Ketoacidosis with canagliflozin

Overview

Erondu N et al.. Diabetes Care 2015 · PMID 26203064 · DOI 10.2337/dc15-1251

Ketoacidosis was rare (7 per 10,000 patients), often with insulin deficiency or triggers.

Does the study support the specific claim? Partly · Methodological quality: low

Original publication

Plain-language summary

Scientific in-depth review

1. Full reference

Authors

Erondu N, Desai M, Ways K, Meininger G

Journal

Diabetes Care

Year

2015

Identifier

PMID 26203064 DOI 10.2337/dc15-1251 · PMC4542268

Verification

verified – PubMed E-utilities efetch, 2026-10-25

Retracted

no

Corrections

none

Extent checked

full text

Conflicts of interest and funding

Supported by Janssen Research & Development (canagliflozin developer, with Mitsubishi Tanabe); all four authors are full-time Janssen employees; sponsor-funded editorial assistance (MedErgy).

2. Research question

Incidence of DKA with canagliflozin

3. Study design and population

Study design

Other – Sponsor-conducted pooled safety analysis (case series with incidence rates) of investigator-reported serious adverse events of DKA and related terms (MedDRA: DKA, ketoacidosis, metabolic acidosis, acidosis) across 17 completed and ongoing randomised controlled trials of canagliflozin; no formal adjudication; descriptive.

Population

Patients with type 2 diabetes in the canagliflozin RCT programme (history of type 1 diabetes or DKA excluded); ~31% on background insulin.

Country

International

Period

Data through 11 May 2015; mean exposure 1.4 years (~24,000 patient-years).

Sample size

17,596 patients (canagliflozin 100 mg 5,337; 300 mg 5,350; comparator 6,909); 12 patients with 13 unblinded serious DKA-related events; 3 further blinded CANVAS events excluded.

Inclusion and exclusion criteria

As described in the abstract; details not available in abstract.

4. Methods and endpoints

Exposure / intervention

Canagliflozin 100 mg or 300 mg.

Comparator

Pooled comparators (placebo, sitagliptin, glimepiride, metformin).

Endpoints

Serious adverse events of DKA and related events; case characteristics.

Statistical methods

Crude incidence proportions and rates per 1,000 patient-years; descriptive case review; no hypothesis testing.

5. Actual results

Results

DKA-related serious events: 4 (0.07%), 6 (0.11%) and 2 (0.03%) with canagliflozin 100 mg, 300 mg and comparator; 0.522, 0.763 and 0.238 per 1,000 patient-years. Six canagliflozin cases had autoimmune diabetes (LADA/type 1/GAD65+); excluding them: 0.130, 0.381 and 0.238 per 1,000 patient-years. Of 10 canagliflozin cases, 8 on insulin (4 with questionable insulin compliance) and most had precipitating factors (infection, surgery, pancreatic cancer, MI, gastroenteritis). Blood glucose at presentation >300 mg/dL in all 10 with values (347-571 mg/dL); 9 of 10 canagliflozin cases >250 mg/dL. Patients with DKA were older, more often male, leaner, with longer diabetes and higher HbA1c.

Absolute risks

0.07% overall (12/17,596); 0.52-0.76 vs 0.24 per 1,000 patient-years (canagliflozin vs comparator).

Authors' conclusion

DKA and related events occurred at low frequency, with an incidence consistent with limited observational data in type 2 diabetes (0.32-2.0 per 1,000 patient-years); underreporting possible; most cases had precipitating factors; low beta-cell reserve (incl. misdiagnosed LADA/type 1) plus acute illness proposed as mechanism.

6. Strengths

• Large randomised exposure base with comparator groups
• Detailed case-level clinical information (glucose, pH, C-peptide, precipitating factors)

7. Weaknesses

• Sponsor analysis by employee authors; no independent adjudication
• Ascertainment by investigator-reported adverse event terms - likely underreporting (authors)
• Very few events (12); no statistical comparison; 3 blinded events excluded
• Pooling of heterogeneous trials and comparators; short mean exposure (1.4 years)
• Comparisons with external background rates from data on file

8. Risk of bias

Orienting tool

none applicable (pooled safety case series from RCTs)

Type of assessment

structured assessment based on the full text by a single reviewer (no formal double review)

Full formal assessment

no

Overall risk of bias

high

Rationale

No standard RoB tool fits a pooled descriptive safety case series; single-reviewer structured judgement from the full text (formal assessment would need two reviewers). Rated high (was 'some_concerns') because of passive ascertainment, sponsor authorship and very few events; the qualitative case information (insulin deficiency, precipitating illness, high glucose) is nevertheless informative.

Full-text review (25 October 2026): Design clarified as sponsor pooled case series; group denominators, rates excluding autoimmune diabetes, insulin use/omission, precipitating factors and glucose at presentation (all >300 mg/dL) added; author employment (all Janssen) documented; structured judgement with domains, overall some_concerns -> high; methodological_quality moderate -> low; relevance to C-INS-08 and euglycaemic aspect of C-ODD-02 clarified.

9. Assessment of the statistical interpretation

Rare events with numerically higher rates on canagliflozin (about 2-3-fold) but too few for precise relative estimates; excluding autoimmune diabetes the difference largely disappears for 100 mg. Cases presented with high glucose, so this programme provides little evidence of euglycaemic DKA.

10. Relevance to the specific claim

Supports C-ODD-02 (SGLT2-associated DKA) but not its 'normal glucose' element within this programme; contradicts C-INS-08 (cases were driven by insulin deficiency/omission).

11. Independent contrary findings

12. Transferability to Germany and Europe

Multinational/US/UK data; transferable to EU practice.

13. Overall methodological judgement

Shows that DKA occurs rarely with canagliflozin, mainly with insulin deficiency (LADA/type 1, insulin omission) and acute illness, typically with marked hyperglycaemia. Supports SGLT2-associated DKA as a rare signal; underlines that insulin deficiency drives DKA.

14. Open questions

None beyond those stated.

15. Linked fact checks

• Diabetic ketoacidosis (DKA) does not kill anyone; it should be treated with corticosteroids and complete fasting from food and drink for days, without insulin and without intravenous fluids; giving insulin in this situation is 'the black disaster'. – Studies contradicting or limiting the claim
• SGLT2 inhibitors such as dapagliflozin (Forxiga) can cause ketoacidosis, sometimes with normal or low blood glucose. – Studies supporting the claim
• Every major class of glucose-lowering drug (metformin, sulfonylureas, thiazolidinediones, DPP-4 inhibitors, SGLT2 inhibitors, GLP-1 agonists) carries a documented signal of serious harm in the peer-reviewed literature. – Studies supporting the claim

Use by the Tayebat archive

Note: according to the archive, the studies were compiled by the archive operators. Whether Dr. Elawady himself cited this study is not established.

What the archive says about the study

Listed in the archive bibliography for 'oral-diabetes-drugs' without its own interpretation; used as support for the claim that every drug class carries documented harm.

A. What was actually studied?

Incidence of DKA with canagliflozin

B. What did the authors find?

12 of 17,596 (0.07%) with serious DKA-related events; 0.52-0.76 vs 0.24 per 1,000 patient-years; most on insulin, with precipitating factors or autoimmune diabetes; glucose >300 mg/dL in all cases with values.

C. What conclusions do the authors themselves draw?

DKA occurred at low frequency, consistent with background rates.

D. What does the archive derive from it?

Support for SGLT2 ketoacidosis.

E. Is this derivation scientifically justified?

Justified as a rare safety signal; the archive omits the low absolute frequency and that cases mostly involved insulin deficiency/omission and high glucose.

Adequacy of the interpretation (by the archive)

essentially correct but incomplete