Dr. Smith and Dr. Gale reporting having no conflict of interest. Dr. Jellinger is on the speakers' bureau for several pharmaceutical companies, including Novo Nordisk, Amylin Lilly, and Takeda.
Four Registry Studies on Insulin Glargine and Cancer Yield Different Results
The first of the four studies, reporting a dose-dependent increase in cancer risk with glargine compared with human insulin in a study of more than 100,000 patients, was submitted to Diabetologia last year, Dr. Gale and Dr. Smith explained in their editorial. Its findings suggested that, compared with people using similar doses of human insulin, out of every 100 people who used Lantus insulin over an average of about 1.5 years, 1 additional person was diagnosed with cancer (Diabetologia 2009 June 26;
However, because of the study's limitations and its enormous implications, the European Association for the Study of Diabetes held the article and requested the three other analyses of data from national diabetes registries in Sweden, Scotland, and Wales to see if the findings could be replicated. All four studies were published simultaneously.
The German Study
The first study included 127,031 insulin-treated diabetic patients from a national health insurance database, all without known malignant disease at baseline and who had received first-time treatment exclusively with either human insulin (95,804), lispro (3,269), aspart (4,103), or glargine (23,855) exclusively (Diabetologia 2009 June 26;
At a mean follow-up of 1.63 years, the unadjusted risk for developing a malignant neoplasm was actually lower in those using all three analogues. But because patients taking a combination of human and analogue insulins had been excluded from the study, the glargine patients were using much lower overall doses than were those on human insulin (median 22 vs. 37 IU/day). After adjustment for daily dose, the risk was significantly increased for those taking glargine, compared with those taking human insulin, with hazard ratios of 1.09 for 10 IU/day, 1.19 for 30 IU/day, and 1.31 for 50 IU/day. No such increases were seen with either of the short-acting analogues lispro or aspart.
The Swedish Study
The Swedish study followed 114,841 individuals aged 35-84 years who had a prescription dispensed for insulin during the latter 6 months of 2005 and linked them with cancer registry data during 2006-2007.
After adjustment for age and sex, the overall rate of malignancy was not elevated for glargine monotherapy, compared with other insulins, nor were the specific rates of prostate or gastrointestinal cancers. However, after adjustment for a variety of other factors, women who took glargine had a significantly higher rate of breast cancer than did women who took other types of insulins as monotherapy (relative risk 1.97).
The Scottish Study
The Scottish group examined a total of 36,254 people using insulin over a 4-month period from a database that includes almost every individual in the country with diabetes. In a 4-year follow-up, the overall group of 3,959 using glargine had the same incidence of all cancers as did those not using glargine (hazard ratio 1.02). However, the subset of 447 patients using glargine as their sole insulin had a significantly higher incidence of all cancers than did the 32,295 using other insulins only (HR 1.55), while those using glargine with other insulins had a slightly lower incidence (HR 0.81).
Overall, there was no increase in breast cancer rates associated with insulin glargine use (HR 1.49), and no differences in breast cancer or all cancers were seen among type 2 diabetic insulin users, they reported.
The authors noted important differences in baseline characteristics between the insulin treatment groups. For example, patients using glargine alone were older than were those on glargine plus other insulins (68 vs. 41 years) and users of other insulins (60 years). Those on glargine alone also were more overweight, more hypertensive, and more likely to be on oral glucose-lowering drugs.
The Welch Study
This study was a retrospective cohort study of 62,809 people treated in U.K. general practices that participate in a national health information network (Diabetologia 2009 June 26;
These patients were all above age 40 at the time of diagnosis and were divided into four treatment groups: monotherapy with metformin or sulfonylurea, combination therapy with the two oral agents, or insulin. The insulin users were further subdivided into users of glargine, long-acting human insulin, biphasic analogue, or human biphasic insulin.
Metformin monotherapy carried the lowest risk of cancer, consistent with previous data suggesting that metformin may have a protective effect against malignancy. Compared with that group, the hazard ratio was 1.08 for those taking metformin plus sulfonylurea, 1.36 for sulfonylurea monotherapy, and 1.42 for insulin-based regimens. The latter finding is also consistent with other data suggesting that insulin use overall increases the risk for malignancy, Dr. Smith and Dr. Gale noted.