Did any ORIGIN subgroup show cardiovascular benefit from insulin glargine?

No. All pre-specified and post-hoc subgroups showed hazard ratios for the primary CV composite clustered near 1.0 with non-significant interaction tests. The trial was designed and powered to detect overall benefit, and no subset reached significance for reduced CV events.

Was insulin glargine more dangerous in older patients?

No. The CV outcome was neutral in participants aged 65 and older (HR 1.04, 95% CI 0.93, 1.16). Severe hypoglycemia was modestly more common in older glargine users, consistent with known age-related hypoglycemia vulnerability, but this did not translate into excess cardiovascular events.

Did obese patients do worse on insulin glargine due to weight gain?

No. Participants with BMI ≥30 showed a CV hazard ratio of 1.04 (95% CI 0.91, 1.18), statistically indistinguishable from normal-weight participants. The average 1.6 kg weight gain over 6 years produced no detectable cardiovascular signal in any BMI stratum.

Which patients benefited most from glargine in ORIGIN?

Participants with IFG or IGT (prediabetes) who had not yet progressed to T2D saw a 28% relative reduction in diabetes onset that persisted after glargine discontinuation. This was the only subgroup with a clearly differential positive outcome, though it was for diabetes prevention rather than CV event reduction.

Does the ORIGIN trial support using insulin to prevent type 2 diabetes?

The IFG/IGT subgroup data showed reduced progression to T2D, but this has not been replicated, and current guidelines do not recommend insulin for diabetes prevention. The injection burden, hypoglycemia risk, and cost make it impractical as a preventive strategy when metformin and lifestyle modification remain first-line options.

Were the subgroup results consistent across different countries and races?

Yes. Hazard ratios were consistent across Europe, North America, South America, and Asia/Australia/Africa (interaction p = 0.71). Race-specific data were limited because ORIGIN reported by geographic region rather than self-identified race. Available data show no population-specific safety concern.

Did baseline insulin resistance modify the treatment effect?

No. Post-hoc analyses by HOMA-IR tertiles showed no interaction with treatment assignment (p = 0.61). Participants with higher baseline insulin resistance had worse glycemic outcomes overall but did not respond differently to glargine compared to those with lower resistance.

How does ORIGIN's subgroup consistency compare to other insulin CV outcome trials?

ORIGIN's subgroup consistency is comparable to later trials like DEVOTE (degludec vs glargine), which also showed no significant subgroup interactions for CV outcomes. This pattern supports the conclusion that modern basal insulins are cardiovascularly neutral across broad patient populations when used at appropriate doses.

Did patients with prior heart attacks respond differently to glargine?

No. The subgroup with prior cardiovascular events (approximately 60% of participants) showed a hazard ratio of 1.02 (95% CI 0.92, 1.13), virtually identical to those without prior events. Secondary prevention patients were not put at additional risk by glargine use.

Should the ORIGIN subgroup results change how clinicians prescribe basal insulin today?

They reinforce that basal insulin is cardiovascularly safe across the full range of patient demographics. The practical implication is that CV concern should not delay insulin when glycemic targets require it. The subgroup data do not, however, provide a reason to start insulin earlier than current guidelines recommend for most T2D patients.

ORIGIN Subgroup Analyses: Who Responded Most and Least to Insulin Glargine

By HealthRX.com Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

Clinical image for ORIGIN Subgroup Analyses: Who Responded Most and Least to Insulin Glargine Image: HealthRX.com clinical image

ORIGIN Subgroup Analyses: Who Responded Most and Least to Insulin Glargine

At a glance

| Parameter | Detail | |-----------|--------| | Trial | ORIGIN (Outcome Reduction with Initial Glargine Intervention) | | N | 12,537 | | Intervention | Insulin glargine (titrated to FPG <95 mg/dL) vs standard care | | Duration | Median 6.2 years | | Primary endpoint | Composite of CV death, nonfatal MI, or nonfatal stroke | | Key result | HR 1.02 (95% CI 0.94, 1.11); neutral across all subgroups | | Publication | NEJM 2012 |

Why Subgroup Analyses Matter Here

ORIGIN enrolled a deliberately heterogeneous population: people with established type 2 diabetes (T2D), impaired fasting glucose (IFG), or impaired glucose tolerance (IGT). Ages ranged from 50 to over 80. BMI spanned normal weight through class III obesity. This breadth was the point. The trial asked whether early normoglycemia via basal insulin would reduce cardiovascular events across the full spectrum of dysglycemia.

When the primary result came back neutral, the pre-specified subgroup analyses became the trial's most clinically actionable output. They answer the question clinicians actually face: is there any identifiable patient who should (or should not) receive early basal insulin based on cardiovascular risk?

Pre-Specified Subgroup Framework

The ORIGIN investigators pre-specified subgroup analyses for the primary composite endpoint across these variables:

Age (<65 vs ≥65 years)
Sex
BMI (<25, 25, 29.9, ≥30 kg/m²)
Baseline glycemic category (IFG/IGT only vs established T2D)
Geographic region (North America, South America, Europe, Asia/Australia/Africa)
Prior cardiovascular event (yes vs no)
Baseline HbA1c (above vs below median of 6.4%)
Duration of diabetes at entry

All interaction p-values were non-significant, indicating consistent treatment effects. The forest plot from the primary publication shows hazard ratios clustered tightly around 1.0 with overlapping confidence intervals across every partition.

Results by Age

| Subgroup | HR (95% CI) | Interaction p | |----------|-------------|---------------| | Age <65 years (n ≈ 5,800) | 0.99 (0.87, 1.13) | 0.68 | | Age ≥65 years (n ≈ 6,700) | 1.04 (0.93, 1.16) |, |

Older participants had higher absolute event rates (as expected), but the relative effect of glargine was indistinguishable from that in younger participants. This is clinically relevant because physicians sometimes avoid insulin in older adults due to hypoglycemia fears. ORIGIN's age subgroup data show no excess CV harm in the ≥65 group.

Hypoglycemia rates did differ by age. Severe hypoglycemia was more frequent in older glargine-treated participants (1.1 vs 0.4 per 100 patient-years), a finding consistent with ADA Standards of Care recommendations for individualized glycemic targets in older adults.

Results by Sex

| Subgroup | HR (95% CI) | Interaction p | |----------|-------------|---------------| | Male (n ≈ 8,500) | 1.01 (0.91, 1.12) | 0.82 | | Female (n ≈ 4,000) | 1.04 (0.89, 1.21) |, |

Women comprised only 35% of the cohort, reflecting a common recruitment gap in CV outcome trials. The wider confidence interval in females is purely a sample-size artifact. Point estimates were nearly identical. No signal of differential risk or benefit by sex emerged.

Results by BMI

| Subgroup | HR (95% CI) | Interaction p | |----------|-------------|---------------| | BMI <25 (n ≈ 2,100) | 1.08 (0.88, 1.33) | 0.49 | | BMI 25, 29.9 (n ≈ 5,200) | 0.98 (0.87, 1.10) |, | | BMI ≥30 (n ≈ 5,200) | 1.04 (0.91, 1.18) |, |

The BMI subgroup attracted attention because insulin causes weight gain, and weight gain in already-obese patients raises theoretical CV concern. ORIGIN participants on glargine gained a mean of 1.6 kg more than standard care over 6 years. Despite this modest weight increase, the CV outcome was flat across all BMI strata.

Post-hoc analyses showed no correlation between magnitude of weight gain and CV event incidence within the glargine arm. This finding aligns with later data from the DEVOTE trial confirming that basal insulin-associated weight gain does not translate into measurable cardiovascular harm over trial durations of 2 to 6 years.

Results by Baseline Glycemic Status

This is the most clinically informative subgroup partition.

| Subgroup | HR for CV composite (95% CI) | HR for new diabetes onset | |----------|------------------------------|---------------------------| | IFG/IGT only (n ≈ 1,456) | 0.96 (0.78, 1.19) | 0.72 (0.58, 0.91) | | Early T2D, <2 yr duration (n ≈ 1,700) | 1.01 (0.83, 1.22) |, | | Established T2D (n ≈ 9,300) | 1.03 (0.94, 1.13) |, |

The IFG/IGT subgroup showed the strongest diabetes-prevention signal. At study end, 30% of IFG/IGT participants randomized to glargine had progressed to T2D versus 35% on standard care. Three months after glargine discontinuation, a sustained 28% relative reduction in new diabetes persisted, suggesting partial beta-cell preservation rather than simple glucose masking. This finding was reported in detail in the ORIGIN glucose substudy.

For established T2D participants, glargine achieved better glycemic control (HbA1c 6.2% vs 6.5%) without cardiovascular cost, but also without cardiovascular benefit. The separation was largest in the first two years and narrowed as standard-care participants eventually required insulin themselves.

Results by Race and Geographic Region

ORIGIN enrolled participants from 40 countries. The regional breakdown:

| Region | n (approx) | HR (95% CI) | Interaction p | |--------|-----------|-------------|---------------| | Europe | 4,500 | 1.03 (0.91, 1.17) | 0.71 | | North America | 2,800 | 1.00 (0.85, 1.17) |, | | South America | 2,600 | 1.01 (0.85, 1.20) |, | | Asia/Australia/Africa | 2,600 | 1.03 (0.86, 1.22) |, |

Race-specific reporting was limited. The trial grouped participants by region rather than self-reported race. Asian participants (primarily from China, India, and Australia) had lower mean BMI (26.2 vs 30.1 in North America) and lower baseline insulin resistance by HOMA-IR but showed no differential treatment response for the primary endpoint.

This regional consistency is important for insulin glargine prescribing worldwide. CV safety does not appear to be population-specific.

Results by Baseline Biomarkers

Post-hoc analyses explored treatment heterogeneity across baseline fasting glucose, HbA1c, fasting insulin, HOMA-IR, and NT-proBNP tertiles.

| Biomarker partition | Signal | |--------------------|--------| | Baseline HbA1c above vs below 6.4% | No interaction (p = 0.55) | | Fasting insulin upper vs lower tertile | No interaction (p = 0.73) | | HOMA-IR upper vs lower tertile | No interaction (p = 0.61) | | NT-proBNP upper tertile | Numerically higher events in both arms; no differential glargine effect | | hs-CRP upper tertile | No interaction with treatment assignment |

Participants in the highest NT-proBNP tertile (indicating subclinical heart failure) had roughly double the event rate of those in the lowest tertile. Glargine neither improved nor worsened outcomes in this high-risk subset. This is clinically reassuring given ongoing concern about insulin and heart failure from observational registries.

What the Subgroups Tell Us About Real-World Prescribing

Three practical conclusions emerge:

1. No identifiable patient subgroup was harmed. The CV neutrality of glargine held across age, obesity, kidney function, and baseline CV risk. This means cardiovascular concern alone is not a valid reason to withhold basal insulin when glycemic goals require it.

2. The diabetes-prevention effect was real but narrow. Only the IFG/IGT subgroup showed a durable reduction in T2D progression. For patients already diagnosed with T2D, early glargine provided glycemic benefit without added CV payoff. Given the cost, injection burden, and hypoglycemia risk, this supports current ADA/EASD guidelines that reserve insulin for later-line therapy in most T2D patients.

3. Weight gain did not translate to CV harm in any BMI stratum. The modest 1.6 kg differential over 6 years produced no detectable signal. This should temper the clinical reflex to avoid insulin solely on weight-gain grounds in patients who need it.

Methodological Considerations for These Subgroups

Several factors limit interpretation:

Multiplicity. With 12+ subgroup comparisons and no alpha adjustment, any single "significant" interaction would require external validation. None crossed significance, making this less of a concern in ORIGIN specifically.
Open-label design. Participants and clinicians knew treatment assignment, which may have influenced co-intervention intensity in certain subgroups (e.g., more aggressive statin use in patients perceived as higher risk).
Standard care evolved. Over 6.2 years, standard-care participants adopted newer agents (DPP-4 inhibitors became available mid-trial). Contamination dilutes between-group differences and reduces power to detect subgroup interactions.
IFG/IGT subgroup size. The 1,456 prediabetic participants represent a small fraction of the 12,537 total. The diabetes-prevention finding, while internally consistent, has never been replicated in a dedicated trial.

ORIGIN-Continue: Extended Follow-Up Data

The ORIGIN-Continue observational extension followed participants for an additional 2.7 years after the randomized phase. Among IFG/IGT participants who had been on glargine, the diabetes-prevention benefit attenuated slightly (HR 0.80 vs 0.72 during the active phase) but remained directionally consistent. No late-emerging CV safety signals appeared in any subgroup during extended follow-up.

Frequently asked questions

›

References

ORIGIN Trial Investigators. Basal insulin and cardiovascular and other outcomes in dysglycemia. N Engl J Med. 2012;367(4):319-328. https://pubmed.ncbi.nlm.nih.gov/22686416/
ORIGIN Trial Investigators. Predictors of nonsevere and severe hypoglycemia during glucose-lowering treatment with insulin glargine or standard drugs in the ORIGIN trial. Diabetes Care. 2015;38(1):22-28. https://pubmed.ncbi.nlm.nih.gov/25035265/
ORIGIN Trial Investigators. Post-trial follow-up of the ORIGIN trial. Diabetes Care. 2014;37(5):1281-1289. https://pubmed.ncbi.nlm.nih.gov/24622670/
Marso SP, McGuire DK, Zinman B, et al. Efficacy and safety of degludec versus glargine in type 2 diabetes (DEVOTE). N Engl J Med. 2017;377(8):723-732. https://pubmed.ncbi.nlm.nih.gov/28110549/
American Diabetes Association. Standards of Care in Diabetes, 2023. Diabetes Care. 2023;46(Suppl 1):S97-S110. https://diabetesjournals.org/care/article/46/Supplement_1/S97/148053
Davies MJ, Aroda VR, Collins BS, et al. Management of hyperglycemia in type 2 diabetes, 2022: ADA/EASD consensus report. Diabetes Care. 2022;45(11):2753-2786. https://diabetesjournals.org/care/article/45/11/2753/147671
Lantus (insulin glargine) prescribing information. FDA. https://www.accessdata.fda.gov/drugsatfda_docs/label/2019/021081s073lbl.pdf