Miriam E Tucker

September 22, 2015

STOCKHOLM — The sodium glucose cotransporter-2 (SGLT2) inhibitor canagliflozin (Invokana, Johnson & Johnson/Janssen) may help improve glycemic control in adults with type 1 diabetes who are inadequately controlled with insulin, a new 22-week phase 2 study suggests.

Canagliflozin is currently approved in the United States, Canada, and Europe for treating type 2 diabetes, but not type 1.

These results, from the first-ever phase 2 study of an SGLT2 inhibitor in patients with type 1 diabetes, were presented September 18 here at the European Association for the Study of Diabetes (EASD) 2015 Meeting by Robert R Henry, MD, chief of the section of endocrinology, metabolism, and diabetes and of the Center for Metabolic Research at the VA Medical Center, San Diego, California.

The study reported on type 1 patients who received 100 mg or 300 mg of canagliflozin or placebo for 18 weeks. Dr Henry concluded that future studies would likely need to test lower doses of the drug, due to the risk/benefit profile.

There was a numeric increase in severe diabetic ketoacidosis (DKA) with canagliflozin compared with placebo, but all cases occurred in the presence of precipitating factors. There have been recent warnings regarding DKA and SGLT2 inhibitors, although this study was initiated in May 2014 and completed June 5, 2015, so the bulk of it was done prior to these warnings

"Implementation of additional mitigation strategies in future studies may substantially reduce DKA in patients with type 1 diabetes treated with canagliflozin," Dr Henry said. He noted, however, that they had suspected this was a possibility, so investigation of "ketone-related adverse events" had been included in the study design.

Asked to comment on the new data, former EASD executive director and clinical diabetologist Viktor Jorgens, MD, told Medscape Medical News, "From what I've seen, I'm not convinced."

He took issue with the study design because it called for patients to reduce their insulin doses by 10% to 20% prior to taking canagliflozin (or placebo), which, in addition to raising the risk of DKA, also confounds the results by worsening control in both the active-drug and placebo groups.

"When you do it double-blind, then you have to reduce the insulin in people who don't need this reduction, and that means you worsen control….The following day the patient not getting the [active] drug will not do well.

"The effect…should be investigated in open studies, where people continue to treat themselves properly [with insulin] and add the [SGLT2 inhibitor]…I wouldn't look at the data with the blinded studies."

Indeed, Dr Henry acknowledged this problem when asked about it during the question-and-answer period and said the company is considering alternative designs for future studies.

But Dr Jorgens also pointed out that there are other side effects of the SGLT2 inhibitor class that could cause particular problems in people with type 1 diabetes, including increased rates of vaginal fungal infections in women and of urinary-tract infections and polyuria in both males and females.

Use of SGLT2 Inhibitors in Type 1 Diabetes "Makes Sense"

Dr Henry began his presentation by noting that, like other SGLT2 inhibitors, canagliflozin lowers plasma glucose by lowering the renal threshold for glucose and increasing urinary glucose excretion. "The mechanism is independent of insulin secretion or action, so it makes reasonably good sense to be able to use it in type 1 diabetes."

The study included an initial 2-week prerandomization phase in which continuous glucose monitor data were collected, 18 weeks of treatment with canagliflozin in doses of either 100 mg or 300 mg or with placebo, and a 2-week safety follow-up.

The 351 patients had all been diagnosed with type 1 diabetes for at least 1 year, had HbA1c levels of 7% to 9% with stable insulin treatment, and had not had DKA or severe hypoglycemic episodes within 6 months prior to randomization.

To avoid hypoglycemia, prior to randomization patients with HbA1c values below 8% were advised to reduce their basal insulin doses by 20%, and those with HbA1c values over 8% to drop their basal levels by 10%.

Of those randomized, 111 completed the entire study with the 100-mg canagliflozin dose, 110 with the 300-mg dose, and 107 taking placebo.

The primary end point, a reduction in HbA1c of at least 0.4 percentage points with no weight gain, was achieved by 37% of those taking 100 mg of canagliflozin and 41% of those taking the 300-mg dose, vs 15% in the placebo group.

A majority of those on canagliflozin — 84% with 100 mg and 96% with 300 mg — did not gain weight, compared with just 49% of the placebo group. The proportions achieving the HbA1c reduction of at least 0.4 percentage points were 45% and 43% for the 100-mg and 300-mg canagliflozin doses, respectively, compared with 23% for placebo.

The mean reductions in HbA1c from baseline were 0.27 and 0.24 percentage points for canagliflozin 100 mg and 300 mg, respectively, both statistically significant compared with the placebo group, in whom HbA1c levels rose 0.01 percentage points.

Both canagliflozin doses reduced body weight, by 3.1% and 5.1% for the 100- and 300-mg doses, respectively from baseline, again both statistically significant compared with the nonsignificant 0.3% weight increase with placebo.

Total insulin doses were reduced by 4.1 units per day with 100-mg canagliflozin and 7.6 units with 300-mg canagliflozin, both significant compared with the placebo group, in whom insulin doses increased slightly.

But Serious DKA, Hypoglycemia, and Infections Increased

Severe hypoglycemia occurred in 3% of patients taking 100-mg canagliflozin, in 7% of those on the 300-mg dose, and in 2% of the placebo group.

Overall, adverse events related to study drug occurred in 16% of the 100-mg group and 29% of those on 300 mg of canagliflozin, compared with 13% of those taking placebo. Serious adverse events occurred in 8% and 7% of the canagliflozin 100-mg and 300-mg groups, respectively, while there were none in the placebo group.

Ketone-related adverse events were seen in 5% of those on 100 mg and 9% of those taking 300-mg canagliflozin, vs none with placebo. Those events were serious DKA episodes requiring hospitalization in 4%, 6%, and 0% of cases, respectively.

All of the patients with serious DKA had precipitating factors that likely contributed to the event, including insulin-pump failure, skipped insulin dose, or concurrent illness, Dr Henry noted.

Urinary-tract infections developed in 4% and 5% of those taking canagliflozin 100 mg and 300 mg, respectively, vs 2% with placebo. Genital fungal infections were not seen in men but occurred in 4% and 21% of the women taking canagliflozin 100 mg and 300 mg, respectively, vs 6% with placebo.

Dr Jorgens pointed out that it would be important to know how the rates of vaginal infections were assessed, since they can be asymptomatic. "If [they are reporting only] the ones diagnosed by a gynecologist, you'll get a lower incidence."

Osmotic-diuresis–related adverse events occurred in 8%, 9%, and 3% of 100-mg canagliflozin, 300-mg canagliflozin, and placebo groups, respectively.

Lower Doses of Canagliflozin for Type 1 in Future Studies

Mitigation strategies are being considered for future trials in type 1 diabetes, including more frequent monitoring of serum ketones, dose interruption during periods of stress or illness, lower doses of canagliflozin, and further education of investigators and patients of potential risks.

"Clearly, it appears from these data that lower doses of canagliflozin will be necessary, as low as 50 or 25 mg," Dr Henry said.

Dr Jorgens advised, "I would try to do longer-term open studies not to have this initial worsening [due to reducing the insulin dose], because it's confusing everything."

However, he also pointed out that, even then, the high cost of canagliflozin might not be outweighed by the benefit.

Dr Henry said the findings of this study and a separate detailed report of each of the DKA cases will be published in upcoming issues of Diabetes Care.

In response to an audience member who questioned the way the insulin dosing was handled, he said, "That's why we do these studies, so that we can design the next set of studies to be more accurate. I'm confident we will have learned a great deal."

Dr Henry is on the advisory board of, is a consultant for, or has received grants from Alere, Amgen, AstraZeneca, Bristol-Myers Squibb, Boehringer Ingelheim, Elcelyx, Gilead, Intercia, Janssen, Merck, Novo Nordisk, Roche-Genentech, Sanofi, Valeritas, Vivus, Ligand, Abbott, Hitachi, Lilly Pharmaceuticals, and ViaCyte. Dr Jorgens has no relevant financial relationships.

European Association for the Study of Diabetes 2015 Meeting; Stockholm, Sweden. Presentation


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.