Current issues of ACP Journal Club are published in Annals of Internal Medicine


Low serum cholesterol levels were associated with an increased risk for cancer among Japanese American men

ACP J Club. 1992 Sept-Oct;117:59. doi:10.7326/ACPJC-1992-117-2-059

Source Citation

Chyou PH, Nomura AM, Stemmermann GN, Kato I. Prospective study of serum cholesterol and site-specific cancers. J Clin Epidemiol. 1992 Mar;45:287-92.



To examine the association between serum cholesterol levels and the occurrence of site-specific cancers among American men of Japanese ancestry.


22-year cohort study with adjustment for age and for cigarette and alcohol use.


The cohort was identified by the Honolulu Heart Program through draft registration files from 1942.


8006 (72%) of 11 148 Japanese American men born from 1900 to 1919 had a serum sample taken between 1965 and 1968. After excluding 245 men with prevalent or histologically unconfirmed cancer and 45 for whom no cholesterol measurement was available, 7716 men were included.

Assessment of risk factors

Cholesterol values for nonfasting frozen serum specimens were determined by the Public Health Service laboratory in San Francisco. Values were divided into 3 groups with cut points at 5.17 mmol/L (200 mg/dL) and 6.08 mmol/L (235 mg/dL). Information on other risk factors was collected at the baseline examination.

Main outcome measure

Association between initial cholesterol level and the development of cancer classified according to primary site. Incident cancer cases were retrieved from the population-based cancer registry, with an estimated coverage of 99% of the study sample.

Main results

Age-adjusted mean baseline serum cholesterol levels were lower for 1380 men who developed any cancer than for the 6336 who did not (5.59 mmol/L vs 5.66 mmol/L, P = 0.013), and lower for 231 men with colon cancer (5.47 mmol/L, P = 0.018) but not for prostate, rectum, stomach, or lung cancer. 402 men with less common cancers (renal pelvis, ureter, bladder, kidney cancer; lymphoma; pancreatic cancer; leukemia; liver, esophageal, laryngeal, gallbladder cancer; skin cancer; or at other or unknown primary sites) had lower cholesterol values (5.48 mmol/L, P = 0.006). Serum cholesterol was not significantly lower for any less common cancer, but the relative risks for oral, pharyngeal, and esophageal cancer, when adjusted for age, smoking, and alcohol consumption, declined from 1.00 in the lowest cholesterol tertile to 0.68 (95% CI 0.35 to 1.32) in the middle range and 0.52 (CI 0.24 to 1.13) in the upper tertile (P for trend, 0.045). The trend was stronger for cancers developing within 10 years (P = 0.012).


Low serum cholesterol levels were associated with the development of any cancer, colon cancer, and less common cancers, during the 22-year follow-up.

Source of funding: National Institutes of Health.

Address for article reprint: Dr. P.H. Chyou, Japan-Hawaii Cancer Study, Kuakini Medical Center, 347 North Kuakini Street, Honolulu, HI 96817, USA.


There is good reason to suspect a relation between cholesterol and colon cancer. An early international study found a correlation (r = 0.6) between the mortality rates from coronary artery disease and colon cancer (1). 2 more recent cohort studies have yielded inconsistent results. The study by Chyou and colleagues suggests that low serum cholesterol is associated with an increased risk for colon cancer. An equally rigorous cohort study, however, showed that high serum cholesterol was associated with an increased risk for colorectal cancer in men (2).

Chyou and colleagues also noted an inverse linear trend of relative risk between serum cholesterol and cancers of the oral cavity, pharynx, and esophagus when these cancer sites were combined; when these sites were considered individually, however, the inverse relation was no longer significant. Previous research has also not shown a clear relation between cancers other than colon cancer and low serum cholesterol levels.

Is the relation between cholesterol and colon cancer causal (in either direction), due to chance, or the result of the presence of a confounder? Confounding could be a problem. The authors controlled for smoking, age, and alcohol, but not for other factors such as dietary fat and individual variations in serum cholesterol during the period of the study.

The controversy remains unsettled: No final conclusion can be reached about an association between serum cholesterol levels and the development of colon and less common site-specific cancers.

James E. Shaw, MD
University of North CarolinaChapel Hill, North Carolina, USA

James E. Shaw, MD
University of North Carolina
Chapel Hill, North Carolina, USA


1. Wynder EL, Shigematsu T. Environmental factors of cancer of the colon and rectum. Cancer. 1967;20:1520-61.

2. Törnberg SA, Holm LE, Carstensen JM, Eklund GA. Risks of cancer of the colon and rectum in relation to serum cholesterol and beta-lipoprotein. N Engl J Med. 1986;315:1629-33.