Am J Epidemiol

Cryptosporidium in tap water: comparison of predicted risks with observed levels of disease.

Perz JF. Ennever FK. Le Blancq SM.
Division of Environmental Health Sciences, School of Public Health, Columbia University, New York, NY 10032, USA.
Waterborne transmission of Cryptosporidium parvum is well-established as a source in outbreaks of cryptosporidiosis; however, the role of tap water in endemic disease is unclear. The authors applied a risk assessment approach incorporating uncertainty analysis to examine the potential role of tap water in the transmission of endemic C. parvum infection. The model had two components: exposure-infection, to relate low-dose exposure to infection; and infection-outcome, to include the probabilities of clinical outcomes leading to case detection and reporting. The population was divided into four subgroups: adults and children with and without acquired immunodeficiency syndrome (AIDS). Because of the high degree of uncertainty associated with available measures, a plausible baseline concentration of oocysts, 1 per 1,000 liters, was assumed for input to the model. In the non-AIDS subgroups, the predicted median annual risk of infection was approximately 1 in 1,000 (non-AIDS adults: 0.0009 infection/person/year, 95% confidence interval (CI) 0.0003-0.0028), while in the AIDS subgroups the predicted risk was 2 in 1,000 (AIDS adults: 0.0019 infection/person/year, 95% CI 0.0003-0.0130). When the risks were applied to the 1995 New York City population, more than 6,000 infections were estimated, with 99% occurring in the non-AIDS categories. Estimates of the overall probabilities that an infection would result in a reported case predicted that three reported illnesses would occur out of every 10,000 infections in non-AIDS adults (95% CI 5 x 10[-5] to 2 x 10[-3]), with a 10-fold higher probability in the non-AIDS pediatric subgroup. In contrast, the majority of infections occurring in the AIDS subgroup were predicted to result in reported cases (AIDS adults: probability = 0.61, 95% CI 0.39-0.80). When the model was applied to the New York City population, the calculated number of tap-water-related cases per year in the non-AIDS subgroups was six (95% CI 1-29), and in the AIDS subgroups it was 34 (95% CI 6-240).

Molecular epidemiology of p53 protein mutations in workers exposed to vinyl chloride.

Smith SJ. Li Y. Whitley R. Marion MJ. Partilo S. Carney WP. Brandt-Rauf PW.
Division of Environmental Health Sciences, School of Public Health, Columbia University, New York, NY 10032, USA.
The production of mutations in cellular tumor suppressor genes such as p53 is involved in the development of many human cancers. These mutations result in the expression of mutant forms of the encoded p53 protein which can potentially serve as a biomarker for this carcinogenic process. Workers exposed to vinyl chloride who are at risk for the development of the sentinel neoplasm angiosarcoma of the liver represent a model population for the study of such a mutant p53 biomarker, since vinyl chloride is known to cause specific p53 mutations in persons with angiosarcoma of the liver. To determine the relation between vinyl chloride exposure and this p53 biomarker, the authors examined serum samples collected between 1987 and 1992 from a cohort of 225 French vinyl chloride workers and 111 unexposed controls (matched according to age, sex, race, smoking, and alcohol drinking) for the presence of mutant p53 protein, using an enzyme-linked immunosorbent assay. Stratification of the exposed workers by quartile of vinyl chloride exposure (in estimated ppm-years) yielded a statistically significant trend of increasing odds ratios for p53 biomarker seropositivity with increasing exposure. These results suggest that this serum biomarker for mutant p53 protein is related to vinyl chloride exposure and may be an early indicator of carcinogenic risk in exposed individuals.

4-Aminobiphenyl DNA damage in liver tissue of hepatocellular carcinoma patients and controls.

Wang LY. Chen CJ. Zhang YJ. Tsai WY. Lee PH. Feitelson MA. Lee CS. Santella RM.
Division of Environmental Health Sciences, School of Public Health, Columbia University, New York, NY, USA.
Epidemiologic studies have suggested that cigarette smoking is a risk factor for the development of hepatocellular carcinoma (HCC). To further investigate this relation, the authors measured levels of 4-aminobiphenyl-DNA adducts by an immunoperoxidase method in surgical liver tissues obtained between 1984 and 1995 from 105 Taiwanese patients with histologically confirmed HCC and 37 Taiwanese patients with metastatic liver tumors or intrahepatic stones. Information on clinicopathologic characteristics, cigarette smoking, and alcohol drinking was abstracted from hospital charts. Mean relative staining intensity for 4-aminobiphenyl-DNA was slightly higher in tumor tissues than in nontumor tissues obtained from HCC patients. Both mean intensities were significantly higher than the mean intensity of control tissues taken from non-HCC patients. However, no difference in mean relative staining intensity was found between smokers and nonsmokers in tissues obtained from non-HCC patients, or in tumor or nontumor tissues taken from HCC cases. After stratification of the relative staining intensities of 4-aminobiphenyl-DNA adduct levels into tertiles according to the total numbers of control tissues analyzed, there was a monotonically increasing risk of HCC. Odds ratios were 4.14 (95% confidence interval (CI) 1.15-15.50) and 9.71 (95% CI 2.82-34.86) for medium and high adduct levels compared with low adduct levels, respectively. The linear relation between adduct levels in liver tissue and HCC risk was also significant after adjustment for covariates, including hepatitis B surface antigen (HBsAg) status. The multivariate adjusted odds ratios were 3.41 (95% CI 0.82-14.25) and 6.48 (95% CI 1.59-26.50) for medium and high adduct levels, respectively. Moreover, there were monotonically increasing HCC risks for higher adduct levels in both HBsAg carriers and noncarriers. The increased risk ratios were more pronounced in noncarriers than in carriers. However, because of the small numbers of subjects, especially controls positive for HBsAg, the interaction between HBsAg status and 4-aminobiphenyl-DNA adduct level was not significant. Among HCC cases, none of the clinicopathologic characteristics were associated with relative staining intensity. These results indicate that 4-aminobiphenyl exposure, which is primarily a result of cigarette smoking, plays a role in the development of HCC in humans.

Prediction and cross-validation of neural networks versus logistic regression: using hepatic disorders as an example.

Year 1998
Duh MS. Walker AM. Pagano M. Kronlund K.
Department of Epidemiology, Harvard School of Public Health, Boston, MA 02115, USA.
The authors developed and cross-validated prediction models for newly diagnosed cases of liver disorders by using logistic regression and neural networks. Computerized files of health care encounters from the Fallon Community Health Plan were used to identify 1,674 subjects who had had liver-related health services between July 1, 1992, and June 30, 1993. A total of 219 subjects were confirmed by review of medical records as incident cases. The 1,674 subjects were randomly and evenly divided into training and test sets. The training set was used to derive prediction algorithms based solely on the automated data; the test set was used for cross-validation. The area under the Receiver Operating Characteristic curve for a neural network model was significantly larger than that for logistic regression in the training set (p = 0.04). However, the performance was statistically equivalent in the test set (p = 0.45). Despite its superior performance in the training set, the generalizability of the neural network model is limited. Logistic regression may therefore be preferred over neural network on the basis of its established advantages. More generalizable modeling techniques for neural networks may be necessary before they are practical for medical research.

Incidence and risk factors for self-reported peptic ulcer disease in the United States.

Year 1998
Everhart JE. Byrd-Holt D. Sonnenberg A.
Division of Digestive Diseases and Nutrition, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD 20892-6600, USA.
Incidence and risk factors for peptic ulcer disease in the United States have not been well defined. During the 1989 National Health Interview Survey, a population-based sample of 42,392 individuals responded to questions regarding doctor-diagnosed ulcers with confirmation by either an upper gastrointestinal series or endoscopy. Ulcers present during the previous 12 months were considered either incident ulcers if diagnosed during this period or chronic active ulcers if diagnosed more than 12 months before the interview. The incidence of ulcers over the year prior to the interview was 5.27 per 1,000 adults. Whereas incident duodenal ulcer cases represented only 2.4 percent of all persons with a history of duodenal ulcer, the corresponding value for gastric ulcer was 8.7 percent. Risk factors for incident ulcers included increasing age, lower income and educational attainment, and musculoskeletal pain or headache. These were similar to risk factors for chronic active ulcers, except smoking was an additional important risk factor for chronic active ulcers. Thus, incident peptic ulcers are common in the United States but represent a small proportion of persons with a history of ulcer disease. Smoking may be a stronger risk factor for chronic ulcers than for new ulcers.

Risk factors for horizontal transmission of hepatitis B virus in a rural district in Ghana.

Year 1998
Martinson FE. Weigle KA. Royce RA. Weber DJ. Suchindran CM. Lemon SM.
Department of Epidemiology, School of Public Health, The University of North Carolina at Chapel Hill 27599-7400, USA.
Most hepatitis B virus (HBV) infections in sub-Saharan African infants and children are acquired through horizontal transmission, but the exact mechanisms of spread have not been documented. The authors conducted a study in rural Ghana which determined seroprevalence in a probability sample of 1,385 individuals of all ages, and evaluated risk factors for horizontal transmission of HBV in a subsample of 547 children aged 1-16 years who were not hepatitis B surface antigen (HBsAg) carriers. Most residents in this district live in compounds which typically contain 2-4 households each. Overall prevalence of HBV seropositives (any HBV marker) was 74.7% (95% confidence interval (CI) 72.5%-76.9%). Prevalence of HBsAg was 20.9% (95% CI 18.8%-23.1%). The data suggest a continuous nonuniform acquisition of HBV infection with advancing age predominantly through horizontal transmission in childhood, with the household, rather than the domestic compound, being the primary place for transmission. The behaviors most strongly associated with prevalence of HBV were sharing of bath towels (OR = 3.1, 95% CI 2.1-4.5), sharing of chewing gum or partially eaten candies (OR = 3.4, 95% CI 2.3-5.0), sharing of dental cleaning materials (OR = 2.5, 95% CI 1.3-4.6), and biting of fingernails in conjunction with scratching the backs of carriers (OR = 2.5, 95% CI 1.6-4.3).

Body mass index, cigarette smoking, and other characteristics as predictors of self-reported, physician-diagnosed gallbladder disease in male college alumni.

Year 1998
Sahi T. Paffenbarger RS Jr. Hsieh CC. Lee IM.
Department of Public Health, University of Helsinki, Finland.
Despite the high prevalence of gallbladder disease in industrialized countries, little is known about the predictors of the disease, especially in men. The authors prospectively studied 16,785 alumni of Harvard University, aged 15-24 years, who were followed for up to 61 years. The health characteristics of these men were ascertained from their college entrance physical examination done in 1916-1950, and updated via mailed questionnaires in 1962 or 1966 (1962/1966). Alumni subsequently self-reported physician-diagnosed gallbladder disease on further mailed questionnaires in 1972 or 1977. Between college time and 1962/1966, 371 gallbladder disease cases occurred. An additional 314 cases occurred after 1962/1966. With respect to college characteristics, after adjustment for potential confounders, the authors found that body mass index (BMI), smoking, physical activity, blood pressure, and consumption of alcohol, coffee, or tea were unrelated to risk. However, BMI in 1962/1966 was directly related to risk of subsequent gallbladder disease, as was BMI gain since college (p, trend = 0.002 and 0.013, respectively). Compared with men with BMI < 22.0, men with BMI > or = 27.0 had a rate ratio of 2.71 (95% confidence interval (CI) 1.57-4.66) for risk of contracting the disease. Men who gained > or = 6.0 BMI units since college had a rate ratio of 1.46 (95% CI 0.86-2.46) compared with men who gained -0.9 to +0.9 BMI units. Compared with never smokers, men who smoked pipes or cigars or < 1 pack of cigarettes daily in 1962/1966 had a rate ratio of 1.43 (95% CI 1.00-2.06), while heavier cigarette smokers had a rate ratio of 1.52 (95% CI 1.03-2.24). Neither physical activity nor physician-diagnosed hypertension or diabetes mellitus in 1962/1966 predicted risk.

Obesity, weight gain, large weight changes, and adenomatous polyps of the left colon and rectum.

Year 1998
Bird CL. Frankl HD. Lee ER. Haile RW.
Department of Preventive Medicine, School of Medicine, University of Southern California, Los Angeles 90033, USA.
Epidemiologic studies of colorectal neoplasia have usually examined body mass index as a risk factor, but not other aspects of obesity. During 1991-1993, the authors obtained weight histories and comprehensive covariate data from men and women aged 50-75 years who underwent sigmoidoscopy at a health maintenance organization in southern California. Using 483 cases with adenomas and 483 controls, measures of obesity (body mass index), positive energy balance (net weight gain in the past 10 years), and weight variability (large weight changes) were each independently related to adenoma prevalence. Compared with subjects in the lowest quartile of body mass index, multivariate-adjusted odds ratios for subjects in increasingly higher quartiles were 2.1 (95% confidence interval (CI) 1.4-2.3), 1.8 (1.1-2.9), and 1.7 (1.0-2.8), respectively. Compared with subjects who reported a net weight loss during the 10 years before sigmoidoscopy, subjects with net weight gains of 1.5-4.5 kg or > or = 4.5 kg had adjusted odds ratios (95% CI) of 2.5 (1.2-5.6) and 1.8 (0.7-4.4), respectively. Compared with subjects who had no large weight changes during adulthood, subjects with 1-2, 3, or > or = 4 changes had adjusted odds ratios (95% CI) of 2.0 (1.0-3.9), 2.5 (1.2-5.5), and 1.5 (0.6-3.6), respectively. Obesity, weight gain, and unstable adult weight may be independently associated with colorectal carcinogenesis.

Cigarette smoking and the colorectal adenoma-carcinoma sequence: a hypothesis to explain the paradox.

Year 1998
Terry MB. Neugut AI.
Division of Epidemiology, School of Public Health, Columbia University, New York, NY, USA.
As recognized precursor lesions to colorectal cancer, colorectal adenomatous polyps have been studied to enhance knowledge of colorectal cancer etiology. Although most of the known risk factors for colorectal cancer are also associated with the occurrence of colorectal adenomas, cigarette smoking has had a strong, consistent relationship with colorectal adenomas but is generally not associated with colorectal cancer. The explanation for this paradox is unknown. With data collected in 1986-1988 during a large case-control study based on colonoscopy results in New York City, New York, the authors investigated the possibility that the paradox may arise because subjects with colorectal adenomas were included in the control group of cancer case-control studies. The authors found a statistically significant increased risk between heavy cigarette smoking (smokers with > or = 40 pack-years of smoking) and risk of adenoma (odds ratio (OR) = 1.61, 95% confidence interval (CI) 1.06-2.44). They saw no increased colorectal cancer risk from heavy cigarette smoking (OR = 1.02, 95% CI 0.52-1.99) using a "manufactured" control group to simulate a typical unscreened, population-based control group. When the authors compared these colorectal cancer cases with an adenoma-free control group examined by colonoscopy in a polytomous model with several case groups (newly diagnosed adenomas, carcinoma in situ, intramucosal carcinoma, and colorectal cancer), they found that the risk for 20-39 pack-years of smoking was elevated, although not statistically significant, and was similar for all four case groups. The risk for the highest smoking category (> or = 40 pack-years) was more strongly elevated in all four case groups, although it was statistically significant for only the newly diagnosed adenoma and the carcinoma in situ cases (adenomas, OR = 1.59, 95% CI 1.05-2.42; carcinoma in situ, OR = 2.05, 95% CI 1.01-4.15; intramucosal carcinoma, OR = 1.30, 95% CI 0.61-2.77; and colorectal cancer, OR = 1.30, 95% CI 0.64-2.65). While the authors' study is weakened by the lack of statistical significance concerning risk for colorectal cancer, these data offer some support for the hypothesis that the association between cigarette smoking and risk of colorectal cancer may have been masked by inclusion in the control group of subjects with adenomas. They also suggest that the major effect of smoking on the colorectal adenoma-carcinoma sequence occurs in the earlier stages of the formation of adenoma and the development of carcinoma in situ.

Meta-analysis of coffee consumption and risk of colorectal cancer.

Year 1998
Giovannucci E.
Department of Medicine, Harvard Medical School and Brigham and Women's Hospital, Boston, MA, USA.
Several studies have found that coffee consumption is related to a lower risk of colorectal cancer, but results have not been consistent. Thus, a meta-analysis of the published articles was conducted to examine this relation. Because of the various ways data were collected and analyzed, a "semiquantitative" approach that compared the high versus the low category of intake for each study was used. The combined results from 12 case-control studies showed an inverse association between coffee consumption and risk of colorectal cancer (pooled relative risk (estimated by odds ratio) for high vs. low category of coffee consumption (RR) = 0.72, 95% confidence interval (CI) 0.61-0.84); the findings were similar in population-based and hospital-based case-control studies. Five cohort studies did not support an association (pooled RR = 0.97, 95% CI 0.73-1.29). The combined results of all studies were driven largely by the case-control studies, which comprised 85 percent of the cases (RR = 0.76, 95% CI 0.66-0.89). The lower risk of colorectal cancer among substantial coffee drinkers was observed in studies from Asia, Northern and Southern Europe, and North America. The results of this meta-analysis indicate a lower risk of colorectal cancer associated with substantial consumption of coffee, but they are inconclusive because of inconsistencies between case-control and prospective studies, the lack of control for important covariates in many of the studies, and the possibility that individuals at high risk of colorectal cancer avoid coffee consumption. Several ongoing prospective cohort studies, based on extensive dietary questionnaires, may provide important new data to evaluate this hypothesis.

Measuring the burden of common morbidities: sampling disease experience versus continuous surveillance.

Year 1998
Morris SS. Santos CA. Barreto ML. Cousens SN. Strina A. Santos LM. Assis AM.
Food Consumption and Nutrition Division, International Food Policy Research Institute, Washington, DC, USA.
Longitudinal prevalence, the proportion of all days of observation that a given individual manifests symptoms of illness, is a measure of disease frequency that is easy to generate from daily morbidity data and has been shown to be strongly related to subsequent health outcome. It is hypothesized that this measure could be derived using a representative sample of days of observation rather than continuous surveillance. The authors use 1990-1991 data from a Brazilian supplementation trial comprising a year's daily records of the occurrence of diarrhea, fever, and cough in 906 children under 5 years of age to examine how many days of morbidity data need to be observed to rank subjects into quintiles of illness frequency. Systematic samples of the full data set, based on every 2nd, 3rd, 5th, 10th, 15th, 20th, and 30th day of data, are compared with the continuous record. For diarrhea and fever, estimates based on less than 72 days of observation result in over one fourth of individuals who should have been in the extreme quintiles of the morbidity distribution being misclassified, and over one fifth of all subjects appear (falsely) to suffer no morbidity. Estimates of longitudinal prevalence should be based on at least 72 days of observation.

Colorectal cancer: another complication of diabetes mellitus?

Year 1998
Will JC. Galuska DA. Vinicor F. Calle EE.
National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, GA 30341-3724, USA.
Delayed stool transit and other gastrointestinal abnormalities are commonly observed in persons with diabetes mellitus and are also known to be associated with colorectal cancer. Previous studies of the contribution of diabetes to colorectal cancer incidence and mortality have been limited by small sample sizes and failure to adjust for covariates. With more than 1 million respondents, the 1959-1972 Cancer Prevention Study provided a unique opportunity to explore whether persons with diabetes (n=15,487) were more likely to develop colorectal cancer during a 13-year follow-up period than were persons without diabetes (n=850,946). After adjustment for colorectal cancer risk factors, such as race, educational level, body mass index, smoking, alcohol use, dietary intake, aspirin use, physical activity, and family history of colorectal cancer, the incidence density ratio comparing colorectal cancer in those with diabetes and those without diabetes was 1.30 (95% confidence interval 1.03-1.65) for men and 1.16 (95% confidence interval 0.87-1.53) for women. However, diabetes was not associated with greater case fatality. Future studies should explore the possibility of a cancer-promoting gastrointestinal milieu, including delayed stool transit and elevated fecal bile acid concentrations, associated with hyperglycemia and diabetic neuropathy.