Worldwide Examination of Patients with CLL Hospitalized for COVID-19

Introduction: Patients (pts) with CLL may be at particular risk of severe COVID-19 given advanced age and immune dysregulation. Two large series with limited follow-up have reported outcomes for pts with CLL and COVID-19 (Scarfò, et al. Leukemia 2020; Mato, et al. Blood 2020). To provide maximal clarity on outcomes for pts with CLL and COVID-19, we partnered in a worldwide effort to describe the clinical experience and validate predictors of survival, including potential treatment effects.

Methods: This international collaboration represents a partnership between investigators at 141 centers. Data are presented in two cohorts. Cohort 1 (Co1) includes pts captured through efforts by European Research Initiative on CLL (ERIC), Italian CAMPUS CLL Program, and Grupo Español de Leucemia Linfática Crónica. The validation cohort, Cohort 2 (Co2), includes pts from US (66%), UK (23%), EU (7%), and other countries (4%). There is no overlap in cases between cohorts.

CLL pts were included if COVID-19 was diagnosed by PCR detection of SARS-CoV-2 and they required inpatient hospitalization. Data were collected retrospectively 2/2020 - 5/2020 using standardized case report forms. Baseline characteristics, preexisting comorbidities (including cumulative illness rating scale (CIRS) score ≥6 vs. <6), CLL treatment history, details regarding COVID-19 course, management, and therapy, and vital status were collected.

The primary endpoint of this study was to estimate the case fatality rate (CFR), defined as the proportion of pts who died among all pts hospitalized with COVID-19. Chi-squared test was used to compare frequencies; univariable and multivariable analyses utilized Cox regression. Predictors of inferior OS in both Co1 and Co2 were included in multivariable analyses. Kaplan-Meier method was used to estimate overall survival (OS) from time of COVID-19 diagnosis (dx).

Results: 411 hospitalized, COVID-19 positive CLL pts were analyzed (Co1 n=281, Co2 n=130). Table 1 describes baseline characteristics. At COVID-19 dx, median age was 72 in Co1 (range 37-94) and 68 in Co2 (range 41-98); 31% (Co1) and 45% (Co2) had CIRS ≥6. In Co1, 48% were treatment-naïve and 26% were receiving CLL-directed therapy at COVID-19 dx (66% BTKi ± anti-CD20, 19% Venetoclax ± anti-CD20, 9.6% chemo/chemoimmunotherapy (CIT), 1.4% PI3Ki, 4% other). In Co2, 36% were never treated and 49% were receiving CLL-directed therapy (65% BTKi ± anti-CD20, 19% Venetoclax ± anti-CD20, 9.4% multi-novel agent combinations, 1.6% CIT, 1.6% PI3Ki, 1.6% anti-CD20 monotherapy, 1.6% other). Most pts receiving CLL-directed therapy had it held at COVID-19 diagnosis (93% in Co1 and 81% in Co2).

Frequency of most COVID-19 symptoms/laboratory abnormalities were similar in the two cohorts including fever (88% in both), lymphocytosis (ALC ≥30 x 10⁹/L; 27% vs. 21%), and lymphocytopenia (ALC < 1.0 x 10⁹/L; 18% vs. 28%), while others varied between Co1 and Co2 (p<0.0001), including cough (61% vs. 93%), dyspnea (60% vs. 84%), fatigue (13% vs. 77%).

Median follow-up was 24 days (range 2-86) in Co1 and 17 days (1-43) in Co2. CFRs were similar in Co1 and Co2, 30% and 34% (p=0.45). 54% and 43% were discharged while 16% and 23% remained admitted at last follow-up in Co1 and Co2, respectively. The proportion of pts requiring supplemental oxygen was similar (89% vs. 92%) while rate of ICU admission was higher in Co2 (20% vs. 48%, p<0.0001). Figure 1 depicts OS in each cohort. Univariable analyses demonstrated that age and CIRS ≥6 significantly predicted inferior OS in both cohorts, while only age remained an independent predictor of inferior OS in multivariable analyses (Table 2). Prior treatment for CLL (vs. observation) predicted inferior OS in Co1 but not Co2.

Conclusions: In the largest cancer dx-specific cohort reported, pts with CLL hospitalized for COVID-19 had a CFR of 30-34%. Advanced patient age at COVID-19 diagnosis was an independent predictor of OS in two large cohorts. This CFR will serve as a benchmark for mortality for future outcomes studies, including therapeutic interventions for COVID-19 in this population. The effect of CLL treatment on OS was inconsistent across cohorts; COVID-19 may be severe regardless of treatment status. While there were no significant differences in distribution of current lines of therapy between cohorts, prior chemo exposure was more common in Co1 vs. Co2, which may account for difference in OS. Extended follow-up will be presented.