Page 164 - Haematologica - Vol. 105 n. 6 - June 2020

E.C. Rotbain et al.
bined with anti-CD20 antibodies (CD20-chlorambucil), is recommended for unfit patients with significant comor- bidity.8 Patients with del(17p)/TP53 mutations are treated with targeted agents (ibrutinib, idelalisib-rituximab or venetoclax).9,10 The Danish guidelines are updated biannu- ally and the changes over time have been described previ- ously.3
Immunoglobulin heavy-chain variable region gene
(IGHV) mutational status is an acknowledged prognostic
factor in CLL and is included in the disease-specific
International Prognostic Index (CLL-IPI).11,12 In previous
studies, patients with unmutated IGHV (U-CLL) had
shorter survival from diagnosis compared with patients
with mutated IGHV (M-CLL), and inferior remission dura-
tion and survival from the start of chemoimmunothera- py.5,6,12-18
We present data on the impact of IGHV mutational sta- tus on overall survival (OS) and treatment-free survival (TFS) from the time of treatment in the world’s largest, nationwide, population-based cohort of consecutive, un- selected patients with CLL receiving different treatment regimens.
Methods
Data sources and study population
The Danish CLL registry contains data on all patients diag- nosed with CLL in Denmark since 2008.19 As of August 2017, the registry contained information on 4,135 CLL patients, who were included in the present study (Figure 1A). The CLL registry con- tains data on sex, dates of birth, diagnosis, and treatment, type of treatment, IGHV mutational status, and other disease charac- teristics including cytogenetics, TP53 mutations, and β2- microglobulin levels at the time of diagnosis. Information on vital status is included in the CLL registry through regular link- age with the Danish Civil Registration System.20,21 Patients with missing data regarding key variables were excluded from the study. Patients were followed from the date of diagnosis in 2008-2017, until the time of death, emigration, or August 2017, whichever came first. All treatments of minimum one series were considered. For the subset of patients who had received first-line treatment at Odense University Hospital, in the Capital Region, or in the Zealand Region between 2008-2016, detailed information on second-line treatment was collected through review of the patients’ clinical records. Together, these regions cover over half of the Danish population. These patients were followed from the date of diagnosis in 2008-2016, until the time of death, emigration, or mid-2018 (ranging from May- November, depending on the date of the patients’ record review), whichever came first.
Statistical analysis
The patients’ characteristics are reported for those with U- CLL and M-CLL and for treatment groups, and compared using parametric or non-parametric descriptive statistics, depending on the data distribution. Kaplan-Meier survival analyses were used to assess survival. TFS from the time of diagnosis (TFSd) was defined as the time to first treatment, end of follow-up, or death¸ whichever came first. OS was determined starting from either the time of diagnosis (OSd) or the time of first-line treat- ment initiation (OSt), until death, or end of follow-up, whichev- er came first. TFSt, defined as the time from initiation of first-line treatment to initiation of second-line treatment, death or end of follow-up, whichever came first, was studied for the sub-popu-
lation with detailed information on second-line treatment from medical record review. OSt and TFSt were the primary endpoints of the study, while OSd and TFSd were secondary endpoints. We explored the prognostic significance of IGHV status, treatment regimen, del(17p) status, elevated β2-microglobulin level, sex, age, and Binet stage for risk of death or treatment, using multi- variable Cox regression models to calculate hazard ratios (HR). All HR presented have been adjusted for these variables, except for TFSt, which was adjusted for sex, age, del(17)p/TP53-muta- tion, and Binet stage. Unadjusted HR were calculated but are not presented in this paper as they were not of clinical relevance. Log-rank tests were used to test for homogeneity of outcomes between exposures. Data analysis was performed using STATA (StataCorp. 2015. Stata Statistical Software: Release 15.1 College Station: StataCorp LP, TX, USA)
Ethics
The study was approved by the Danish Health and Medicine Authorities (jr. n. 3-3013-1141/1) and the Danish Data Protection Agency (jr. n. RH-2015-96 03856). Results for subgroups includ- ing fewer than five patients were reported as “less than five” to ensure anonymity of individual patients, in accordance with Danish legislation.
Results
Characteristics at time of diagnosis of chronic lympho- cytic leukemia
In total, 4,135 patients with a median follow-up time of 3.5 years were available for analysis, of whom two were excluded because of incomplete data. Information on IGHV mutational status was available for 3,197 (77%) patients, of whom 1,017 (32%) had U-CLL and 2,180 (68%) had M-CLL (Figure 1A). The characteristics of the patients, divided according to mutational status, are listed in Table 1. Among patients with unknown IGHV status, 255 (27%) received treatment during follow-up, compared with 481 (47%) of U-CLL and 369 (17%) of M-CLL patients. Distributions of sex and age at diagnosis were comparable between U-CLL and M-CLL patients, whereas prognostic factors were unevenly distributed, with del(13q) found in 28% of U-CLL patients and 53% of M- CLL patients. The prevalences of del(17p) (7%), del(11q) (16%) and trisomy(12) (16%) were higher among U-CLL patients than among M-CLL patients (4%, 2% and 11%, respectively). Of the U-CLL patients, 30% were catego- rized as having Binet stage B/C, compared with 13% of M- CLL patients, and 19% had a high level of β2-microglobulin (>4.0 mg/L), compared with 10% of M-CLL patients.
IGHV status and prognosis from the time of diagnosis
Patients with unmutated IGHV had shorter OSd [HR=1.23, 95% confidence interval (95% CI): 1.01-1.50], compared with patients with M-CLL, and shorter TFSd (HR=2.24, 95% CI: 1.95-2.57) (Figure 2A, B). The 5-year OSd was 71% (95% CI: 68-74) for U-CLL patients and 81% (95% CI: 79-83) for those with M-CLL. The 5-year TFSd for U-CLL patients was 31% (95% CI: 27-35), com- pared with 68% (95% CI: 65-70) for those with M-CLL. Patients with unknown IGHV status had a shorter 5-year OSd (61%, 95% CI: 57-64) than patients with U-CLL or M-CLL, while the 5-year TFSd (45%, 95% CI: 41-49) in this group was between that of U-CLL and M-CLL patients (data not shown). Overall, 92 (9%) patients with U-
1622
haematologica | 2020; 105(6)