A recent study, from New Zealand, has added further evidence to support a link between taking statins and new onset diabetes.

Data from the national primary care electronic database was used to create a prospective cohort study to compare the incidence of new diabetes in groups of patients taking statins and antihypertensive drugs.

The study was started in 2005 when patients aged 40 to 60 years who had received a first prescription for one of the study drugs that year were taken as the initial cohort. The age limit of 60 was chosen to limit the inclusion of cardiovascular drugs prescribed for other conditions (e.g. heart failure).

Participants were assigned to four groups. Patients on antihypertensives thought to be associated with an increase in diabetes risk – thiazide diuretics and beta-blockers (TB); those on antihypertensives thought less likely to induce diabetes – ACE inhibitors, angiotensin II receptor blockers and calcium channel blockers (AAC); patients on statins; and a control group, comprising a healthy population in terms of cardiovascular risk, prescribed diclofenac.

Patients who were receiving any of the study drugs before 2005 were excluded as were those on oral corticosteroids, or with any prior prescription of drugs for diabetes.

The outcome measure was the development of type 2 diabetes. The marker for this was a prescription for metformin, the first-line diabetes treatment in New Zealand.

There were 7,140 patients in the diclofenac group, 5,769 patients in the higher-risk antihypertensive group, 6,565 patients in the lower-risk antihypertensive group and 12,612 patients in the statin group.

The demographics of the groups varied. There were more women in the high-risk antihypertensive group and more men in the statin group. Patients in the cardiovascular risk modifying groups were older than those in the diclofenac group, which also contained more Maori patients. The differences were allowed for in the multivariable analysis.

The groups were followed up from 2005 to 2011. Loss to follow-up could not be directly monitored by prescription analysis, and was measured by proxies of medication persistence, death and emigration rates. Persistence with medicine was taken as at least one prescription a year. The highest drop out rate in this respect was in the diclofenac group, which is unsurprising as the drug is often given for temporary symptoms.

Overall 710 patients received first prescriptions for metformin in the study period. The incidence rates for first prescription of metformin for the diclofenac, TB, AAC and statin groups were 2.5, 2.8, 4.2 and 5.5 cases per 1,000 person-years respectively.

In the multivariable analysis, patients started on statins had the highest risk of receiving a first prescription of metformin in six years of exposure, compared with the control group, see below.

• Statins: hazard ratio (HR) 3.31

(95% CI:2.56-4.30; P < 0.01)

• AAC: HR 2.32 (95% CI: 1.74-3.09; P < 0.01)

• TB: HR 1.59 (95% CI:1.15-2.20; P < 0.01)

There was a lower risk in female patients, and a higher risk in patients from a non-European ethnic background.  Age was not related to risk.

The authors discuss some possible limitations of their study, in particular the inability to compare cohort characteristics such as BMI and family history. There was no distinction in the study between doses and formulation of the statins.

It is also possible that the patients receiving statins may have been exhibiting the metabolic syndrome and therefore automatically at risk of developing diabetes.

However, the study results confirm that taking cardiovascular disease risk modifying drugs increases the risk of diabetes compared with taking diclofenac, and that statins increase this risk more than any family of antihypertensive agent.

The results support the conclusion of several recent meta-analyses that have suggested that statins are associated with diabetes.

The NAVIGATOR trial studied the effect of valsartan and nateglinide on the progression from impaired glucose tolerance to diabetes, and found no evidence of benefit from either drug.

A further analysis of the data from this trial concluded that statins increased the risk of diabetes in patients with impaired glucose tolerance and other cardiovascular risk factors.1 Thiazide diuretics were also associated with an increased risk, but beta-blockers were not.

As GPs we are watching the evolving debate about the use of statins for primary prevention.

The evidence from this paper seems to be consistent with other published studies, and suggests that taking statins is associated with an increased risk of new diabetes.

The safest intervention for our patients at cardiovascular risk is for them to improve their lifestyle. If we do offer statins, we should include this concern in the discussion with our patients.

Dr Matthew Lockyer