Key Points

  • The advantages of first-line FCR over BR seen in clinical trials were not seen in community practice in this Connect CLL Registry study.

  • Novel agents, especially ibrutinib, improved outcomes for relapsed/refractory CLL, supporting their inclusion in the treatment paradigm.

Abstract

Optimal treatment of chronic lymphocytic leukemia (CLL) remains unclear. The Connect CLL Registry, a United States–based multicenter prospective observational cohort study, enrolled 1494 patients between 2010 and 2014 from predominantly community-based settings. Patients were grouped by line of therapy (LOT) at enrollment. With a median follow-up of 46.6 months (range, 0-63.0 months), median overall survival (OS) was not reached in LOT1, 63.0 months (95% confidence interval [CI], 46.0-63.0 months) in LOT2, and 38.0 months (95% CI, 33.0-47.0 months) in LOT≥3. Bendamustine and rituximab (BR; 33.5%); fludarabine, cyclophosphamide, and rituximab (FCR; 21.4%); and rituximab monotherapy (18.5%) were the most common regimens across LOTs. Median event-free survival (EFS) was similar in patients treated with BR (59.0 months) and FCR (55.0 months) in LOT1; median OS was not reached. In multivariable analysis, BR or FCR vs other treatments in LOT1 was associated with improved EFS (hazard ratio [HR], 0.60; P < .0001) and OS (0.67; P = .0162). Using the Kaplan-Meier product limit, ibrutinib vs other treatments improved OS in LOT2 (HR, 0.279; P = .009), LOT3 (0.441; P = .011), and LOT≥4 (0.578; P = .043). Prognostic modeling of death at 2 years postenrollment identified 3 risk groups: low (mortality rate, 6.2%), medium (14.5%), and high (27.4%). The most frequent adverse events across LOTs were pneumonia (11.6%) and febrile neutropenia (6.2%). These data suggest that advantages of LOT1 FCR over BR seen in clinical trials may not translate to community practice, whereas receiving novel LOT2 agents improved outcomes. This trial was registered at www.clinicaltrials.gov as NCT01081015.

Introduction

Recent advances have transformed the treatment of chronic lymphocytic leukemia (CLL).1  In clinical trials, standard first-line treatment with fludarabine, cyclophosphamide, and rituximab (FCR) is associated with median progression-free survival (PFS) of 4.3 to 8.0 years2-5  and median overall survival (OS) of 8.4 to 12.7 years,2,4  whereas treatment with bendamustine and rituximab (BR) leads to median PFS of ∼3.5 years.3,6,7  However, relapse is common, particularly in patients with poor prognostic features, such as unmutated immunoglobulin heavy chain (IgHV) genes and del(17p).2-4  Ibrutinib, obinutuzumab, and venetoclax are all indicated as first-line therapy, whereas in the relapsed/refractory (R/R) setting, ibrutinib, idelalisib (with or without rituximab [R]), obinutuzumab, ofatumumab, duvelisib, and venetoclax (with or without R) are indicated.8-10  Treatment choice is based on patient age and fitness, disease characteristics (eg, lymph node size), molecular and genetic factors (eg, IgHV mutation status, TP53 mutations, del[17p], del[11q]), response to prior treatment, treatment goals, and patient preference.9-12 

Despite novel treatments, such as ibrutinib (US Food and Drug Administration [FDA] approved in February 2014) and venetoclax (FDA approved in April 2016),8,13  many questions remain regarding optimal treatment choice and sequencing.14,15  Clinical trial results may differ from those in daily practice, and patients enrolled in trials tend to be younger, more motivated, and healthier than those seen in daily practice, complicating treatment selection and extrapolation of clinical trials results to general practice.16  Patient age and health are particularly relevant in CLL, where the median age at diagnosis is 70 years,17  and treatment decisions are often complicated by comorbidity and polypharmacy.11  Additionally, treatment patterns and dose intensity in real-world settings can vary considerably.18-20 

The Connect CLL Registry is a large US-based multicenter prospective observational cohort study of patients initiating therapy for CLL, designed to collect information on treatment patterns and outcomes in a real-world setting.21  Here, we present the final data from the Connect CLL Registry. Survival outcomes (event-free survival [EFS] and OS) are described for patients with CLL undergoing first-, second-, or subsequent-line therapy, including factors associated with improved EFS and OS. Because the Registry enrolled patients between 2010 and 2014 before the publication of data showing superiority of novel agents over chemoimmunotherapy (CIT), our analyses focused on the most common CIT regimens, as well as ibrutinib-containing regimens in the R/R setting. Real-world safety outcomes are also described.

Patients and methods

Connect CLL Registry

Full details of the Connect CLL Registry design and conduct have been described previously.21  Briefly, study centers were invited to enroll up to 30 eligible patients per site. Eligible patients were age ≥18 years, had CLL (defined by International Workshop on Chronic Lymphocytic Leukemia 2008 guidelines22 ), and had initiated a new line of therapy (LOT) within 60 days before enrollment. Exclusion criteria were prolymphocytic leukemia or small lymphocytic lymphoma and terminal disease (<6 months to live).

Participation was voluntary, and all patients provided written informed consent. Patient data were captured in an electronic data-capture system at baseline (enrollment) and every 3 months, for up to 5 years or until early discontinuation because of study termination, patient withdrawal, loss to follow-up, or death. This noninterventional study was conducted in accordance with the Declaration of Helsinki, and the Registry protocol was approved by a central institutional review board (IRB) (Quorum Review IRB, Seattle, WA) or the respective IRB at each site. This study report is aligned with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.23 

Survival and response analyses

Patients were grouped into 3 cohorts defined by LOT initiated at enrollment: LOT1, LOT2, and LOT≥3; for the ibrutinib subanalysis, patients were further grouped into a LOT≥4 cohort, as described in the following paragraph. Treatment-free survival (TFS) was calculated as time to next CLL treatment or death. EFS was calculated as time from Registry enrollment to first event; event was defined as the first sign of progressive disease (PD), treatment of relapse, transformation, or death. OS was defined as time from enrollment to death. Patients without a known date of death were censored at last known date alive, study discontinuation date, or Registry end date (whichever was earliest). Comparisons of EFS and OS between groups were assessed using the Kaplan-Meier product-limit method.

OS was also compared between patients who received ibrutinib and those who did not receive kinase inhibitors. In this subanalysis, OS was redefined as time from first treatment with a specific LOT to death. Patients were grouped by LOT; because no patients received ibrutinib in LOT1, groups based on LOT2, LOT3, and LOT≥4 were used. In the analysis of the LOT2 group, patients enrolled in LOT2 who never received kinase inhibitors in any subsequent LOT were compared with patients treated with ibrutinib in LOT2. The LOT3 and LOT≥4 groups were analyzed similarly.

Investigator-reported response was assessed in the per-protocol population by overall response rate, defined as partial response (PR) or better and complete response (CR) rates. Confirmatory bone marrow samples for CR assessment were not required.

Statistical analyses

Cox regression analysis was used to identify factors associated with improved EFS and OS. Variables associated with improved EFS and OS included demographic and baseline characteristics: age, sex, region, Eastern Cooperative Oncology Group performance status (ECOG PS) score, practice type (academic or community), Charlson Comorbidity Index (CCI) score, risk group as defined by molecular markers del(17p) and del(11q), and enrollment therapy. Predictors demonstrating a statistically significant association with outcome at P < .1 in the univariate regression model were included in the multivariable regression model. The final set of independent predictors was derived using a score-based selection process.

The target enrollment of 1500 patients from 200 treatment centers was not based on statistical consideration, but was deemed sufficient to meet the study objectives. All statistical analyses were conducted using SAS software (version ≥9.2; SAS Institute, Cary, NC).

Prognostic modeling of early death within 2 years

A prognostic model using univariate logistic regression was developed to identify patient characteristics associated with risk of death within 2 years of enrollment; variables included enrollment site, race, age, insurance coverage, household income, enrollment regimen, CCI score, ECOG PS score at baseline, and Rai stage at baseline. Modeling was performed using a bootstrap model selection method24  by drawing repeated bootstrap samples from the original data set. A parsimonious model was fit using multivariable logistic regression with backward variable selection. A total of 200 bootstrap samples were derived by removing 40 randomly selected patients from the initial data set (ie, ∼5% of the total sample), and a predictive model was generated using the remaining 95% of each sample. The predictive ability of each model was assessed using the c-index,25  which is identical to the area under the receiver operating characteristic curve. The 50 models with the most discriminatory power (of the 200 bootstrap samples) were identified by the c-index. Variables associated with early death in ≥1 model were assigned a score of 1 to 3 based on relative magnitude of effect.26  Total scores were calculated for each patient, and patients were classified as being low (score ≤1), medium (score 2-4), or high risk (score ≥5). Prognostic models were validated by a similar bootstrap procedure and by multivariable analysis of all patients with a grouping variable and interaction terms with each of the covariates.

Safety

Serious adverse events (SAEs; graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events [version 4.0]) or potential SAEs that occurred during enrollment in the Registry were recorded. Tumor lysis syndrome, tumor flare reaction, and second primary malignancy (SPM) were recorded as adverse events (AEs) of interest. Hospitalization or intensive care unit admission and duration of hospitalization were also recorded.

Results

Patient baseline characteristics

Between March 2010 and January 2014, 1494 patients were enrolled in the Connect CLL Registry from 199 US sites. A majority of patients (n = 1311 [87.8%]) were treated in 179 community hospitals, whereas 155 patients (10.4%) received therapy in 17 academic institutions, and 28 patients (1.9%) received therapy in 3 government centers (supplemental Figure 1; supplemental Table 1). Median age was 69 years (range, 22-99 years), and 953 patients (63.8%) were male. Median time from CLL diagnosis to enrollment was 3.7 years (range, 0-32 years): 1.5 years in patients receiving LOT1; 5.6 years for LOT2; and 8.3 years for LOT≥3 (Table 1).

Table 1.

Patient demographics and characteristics at enrollment in the Registry by LOT and overall

LOT1 (n = 889)LOT2 (n = 261)LOT≥3 (n = 344)All patients (N = 1494)
Age, y     
 Median 68.0 71.0 69.0 69.0 
 Range 22-99 34-96 34-93 22-99 
Male sex 566 (63.7) 171 (65.5) 216 (62.8) 953 (63.8) 
Time from CLL diagnosis to Registryenrollment, y     
 Median 1.5 5.6 8.3 3.7 
 Range 0-32 0-32 0-32 0-32 
Race n = 863 n = 251 n = 330 n = 1444 
 White 798 (92.5) 234 (93.2) 301 (91.2) 1333 (92.3) 
 Black 56 (6.5) 15 (6.0) 27 (8.2) 98 (6.8) 
 Asian 3 (0.3) 0 (0) 1 (0.3) 4 (0.3) 
 American Indian/Alaskan native 0 (0) 1 (0.4) 0 (0) 1 (0.1) 
 Other 6 (0.7) 1 (0.4) 1 (0.3) 8 (0.6) 
US geographic region n = 881 n = 260 n = 342 n = 1483 
 Northeast 112 (12.7) 42 (16.2) 53 (15.5) 207 (14.0) 
 Midwest 277 (31.4) 68 (26.2) 114 (33.3) 459 (31.0) 
 South 352 (40.0) 113 (43.5) 126 (36.8) 591 (39.9) 
 West 140 (15.9) 37 (14.2) 49 (14.3) 226 (15.2) 
Institution type     
 Academic 86 (9.7) 28 (10.7) 41 (11.9) 155 (10.4) 
 Community 787 (88.5) 228 (87.4) 296 (86.0) 1311 (87.8) 
 Government 16 (1.8) 5 (1.9) 7 (2.0) 28 (1.9) 
Insurance type     
 Medicare 512 (57.6) 171 (65.5) 231 (67.2) 914 (61.2) 
 Medicaid 42 (4.7) 11 (4.2) 12 (3.5) 65 (4.4) 
 Supplemental coverage 178 (20.0) 60 (23.0) 88 (25.6) 326 (21.8) 
 Private health insurance 403 (45.3) 99 (37.9) 125 (36.3) 627 (42.0) 
 Military 15 (1.7) 3 (1.1) 8 (2.3) 26 (1.7) 
 Self-pay 13 (1.5) 3 (1.1) 3 (0.9) 19 (1.3) 
 Other 13 (1.5) 3 (1.1) 8 (2.3) 24 (1.6) 
 Not specified 20 (2.2) 13 (5.0) 8 (2.3) 41 (2.7) 
ECOG PS score n = 690 n = 183 n = 250 n = 1123 
 0 347 (50.3) 77 (42.1) 104 (41.6) 528 (47.0) 
 1 296 (42.9) 94 (51.4) 123 (49.2) 513 (45.7) 
 2-4 47 (6.8) 12 (6.6) 23 (9.2) 82 (7.3) 
CCI score     
 Median 2.0 2.0 2.0 2.0 
 Range 2.0-10.0 2.0-9.0 2.0-13.0 2.0-13.0 
LOT1 (n = 889)LOT2 (n = 261)LOT≥3 (n = 344)All patients (N = 1494)
Age, y     
 Median 68.0 71.0 69.0 69.0 
 Range 22-99 34-96 34-93 22-99 
Male sex 566 (63.7) 171 (65.5) 216 (62.8) 953 (63.8) 
Time from CLL diagnosis to Registryenrollment, y     
 Median 1.5 5.6 8.3 3.7 
 Range 0-32 0-32 0-32 0-32 
Race n = 863 n = 251 n = 330 n = 1444 
 White 798 (92.5) 234 (93.2) 301 (91.2) 1333 (92.3) 
 Black 56 (6.5) 15 (6.0) 27 (8.2) 98 (6.8) 
 Asian 3 (0.3) 0 (0) 1 (0.3) 4 (0.3) 
 American Indian/Alaskan native 0 (0) 1 (0.4) 0 (0) 1 (0.1) 
 Other 6 (0.7) 1 (0.4) 1 (0.3) 8 (0.6) 
US geographic region n = 881 n = 260 n = 342 n = 1483 
 Northeast 112 (12.7) 42 (16.2) 53 (15.5) 207 (14.0) 
 Midwest 277 (31.4) 68 (26.2) 114 (33.3) 459 (31.0) 
 South 352 (40.0) 113 (43.5) 126 (36.8) 591 (39.9) 
 West 140 (15.9) 37 (14.2) 49 (14.3) 226 (15.2) 
Institution type     
 Academic 86 (9.7) 28 (10.7) 41 (11.9) 155 (10.4) 
 Community 787 (88.5) 228 (87.4) 296 (86.0) 1311 (87.8) 
 Government 16 (1.8) 5 (1.9) 7 (2.0) 28 (1.9) 
Insurance type     
 Medicare 512 (57.6) 171 (65.5) 231 (67.2) 914 (61.2) 
 Medicaid 42 (4.7) 11 (4.2) 12 (3.5) 65 (4.4) 
 Supplemental coverage 178 (20.0) 60 (23.0) 88 (25.6) 326 (21.8) 
 Private health insurance 403 (45.3) 99 (37.9) 125 (36.3) 627 (42.0) 
 Military 15 (1.7) 3 (1.1) 8 (2.3) 26 (1.7) 
 Self-pay 13 (1.5) 3 (1.1) 3 (0.9) 19 (1.3) 
 Other 13 (1.5) 3 (1.1) 8 (2.3) 24 (1.6) 
 Not specified 20 (2.2) 13 (5.0) 8 (2.3) 41 (2.7) 
ECOG PS score n = 690 n = 183 n = 250 n = 1123 
 0 347 (50.3) 77 (42.1) 104 (41.6) 528 (47.0) 
 1 296 (42.9) 94 (51.4) 123 (49.2) 513 (45.7) 
 2-4 47 (6.8) 12 (6.6) 23 (9.2) 82 (7.3) 
CCI score     
 Median 2.0 2.0 2.0 2.0 
 Range 2.0-10.0 2.0-9.0 2.0-13.0 2.0-13.0 

Data are n (%) unless otherwise noted.

At enrollment, 889 patients (59.5%) were receiving LOT1, 261 (17.5%) were receiving LOT2, and 344 (23.0%) were receiving LOT≥3. At end of study, 520 patients (34.8%) had completed 5 years of follow-up. The most common reason for early discontinuation was death (n = 470 [31.5%]), followed by study termination (n = 116 [7.8%]), loss to follow-up (n = 113 [7.6%]), and withdrawal of informed consent (n = 111 [7.4%]). The 470 patients who died were included in the statistical analysis. Median duration of follow-up was 54.4 months (range, 1-63 months) in LOT1, 33.9 months (range, 0-63 months) in LOT2, and 27.1 months (range, 0-60 months) in LOT≥3.

Overall, the most frequently used regimens (including follow-up therapy) were BR (n = 500 [33.5%]), FCR (n = 320 [21.4%]), and R monotherapy (n = 276 [18.5%]; Table 2). These were also the most frequently used regimens in LOT1, with 230 patients (25.9%) receiving FCR, 188 (21.1%) receiving BR, and 103 (11.6%) receiving R monotherapy. Ibrutinib was used by 249 patients (16.7%) with R/R disease; no patients received kinase inhibitors in LOT1.

Table 2.

Top 15 most frequently used treatment regimens overall and by LOT, including follow-up therapy

Treatment*LOT1 (n = 889)LOT2 (n = 261)LOT≥3 (n = 344)Total (N = 1494)
BR 188 (21.1) 151 (57.9) 180 (52.3) 500 (33.5) 
FCR 230 (25.9) 59 (22.6) 39 (11.3) 320 (21.4) 
R monotherapy 103 (11.6) 105 (40.2) 116 (33.7) 276 (18.5) 
Ibrutinib  72 (27.6) 183 (53.2) 249 (16.7) 
Bendamustine 31 (3.5) 18 (6.9) 67 (19.5) 114 (7.6) 
FR 56 (6.3) 23 (8.8) 27 (7.8) 104 (7.0) 
Ofatumumab  13 (5.0) 79 (23.0) 91 (6.1) 
Chlorambucil 41 (4.6) 14 (5.4) 31 (9.0) 81 (5.4) 
Alemtuzumab  9 (3.4) 49 (14.2) 62 (4.1) 
Investigational product 14 (1.6) 8 (3.1) 37 (10.8) 55 (3.7) 
R-CVP 23 (2.6) 18 (6.9)  50 (3.3) 
R-CP 24 (2.7) 12 (4.6) 12 (3.5) 48 (3.2) 
Fludarabine 12 (1.3) 14 (5.4)  33 (2.2) 
Lenalidomide   23 (6.7) 32 (2.1) 
R plus investigational product 17 (1.9)   27 (1.8) 
Treatment*LOT1 (n = 889)LOT2 (n = 261)LOT≥3 (n = 344)Total (N = 1494)
BR 188 (21.1) 151 (57.9) 180 (52.3) 500 (33.5) 
FCR 230 (25.9) 59 (22.6) 39 (11.3) 320 (21.4) 
R monotherapy 103 (11.6) 105 (40.2) 116 (33.7) 276 (18.5) 
Ibrutinib  72 (27.6) 183 (53.2) 249 (16.7) 
Bendamustine 31 (3.5) 18 (6.9) 67 (19.5) 114 (7.6) 
FR 56 (6.3) 23 (8.8) 27 (7.8) 104 (7.0) 
Ofatumumab  13 (5.0) 79 (23.0) 91 (6.1) 
Chlorambucil 41 (4.6) 14 (5.4) 31 (9.0) 81 (5.4) 
Alemtuzumab  9 (3.4) 49 (14.2) 62 (4.1) 
Investigational product 14 (1.6) 8 (3.1) 37 (10.8) 55 (3.7) 
R-CVP 23 (2.6) 18 (6.9)  50 (3.3) 
R-CP 24 (2.7) 12 (4.6) 12 (3.5) 48 (3.2) 
Fludarabine 12 (1.3) 14 (5.4)  33 (2.2) 
Lenalidomide   23 (6.7) 32 (2.1) 
R plus investigational product 17 (1.9)   27 (1.8) 

Data are n (%). Patients could receive any treatment regimen during multiple LOTs.

FR, fludarabine and rituximab; R-CP, rituximab, cyclophosphamide, and pentostatin; R-CVP, rituximab, cyclophosphamide, vincristine, and prednisone.

*

Patients could receive >1 therapy.

Not in top 15 most frequently used regimens.

A higher proportion of patients with Rai stage 3 to 4 received FCR (28.5%) or BR (20.5%) compared with patients with Rai stage 0 to 2 (23.6% and 16.1%, respectively). Conversely, 9.8% of patients with Rai stage 0 to 2 received R monotherapy in LOT1 compared with 6.1% of patients with Rai stage 3 to 4.

Response

Response rates in LOT1 for the 3 most common regimens received are shown in Figure 1. Rates of CR were similar between patients receiving FCR and BR (44.3% vs 44.0%) but higher than in those receiving R monotherapy (15.8%). Rates of PR were higher in patients receiving FCR (24.5%) than in those receiving BR (18.2%) or R monotherapy (13.3%). Notably, response rates were lower than reported in clinical trials of these regimens. When response rates based on all assessments within the enrollment LOT were considered, rates of CR and PR were higher in patients receiving FCR than in patients receiving BR (supplemental Figure 2A). There were no differences in response rates when a high-risk population of patients with del(17p) mutation were excluded (supplemental Figure 2B).

Figure 1.

Patient response to enrollment regimen at LOT1. Response rates for patients receiving FCR, BR, or R monotherapy in LOT1 (based on all response assessments during LOT1). Bone marrow confirmation was not required for CR or PR assessment. NE, not evaluable; SD, stable disease.

Figure 1.

Patient response to enrollment regimen at LOT1. Response rates for patients receiving FCR, BR, or R monotherapy in LOT1 (based on all response assessments during LOT1). Bone marrow confirmation was not required for CR or PR assessment. NE, not evaluable; SD, stable disease.

TFS

TFS was significantly longer for patients receiving LOT1 (median, 42.0 months) and decreased as patients received LOT2 (17.0 months) and LOT≥3 (13.0 months; P < .0001; supplemental Figure 3A). Median TFS was similar in patients receiving BR or FCR in LOT1 (54.0 vs 52.0 months; supplemental Figure 3B).

EFS

Overall, 934 patients (62.5%) experienced an EFS event: PD (n = 519 [34.7%]), treatment of relapse (n = 208 [13.9%]), death (n = 188 [12.6%]), or transformation (n = 19 [1.3%]). Median EFS was 29.0 months (95% confidence interval [CI], 26.0-32.0 months) in the total population. As expected, EFS was longer in patients receiving LOT1 and decreased as patients received subsequent LOTs (Figure 2A). Median EFS was similar in patients receiving BR and patients receiving FCR but was significantly longer than in patients receiving other treatment in LOT1 (Figure 2B); median duration of treatment was 4.67, 4.12, and 1.61 months in patients receiving BR, FCR, and other treatment in LOT1, respectively. Similar results were seen when a high-risk population of patients with del(17p) were excluded from the analyses (supplemental Figure 4A).

Figure 2.

Survival probability of patients. EFS of patients by LOT (A) and LOT1 regimen (B). OS of patients by LOT (C) and LOT1 regimen (D).

Figure 2.

Survival probability of patients. EFS of patients by LOT (A) and LOT1 regimen (B). OS of patients by LOT (C) and LOT1 regimen (D).

In LOT1, age <75 years (vs ≥75 years; hazard ratio [HR], 0.74), treatment with BR or FCR (vs other; HR, 0.60), and absence of del(17p) (vs presence of del[17p]; HR, 0.54) were significantly associated with improved EFS in multivariable analyses (Table 3). In LOT≥2, treatment with BR or FCR (vs other; HR, 0.64) and absence of del(17p) (vs presence of del[17p]; HR, 0.51) were significantly associated with improved EFS in multivariable analyses (Table 3).

Table 3.

Predictors of EFS in univariate and multivariable analyses

VariableLOT1LOT≥2
Univariate analysisMultivariable analysisUnivariate analysisMultivariable analysis
HRPHR (95% CI)PHRPHR (95% CI)P
Age category, y         
 <75 vs ≥75 0.62 <.0001 0.74 (0.572-0.954) .0205 0.83 NS — — 
Treatment         
 BR or FCR vs other 0.55 <.0001 0.60 (0.477-0.762) <.0001 0.63 <.0001 0.64 (0.487-0.842) .0014 
del(17p)         
 No vs yes 0.53 .0002 0.54 (0.382-0.757) .0004 0.49 .0003 0.51 (0.345-0.748) .0006 
Insurance         
 Other vs private 1.36 .0012 — — 0.93 NS — — 
Race         
 Other vs white 1.34 .0445 — — 1.01 NS — — 
ECOG PS         
 ≤1 vs ≥2 1.05 NS — — 1.30 .0160 — — 
Site         
 Academic vs community/government 1.02 NS — — 0.86 NS — — 
Sex         
 Female vs male 0.94 NS — — 0.91 NS — — 
Rai stage         
 ≤1 vs ≥2 0.92 NS — — 0.92 NS — — 
Prior malignancy         
 No vs yes 0.86 NS — — 1.00 NS — — 
CD38         
 Negative vs positive 0.86 NS — — 0.77 .0087 — — 
CCI score         
 ≤2 vs ≥3 0.80 .0141 — — 0.85 NS — — 
Region         
 Midwest vs west 0.80 NS — — 1.02 NS — — 
 South vs west 0.78 NS — — 1.10 NS — — 
 Northeast vs west 0.78 NS — — 0.97 NS — — 
VariableLOT1LOT≥2
Univariate analysisMultivariable analysisUnivariate analysisMultivariable analysis
HRPHR (95% CI)PHRPHR (95% CI)P
Age category, y         
 <75 vs ≥75 0.62 <.0001 0.74 (0.572-0.954) .0205 0.83 NS — — 
Treatment         
 BR or FCR vs other 0.55 <.0001 0.60 (0.477-0.762) <.0001 0.63 <.0001 0.64 (0.487-0.842) .0014 
del(17p)         
 No vs yes 0.53 .0002 0.54 (0.382-0.757) .0004 0.49 .0003 0.51 (0.345-0.748) .0006 
Insurance         
 Other vs private 1.36 .0012 — — 0.93 NS — — 
Race         
 Other vs white 1.34 .0445 — — 1.01 NS — — 
ECOG PS         
 ≤1 vs ≥2 1.05 NS — — 1.30 .0160 — — 
Site         
 Academic vs community/government 1.02 NS — — 0.86 NS — — 
Sex         
 Female vs male 0.94 NS — — 0.91 NS — — 
Rai stage         
 ≤1 vs ≥2 0.92 NS — — 0.92 NS — — 
Prior malignancy         
 No vs yes 0.86 NS — — 1.00 NS — — 
CD38         
 Negative vs positive 0.86 NS — — 0.77 .0087 — — 
CCI score         
 ≤2 vs ≥3 0.80 .0141 — — 0.85 NS — — 
Region         
 Midwest vs west 0.80 NS — — 1.02 NS — — 
 South vs west 0.78 NS — — 1.10 NS — — 
 Northeast vs west 0.78 NS — — 0.97 NS — — 

NS, not significant.

OS

In total, 470 patients (31.5%) died during the study: 191 patients in LOT1 (21.5%), 106 in LOT2 (40.6%), and 173 in LOT≥3 (50.3%). The most frequent causes of death were PD (n = 175 [37.2%]), infection (n = 101 [21.5%]), and SPM (n = 41 [8.7%]). Median OS in the total population was 63.0 months (95% CI, 63.0 months to not reached). As expected, OS was longer in patients receiving LOT1 and decreased as patients received subsequent LOTs (Figure 2C). Median OS was not reached in patients receiving BR, FCR, or other treatment in LOT1 (Figure 2D). OS was similar when a subset of high-risk patients with del(17p) were excluded (supplemental Figure 4B).

In LOT1, ECOG PS score 0 or 1 (vs ≥2; HR, 1.45), CCI score ≤2 (vs ≥3; HR, 0.58), treatment with BR or FCR (vs other; HR, 0.67), and age <75 years (vs ≥75 years; HR, 0.48) were significantly associated with improved OS in multivariable analyses (Table 4). In LOT≥2, ECOG PS score 0 or 1 (vs ≥2; HR, 1.85), absence of del(17p) (vs presence of del[17p]; HR, 0.47), age <75 years (vs ≥75 years; HR, 0.62), and receiving treatment in an academic setting (vs community or government setting; HR, 0.34) were significantly associated with improved OS in multivariable analyses (Table 4).

Table 4.

Predictors of OS in univariate and multivariable analyses

VariableLOT1LOT≥2
Univariate analysisMultivariable analysisUnivariate analysisMultivariable analysis
HRPHR (95% CI)PHRPHR (95% CI)P
ECOG PS         
 ≤1 vs ≥2 1.82 .0003 1.45 (1.032-2.028) .0320 1.76 <.0001 1.85 (1.205-2.827) .0048 
Treatment         
 BR or FCR vs other 0.60 .0003 0.67 (0.482-0.928) .0162 0.70 .0036 — — 
CCI score         
 ≤2 vs ≥3 0.49 <.0001 0.58 (0.420-0.803) .0010 0.66 .0003 — — 
Age category, y         
 <75 vs ≥75 0.36 <.0001 0.48 (0.345-0.672) <.0001 0.53 <.0001 0.62 (0.408-0.928) .0205 
Site         
 Academic vs community/government 1.03 NS — — 0.55 .0097 0.34 (0.146-0.770) .0100 
del(17p)         
 No vs yes 0.70 NS — — 0.40 <.0001 0.47 (0.284-0.781) .0035 
Insurance         
 Other vs private 2.33 <.0001 — — 1.34 .0167 — — 
Race         
 Other vs white 1.34 NS — — 0.99 NS — — 
Sex         
 Female vs male 1.06 NS — — 0.78 .0430 — — 
Region         
 South vs West 1.04 NS — — NA NA — — 
 Midwest vs West 0.88 NS — — NA NA — — 
 Northeast vs West 0.76 NS — — NA NA — — 
CD38         
 Negative vs positive 0.77 NS — — 0.87 NS — — 
Rai stage         
 ≤1 vs ≥2 0.75 NS — — 0.88 NS — — 
Prior malignancy         
 No vs yes 0.70 .0230 — — 0.81 NS — — 
VariableLOT1LOT≥2
Univariate analysisMultivariable analysisUnivariate analysisMultivariable analysis
HRPHR (95% CI)PHRPHR (95% CI)P
ECOG PS         
 ≤1 vs ≥2 1.82 .0003 1.45 (1.032-2.028) .0320 1.76 <.0001 1.85 (1.205-2.827) .0048 
Treatment         
 BR or FCR vs other 0.60 .0003 0.67 (0.482-0.928) .0162 0.70 .0036 — — 
CCI score         
 ≤2 vs ≥3 0.49 <.0001 0.58 (0.420-0.803) .0010 0.66 .0003 — — 
Age category, y         
 <75 vs ≥75 0.36 <.0001 0.48 (0.345-0.672) <.0001 0.53 <.0001 0.62 (0.408-0.928) .0205 
Site         
 Academic vs community/government 1.03 NS — — 0.55 .0097 0.34 (0.146-0.770) .0100 
del(17p)         
 No vs yes 0.70 NS — — 0.40 <.0001 0.47 (0.284-0.781) .0035 
Insurance         
 Other vs private 2.33 <.0001 — — 1.34 .0167 — — 
Race         
 Other vs white 1.34 NS — — 0.99 NS — — 
Sex         
 Female vs male 1.06 NS — — 0.78 .0430 — — 
Region         
 South vs West 1.04 NS — — NA NA — — 
 Midwest vs West 0.88 NS — — NA NA — — 
 Northeast vs West 0.76 NS — — NA NA — — 
CD38         
 Negative vs positive 0.77 NS — — 0.87 NS — — 
Rai stage         
 ≤1 vs ≥2 0.75 NS — — 0.88 NS — — 
Prior malignancy         
 No vs yes 0.70 .0230 — — 0.81 NS — — 

NA, not available.

Prognostic modeling of early death within 2 years

Prognostic modeling was performed on 791 patients with CLL; 100 patients (12.6%) died within 2 years of enrollment, and 691 (87.4%) were followed for ≥2 years. Six variables were identified as independent predictors of early death and included in the prognostic model: Rai stage >1, race other than white, ECOG PS score >1, presence of CD38 expression, CCI score ≥3, and age >75 years. Median discriminatory power observed during validation of the model was c = 0.72 (mean, 0.72; interquartile range, 0.71-0.73). Based on magnitude of effect, age >75 years, Rai stage >1, CCI score ≥3, CD38 expression, and ECOG PS score >1 were assigned a score of 1, whereas race was assigned a score of 2. Total score was calculated for each patient, and patients were classified as low (score ≤1), medium (score ≤4), or high risk (score ≥5). After stratifying the analysis population by risk group, mortality rate was 6.2% for low-risk (n = 323), 14.5% for medium-risk (n = 373), and 27.4% for high-risk patients (n = 95; χ2P < .0001). In patients alive after 2 years, OS was found to be significantly lower in the high-risk group compared with patients in the medium- and low-risk groups (Figure 3).

Figure 3.

OS in patients who survived ≥2 years after enrollment, stratified by risk group.

Figure 3.

OS in patients who survived ≥2 years after enrollment, stratified by risk group.

Patients receiving ibrutinib

OS was significantly better in patients who received ibrutinib in the R/R setting than in those who did not. HRs for OS associated with receipt of ibrutinib during LOT2, LOT3, or LOT≥4 were 0.279 (95% CI, 0.106-0.730; P = .009), 0.441 (95% CI, 0.236-0.826; P = .011), and 0.578 (95% CI, 0.341-0.980; P = .043), respectively. Similarly, OS was improved for patients receiving ibrutinib in LOT2 vs those who received FCR or BR (HR, 0.461; 95% CI, 0.214-0.998; P = .049). A sensitivity analysis that excluded patients who received ibrutinib in LOT≥3 confirmed this result (HR, 0.321; 95% CI, 0.148-0.697; P = .0041).

Safety

Overall, 877 patients (58.7%) experienced ≥1 SAE: 449 patients (50.5%) in LOT1, 184 (70.5%) in LOT2, and 244 (70.9%) in LOT≥3 (supplemental Table 2). SAEs increased with age: 51.4% (n = 271) of patients age ≤65 years, 59.0% (n = 302) of those age 65 to 75 years, and 66.8% (n = 304) of those age ≥75 years experienced ≥1 SAE. In LOT1, 118 patients (48.0%) receiving FCR, 99 (49.7%) receiving BR, and 54 (52.4%) receiving R monotherapy experienced ≥1 SAE. The most frequent SAEs were pneumonia (11.6%), febrile neutropenia (6.2%), and pyrexia (3.8%). Overall, 802 patients (53.7%) experienced ≥1 SAE of grade ≥3.

In total, 248 patients (16.6%) experienced an SPM, most frequently squamous cell carcinoma (n = 101) and basal cell carcinoma (n = 56). The most frequent nondermatologic neoplasms were lung/bronchus cancer (n = 31), myelodysplastic syndrome (n = 17), and prostate cancer (n = 14). The most frequent infectious SAEs were pneumonia (n = 71), sepsis (n = 24), cellulitis (n = 10), and urinary tract infection (n = 10). Tumor lysis syndrome occurred in 18 patients (1.2%), and 13 patients (0.9%) experienced tumor flare reaction.

Overall, 301 patients (20.1%) died as a result of an SAE, most frequently because of neoplasm (n = 128), infection (n = 75), cardiac disorder (n = 26), or respiratory disease (n = 25). The most common fatal neoplasms were CLL (n = 89), Richter syndrome (n = 5), acute myeloid leukemia (n = 4), metastatic lung cancer (n = 3), and metastatic squamous cell carcinoma (n = 3). In the 75 patients who died as a result of infection, pneumonia (n = 34), sepsis (n = 14), and septic shock (n = 7) occurred most frequently.

Hospitalization

Overall, 412 patients (46.3%) in LOT1, 165 (63.2%) in LOT2, and 226 (65.7%) in LOT≥3 were admitted to the hospital during the study; 47 patients (9.3%) in LOT1, 35 (15.5%) in LOT2, and 53 (17.0%) in LOT≥3 died there. Fewer admissions in LOT1 (38.1%) were related to CLL or CLL treatment than in LOT2 (50.9%) or LOT≥3 (54.4%). ICU admission occurred in 68 patients (16.5%) in LOT1, 38 (23.0%) in LOT2, and 38 (16.8%) in LOT≥3. Median duration of hospitalization was 9.0 days (range, 1-107 days) in LOT1, 13.0 days (range, 1-71 days) in LOT2, and 13.0 days (range, 1-86 days) in LOT≥3.

Discussion

In this study from the Connect CLL Registry, treatment with BR or FCR was associated with improved EFS vs other treatments, including R monotherapy. In contrast, OS was similar in patients treated with FCR and BR in LOT1. As expected, EFS and OS became shorter and response rates decreased as LOT increased. In addition, ibrutinib therapy improved OS in all R/R patients regardless of the LOT in which it was received (LOT2, LOT3, or LOT≥4), even though it was approved by the FDA after the start of the Registry. Furthermore, our predictive model identified a group of patients at high risk of death within 2 years, with a mortality rate more than twice that of the overall population and 4 times higher than that of the low-risk group. This model was able to discriminate clearly between patients with a substantially increased risk of early death based on disease characteristics at enrollment in the Registry. The c-index of the predictive model (0.72) was similar to that of the CLL International Prognostic Index, which has a c-index of 0.71 for predicting OS in patients with newly diagnosed CLL.27  In a previous study from the Connect CLL Registry, a predictive model identified 3 predictors of early death from CLL or infection in older patients with CLL: time from diagnosis, enrollment therapy other than BR, and anemia. This model identified a group of older patients at significantly increased risk of death (P = .0002).28 

Prescribing patterns for FCR and BR in the Registry were broadly in line with observational results from a Surveillance, Epidemiology, and End Results (SEER)–Medicare database study that compared LOT1s of patients diagnosed in 2008 with those of patients diagnosed in 2014.29  FCR use decreased from 25.8% to 11.9%, whereas R monotherapy use remained similar (2008, 14.7%; 2014, 14.8%). BR was used by 35.5% of patients in 2014. Data for 2008 were not shown, suggesting that BR was not among the top 4 treatments reported and that <9.5% of patients used BR in 2008, because this was the lowest prevalence of medication use reported in the 2008 data.

Our data showing similar EFS in patients receiving FCR or BR in LOT1 are in contrast with PFS data from the German CLL Study Group (GCLLSG) CLL10 trial, where median PFS after FCR in LOT1 was significantly longer than after BR in LOT1 in patients without del(17p) (55.2 vs 38.5 months; P = .001). As in our study, no OS differences were observed between treatment arms.3 

Similar results showing poorer response rates with increasing LOT were observed in a SEER-Medicare study of 3214 patients newly diagnosed with CLL between 2007 and 2011, where median OS was 52.4 months in LOT1 vs 33.7 months in LOT2.30  In other studies, OS from initiation of salvage therapy ranged from 23 to 75 months.31,32  In another SEER-Medicare study of 1974 older patients diagnosed with CLL/small lymphocytic lymphoma from 1992 to 2011, a longer time from LOT1 to LOT2 was associated with better prognosis.33 

In our study, factors associated with increased EFS were absence of del(17p), age <75 years, and treatment with BR or FCR (vs other treatments). Conversely, in the GCLLSG CLL10 trial, factors predicting shorter PFS were treatment with BR (vs FCR), increased serum thymidine kinase, presence of del(11q), and unmutated IgHV genes.3  Factors associated with improved OS in our study were lower ECOG PS and CCI scores, treatment with BR or FCR (vs other treatments), and younger age in LOT1 and lower ECOG PS score, absence of del(17p), younger age, and receiving treatment in an academic setting (vs community or government setting) in LOT≥2. Similar results were seen in other studies, where older age was associated with worse OS after LOT1 and LOT2.3,30  Because CIT is more suitable for younger and fit patients, this survival disadvantage for older patients is expected.3,9,34  In addition, receiving chlorambucil monotherapy (vs R monotherapy or R-containing CIT), male sex, northeast location (vs west), receiving Medicare Part D low-income subsidies, higher National Cancer Institute comorbidity index score, and presence of disability were associated with worse OS after LOT1. Of note, presence of del(17p) was a significant predictor of OS after LOT≥2, but not in LOT1. Another SEER-Medicare database study showed that median OS for patients receiving R-containing CIT in LOT1 was similar (52 months) to that for patients receiving R monotherapy (53 months), but longer compared with that for patients receiving chemotherapy alone (34 months).20  This finding may be due to less rigorous use of FCR in the real-world setting than in clinical trials, a factor that may also influence outcomes in the Connect CLL Registry.

Our analysis demonstrated superior OS in patients treated with ibrutinib in the R/R setting compared with patients who did not receive ibrutinib. Patients who received ibrutinib in LOT2 after receiving FCR or BR in LOT1 had longer OS than patients receiving FCR or BR during LOT2, indicating that kinase inhibitors in LOT2 improved outcomes even when the most efficacious LOT1s were used. These results support the general agreement from multiple phase 2 and 3 trials that novel agents, including ibrutinib and venetoclax with or without R, are the optimal approach for patients with R/R CLL after CIT in LOT18,11  rather than recycling CIT, as was traditionally done before the introduction of novel agents. In an observational analysis of patients with CLL without del(17p), OS was similar in patients treated with ibrutinib and BR in LOT2 (3-year OS, 63% vs 74%, respectively; P = .146) after first-line CIT.31  In the phase 3 MURANO trial of patients with R/R CLL, improved survival rates were seen after therapy with venetoclax plus R vs BR, with 2-year EFS rates of 84.9% and 34.8%, respectively; 24-month OS rates were 91.9% and 86.6%, respectively.35 

Because the Registry was conducted from 2010 to 2014, no patients received ibrutinib in the frontline setting; therefore, no data on outcomes for novel therapies used in LOT1 are available. However, in a recent phase 3 study of BR, ibrutinib, or ibrutinib plus R in patients with previously untreated CLL, PFS was higher in patients treated with ibrutinib (HR, 0.39; P < .001) and ibrutinib plus R (HR, 0.38; P < .001) than in patients treated with BR.36  In another phase 3 study of patients treated with FCR or ibrutinib plus R, PFS and OS were longer in patients treated with ibrutinib plus R than FCR (HR, 0.35; P < .001 and HR, 0.17; P < .001, respectively).37  These data suggest that, although FCR and BR may still be recommended for younger, fitter patients without del(17p),9  their utility in the treatment of patients with CLL is decreasing, limiting the ability of our data to influence treatment decisions for patients initiating therapy today. However, because real-world data collected in the context of prospective registries are of increasing importance to the CLL community and the FDA, these data may serve as an important historical control, because data from more modern registries report on outcomes for patients with CLL treated with novel agents.

In the Connect CLL Registry, safety profiles were similar to those observed in other studies of FCR3,38  and BR.3,31,39  In line with clinical trial results,3  incidence of SAEs increased with increasing age and increasing LOT.

Our study has some limitations. As with all nonrandomized observational studies, care must be taken when extrapolating results to the general population. On average, 7.5 patients were included per site. Although consecutive patients presenting to the sites were evaluated for enrollment in the Registry and invited to participate, a selection bias may have occurred as physicians selected patients for enrollment. Despite the population being elderly, there were very few patients with ECOG PS ≥2. However, these data were similar to the large SEER data set, suggesting they may be illustrative of the CLL population at large.29  Cause-of-death data were limited by the information provided by each site and therefore may lack detail. There is also potential for missing data in registries; however, the Registry had the ability to query sites for more information on questionable data. Because of the lack of scheduled clinic visits, evaluation intervals may have varied between patients, potentially affecting assessment of EFS and limiting comparisons of these observational data with clinical trial results. Response assessment for patients treated in clinical studies and those treated in clinical practice may be different; unlike a clinical trial where CR is confirmed with a bone marrow assessment, confirmatory bone marrow samples for CR assessment were not required in the Registry. Finally, this study was initiated before the introduction of ibrutinib and other novel agents as LOT1s for CLL and may therefore serve as a benchmark for newer treatments. Despite these limitations, this final analysis from the Connect CLL Registry shows the value of Registry data in highlighting differences in clinical practice and outcome in community practice, where age and comorbidity are important predictors of EFS and OS vs clinical trials. The proven clinical superiority of FCR over BR3  was not replicated in this analysis, suggesting that clinical trial results may not be directly applicable to community practice. Real-world patient outcomes, including survival and AEs, vary based on treatment site, LOT, patient age, and type of treatment received. Receiving novel agents in LOT2 improved outcomes in patients who received previous CIT, supporting the inclusion of ibrutinib, idelalisib, duvelisib, and venetoclax in the treatment paradigm for patients with R/R CLL.8  However, additional randomized clinical trials and real-world data are needed to establish the role of non-CIT regimens in the treatment of patients with R/R CLL.

Bristol-Myers Squibb is committed to responsible and transparent sharing of clinical trial data with patients, health care practitioners, and independent researchers to both improve scientific and medical knowledge as well as foster innovative treatment approaches. Researchers interested in obtaining access to documents and/or data can make their requests at https://www.vivli.org.

Acknowledgments

The authors thank all the patients and their families who participated in the Connect CLL Registry. The Connect CLL Scientific Steering Committee acknowledges the contributions of all past and current members of the committee for their guidance in the design of the Registry, and participation in analysis of the data, including Matthew S. Davids, Charles M. Farber, Ian Flinn, Christopher R. Flowers, David L. Grinblatt, Neil E. Kay, Michael Keating, Thomas J. Kipps, Mark F. Kozloff, Nicole Lamanna, Susan Lerner, Anthony Mato, Chadi Nabhan, Chris L. Pashos, Jeff P. Sharman, and Mark Weiss.

The Connect CLL Registry is sponsored and funded by Bristol-Myers Squibb. The sponsor supported the authors in collecting and analyzing the data reported in this Registry. The authors received medical writing support in the preparation of this manuscript from Victoria Edwards and Nicky Dekker of Excerpta Medica BV, supported by Bristol-Myers Squibb.

Authorship

Contribution: A.M., C.N., N.L., N.E.K., D.L.G., C.R.F., C.M.F., M.S.D., and J.P.S. recruited patients to the Registry; P.K. and A.S.S. completed the statistical analyses; and all authors interpreted the data, directed the development, review, and approval of this manuscript, and are fully responsible for all content and editorial decisions.

Conflict-of-interest disclosure: A.M. has been a consultant for AbbVie, Adaptive, AstraZeneca, Celgene Corporation (including data safety monitoring board), Genentech, Johnson & Johnson, LOXO, Regeneron, Pharmacyclics, Sunesis, and TG Therapeutics (including data safety monitoring board); and has received research funding from AbbVie, Acerta, Adaptive, DTRM, Genentech, Johnson & Johnson, Regeneron, LOXO, Pharmacyclics, Portola, and TG Therapeutics. C.N. is employed by Aptitude Health. N.L. has received research funding from AbbVie, AstraZeneca, BeiGene, Genentech, Gilead, Infinity Pharma, Juno, Ming, ProNAi, and TG Therapeutics; been a consultant for AbbVie, AstraZeneca, BeiGene, Genentech, Gilead, Juno, ProNAi, and Pharmacyclics; and served on an advisory committee for Celgene Corporation. N.E.K. has received research funding from Gilead, Celgene Corporation, Hospira, Genentech, and Pharmacyclics and been on advisory committees for Gilead and Celgene Corporation. D.L.G. has been a consultant and member of a speaker’s bureau for Celgene Corporation. C.R.F. has received research funding from Gilead, Spectrum, Millennium, Janssen, Infinity Pharma, AbbVie, Acerta, Pharmacyclics, and TG Therapeutics; served as a consultant for Celgene Corporation, OptumRx, Gilead, Seattle Genetics, Millennium, and Genentech/Roche; and received honoraria from Celgene Corporation. C.M.F. has received research funding from Genentech and Gilead; been a consultant and part of a speaker’s bureau for Celgene Corporation, Gilead, Janssen, Pharmacyclics, Seattle Genetics, and Genentech; served on an advisory committee for Celgene Corporation; and received honoraria from Janssen. M.S.D. has served as a consultant for AbbVie, Acerta Pharma, Adaptive Biotechnologies, AstraZeneca, Genentech, Gilead, Janssen, MEI Pharma, Pharmacyclics, Research to Practice, Syros Pharmaceuticals, TG Therapeutics, and Verastem and received research funding from Acerta Pharma, Ascentage Pharma, Bristol-Myers Squibb, Genentech, MEI Pharma, Pharmacyclics, Surface Oncology, TG Therapeutics, and Verastem. A.S.S., K.S., E.D.F., M.G., and P.K. are employees of Bristol-Myers Squibb and have equity. S.M.G.U. is a former employee of Celgene Corporation (now Bristol-Myers Squibb). J.P.S. has received honoraria from Genentech, Gilead, and TG Therapeutics; been a consultant for Celgene Corporation, Genentech, Gilead, and Pharmacyclics; been a member of a speaker’s bureau for Gilead; received research funding from Celgene Corporation, Genentech, Gilead, Pharmacyclics, TG Therapeutics, Seattle Genetics, and Acerta; and received travel expenses from Celgene Corporation and Gilead.

Correspondence: Anthony Mato, CLL Program, Leukemia Service, Memorial Sloan Kettering Cancer Center, 1275 York Ave, New York, NY 10065; e-mail: matoa@mskcc.org.

References

References
1.
Siegel
RL
,
Miller
KD
,
Jemal
A
.
Cancer statistics, 2018
.
CA Cancer J Clin
.
2018
;
68
(
1
):
7
-
30
.
2.
Byrd
JC
,
Ruppert
AS
,
Heerema
NA
, et al
.
Lenalidomide consolidation benefits patients with CLL receiving chemoimmunotherapy: results for CALGB 10404 (Alliance)
.
Blood Adv
.
2018
;
2
(
14
):
1705
-
1718
.
3.
Eichhorst
B
,
Fink
AM
,
Bahlo
J
, et al;
German CLL Study Group (GCLLSG)
.
First-line chemoimmunotherapy with bendamustine and rituximab versus fludarabine, cyclophosphamide, and rituximab in patients with advanced chronic lymphocytic leukaemia (CLL10): an international, open-label, randomised, phase 3, non-inferiority trial
.
Lancet Oncol
.
2016
;
17
(
7
):
928
-
942
.
4.
Thompson
PA
,
Tam
CS
,
O’Brien
SM
, et al
.
Fludarabine, cyclophosphamide, and rituximab treatment achieves long-term disease-free survival in IGHV-mutated chronic lymphocytic leukemia
.
Blood
.
2016
;
127
(
3
):
303
-
309
.
5.
Hallek
M
,
Fischer
K
,
Fingerle-Rowson
G
, et al;
German Chronic Lymphocytic Leukaemia Study Group
.
Addition of rituximab to fludarabine and cyclophosphamide in patients with chronic lymphocytic leukaemia: a randomised, open-label, phase 3 trial
.
Lancet
.
2010
;
376
(
9747
):
1164
-
1174
.
6.
Michallet
A-S
,
Aktan
M
,
Hiddemann
W
, et al
.
Rituximab plus bendamustine or chlorambucil for chronic lymphocytic leukemia: primary analysis of the randomized, open-label MABLE study
.
Haematologica
.
2018
;
103
(
4
):
698
-
706
.
7.
Woyach
JA
,
Ruppert
AS
,
Heerema
NA
, et al
.
Ibrutinib alone or in combination with rituximab produces superior progression free survival (PFS) compared with bendamustine plus rituximab in untreated older patients with chronic lymphocytic leukemia (CLL): results of Alliance North American Intergroup Study A041202 [abstract]
.
Blood
.
2018
;
132
(
suppl 1
).
Abstract 6
.
8.
Yosifov
DY
,
Wolf
C
,
Stilgenbauer
S
,
Mertens
D
.
From biology to therapy: the CLL success story
.
Hemasphere
.
2019
;
3
(
2
):
e175
.
9.
National Comprehensive Cancer Network, Inc.
NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®) for Chronic Lymphocytic Leukemia/Small Lymphocytic Leukemia, version 4.
2020
. Plymouth Meeting, PA: National Comprehensive Cancer Network, Inc; 2020. https://www.nccn.org/professionals/physician_gls/pdf/cll.pdf. Accessed 6 January 2020. © National Comprehensive Cancer Network, Inc. 2019. All rights reserved. Referenced with permission. To view the most recent and complete version of the guideline, go online to NCCN.org. NCCN makes no warranties of any kind whatsoever regarding their content, use or application and disclaims any responsibility for their application or use in any way.
10.
Eichhorst
B
,
Robak
T
,
Montserrat
E
, et al;
ESMO Guidelines Committee
.
Chronic lymphocytic leukaemia: ESMO clinical practice guidelines for diagnosis, treatment and follow-up
.
Ann Oncol
.
2015
;
26
(
suppl 5
):
v78
-
v84
.
11.
Jain
N
,
Thompson
P
,
Ferrajoli
A
,
Nabhan
C
,
Mato
AR
,
O’Brien
S
.
Approaches to chronic lymphocytic leukemia therapy in the era of new agents: the conundrum of many options
.
Am Soc Clin Oncol Educ Book
.
2018
;
38
(
38
):
580
-
591
.
12.
Mansfield
C
,
Masaquel
A
,
Sutphin
J
, et al
.
Patients’ priorities in selecting chronic lymphocytic leukemia treatments
.
Blood Adv
.
2017
;
1
(
24
):
2176
-
2185
.
13.
Hallek
M
.
Chronic lymphocytic leukemia: 2020 update on diagnosis, risk stratification and treatment
.
Am J Hematol
.
2019
;
94
(
11
):
1266
-
1287
.
14.
Brown
JR
,
Hallek
MJ
,
Pagel
JM
.
Chemoimmunotherapy versus targeted treatment in chronic lymphocytic leukemia: when, how long, how much, and in which combination?
Am Soc Clin Oncol Educ Book
.
2016
;
35
:
e387
-
e398
.
15.
Sarraf Yazdy
M
,
Mato
AR
,
Cheson
BD
.
Combinations or sequences of targeted agents in CLL: is the whole greater than the sum of its parts (Aristotle, 360 BC)?
Blood
.
2019
;
133
(
2
):
121
-
129
.
16.
Abel
GA
.
The real world: CLL
.
Blood
.
2011
;
117
(
13
):
3481
-
3482
.
17.
National Cancer Institute Surveillance, Epidemiology, and End Results. Cancer stat facts: leukemia–chronic lymphocytic leukemia. https://seer.cancer.gov/statfacts/html/clyl.html. Accessed 11 December 2018.
18.
Flowers
CR
,
Nabhan
C
,
Kay
NE
, et al
.
Reasons for initiation of treatment and predictors of response for patients with Rai stage 0/1 chronic lymphocytic leukemia (CLL) receiving first-line therapy: an analysis of the Connect® CLL cohort study
.
Leuk Lymphoma
.
2018
;
59
(
10
):
2327
-
2335
.
19.
Shenkel
B
,
Rider
A
,
Macomson
B
,
Hallworth
P
.
Real-world treatment patterns and characteristics in patients with chronic lymphocytic leukemia [abstract]
.
J Clin Oncol
.
2017
;
35
(
suppl 8
).
Abstract 65
.
20.
Danese
MD
,
Griffiths
RI
,
Gleeson
M
, et al
.
An observational study of outcomes after initial infused therapy in Medicare patients diagnosed with chronic lymphocytic leukemia
.
Blood
.
2011
;
117
(
13
):
3505
-
3513
.
21.
Mato
A
,
Nabhan
C
,
Kay
NE
, et al
.
Real-world clinical experience in the Connect® chronic lymphocytic leukaemia registry: a prospective cohort study of 1494 patients across 199 US centres
.
Br J Haematol
.
2016
;
175
(
5
):
892
-
903
.
22.
Hallek
M
,
Cheson
BD
,
Catovsky
D
, et al;
International Workshop on Chronic Lymphocytic Leukemia
.
Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines
.
Blood
.
2008
;
111
(
12
):
5446
-
5456
.
23.
University of Bern
.
STROBE statement: Checklist of items that should be included in reports of cohort studies
. https://www.strobe-statement.org/fileadmin/Strobe/uploads/checklists/STROBE_checklist_v4_cohort.pdf.
Accessed 11 December 2018
.
24.
Austin
PC
,
Tu
JV
.
Bootstrap methods for developing predictive models
.
Am Stat
.
2004
;
58
(
2
):
131
-
137
.
25.
Harrell
FE
Jr.
Regression modeling strategies
.
New York
:
Springer-Verlag
;
2001
.
26.
Sullivan
LM
,
Massaro
JM
,
D’Agostino
RB
Sr.
Presentation of multivariate data for clinical use: The Framingham Study risk score functions
.
Stat Med
.
2004
;
23
(
10
):
1631
-
1660
.
27.
Gentile
M
,
Shanafelt
TD
,
Rossi
D
, et al
.
Validation of the CLL-IPI and comparison with the MDACC prognostic index in newly diagnosed patients
.
Blood
.
2016
;
128
(
16
):
2093
-
2095
.
28.
Nabhan
C
,
Mato
A
,
Flowers
CR
, et al
.
Characterizing and prognosticating chronic lymphocytic leukemia in the elderly: prospective evaluation on 455 patients treated in the United States
.
BMC Cancer
.
2017
;
17
(
1
):
198
.
29.
Seymour
EK
,
Ruterbusch
JJ
,
Beebe-Dimmer
JL
,
Schiffer
CA
.
Real-world testing and treatment patterns in chronic lymphocytic leukemia: A SEER patterns of care analysis
.
Cancer
.
2019
;
125
(
1
):
135
-
143
.
30.
Mato
A
,
Jahnke
J
,
Li
P
, et al
.
Real-world treatment and outcomes among older adults with chronic lymphocytic leukemia before the novel agents era
.
Haematologica
.
2018
;
103
(
10
):
e462
-
e465
.
31.
Cuneo
A
,
Follows
G
,
Rigolin
GM
, et al;
GIMEMA, European Research Initiative on CLL (ERIC) and UK CLL forum
.
Efficacy of bendamustine and rituximab as first salvage treatment in chronic lymphocytic leukemia and indirect comparison with ibrutinib: a GIMEMA, ERIC and UK CLL FORUM study
.
Haematologica
.
2018
;
103
(
7
):
1209
-
1217
.
32.
Joffe
E
,
Goldschmidt
N
,
Bairey
O
, et al;
Israeli CLL Study Group
.
Outcomes of second-line treatment after fludarabine cyclophosphamide and rituximab in patients with chronic lymphocytic leukemia outside clinical trials
.
Eur J Haematol
.
2018
;
101
(
3
):
399
-
406
.
33.
Ammann
EM
,
Shanafelt
TD
,
Larson
MC
, et al
.
Time to second-line treatment and subsequent relative survival in older patients with relapsed chronic lymphocytic leukemia/small lymphocytic lymphoma
.
Clin Lymphoma Myeloma Leuk
.
2017
;
17
(
12
):
e11
-
e25
.
34.
Stauder
R
,
Eichhorst
B
,
Hamaker
ME
, et al
.
Management of chronic lymphocytic leukemia (CLL) in the elderly: a position paper from an international Society of Geriatric Oncology (SIOG) Task Force
.
Ann Oncol
.
2017
;
28
(
2
):
218
-
227
.
35.
Seymour
JF
,
Kipps
TJ
,
Eichhorst
B
, et al
.
Venetoclax-rituximab in relapsed or refractory chronic lymphocytic leukemia
.
N Engl J Med
.
2018
;
378
(
12
):
1107
-
1120
.
36.
Woyach
JA
,
Ruppert
AS
,
Heerema
NA
, et al
.
Ibrutinib regimens versus chemoimmunotherapy in older patients with untreated CLL
.
N Engl J Med
.
2018
;
379
(
26
):
2517
-
2528
.
37.
Shanafelt
TD
,
Wang
XV
,
Kay
NE
, et al
.
Ibrutinib-rituximab or chemoimmunotherapy for chronic lymphocytic leukemia
.
N Engl J Med
.
2019
;
381
(
5
):
432
-
443
.
38.
Fischer
K
,
Bahlo
J
,
Fink
AM
, et al
.
Long-term remissions after FCR chemoimmunotherapy in previously untreated patients with CLL: updated results of the CLL8 trial
.
Blood
.
2016
;
127
(
2
):
208
-
215
.
39.
Fischer
K
,
Cramer
P
,
Busch
R
, et al
.
Bendamustine combined with rituximab in patients with relapsed and/or refractory chronic lymphocytic leukemia: a multicenter phase II trial of the German Chronic Lymphocytic Leukemia Study Group
.
J Clin Oncol
.
2011
;
29
(
26
):
3559
-
3566
.

Author notes

The full-text version of this article contains a data supplement.

Supplemental data