Meeting News CoveragePerspective

Metformin may not improve survival in pancreatic cancer

PHILADELPHIA — The diabetes drug metformin was not associated with improved survival outcomes for patients with pancreatic cancer, according to retrospective study data presented at the American Association for Cancer Research Annual Meeting.

Metformin is being tested in cancer trials — more than 20 of which remain open — because epidemiologic studies have suggested that its use is associated with a reduction in cancer deaths.

However, Roongruedee Chaiteerakij, MD, PhD, of the division of gastroenterology and hepatology at the Mayo Clinic Cancer Center in Rochester, Minnesota, and colleagues identified inconsistencies and design flaws in many of these trials. Thus, Chaiteerakij and colleagues sought to evaluate metformin use and survival of patients with pancreatic ductal adenocarcinoma and address potential biases that are often part of retrospective clinical trials.

“Epidemiologic studies of medication exposure and cancer survival warrant very careful and detailed data collection and analysis to minimize biases,” Chaiteerakij said during a press conference. “Researchers should exercise caution when initiating clinical trials based on retrospective epidemiologic studies.”

The researchers identified 1,360 patients with pancreatic ductal adenocarcinoma and diabetes from the Mayo Clinic’s Specialized Programs of Research Excellence (SPORE) in pancreatic cancer database who were treated between 2000 and 2011.

OS was the primary outcome of the study.

Patients were categorized into five groups based on whether they had never used metformin (n = 908; reference group) or whether they started metformin more than 1 year before pancreatic cancer diagnosis (n = 84), within 1 year before pancreatic cancer diagnosis (n = 212), within 30 days after pancreatic cancer diagnosis (n = 104), or more than 30 days after pancreatic cancer diagnosis (n = 34).

The median survival for those patients who did not use metformin was 308 days compared with 292 days for those who had some metformin exposure.

Among 413 patients with resectable disease, those who used metformin demonstrated prolonged survival (median, 782 days) compared with those who did not use metformin (median, 612 days; P = .07). However, patients who started metformin more than 30 days after diagnosis had a median survival of 818 days, approximately 3.5 times longer than the median survival of any of the other groups that used metformin, which researchers noted skewed the survival data.

None of the other metformin use groups had a median survival as long as the never-users of the drug, and there was no statistical survival difference.

“These patients already survived more than 30 days, suggesting that there is an inherent survival bias in this group of patients,” Chaiteerakij said in a press release. “After accounting for these unintended biases, the benefit of metformin was not confirmed in our study.”

When compared with never-users, those who started metformin more than 1 year before diagnosis had an HR of 1.08 (95% CI, 0.85-1.37). Researchers calculated an HR of 0.99 (95% CI, 0.85-1.17) for those who started metformin within 1 year before diagnosis 1.04 (95% CI, 0.83-1.31) for those who started metformin within 30 days after diagnosis and 0.49 (95% CI, 0.33-0.74) for those who started metformin more than 30 days after diagnosis.

Chaiteerakij said that an analysis adjusted for several factors including BMI, gender, age and state of disease demonstrated no association between metformin use and survival.

“When we looked at most of the retrospective studies when they showed the benefit of metformin use in cancer survival, most of them classified the use of metformin as ‘ever’ and ‘never’ and most of them looked at metformin exposure prior to metformin use before diagnosis of any cancer,” Chaiteerakij said. “In reality, when you decide to do a clinical trial, you start the patient from the cancer diagnosis and put the patient on the drug. So, the study data doesn’t always match in the retrospective cohort.”  – by Anthony SanFilippo

Reference:

Chaiteerakij R, et al. Abstract 8687. Presented at: American Association for Cancer Research Annual Meeting; April 18-22, 2015; Philadelphia.

Disclosure: Chaiteerakij reported no relevant financial disclosures. HemOnc Today was unable to obtain a list of relevant disclosures for the other researchers.

PHILADELPHIA — The diabetes drug metformin was not associated with improved survival outcomes for patients with pancreatic cancer, according to retrospective study data presented at the American Association for Cancer Research Annual Meeting.

Metformin is being tested in cancer trials — more than 20 of which remain open — because epidemiologic studies have suggested that its use is associated with a reduction in cancer deaths.

However, Roongruedee Chaiteerakij, MD, PhD, of the division of gastroenterology and hepatology at the Mayo Clinic Cancer Center in Rochester, Minnesota, and colleagues identified inconsistencies and design flaws in many of these trials. Thus, Chaiteerakij and colleagues sought to evaluate metformin use and survival of patients with pancreatic ductal adenocarcinoma and address potential biases that are often part of retrospective clinical trials.

“Epidemiologic studies of medication exposure and cancer survival warrant very careful and detailed data collection and analysis to minimize biases,” Chaiteerakij said during a press conference. “Researchers should exercise caution when initiating clinical trials based on retrospective epidemiologic studies.”

The researchers identified 1,360 patients with pancreatic ductal adenocarcinoma and diabetes from the Mayo Clinic’s Specialized Programs of Research Excellence (SPORE) in pancreatic cancer database who were treated between 2000 and 2011.

OS was the primary outcome of the study.

Patients were categorized into five groups based on whether they had never used metformin (n = 908; reference group) or whether they started metformin more than 1 year before pancreatic cancer diagnosis (n = 84), within 1 year before pancreatic cancer diagnosis (n = 212), within 30 days after pancreatic cancer diagnosis (n = 104), or more than 30 days after pancreatic cancer diagnosis (n = 34).

The median survival for those patients who did not use metformin was 308 days compared with 292 days for those who had some metformin exposure.

Among 413 patients with resectable disease, those who used metformin demonstrated prolonged survival (median, 782 days) compared with those who did not use metformin (median, 612 days; P = .07). However, patients who started metformin more than 30 days after diagnosis had a median survival of 818 days, approximately 3.5 times longer than the median survival of any of the other groups that used metformin, which researchers noted skewed the survival data.

None of the other metformin use groups had a median survival as long as the never-users of the drug, and there was no statistical survival difference.

“These patients already survived more than 30 days, suggesting that there is an inherent survival bias in this group of patients,” Chaiteerakij said in a press release. “After accounting for these unintended biases, the benefit of metformin was not confirmed in our study.”

When compared with never-users, those who started metformin more than 1 year before diagnosis had an HR of 1.08 (95% CI, 0.85-1.37). Researchers calculated an HR of 0.99 (95% CI, 0.85-1.17) for those who started metformin within 1 year before diagnosis 1.04 (95% CI, 0.83-1.31) for those who started metformin within 30 days after diagnosis and 0.49 (95% CI, 0.33-0.74) for those who started metformin more than 30 days after diagnosis.

Chaiteerakij said that an analysis adjusted for several factors including BMI, gender, age and state of disease demonstrated no association between metformin use and survival.

“When we looked at most of the retrospective studies when they showed the benefit of metformin use in cancer survival, most of them classified the use of metformin as ‘ever’ and ‘never’ and most of them looked at metformin exposure prior to metformin use before diagnosis of any cancer,” Chaiteerakij said. “In reality, when you decide to do a clinical trial, you start the patient from the cancer diagnosis and put the patient on the drug. So, the study data doesn’t always match in the retrospective cohort.”  – by Anthony SanFilippo

Reference:

Chaiteerakij R, et al. Abstract 8687. Presented at: American Association for Cancer Research Annual Meeting; April 18-22, 2015; Philadelphia.

Disclosure: Chaiteerakij reported no relevant financial disclosures. HemOnc Today was unable to obtain a list of relevant disclosures for the other researchers.

    Perspective
    Igor Astsaturov

    Igor Astsaturov

    Metformin is a common drug used in the treatment of type 2 diabetes that acts through activation of AMP-dependent kinase, a negative regulator of mTOR pathway signaling. The link between diabetes and pancreatic cancer has been long established. Several epidemiological studies indicated that administration of metformin reduces the incidence of and improves prognosis in pancreatic cancer patients. One study showed a 62% pancreatic cancer risk reduction in patients with type 2 diabetes associated with metformin use (Li D, et al. Gastroenterology. 2009;doi: 10.1053/j.gastro.2009.04.013.). A number of preclinical studies showed metformin suppressed growth of pancreatic cancer cells in vitro and in xenografts by blocking G-coupled protein receptor and mTOR signaling. Ongoing clinical trials are testing the addition of metformin to chemotherapy.

    The results of a retrospective analysis by Chaiteerakij and colleagues presented at the American Association for Cancer Research 2015 Annual Meeting are certainly discouraging and indicate that the effects of metformin were not strong enough to demonstrate the survival advantage. While prospective validation of the metformin activity is certainly warranted, it raises a concern that the ongoing metformin trials are standing a relatively small chance for success.  Therapy advancement for pancreatic cancer has been hampered by the lack of actionable targets (the founding mutations in KRAS and TP53 are not drug-amenable) and by very poor penetration of therapeutic agents to the highly fibrotic tumor. These critical mechanisms are in the crosshair of the current fight against this devastating cancer.

    • Igor Astsaturov, MD, PhD
    • Fox Chase Cancer Center

    Disclosures: Astsaturov reports no relevant financial disclosures.

    Perspective

    For many of our most common cancers — breast cancer, prostate cancer and colorectal cancer —there is a pretty consistent story of obesity and being overweight related to increased cancer risk or a worse outcome for those already diagnosed. Many times that is or isn’t associated with formal diabetes. One wonders, if someone is overweight with a BMI of 26 kg/m2 or greater, does the amount of insulin that is produced to maintain a normal blood sugar at that weight drive cancer signaling, since the PI3K is downstream of insulin signaling? One of the thoughts is that metformin may be used in that context, but that’s not the way that it has been used, so you can’t get that from the epidemiology and you have to do a clinical trial. The data collected here and the analysis qualifies the enthusiasm for the 11 clinical trials ongoing with metformin for pancreatic cancer. There are many metformin clinical trials ongoing for other cancers. How pervasive is this bias that was identified in epidemiologic studies? Should that collar other clinical trials expectations for other cancers for the same reasons?

    • William Nelson, MD, PhD
    • Sidney Kimmel Comprehensive Cancer Center Johns Hopkins University

    Disclosures: Nelson reports consultant/advisory roles with and stock ownership in Cepheid, CHD Biosciences, Clontech, Digital Harmonics, MDxHealth and ProQuest Investments.

    Perspective

    Chaiteerakij and colleagues have correlated metformin use with pancreatic cancer survival using a large database of 1,360 patients with diabetes, a known risk factor for pancreatic cancer. The patients all had pancreatic cancer and were divided in five groups based on metformin use. The group with the longest survival (who started metformin > 30 days after diagnosis of pancreatic cancer) was comprised of only 34 patients. The researchers reported that use of metformin has no OS benefit in patients with pancreatic cancer. This is contrary to the retrospective study by Sadeghi and colleagues from The University of Texas MD Anderson Cancer Center on 302 patients, who demonstrated improved survival in diabetic patients with pancreatic cancer. Although this study is larger, there is considerable heterogeneity in the duration with respect to pancreatic cancer diagnosis in the population and this may have impacted the study results.  Since metformin decreases the cardiovascular morbidity and mortality, cancer-specific survival would have been a more informative endpoint. While there is strong preclinical rationale to believe that metformin has dose-dependent antitumor effects in preclinical models, the many co-morbidities that co-exist with diabetes that impact treatment decision making and the poor survival following pancreatic cancer diagnosis regardless of stage make it challenging to study the impact of metformin on pancreatic cancer survival. It is commendable that the researchers used an available database and presented this hypothesis-generating data. Prospective studies evaluating the role of metformin following cancer diagnosis controlling for known factors are needed to help answer this question. The role of dose of metformin must also be carefully studied, given that metformin is believed to exert its effect on different molecular pathways through dose-dependent mechanisms.

    References:

    Sadeghi N, et al. Clin Cancer Res. 2012;doi:10.1158/1078-0432.CCR-11-2994.

    Johnson JA, et al.  Diabetic Med. 2005;22:497-502.

    Ming M, et al. PLoS ONE. 2014: doi:10.1371/journal.pone.0114573.

    • Hassan Hatoum, MD, and Renuka Iyer, MD
    • Roswell Park Cancer Institute

    Disclosures: Hatoum and Iyer report no relevant financial disclosures.

    See more from American Association for Cancer Research Annual Meeting