Many therapeutic agents can predispose to or precipitate diabetes, especially when pre-existing risk factors are present, and these may cause glucose control to deteriorate if administered to those with existing diabetes. They may act by increasing insulin resistance, by affecting the secretion of insulin, or both. For convenience, these agents may be subdivided into widely used medications that are weakly diabetogenic, and drugs used for special indications that are more strongly diabetogenic. Examples of the former include antihypertensive agents and statins, and examples of the latter include steroids, antipsychotics and a range of immunosuppressive agents. There are also a number of known beta cell poisons including the insecticide Vacor, alloxan and streptozotocin.
A wide range of therapeutic agents may affect glucose tolerance, and the list of known or suspected drugs is lengthy. This entry summarizes evidence concerning the agents most frequently implicated.
Widely used medications
A number of drugs used to reduce cardiovascular risk also predispose to the development of diabetes. These include the thiazide diuretics, beta-blockers and statins. It should however be appreciated that these are commonly offered to individuals who are at increased risk of diabetes by virtue of risk factors such as obesity and hypertension, and that risk association does not necessarily mean causation.
Thiazides: Thiazide diuretics revolutionized the treatment of hypertension in the 1960s, but were soon noted to increase the risk of diabetes. Subsequent experience showed that that this risk is greatly reduced by low-dose therapy, whose benefits therefore outweigh its risks. The thiazides have a weak inhibitory effect upon release of insulin from the beta cell. This effect is so marked in one member of the family, diazoxide, that this is used to control excessive insulin secretion by unresectable insulinomas.
Beta-blockers: These impair insulin release, especially agents that are not selective for the β1-receptor subtype. Several studies have linked chronic use of β-blockers with an increased risk for the development of diabetes. The Atherosclerosis Risk in Communities (ARIC) study found that the risk of diabetes in was increased by 28% in those taking a beta-blocker for hypertension over a 6-year period, as compared with other medications. The risk conferred by non-selective beta blockers may be enhanced if these are combined with a thiazide diuretic.
Statins: Meta-analysis has shown an excess risk of ~9% of progression to diabetes in those taking statins. A more recent comparison of high-dose versus moderate dose statin therapy showed that statin use was associated with an increased risk of diabetes equivalent to 2 cases per 1000 patient/yrs treated, but resulted in 6.5 fewer cardovascular events per 1000 patient/yrs.
Drugs conferring a high risk of diabetes
Steroids: Steroid drugs can induce a form of iatrogenic Cushing's syndrome, and are probably the most widely used drugs which confer a high risk of diabetes. A Canadian paediatric study found that 0.4% of childhood diabetes was ascribed to medication, and 55 of 56 children in this category were on steroids. Monogenic diabetes was diagnosed in only 0.2% of children.
Anabolic steroids (synthetic androgens mimicking testosterone or dihydrotestosterone) should not be confused with glucocorticoids. Although they have many unwanted effects, diabetes is not among them.
The association between anti-psychotic agents and diabetes is complex and disputed. Diabetes is, for example, considered to be 2-3 times as common in people with schizophrenia as in the background population, probably because of lifestyle factors such as obesity and lack of exercise, although a family history of diabetes is common in those with schizophrenia. This increase in baseline risk must be taken into account when considering the effect of any anti-psychotic agent upon risk of diabetes.
Although the effect of anti-psychotic drugs in general may have been over-estimated by studies which have not corrected for underlying predisposition to diabetes, there is little doubt that they are associated with an increased risk of diabetes. An unexplained feature of their action is that presentation in diabetic ketoacidosis is relatively common in individuals who otherwise appear to have type 2 diabetes. Women are more commonly affected than men.
Since these drugs predispose to weight gain and are not known to affect insulin secretion, their diabetogenic effect was assumed to be due to weight gain and insulin resistance. Weight gain does not in itself provide a sufficient explanation of the phenomenon, however, and a unified mechanistic explanation is still lacking.
"Atypical" antipsychotic agents were introduced in he attempt to reduce the high frequency of extrapyramidal side effects associated with use of conventional antipsychotics. Although they have brought undoubted benefits to otherwise untreatable patients, they are ~ 50% more likely to result in diabetes.
The agents most commonly implicated have been clozapine, olanzapine and risperidone; other agents should not however be assumed to be safer until more experience has been obtained concerning their use.
A number of immunosuppressive agents affect β-cell growth, proliferation and function. The most potent of these are the calcineurin inhibitors, which also affect the nuclear factor of activated T-cells (NFAT) pathway; examples are tacrolimus and ciclosporin A. This limits the application of these agents in the prevention of type 1 diabetes, and helps explain the high prevalence of post-transplantation diabetes, observed in 13.4% of patients following solid-organ transplantation (16.6% for tacrolimus vs 9.8% for ciclosporin).
Protease inhibitors such as ritonavir form an important element in highly active antiretroviral therapy (HAART) and strongly predispose to the development of diabetes by increasing insulin resistance and possibly by other direct effects upon the β-cell. The Multicenter AIDS Cohort Study found that 14% of HAART treated men developed diabetes, with an annual rate of 4.7 per 100 person-years as against 1.4/100 person-years for seronegative controls (rate ratio = 4.11; 95% confidence interval, 1.85-9.16, adjusted for age and body mass index)
Pentamidine, used by iv injection for the treatment of _Pneumocystis carinii _pneumonia, can cause transient hyperinsulinaemia associated with hypoglycaemia, followed by persistent beta cell failure requiring insulin treatment.
Nicotinic acid, used as a lipid-lowering agent, can induce diabetes secondary to increased insulin resistance.
^ Wolff FW et al. Diabetogenic activity of long-term administration of the benzothiadiazines. JAMA 1963;185:568-74
^ Greiss TW et al. Hypertension and antihypertensive therapy as risk factors for type 2 diabetes mellitus. N Engl J Med 2000;342:905.
^ Preiss D et al. Risk of incident diabetes with intensive-dose as compared with moderate-dose statin therapy: a meta-analysis. JAMA 2011;305:2556-64
^ Holt RIG, Peveler RC. Association between anti-psychotic drugs and diabetes. Diabetes, Obesity and Metabolism 2006;8:125-135
^ Seeman MV. Secondary effects of antipsychotics: women at greater risk than men. Schizophr Bull 2009;35:937-48
^ Sowell MO et al. Hyperglycemic clamp assessment of insulin secretory responses in normal subjects treated with olanzapine, risperidone, or placebo. JCEM 2002;87:2918-23
^ Heisel O et al. New onset diabetes mellitus in patients receiving calcineurin inhibitors: a systematic review and meta-analysis. Am J Transplant 2004;4:583-95
^ Brown TT et al. Antiretroviral therapy and the prevalence and incidence of diabetes mellitus in the Multicenter AIDS Cohort Study. Arch Intern Med 2005;165:1179-84