Abstract Reprints

1. Increasing incidence of diabetes in successive birth cohorts of Pima Indians

The incidence of non-insulin-dependent diabetes mellitus (NIDDM) has increased in recent years among Pima Indians, presumably because of lifestyle changes. Analysis of diabetes incidence by birth cohort could yield insight into the nature of the lifestyle factors that result in NIDDM. Incidence of NIDDM was determined among 2313 adult (age 20 yrs) Pima Indians participating in a longitudinal study from 1965 to 1994. Person-years were stratified by categories of age, sex, body mass index (BMI) and by eight 10-year birth cohorts (1895-1974). Because cohorts enter the study at different ages, conventional survival analysis and direct standardization of rates are of limited utility in comparing cohorts. Consequently, Poisson regression was used to calculate incidence rate ratios (IRR) and test hypotheses. NIDDM developed among 756 (33%) subjects during 20,615 person-years of follow-up. For most age groups, incidence rates were higher in more recent birth cohorts. Controlled for age and sex, there was a statistically significant cohort effect (2 = 16.1, df = 7, p = 0.02). This effect largely reflected a linear increase in incidence (2 = 10.1, df = 1, p < 0.01) with little evidence for an additional non-linear effect (2 = 6.0, df = 6, p = 0.42). The model indicated that incidence of NIDDM increased by 13% in each successive birth cohort (IRR = 1.13, 95% confidence interval (CI) 1.05-1.23). When controlled for BMI, which has also increased in recent cohorts, the IRR was reduced to 1.07 (95% CI, 0.99-1.16). During 30 years of observation, incidence of NIDDM in Pima Indians has increased with successive birth cohorts. This suggests that exposure to diabetogenic lifestyle factors may have increased progressively over the 80-year span of birth cohorts.

2. Cardiovascular disease in Navajo Indians with type 2 diabetes

Rates of both type 2 diabetes and cardiovascular disease have risen sharply in recent years among Navajo Indians, the largest reservation-based American Indian tribe, but the association between the two conditions is not entirely clear.

Rates of cardiovascular disease and some possible associations in several hundred diabetic and nondiabetic Navajos were estimated. Nearly one-third (30.9 percent) of those with diabetes had formal diagnoses of cardiovascular disease-25.3 percent had heart disease, 4.4 percent had cerebrovascular disease, and 4.1 percent had peripheral vascular disease. (The percentages exceed the total because some people had more than one diagnosis.) Age-adjusted rates were 5.2 times those of nondiabetics for heart disease, 10.2 times for cerebrovascular disease, and 6.8 times for peripheral vascular disease.

Accentuation of risk was most marked in young diabetics and in female diabetics. Hypertensive diabetics had a twofold increase in heart disease and more than a fivefold increase in cerebral and peripheral vascular disease over nonhypertensive diabetics. Age, blood pressure, cholesterol levels, and albumenuria were independent risk factors for cardiovascular disease. Triglyceride levels or body weight were not. Male sex and diabetes duration were independent risk factors for cerebral and peripheral vascular disease but not for heart disease.

In view of the impressive segregation of cardiovascular disease in the diabetic Navajo population, the prevention of diabetes through population-based health promotion seems basic to its containment. Over the short term, vigorous treatment of hypertension in subjects who are already diabetic is mandatory.

3. Coronary heart disease prevalence and its relation to risk factors in American Indians: the Strong Heart Study

Although coronary heart disease (CHD) is currently the leading cause of death among American Indians, information on the prevalence of CHD and its association with known cardiovascular risk factors is limited. The Strong Heart Study was initiated in 1988 to quantify cardiovascular disease and its risk factors among three geographically diverse groups of American Indians. Members of 13 Indian communities in Arizona, Oklahoma, and South and North Dakota between 45 and 74 years of age underwent a physical examination that included medical history; an electrocardiogram; anthropometric and blood pressure measurements; an oral glucose tolerance test; and measurements of fasting plasma lipoproteins, fibrinogen, insulin, hemoglobin A1c, and urinary albumin. Prevalence rates of definite myocardial infarction and definite CHD were higher in men than in women in all three centers (p < 0.0001) and higher in those with diabetes mellitus (p = 0.002 in men and p = 0.0003 in women). Diabetes was associated with relatively higher prevalence rates of myocardial infarction (diabetic : nondiabetic prevalence ratio = 3.8 vs. 1.9) and CHD (prevalence ratio = 4.6 vs. 1.8) in women than in men. Prevalence rates of heart disease were lowest in the communities in Arizona; prevalence rates were similar in Oklahoma and South Dakota/North Dakota and were two- to threefold higher than those in Arizona. By logistic regression, prevalent CHD among American Indians was significantly and independently related to age, diabetes, hypertension, albuminuria, percentage of body fat, smoking, high concentrations of plasma insulin, and low concentrations of high density lipoprotein cholesterol. In contrast to reports from other non-Indian populations, diabetes was the strongest risk factor. The lower prevalence of CHD among Indians in Arizona is distinctive in view of their higher rates of diabetes, obesity, hypertension, and albuminuria, but it may be partly related to their low frequency of smoking and their low concentrations of total and low density lipoprotein cholesterol. These findings from the initial Strong Heart Study examination emphasize the importance of diabetes and its associated variables as risk factors for CHD in Native American populations.

4. Incidence and predictors of diabetes in Japanese-American men: the Honolulu Heart Program

Reports on the incidence and predictors of diabetes in minority populations are infrequent. The 6-year cumulative incidence of diabetes between 1965 and 1974 was estimated among 7210 Japanese-American men aged 45 to 68 years who were enrolled in the Honolulu Heart Program and were free of clinically recognized diabetes at baseline. The incidence of "possible" diabetes (based on history, medication, or hospital diagnosis) was 12.8% and the incidence of "probable" diabetes (based on diabetic medication) was 5.7%. Estimates of incidence in subjects with a nonfasting glucose concentration less than 225 mg/dL 1 hour after a 50-g load were 9.7 and 4.0%, respectively. Multivariate adjusted odds ratio (ORs) for probable diabetes in all subjects comparing the upper quintile with the lower four quintiles combined for continuous variables indicated statistically significant direct associations with body mass index (OR, 1.69; 95% confidence interval (CI), 1.31 to 2.18), 1-hour postchallenge glucose level (OR, 5.79; 95% CI, 4.58 to 7.33), triglyceride levels (OR, 1.47; 95% CI, 1.14 to 1.91), systolic blood pressure (OR, 1.36; 95% CI, 1.05 to 1.76), and parental history of diabetes (OR, 1.73; 95% CI, 1.29 to 2.33), and an inverse association with physical activity (OR, 0.49; 95% CI, 0.34 to 0.72), using logistic regression models including these variables as well as age, subscapular/triceps skinfold ratio, and hematocrit simultaneously. Associations were similar but slightly weaker in men with glucose levels less than 225 mg/dL and in those who remained free of cardiovascular disease. When older men (55 to 68 years old) were compared with younger (45 to 54 years old) men, associations among the older group were stronger for body mass index, physical activity, and systolic blood pressure and they were weaker for glucose levels, triglyceride values, and parental diabetes. Results suggest that body mass index, physical inactivity, glucose level, and parental diabetes appear to be independent risk factors for diabetes, while triglyceride and systolic blood pressure levels may be markers for an adverse cardiovascular risk factor profile associated with diabetes and may reflect an insulin resistance syndrome.

5. Increase in incidence of insulin-dependent diabetes mellitus among children in Finland

Background. In Finland, the incidence of insulin-dependent diabetes mellitus (IDDM) in children aged <15 years is the highest in the world. The aim of this study was to determine the temporal variation in incidence and the age distribution at diagnosis of IDDM.

Subjects and Methods. Data on incidence of IDDM in Finland nationwide were obtained from two sources: the Central Drug Registry for the years 1965-1986 (6195 IDDM cases) and the prospective IDDM registry for the years 1987-1992 (2062 IDDM cases). The annual incidence rates were calculated per 100 000 population. The increase in incidence from 1965 to 1992 was estimated by fitting the linear regression with the annual incidence data.

Results. The overall incidence of IDDM between 1987 and 1992 was 36 per 100 000/year. During 1965-1992 the increase was almost linear. The regression-based change in incidence was 2.8% per year. In the 1970s the increase in incidence was steepest in 5-9 year olds and since the mid-1980s in those <5 years old at diagnosis.

Conclusions. The incidence of IDDM in Finnish children seems to increase further. During the last decades the increase in incidence has been almost linear with occasional peaks. The age-at-diagnosis of IDDM has been moving towards the younger ages, and differences in incidence between age groups have now almost disappeared among Finnish children aged 1-14 years.

6. International analysis of insulin-dependent diabetes mellitus mortality: a preventable mortality perspective. The Diabetes Epidemiology Research International (DERl) Study

Differential survival associated with insulin-dependent diabetes mellitus (IDDM) was evaluated in a cross-country study using four population-based IDDM cohorts from Japan (n = 1,374), Israel (n = 610), Allegheny County, Pennsylvania (n = 995), and Finland (n = 5,144). For the purpose of this cross-country comparison, the Allegheny County cohort was taken to be representative of the United States. The mortality status as of January 1, 1990, was determined for all individuals who were diagnosed with diabetes at the age of less than 18 years between 1965 and 1979 and who were taking insulin at the time of hospital discharge. The results showed that the mortality experience for IDDM individuals in Japan and the United States was much worse than that in Finland and Israel. The age-adjusted mortality rates (per 100,000 person-years) for the four cohorts were 760 (Japan), 158 (Israel), 408 (Allegheny County), and 250 (Finland). By using the mortality data from Allegheny County, Pennsylvania, to extrapolate to the US IDDM mortality experience, the authors estimated 2,396 deaths among individuals with IDDM in the United States. It was calculated that 1,261 (53%) of these deaths would not have occurred in the United States given Finland's mortality rates. It is critical to determine why individuals with IDDM in the United States have a poorer outcome.

7. Mortality attributable to tobacco use in Canada and its regions, 1991

Using the data from the 1991 General Social Survey, Canadian Mortality Database, the 1991 Annual Report of Fire Losses in Canada and previously published estimates of lung cancer deaths attributable to passive smoking, the Smoking-Attributable Mortality, Morbidity and Economic Cost method was used to estimate national and regional smoking-attributable mortality (SAM) for 1991. The results indicate that 27,867 men and 13,541 women died as a result of smoking, including 171 children under the age of one. This total of 41,408 deaths represents an increase of 3,051 deaths since 1989, with women accounting for 2,721 of these increased deaths. The increase in female mortality is almost entirely due to adult diseases, equally divided between neoplasms (990), cardiovascular diseases (927) and respiratory diseases (821).

8. Parental smoking and risk of otitis media in pre-school children

This study was designed to estimate the role of passive smoking in the occurrence of repeated acute otitis media among pre-school children. The parents of 918 children, who were part of a birth cohort, were interviewed when the children were four years old. The frequency of episodes of otitis media was determined, as well as the age at which the first such episode occurred. Information about smoking in the home had been obtained when the child was two weeks old and again at 6, 12, 18 and 40 months of age. In all, 119 mothers (13%) smoked at least 20 cigarettes a day on average, 205 (22%) smoked less than 20, and 593 (65%) stated that they did not smoke at all. Children of mothers who smoked 20 or more cigarettes per day were at significantly increased risk of having four or more episodes of acute otitis media; relative risk (RR) = 1.8, 95% confidence interval (CI) = 1.1-3.0 and of having their first episode earlier in life RR = 1.3 (CI = 1.0-1.8), after adjustment. Moreover, the risk of recurrent (³ four episodes) otitis media increases with the amount of cigarettes smoked. These findings clearly strengthen the case for persuading parents not to smoke in the presence of young children.

9. The impending Canadian prostate cancer epidemic

Purpose: To model and forecast prostate cancer incidence and mortality in Canada to the year 2016.

Methods: Bivariate multiplicative models of prostate cancer incidence and mortality for Canadian men aged 45 years or older, linear in time and Weibull in age, were fitted using weighted non-linear regression.

Results: The number of incident cases of prostate cancer is forecast to increase from 11,355, observed in 1990, to 26,900 by the year 2010 and to 35,200 by the year 2016. The number of deaths are estimated to climb from 3,424, observed in 1991, to an estimated 6,300 by the year 2010, and to 7,800 by the year 2016.

Conclusions: The dramatic increase in prostate cancer rates with age, coupled with the expected large increase in the elderly Canadian male population and steadily increasing prostate cancer incidence rates will produce very large increases in the number of men who will have prostate cancer over the next 20 years. This has important implications for health care delivery in the future.

10. Hereditary aspects of prostate cancer

Objective: To review current literature on the hereditary aspects of prostate cancer and to evaluate the importance of family history in history taking and screening for prostate cancer.

Data sources: MEDLlNE was searched for articles in English or French published between Jan. 1, 1956, and Oct. 31, 1994, with the use of MeSH headings "prostatic neoplasms," "genetics" and "chromosomes." Additional references were selected from the bibliographies of articles found during the search.

Study selection: Case-control studies involving the incidence of prostate cancer and relative risk (RR) of such cancer in the families of men with this disease, compared with a control group, were included. Only studies in which prostate cancer was diagnosed on the basis of histologic tests were included. Animal investigations were excluded.

Data extraction: Ten case-control studies were evaluated critically in terms of design, case and control groups, the size of the samples and statistical results. The incidence of prostate cancer in the families of cases, compared with that in the families of controls, and differences in RR were reviewed.

Data synthesis: The lifetime risk of prostate cancer is 9.5% and of death from prostate cancer is 2.9% for a man 50 years of age. For first-degree male relatives of men with prostate cancer, the calculated RR ranges from 1.7 to 8.73. "Hereditary" prostate cancer is a term applied to a specific subset of patients with prostate cancer. This form of prostate cancer is transmitted by a rare, autosomal, dominant allele with high penetrance; it accounts for an estimated 43% of early-onset disease (affecting men less than 55 years of age) but only 9% of all prostate cancer in men up to 85 years of age. A greater number of affected family members and early onset among family members are the most significant predictors of risk.

Conclusions: Recent confirmation of the familial clustering and Mendelian inheritance patterns of some prostate cancer has important implications. It increases the potential for directed research into the causes of prostate cancer and for refinements in the current screening practices to detect this common disease. Manoeuvres to detect prostate cancer should be started earlier among men with one or more first-degree relatives with the disease than among other men.

11. Physical activity, medical history, and risk of testicular cancer (Albert and British Columbia, Canada)

In order to evaluate risk factors for germ cell cancers, we conducted a case-control study of 510 men with testicular cancer aged 15 to 79 years and 996 randomly selected age-matched controls in the provinces of British Columbia and Alberta, Canada. Subjects completed a mailed questionnaire providing data on education level, ethnic origin, medical history, smoking, occupation, and recreational and sports activity. The response rate among cases was 80.3 percent and among controls was 68.1 percent. After controlling for age and ethnic origin, undescended testis was associated positively with risk of testicular cancer (odds ratio [OR] = 3.5; 95 percent confidence interval [CI] = 2.2-5.7) as was inguinal hernia requiring surgery (OR = 2.0, CI =1.3-2.9), and hydrocoele (OR = 2.6, CI = 1.4-5.1). Risk of testicular cancer increased with height, with subjects taller than 180 cm having a significantly increased risk compared with those 174 cm or less (OR = 1.5, CI = 1.1-2.1). A moderate to high level of recreational activity level was associated inversely with testicular cancer risk (OR = 0.6, CI = 0.5-0.8).

12. Rates of transcervical and pertrochanteric hip fractures in the province of Quebec, Canada, 1981-1992

Two distinct subtypes of hip fracture, transcervical and pertrochanteric, can be distinguished on the basis of the anatomical location of the injury. While the epidemiology of hip fractures has been well described, typically, little or no distinction is made between these subtypes. The objective of this study was to compare and contrast age- and sex-specific rates of transcervical and pertrochanteric fractures in Quebec, Canada. The data for this study were obtained from a database containing records of all persons discharged from all hospitals in Quebec from 1981 to 1992. Rates of hip fracture were calculated by using the population aged 50 years and older as the denominator, and changes in rates over time were assessed using Poisson regression. There were no statistically significant trends in the changes in rates over time (i.e., 95 percent confidence intervals overlapped the null value). Among women below age 70 years, transcervical fractures were more common, whereas among older women, pertrochanteric fractures predominated. Among men, pertrochanteric fractures predominated at all ages. There was a marked seasonal variation in the occurrence of all hip fractures combined: Compared with the summer months, the relative risk of all hip fractures during the winter was 1.32 (95 percent confidence interval 1.28-1.36). The results of this study indicate that the two subtypes of hip fracture, transcervical and pertrochanteric, have different patterns of occurrence, suggesting different risk factor profiles. Clearly, a multidisciplinary research approach is needed before it will be possible to untangle the complex relation between the metabolic processes occurring at the level of the individual and the distribution of the disease in the population.

13. Accounting for cluster randomization: a review of primary prevention trials, 1990 through 1993

Objectives. This methodological review aims to determine the extent to which design and analysis aspects of cluster randomization have been appropriately dealt with in reports of primary prevention trials.

Methods. All reports of primary prevention trials using cluster randomization that were published from 1990 to 1993 in the American Journal of Public Health and Preventive Medicine were identified. Each article was examined to determine whether cluster randomization was taken into account in the design and statistical analysis.

Results. Of the 21 articles, only 4 (19%) included sample size calculations or discussions of power that allowed for clustering, while 12 (57%) took clustering into account in the statistical analysis.

Conclusions. Design and analysis issues associated with cluster randomization are not recognized widely enough. Reports of cluster randomized trials should include sample size calculations and statistical analyses that take clustering into account, estimates of design effects to help others planning trials, and a table showing the baseline distribution of important characteristics by intervention group, including the number of clusters and average cluster size for each group.

14. Pediatric mortality and hospital use in Canada and the United States, 1971 through 1987

Since 1971 pediatric mortality rates have decreased markedly but differently in Canada and the United States. These trends were examined in light of changes in hospital use and health care financing. Annual mortality and hospital use rates for children aged 14 years and younger were calculated. Between 1971 and 1987, all-cause mortality in Canada fell from 165 to 74 per 100 000; the American rate fell from 172 to 96 per 100 000. American hospitalization rates remained essentially constant until 1983 and then fell by 27.5%, while Canadian hospitalization rates declined throughout. In 1987 Canadian children had higher hospitalization rates, while American children had higher mortality rates. These differences may be associated with differences in health financing; the adoption of US prospective payment systems was temporally coincident with sharp declines in hospitalization rates for American children.

15. Socioeconomic indicators and mortality from coronary heart disease and cancer: a 22-year follow-up of middle-aged men

Objectives. Data from the Western Collaborative Group Study were used to determine the extent to which the inverse association between socioeconomic status (SES) and mortality can be explained by risk factors for major causes of mortality.

Methods. The relation of education and income to subsequent mortality was studied in 3154 employed, middle-aged men over 22 years of follow-up. Results. Over the follow-up period, 584 (18.5%) men died, 214 (6.8%) from coronary heart disease and 70 (2.2%) from lung cancer. A significant inverse association with systolic blood pressure, serum cholesterol, and smoking was found only for education. For education, adjustment for risk factors reduced the relative risk for coronary heart disease mortality from 1.80 (95% confidence interval = 1.33, 2.44) to 1.54 (1.13, 2.09), for lung cancer mortality from 1.60 (0.95, 2.70) to 1.38 (0.81, 2.34), and for all-cause mortality from 1.49 (1.09, 1.13) to 1.33 (1.12, 1.60). For income, adjustment for risk factors did not change relative risk for mortality from coronary heart disease (1.27 [0.97, 1.66]) and all causes (1.21 [1.03, 1.43]), but it did increase the relative risk for lung cancer mortality from 1.68 (1.05, 2.68) to 1.83 (1.13, 2.96).

Conclusions. In middle-aged, employed men, the association between SES and mortality is partially but not completely accounted for by major risk factors for mortality.

16. Poverty and death in the United States-1973 and 1991

We conducted a survival analysis to determine the effect of poverty on mortality in a national sample of blacks and whites 25-74 years of age (in the First National Health and Nutrition Examination Survey and National Health Examination Follow-up Survey). We estimated the proportion of mortality associated with poverty from 1973 through 1984 and in 1991 by calculating the population attributable risk. We assessed confounding by major known risk factors, such as smoking, serum total cholesterol, and inactivity. In 1973, 16.1% of U.S. mortality among black and white persons 25-74 years of age was attributable to poverty; in 1991, the proportion increased to 17.7%. In 1991, the population attributable risk of poverty on mortality was lowest for white women. 1.7 times higher for white men, 2.6 times higher for black women, and 3.6 times higher for black men. Potential confounders explained 40% of the effect of poverty on mortality among women.

The proportion of mortality attributable to poverty among U.S. black and white adults has increased in recent decades and is comparable to that attributable to cigarette smoking. The effect of poverty on mortality must be explained by conditions other than commonly recognized risk factors.

17. The sex ratio of mortality and its secular trends

Background. The mortality level from all causes is different between populations and it has decreased for both men and women in most countries in the last decades. However, there is a difference in the male/female sex ratio of mortality between populations and its time trends and the reasons for these differences remain unclear.

Methods. The sex ratio of all-cause mortality and the main causes of death, i.e. total cardiovascular disease and cancer, for 30 populations in 1988 (mean of 1987-1989), and the time trends of the sex ratio for 27 populations are analysed.

Results. Large differences in the sex ratio of mortality exist among the studied populations. The sex ratio of all-cause, total cardiovascular and cancer mortality markedly increased in most countries during recent decades.

Conclusions. The sex ratio of all-cause mortality and its time trends correlated significantly and positively with the sex ratio of mortality and its time trends from total cardiovascular disease and cancer. The differences of the sex ratio of mortality and their time trends between populations cannot be explained by genetic factors. They could be attributed to differences in life style. A different exposure and different reaction to the risk factors of cardiovascular diseases and cancer, e.g. saturated fat intake, alcohol intake and smoking habits, between men and women are considered to be the main causes for these differences in the sex ratio of mortality.

18. Epidemiology for prevention

This paper illustrates the basis of, expectations for and evaluation of prevention from an epidemiological perspective. Specifically, the extent to which epidemiologists could and should be involved in designing and evaluating public health interventions is addressed. Changes in the view on disease causation and epidemiology's role for the future of public health are discussed. Examples from cardiovascular epidemiology are used to illustrate the ever more complex, but still incomplete, knowledge on which prevention is based. A few current examples illustrate paradoxes where health information must balance academic discord. Methodological problems in the evaluation of intervention studies often fail to live up to the expectations of prevention. Outcome indicators of preventive projects must be developed and traditional appraisals of effects be supplemented with process analyses using both quantitative and qualitative methods. Social conditions for prevention are discussed and areas for further research are suggested.