Close X

Drugs that cause diabetes

18th May 2022
female about to swallow a pill held to the lips

Steroids, antipsychotics, antiretroviral therapy for HIV and immune checkpoint inhibitors against cancer: all of these may cause diabetes. A new review in Diabetologia looks at these four classes of diabetogenic drugs and advises careful selection and monitoring to reduce the risk. Dr Susan Aldridge reports.

Various drugs commonly used in clinical practice can cause hyperglycaemia or even new-onset diabetes. They may also worsen glycaemic control in people who already have diabetes. This can occur by various different mechanisms, depending on the drug involved, including promotion of weight gain, impaired insulin secretion, increased hepatic glucose production and cytotoxicity towards beta cells.

The four main drug classes of particular concern in this respect are: glucocorticoids, antipsychotics, antiretroviral therapy and immune checkpoint inhibitors. It’s also worth noting that a meta-analysis has shown a 9 to 33% increased risk of new-onset diabetes with statins – which are very widely used in people with, or at risk of, diabetes. The deterioration in glucose control with statins is, however, generally mild and the cardiovascular protection they afford outweighs the diabetes risk.

Drug-induced diabetes may be reversible once the medication is discontinued, or it may persist, depending on the characteristics of both the medication and the patient. Recommendations for the management of drug-induced diabetes are generally the same as for any other type of diabetes and focus upon lifestyle modification and, if necessary, glucose-lowering treatment, starting with metformin.


Glucocorticoid- or steroid-induced diabetes has been known of for over 60 years. Its prevalence varies from 2% in outpatient general practice to 32% for those on steroids for rheumatoid arthritis or following an organ transplant and it generally occurs early in the course of treatment. Glycaemic response to steroids injected into joints is variable, depending on the number of joints injected and the type and dose of steroid. Steroids can induce diabetes in a number of ways, including by increasing hepatic glucose production, activating enzymes involved in glucose metabolism and reducing insulin sensitivity in muscle.

There is no consensus on how to manage steroid-induced diabetes. Metformin should be prescribed for mild hyperglycaemia, progressing to sulphonylureas or insulin, if it is more severe. When the steroids are being reduced, or stopped, it’s important to avoid hypoglycaemia.


There are two main types of antipsychotic drug, depending on their mode of action – the first-generation or typical antipsychotics and the second-generation or atypical antipsychotics. The prevalence of diabetes among people taking antipsychotics is around 10%. A meta-analysis covering 438,245 people with severe mental health problems showed that the prevalence of diabetes prior to receiving antipsychotic therapy was 2.9%, increasing to 11.3% after treatment was started.

All of the antipsychotics can lead to significant weight gain which, in itself, increases the risk of diabetes. However, the typical antipsychotics come with a lower risk of diabetes. Among the atypical antipsychotics, clozapine and olanzapine have the highest risk, followed by risperidone and quetiapine, while aripiprazole and ziprasidone (which have a different mode of action from the first four) have the lowest risk. This ranking can help guide the selection of an antipsychotic for people already at risk of diabetes.

A rapid rise in blood glucose and incidence of diabetes is observed in the atypical antipsychotics even before weight gain, suggesting a direct effect of the drug on diabetes risk. There have been reports of diabetic ketoacidosis (DKA), suggesting an adverse impact on beta cells and insulin secretion.

Lifestyle advice to minimise weight gain and the development of diabetes should be given when an antipsychotic is prescribed. Monitoring blood glucose for 12 weeks following initiation of therapy, and annually thereafter, is also recommended for those without diabetes. And if blood glucose does rise, a switch to a less diabetogenic antipsychotic might be appropriate. Finally, given the high prevalence of obesity, the use of glucose-lowering agents that cause weight loss, like the GLP-1 receptor agonists or SGLT-2 inhibitors, could be helpful. In one study, GLP-1 receptor agonists were shown to reduce weight gain in people treated with clozapine and olanzapine, with accompanying improvement in blood glucose levels.

Antiretroviral therapy

There are four classes of antiretroviral therapy (ART) that are used to treat HIV infection. The prevalence of ART-induced diabetes has evolved in recent years, mainly because of change in usage of these drugs. The global cumulative incidence worldwide averaged 4.9%. Factors associated with ART-induced diabetes include being overweight or obese, dyslipidaemia and disturbances of fat distribution (lipoatrophy and/or lipohypertrophy). Diabetes is generally asymptomatic in people who have HIV, so screening is important after ART initiation or modification.

The mechanisms by which ART cause diabetes are complex and not completely understood. They do cause increased adiposity, which has a major impact on insulin sensitivity. There is also evidence for alteration in beta cell function.

Fasting plasma glucose should be measured every six to 12 months in all individuals treated for HIV. Both lifestyle modification and metformin have been shown to improve symptoms of metabolic syndrome in these individuals. And, as with the antipsychotics, avoiding the ART medications with higher diabetes risk, wherever possible, is important.

Immune checkpoint inhibitors

In cancer treatment, immune checkpoint inhibitors (ICIs) lead to activation of the immune system, which overcomes tumour-induced immunosuppression. However, these drugs also lead to a number of immune-related adverse events – including a form of new-onset diabetes that resembles autoimmune type 1 diabetes.

A retrospective study from the FDA identified 735 cases of new-onset diabetes associated with ICIs between 1 January 2015 and 31 December 2019, with an estimated overall incidence of 1.27%. Of these, 183 had fulminant type 1 diabetes, 338 presented with DKA, 183 had life-threatening outcomes and 41 died. There was also a consistent increase in reporting over this period of time. There are differences between the ICIs, with those treated with anti-CTLA-4 therapy (ipilimumab) being less likely to develop diabetes than those on anti-PD-1 (pembrolizumab, nivolumab) or anti-PDL-1 (atezolizumab, avelumab, durvalumab) therapy. According to the World Health Organisation, only 12 of 283 cases of new-onset diabetes were treated with anti-CTLA-4 therapy alone, while anti-PD-1 monotherapy accounted for 76% of all these cases.

ICI-associated diabetes often resembles type 1 in its pathophysiology and clinical features. Its presentation varies from asymptomatic hyperglycaemia through to classical symptoms of diabetes like polydipsia and polyuria or even life-threatening DKA. The time from ICI initiation to diabetes diagnosis also varies, ranging from a few days to several months.  

In this form of diabetes, ICI activation of the immune system is directed towards the beta cells, although the precise mechanism is not well understood. It has been suggested that people with islet-cell antibodies could be more at risk. Studies so far show only half those with ICI-related diabetes were positive for at least one antibody, with anti-GAD being the most common. 

People about to start ICI therapy should be told about the potential risk of diabetes and educated about the symptoms and signs of hyperglycaemia and DKA. Blood glucose and HbA1c should also be checked regularly. People with mild hyperglycaemia can continue on their therapy, but with close monitoring. If hyperglycaemia is moderate, continuation of the ICI is feasible, if they start on insulin. And if hyperglycaemia is severe and, of course, in the case of DKA, the ICI should be stopped, at least temporarily, and insulin treatment is essential. After recovery, restarting ICIs can be considered, with close glucose monitoring. ICI-related diabetes is irreversible and will usually require life-long insulin therapy. Management should be multidisciplinary, including endocrinologists and oncologists.


Drug-induced diabetes is a growing concern, owing to the increased use of glucocorticoids, antipsychotics, ART and ICIs. It’s not clear whether the resulting hyperglycaemia arises directly from drug action or whether the drug somehow uncovers pre-existing diabetes in high-risk individuals. Further research is needed, but there are multiple mechanisms at play, which differ between the four classes of drug, although increased adiposity clearly plays a role with the antipsychotics. 

Despite the prospect of new-onset or worsening diabetes, the benefits of the four classes of drugs described here generally outweigh the risks where prescribed appropriately. But clinicians should prescribe the medication with the lowest diabetes risk, wherever possible. And careful monitoring – particularly of high-risk individuals – is essential if they are receiving agents known to impair glucose metabolism. 

To read this paper, go to: Fève B, Scheen A. When therapeutic drugs lead to diabetes. Diabetologia online 4 March 2022.  

Any opinions expressed in this article are the responsibility of the EASD e-Learning Programme Director, Dr Eleanor D Kennedy.