ADVANCE: intensive control to HbA1c 6.5%

Overview

ADVANCE Collaborative Group; Patel A et al.. N Engl J Med 2008 · PMID 18539916 · DOI 10.1056/NEJMoa0802987

Fewer kidney complications and no extra deaths, but more severe hypoglycaemia.

Does the study support the specific claim? No · Methodological quality: high

Original publication

Plain-language summary

Scientific in-depth review

1. Full reference

Authors

ADVANCE Collaborative Group; Patel A et al.

Journal

N Engl J Med

Year

2008

Identifier

PMID 18539916 DOI 10.1056/NEJMoa0802987

Verification

verified – PubMed efetch 2026-10-25 (this agent)

Retracted

no

Corrections

none

Extent checked

full text

Conflicts of interest and funding

Supported by Servier (major sponsor; manufacturer of gliclazide MR and perindopril-indapamide) and NHMRC Australia; investigator-initiated, designed, analysed and interpreted independently of sponsors, data not shared with sponsors; many writing-committee members report fees/grants from Servier and other companies.

2. Research question

Effect of intensive glucose control (gliclazide MR-based, HbA1c <=6.5%) on vascular outcomes

3. Study design and population

Study design

Randomised controlled trial – Investigator-initiated, factorial, open-label (glucose arm) randomised controlled trial (ADVANCE, NCT00145925); central computer randomisation after 6-week active run-in; intention-to-treat; blinded end-point adjudication.

Population

Type 2 diabetes diagnosed at >=30 years, age >=55 (mean 66), with history of major macro- or microvascular disease or >=1 other vascular risk factor; mean diabetes duration 7.9 years, mean HbA1c 7.5%; 91% on oral agents, ~1.5% on insulin; definite indication for long-term insulin excluded.

Country

215 centres in 20 countries (Asia 37%, Europe 46%, Australia/New Zealand 13%, North America 4%)

Period

Recruitment June 2001 - March 2003; glucose comparison completed January 2008; median follow-up 5.0 years.

Sample size

12,877 registered; 1,737 withdrawn in run-in; 11,140 randomised (intensive 5,571, standard 5,569); 17 lost to follow-up.

Inclusion and exclusion criteria

As above; no HbA1c-based inclusion/exclusion criteria.

4. Methods and endpoints

Exposure / intervention

Intensive strategy: gliclazide MR 30-120 mg (90% still on it at final visit, 70.4% at 120 mg) plus metformin, thiazolidinediones, acarbose or insulin as needed to HbA1c <=6.5%; insulin used by 40.5% at study end (vs 24.1% standard).

Comparator

Standard glucose control per local guidelines (other sulfonylurea allowed; 57.1% on other sulfonylurea at end).

Endpoints

Primary: composite of major macrovascular events (CV death, non-fatal MI, non-fatal stroke) and major microvascular events (new/worsening nephropathy or retinopathy), jointly and separately; prespecified secondary outcomes incl. all-cause death, hypoglycaemia, hospitalisation.

Statistical methods

Unadjusted Cox models, ITT; NNT from absolute risk differences; no multiplicity adjustment; protocol amended mid-trial (blinded to results) to add joint primary analysis and extend follow-up by 18 months.

5. Actual results

Results

HbA1c at end 6.5% vs 7.3% (time-weighted difference 0.67 percentage points, 95% CI 0.64-0.70). Combined major macro-/microvascular events 18.1% vs 20.0%, HR 0.90 (0.82-0.98; P=0.01), NNT 52 over 5 years (30-213). Microvascular 9.4% vs 10.9%, HR 0.86 (0.77-0.97); nephropathy 4.1% vs 5.2%, HR 0.79 (0.66-0.93), mainly macroalbuminuria (HR 0.70, 0.57-0.85); doubling of creatinine HR 1.15 (0.82-1.63); renal replacement/renal death HR 0.64 (0.38-1.08); retinopathy not reduced (P=0.50). Macrovascular HR 0.94 (0.84-1.06); CV death HR 0.88 (0.74-1.04); all-cause death 8.9% vs 9.6%, HR 0.93 (0.83-1.06). Severe hypoglycaemia 2.7% vs 1.5%, HR 1.86 (1.42-2.40); 0.7 vs 0.4 events per 100 patient-years. Hospitalisation 44.9% vs 42.8%, HR 1.07 (1.01-1.13). Weight +0.7 kg; SBP 1.6 mm Hg lower.

Absolute risks

Combined events 18.1% vs 20.0% (NNT 52 over 5 years); death 8.9% vs 9.6%; severe hypoglycaemia 2.7% vs 1.5% (about 3 additional severe events per 1,000 patients per year).

Authors' conclusion

An intensive gliclazide MR-based strategy lowering HbA1c to 6.5% yielded a 10% relative reduction in combined major macro- and microvascular events, mainly via a 21% reduction in nephropathy; no significant effect on macrovascular events or death; no evidence of increased mortality; more severe hypoglycaemia.

6. Strengths

• Large, international RCT with 5-year follow-up and near-complete vital status
• Investigator-led with data independent of sponsors
• Blinded adjudication of primary end points and deaths
• Broad population, mostly European/Asian

7. Weaknesses

• Open-label glucose comparison; more visits in intensive arm (31 vs 11) - co-intervention (e.g. lower BP)
• Mid-trial protocol amendment adding joint primary outcome (stated as blinded to results)
• Microvascular benefit driven by macroalbuminuria (surrogate-like), not by hard renal end points
• Underpowered for macrovascular events (lower event rate than expected)
• Not an insulin-specific comparison (insulin used in 40.5% vs 24.1%)
• Sponsor manufactures the study drug; author COI

8. Risk of bias

Orienting tool

RoB 2

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 RoB 2-oriented judgement from the full text; formal assessment requires two independent reviewers and the protocol. Changed from 'low' (abstract) to 'some_concerns' because of the open-label design, differential visit frequency and the mid-trial change to a joint primary outcome; mortality and macrovascular results remain at low risk of bias.

Full-text review (25 October 2026): Population, run-in, drug use at end (insulin 40.5% vs 24.1%; gliclazide 90%), NNT 52, renal component breakdown (macroalbuminuria vs creatinine doubling), hospitalisation, hypoglycaemia rates added; funding/independence (Servier, NHMRC) and COI documented; RoB 2 overall low -> some_concerns (open-label, mid-trial joint-outcome amendment).

9. Assessment of the statistical interpretation

Modest relative benefit (HR 0.90) on a composite driven by albuminuria; no effect on macrovascular events or mortality (CIs compatible with modest benefit or small harm); clear but low absolute hypoglycaemia excess.

10. Relevance to the specific claim

Limits C-INS-04 (intensive control towards normal HbA1c was not harmful in ADVANCE, unlike ACCORD); indirectly limits C-ODD-11 (a gliclazide-based strategy did not increase CV death or mortality), though it does not compare sulfonylurea with metformin.

11. Independent contrary findings

ACCORD (A-INS-01).

12. Transferability to Germany and Europe

Multinational trial including European sites or directly applicable to EU practice.

13. Overall methodological judgement

High-quality RCT showing that intensive glucose lowering to HbA1c 6.5% with a gliclazide-based strategy (40% eventually on insulin) did not increase mortality, reduced nephropathy (mainly albuminuria) and increased severe hypoglycaemia. Shows that intensive control is not uniformly harmful (contrast with ACCORD).

14. Open questions

None beyond those stated.

15. Linked fact checks

• Large trials in type 2 diabetes showed that intensive glucose lowering towards normal HbA1c can be harmful: ACCORD was stopped after intensive therapy raised all-cause mortality by 22%, and severe hypoglycaemia was associated with roughly doubled mortality. – Studies contradicting or limiting the claim
• Sulfonylureas cause more myocardial infarctions, strokes and deaths than metformin (21% higher in 253,690 veterans). – Studies contradicting or limiting the claim