International Agency for Research on Cancer (IARC) - Summaries & Evaluations


VOL.: 31 (1983) (p. 95)

5. Summary of Data Reported and Evaluation

5.1 Experimental data

Cholesterol was tested for carcinogenicity in mice by administration in the diet, by subcutaneous administration and by bladder implantation. These studies were all inadequate for evaluation. Cholesterol was also tested in combination with various carcinogens, but the results are insufficient to assess the co-carcinogenic potential of this compound.

Cholesterol, when free of oxidation products, was not mutagenic in bacteria and did not cause cell transformation. The data were inadequate to evaluate the activity of cholesterol in short-term tests.

Cholesterol induced cleft palate in rats.

5.2 Human data

Cholesterol and cholesterol derivatives are widely distributed in human tissues and in the human diet. Cholesterol is used in a multitude of pharmaceutical and cosmetic preparations.

As in the preceding presentation of epidemiological data, the evidence is summarized according to the three contexts in which cholesterol has been measured, i.e., in the diet, the blood and the faeces.

Few epidemiological studies of cancer have examined dietary cholesterol specifically, although many have examined total saturated fat intake. At the population level, cancers of the colon and female breast (and, to a lesser extent, the prostate, endometrium, ovary, pancreas and rectum) are strongly positively correlated with estimated per-caput intake of dietary saturated fat and, where studied, cholesterol. In the few case-control studies that have estimated individual dietary cholesterol intake, cancers of the breast and colon-rectum have tended to be positively associated with dietary total fat, saturated fat and cholesterol. Although there is moderately consistent, albeit limited, evidence of an association between high dietary intake of cholesterol and cancers of the colon, breast and, less convincingly, some other cancers, a causal relationship cannot be inferred from these epidemiological data. In particular, there is a difficulty in distinguishing between any effect of cholesterol and the effects of other dietary factors with which cholesterol is positively or negatively correlated.

With respect to serum cholesterol, the findings from seven cholesterol-lowering intervention trials (five by diet, two by drugs) indicate no alteration in cancer risk consequent upon a reduction in individual serum cholesterol. However, long-term observational follow-up studies have shown either an inverse relationship between individual serum cholesterol (measured at entry into the study) and subsequent total cancer risk (mortality or incidence) or no such relationship. The lack of consistency of these studies, and the partial attributability of the observed inverse relationships to a presumed cholesterol-lowering effect of subclinical cancer (the 'undetected cancer effect'), preclude any conclusion that there is an increased general cancer risk consequent upon low serum cholesterol.

It seems clear that serum cholesterol does not have a strong or direct relationship with human carcinogenesis in general, of the kind it apparently has with coronary heart disease. The inverse relationship with total cancer risk, when present, is not strong, is not graded (but, instead, tends to occur in stepwise fashion below concentrations of 180-190 mg/dl), and applies predominantly to men. For cancer of the colon, specifically, the prospective observational studies show a moderately consistent inverse association with serum cholesterol.

No data were available to assess the teratogenicity or chromosomal effects of this compound in humans.

5.3 Evaluation

No evaluation of the carcinogenicity of cholesterol to experimental animals could be made.

There is inadequate evidence from epidemiological studies that cholesterol as such is carcinogenic to humans.

There is limited evidence to indicate that raised dietary intake of cholesterol by individuals is associated with an increased risk of breast and colo-rectal cancer.

There is strong evidence that a low serum cholesterol level attained by dietary or pharmacological means is not per se associated with an increased risk of cancer.

There is limited evidence that, within the populations studied, those male individuals with relatively low serum concentrations have an increased risk of colon cancer; the data pertaining to women were inadequate for evaluation. There is inadequate evidence that a low serum cholesterol level increases the risk of cancer at sites other than the colon.

The available epidemiological and experimental studies do not permit an evaluation of the carcinogenicity to humans of cholesterol per se.

N.B. - The pattern of associations between cholesterol and human cancer risk contains some apparent contradictions. Depending on whether cholesterol is measured in the diet, blood or faeces, its association with cancer risk may tend to be either positive or negative. Because the biological significance, in relation to carcinogenesis, of cholesterol in each of these basically different contexts is not sufficiently understood, it is not yet possible to reconcile these apparently divergent findings.

For definition of the italicized terms, see Preamble Evaluation.

Previous evaluation: Vol. 10 (1976)

Subsequent evaluation: Suppl. 7 (1987)

Last updated: 16 April 1998

    See Also:
       Toxicological Abbreviations