Insulin, metformin and cancer risk: A discussion of possible links
Diabetes, its treatments and their relationship to malignancies is an area of considerable complexity, but further research is required to determine whether these concerns and benefits are valid.
Photo by Phillip Scheurmann
Data from a number of recent studies have suggested that patients with diabetes taking insulin are at risk for increased rates of cancer diagnosis, that there is no additional risk for cancer with insulin glargine, and that metformin and thiazolidinediones may have a protective effect in patients with and without diabetes.
What has emerged in the past few months is a whole new area of investigation with regards to diabetes and cancer, Edwin A.M. Gale, MD, editor of Diabetologia, said during a press conference at the 45th Annual Meeting of the European Association for the Study of Diabetes (EASD). At the beginning of , I thought there were not many areas of diabetes about which I was entirely ignorant and I was wrong.
The link between diabetes therapies and cancer has recently caused a firestorm of attention, according to Jeffrey A. Johnson, PhD, of the University of Alberta.
When looking at the literature, the link is fairly well established, but people just werent talking about it as much until recently, Johnson told HemOnc Today.
Specifically, in June, concern about a possible link between insulin and cancer risk was raised after four studies linking the two conditions were published in Diabetologia.
The first was a German study of 127,031 patients with diabetes included in an insurance database. In it, Hemkens et al identified a statistically significant, dose-dependent link between cancer diagnosis and patients who had used insulin glargine (Lantus, Sanofi-Aventis) compared with patients who had used human insulin. Patients who had taken a daily dose of 10 IU glargine had an increased risk for cancer (HR=1.09). When the dose was increased to 50 IU, the risk for cancer increased to 31%. No association was reported with insulin aspart or lispro in this study.
These results prompted the editors of Diabetologia and the EASD to convene and commission three additional observational analyses that examined the safety of insulin before publishing the German database study findings. Further research was carried out using databases from Scotland, Sweden and the United Kingdom.
According to Gale, the editors expected the results of the additional studies to be negative; however, the Swedish study showed a significantly increased risk for breast cancer in about 6,000 patients.
This population-based study followed nearly 115,000 adults in Sweden who were issued a prescription for insulin in 2005. The researchers used data from the national cancer registry for 2006 and 2007. During this time, patients using glargine alone had an increased incidence of breast cancer vs. those who used other insulins (RR=1.99), after adjustment for BMI, age at diabetes onset, cardiovascular disease and other factors. Monotherapy with glargine was not associated with an increased risk for gastrointestinal cancer (RR=0.93), prostate cancer (RR=1.27) or overall rate of malignancy (RR=1.07), after adjustment for age and sex.
A Scottish study by Colhoun and colleagues evaluated 36,254 insulin-users included in a large clinical diabetes database. At the time of the study, relatively few patients were taking insulin glargine: 447 on glargine monotherapy vs. 3,512 on glargine plus other insulins. During four years of follow-up, the incidence of all cancers was the same among the 3,959 patients using glargine and patients using other types of insulin (HR=1.02). However, the 447 patients who used glargine alone had a significantly higher incidence of all cancers than the 32,295 using other insulins (HR=1.55 vs. HR=0.81).
The fourth study, conducted in the United Kingdom by Craig J. Currie, PhD, and colleagues, was a retrospective cohort study of 62,809 people treated in U.K. general practices. All were aged 40 years and older and started treatment with oral agents or insulin after 2000. They were divided into four treatment groups: metformin monotherapy, sulfonylurea monotherapy, combination therapy with metformin and sulfonylurea, or insulin. Insulin users were subdivided into glargine, long-acting human insulin, biphasic analogue or human biphasic insulin.
Risk for cancer with basal human insulin alone vs. glargine alone was 1.24. Insulin therapy was associated with an increased risk for colorectal (HR=1.69) or pancreatic cancer (HR=4.63) but not breast or prostate cancer when compared with metformin. Results suggested a protective effect of metformin metformin alone was associated with the lowest risk for cancer.
Metformin plus insulin was associated with a reduced progression to cancer (HR=0.54). Compared with metformin monotherapy, the HR for cancer was 1.42 for the insulin group, 1.36 with sulfonylurea monotherapy and 1.08 with metformin plus sulfonylurea. Patients on insulin or secretagogues were more likely to develop solid cancers than those on metformin, and combination with metformin abolished most of this excess risk, the researchers concluded.
The message published in Diabetologia was a mixed one, Gale said.
Interpreting the risk
At the EASD meeting, experts acknowledged these data but said the risk for cancer is no greater with glargine compared with other types of insulin. Despite this, the four studies expose the possibility that insulin could cause existing cancer cells to grow and divide more rapidly, which may explain why more cancers were diagnosed over one to three years of observation.
Jay S. Skyler, MD, who spoke at the EASD press conference and a related symposium on behalf of Sanofi-Aventis, said careful examination of the four studies did not support the headlines generated after the publication of these studies.
If we discard the supposed dose effect in the German study, which several observers feel is not valid, none of the studies show a relationship between insulin glargine and cancer, Skyler, professor of medicine at the University of Miami Leonard M. Miller School of Medicine, said in a statement.
Further, he said the link between cancer and insulin glargine based on available evidence is unsubstantiated, unwarranted and unproven, and a randomized, controlled trial would be the best way to establish or refute whether there is a relationship between insulin glargine and cancer.
In addition to these four studies, there are more than 120 published papers showing that people with type 2 diabetes are at increased risk for various cancers and cancer mortality, according to Johnson. Although the exact mechanisms are not known, insulin especially in high concentrations may likely have a role in the regulation of cell growth and differentiation, he said.
The association of diabetes and cancer has been studied for many years, both epidemiologically and at the basic science level, to attempt to understand if there is causality, Derek LeRoith, MD, PhD, wrote in an editorial published earlier this year in HemOnc Today.
Both obesity and type 2 diabetes (not type 1 diabetes) are associated with an increased risk for cancer and increased mortality rates. Many possible factors are potentially involved in this association. Insulin resistance and hyperinsulinemia are common to obesity and type 2 diabetes, but, similarly, inflammatory cytokines, dyslipidemia and other adipocytokines are all possible factors that have been shown in cell culture and animal models to stimulate cancer growth, he said.
Andrew G. Renehan, PhD, and colleagues conducted a systematic review and meta-analysis of prospective, observational studies to determine the risk for cancer associated with increased BMI. After reviewing 141 articles that included 282,137 incident cases of cancer, the researchers concluded that increased BMI is associated with increased risk for malignancies.
Men with a 5 kg/m2 increase in BMI were at increased risk for esophageal adenocarcinoma (RR=1.52), thyroid (RR=1.33), colon (RR=1.24) and renal (RR=1.24) cancers. Women with a 5 kg/m 2 increase in BMI were at increased risk for esophageal adenocarcinoma (RR=1.51), endometrial (RR=1.59), gallbladder (RR=1.59) and renal (RR=1.34) cancers. The findings were published in 2008 in The Lancet.
Warning against overreaction
Since publication of the Diabetologia studies, major organizations such as the American Diabetes Association released statements that caution against overreaction until further data are available. The FDA also released a statement that said, based on the currently available data, the FDA recommends that patients should not stop taking their insulin therapy without consulting a physician, since uncontrolled blood sugar levels can have both immediate and long-term serious adverse effects.
A new analysis of the Sanofi-Aventis pharmaco-vigilance database for all randomized controlled trials comparing insulin glargine with diabetes comparators revealed no increased incidence of cancer, including breast cancer. Of 10,880 people included in the analysis (glargine=5,657), 0.8% taking glargine reported 52 cases of malignant cancer and 0.9% taking comparators reported 48 cases.
One point has become abundantly clear, and this is that cancer must now be numbered among the complications of diabetes, Gale and Ulf Smith, MD, PhD, president of the EASD, wrote in a Diabetologia editorial.
In addition to the type of insulin, the amount of insulin a patient with diabetes is taking may affect cancer risk. At the EASD meeting, Currie, of Cardiff University in the United Kingdom, and Gale presented the findings of an extension study that explored the relationship between insulin dose and cancer risk in patients with type 2 diabetes. The study analyzed 4,829 patients receiving insulin alone, 5,035 receiving insulin plus metformin and 31,421 receiving metformin alone. Insulin exposure was measured in terms of prescriptions dispensed per year: fewer than seven, seven to 10, 11 to 15 and more than 15.
Cancer risk was found to increase with increasing doses of insulin, with a sixfold increased cancer risk among patients taking the highest insulin dose (more than 15 prescriptions, HR=5.73) vs. metformin alone.
The addition of metformin to insulin treatment canceled out the increased cancer risk with lower doses of insulin and halved the cancer risk at the highest insulin dose group (HR=3.2).
Currie reported a dose-response association between insulin and cancer incidence, but data do not prove a causal relationship. This finding may be a consequence of insulin resistance rather than a direct effect of the insulin itself, he said.
These findings add to the rapidly accumulating evidence that metformin can play an important role in protection from cancer, Currie said. If this observation is confirmed, metformin could potentially prevent many thousands of cancers in patients with type 2 diabetes.
A protective effect
Despite the possible negative effects that insulin may have on cancer risk, several studies suggest that metformin could play an important role in the prevention of cancer particularly pancreatic cancer and treatment is being considered outside the area of diabetes in nondiabetics, Gale said.
Data presented by Johnson and colleagues at the ADA 69th Scientific Sessions in June revealed a 25% reduction in cancer mortality among people with type 2 diabetes using both metformin and thiazolidinedione. The data come from an analysis using administrative databases in Saskatchewan, Canada.
The association between metformin and decreased cancer risk was reported as early as 2005, when a study published in the British Medical Journal reported a 25% reduced risk for developing cancer when patients with type 2 diabetes used metformin.
Now, researchers in oncology are investigating the use of metformin in nondiabetics.
The story with metformin is extremely exciting because clinical studies showing that patients treated with chemotherapy for breast cancer do considerably better if they are on metformin as compared to other treatment, Smith said during the EASD meeting. Another exciting paper [published recently] shows that the pathogenesis of metastases is linked to cancer stem cells that remain after chemotherapy.
We do not know the mode of action of metformin. But if the cancer field starting using this drug as an anticancer agent, then I think it is very, very telling, he said.
Epidemiologic literature also suggests a positive role of thiazolidinedione in terms of cancer risk but a negative role of sulfonylureas. Drug therapies that increase circulating insulin such as sulfonylureas may have a negative effect on hastening tumor growth, and drugs that improve insulin sensitivity such as thiazolidinedione may have a positive effect on reducing cancer outcomes, Johnson said.
In a 2006 Diabetes Care study, Johnson and colleagues compared metformin and sulfonylurea exposure in patients with type 2 diabetes. They found that relative to metformin use, people using sulfonylureas had about a 30% increased risk for death from cancer. In the same analysis, those who used insulin had about a 90% increased risk for death from cancer.
Its hard to say whether metformin is good or sulfonylureas are bad in these analyses, and many of the studies being done are confounded by drug exposures to other antidiabetic agents that may also have a role, Johnson said. But epidemiologic evidence to date would continue to support that metformin potentially has greater benefits that we thought.
Current evidence for antidiabetic therapies in the diabetes/cancer relationship is limited to epidemiologic studies, which some have criticized.
To the extent when an association is a safety concern, we have to rely on the best available evidence, which may be epidemiologic, Johnson said. There is a hierarchy of evidence but sometimes that is not always possible, like in the case of a randomized, controlled trial, and we have to make our decisions on the best available evidence.
Wait for data
For now, the data are not definitive, Currie said. We have to be highly responsible in the way we report this; it needs to be handled with great caution and care.
Gale said the insulin analogue issue may be seen as a side issue in the future because the focus may shift to examining possible links between insulin resistance, cell metabolism and cell turnover; possible nondiabetic use of metformin and glitazones; and targeted screening of high-risk groups.
This is a new area and we have not had time to answer all of the questions, Smith said.
However, based on evidence to date, the experts agreed that patients and health care professionals should not change therapies.
The pieces of the puzzle are starting to come together, Johnson said. But the story of the potential risks for cancer with antidiabetic agents is not yet fully understood.
In conjunction with the EASD meeting, the European Foundation for the Study of Diabetes, the research foundation of EASD, announced that it will support research into diabetes and cancer with up to $4.4 million.
The program will consider all approaches to further elucidate the relationship between diabetes and cancer, as well as the relationship between different diabetes treatments and cancer, either in human studies or suitable human models.
Additionally, Sanofi-Aventis announced the launch of a major insulin research program to provide methodologically robust research to contribute to the debate over insulin safety.
When weve been treating diabetes, we have been focused on pressing a button to control blood glucose, and we have perhaps forgotten that in pressing that button we are inducing quite profound metabolic changes, Gale said. by Katie Kalvaitis
For more information:
- Bowker SL. Diabetes Care. 2006;29:254-258.
- Colhoun HM. Diabetologia. 2009;doi:10.1007/s00125-009-1453-1.
- Currie CJ. Diabetologia. 2009;doi:10.1007/s00125-009-1440-6.
- Gale EAM. Diabetes therapy and cancer. Presented at: the 45th Annual Meeting of the European Association for the Study of Diabetes; Sept. 29-Oct. 2, 2009; Vienna.
- Hemkens LG. Diabetologia. 2009;doi:10.1007/s00125-009-1418-4.
- Home PD. Diabetologia. 2009;52:2499-2506.
- Jonasson JM. Diabetologia. 2009;doi:10.1007/s00125-009-1444-2.
- Renehan AG. Lancet. 2008;371:569-578.
- Smith U. Diabetologia. 2009;doi: 10.1007/s00125-009-1441-5.