Plasma noncholesterol sterols as indicators of cholesterol absorption

A publication by Jakulj et al. (1) in this issue of the Journal of Lipid Research calls into question the validity of plasma noncholesterol sterols for the evaluation of cholesterol absorption in humans. This method was first proposed by Miettinen and coworkers (2^–5), who showed that either sitosterol/total cholesterol or campesterol/lathosterol ratios reflect amounts of cholesterol absorbed by the intestine. Normally only small amounts of plant sterols (sitosterol and campesterol) are absorbed, but Miettinen and associates claimed that quantities absorbed are proportional to amounts of cholesterol absorbed. In turn, the greater the total of plant sterol absorbed, the higher will be the plasma level, normalized for plasma cholesterol levels. These investigators (4, 6, 7) further showed that plasma levels of a cholesterol synthesis precursor, lathosterol, reflect the body's cholesterol synthesis rate. Because absorbed cholesterol returns to the liver and suppresses cholesterol synthesis, a low level of plasma lathosterol will denote reduced cholesterol synthesis because of more cholesterol absorption. Consequently, the campesterol/lathosterol ratio should better indicate cholesterol absorption than the sitosterol/cholesterol ratio.

Jakulj et al. (1) attempted to validate plant sterols in plasma as a marker of cholesterol absorption. To do so, they compared plasma campesterol/cholesterol ratios with cholesterol absorption in mildly hypercholesterolemic subjects. Fractional absorption of dietary cholesterol was determined by comparing isotope ratios of plasma cholesterol following oral and intravenous administration. They found no difference in fractional cholesterol absorption in those with the highest and lowest campesterol/cholesterol ratios. Nor was there a correlation between these ratios and fractional absorption in the total population. The authors concluded that previous studies claiming a relationship between plasma sterol ratios and cholesterol absorption are of questionable validity.

It might be noted that Jakulj et al. (1) did not measure a marker of cholesterol synthesis, e.g., lathosterol. Because of the inverse relation between cholesterol absorption and synthesis, the addition of a synthesis precursor could have helped to validate the use of noncholesterol sterols as a marker for cholesterol absorption as well as synthesis.

The overall design of the study of Jakulj et al. (1) nonetheless is of interest. Testing whether subgroups with the highest and lowest plant sterol levels have a different fractional absorption of cholesterol is innovative and worthy of further evaluation. On the other hand, methodological issues raised by this study deserve consideration.

The method for measuring fractional absorption of cholesterol employed a dual stable-isotope method (1). [C¹³]cholesterol was given intravenously, and 50 mg of [H²]cholesterol was administered in a stomach-soluble gelatin capsule together with a standardized breakfast. Seemingly, [H²]cholesterol was encapsulated in crystalline form and was not dissolved in oil before administration. This method of administration differed from that reported by Lütjohann et al. (8); these workers dissolved small quantities of [H²]cholesterol and [H²]sitosterol in sunflower oil by gentle heating and stirring before adding it to stomach-soluble gelatin capsules. Cholesterol absorption was determined by differences in isotope ratios between oral administration and fecal sterols. This method has the potential advantage of requiring much smaller amounts of isotopic cholesterol than was employed by Jakulj et al. (1); presumably larger amounts of oral [H²]cholesterol were required to label all body pools. This could create a solubility problem and reduce the bioavailability of the oral label.

A need for large amounts of oral [H2]cholesterol might account for the relatively low estimates of fractional cholesterol absorption compared with previous reports; Jakulj et al. (1) found absorption in the range of 25%, which contrasts to many previous reports in which absorption generally ranged from 40% to 60% (2, 6, 9^–14). However, Jakulj et al. (1) were aware of this potential issue; they performed a substudy that compared absorption measured with crystalline [H²]cholesterol and with [H²]cholesterol dissolved in coconut oil. The data were presented in supplemental material. Similar results were obtained, i.e., fractional absorption remained in the range of 25%. Because of these apparently low estimates of fractional absorption, further studies will be required to determine whether 50 mg of [H²]cholesterol can be adequately dissolved to ensure that it reaches the intestine in bioavailable form.

Several isotopic methods have been employed for measurement of cholesterol absorption. Some used a single dose of oral cholesterol, as used by Jakulj et al. (1); others administered the labeled cholesterol multiple times a day over a period of a week or more (9, 15). The latter carries the potential advantage of giving a more complete estimate of average daily absorption. A rigorous comparison of single and continuous dosing remains to be done; nonetheless, both separately have given a fractional absorption of 40% to 60% (2, 6, 9^–14, 16).

Miettinen et al. (17) have acknowledged the limitations of noncholesterol sterols for estimating cholesterol absorption. They suggest that in evaluating cholesterol metabolism, it is preferable to employ several cholesterol precursors and plant sterols in plasma to mark absorption; and confidence will be enhanced by using at least one absolute measurement. They further note that during consumption of plant sterol-enriched diets, noncholesterol sterols do not adequately reflect cholesterol metabolism.

The limitation of fractional cholesterol absorption as a marker of absolute cholesterol absorption must be recognized as well. Absorbed cholesterol is derived from both dietary cholesterol and biliary cholesterol. If biliary cholesterol is high, as occurs with obesity, dietary cholesterol will be diluted and fractional absorption will be low. And conversely, if biliary cholesterol is low, as occurs in sitosterolemia, fractional absorption will be relatively high. To determine absolute absorption, it is necessary to utilize intestinal perfusion techniques that account for both oral and biliary cholesterol (18).

To the knowledge of this author, no category of subjects consistently exhibit fractional cholesterol absorption above 50–60%. An exception may be subjects with differences in apoE genotypes; those with the E4 genotype seemingly had higher cholesterol absorption than those with the more common E3 genotype (12). Otherwise, the primary use of noncholesterol sterols has been to identify subjects with a reduced cholesterol absorption. This is especially the case for testing of agents or identifying conditions that interfere with cholesterol absorption. For example, Miettinen and colleagues (19) reported a reduction in plasma noncholesterol sterols in a variety of circumstances in which a reduction of cholesterol absorption was plausible or certain: during administration of plant stanols; after Roux-en-Y gastric bypass (20); in primary biliary cirrhosis (21); and in those with genetic forms of low cholesterol absorption (22). Reductions in absorption likewise have been reported by others during treatment with plant stanols (23), ezetimibe (24^–26), surfomer (AOMA) (10), sucrose polyester (olestra) (27), and Lactobacillus reuteri (28).

In a noteworthy study, Sudhop et al. (24) tested effects of ezetimibe on absorption of cholesterol using the continuous-isotope method. On placebo, fractional cholesterol absorption averaged 50%, and on ezetimibe therapy it fell to 23%. Plasma noncholesterol plant sterols, adjusted for total cholesterol, fell in parallel. This confirms a strong relationship between cholesterol absorption and plant sterol levels.

There is other evidence that supports the utility of the campesterol/lathosterol ratio as tracer of cholesterol absorption, for example, in comparison of subjects with NPC-1L1 mutations who carry a defect in intestinal absorption of cholesterol (29). Campesterol/lathosterol ratios are relatively high in unaffected individuals and are low in homozygotes for the mutation. Ratios are intermediate in heterozygotes.

In summary, a wealth of data from several laboratories indicates that plasma noncholesterol sterols reflect cholesterol absorption. This relationship has been confirmed in many studies. This does not mean that fractional cholesterol absorption can be quantitatively measured by plasma noncholesterol sterols. It cannot. Plasma noncholesterol sterols are most useful in comparison studies, especially those that test effects of various agents, diseases, or genetic factors that may interfere with cholesterol absorption. As Miettinen et al. (17) have pointed out, differences in plasma noncholesterol sterols are most useful when confirmed by more-direct measures of cholesterol absorption. But it must be kept in mind that isotopic measures of fractional cholesterol absorption fail to quantify absolute rates of cholesterol absorption because of discrepancies introduced by variation in biliary cholesterol secretion.