The Quest for a Cholesterol-decreasing Diet: Should We Subtract, Substitute, or Supplement?

Recent reports from the National Cholesterol Education Program have documented the continued decrease of serum cholesterol levels in U.S. adults. Using data from the National Health and Nutrition Examination Surveys, the mean cholesterol level for adults, ages 20 to 74 years, was shown to have decreased from 5.69 mmol/L (220 mg/dL) in 1960 to 1962 to 5.30 mmol/L (205 mg/dL) in 1988 to 1991 [1]. This decrease was observed in essentially all race and sex subgroups and was most marked in the past 10 years. Although cholesterol-decreasing drugs have increased in use, it is likely that most of this decrease in cholesterol levels was because of modifications of lifestyle, especially diet. Despite the optimistic data on decreases in population-wide cholesterol levels, it is still estimated that only 49% of U.S. adults have desirable cholesterol levels [2]. Of the remaining, at least 29% (52 million U.S. adults) would require dietary therapy.

Thus, the proportion of patients in the typical internist's office who require dietary counseling for increased serum cholesterol levels is likely to be 30% to 50%. In addition, the public is bombarded with new findings related to nutrition and health, in general, and cholesterol, in particular, creating considerable confusion about what should comprise a cholesterol-decreasing diet. Patients promptly query their physicians about these newest findings, hoping to find the next dietary panacea. Of greater concern, however, is that these new findings may be used indiscriminately by patients instead of in a medically directed program that is integrated into the patient's total health care program.

This issue of Annals contains two research papers and a meta-analysis [3-5] about dietary factors that affect serum cholesterol levels. These papers are representative of the increasing presence of nutritional research appearing in the medical literature. They also illustrate two of the several options for patients needing or desiring a diet to decrease serum cholesterol levels: namely, to subtract those dietary constituents that increase cholesterol levels, to substitute cholesterol-increasing foods for cholesterol-decreasing or cholesterol-neutral foods, and to supplement the diet with substances that actually decrease cholesterol levels. The goal of our editorial is to put these studies into perspective in the context of current nutritional trends and guidelines, rather than to comment on specific methodologic issues raised by these reports.

Clearly, the major theme of the National Cholesterol Education Program is subtraction. The goal of the Population Panel's recommendation is to decrease dietary intake of total fat from the current 36% to 38% of total calories to 30% or less, to decrease saturated fat from 14% or 16% to 10% or less, and to decrease dietary cholesterol from the current 400 mg daily or more to 300 mg daily or less [6]. This diet is recommended for all U.S. adults and is identical to the step I diet for hypercholesterolemic patients. The step II diet is the therapeutic diet recommended by the second report of the Adult Treatment Panel [7]; this diet recommends further decreases in saturated fat to 7% of calories and decreases in dietary cholesterol to less than 200 mg daily. This step II diet requires much more discipline and skill in food purchasing and preparation, is infrequently prescribed in the physician's office, and often requires the services of a registered dietitian. If instituted, the deletions in fat using the step I diet can lead to a 10% or greater decrease in cholesterol levels; additional decreases in cholesterol levels are possible after using the step II diet [7]. These national recommendations serve as the bench marks for comparison of the other dietary strategies, including substitution [3] and supplementation [4, 5].

The report by Davis and colleagues [3] describes the effects of substituting a modified-fat cheese (45% polyunsaturated fat) for that containing predominantly butter fat (2.4% polyunsaturated fat). Statistically significant decreases in total cholesterol and low-density lipoprotein cholesterol levels were obtained, which were expected after replacing saturated with polyunsaturated fat. Indeed, much of the decrease in serum cholesterol levels during the past 20 years is probably due to such a substitution, because total fat consumption in the United States has not decreased during this period but rather has shown a replacement of foods high in saturated fats with those high in polyunsaturated fats [8]. If a step II diet is needed by 52 million U.S. adults, such products that replace traditional foods high in saturated fat will be needed. The importance of this report [3] is underscored by the observation that cheese is the single largest source of saturated fat in the diets of U.S. women [9].

The problem with the substitution of one fat for another is the concern about the safety of that fat replacing the saturated fat. The fat-modified cheese in this study [3] increased the polyunsaturated fat intake by 13%. Diets high in polyunsaturated fat have been linked to increased rates of cancera concern that led to the recommendation for decreasing total fat rather than just saturated fat. Polyunsaturated fat may also decrease high-density lipoprotein cholesterol levels [10]; the importance of this reduction is unclear, but it certainly does not fuel the enthusiasm for markedly increasing polyunsaturated fat consumption.

Two additional papers [4, 5] in this issue address nutritional supplements. Psyllium has had wide-spread use as a stool softener, and its cholesterol-decreasing capabilities have been explored previously. Sprecher and colleagues [4] describe psyllium's ability to decrease total and low-density lipoprotein cholesterol in the presence of either a high- or low-fat diet. Psyllium, at 10.2 g daily, had a small (6.4% to 7.2%) effect on low-density lipoprotein cholesterol in persons eating either high- or low-fat diets, respectively. A substantial number of persons taking psyllium had a decrease in serum cholesterol levels, which moved them to a lower risk strata for cardiac events. This supplement, then, should provide about 50% of the decrease in cholesterol levels observed when changing from the current U.S. diet to a step I diet.

A larger effect was observed for supplementation with garlica practice popular in Europe. The meta-analysis by Warshafsky and colleagues [5], of five carefully selected studies according to preset criteria, estimated the effect size of supplementing the daily diet with approximately one clove or less of garlic per day. The effect, a decrease in serum cholesterol levels by 0.59 mmol/L (23mg/dL [9%]) on the average, will be similar to that of a number of cholesterol-decreasing agents currently prescribed for the correction of hypercholesterolemia, such as bile acidbinding resins and fibric acid derivatives.

These two supplements (psyllium and garlic) potentially have great appeal, because they are available without prescription and they are natural products. However, their role in cholesterol-decreasing therapy needs to be clarified. Their exact modes of action have not been elucidated, making prediction of their interaction with other dietary constituents difficult. In this regard, it is useful that psyllium's effect is observed across a wide range of dietary fat levels [4]. Similarly, their interaction with other cholesterol-decreasing drugs needs to be described, because the patients who might use these supplements may also be those who are prescribed prescription drugs for the same purpose. For example, psyllium products are often recommended along with bile acid-binding resins. Finally, their natural, nonprescription qualities do not always translate into lower costs. If a decrease in cholesterol levels is the goal, the cost of these supplements should be compared with other cholesterol-decreasing agents such as generic niacin, which costs approximately $6 to $7 per month and has a proven track record of ability to lower cholesterol levels and to reduce cardiac events.

One final issue deals with the use of these dietary agents by the patient. Low-fat cheeses, psyllium-containing products, and even garlic tablets are and will be available and marketed for various beneficial effects. For the person with a high-risk cholesterol level, they can probably serve useful purposes as part of an overall dietary modification program. One concern is that their use will delay or replace the initiation of such a program, with the patient depending on these agents instead of on comprehensive medical care. The main message to the patient should be subtraction of dietary fat, as advised by physicians and nutritionists, rather than substitution or supplementation. These additional foods can then serve as adjuncts to an overall dietary program, rather than as the primary intervention. However, this dietary program will likely also contain soluble fiber and modified-fat foodsand perhaps, will be strongly flavored with garlic.

• Copyright 2004 by the American College of Physicians