Chronic Lymphocytic Leukemia (CLL) displays remarkable ethnic predisposition for Caucasians with a lower incidence of disease in African Americans and relative sparing of Asian Populations. In addition, CLL displays one of the highest familial predispositions of all hematologic malignancies, yet the genetic basis for these differences is not fully defined. Large scale genome wide association studies (GWAS) of CLL have identified the Human Leukocyte Antigen (HLA) locus and other loci associated with the development of CLL but the contribution of individual HLA alleles has not been extensively examined. The National Marrow Donor Program (NMDP) registry contains a large number of disease cases and donor controls across multiple populations as well as high quality classical HLA typing, making it a powerful source of data for inferring HLA associations for hematologic diseases. Using the NMDP dataset, we report the largest study of the HLA associations for chronic lymphocytic leukemia (CLL) to date.


Cases consisted of 2,689 US Caucasian and 298 African-American patients with CLL who underwent HLA typing performed for allogeneic donor search reported to the National Marrow Donor Program (NMDP). 50,000 controls were randomly selected for each population from US individuals who voluntarily enrolled in the NMDP unrelated donor registry. To address ambiguous haplotype phasing inherent to HLA typing, we used population haplotype frequencies to calculate probabilities for genotypes for each possible pair of HLA haplotypes. We then incorporated this typing ambiguity into a statistical model using multiple imputation, allowing us to report high-resolution allele, haplotype, and genotype associations. Patients and controls were compared using multivariate logistic regression using a generalized linear model for the covariates of age, birthdate, gender, and geographical location. To control for multiple testing, we adjusted p-values using False Discovery Rate with a cutoff of 0.05.


We identified several HLA alleles that are associated with an increased susceptibility to the development of CLL as well as several alleles that are protective in the Caucasian as well as the African American (AA) US population. In Caucasians, A*02:01 (OR = 1.203, p = 0.00053), B*15:01 (OR = 1.264, p = 0.0105), B*38:01 (OR = 1.396, p = 0.0198), DRB1*04:02 (OR = 1.675, p = 0.0006) and DRB1*07:01 (OR = 1.223, p = 0.0006) were predisposing and A*01:01 (OR = 0.849, p = 0.01077), B*08:01 (OR = 0.853, p = 0.0437), B*27:05 (OR = 0.743, p = 0.0356) were protective. For African Americans, we find a different collection of high resolution HLA alleles: B*37:01, B*15:17 and DRB1*09:01 were predisposing to the development of CLL (OR 3.017, 3.089 and 2.144, respectively) with highly statistically significant p-values (0.0344, 0.0344, 0.00057, respectively). At the allele family level, B*44 was found to be protective in both Caucasian (OR = 0.843, p = 0.0115) and AA populations (OR = 0.485, p = 0.0309). The allele family C*07 (OR 0.848, p = 0.0046) was protective in Caucasians while C*04 was protective in AA populations (OR = 0.645, p = 0.0422). Association results also revealed additive effects of individual predisposing alleles within the same haplotype, such as the A*02:01∼C*06:02∼B*13:02∼DRB1*07:01 haplotype, which had a higher odds ratio (OR = 2.2, p = 0.000086) than either A*02:01 or DRB1*07:01. We performed a population risk analysis by applying the currently identified allelic disease associations for Caucasians to the known allele frequency seen in other populations, and determined that HLA allele frequency differences contribute to the lower risk of CLL development for both the AA and Asian populations.


We have identified several HLA associations that confer increased risk of development of CLL in both U.S. Caucasians and African Americans as well as several protective alleles. These results confirm the previously identified association with HLA*02:01 (OR 1.32, Di Bernardo, et al) with highly similar odds ratio. In addition, we identify many HLA allele associations that have not been previously observed due to low statistical power. Differences in HLA predispositions and HLA allele frequency between Caucasian, AA and Asian populations contribute to the difference in incidence of CLL in these populations and may confer inherent biologic differences due to their impact on immune surveillance.


Hill:Celgene: Honoraria, Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.