ORIGIN n-3 Fatty Acids

Published

Bosch J, et al. "n-3 fatty acids and cardiovascular outcomes in patients with dysglycemia". The New England Journal of Medicine. 2012. 367(4):309-318.
PubMed • Full text • PDF

Clinical Question

Among those with prediabetes or diabetes does supplementation with n-3 polyunsaturated fatty acids reduce CV mortality when compared to placebo?

Bottom Line

Among those with prediabetes or diabetes, supplementation with n-3 polyunsaturated fatty acids (ie, fish oil) does not reduce CV mortality.

Major Points

Omega-3 (n-3) polyunsaturated fatty acids (PUFA) include linolenic acid, eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA). Linolenic acid comes from vegetable sources like soybeans, canola oils, and nuts, while EPA and DHA are fish oils. Interest in n-3 PUFAs for prevention of CVD began in the 1970s when observational studies found that the Inuit people of Greenland, who consumed a very high amount of fish, had low LDL-cholesterol, high HDL-cholesterol, high levels of EPA, and low incidence of CVD. These individuals also had a low serum level of the n-6 PUFA arachadonic acid (AA). It was theorized that a high ratio of n-3:n-6 PUFAs would be beneficial given the lower activity of prostacyclin and subsequent reduced risk of platelet aggregation. Early observational studies and comparative trials in the pre-aspirin era yielded conflicting results for the effects of fish-based n-3 PUFA in CVD. More recently, the 1999 GISSI-Prevenzione trial found a reduction in all-cause mortality, MI, and stroke when n-3 PUFAs were given to men with CVD. However, the GISSI study was limited by high dropout and its open-label design. As such, the role of n-3 PUFA in individuals at high risk for CVD remained unclear.

Published in 2012, the ORIGIN n-3 Fatty Acid trial randomized 12,536 individuals with insulin resistance or diabetes to a n-3 PUFA tablet (465 mg of EPA and 375 mg of DHA) or placebo (olive oil). With a median follow-up of 6.2 years, there was no difference in the primary outcome of CV mortality between the two groups. Secondary analyses of CV endpoints also did not demonstrate a benefit. There was a modest improvement in triglyceride levels. A 2012 meta-analysis incorporating the outcomes of this trial found no benefit of n-3 PUFA supplementation for mortality or CV endpoints.

This trial joins several other negative studies from the 1990s-2000s that to attempted to validate CVD nutritional theory (eg, benefits of associated with various dietary patterns) via administration of dietary supplements. For example, diets high in beta-carotene had previously been associated with lower rates of cancer and CVD, but supplementation with beta-carotene provided no benefit in the 1996 Physician's Health Study (and increased risk for lung cancer in smokers in the 1994 ATBC trial). The antioxidant vitamin E was hypothesized to reduce CV event rates thought to be associated with oxidative stress; yet a 2005 meta-analysis found increased mortality with the administration of high-dose vitamin E. Combination folic acid, B6, and B12 for secondary prevention in MI was associated with increased frequency of CVD endpoints in the 2006 NORVIT Trial. In contrast to supplement trials, trials of dietary modification have been modestly beneficial. This includes BP reduction and improvement in cholesterol profiles with the 1997 DASH diet (more vegetables, fruit, grains, and low-fat dairy, less sweetened beverages and red meat) and reduction in CV event rates with the Mediterranean diet studied in the 2013 PREDIMED study. Perhaps a reductive approach to nutrition is incorrect. Rather than eating individual supplements, the focus should be on eating whole foods that have been associated with positive health outcomes. As advocated by cardiologist Ellis Lader, maybe we should all eat more fish.

Guidelines

AHA/ACC lifestyle management to reduce CV risk (2013, adapted)

• For those who would benefit from lowering of LDL-cholesterol or BP lowering, advise that they eat a diet emphasizing vegetables, fruits, and whole grains, includes low-fat dairy, poultry, fish, legumes, nontropical vegetable oils, nuts, and limit sweets, sugar-sweetened beverages, and read meats (NHLBI grade A, ACC/AHA COR I, ACC/AHA LOE A)

Design

• Multicenter, double blind, 2x2 factorial, randomized controlled trial
• N=12,536 (15,392 screened)
  • n-3 fatty acids (n=6,281)
  • Placebo (n=6,255)
• Setting: 573 centers in 40 countries
• Enrollment: 2003-2005
• Median follow-up: 6.2 years
• Analysis: Intention-to-treat
• Primary outcome: CV mortality

Population

Inclusion Criteria

Described in the protocol.

• Age ≥50 years
• One of the following:
  • Impaired glucose tolerance - Postprandial glucose 140-200 mg/dL[7.7-11.1 mmol/L] and fasting glucose <126 mg/dL[<7 mmol/L]
  • Impaired fasting glucose - Fasting glucose 110-126 mg/dL[6.1-7 mmol/L) and postprandial glucose <200 mg/dL[<11.1 mmol/L]
  • Early T2DM - Fasting glucose ≥126 mg/dL[≥7 mmol/L], postprandial glucose ≥200 mg/dL[≥11.1 mmol/L], or diagnosis of DM and one of the following:
    • No diabetes medications in the prior 10 weeks and A1c <9%
    • One diabetic medication at stable dose in the prior 10 weeks and A1c <8% (if medication ≥50% maximum dose) or A1c <8.5% (if medication <50 maximum dose)
• At risk for CVD, defined by ≥1 of the following:
  • MI, not in the prior 5 days
  • CVA, not in the prior 5 days
  • Revascularization procedure
  • Angina with ischemic changes on testing or UA
  • Micro- or macroalbuminuria
  • LVH on EKG or echocardiogram
  • ≥50% stenosis of an artery (coronary, carotid, lower extremity) on angiography
  • ABI <0.9
• Informed consent
• Able and interested to comply with study procedures
• Non-pregnant women

Exclusion Criteria

• T1DM (or positive anti-GAD antibody) or other diabetes requiring insulin
• A1c ≥9%
• Nonadherence or unwilling to participate
• CABG in prior 4 years unless MI or angina since CABG <4 years prior
• Creatinine >2 mg/dL, ALT or AST >2.5x ULN

• Liver disease
• NYHA class ≥III
• Excluding CV causes, expected survival <3 years
• Cancer in prior 3 years except non-melanoma skin cancer
• Participation in this or another trial
• Known hypersensitivity to study products
• Heart transplant recipient or on waiting list

Baseline Characteristics

From the n-3 fatty acids group.

• Demographics: Age 64 years, female 35%
• PMH: MI, CVA, or revascularization 59%, HTN 79%, active smoking 12%, micro or macroalbuminuria 16%, cancer 4%
• Health data: BMI 30 kg/m², HR 70, BP 146/84, BMI 1.16
• Labs: Tchol 189 mg/dL[4.8 mmol/L], LDL 112 mg/dL[2.9 mmol/L], HDL 46 gm/dL[1.2 mmol/L], TG 142 mg/dL[1.6 mmol/L], creatinine 1 mg/dL[88.4umol/L] (eGFR 77 mL/min/1.73 m²), albumin:creatinine 5.2, fasting glucose 125 mg/dL[6.9 mmol/L], A1c 6.4%
• Medications: ACE-inhibitor or ARB 69%, thiazide 19%, antiplatelet 70%, anticoagulant 7%, beta blocker 53%, CCB 27%, statin 53%
• Average EPA-DHA intake 210 mg/day

Interventions

• Following a 10 day run-in, participants were randomized to a group:
  • n-3 fatty acids - 1g tablet with 465 mg of EPA and 375 mg of DHA
  • Placebo - 1 g olive oil
• Use of n-3 PUFA supplements were discouraged
• A dietary questionnaire was completed at randomization, year 2, and at the end of the stuyd

Participants were also randomized to receive insulin glargine or standard of care in this 2x2 factorial study. Those results are out of the scope of this review.

Outcomes

Presented as n-3 fatty acids vs. placebo.

Primary Outcome

CV mortality

9.1% vs. 9.3% (adjusted HR 0.98; 95% CI 0.87-1.10; P=0.72)

1.55 vs. 1.58 events/100 patient-years

Secondary Outcomes

MI, CVA, or CV mortality

Of note, this was not identified as a secondary outcome in the protocol.

16.5% vs. 16.3% (adjusted HR 1.01; 95% CI 0.93-1.10; P=0.81)

All-cause mortality

15.1% vs. 15.4% (adjusted HR 0.98; 95% CI 0.89-1.07; P=0.63)

Arrhythmia mortality

Sudden unexpected death, nonsudden death, unwitnessed death, or resuscitation after cardiac arrest

4.6% vs. 4.1% (adjusted HR 1.10; 95% CI 0.93-1.30; P=0.26)

Additional Outcomes

MI

5.5% vs. 5.1% (adjusted HR 1.09; 95% CI 0.93-1.27; P=0.28)

CVA

5.0% vs. 5.4% (adjusted HR 0.92; 95% CI 0.79-1.08; P=0.32)

HF hospitalization

5.3% vs. 5.1% (adjusted HR 1.02; 95% CI 0.88-1.19; P=0.76)

Revascularization

13.8% vs. 14.3% (adjusted HR 0.96; 95% CI 0.87-1.05; P=0.39)

Angina

New, worsening, or unstable.

11.5% vs. 11.6% (adjusted HR 1.00; 95% CI 0.90-1.10; P=0.94)

Amputation of digit or limb for ischemia

0.8% vs. 0.8% (adjusted HR 1.09; 95% CI 0.74-1.62; P=0.67)

CV hospitalization

32.7% vs. 33.4% (adjusted HR 0.98; 95% CI 0.92-1.04; P=0.50)

Change from baseline

SBP: -4.37 vs. -4.51 mmHg (P=0.75)

DBP: -4.93 vs. -4.96 mmHg (P=0.91)

HR: -0.13 vs. +0.25 BPM (P=0.13)

Tchol: -15.7 vs. -14.6 mg/dL (P=0.17)

LDL: -11.8 vs. -12.4 mg/dL (P=0.44)

HDL: -0.1 vs. -0.2 mg/dL (P=0.78)

TG: -23.5 vs. -9.0 mg/dL (P<0.001)

Cognitive function

Not reported (identified as outcome in the protocol.)

Hgb A1c, reversion to normal glucose tolerance

Not reported (identified as outcome in the protocol.)

Criticisms

• The placebo group was given olive oil, which may have conferred some CV protection

Funding

• Sanofi
• Study medication from Pronova BioPharma Norge