Grant N. Stemmermann

A prospective study of diet, smoking, and lower urinary tract cancer☆

The association of diet and smoking with bladder cancer was investigated in a cohort study conducted in Hawaii. The study included 7995 Japanese-American men who were born between 1900 and 1919, and were examined from 1965 to 1968. After 22 years of follow-up, 96 incident cases of bladder cancer were diagnosed. Current cigarette smokers had a 2.9-fold risk of bladder cancer, compared with nonsmokers. A direct dose-response relation was observed, based on pack-years of cigarette smoking. Consumption of fruit was inversely associated with the risk of bladder cancer (P = 0.038). The relative risk was 0.6 among subjects who had the most frequent (> or = 5 times/wk) intake of fruits compared to those with the least intake (< or = 1 time/wk). A weaker inverse association with milk intake was also observed (P = 0.07). Frequent consumption of fried vegetables, pickles, or coffee increased the risk of bladder cancer, but none of these foods showed a significant dose-response relationship. There was no association of other selected foods, alcohol, total calories, protein, fat, or carbohydrates with bladder cancer risk.

Pathogenesis of Hyperplastic Polyps of the Colon: A Hypothesis Based on Ultrastructure and in vitro Cell Kinetics

The morphology of hyperplastic polyps and the cell kinetics of their cell populations suggests that they result from hypermaturation of epithelium on the colon surface. The morphological features which support this hypothesis are the superficial location of maximal change, cell elongation, increase in the number and length of microvilli, accentuation of lateral intercellular interdigitations, and an increase in the breadth of basal lamina attachment. Cell kinetics studies indicate that their mode of cell renewal is the same as that of normal mucosa, but with longer turnover time and delayed migration.

Elevated serum pepsinogen I and II levels differ as risk factors for duodenal ulcer and gastric ulcer.

We investigated the possibility that serum pepsinogen I (PG I) and pepsinogen II (PG II) levels might differ as risk factors for duodenal ulcer and gastric ulcer. From 1967 to 1970, serum was obtained from 7498 Japanese men in Hawaii, and the cohort was followed up to 1981 for the development of peptic ulcer disease. Pepsinogen I and PG II levels in stored serum were significantly higher in the subjects who developed duodenal ulcer (n = 43) or gastric ulcer (n = 115) than in 212 control subjects. The linear trend in risk of each type of ulcer was highly significant for both PG I and PG II. An elevated serum PG I level (greater than or equal to 130 micrograms/L), however, was associated with about a threefold higher odds ratio for duodenal ulcer than for gastric ulcer (8.37 vs. 2.83), whereas an elevated PG II level (greater than or equal to 30 micrograms/L) was associated with about a threefold higher odds ratio for gastric ulcer than for duodenal ulcer (18.21 vs. 6.49). In contrast, the PG I/PG II ratio was significantly lower in the gastric ulcer than in the control and duodenal ulcer cases, and showed a significant linear trend in risk only for gastric ulcer. In addition, a PG I/PG II ratio of less than 4.0, which has been shown previously to be indicative of chronic gastritis, was associated with an almost 10-fold higher odds ratio for gastric ulcer than for duodenal ulcer (7.35 vs. 0.79). The results indicate that an elevated serum PG I level is a major risk factor for duodenal ulcer, whereas an elevated serum PG II level and a low PG I/PG II ratio are major risk factors for gastric ulcer.

Prospective study of alcohol intake and large bowel cancer.

The alcohol intake of a cohort of Japanese men in Hawaii is directly and significantly related to the risk of developing rectal cancer, whether assessed on the basis of amount consumed or as a percent of total calories. Wine and whiskey are directly related to rectal cancer, but beer is the only alcoholic beverage that displays a statistically significant dose-response (P=0.008) Colon cancer risk also is related directly to alcohol intake, but the association is statistically significant only when measured as a percent of energy intake. This suggests that alcohol might displace cancer inhibitors from the diet. Calcium, vitamin C, and dietary fiber are inversely related to colon cancer risk in this cohort, and each of these micronutrients displays statistically significant negative correlation with alcohol intake. A possible positive association between alcohol and lung cancer was ruled out after adjusting for cigarette smoking. Cancers of the prostate and stomach were unrelated to alcohol intake, but the risk of acquiring cancer at all other sites combined was strongly related to alcohol intake.

Prospective study of clinical gallbladder disease and its association with obesity, physical activity, and other factors

The association of anthropometric measurements, serum tests, and life-style factors with the risk of clinical gallbladder disease was investigated in a prospective study of 7831 American men of Japanese ancestry in Hawaii. After 152,831 person-years of observation, 471 incident cases of gallbladder disease were diagnosed by histology or radiology. There was an increase in risk of gallbladder disease with a progressive increase in body mass index, height, pack-years of cigarette smoking, and diastolic blood pressure. There was a decrease in risk with an increase in physical activity, after controlling for the effect of other variables. There was also an inverse association with total caloric intake that could be related to dietary restriction in obese subjects or to the limitations of the 24-hr diet history. Serum cholesterol levels and the intake of total fat were not related to clinical gallbladder disease in this cohort.

The influence of dairy and nondairy calcium on subsite large-bowel cancer risk

This prospective study assesses the impact of fat and calcium intake on the risk of developing cancer in each large-bowel subsite. The study population is a cohort of Hawaii Japanese men who experience high rates of colon cancer, especially of the sigmoid segment. Total calcium intake is not related to the risk of colon cancer, and separation of calcium into dairy and nondairy sources does not alter the result. There is, however, a significant, monotonic increase in sigmoid colon cancer risk with decreasing total calcium intake. Similar trends are shown for both dairy and nondairy calcium. Dietary calcium is not consumed in large quantities among the Hawaii Japanese, partly because of their limited consumption of milk due to lactose intolerance. If calcium plays a protective role against sigmoid colon cancer, this effect is unlikely to be related to fat intake. Sigmoid colon cancer subjects had lower intakes of fat than other cohort men, and a statistical test for the interaction effect of total calcium and fat intake on colon cancer risk was statistically insignificant (P=0.2).

Cryptosporidiosis Report of a fatal case complicated by disseminated toxoplasmosis

A middle-aged woman suffered from chronic diarrhea and malnutrition as a result of a small intestinal infection with a coccidial protozoan--Cryptosporidium. This parasite has been found among a wide range of reptilian, avian and mammalian hosts, but rarely in human beings. This woman ultimately died as a result of disseminated toxoplasmosis of the type usually encountered in an adult whose immune responses were compromised. There was clinical evidence of both humoral and cellular immunologic incompetence, such as depressed levels of complement and immunoglobulins, and decreased T cell reactivity, respectively. The former could be accounted for on the basis of malnutrition causing deficient protein synthesis, and the latter may have been a manifestation of altered function of transformed T cells.

Serum pepsinogens I and II and stomach cancer

Prospective epidemiologic studies among Hawaiian Japanese men have shown that a serum pepsinogen I (PG I) level below 20 ng/ml is highly specific for extensive intestinal metaplasia of the stomach and for the intestinal type of stomach cancer. The test, however, shows a low level of sensitivity as a predictor of the intestinal type stomach cancer. Since antralization of the corpus, as encountered in Type A atrophic gastritis and late stage Type B gastritis, results in an increase of PG II relative to PG I, we examined the sensitivity and specificity of the PG I/PG II ratio as a predictor of gastric cancer. Using a cut-off point of 2.0 in the ratio to separate subjects at high and low risk for stomach cancer, we found a modest improvement in the level of sensitivity with a small decrease in specificity. Abnormally low PG I levels and PG I/PG II ratios were associated mainly with advanced stages of the intestinal type of cancer and therefore, are not useful screening tests to identify early cases of gastric cancer. The use of the PG I/PG II ratio resulted in a modest improvement in the level of sensitivity with a small decrease in specificity.

Colchicine intoxication. A reappraisal of its pathology based on a study of three fatal cases.

Abstract The tissue changes found with colchicine toxicity are described and correlated with the clinical manifestations of this state. Intercellular and intracellular edema, depressed enzyme activity, and structural distortions of the cytoplasm and nuclei of differentiated nondividing intestinal mucosal cells account for the dehydration and hyponatremia found in two patients who died during acute toxic episodes. These changes would appear to be a more immediate threat to life than disturbances of cell replication. The latter, however, can explain the associated thrombocytopenia and leukopenia. Mucosal damage, inhibition of granulocyte function, and a possibly altered intestinal flora explain the development of staphylococcal septicemia in a third patient.

Prospective study of serum cholesterol and site-specific cancers

From 1965 to 1968, 7716 Japanese-American men were examined and tested for serum cholesterol. After 22 years, 1380 incident cancer cases were identified. Of the site-specific cancers, only colon cancer cases had a significantly lower mean serum cholesterol value than that of noncases (213.0 mg/dl vs 219.0 mg/dl). When the study subjects were separated into either a low, middle or high group, based on their serum cholesterol values, there was a significant inverse trend for cases of oral/pharyngeal/esophageal cancer combined. The association was present for cases diagnosed within 10 years of examination (p = 0.012), but not for cases diagnosed after 10 years. This suggests that the inverse association is due to the metabolic effects of undiagnosed oral/pharyngeal/esophageal cancer upon serum cholesterol levels. These results are discussed in relation to other studies on serum cholesterol.