Swedish study of glargine and cancer

Overview

Jonasson JM et al.. Diabetologia 2009 · PMID 19588120 · DOI 10.1007/s00125-009-1444-2

Overall cancer rates were the same; women using only glargine had more breast cancer (25 cases), which the authors say may be chance.

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

Original publication

Plain-language summary

Scientific in-depth review

1. Full reference

Authors

Jonasson JM, Ljung R, Talbäck M, Haglund B, Gudbjörnsdòttir S, Steineck G

Journal

Diabetologia

Year

2009

Identifier

PMID 19588120 DOI 10.1007/s00125-009-1444-2

Verification

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

Retracted

no

Corrections

none

Extent checked

full text

Conflicts of interest and funding

No funding or competing-interest statement in the available full text; study performed at the request of the EASD by Swedish university and National Board of Health and Welfare researchers.

2. Research question

Short-term cancer incidence in glargine users (requested by EASD)

3. Study design and population

Study design

Cohort study – Nationwide fixed-cohort study linking seven Swedish population registers (Prescribed Drug, Cancer, Causes of Death, National Diabetes, Patient, Medical Birth, Education registers); performed at the request of the EASD.

Population

All residents aged 35-84 with >=1 insulin prescription dispensed 1 July-31 December 2005 and alive on 1 January 2006; 61% classified type 2, 15% type 1, 23% unknown diabetes type.

Country

Sweden

Period

Exposure window July-December 2005; follow-up 1 January 2006 - 31 December 2007 (2 years).

Sample size

114,841: glargine alone 5,970 (5.2%), glargine plus other insulins 20,316 (17.7%), other insulins only 88,555 (77.1%). Breast-cancer analyses: 47,776 women.

Inclusion and exclusion criteria

Persons with a prior record of the outcome cancer (since 1958) excluded from the respective analysis; loss to follow-up defined as no dispensed prescription in 2008.

4. Methods and endpoints

Exposure / intervention

Glargine alone, or glargine with other insulins, classified from prescriptions dispensed in the second half of 2005 (no dose or duration information in main analysis).

Comparator

Users of insulins other than glargine.

Endpoints

All malignancies (incl. some benign/precancerous lesions per Cancer Register routines), all malignancies plus in situ, breast, prostate and gastrointestinal adenocarcinoma; also mortality and acute myocardial infarction.

Statistical methods

Poisson regression (incidence rate ratios, reported as RR) with person-years offset; adjustment for age, sex, BMI, smoking, age at onset of diabetes, CVD, age at first birth, oestrogen, metformin, education; missing values as separate category; sensitivity analyses (latency, extended exposure window, negative binomial).

5. Actual results

Results

Glargine alone vs other insulins (adjusted for age, sex): all malignancies RR 1.07 (95% CI 0.91-1.27) [multivariable 1.06 (0.90-1.25)]; breast cancer 1.99 (1.31-3.03) [unadjusted 1.91 (1.25-2.89); multivariable 1.97 (1.29-3.00)]; prostate 1.27 (0.89-1.82); gastrointestinal 0.93 (0.61-1.40). Glargine combined with other insulins: all malignancies 0.95 (0.85-1.07), breast 1.10 (0.77-1.56). Breast cancer with 1-year latency: 2.22 (1.24-3.99); any glargine use combined: 1.40 (1.04-1.89). No significant trend with number of defined daily doses (data not shown). Female glargine-alone users: mortality ratio 0.83 (0.71-0.96), AMI 0.77 (0.59-1.00).

Absolute risks

Incidence per 1,000 person-years, glargine alone vs other insulins: all malignancies 14.4 vs 14.8; breast cancer 5.0 (25 cases) vs 2.6 (183 cases); prostate 5.3 vs 4.7; gastrointestinal 2.1 vs 2.6.

Authors' conclusion

Women using glargine alone had an increased breast cancer incidence rate in 2006-2007; this may be random fluctuation; no significant results for other sites or all malignancies; no definitive conclusions on causality can be drawn; data from other settings needed.

6. Strengths

• Complete nationwide registers with high cancer-register completeness
• Fixed-cohort design avoiding reverse causation
• Adjustment for BMI, smoking and reproductive factors (where available); multiple sensitivity analyses
• Absolute incidence rates reported

7. Weaknesses

• Only 2 years of follow-up; glargine on Swedish market since 2003, implausibly short latency for carcinogenesis (authors)
• Breast-cancer signal based on 25 cases in one of several subgroups and sites; no correction for multiple comparisons
• No dose-response with defined daily doses; signal absent in glargine+other insulin users
• BMI and smoking missing for about half of participants (handled as missing category); no data on alcohol, genetics, screening
• Exposure classified from 6 months of prescriptions only (misclassification)
• Comparator group older with higher mortality - channelling possible

8. Risk of bias

Orienting tool

ROBINS-E

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

some concerns

Rationale

Single-reviewer ROBINS-E-oriented judgement from the full text; formal assessment requires two independent reviewers. Changed from 'high' (abstract) to 'some_concerns': the full text shows good register-based design and adjustment; the main problem is chance/multiple testing and very short follow-up, which the authors themselves emphasise.

Full-text review (25 October 2026): Group sizes, absolute incidence rates per 1,000 PY, combined-group and sensitivity results (latency RR 2.22, any glargine 1.40, no DDD trend), mortality/AMI added; no COI/funding statement found; bias overall high -> some_concerns (register design sound, main issue chance/multiplicity); methodological_quality low -> moderate; interpretation_adequacy set to partly_overinterpreted for archive use.

9. Assessment of the statistical interpretation

Overall cancer incidence was essentially equal (14.4 vs 14.8 per 1,000 person-years). The breast cancer RR of about 2 rests on 25 cases, is not replicated in glargine+other insulin users, shows no dose-response, and arises within 2 years - patterns more compatible with chance or confounding than causation, as the authors state.

10. Relevance to the specific claim

Limits C-INS-06: no overall or dose-dependent cancer increase; isolated breast-cancer subgroup signal.

11. Independent contrary findings

I-INS-05, I-INS-06.

12. Transferability to Germany and Europe

Swedish data; directly relevant.

13. Overall methodological judgement

Overall null for cancer with a single, unconfirmed breast-cancer subgroup signal that the authors themselves consider possibly random; does not support a dose-dependent cancer effect of glargine.

14. Open questions

None beyond those stated.

15. Linked fact checks

• Insulin glargine increases cancer risk in a dose-dependent way (and long-term use increases breast cancer). – Studies with mixed or unclear results

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 'insulin-injections' without its own interpretation; used as general support for the claim that insulin intensification harms.

A. What was actually studied?

Short-term cancer incidence in glargine users (requested by EASD)

B. What did the authors find?

All malignancies RR 1.07 (0.91-1.27); breast cancer RR 1.99 (1.31-3.03) only in women using glargine alone (25 cases), not in glargine+other insulin users (1.10); no dose trend; 2-year follow-up.

C. What conclusions do the authors themselves draw?

Women using glargine alone had higher breast cancer incidence; may be random fluctuation; no definitive conclusion on causality.

D. What does the archive derive from it?

Implicit support for glargine-cancer concern.

E. Is this derivation scientifically justified?

Only partly: overall malignancy was not increased, the breast finding appeared only in one subgroup without dose-response, and the authors explicitly call it possibly random and decline any causal conclusion.

Adequacy of the interpretation (by the archive)

partly correct but over-interpreted