Insulin and cancer
Insulin has long been known to possess growth-promoting properties, including enhanced cell proliferation (i.e. it is mitogenic). Exposure to high levels of insulin promotes expansion of some cancer cell lines in vitro and cancer growth in some experimental situations. Insulin does not however cause malignant transformation of healthy cells (i.e. it is not mutagenic). Exogenous insulin is often needed to treat type 2 diabetes. In contrast with endogenous insulin, which is largely cleared by the liver, exogenous insulin is delivered into the systemic circulation and thus increases the exposure of peripheral tissues to its effects. Exogenous insulin might thus have the potential to promote the growth of pre-existing cancer foci in certain locations. In line with this, some observational studies report that cancer risk is increased in insulin-treated type 2 diabetes, as compared with other glucose-lowering therapies. This observation has not been confirmed in other studies, however, and the only cancer unequivocally linked to insulin use is carcinoma of the pancreas, almost certainly because this tumour causes worsening of diabetes (reverse causation). This entry reviews the evidence concerning insulin use and cancer; daughter pages consider the evidence relating specifically to human and analogue insulins.
Figure 1: Modified from Pollak M. Nature Rev Cancer 8:915-928, 2008Insulin and the insulin growth factors IGF-1 and IGF-2 share a common evolutionary origin, and insulin has growth-promoting as well as metabolic actions. The insulin-IGF axis plays a major role in promoting growth and development of those cancers which overexpress the receptor family on their surface, and hyperinsulinemia has been implicated as a possible causal factor linking obesity and diabetes to cancer (see Hyperinsulinemia and cancer).
It has been known for more than 50 years that insulin is mitogenic in vitro and can promote tumour growth in experimental animals. Furthermore, some observational studies have reported increased cancer mortality in insulin-treated type 2 diabetes. These considerations have given rise to concerns that insulin use might promote the development of some cancers, and/or be associated with increased mortality.
The possibility that insulin use might be associated with increased cancer incidence and/or mortality is both sensitive and controversial. The issues are complex, and problems have arisen both with regard to measuring and interpreting the association. Some relevant considerations are as follows:
- The comparison between insulin and oral agents in the treatment of diabetes does not distinguish between a possible protective effect of oral agents such as metformin as against an adverse effect of the insulin. Equally, oral agents which promote insulin secretion might have risks similar to those of insulin itself.
- There are many potential confounders: people on insulin, for example, tend to be older, to have a longer duration of diabetes and greater hyperglycemic exposure than those on tablets. BMI is another potential confounder.
- It is important to distinguish between increased cancer incidence and increased cancer mortality. For example, those on insulin have a reduced overall survival, and reduced survival with cancer might simply reflect this. Alternatively, their worse metabolic condition might have an adverse effect on the outcome of cancer treatment. Cancer mortality is an unreliable index of susceptibility to cancer.
- The timing and duration of exposure must be taken into consideration. If insulin accelerates cancer progression, for example, one would expect the incidence of cancer to be positively associated with the duration of exposure to insulin.
- Conversely, reverse causation must be considered: i.e. substances produced by the tumour might promote metabolic deterioration, thus generating the need for insulin therapy. (See Pancreatic cancer).
Although a number of observational studies have reported an increased overall cancer risk in insulin-treated diabetes, a recent meta-analysis found no increase in insulin treated patients other than in the recently diagnosed, in whom a peak cancer incidence was seen following diagnosis but diminished with longer duration of diabetes. This effect reached significance for pancreatic cancer with both human and glargine insulin, and was suggestive of reverse causation, although detection bias might also have contributed.
Previous subanalyses of individual cancers show no relation between insulin treatment and breast cancer. Inhaled but not injected insulin has been associated with lung cancer in small studies, and the association with colorectal cancer reported from several studies remains controversial. In sum, this area remains controversial but there is currently no convincing evidence to support a causal relationship between insulin treatment and cancer (see Human insulin and cancer).
Analogue Insulin and Cancer
One of the first insulin analogs, AspB10, was withdrawn from development because of an excess of mammary cancers in rats. This appeared to be related to increased mitogenicity relative to human insulin, and there have been concerns about insulin glargine which shares this property. Insulin glargine is however partially broken down into less mitogenic molecules following injection, so its in vivo properties cannot be inferred from laboratory studies. One epidemiologic analysis suggested an overall increase in cancer risk in patients taking insulin glargine, but subsequent epidemiologic and trial data have largely dispelled this concern. Furthermore, no overall increase in cancer risk was shown in ORIGIN, a large prospective trial which included patients on glargine, although interpretation of this trial is complicated by concurrent use of other medications (e.g. metformin) and was therefore not a direct comparison between insulin and other therapies.
Recent systematic review and meta-analysis found that both human insulin and insulin glargine were associated with recent onset carcinoma of the pancreas, quite possibly due to reverse causation. Insulin glargine was also observed to be associated with a reduced rate of colorectal cancer, but an increased rate of prostate cancer, but these findings need to be interpreted with caution.
Several epidemiologic analyses have suggested a small increase in risk of post-menopausal breast cancer in glargine users, whereas other studies have been negative. The possibility of a small excess risk of this cancer cannot therefore be excluded (see Insulin analogs and cancer).
Overall, current evidence suggests that insulin treatment (animal, human or analogue) is not causally related to any form of cancer. Notably, "second generation" studies have tended to produce smaller or negative risk estimates as compared with some of the earlier studies. Nonetheless, further studies will be needed before this issue can be clarified beyond doubt.
The most striking development has been the observation that cancer risk appears to be negatively related to duration of exposure. This might be due to detection bias (increased contact with medical staff) or reverse causation (tumour-related metabolic decompensation prompting the prescription of insulin), but the observation argues against a causal association between insulin treatment and the risk of development or progression of cancer.
^ Renehan AG et al. Diabetes and Cancer (2): evaluating the impact of diabetes on mortality in patients with cancer. Diabetologia 2102;55:1619-32
^ Colmers IN et al. Insulin use and cancer risk in patients with type 2 a systematic review and meta-analysis of observational studies. Diabetes Metab 2012;485-506