Newsletters & Publications

Newsletters & Publications

Nutrition Close-Up

Special Report – Fall 1997

Meta-Analyses of Plasma Lipoprotein
Responses to Changes in Dietary Fat and Cholesterol

“…there’s no connection whatsoever between cholesterol in food
 and cholesterol in the blood. None. And we’ve known that all  along.”

Dr. Ancel Keys
Eating Well
[March/April 1997]


In 1997 two independent research groups published detailed meta-analyses of the effects of dietary fat and dietary cholesterol, on plasma lipid and lipoprotein levels in humans. The analysis by Clarke et al. [1] used data from 82 metabolic ward studies and Howell et al. [2] included findings from 224 reports involving more than 8,000 study subjects from metabolic ward, controlled feeding and free-living population studies. These two reports provide the most up-to-date and extensive analyses of the effects of dietary lipids on plasma lipid and lipoprotein levels. The results of these studies provide predictive equations to estimate the population-wide plasma lipid responses to implementation of dietary recommendations to reduce intakes of total fat, saturated fat, and cholesterol.

What is a Meta-Analysis?

Meta-analysis is a statistical method of quantitatively combining and synthesizing results from individual studies to derive summary findings based on the currently available evidence. It is literally the taking of all the appropriate pieces of the puzzle (the individual studies) and trying to put the picture together. In the report by Howell et al. this process entailed an initial screening of 12,520 citations on dietary lipids and plasma lipids published between 1966 and 1994. Through a selection process based on defined criteria the authors selected 224 studies for inclusion in the meta-analysis. Clarke et al. limited their analysis to 80 metabolic ward studies.

How do the Two Analyses Compare?

The results of the two analyses are very comparable. The predicted changes in plasma total cholesterol (dTC), LDL cholesterol (dLDL) and HDL cholesterol (dHDL) with changes in percent of calories from total fat (dFAT), saturated fat (dSFA) and polyunsaturated fat (dPUFA) and mg/day cholesterol (dC) for Clarke et al. [1] and Howell et al. [2] are:

[1] dTotal Cholesterol = 2.01 dSFA – 1.01 dPUFA + 0.027 dC
[2] dTotal Cholesterol = 1.92 dSFA – 0.90 dPUFA + 0.022 dC
[1] dLDL Cholesterol = 1.39 dSFA – 0.85 dPUFA + 0.019 dC
[2] dLDL Cholesterol = 1.81 dSFA – 0.50 dPUFA
[1] dHDL Cholesterol = 0.50 dSFA + 0.39 dFat
[2] dHDL Cholesterol = 0.29 dSFA + 0.19 dFat

It is clear from both sets of predictive equations that changes in dietary fat quality and quantity have the major effects on plasma total and lipoprotein cholesterol concentrations. Exchanging 1% of SFA calories for PUFA calories would lower plasma total cholesterol and LDL cholesterol by 3.0 mg/dl and 2.2 mg/dl, respectively. Lowering dietary cholesterol by 50 mg/day would result in a 1 mg/dl decrease in total cholesterol and 1 mg/dl in LDL cholesterol. Reductions in the percent of calories from fat lower plasma HDL cholesterol levels. The results of the meta-analyses are consistent with epidemiological studies indicating that plasma HDL cholesterol levels decrease as fat calories are reduced and replaced with carbohydrate calories. This is often related to associated increases in plasma triacylglycerol levels with increasing intakes of simple carbohydrates.

What are the Changes in Plasma Cholesterol if the Public follows AHA’s Dietary Guidelines?

Based on the predictive equations generated from these meta-analyses, it can be estimated that a reduction in SFA intake from 13% to 10% of calories will lower plasma cholesterol by 5.7 mg/dl; an increase in PUFA intake from 7% to 10% of calories will lower cholesterol levels by 2.7 mg/dl; and a 100 mg/day decrease in dietary cholesterol will lower the plasma level by 2.2 mg/dl. Overall, modification of the present diet to one with 30% of calories as fat (10% SFA and 10% PUFA) and 300 mg/day cholesterol will lower plasma cholesterol levels by 10.6 mg/dl, a 5% reduction in the mean plasma cholesterol level of the population. These dietary changes would be predicted to lower plasma LDL cholesterol by 7 mg/dl and HDL cholesterol by 2 mg/dl. From epidemiological data it can be estimated that a 10 mg/dl reduction in plasma cholesterol would result in a 8% reduction in CVD risk [3].

What about Eggs and Plasma Cholesterol?

According to the results of these meta-analyses, a decrease in 100 mg/dl of dietary cholesterol would result in 2-3 mg/dl reduction of plasma cholesterol levels. What does this mean in terms of egg consumption? In 1945 per capita egg consumption was 405 eggs per year. In the last fifty years per capita intake decreased to 235 eggs per year, a reduction of 170 eggs per person per year. [Note: Some advocates of the dietary cholesterol – plasma cholesterol relationship point to a corresponding fall in the average plasma cholesterol from 235 mg/dl in the early 50s to an average of 205 mg/dl today and assume egg restrictions played a role.]. Using the findings from the two meta-analyses it is possible to calculate what the effect of limiting egg consumption has been on plasma cholesterol levels. A 170 eggs per year fall in consumption equals a 0.47 eggs per day decrease resulting in a 100 mg/day reduction in dietary cholesterol (0.47 eggs/day x 215 mg cholesterol/large egg). Using the average response factor for dietary cholesterol [0.025 mg/dl change in plasma cholesterol per mg/day change in dietary cholesterol] it can be estimated that this 42% reduction in per capita egg consumption would lower plasma cholesterol levels by 2.5 mg/dl. Given an average plasma cholesterol level of 235 mg/dl in 50s, the 42% decrease in egg consumption resulted in a 2.5 mg/dl (-1.1%) decrease in the average cholesterol level and can account for less than 10% of the 30 mg/dl reduction over a forty year period. Obviously other factors, including reductions in saturated fat in the diet, are responsible for the observed decrease in the cholesterol levels of the population.

What about Individual Patient Responses to Changes in Dietary Lipids?

Numerous studies have documented a large degree of individual heterogeneity of plasma lipid responses to changes in dietary fat and cholesterol. The results of these meta-analyses apply only to average plasma cholesterol changes in the population and cannot be applied to individual patients. For example, research has shown that in the majority of individuals there is little effect of dietary cholesterol on plasma cholesterol, whereas in 15-20% of the population the change in plasma cholesterol levels in response to changes in dietary cholesterol can be significant. The dietary cholesterol response factor estimated by meta-analysis provides only an average value thus individual responses can vary widely from this mean. The results of these meta-analyses are applicable to population based approaches for dietary interventions to reduce plasma cholesterol levels and CVD risk but have no applicability to the efficacy of interventions in individual patients.


1. Clarke R, Frost C, Collins R, Appleby P, Peto R. Dietary lipids and blood cholesterol: Quantitative meta-analysis of metabolic ward studies. BMJ 1997;314:112-117.

2. Howell WH, McNamara DJ, Tosca MA, Smith BT, Gaines JA. Plasma lipid and lipoprotein responses to dietary fat and cholesterol: A meta-analysis. Am J Clin Nutr 1997;65:1747-1764.

3. Verschuren WMM, Jacobs DR, Bloemberg BPM, et al. Serum total cholesterol and long-term coronary heart disease mortality in different cultures: Twenty-five year follow-up of the seven countries study. JAMA 1995;274:131-136.


Back to Top


Scroll to Top