Replication of Genetic Markers Associated with Clozapine-Induced Agranulocytosis.

Clozapine is a highly efficacious drug for the treatment of schizophrenia, but its use is limited, in part due to the side effect of agranulocytosis. In order to reduce the incidence of clozapine-induced agranulocytosis (CIA), patients are required to submit to a blood monitoring program. A genetic component to CIA is suggested by published associations in a number of genes. A genetic test assessing the risk of CIA might allow safer and more broad use of clozapine. We previously conducted a candidate gene-based case-control study to discover genes associated with CIA. Cases were patients with an ANC<500 during clozapine treatment, and controls were patients treated for at least one year without a significant reduction in WBC or ANC. For each case, we attempted to enroll 2 age-, gender- and ethnicity-matched controls. We collected blood, medical histories and informed consent from 33 CIA cases and 54 controls. Seventy-four candidate genes were sequenced in the exons, intron-exon boundaries, 5′ untranslated region and promoter region for each of the cases and controls, and a haplotype analysis was conducted using logistic regression. A total of 20 haplotype markers in HLA-DQB1, 1 in HLA-C, 1 in NTSR1, 1 in DRD1 and 6 in CSF2RB were significant with permutation test adjusted p-values under 0.05.

A replication analysis was performed on an independently collected cohort of 49 cases and 78 controls from Germany. Cases were patients who developed an ANC<500 during clozapine treatment, and controls were patients treated for at least two years without an adverse effect on the WBC or ANC. We attempted to replicate 28 haplotype markers from 4 of the 5 genes. We did not attempt to replicate the findings from HLA-C. Permutation tests were performed for markers within genes, and the Benjamini-Hochberg FDR method was applied to the best permutation p-values from the four genes. Markers with p-values less than 0.05 after these adjustments were considered replicated. One genetic marker in HLA-DQB1 replicated further implicating this as the strongest genetic association with CIA. Markers in CSF2RB were not statistically significant, but warrant further study. The strength of the findings is sufficient to develop a diagnostic test for CIA risk. These findings may also be applicable to agranulocytosis and neutropenia induced by other drugs.