Cause of death in patients with newly diagnosed chronic lymphocytic leukemia (CLL) stratified by the CLL-International Prognostic Index

8026 Background: CLL progression and CLL-related complications (infections and second malignancies) were the leading cause of death (COD) in a prospective cohort of CLL patients (Strati, BJH 2017). The CLL-IPI integrates major clinical and molecular prognostic factors and stratifies patients into 4 risk groups with distinct prognosis. It is unknown if COD differs according to CLL-IPI risk group in patients with newly diagnosed CLL. Methods: Patients diagnosed with CLL between 1/2000-12/2019 and seen within 1 year of diagnosis were identified from the Mayo Clinic CLL database. Cumulative incidences of cause-specific death were analyzed using Gray’s test, with deaths from different causes treated as competing events and deaths from unknown causes excluded. Results: 1276 patients were included in this study. The median age at diagnosis was 63 years (range 24-92), and 880 (69%) were male. Based on CLL-IPI score, 449 (35%) had low risk disease, 443 (35%) had intermediate risk disease, and 384 (30%) had high/very high risk disease. Median follow-up time for the study was 6 years; 286 deaths occurred. The COD was CLL progression in 99 (35%), infection in 16 (6%), second malignancy in 47 (16%), CLL-unrelated in 59 (21%), and unknown in 65 (23%) patients. The rates of death due to CLL progression were higher (17.3% at 5 years; 30.3% at 10 years) than the rates due to CLL-related complications (5.7% at 5 years; 12.9% at 10 years) or due to CLL-unrelated causes (8.6% at 5 years; 16.9% at 10 years) in the CLL-IPI high/very high risk group, but not the CLL-IPI low or intermediate risk group (Table). A higher CLL-IPI risk group was associated with a higher rate of death due to CLL progression ( P < 0.001), as well as a higher rate of death due to CLL-related complications ( P = 0.013), and CLL-unrelated causes ( P < 0.001). Conclusions: Causes of death in newly diagnosed CLL patients differ according to their CLL-IPI risk group. In patients with high/very high risk CLL, improving CLL disease control with novel agents seems justified. In patients with low/intermediate risk CLL, there should be increased efforts to reverse immune dysfunction to reduce infections and second malignancies. [Table: see text]


Dear Editor,
Chronic lymphocytic leukemia (CLL) is an indolent B-cell lymphoproliferative disorder that is considered incurable [1]. The clinical presentation of CLL is heterogeneous and the rate of progression is variable. Many patients have an indolent disease course and can be observed for years, while a subset of patients experience rapid disease progression with symptoms and/or bone marrow failure requiring treatment. A recent advance in risk stratification has been the CLL-International Prognostic Index (CLL-IPI), which integrates major clinical and molecular prognostic factors into a single risk score. Specifically, CLL-IPI uses age, Binet or Rai stage, beta-2 microglobulin, immunoglobulin heavy-chain gene variable region (IGHV) mutation status, and TP53 disruption status (17p deletion or somatic TP53 mutation) to stratify patients with newly diagnosed CLL into four risk groups with distinct prognosis [2]. Subsequent studies have shown that CLL-IPI can predict time to first treatment (TTFT) and overall survival (OS) in newly diagnosed CLL [2][3][4][5]. Although CLL primarily affects older adults in whom comorbidities can be competing causes of mortality, we previously showed that CLL progression and CLL-related complications (including infections and second malignancies) were the leading causes of death in patients with newly diagnosed CLL, regardless of age and comorbidities [6]. In this study, we sought to investigate if the causes of death in patients with newly diagnosed CLL would differ according to the CLL-IPI risk group.
This study was approved by the Mayo Clinic Institutional Review Board. Patients diagnosed with CLL between January 1, 2000 and December 26, 2019 were identified from the Mayo Clinic CLL Database, which includes consecutive CLL patients who were evaluated within one year of diagnosis in the Division of Hematology at Mayo Clinic, Rochester, MN [7]. All patients provided an informed consent form. Clinical characteristics, follow-up data, and survival outcomes were abstracted from medical record. Cause of death was classified into one of four categories: CLL progression, CLL-related complications (infection or second malignancy), CLL-unrelated, and unknown, consistent with our previously published approach [6]. Cumulative incidences of cause-specific death were analyzed using Gray's test, with deaths from different causes treated as competing events and deaths from unknown causes excluded. All P values were twosided and considered significant if <0.05. Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).
Our data suggest that the cause of death in patients with newly diagnosed CLL differs according to their CLL-IPI risk group. In patients with CLL-IPI low or intermediate-risk disease, the risk of dying from CLL progression is similar to that of dying from CLLrelated complications (including infections and second malignancies) or CLL-unrelated causes. However, patients with CLL-IPI high-or very-high-risk disease have a~3-fold higher risk of dying from CLL progression or associated complications such as infection or second malignancy compared to non-CLL-related reasons. Similar result trends were observed in follicular lymphoma (FL), another indolent B-cell malignancy, where risk of FL-related death was higher than risk of FL-unrelated death in patients with higher FLIPI (2 or 3-5) but not low FLIPI (0-1), highlighting the importance of prognostic index in informing the risk of death from different causes [8].
Results of our study have several implications in clinical practice. First, given the value of CLL-IPI in predicting TTFT, OS, and cause of death, diagnostic tests that determine the CLL-IPI risk group should be strongly considered in all patients with newly diagnosed CLL. From a patient counseling standpoint, it is important to discuss the heterogeneity of clinical outcomes and provide individualized counseling regarding survival. Historically, many patients were told that they were "more likely to die with CLL than to die of CLL". Our data suggest that this is not true for patients with CLL-IPI high-or very-high-risk disease, which represent approximately one third of all newly diagnosed CLL patients in our study. Second, the increased risk of dying from CLL disease progression in CLL-IPI high-/very-high-risk groups suggest that improving CLL disease control is of utmost importance in these patients. Early intervention trials using novel targeted agents such as the CLL12 trial comparing ibrutinib to placebo [9] or the EVOLVE study (NCT04269902) comparing early vs. delayed venetoclax and obinutuzumab seem prescient in patients with high-and very-high-CLL-IPI risk groups [5]. Third, as 63/221 (29%) of deaths with a known cause were due to infections or second malignancies, there should be increased efforts to reduce the risks of the same, e.g., diligence in routine vaccinations and age-appropriate cancer screenings. Understanding the mechanisms of immune dysfunction and then developing strategies to restore more normal immune system in CLL patients are critical to improving their clinical outcomes. Finally, data from this study can be used in publicly funded health systems to guide decisions about obtaining molecular studies in patients with newly diagnosed CLL. Identifying patients with high-and very-high-risk CLL at the time of diagnosis may allow better counselling for such patients, an enhanced surveillance program for detection of progressive disease, and potential enrollment in early intervention clinical trials. On the other hand, testing to ascertain the CLL-IPI may provide reassurance to patients who are found to have low-and intermediate-risk disease.
The strengths of this study include a large prospectively followed cohort of consecutive CLL patients, comprehensive clinical information, follow-up duration, and competing risk analysis. The limitations include heterogenous therapies,~2decade time interval during which patients were studied when there were significant changes in treatments, and unknown cause of death in~20% of the patients.
In summary, the CLL-IPI risk score has an important role in predicting causes of death in newly diagnosed CLL and has