Despite extensive research into diabetes–cancer connection, answers remain elusive

Since 2000, there have been approximately 950 published papers that examined the association between diabetes and cancer, a PubMed search reveals.

The investigations — primarily epidemiological or observational in design — have generated dozens of hypotheses about biological links between the two diseases, as well as the possibility that certain medications intended to regulate blood glucose levels could increase risk for malignancy.

Yet, the evidence remains far from conclusive.

In August, the American Association of Clinical Endocrinologists (AACE) released a consensus statement developed in response to increasing concern about the potential relationship between diabetes, diabetes treatments, obesity and cancer prevalence.

A task force commissioned by AACE and the American College of Endocrinologists (ACE) explored possible synergistic mechanisms of cancer and diabetes development on the molecular level. Members also evaluated the effect of several diabetes treatments — including metformin and glucagon-like peptide 1 (GLP-1) agonists — on cancer incidence and mortality.

The report concluded that diabetes and obesity are associated with statistically significant increased risks for several types of cancer. However, the report’s authors noted “complex and inconclusive findings” regarding the effects of antidiabetes pharmacologic agents.

“The implication of various medications’ role in the development of cancer has concerned physicians and patients alike, but the sum of evidence presents a very compelling case and suggests the risk of cancer is unproven,” said Yehuda Handelsman, MD, FACP, FACE, FNLA, medical director of the Metabolic Institute of America and co-chair of the task force that developed the consensus statement.

“None of those headlines should dissuade us,” Handelsman told HemOnc Today. “None of the evidence says we shouldn’t use accepted medications based on how the FDA says we should use them.”

HemOnc Today spoke with several clinicians and researchers about the current state of the science regarding the diabetes–cancer connection, the need for — and barriers associated with — large-scale randomized trials, and the ways in which physicians can use the available evidence to improve patient care.

Role of obesity, cholesterol

Obesity is the least disputed component of the diabetes–cancer connection.

“Many proposed biological mechanisms link obesity to cancer development through the direct or indirect effects of obesity on insulin and insulin-like growth factor I, sex hormones, adipokines and inflammation,” they wrote. “The collective activation of these individual mechanisms promotes an environment of increased proliferation, inhibited apoptosis and increased genomic instability. Recent tissue-based breast cancer studies have provided support for hypothetical obesity-related cancer mechanisms in humans. Breast tissue samples obtained from women undergoing surgery for breast cancer have shown a significant direct correlation between BMI and inflammation, adipocyte size, and aromatase expression and activity.”

The data suggest that patients with diabetes and those who are obese should undergo regular cancer screening and counseling on lifestyle changes, Handelsman and colleagues concluded. In addition, those who develop “typical obesity-related cancers” — especially at a younger age — should undergo screening for diabetes, insulin resistance and other metabolic abnormalities.

Obesity and diabetes also appear to adversely affect survival in patients with cancer.

In a study published this past summer in Annals of Oncology, Jiralerspong and colleagues evaluated 6,342 women with stage I to stage III breast cancer.

Among overweight women, HRs were 1.18 (95% CI, 1.02-1.36) for RFS, 1.2 (95% CI, 1-1.42) for OS and 1.21 (95% CI, 0.98-1.48) for breast cancer-specific survival (BCCS) compared with normal-weight women, results of a multivariable analysis showed. Among obese women, HRs were 1.13 (95% CI, 0.98-1.31) for RFS, 1.24 (95% CI, 1.04-1.48) for OS and 1.23 (95% CI, 1-1.52) for BCCS.

Among women with diabetes, HRs were 1.21 (95% CI, 0.98-1.49) for RFS, 1.39 (95% CI, 1.1-1.77) for OS and 1.04 (95% CI, 0.75-1.45) for BCCS.

“There is no real question that diabetes is associated with a higher risk of certain types of cancer, and with a higher mortality following cancer diagnosis. The obvious question is why,” Prof. Edwin Gale, MB, BChir(Cantab), FRCP, emeritus professor at the University of Bristol in the United Kingdom, told HemOnc Today. “Type 2 diabetes, obesity and insulin resistance associate together. Obesity and markers of insulin resistance are each associated with an increased risk of cancer, but most of the evidence suggests that hyperglycemia itself is not. Since hyperglycemia is the hallmark of diabetes, there is still an open question as to whether the increased risk seen in diabetes is secondary to its metabolic syndrome features or not. In my view, it is — though there may be exceptions for individual cancers.”

Cholesterol also may play a role

Jamie Morton, MBBS, FRCP, PhD, of The Heart Research Institute in Sydney, Australia, and colleagues followed 11,140 patients with type 2 diabetes for a median of 5 years. They aimed to evaluate possible links between baseline HDL cholesterol levels and cancer incidence and mortality risk. Low HDL level is a known risk factor for cardiovascular disease in patients with type 2 diabetes and low HDL has been associated with cancer incidence in patients without diabetes. However, it was previously unknown whether low HDL also would be a risk factor for patients with diabetes.

The final analysis included 699 patients who developed cancer. For every 0.4 mmol/L lower baseline HDL cholesterol, researchers observed a 16% increase in cancer incidence (HR=1.16; 95% CI, 1.06-1.28). The trend remained significant after adjustment for confounding.

The association was even stronger when only looking at cancers that occurred early following the HDL measurement and at this time point was significantly associated with cancer mortality. However, this association was attenuated in cancers that occurred later.

“Complicating the relationship between cancer and total cholesterol is the understanding that the respective levels of high- and low-density lipoproteins that comprise total cholesterol level may have vastly different effects on cancer risk, in much the same way as is observed for cardiovascular disease risk,” Morton said. “HDL exerts anti-inflammatory effects, raising the possibility that HDL could interrupt inflammation-driven carcinogenesis. The antioxidant effect of HDL could also provide protection against oxidative stress, known to be closely related to cancer and inflammation.”

Conversely, Morton cited basic science findings that indicate a role of HDL in promoting cancer-related cellular mechanisms.

“HDL can promote vascular endothelial cell proliferation and migration and reduce apoptosis,” he said. “In the atherosclerotic plaque, these effects have a positive effect on endothelial integrity; however, in cancer, these events could drive tumorigenesis.”

The conclusions of the study were that low HDL is associated with cancer in patients with type 2 diabetes but is probably not causal.

Conflicting results

The effect of diabetes in patients who are not obese also has been investigated, but results were far from conclusive.

Yerrabothala and colleagues examined the effect of diabetes in a cohort of 255 women with various stages of breast cancer. More women without diabetes presented with early-stage cancer than women with diabetes (54.1% vs. 41.2%), a difference researchers noted was nonsignificant (P=.068).

However, researchers reported a significant survival difference among patients with diabetes compared with those who did not have diabetes (P=.001). After adjustments for covariates and stratification for BMI, patients with diabetes demonstrated poorer prognosis regarding survival than patients who did not have diabetes.

A population-based and prospective cohort study by Backemar and colleagues, however, suggested the opposite outcome in patients who underwent surgical resection for esophageal or cardia cancer.

The study, which included 609 patients treated in Sweden during a 4-year period, showed no increase in mortality risk among the 67 patients with diabetes compared with those who did not have diabetes (HR=0.81; 95% CI, 0.60-1.09).

“[These findings are] also complicated,” Gale said. “The key point to remember is that Mr. A, who has diabetes, is more likely to die in the next 5 years than Mr. B, who does not. If both get cancer, Mr. A is still more likely to die first. Thus, it is no surprise that people with diabetes have an increased mortality following a cancer diagnosis.

“The real question is whether they have a higher cancer-related mortality,” Gale added. “There are several possible reasons why they might — for example, diabetes or its complications puts limits on the use of some types of cancer therapy. But, in general, the risk of cancer-related death is not greatly increased.”

Antidiabetes treatments

The influence antidiabetes treatments may have in the development of cancer incidence is much less clear.

In 2011, the FDA announced a warning would be added to the label of pioglitazone (Actos, Takeda) noting that high-dose use of the drug for more than 2 years may result in an increased risk for bladder cancer.

A 2012 meta-analysis by Colmers and colleagues concluded long-term use of pioglitazone — a prescription drug in the thiazolidinedione (TZD) class — was associated with a significantly increased risk for bladder cancer. Two other retrospective studies found a similar association when pioglitazone exposure exceeded 24 months or when cumulative doses exceeded 28,000 mg.

However, in their consensus statement, Handelsman and colleagues acknowledged emerging evidence that suggests TZDs, as well as metformin, may reduce risk for certain malignancies. Data suggest these drugs potentially could serve as adjunctive therapy for cancer management.

“Nonetheless, it is premature to prescribe metformin and TZDs solely for these as yet unproven indications,” they wrote.

Conversely, the task force members said certain antihyperglycemic agents have been shown to lead to the development of some malignancies.

“However, the evidence implicating these medications is primarily based on basic research, animal data and descriptive epidemiologic studies useful to formulate, not test, hypotheses,” they wrote. “To detect reliably the most plausible small to moderate effects requires large-scale randomized evidence.”

Current evidence is not sufficient to change clinical practice, so clinicians should continue to prescribe all FDA-approved antihyperglycemic drugs according to available recommendations, Handelsman said. However, they should be aware of potential risks and monitor patients more closely.

“Clinicians (also) should exercise caution when choosing medications implicated in the etiology of cancer for patients with the specific organ-related risk,” Handelsman and colleagues wrote.

Of all antidiabetic agents, metformin has been the subject of the most intense investigation regarding its effect on cancer incidence.

In a study published in July in the Journal of the National Cancer Institute, Margel and colleagues examined a potential link between metformin use and prostate cancer risk among 119,315 men with diabetes. The study included 5,306 men with prostate cancer and 26,530 matched controls.

Researchers observed no connection between metformin use and risk for any prostate cancer (adjusted OR [aOR]=1.03; 95% CI, 0.96-1.1), nor did they find an associations with high-grade disease (aOR=1.13; 95% CI, 0.96-1.32), low-grade disease (aOR=0.94; 95% CI, 0.82-1.06) or biopsy-diagnosed disease (aOR=0.98; 95% CI, 0.84-1.02).

Margel and colleagues also conducted a population-based, retrospective study that included 3,837 patients with prostate cancer. In that analysis — results of which were published in September in the Journal of Clinical Oncology — they determined cumulative duration of metformin exposure after prostate cancer diagnosis was linked to decreased all-cause and prostate cancer-specific mortality.

Tseng and colleagues conducted a prospective study that included 86,939 patients with type 2 diabetes. All patients were aged at least 25 years.

The researchers investigated a possible association between insulin use and bladder cancer mortality. They also evaluated whether a synergistic effect exists among those who smoke and use insulin.

Multivariate analysis results indicated that older age, male sex and insulin use consistently predicted bladder cancer mortality.

Depending on the statistical model, adjusted HRs for bladder cancer death among insulin users compared with nonusers ranged from 1.877 to 2.502 (P<.05 for all).

“Clinicians should be careful when reading reports that metformin or TZDs may be beneficial,” Handelsman said. “We stressed and highlighted the fact that basic science research and long-term, double blind, randomized studies in large populations are necessary in this area.”

Mechanism of action

There are several theories regarding the potential mechanism of action.

“The most plausible is activation of the insulin and insulin-like growth factor (IGF) axis,” Morton told HemOnc Today. “Although glucose uptake in cancer cells is constitutively high and independent of insulin, stimulation of these receptors by insulin acts as an abnormal growth signal by activating downstream signaling pathways. This may stimulate cancer cell proliferation and metastasis and protect against apoptosis.”

Data has suggested that insulin and IGF-I has mitogenic and antiapoptotic properties. Hyperinsulinemia also may affect estrogen levels, which could stimulate tumors that are responsive to sex hormones, Morton said.

“Another well-characterized pathway is through modulation of inflammatory cytokines,” he said. “The adipose tissue in diabetes is not an inert collection of fat but a highly active endocrine organ producing, among others, tumor necrosis factor alpha (TNF-alpha), interleukin-6 (IL-6), monocyte chemoattractant protein 1 (MCP-1) and platelet activator inhibitor 1 (PAI-1). The first two, TNF-alpha and IL-6, are two of the most studied inflammatory cytokines in cancer and integral to the concept of cancer-related inflammation. Recruited inflammatory cells play an indispensable role in tumor proliferation, survival and metastasis.”

Handelsman said he wondered whether glucose itself may add to cancer risk.

“People with diabetes have more cancer than people with obesity,” Handelsman said. “I wonder if glucose is causing that. There is no direct evidence that glucose causes cancer. However, it might be inflammation that’s causing cancer. It is possible that cancer cells are growing in a different way.”

Derek LeRoith, MD, professor of medicine, endocrinology, diabetes and bone disease at Mount Sinai Hospital, also addressed the subject.

“The general consensus is that there is no strong evidence to suggest a connection between injected insulin and cancer risk, while there is for endogenous hyperinsulinemia of the individuals,” LeRoith said.

The AACE/ACE consensus statement, however, provides — at least for now — the final word on glycemic control.

“The totality of available evidence supports the need for astute clinical judgment in which remote yet plausible cancer risks are weighed against suboptimal glycemic control and higher likelihoods of diabetes complications — especially microvascular, but also macrovascular complications,” the statement reads. “When prescribing antihyperglycemic medications, a comprehensive risk–benefit analysis must be performed to include an assessment of the baseline personal and familial risk of malignancies in specific organ systems.”

The statement authors added that “rigorous and multifactorial approaches” should be taken in patients with diabetes who receive treatment for cancer.

“For inpatients, aggressive glycemic management has been associated with improved outcomes,” they wrote.

GLP-1 agonists

Peter Butler, MD, chief of endocrinology, diabetes and hypertension at David Geffen School of Medicine, and colleagues conducted a study that focused on the potential associations between GLP-1 receptor agonists and pancreatic and thyroid cancers.

“Biological hypotheses and animal data suggest that pancreatitis was a potential human side effect of these drugs,” researcher Michael Elashoff, PhD, told HemOnc Today.

The researchers used an uncontrolled FDA adverse events database to perform a controlled analysis that compared three adverse events for GLP-1 drugs vs. non–GLP-1 diabetes drugs.

The study, published in Diabetes Care, included a secondary level of control that evaluated for adverse events that were not associated with GLP-1 or non–GLP-1 drugs to control for the potential that newer drugs might be the subject of more reports within the database.

“We found a strong indicator that there was an elevated number of reports for pancreatitis with GLP-1 drugs compared with other diabetes drugs,” Elashoff said. “There was a significant OR. We also found a modest increase in pancreas cancer and thyroid cancer for these drugs, with an OR of about 2.”

Those two cancers were preselected based on a plausible biological mechanism, Elashoff said.

“This makes it more involved than just a data mining exercise,” he said.

Elashoff noted the OR for pancreatic and thyroid cancers was not as strong as the effect for pancreatitis, which was linked to an OR of approximately 6, but there still was a signal.

“That kind of database analysis does not establish causality,” he said. “We merely observed a potential association.”

The rareness of pancreas and thyroid cancers also complicates the analysis.

“The ideal way to look at this is a randomized controlled trial, but of course you would need enormous numbers of patients,” Elashoff said. “So although there is clearly potential in this FDA database because it draws from millions of patients and millions of reports, we don’t have a conclusive answer.”

Gale also addressed the controversy surrounding GLP-1 drugs.

“There is a signal for pancreatic cancer for multiple agents in the class in the FDA, EMA and WHO databases,” Gale said. “There is a plausible mechanism, and the autopsy pancreas specimens jointly reported by Dr. Butler’s group and [the group led by Mark Atkinson, PhD, of the University of Florida] are very disturbing.”

The American Diabetes Association in June asked all pharmaceutical companies involved in the development or marketing of incretin-based medications — which include GLP-1 receptor agonists and dipeptidyl-peptidase IV (DPP-IV) inhibitors — to help provide more definitive evidence about whether those therapies contribute to the development of pancreatic cancer or pancreatitis.

A ‘raw area’

Despite the ongoing uncertainty surrounding this topic, Handelsman and his consensus statement co-authors provided two concrete suggestions.

The first relates to the fundamentals of good health.

“Based on currently understood mechanisms for the development of cancer in obesity and diabetes, proper nutrition management, weight loss, and exercise are equally important to the management of people with cancer as it is to people with obesity and diabetes,” they wrote.

The second focused on cancer screening. They suggested that screenings with proven benefit in certain high-risk populations should apply to obese individuals, as well as those with clinical symptoms of diabetes that may put them at risk for cancer.

Gale said he found these and most other recommendations in the consensus statement to be valuable.

“The AACE summary statement is a reasonable reflection of current thinking in this area,” he said. “It confirms that there is an increased cancer risk in type 2 diabetes, probably mediated by obesity rather than hyperglycemia, and it reflects growing skepticism about the use and interpretation of observational studies to implicate individual therapies.

“But we shouldn’t throw the baby out with the bathwater,” Gale said. “Insulin resistance is in some way implicated in cancer risk, and this requires further investigation. Also, the jury is still out on the incretins, which, unlike the other agents considered, have a plausible mechanism for harm and a strong safety signal in regulatory databases.”

LeRoith agreed.

“There is not enough evidence,” he said. “We need more studies to be 100% sure.”

Answers are unlikely in the short term, Handelsman said.

“This is a raw area,” Handelsman said. “The study of diabetes and cancer is not a field as yet. There are no true experts. It may be some time before these answers can be found.” – by Rob Volansky

References:

American Association of Clinical Endocrinologists and American College of Endocrinology. Diabetes and cancer: An AACE/ACE Consensus Statement. 2013. Available at: www.aace.com/files/position-statements/diabetes-and-cancer-consensus-statement.pdf. Accessed on Sept. 12, 2013.

Backemar L. Am J Surg. 2013;doi:10.1016;j.amjsurg.2013.01.035.

Butler PC. Diabetes Care. 2013;36:2118-2125.

Colmers IN. Diabetes Metab. 2012;38:475-484.

Jiralerspong S. Ann Oncol. 2013;published online ahead of print July 1.

Margel D. J Clin Oncol. 2013;31:3069-3075.

Margel D. J Natl Cancer Inst. 2013;doi:10.1093/jnci/djt170.

Morton J. Cancer Epidemiol Biomarkers Prev. 2013;22:1628-1633.

Tseng CH. Clin Genitourin Cancer. 2013;doi:10.1016/j.clgc.2013.04.019.

Yerrabothala S. Pathol Oncol Res. 2013;published online ahead of print July 6.

Zhu Z. BMC Cancer. 2013;13:310.

For more information:

Michael Elashoff, PhD, can be reached at elashoff@gmail.com.

Prof. Edwin Gale, MB, BChir(Cantab), FRCP, can be reached at edwin.gale@bristol.ac.uk.

Yehuda Handelsman, MD, FACP, FACE, FNLA, can be reached at yhandelsman@yahoo.com.

Derek LeRoith, MD, can be reached at derek.leroith@mssm.edu.

Jamie Morton, MBBS, FRCP, PhD, can be reached at jamie.morton@hri.org.au.

Disclosure: Elashoff, Gale, Handelsman, LeRoith and Morton report no relevant financial disclosures.

 

Should an independent review be conducted to determine whether incretin therapy contributes to the development of pancreatic cancer or pancreatitis?

The additional information would help settle the question.

The American Association of Clinical Endocrinologists (AACE) is one of multiple organizations that recently issued press releases regarding the relationship between incretin therapy and pancreatic disease. The American Diabetes Association — in conjunction with the European Association for the Study of Diabetes and the International Diabetes Federation — issued one, as did The Endocrine Society.

In all of them, the common factor is that there is insufficient evidence to date to alter practice guidelines or change how patients and providers treat diabetes. We’ve all recommended that people who are currently being treated with incretins speak with their health care providers about the risks and benefits, then make individualized decisions.

Having said that, there is still some question. We are always trying to clarify both the safety and efficacy of medications. This is not to say there’s a problem; rather, we’re saying we need to have all of the available information in order to make rational decisions.

The first thing we need to do is to be able to obtain the patient-level data from as many of the studies that have been performed as possible. We’re looking at what is clearly a rare event, which means we have to have large patient populations to be able to see whether it’s occurring by chance or if there is some association with the intervention. This is an effort to increase the numbers of individuals so that we might be able to get more definitive data. We’re not looking to question any of those data. We simply want to pull it all together to increase the likelihood of our being able to answer the question.

We are still negotiating with our industry colleagues about how we can obtain the data, what data will be available and what format they will be in. All of that needs to be identified and clarified before we release the RFA for the analysis, but those meetings and discussions are ongoing.

The AACE consensus statement says there is insufficient evidence of a problem to change therapy, and we completely agree. All we’re asking for is further analyses to clarify the occurrence of rare events.

Robert E. Ratner, MD, FACP, FACE, is the chief scientific and medical officer for the American Diabetes Association. He can be reached at American Diabetes Association, 1701 N. Beauregard St., Alexandria, VA 22311; email: rratner@diabetes.org. Disclosure: Ratner reports no relevant financial disclosures.

 

Based on current data, it is much too early to estimate such a risk.

In my view, there is no positive data demonstrating that these medications would directly cause pancreatic cancer. However, we cannot exclude any long-term risk beyond current exposure.

With the knowledge we have today from various sources, animal experiments, epidemiological studies and mechanistic considerations, I do not see a solid ground for restricting the use of incretin therapies.

The recent AACE consensus statement said there is not enough evidence to change therapy, which is what we already know. There is not sufficient evidence to make a strong point that we should restrict the use of these drugs in general or in certain populations, nor does it finalize the discussion with some agreement that this was accepted universally no matter the point of view.

The issue needs to be looked at as a larger picture, in terms of the drug class. For instance, we must look at the proven benefits from incretin-based therapies, and influences on soft-surrogate parameters. They tend to result in reduced HbA1c, body weight and blood pressure. If we move on to the GLP-1 receptor agonists, however, no data available to date has proven that the class translates to fewer cardiovascular events, prolonged life, etc.

Therefore, a very critical position could be that as long as there is no proven benefit, I hesitate to use a drug that has a definite risk or some undecided risk that we cannot exclude at present if we are talking about a life-threatening disease such as pancreatitis and pancreatic cancer. However, I do not have this position.

I think it is reasonable to assume if you reduce glucose, blood pressure and body weight, it will be a basis for improved health outcomes. However, that has not been proven for any of these drugs, so we are waiting for results that finally allow us to make the statement: “Yes, this is the treatment that benefits the patient in the following way.”

Prof. Michael A. Nauck, MD, PhD, is head of the Diabetes Center Bad Lauterberg in Germany. He can be reached at Diabeteszentrum Bad Lauterberg, Kirchberg 21, D-37431 Bad Lauterberg im Harz, Germany; email: nauck@diabeteszentrum.de. Disclosure: Nauck reports research support from, advisory board roles with, and honoraria or consulting fees from AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, Hoffmann-La Roche, Janssen, Merck Sharp & Dohme, Novo Nordisk, Sanofi-Aventis, Takeda and several other companies.