GISSI-HF

In Review

Tavazzi L, et al. "Effect of n-3 polyunsaturated fatty acids in patients with chronic heart failure". The Lancet. 2008. 372(9645):1223-1230.
PubMed • Full text • PDF

Clinical Question

In patients with heart failure, does n-3 PUFA reduce all-cause mortality and hospital admission for CV disease, as compared to placebo?

Bottom Line

In patients with heart failure, n-3 PUFA reduces all-cause mortality and hospital admission for CV disease, as compared to placebo.

Major Points

The GISSI-HF trial studied the effect of n-3 polyunsaturated fatty acids (PUFA) in patients with symptomatic heart failure (NYHA Class II-IV), with the co-primary outcomes being all-cause mortality, and all-cause mortality or hospital admission for CV reasons.

Patients were randomized to receive n-3 PUFA 1 g daily (n=3,494) or placebo (n=3,481). Patients who received PUFA had a lower all-cause mortality (27% vs. 29% in placebo; adjusted HR 0·91; 95·5% CI 0·833-0·998; P=0·041; NNT=56). A benefit on all-cause mortality or hospital admission was seen as well (57% vs. 59% in placebo; adjusted HR 0·92; 99% CI 0·849-0·999; P=0·009; NNT=44). The most frequent adverse reaction, gastrointestinal symptoms, were observed equally in both group (3%).

In a substudy involving 608 heart failure patients, the authors reported that n-3 PUFA was associated with significant LVEF improvement (11.5% vs. 9.9% at 3-years in n-3 PUFA vs. placebo; P=0.005). In a separate paper, the investigators also reported that rosuvastatin did not reduce all-cause mortality or hospital admission for CV reasons in patients with heart failure. Similarly, the CORONA trial did not observe any benefit of rosuvastatin on the primary outcome of CV death, nonfatal MI, or nonfatal stroke. However, hospitalization for CV reason was reduced.

A meta-analysis published in 2012 reported the effect of n-3 PUFA on all-cause mortality (17 trials, n=51,264; RR 0.95, 95% CI 0.89-1.01) and CV mortality (14 trials, n=48,500; RR 0.89, 95% CI 0.83-0.96). The maximal benefits were seen at a mean daily intake of 0.20 grams.

Guidelines

ACCF/AHA Guideline for the Management of Heart Failure (2013, adapted)

• n-3 PUFA is reasonable as adjunctive therapy in patients HFrEF or HFpEF (NYHA class II-IV) if not contraindicated (Class IIa, Level of Evidence B)

Design

• Multicenter, double-blind, randomized, placebo-controlled trial
• N=6,975
  • n-3 PUFA 1 g daily (n=3,494)
  • Placebo (n=3,481)
• Setting: 357 centers in Italy
• Enrollment: 2002-2005
• Follow-up: median 3.9 (3-4.5) years
• Analysis: intention-to-treat
• Primary outcomes:
  1. all-cause mortality
  2. all-cause mortality or hospital admission for CV reasons

Population

Inclusion Criteria

• age ≥18 years
• heart failure
  • NYHA class II-IV symptoms
  • LVEF measured within 3 months prior to enrolment. For patients with LVEF >40%, they need to have ≥1 hospital admission for heart failure in the preceding year

Exclusion Criteria

• indication or contraindication to n-3 PUFA
• hypersensitivity to n-3 PUFA
• noncardiac illness that may limit long-term follow-up
• investigational treatment within 1 month prior to randomization
• ACS or revascularization within 1 month prior to randomization
• cardiac surgery scheduled within 3 months after randomization
• significant liver disease
• pregnancy or lactation

Baseline Characteristics

From the n-3 PUFA group

• Demographics: age 67±11 years, females 22.2%
• Heart failure:
  • NYHA Class II (63·7%), III (33·7%), IV (2.6%)
  • LVEF 33±8.5%
  • Admission for HF in the previous year 50%
  • Etiology: ischemic 49.1%, dilatative 30.1%, hypertensive 14.1%, other 3.1%, unknown 33.6%
• Medical history:
  • MI 41.8%, PCI 12.2%, CABG 17.6%
  • hypertension 54%, DM 28.4%, hypercholesterolemia 63.3%, peripheral vascular disease 8.4%, stroke 4.8%, current smoker 14.4%, AF 19.5%; ICD 7.1%, pacemaker 13.5%
  • COPD 21.2%, neoplasia 3.6%
• Clinical measurements: systolic BP 126±18 mm Hg, diastolic BP 77±10 mm Hg; heart rate 72±13 beats/min; BMI 27±5 kg/m²
• Medications: ACE inhibitor 77.2%, ARB 19.3%, antiplatelet agents 48.8%, statin 22.3%, beta-blocker 65.1%, spironolactone 38.6%, diuretics 89.5%, digoxin 37.1%, anticoagulant 29.4%, calcium-channel blocker 9.8%, nitrates 35.4%, amiodarone 19.1%

Interventions

• Patients randomized to receive n-3 PUFA PO 1 g (850-882 mg eicosapentaenoic acid and docosahexaenoic acid as ethyl esters in 1:1·2 average ratio) or placebo.
• Follow-up at 1, 3, 6, 12 months and every 6 months thereafter
• Compliance was defined as administration of drug for ≥80% observed days
• Patients also received optimal medical therapy for heart failure
• In the study, patients were also randomized to receive rosuvastatin 10 mg PO or placebo.

Outcomes

Comparisons are n-3 PUFA vs. placebo

Primary Outcome

All-cause mortality

27% vs. 29% (adjusted HR 0·91; 95·5% CI 0·833-0·998; P=0·041; NNT=56); (unadjusted HR 0·93; 95·5% CI 0·852-1·021; P=0·124)

heart failure accounted for the most number of deaths: 9.1% vs. 9.5%

All-cause mortality or hospital admission for CV reasons

57% vs. 59% (adjusted HR 0·92; 99% CI 0·849-0·999; P=0·009; NNT=44); (unadjusted HR 0·94; 99% CI 0·869-1·022; P=0·059)

Secondary Outcomes

CV mortality

20·4% vs. 22·0% (adjusted HR 0·9; 95% CI 0·81-0·99; P=0·045); (unadjusted HR 0·92; 95% CI 0·83-1·02; P=0·121)

Sudden cardiac death

8.8% vs. 9.3% (adjusted HR 0·93; 99% CI 0·79-1.08; P=0·333); (unadjusted HR 0·94; 95% CI 0·8-1·1; P=0·413)

Hospitalization

56.8% vs. 58.3% (adjusted HR 0·94; 95% CI 0·88-1·00; P=0·049); (unadjusted HR 0·96; 95% CI 0·90-1·02; P=0·178)

Hospitalization for CV reason

46.8% vs. 48.5% (adjusted HR 0·93; 95% CI 0·87-0·99; P=0·026); (unadjusted HR 0·95; 95% CI 0·89-1·02; P=0·122)

Hospitalization for heart failure

28% vs. 28.6% (adjusted HR 0·94; 95% CI 0·86-1·02; P=147); (unadjusted HR 0·97; 95% CI 0·89-1·06; P=0·511)

CV mortality or hospitalization due to any reason

61.7% vs. 63.3% (adjusted HR 0·94; 95% CI 0·89-0.99; P=0·043); (unadjusted HR 0·96; 95% CI 0·90-1·02; P=0·159)

Fatal and non-fatal MI

3.1% vs. 3.7% (adjusted HR 0.82; 95% CI 0·63-1.06; P=0·121); (unadjusted HR 0·82; 95% CI 0·64-1·06; P=0·135)

Fatal and non-fatal stroke

3.5% vs. 3% (adjusted HR 1.16; 95% CI 0·89-1.51; P=0·271); (unadjusted HR 1.18; 95% CI 0·91-1·53; P=0·225)

ischemic: 2.8% vs. 2.3%

hemorrhagic: 0.4% vs. 0.3%

Subgroup Analysis

No significant interaction for age, LV function, etiology of heart failure, NYHA class, the presence of diabetes, total cholesterol level.

Adverse Events

Resulting in withdrawing from treatment

2.9% vs. 3% (P=0.87)

• GI disorder: 96 vs. 92 patients
• Allergy: 3 vs. 9 patients
• Liver dysfunction: 1 vs. 1 patient
• Lipid abnormality: 0 vs. 1 patient
• Hepatocellular jaundice: 0 vs. 1 patient
• Subdural hematoma: 1 vs. 0 patient
• Muscle-related symptom: 1 vs. 0 patient
• Subdural hematoma: 1 vs 0 patient

Other outcomes (presented in other papers)

LVEF increase from baseline (30%)

comparisons are n-3 PUFA (n = 312) vs. placebo (n = 296):

8.1% vs. 6.3% at 1 year, 11.1% vs. 8.2% at 2 years, 11.5% vs. 9.9% at 3 years (P=0.005)

comparisons are rosuvastatin 10 mg daily (n = 212) vs. placebo (n = 207):

10.6% vs. 6.7% at 1 year, 12.9% vs. 11.2% at 2 years, 14.4% vs. 13.4% at 3 years (P=not significant)

All-cause mortality

comparisons are rosuvastatin 10 mg daily (n = 2,285) vs. placebo (n = 2,289):

29% vs. 28% (adjusted HR 1·00; 95·5%CI 0·898-1·122; P=0·943)

All-cause mortality or hospitalization for CV reasons

comparisons are rosuvastatin 10 mg daily (n = 2,285) vs. placebo (n = 2,289):

57% vs. 56% (adjusted HR 1·01; 95·5%CI 0·908-1·112; P=0·903)

Criticisms

• It has been questioned if dietary intake of n-3 PUFA may confound the study findings. Data on dietary habits of the trial patients were not published together with the main paper.
• The use of glitazones, which may increase the risk of heart failure, was not recorded in the study.

Funding

• Società Prodotti Antibiotici, Pfizer, Sigma Tau, and AstraZeneca