Mitochondrial Genetic Polymorphisms Do Not Predict Survival in Patients with Pancreatic Cancer

Pancreatic cancer is expected to affect 37,170 patients in the United States in 2007 and 33,370 will to succumb to the disease (1). Mutations of mitochondrial DNA have frequently been observed in human cancers. Recent data suggest that certain mitochondrial DNA single nucleotide polymorphisms (mtSNP) may correlate with poorer prognosis in patients with advanced pancreatic cancer (2, 3). It was reported that a mtSNP in the 16519 mitochondrial DNA nucleotide was found to be associated with worse prognosis; a subgroup of patients without distant metastases or vascular invasion seemed to have worse survival in the presence of the 16519T allele when compared with the 16519C allele (3). MtSNPs (16,519) have been associated with endometrial cancer (4) and with nonmalignant conditions such as type 2 diabetes mellitus (5) and iron overload disorders (6). We examined the association of 24 common mtSNP, including 16519 mitochondrial DNA, with prognosis in a large cohort of patients with pancreatic cancer.

Adult patients with histologically confirmed pancreatic adenocarcinoma seen at Mayo Clinic were identified through an ultrarapid identification system and invited to participate in a prospective study and registry of pancreatic cancer. Blood samples, clinical information, and risk factor data were collected. The registry performs follow-up of all patients by chart review, mailed surveys, and linkages to the Mayo Tumor Registry and the National Death Index (7). Twenty-four SNPs were selected based on allele frequency, haplogroup, and tagSNPs. There were 17 SNPs in the mitochondrial coding region and 7 common SNPs in the regulatory region (displacement loop). Genotyping methods using GenomeLab SNPstream system (Beckman Coulter, Inc.) were reported in detail previously (8).

Kaplan-Meier survival analyses [time to event (death) was the time between the diagnosis of pancreatic cancer and death] were implemented to examine the univariate associations between mtSNP allele and survival postdiagnosis of pancreatic cancer. Cox proportional hazards regression models were used to test the association after adjusting for covariates [age, gender, smoking history (ever/never), stage at diagnosis, and treatment received chemotherapy and/or radiation (no/yes)]. Analyses were done on the overall group as well as stratified by whether surgery was done (no/yes). HRs and 95% confidence intervals were used to quantify any significant associations. To account for multiple testing, Bonferroni correction was used. Analyses were done using SAS version 9.1.3.

Patients (990) with pancreatic cancer seen at Mayo Clinic between October 1, 2000, and February 23, 2006, were included in our analysis (Table 1). The call rates for the genotyping of each of the mtSNPs ranged from 94.2% to 99.6%. Information regarding the patients' vital status through February 6, 2007, was ascertained. Twelve-month survival status was available for 913 of the patients (92%). We found no evidence that any of the mtSNPs or haplogroups predicted survival when the entire cohort was evaluated before or after adjusting for covariates (age at diagnosis, sex, stage, smoking status, chemotherapy, and radiation therapy).

When patients who did not undergo surgery were analyzed separately, only two mtSNPs showed a trend toward a correlation with survival [16189G: hazard ratio (HR), 1.25; P = 0.06; and 1719G: HR, 0.70; P = 0.07 after adjusting for covariates; Table 2]. No such correlation was found for patients undergoing surgery. The 16519T allele previously reported by other investigators to be associated with prognosis (3) did not predict survival in patients with pancreatic cancer (HR, 0.96; P = 0.11) when the entire cohort was analyzed or when patients were analyzed according to whether they underwent resection or not. Patients with resectable tumors were, however, less likely to have the 16519T allele when compared with patients not undergoing resection (30% versus 39%).

When patients surviving 6 months were compared with those surviving >6 months, an association was found for 2 alleles. The patients surviving >6 months were more likely to have the 709G allele (89% versus 82%; P = 0.022) and the 13368C allele (93% versus 89%; P = 0.027), but after adjusting for multiple comparisons, this difference was not significant.

This large study evaluated 24 mitochondrial SNPs in 990 patients with pancreatic cancer and did not show any relationship between the evaluated mtSNPs or haplogroups and survival. Although there was a trend toward significance for the 1719G and 16189G alleles, this disappeared after adjusting for multiple comparisons. The 16189 nucleotide was of potential interest because it is located within the highly polymorphic displacement loop of the mitochondrial DNA. However, using our large sample of patients, we did not observe any difference in survival among patients carrying these mtSNPs, or 16519, even after analyzing patients with resectable cancers separately.

In conclusion, our findings indicate that the evaluated mtSNPs have no clinically significant value in predicting prognosis in patients with pancreatic cancer.

No potential conflicts of interest were disclosed.