Clonal hematopoiesis of indeterminate potential (CHIP) and risk of non‐Hodgkin lymphoma: A community‐based cohort study

Clonal hematopoiesis of indeterminate potential (CHIP), characterized by the expansion of hematopoietic cell clones carrying somatic mutations,1, 2 is considered as a risk factor for various diseases, especially hematological malignancies.3-6 In contrast to the well-established association between CHIP and myeloid malignancies, large knowledge gaps exist for the association between CHIP and the risk of lymphoid malignancy.7 Non-Hodgkin lymphoma (NHL) is a heterogeneous group of hematopoietic malignancies and accounts for about 90% of all lymphomas, which arise from the malignant transformation of lymphocytes.8 Although several factors (e.g., immune disorders, infections, medicines, genetics, and lifestyle) have been suggested as individual risk factors of NHL, a significant proportion of NHL patients cannot be attributed to provoked or known risk factors.9 Lymphoid CHIP has been revealed to be associated with a higher incidence of lymphoid malignancies;10 however, the majority of CHIP is myeloid, and the relationship between CHIP and NHL remains to be elucidated. Thus, we conducted a large-scale community-based cohort study using data from the UK Biobank and investigated whether CHIP is a risk factor for NHL. CHIP was ascertained by analyzing whole-exome sequencing (WES) data. We used the Cox regression model to estimate hazard ratios (HRs) with 95% confidence intervals (CIs) of NHL associated with CHIP. The detailed methods can be seen in the supplementary information, Figure S1, and Table S1. Among the 437,219 participants from the UK Biobank, we identified 13,068 individuals with CHIP and 424,151 without CHIP. The demographic and clinical characteristics of the cohort are summarized in Table S2. Individuals with CHIP were older, with a median age of 62 years, compared to those without CHIP, who had a median age of 57 years. Furthermore, participants with CHIP were more inclined to be male, of White ethnicity, overweight, and have a history of smoking (either previously or currently) compared to those without CHIP. The median follow-up was 12.6 years for individuals without CHIP and 12.8 years for those with CHIP. Among individuals with CHIP, there were 150 incident cases of any NHL (incidence rate [IR], 93.22 per 100,000 person-years), while 2653 cases (IR, 48.35 per 100,000 person-years) were identified among individuals without CHIP (i.e., reference). Individuals with CHIP had a higher risk of any NHL than those in the non-CHIP group (hazard ratio [HR]: 1.52, 95% confidence interval [CI]: 1.30–1.78) (Table 1). In the analysis by subtypes of NHL (Table 1), we found heightened risks in relation to CHIP for most studied NHL subtypes, especially for CLL/SLL (HR: 1.64, 95% CI: 1.21–2.21), FL (HR: 1.63, 95% CI: 1.06–2.51), and mature T/NK-cell lymphomas (HR: 2.79, 95% CI: 1.57–4.96). Focusing on the specific CHIP gene mutations, we revealed that CHIP individuals with different mutations also show general increased risks for NHL. CHIP harboring TET2 mutations (HR: 1.76, 95% CI: 1.30–2.40) or other mutations (HR: 2.37, 95% CI: 1.73–3.24) significantly enhanced the risk of NHL. Additionally, CHIP with mutations in DNMT3A (HR: 1.22, 95% CI: 0.95–1.58) was also suggested as an individual risk factor for NHL, although no statistical significance was found. We analyzed the variant allele fractions (VAFs) of the mutant genes and observed that the VAFs are different among the mutant genes (Table S3). Furthermore, our analysis stratified by the VAF of CHIP revealed elevated risks for NHL in both CHIP cases with a VAF of ≥0.1 (HR: 1.55, 95% CI: 1.27–1.88) and those with a VAF of <0.1 (HR: 1.38, 95% CI: 1.02–1.88) (Figure 1). We detected potential modifiers of the associations between CHIP and NHL, through subgroup analyses stratified by sex (female or male), age (less than 60 years or 60 years and older), ethnicity (white, other, or unknown), Townsend deprivation index quartiles (Q1, Q2, Q3, or Q4), educational attainment (college or university degree, others, or unknown), BMI categories (less than 18.5, 18.5 to less than 25, 25 to less than 30, or 30 and higher), and smoking status (never smoked, previously smoked, currently smoking, or unknown). The overall association between CHIP and NHL remained largely unchanged across the above subgroups (Table S4). We also conducted a sensitivity analysis by excluding the overlapping genes between CHIP and lymphoid malignancy from a comprehensive summary,11 to mitigate any interference these genes might have on the assessment of NHL risk. The results (Table S5) showed that the overall correlation between CHIP and NHL remained largely stable even after excluding CHIP cases harboring driver mutations associated with lymphoid malignancy (HR: 1.30, 95% CI: 1.06–1.59, p = 0.01). To further evaluate the robustness of our findings against potential reverse causality, we conducted a sensitivity analysis incorporating a 3- and 5-year lag period of follow-up after blood sampling. The analysis revealed largely comparable results for any NHL when using either a 3-year (HR: 1.46, 95% CI: 1.22–1.75) or a 5-year (HR: 1.50, 95% CI: 1.23–1.83) lag time in the analysis (Table S6). Given that lymphoid malignancies and myeloid malignancies constitute two distinct categories of hematologic malignancies and that CHIP has been implicated as a risk factor for myeloid malignancy,3 we conducted a stratified analysis focusing on the association between myeloid malignancy and NHL. Positive correlations were observed between myeloid malignancy and NHL, not only in the CHIP group (odds ratio [OR]: 1.82, 95% CI: 1.05–3.16, p = 0.03) but also in the non-CHIP group (OR: 1.88, 95% CI: 1.26–2.80, p = 0.002). (Table S7). Since myeloid skewing is a notable feature in individuals with CHIP,12 we further conducted stratified analysis by myeloid skewing and myeloid malignancy. We found a negative association between standard deviation increase of myeloid skewing and NHL referred to individuals with myeloid malignancy (i.e., reference) in both CHIP group (OR: 0.60, 95% CI: 0.45–0.78, p < 0.001) and non-CHIP group (OR: 0.68, 95% CI: 0.62–0.74, p < 0.001) (Table S8), suggesting more tendency of myeloid skewing in myeloid malignancy compared to NHL. Clonal cytopenia of undetermined significance (CCUS), a prevalent condition characterized by persistent hematopoietic clonality and cytopenia, poses a risk for progression to myeloid malignancy.13 Thus, we conducted a stratified analysis by CCUS to investigate the association between CCUS and NHL. Compared to individuals without CCUS (HR: 1.44, 95% CI: 1.21–1.72, p < 0.001), a higher increased risk of NHL was found for individuals with CCUS (HR: 2.58, 95% CI: 1.64–4.06, p < 0.001) (Table S9). In this extensive community-based cohort study, we demonstrated that the presence of CHIP in healthy individuals is associated with an elevated risk of NHL. All individuals with CHIP, regardless of the varying VAFs they carry, exhibited an increased risk of developing NHL. This connection unveils a new chapter in the study of how normal hematopoiesis can deviate towards malignancy, providing a novel framework for exploring the molecular and cellular mechanisms underlying lymphoma. In clinical settings, the identification of CHIP as a risk for NHL may help clinicians to stratify individuals at a higher risk of developing NHL. This could lead to more personalized surveillance strategies, such as earlier and more frequent screening for NHL in individuals with CHIP. High-risk individuals identified by CHIP screening could be considered for preventive interventions, such as lifestyle modifications or targeted therapies, in the future. Some limitations exist in our study. First, only a single cohort was used for our study, limiting validation from the external cohort. Future studies with multiple cohorts of different backgrounds are needed to validate our findings. Second, our study cannot obtain dynamic repeated WES data to evaluate time-varying exposure, resulting in participants who developed CHIP during follow-up not being identified. Consequently, our observed association might be lower than it is. Third, despite a large sample being analyzed in our study, the limited number of individuals was insufficient for statistical power to reveal the association between different subtypes of NHL and CHIP. Beyond its contribution to myeloid malignancy, our research highlighted CHIP as a risk factor for NHL, facilitating novel knowledge in the field. This research offers new insights into the genetic underpinnings of lymphoma development and presents potential avenues for early detection and prevention. Future research endeavors should prioritize translating these findings into clinical applications, ultimately aiming to enhance patient outcomes. This research has been conducted using the UK Biobank Resource under application number [106912]. Mingcheng Liu: Writing—original draft; investigation; methodology; visualization. Luyao Zhou: Methodology; investigation; visualization. Qiaoxue Liu: Software; methodology. Xiaojing Wang: Writing—review and editing; validation; supervision; visualization; conceptualization; project administration. Hui Wei: Writing—review and editing; project administration; funding acquisition; resources; supervision; conceptualization; software; formal analysis. Qianwei Liu: Writing—review and editing; project administration; resources; supervision; data curation; methodology; conceptualization. The authors declare no conflict of interest. This article was funded by National Key Research and Development Program of China (2023YFC2508900); National Natural Science Foundation of China (82370183); Tian Jin Natural Science Foundation (23JCZXJC00310); CAMS Innovation Fund for Medical Sciences (2024-I2M-ZH-015); Haihe Laboratory of Cell Ecosystem Innovation Fund (22HHXBSS00040); and Beijing Xisike Clinical Oncology Research Foundation (Y-SYBLD2022ZD-0031). The data that support the findings of this study are available in UK Biobank at https://www.ukbiobank.ac.uk/. These data were derived from the following resources available in the public domain: - DNAnexus, https://platform.dnanexus.com/ This study utilized the UK Biobank Resource, under Application Number 106 912. Researchers can access data from UK Biobank at www.ukbiobank.ac.uk. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.