.
Patients also reported their self-rated health on a visual analogue scale (VAS). Lower scores on the EQ-5D-3L domains and higher scores on the EQ-5D-3L index and VAS correspond to better HRQoL.

Statistical analysis
The study closed 31 December, 2017. Mean HRQoL changes from baseline and differences between patient clusters over time were analyzed using a repeated measures regression model. To determine whether changes in HRQoL scores were clinically meaningful, the MID for each instrument was prespecified and defined in accordance with previously published limits [22,23]. For the FACT-Leu Total scores, the MID was 11 [23]. For the EQ-5D-3L index-based scores, the MID was 0.06 [22].
Univariate and multivariable logistic regression analyses were performed to identify factors associated with clinically meaningful improvements in HRQoL on the FACT-Leu Total scores and EQ-5D-3L index scores from baseline to 12 months (see Supporting Information Methods   In the 828 patients who completed both questionnaires at baseline, baseline characteristics were similar to those from the full LOT1 cohort (Table 1).  Abbreviations: EQ-5D-3L, EuroQol 5-dimension 3-level questionnaire; FACT-Leu, Functional Assessment of Cancer Therapy-Leukemia; HRQoL, healthrelated quality of life; LOT1, first-line therapy.

Completed FACT-Leu Total Score at baseline N=851
Completed FACT-Leu Total Score at baseline + month 12 N=526

Completed baseline EQ-5D and 5 FACT-Leu components at baseline* N=828
Completed EQ-5D and FACT-Leu Total Score at baseline + month 12 N=493 Patients completing both questionnaires Completed EQ-5D at baseline + month 12 N=513

EQ-5D-3L
Among 852 patients who completed the EQ-5D-3L at baseline, mean EQ-5D-3L index scores were 0.87 at baseline and 0.87 at 12 months In the 852 patients who completed baseline EQ-5D-3L questionnaires, patients were clustered into two groups based on baseline EQ-5D-3L scores; inferior and superior ( Figure 1). EQ-5D-3L scores by cluster are shown in Figure 2B. The superior EQ-5D-3L cluster consisted of 609 patients (71.5%), with a mean EQ-5D-3L index score of 0.92 over the first 12 months. The inferior EQ-5D-3L cluster included 243 patients (28.5%) whose mean EQ-5D-3L index score was 0.72. Differences between baseline characteristics of patients in each cluster are described in the Supporting Information Results.
Cluster analyses showed differences in HRQoL between the clusters across most EQ-5D-3L components except self-care ( Figure 4B). In the inferior EQ-5D-3L cluster, there was a trend toward improved EQ-5D-3L domain and index scores in the first 3-6 months, with smaller improvements seen in months 6-12 on most scores except self-care.
In the superior EQ-5D-3L cluster, there was a trend toward worsened EQ-5D-3L component and index scores in the first 3-6 months, with stable scores seen thereafter.

Treatment received and HRQoL
Of the 851 patients completing FACT-Leu Total scores at baseline, 240    symptoms and experience of their disease. Furthermore, the trend toward an improvement in HRQoL, as measured by FACT-Leu, in FCR-treated patients may be due to a greater sensitivity in detecting leukemia-specific improvements in HRQoL with this leukemiafocused instrument. The Additional Concerns items, such as "lumps or swelling," which would likely be treatment-sensitive, showed the greatest improvement in the FACT-Leu score.

Predictors of HRQoL improvement
Data regarding the effect of treatment initiation in CLL on HRQoL are mixed. In the GCLLSG CLL8 trial, no significant differences in HRQoL were observed between fludarabine and cyclophosphamide (FC) and FCR during treatment or follow-up as measured by the EORTC QLQ-C30 questionnaire, despite a higher adverse event burden with FCR [6]. Similarly, in the GCLLSG CLL4 trial there was no difference in HRQoL, as assessed by EORTC QLQ-C30, between fludarabine monotherapy and FC [24]. Two trials of lenalidomide and ofatumumab showed either no effect or a slight positive effect of treatment on HRQoL, respectively [15,25]. However, in two population-based studies from the Netherlands, initiation of chemotherapy or CIT was associated with a considerable worsening of HRQoL [14,26]. Differences in setting, study design, and population should be considered when interpreting these data. Furthermore, the long-term effect of treatment must not be discounted. EORTC QLQ-C30 data comparing chlorambucil monotherapy, fludarabine monotherapy, and FC showed initial differences in HRQoL between patients receiving fludarabine and those on chlorambucil alone. However, these differences were transient and all patients experienced a positive effect on HRQoL when remission was ultimately achieved [27]. These factors may also explain why there was no significant improvement in HRQoL among patients receiving first-line treatment in the Connect CLL Registry. All patients in the Registry received chemotherapy regimens at LOT1, which are known to have a high adverse event burden [28] and require hospital visits in order to receive the drug infusions. These factors may negatively affect the patient perception of their health or the impact of the disease on their daily life. However, all patients were receiving firstline treatment for CLL, reducing confounding caused by prior therapies and patients with relapsed/refractory disease. Therefore, these findings provide important insight into the real-world experiences of patients undergoing first-line treatment with chemotherapy regimens.
Our study has several limitations. Selection bias can occur as physicians may select certain patients for enrollment. Additionally, median time between CLL diagnosis and Registry enrollment, for which treatment initiation is a requirement, was relatively short in the Registry patients compared with the overall CLL population [29], suggesting a possible bias toward a higher-risk population. To minimize these limitations, consecutive patients presenting to the sites were evaluated for enrollment and invited to participate. However, the observed improvement in mean HRQoL scores over time may have been influenced by early patient dropout, that is patients with poorer health dropping out of the Registry or declining further HRQoL participation earlier than healthier patients, which could bias the mean score upward. Subset sensitivity analyses showed that patients with lower ECOG PS scores, indicating better health, were more likely to complete both questionnaires at 12 months. However, baseline HRQoL scores themselves were not associated with 12-month completion rates. Additionally, the repeated measures regression model applied is robust to data missing not at random. Treatment at academic centers, younger patient age, and non-white race were all associated with non-compliance with HRQoL questionnaires. This may indicate an unmet need for greater patient education to encourage questionnaire completion or to identify those patients requiring more support to complete questionnaires.
Due to the timing of the Connect ® CLL Registry, no patients received first-line treatment with novel agents. Novel agents such as ibrutinib, idelalisib, and venetoclax are now commonly used following approval by the US Food and Drug Administration for previously untreated and relapsed/refractory patients. A recent analysis of HRQoL in the HELIOS study showed that there was no change in HRQoL between patients with relapsed CLL receiving ibrutinib plus BR (followed by ibrutinib alone) versus placebo plus BR (followed by ibrutinib alone). However, the subset of patients who had worse wellbeing, physical functioning, and fatigue at baseline experienced greater improvements in these HRQoL outcomes with ibrutinib plus BR versus placebo plus BR [4]. Similarly, in the COMPLEMENT2 trial, the addition of the anti-CD20 antibody ofatumumab to FC did not affect HRQoL, as measured by the EORTC QLQ-C30 and EORTC Quality of Life Questionnaire -CLL module, while small improvements in HRQoL were maintained after treatment [7].
As with any HRQoL study, it is important to consider whether statistically significant differences are relevant and meaningful for patients.
While a MID threshold of 0.06 is generally used for EQ-5D-3L index scores in US patients [22], no clear MID thresholds have been defined for the individual EQ-5D-3L domains. Our data also suggest that the leukemia-specific FACT-Leu measure may be more sensitive at detecting incremental HRQoL changes following initiation of a new treatment. While the EQ-5D-3L instrument has been used in other studies of CLL [4,[14][15][16], patients may have felt that the generic EQ-5D-3L questionnaire did not accurately capture their experiences of living with CLL, as reflected in the lower EQ-5D-3L completion rates. The FACT-Leu questionnaire is a relatively new HRQoL instrument, which has been validated in adult patients with acute and chronic leukemia [9]. Few studies have reported the use of the FACT-Leu in patients with CLL [12,15,16], but it has been used in patients with chronic myeloid leukemia [30][31][32][33][34] and acute myeloid leukemia [35,36]. Increased use of this measure in real-world and clinical trial settings will help define its utility and the impact of treatment on HRQoL in patients with CLL.
In the Connect ® CLL Registry in which patients were predominantly treated with CIT in the first-line, HRQoL remained stable to slightly improved over 12 months of follow-up. No significant increases in HRQoL were observed, mirroring previous reports in which treatment initiation did not have a significant impact on HRQoL. It will be important to assess these indices in the era of novel targeted therapies and to validate how FACT-Leu and EQ-5D-3L values in CLL compare with other chronic medical conditions. ing from AbbVie, Celgene Corporation, Gilead, Pronai and TG Therapeutics; has been a consultant for AbbVie; and has been a member of a speakers' bureau for Celgene Corporation.