Is the recent concern over insulin glargine and malignancy warranted?

Over the past few months, there has been a major — or minor — storm over the issue of insulin glargine and malignancy. The controversy began with the June publication of Diabetologia, which included four observational studies on this association.

The first study, submitted by Hemkens and colleagues, included 127,000 patients in a German health insurance fund who were using human insulin or insulin analogues glargine (Lantus, Sanofi-Aventis), aspart or lispro; 24,000 were taking insulin glargine. The primary outcome looked at the risk for malignancy, and secondary measures looked at all-cause mortality; neither endpoint was increased. However, when the hazard ratio was analyzed according to dose, the researchers reported an excess increase in cancer in those receiving more than 40 IU or 50 IU of insulin glargine. The authors acknowledge the lack of data on patients, such as BMI and duration of diabetes, which may affect any interpretation of the study. In addition, many patients on combined insulin therapies were not included in the analysis.

Derek LeRoith

The European Association for the Study of Diabetes requested similar studies from other registries, which formed the basis for the other three articles. Results of a Swedish-based study of 114,000 patients demonstrated decreased overall cancer rates and reduced overall mortality; nevertheless, there was a relative risk of 1.99 (95% CI, 1.31-3.03) for breast cancer among patients taking insulin glargine alone, but no increased risk with insulin glargine in combination with other insulins. The third study (36,000 patients in Scotland) and fourth study (63,000 patients in the United Kingdom) did not confirm these results. Interestingly, patients taking metformin had a lower risk for developing cancer than those on sulfonylurea and insulin, confirming other epidemiological study data.

Questions remain

Many questions have been asked after these results were published. Are the data from observational studies meaningful? Is there a potential relationship between insulin and cancer? In addition, what should health care professionals and patients do?

Numerous epidemiologists and well-renowned trialists have attacked the studies for a number of fundamental limitations. Studies that include patients in registries are purely observational and lacking randomization. Very limited data are presented on the patient characteristics and allocation bias, short exposure to insulin and misclassification, to name a few, and add to the difficulty in interpreting such results. In support of this, an analysis has been performed on a postmarket, controlled study comparing NPH insulin with insulin glargine in more than 1,000 patients; after five years, no increased risk for cancer was observed in the patients using insulin glargine.

The association of diabetes and cancer, and even the use of insulin and cancer, has been studied for many years, both epidemiologically and at the basic science level, to attempt to understand if there is causality. Both obesity and type 2 diabetes (not type 1 diabetes) are associated with an increased risk for cancer and increased mortality rates. Indeed, bariatric surgery for obesity reduces the rates for developing cancer and cancer- related mortality. 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. Thus, there is a developing theme that obesity and type 2 diabetes are associated with cancer risk — and the exact mechanisms need to be determined. Some epidemiological studies have shown that increased risk for breast cancer, for example, is associated with insulin/C-peptide levels in obese individuals.

On the other hand, the relationship between insulin therapy and cancer in type 2 diabetes is more tenuous and purely observational. Many of the studies are from registries that predate the use of the newer, long-acting analogues such as insulin glargine and insulin detemir. The studies suggest that sulfonylureas and insulin in patients are associated with higher cancer risk and that metformin reduces this risk. However, these observational studies do not deal with other confounding variables, such as severity and length of disease, which may speak to the dyslipidemia, inflammatory cytokine levels, hyperglycemia and insulin resistance.

At the basic level, all insulin analogues have altered characteristics that affect their binding to the insulin receptor and the insulin-like growth factor 1 receptor; the former being the metabolic receptor, the latter being a mitogenic receptor. Indeed, AspB10 insulin analogue caused cancer in preclinical studies and was immediately removed from development. Although insulin glargine has higher affinity for the IGF-1 receptor, this affinity is many fold lower than IGF-1 itself. Studies have suggested that insulin glargine and determir are more mitogenic than human insulin in cell culture, but less than IGF-1 itself. On the other hand, insulin glargine given subcutaneously is metabolized to two active metabolites (M1 and M2) that reach the insulin receptor to enhance metabolic function. Furthermore, the fast dissociation rates of insulin glargine and, importantly, the M1 metabolite from the insulin receptor should not result in excessive mitogenic activity.

What to do?

Finally, for the heath care professional, the important question is, “What to do?” Although the study investigators would not presuppose to suggest what health care professionals and patients should do at this stage, the major organizations (FDA, American Diabetes Association, American Association for Clinical Endocrinologists) and their European counterparts have all stated that the four studies published in

Diabetologia are confusing, unconvincing, do not demonstrate causality and uniformly suggest that health professionals and patients do not change their use of insulin — nor is there any reason to stop using insulin glargine. Keep in mind the value of insulin in both type 1 and type 2 diabetes and the extra advantage of basal insulins in our armamentarium.

Furthermore, the studies confirm the value of metformin with insulin in most patients with type 2 diabetes whenever possible.

As many experts have stated, postmarketing vigilance is very appropriate with the introduction of any new medication and antidiabetic agents should be considered similarly.

Derek LeRoith, MD, PhD, is Chief of the Division of Endocrinology, Diabetes and Bone Diseases at Mount Sinai School of Medicine.

Exciting, just-published data in Cancer Research (Sept 14. online) from Harvard investigators adds punch to this editorial. Metformin has just been reported to destroy breast cancer stem cells in vitro. Moreover, when added to chemotherapy, metformin markedly inhibited regrowth of tumors in mice, whereas chemotherapy or metformin alone was ineffectual. A phase-2 study of metformin as adjuvant therapy in breast cancer is in the works.

– Harry S. Jacob, MD

HemOnc Today Chief Medical Editor

For more information:

• Colhoun HM. Diabetologia. 2009;doi:10.1007/s00125-009-1453-1.
• Currie CJ. Diabetologia. 2009;doi:10.1007/s00125-009-1440-6.
• Hemkens LG. Diabetologia. 2009;doi:10.1007/s00125-009-1418-4.
• Jonasson JM. Diabetologia. 2009;doi:10.1007/s00125-009-1444-2.
• Rosenstock J. Diabetologia. 2009;doi:10.1007/s00125-009-1415-7.