• Included 9 RCTs
• Found no significant reduction in:
  • all-cause mortality, cardiovascular mortality, myocardial infarction, coronary artery events
• Stroke could not be meaningfully evaluated because too few stroke events were reported.
• Most included trials were secondary prevention studies, and very few participants were on statins.

• Included 17 RCTs total (fewer per outcome/intervention)
• Overall, reducing or modifying saturated fat intake trended toward lower all-cause mortality, cardiovascular mortality, nonfatal MI, and stroke
  • but for all four outcomes, the pooled confidence intervals included no effect.
• In 5 year period, estimated absolute effect varied by baseline cardiovascular risk. With ‘predefined thresholds’:
  • In low cardiovascular risk groups, the point estimate suggests an “unimportant effect”.
  • In high cardiovascular risk groups, the point estimate suggests an “important effect”.

• Study published in 1989, trial began in 1968

• double blinded, parallel group, randomized controlled dietary intervention trial

• 4393 Institutionalized men and 4664 Institutionalized women (mental institutions, nursing home)

• Participants could decline to enroll or discontinue at any time.

• corn oil / corn-oil margarine vs higher-saturated-fat institutional diet; tested cholesterol, cardiovascular events, cardiovascular death, total death.

• compared the effects of a

  • 39% fat control diet (18% saturated fat, 5% polyunsaturated fat, 16% monounsaturated fat, 446 mg dietary cholesterol per day)
  • 38% fat treatment diet (9% saturated fat, 15% polyunsaturated fat, 14% monounsaturated fat, 166 mg dietary cholesterol per day)

• mean duration of time on the diets was 384 days, with 1568 subjects consuming the diet for over 2 years

• For the entire study population, no differences between the treatment and control groups were observed for cardiovascular events, cardiovascular deaths, or total mortality.

• team recovered study materials stored on two nine-track magnetic tapes plus paper documents

• cross-referenced discrepancies in older reporting with the Broste thesis (a Master’s thesis written in 1981).

• re-evaluated results using recovered data

• “longitudinal data for the 2355 participants who received the intervention diet for a year or more”

  • Show fig 5, the Kaplan Meier life table graphs of cumulative mortality

• Participants were followed only while in hospital, and only about a quarter of randomized participants remained in the study for a year or longer.

• “…we are not able to determine if the effects of the high linoleic acid diet varied by smoking status, pre-existing coronary heart disease, psychiatric history, or drug use.

“confounding by dietary trans fat is an exceedingly unlikely explanation for the lack of benefit of the intervention diet.”

• Original Paper published in 1978, trial began in 1966
• secondary intervention (all already had heart attack)
• 458 men with coronary heart disease, aged 30-59
• 2 to 7 years.

• intervention: safflower oil / safflower margarine advice vs usual diet (restrict calories if overweight);

• outcome: tested all-cause, cardiovascular, and CHD mortality

  • specific, high-LA safflower intervention, not just a reduce SF trial

• diets assessed by interview and/or food log three times during the first year and twice yearly thereafter

• Group P: saturated fatty acids contributed 13.5% and polyunsaturated fatty acids 8.9% of total calories.

• Group F: consumed a derived 9.8% of calories from saturated fatty acids and 15.1% from polyunsaturates.

• data recovered from the original trial dataset stored on a nine-track magnetic tape
• obtained from original investigator
• median follow-up of 39 months

• intervention changed linoleic acid and saturated fat, but also monounsaturated fat and dietary cholesterol,
• trans fat
• Original authors say the study effectively became a multifactorial study.
  • large changes in smoking, diet, body weight, lifestyle, and physical activity before and after entry.
  • number of subjects and the duration may have been insufficient.
• prognosis was driven mainly by severity of underlying coronary/myocardial disease,
  • dietary variables not showing predictive power in their multivariate analysis.
• main results borderline significant

• Published in 1968/9, trial began in 1959
• 846 male veterans, aged 54-88 (mean 65.5)
• duration varied by entry (participants could enter anytime)
• The diet intervention:
  • intervention diet: total fat same, substituted vegetable unsaturated fat for two-thirds of the animal fat
  • intervention diet: about 38.4 g/day linoleic acid and 365.4 mg/day cholesterol.
  • control diet: conventional institutional diet, about 40% of calories from fat
  • control diet: 10.0 g/day linoleic acid and 652.7 mg/day cholesterol
• Primary outcome: ischemic heart disease manifested by sudden death or myocardial infarction.
• Secondary outcomes: cerebral infarction, amputation due to peripheral atherosclerosis, ruptured aortic aneurysm, and intestinal infarction.

• intervention diet lowered serum cholesterol
• primary endpoint was not statistically significant
• some combined endpoint analyses favored intervention diet (p=0.04), especially in younger half of the cohort
• Fatal atherosclerotic events were lower in the intervention group; total mortality was similar between groups.

Problems:

• Only partial diet control: participants ate only about half of their meals under study conditions.

• Adherence was limited and was tracked indirectly, mainly by dining hall attendance.

• More people dropped out of the intervention group than the control group.

• Study group was mixed: some participants already had signs or history of atherosclerotic disease.

• Diet change was not just “less saturated fat”:

  • cholesterol intake also changed
  • special oils and margarines were used
  • other parts of the diet may have changed at the same time

• Trans-fat exposure is a concern because margarines from that era were used, but it was not clearly measured.

• Study was relatively small for judging hard clinical outcomes.

• Some nonfatal events may have been missed when participants were away from the center.

• Main positive result came from combining several outcomes, not from a clear result on the main endpoint alone.

• late in the trial, excess non-atherosclerotic mortality in the intervention group

  • driven by small numbers of trauma and carcinoma deaths,
  • authors argued it probably did not reflect toxicity.

The carcinoma issue:

• fewer fatal atherosclerotic events on the intervention diet, but a borderline excess of carcinoma deaths.

• trial authors treated that cancer signal as concerning but inconclusive

• overall: 31 carcinoma deaths in intervention group; 17 in the control group (P=0.06). Mortality almost equal.

• Figure 14 in main study:

  • not cancer-specific curve, ‘death due to other cause’ (other than atherosclerotic complication)
  • late excess was small in absolute numbers
  • among deaths after 6 years, total excess non-atherosclerotic mortality in intervention group was 9 cases
    • 4 of those deaths were due to trauma.

• many cancer deaths in intervention group occurred among men who did not adhere closely to the diet
  • reduced likelihood that the polyunsaturated-oil diet itself explained the excess carcinoma mortality
• excess was only of borderline statistical significance
• post-hoc anlaysis