Revisiting the diet-heart hypothesis: critical appraisal of the Minnesota Coronary Experiment
Several recently published articles have concluded that associations between high cholesterol and premature death are lacking and advocate that current guidance advising low (<10%) saturated fat consumption should be revised(1-3). One such study by Ramsden et al., recovered 45 year-old data from the Minnesota Coronary Experiment (MCE), a randomized controlled trial performed in state mental hospitals and nursing homes in Minnesota, USA, and focused on reassessing the effect of a linoleic acid-rich diet on serum cholesterol and mortality(3). The authors reported that such a diet lowered total cholesterol, but did not reduce the risk of premature mortality, and might even have increased it.
However, to propagate a change in the current consensus, these claims should be based upon the valid and reliable evidence. Therefore we performed a critical appraisal of MCE consulting the methods and data provided in three publicly available descriptions of this trial: Ramsden et al., and Frantz et al., both peer-reviewed papers, and a thesis by Broste, on which the primary data in the analysis by Ramsden et al., were based(3-5).
We used the Cochrane risk of bias tool to assess the internal validity of data from the original MCE (Table 1)(6).
Several factors that weaken the validity of conclusions drawn from the MCE study were identified. Here we focus on the areas of most concern - discrepancies in data and selective outcome reporting, high attrition and poor generalisability.
Discrepancies in data and selective outcome reporting
The total number of MCE participants and event rates differed in each report of the same data: 9570 participants and 517 deaths in Ramsden et al., 9057 and 517 in Frantz et al. and 9570 and 477 in Broste’s thesis. The difference in event rates between intervention and control groups is driven by events occurring in those aged >65 years(4). We again observed a discrepancy in the numbers analysed for this age group: 626 participants were left after year 1 in Broste’s thesis, whilst only 595 are reported by Ramsden et al. These inconsistencies in participant and event rates cast doubt over the overall precision of data collection and their recovery.
Ramsden et al. focused on one statistically significant mortality association – with serum cholesterol concentrations. However, smoking, a higher BMI, and a higher diastolic blood pressure were each associated with a lower mortality risk in Broste’s thesis and also substantially contradict our current knowledge(4). As outcomes and statistical analysis methods in original MCE were not clearly pre-specified a priori, any subsequent statistical sub-analyses of MCE data should have been adjusted for multiple analysis inflation. This was not performed nor acknowledged in Ramsden et al., and the resultant observed associations could have arisen by chance.
We had to turn to data in Broste’s original thesis to analyse the extent of attrition in each group(4). Firstly, more than 83% of participants are lost to follow-up from both study arms (4028 control, 3953 diet). All people lost to follow-up were excluded from further analysis and per protocol analysis was performed. Such analysis is likely to inflate estimated effect sizes(7).
Secondly, 21 more deaths were observed in the diet group (269 vs 248) and this difference is exclusively driven by a higher mortality rate in the >65 year old population (190/953 = 19.9%, 162/958 = 16.9% in the diet and control groups, respectively)(3, 5). We manually counted events from Kaplan-Meier curves for >65 year old participants remaining in study for >1 year (when mortality difference between diet and control groups becomes established). 27 patients in the control arm and 40 patients in the diet group died during year 1-2 period. However, during the same period 15 more control group participants are lost to follow-up without explanation (81 and 96). It is likely that some of these 15 participants could have actually died, as year 1-2 per protocol mortality rate in control group was 12% (27/225). Treating excessive control group’s attrition cases as deaths significantly alters the mortality rate estimates for >65 group and halves the relative risk reduction to 0.075.
Unfortunately none of the MCE reports provide data to ascertain if prognostic equivalence between both >65 groups has been maintained throughout follow-up, which significantly impairs the validity of any demonstrated effect(s)(8). Incidentally, equivalent compensation for evident higher attrition in the control group in the overall MCE population would actually demonstrate a higher mortality rate within the control group.
Generalisability of findings
Apart from highly selective study population (mental hospital inpatients), the dietary interventions present a major concern for the external validity from the MCE data. Current World Health Organisation dietary recommendations are that no more than 30% of daily calories come from fat(9). In MCE, the intervention and control diets, respectively, contained 38% and 39% of calories from fat. AMA recommends a daily intake of linoleic acid of between 5% and 10% of the energy content of the diet, with a note of caution for higher intakes (>11%) due to an observed association with increased lipid peroxidation(10, 11). The proportion of energy from linoleic acid in the MCE intervention arm was 13%. This was achieved by adding large quantities of liquid corn oil - practically the sole source of linoleic acid in MCE. Such a diet may be harmful for reasons unrelated to potential cholesterol mechanism and does not reflect current dietary recommendations either in terms of proportion of energy from linoleic acid or its consumption in the general population. Even if correct and correctly interpreted, calls for a change to dietary guidelines based on MCE data are remiss as recommendations do not reflect the diets offered to control group participants in the MCE study.
Well supported developments in our understanding of the interaction between dietary fat and human health are needed. However, data from the MCE suffer from multiple methodological shortcomings such as discrepant reporting, heavy attrition and poor generalisability. A detailed re-analysis empirically demonstrates that any claims to change our current understanding of the relationship between saturated fat intake and mortality are not fully supported by data from MCE.
Tumas Beinortas1*, Kamal R. Mahtani2, David Nunan2
1 St. James Hospital, Leeds Teaching Hospitals NHS Trust
2 Centre for Evidence Based Medicine, Department of Primary Care Health Sciences, University of Oxford
St. James Hospital, Leeds Teaching Hospitals NHS Trust
1. de Souza RJ, Mente A, Maroleanu A, Cozma AI, Ha V, Kishibe T, et al. Intake of saturated and trans unsaturated fatty acids and risk of all cause mortality, cardiovascular disease, and type 2 diabetes: systematic review and meta-analysis of observational studies. BMJ (Clinical research ed). 2015;351:h3978.
2. Ramsden CE, Zamora D, Leelarthaepin B, Majchrzak-Hong SF, Faurot KR, Suchindran CM, et al. Use of dietary linoleic acid for secondary prevention of coronary heart disease and death: evaluation of recovered data from the Sydney Diet Heart Study and updated meta-analysis. BMJ (Clinical research ed). 2013;346:e8707.
3. Ramsden CE, Zamora D, Majchrzak-Hong S, Faurot KR, Broste SK, Frantz RP, et al. Re-evaluation of the traditional diet-heart hypothesis: analysis of recovered data from Minnesota Coronary Experiment (1968-73). BMJ (Clinical research ed). 2016;353:i1246.
4. Broste SK. Lifetable analysis of the Minnesota Coronary Survey: The University of Minnesota; 1981.
5. Frantz ID, Jr., Dawson EA, Ashman PL, Gatewood LC, Bartsch GE, Kuba K, et al. Test of effect of lipid lowering by diet on cardiovascular risk. The Minnesota Coronary Survey. Arteriosclerosis (Dallas, Tex). 1989;9(1):129-35.
6. Higgins JP, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ (Clinical research ed). 2011;343:d5928.
7. Abraha I, Cherubini A, Cozzolino F, De Florio R, Luchetta ML, Rimland JM, et al. Deviation from intention to treat analysis in randomised trials and treatment effect estimates: meta-epidemiological study. BMJ (Clinical research ed). 2015;350:h2445.
8. Dumville JC, Torgerson DJ, Hewitt CE. Reporting attrition in randomised controlled trials. BMJ (Clinical research ed). 2006;332(7547):969-71.
9. World Health Organization. Healthy diet Factsheet. 2015.
10. Harris WS, Mozaffarian D, Rimm E, Kris-Etherton P, Rudel LL, Appel LJ, et al. Omega-6 fatty acids and risk for cardiovascular disease: a science advisory from the American Heart Association Nutrition Subcommittee of the Council on Nutrition, Physical Activity, and Metabolism; Council on Cardiovascular Nursing; and Council on Epidemiology and Prevention. Circulation. 2009;119(6):902-7.
11. Food and Agriculture Organization of the United Nations. Fats and fatty acids in human nutrition: report of an expert consultation. FAO Food and Nutrition Paper 91. Rome: 2010.
Competing interests: No competing interests