Common genetic variation and survival after colorectal cancer diagnosis: a genome-wide analysis

Abstract

Genome-wide association studies have identified several germline single nucleotide polymorphisms (SNPs) significantly associated with colorectal cancer (CRC) incidence. Common germline genetic variation may also be related to CRC survival. We used a discovery-based approach to identify SNPs related to survival outcomes after CRC diagnosis. Genome-wide genotyping arrays were conducted for 3494 individuals with invasive CRC enrolled in six prospective cohort studies (median study-specific follow-up = 4.2-8.1 years). In pooled analyses, we used Cox regression to assess SNP-specific associations with CRC-specific and overall survival, with additional analyses stratified by stage at diagnosis. Top findings were followed-up in independent studies. A P value threshold of P < 5×10(-8) in analyses combining discovery and follow-up studies was required for genome-wide significance. Among individuals with distant-metastatic CRC, several SNPs at 6p12.1, nearest the ELOVL5 gene, were statistically significantly associated with poorer survival, with the strongest associations noted for rs209489 [hazard ratio (HR) = 1.8, P = 7.6×10(-10) and HR = 1.8, P = 3.7×10(-9) for CRC-specific and overall survival, respectively). No SNPs were statistically significantly associated with survival among all cases combined or in cases without distant-metastases. SNPs in 6p12.1/ELOVL5 were associated with survival outcomes in individuals with distant-metastatic CRC, and merit further follow-up for functional significance. Findings from this genome-wide association study highlight the potential importance of genetic variation in CRC prognosis and provide clues to genomic regions of potential interest.

Figure 1.

Association between dose of rs1442089 minor allele and overall survival among those with distant-metastatic colorectal cancer by study population. Estimates for COIN/COIN-B are adjusted for treatment arm, chemotherapy regimen, age at randomization, sex and time from diagnosis to randomization. Estimates for all other studies are adjusted for age at diagnosis, sex and the first three principal components of genetic ancestry.

Figure 2.

Association between dose of rs17544464 minor allele and overall survival among those with distant-metastatic colorectal cancer by study population. Estimates for COIN/COIN-B are adjusted for treatment arm, chemotherapy regimen, age at randomization, sex and time from diagnosis to randomization. Estimates for all other studies are adjusted for age at diagnosis, sex and the first three principal components of genetic ancestry.