Annals of Epidemiology

We conducted a retrospective cohort mortality study to determine the postservice suicide risk of recent wartime veterans comparing them with the US general population as well as comparing deployed veterans to nondeployed veterans. Veterans were identified from the Defense Manpower Data Center records, and deployment to Iraq or Afghanistan war zone was determined from the Contingency Tracking System. Vital status of 317,581 deployed and 964,493 nondeployed veterans was followed from the time of discharge to December 31, 2009. Underlying causes of death were obtained from the National Death Index Plus. Based on 9353 deaths (deployed, 1650; nondeployed, 7703), of which 1868 were suicide deaths (351; 1517), both veteran cohorts had 24% to 25% lower mortality risk from all causes combined but had 41% to 61% higher risk of suicide relative to the US general population. However, the suicide risk was not associated with a history of deployment to the war zone. After controlling for age, sex, race, marital status, branch of service, and rank, deployed veterans showed a lower risk of suicide compared with nondeployed veterans (hazard ratio, 0.84; 95% confidence interval, 0.75-0.95). Multiple deployments were not associated with the excess suicide risk among deployed veterans (hazard ratio, 1.00; 95% confidence interval, 0.79-1.28). Veterans exhibit significantly higher suicide risk compared with the US general population. However, deployment to the Iraq or Afghanistan war, by itself, was not associated with the excess suicide risk. Published by Elsevier Inc.

To examine age-related effects of smoking on lung cancer mortality in a large population-based case-control study that was incorporated into a nationwide retrospective survey of mortality in China. Two data sets were pooled for this analysis: national mortality survey data and data from a nationwide case-control comparison. These pooled data were used to calculate age-specific lung cancer death rates by smoking status and expected years of lost life expectancy for each smoking-associated death. There was a significant excess of deaths (54% of urban deaths, 51% of rural deaths) at 35 to 69 years of age among male smokers and the average loss of life expectancy per smoking-associated death at these ages was 18.3 years. For women ages 35 to 69 years of age, 29% and 11% excess lung cancer mortality was observed in urban and rural smokers, respectively, with an average loss of life expectancy per smoking-associated death of 21.3 years. Tobacco smoking was associated with a large number of deaths from lung cancer. Utilizing information from a population-based retrospective mortality survey for conducting an analytic epidemiological study of main determinants of disease is feasible and can generate important results.

As a nation, we must accept and appreciate the fact of the racial and cultural diversity that characterizes America. Such diversity dictates that fundamental approaches and solutions to social problems, including health, need to be specific to local areas. From addressing health problems, the least controversial of social challenges, a degree of mutual respect can evolve that permits other issues to be addressed as well. Resolving health problems will require a coordinated effort of local, city, state and federal resources, both medical and non-medical. In addressing such problems, three important principles need to be embraced: 1) the development of a surveillance system to measure the problem and assess progress in terms of ultimate outcomes; 2) the continuing use of surveillance data to assess and modify strategy and to allocate resources as needed; and 3) the need in health programs in particular to assign far higher priority to "consumer satisfaction."

Although rarely available, detailed analyses of attrition in psychiatric surveys are important because surveys of this type might be more vulnerable to follow-up losses. In this report the demographic characteristics, as well as history of alcohol problems and psychiatric disorders of responders were compared to nonresponders in an 11-year follow-up study. Data revealed few differences between responders and nonresponders. Men, those less educated, and low users of medical care were more likely to be nonresponders, as were those reporting driving trouble when drinking or a history of barbiturate abuse or dependence. A history of other psychiatric disorders was not associated with nonresponse. Refusal conversion did not change the findings; those who were converted (25% of initial refusals) had demographic characteristics, symptoms of alcohol abuse, and psychiatric histories comparable to those who resisted conversion. These findings suggest that efforts to convert refusals to responders might not be necessary. The results also support community psychiatric research by providing evidence that those with a history of psychiatric disorder are not more difficult to recruit than their unaffected counterparts.

To investigate the impact of attrition on prevalence and associations between variables across four waves of a longitudinal study. Prevalence of socio-demographic and health characteristics were estimated for respondents to one, two, three or all four waves of the Australian Longitudinal Study of Women's Health cohort born between 1973 and 1978. Associations with self-rated General Health (GH) and Mental Health (MH) were compared using fixed effects in separate mixed models for respondents to at least one wave, at least two waves, at least three waves, or four waves of the longitudinal study. 14,247 women aged 18-23 years responded to Wave 1 in 1996. Respondents to all waves were more educated, and less likely to be stressed about money, to be smokers and to have children than respondents to some waves. Across all models, better GH was consistently associated with more education, no monetary stress, being married, having children, fewer visits to the doctor and not smoking. Similar results were obtained for MH. Although the potential for bias due to attrition must be considered, the current paper contributes to the growing body of evidence that suggests such biases are insufficient to preclude meaningful longitudinal analyses.

To assess ethnic differences in the risk of post-traumatic stress disorder (PTSD) after a disaster, and to assess the factors that may explain these differences. We used data from a representative survey of the New York City metropolitan area (n=2,616) conducted 6 months after September 11, 2001. Linear models were fit to assess differences in the prevalence of PTSD between different groups of Hispanics and non-Hispanics and to evaluate potential explanatory variables. Hispanics of Dominican or Puerto Rican origin (14.3% and 13.2%, respectively) were more likely than other Hispanics (6.1%) and non-Hispanics (5.2%) to report symptoms consistent with probable PTSD after the September 11 terrorist attacks. Dominicans and Puerto Ricans were more likely than persons of other races/ethnicities to have lower incomes, be younger, have lower social support, have had greater exposure to the September 11 attacks, and to have experienced a peri-event panic attack upon hearing of the September 11 attacks; these variables accounted for 60% to 74% of the observed higher prevalence of probable PTSD in these groups. Socio-economic position, event exposures, social support, and peri-event emotional reactions may help explain differences in PTSD risk after disaster between Hispanic subgroups and non-Hispanics.

This study re-examined the differential effect of socioeconomic status on the survival of women with breast cancer in Canada and the United States. Ontario and California cancer registries provided 1,913 cases from urban and rural places. Stage-adjusted cohorts (1998-2000) were followed until 2006. Socioeconomic data were taken from population censuses. SES-survival associations were observed in California, but not in Ontario, and Canadian survival advantages in low-income areas were replicated. A better controlled and updated comparison reaffirmed the equity advantage of Canadian health care.

To assess the relationships between lipid levels and sexual maturity, independently of age-related differences, and to investigate possible differences related to sexual maturity across the percentiles of the lipid distributions. Fasting serum total cholesterol and triglyceride concentrations were measured in 6577 boys and 6605 girls, aged from 10 to 13 years, with different Tanner stages. The total cholesterol and triglyceride mean and percentile levels were estimated according to age and Tanner stage by ordinary least squares and percentile regression models, separately in both sexes. In boys and girls, total cholesterol levels were significantly associated with pubertal stage after controlling for age. At age 12, the estimated mean levels in boys varied from 4.82 mmol/L for Tanner 1 to 4.41 for Tanner 5. The corresponding values were 5.05 and 4.62 mmol/L in girls, for whom the association with maturity was stronger in the upper than in the lower percentiles (p < 0.0001); between the extreme Tanner stages, the 95th percentiles of total cholesterol differed by 0.80 mmol/L, in comparison to 0.19 mmol/L for the 5th percentiles. Therefore, 1. 8% of girls and 0.7% of boys were classified differently whether Tanner stage was used or not to assess hypercholesterolemia (concentrations in the upper 5% of the distributions). Triglycerides were positively related to sexual maturity independently of age, but the discrepancies between classifications were lower; 1.1% in girls and 0.4% in boys. Our findings emphasize the importance of sexual maturity, even for a given age, for interpreting lipid levels in children.

Earthquakes pose a persistent but unpredictable health threat. Although knowledge of geologic earthquake hazards for buildings has increased, spatial relations between injuries and seismic activity have not been explained. Fatal and hospital-admitted earthquake injuries due to the 1994 Northridge Earthquake were identified. Geographical Information Systems software was used to map all injury locations. Injuries were analyzed with regard to distance from the earthquake epicenter, the Modified Mercalli Intensity Index, peak ground acceleration, and proportion of damaged residential buildings. Injury severity was inversely related to distance from the epicenter and increased with increasing ground motion and building damage. However, injury incidence and severity were not completely predicted by seismic hazard and building damage, and injuries of all severities occurred in a large geographic area. Average distance to the epicenter was smallest for injuries related to falling building parts and largest for cutting/piercing injuries and falls. The injuries from the Northridge Earthquake extended beyond the areas of highest environmental activity. Factors such as age and activity during the earthquake may be equally important in predicting injury from earthquakes as seismic features.

To determine the effects of average alcohol consumption and changes in alcohol intake on the insulin resistance syndrome parameters in a 3-year follow-up study. Longitudinal study of 1856 and 1529 alcohol drinking men and women in the French DESIR study (Data from an Epidemiological Study on the Insulin Resistance syndrome), aged 30 to 64 years. In men, fasting glucose, body mass index, waist circumference, systolic blood pressure, and HDL-cholesterol were positively associated with average alcohol consumption while there was no association with insulin or triglycerides concentrations. A change in alcohol intake was positively associated with HDL-cholesterol concentration and systolic blood pressure at follow-up. These effects of alcohol could not be attributed specifically to the intake of wine. In women, while the alcohol HDL-cholesterol relation was similar to that found in the men, the only significant effect of average alcohol intake was an increase in systolic blood pressure, with a spurious decrease in blood pressure related to a 3-year increase in alcohol intake. Alcohol only provided a beneficial effect on HDL-cholesterol. The beneficial effect seen by other authors of moderate alcohol drinking on diabetes and cardiovascular risk may be due to effects on parameters other than those included in the current definitions of the insulin resistance syndrome.

To describe the distribution and to assess the association of serum total homocysteine (tHcy) concentration with variables associated with insulin resistance syndrome in adolescent boys and girls. In the Third National Health and Nutrition Examination Survey, adolescents aged 12 to 16 years (n = 941) had parental medical history ascertained and glycated hemoglobin, systolic blood pressure, body mass index (BMI), body fat distribution, HDL cholesterol, and serum tHcy concentrations were measured. Cumulative distribution of serum tHcy in boys was shifted to the right compared with that in girls. A parental history of high blood pressure or stroke before age 50 was significantly positively associated with the subjects' log serum tHcy after adjustment for confounders (beta 0.156, p = 0.003). Log serum tHcy was significantly positively associated with systolic blood pressure in boys after adjustment for confounders (beta = 0.21, p = 0.03). Log serum tHcy was not significantly associated with glycated hemoglobin percent or most other risk factors other than age. Log serum tHcy concentration showed borderline significant (r = -0.15, p = 0.044) inverse association with BMI in girls. tHcy was associated with parental history of high blood pressure or stroke before age 50 and with systolic blood pressure in adolescent boys.

The Multiple Risk Factor Intervention Trial (MRFIT), a randomized clinical trial for the primary prevention of coronary heart disease, enrolled 12,866 men (including 8194 cigarette smokers) aged 35-57 years at 22 clinical centers across the United States. Participants were randomized either to special intervention (SI), which included an intensive smoking cessation program, or to usual care (UC). After 16 years of follow-up, lung cancer mortality rates were higher in the SI than in the UC group. Since rates of smoking cessation in SI were higher than those for UC for the 6 years of the trial, and since risk of lung cancer mortality is known to decline with smoking cessation, these results were unexpected. The purpose of the present study was to investigate hypotheses that could explain the higher observed lung cancer mortality rates in the SI as compared with the UC group. Analytic methods were employed to determine whether SI and UC participants differed either in baseline characteristics or in characteristics that changed during the trial and to determine whether these differences could explain the higher rates of lung cancer mortality among SI as compared to UC participants. Rates of mortality from coronary heart (CHD) were examined to explore the possibility that prevention of CHD death may have contributed to greater mortality due to lung cancer in the SI group. From randomization through December 1990, 135 SI and 117 UC participants died from lung cancer. The relative difference between the SI and U groups adjusted for age and number of cigarettes smoked per day, was 1.17 (95% CI:0.92-1.51). The greatest difference between the SI and UC groups in lung cancer mortality rates occurred among the heaviest smokers at baseline who did not achieve sustained smoking cessation by year 2. In this group the rates of death from CHD were approximately the same among the SI and UC subjects. No differences in baseline characteristics were found between SI and UC smokers who did not achieve sustained cessation by year 2, and there were no differences in follow-up characteristics between the two study groups that could explain the difference in lung cancer mortality. None of the hypotheses proposed to explain the unexpected higher rates of lung cancer mortality among SI as compared with UC subjects were sustained by the data. Thus we conclude that the difference observed is due to chance, and that a longer period of sustained smoking cessation plus follow-up is necessary to detect a reduction in lung cancer mortality as a result of smoking cessation intervention in a randomized clinical trial.

Medical late effects among cancer survivors may result in impairments that limit physical performance and activities necessary for normal participation in daily life. The aim of this analysis was to estimate the prevalence of physical performance limitations and participation restrictions among recent (< 5 years since diagnosis), and long-term (> or = 5 years) cancer survivors. Data from the 1999-2002 National Health and Nutrition Examination Survey were analyzed to compare the proportions of physical performance limitations and participation restrictions among 279 recent and 434 long-term cancer survivors, and among 9370 persons with no reported cancer history. Multivariable logistic regression was used to calculate adjusted prevalence odds ratios. Physical performance limitations were 1.5-1.8 times (53% versus 21%) and participation restrictions 1.4-1.6 times (31% versus 13%) more prevalent in cancer survivors than in those with no cancer history. Recent cancer history was associated with increased prevalence of physical performance limitation and participation restriction, particularly in survivors aged 40-49 years. Over half of the cancer survivors reported physical performance limitations; one third reported participation restrictions. Deficits were present many years following cancer diagnosis, even among survivors who were not elderly. Cancer survivors may benefit from evaluation for rehabilitation services long after treatment for their original disease.

Using three-dimensional (3D) scanning data along with other existing subject's medical profiles to search for better anthropometric markers in association with type 2 Diabetes Mellitus (DM). In this cross-sectional study with 6007 subjects undergoing health examination in a period of 3 years, the authors adopted data from 3D scanning with hundreds of body measures and conducted factor analysis to search for practical indicators better associated with type 2 DM. A multiple logistic regression model was used to analyze strength of association between indicators and presence of type 2 DM. The trunk component derived from factor analysis was positively associated with type 2 DM, regardless of obesity status. However, lower limbs component was found to be negatively associated with type 2 DM in the same stratifications. Waist was thus found as the strongest indicator among practical measures of trunk component. The strength of association between thigh and type 2 DM was found to be the highest among practical measures of lower limbs component. A marker from taking the ratio of waist to thigh (WTR) was derived from the approach which was found to be the best indicator for association with type 2 DM while comparing to body-mass index, waist circumference, or waist-hip ratio. This study offers a low-cost, noninvasive practical marker that is better associated with the presence of type 2 DM. WTR, with further study, may be used in clinical practice, epidemiological study, and preventive medicine in the future.

To investigate the relationship between serum sialic acid level and risk of coronary heart disease (CHD) and stroke in men and women without diagnosed cardiovascular disease. A prospective case-cohort study over the period 1981 to 1998 involving 151 CHD cases, 87 stroke cases, and a random sub-cohort of 340 was used. Sialic acid levels were determined by enzymatic method from frozen serum. Cox proportional hazards regression was used to estimate the relative risks of CHD and stroke for sialic acid tertiles and for continuous sialic acid level after adjustment for age, blood pressure, body mass index, cholesterol, triglycerides, diabetes, and smoking. The multivariate-adjusted relative risk of CHD associated with a 25 mg/dl increase in sialic acid was 1.22 (95% CI: 1.02-1.45) overall, 1.40 (95% CI: 1.11-1.76) in women, and 1.06 (95% CI: 0.82-1.37) in men. The overall relative risk for stroke was 1.13 (95% CI: 0.87-1.46) and for CHD and stroke combined it was 1.17 (95% CI: 0.99-1.37) Serum sialic acid may be a long-term predictor of CHD events in adults (especially women) who are currently clinically free of cardiovascular disease. Further studies are needed to determine whether this association can be explained by sialic acid being a marker of accelerated atherosclerosis or inflammation.

Specific events in the mother's reproductive history and certain birth characteristics have been associated with childhood leukemia. Few studies have explored these associations specifically in infants. The Children's Cancer Group (CCG) conducted three separate case-control studies of childhood leukemia that involved similar methodologies and data collection. Data from interviews of the mothers of a total of 303 children diagnosed with leukemia at 1 year of age or younger and their matched controls (n = 468) were available from the three studies. These data included maternal reproductive history (stillbirths, abortions, and miscarriages) and certain birth characteristics of the index child. Compared with controls, cases were significantly more likely to be female (P < 0.01) and were more often heavier at birth (particularly cases diagnosed after 6 months of age (odds ratio, 4.18; 95% confidence interval, 1.75-10.02)). Overall, there were no statistically significant differences between cases and controls in regard to maternal report of any type of previous fetal loss. Finally, being a later-born child was associated with an increased risk of acute myeloid leukemia but not of acute lymphoblastic leukemia. The relationships among birthweight, prior fetal loss, and risk of infant leukemia appear to be complex. Further studies of infant leukemia that incorporate molecular as well as epidemiologic data may help to elucidate these differences.

The Women's Health Initiative (WHI) is the largest research program ever initiated in the United States with a focus on diet and health. Therefore, it is important to understand and document the measurement characteristics of the key dietary assessment instrument: the WHI food frequency questionnaire (FFQ). Data are from 113 women screened for participation in the WHI in 1995. We assessed bias and precision of the FFQ by comparing the intake of 30 nutrients estimated from the FFQ with means from four 24-hour dietary recalls and a 4-day food record. For most nutrients, means estimated by the FFQ were within 10% of the records or recalls. Precision, defined as the correlation between the FFQ and the records and recalls, was similar to other FFQs. Energy adjusted correlation coefficients ranged from 0.2 (vitamin B12) to 0.7 (magnesium) with a mean of 0.5. The correlation for percentage energy from fat (a key measure in WHI) was 0.6. Vitamin supplement use was common. For example, almost half of total vitamin E intake was obtained from supplements. Including supplemental vitamins and minerals increased micronutrient correlation coefficients, which ranged from 0.2 (thiamin) to 0.8 (vitamin E) with a mean of 0.6. The WHI FFQ produced nutrient estimate, that were similar to those obtained from short-term dietary recall and recording methods. Comparison of WHI FFQ nutrient intake measures to independent and unbiased measures, such as doubly labeled water estimates of energy expenditure, are needed to help address the validity of the FFQ in this population.

To determine whether perinatal exposure to "background" levels of polychlorinated biphenyls (PCBs) or dichlorodiphenyldichloroethene (DDE) affected child development, we followed a cohort of children through the age of 2 years; 676 were tested at 18 months and 670 at 24 months with the Bayley Scales. Transplacental and cumulative exposures from breast milk to both chemicals were estimated for each child from analyses of breast milk and other samples. Bayley testing was done without knowing the chemical levels. At 18 and 24 months, adjusted scores on the psychomotor scales were 4 to 9 points lower among children in the top fifth percentile of transplacental PCB exposure, significantly so at 24 months. There were no consistent effects apparent from exposure to PCBs through breast milk, nor from DDE exposure. We conclude that there is a small delay in motor maturation attributable to transplacental exposure to PCBs that is still detectable at 24 months. There is no evidence of an effect from the larger but later exposure through breast milk, although effects may emerge as the children age.

Although eliminating health disparities by race, ethnicity, and socioeconomic status (SES) is a top public health priority internationally and in the United States, weight-related racial/ethnic and SES disparities persist among adults and children in the United States. Few studies have examined how these disparities have changed over time; these studies are limited by the reliance on rate differences or ratios to measure disparities. We sought to advance existing research by using a set of disparity metrics on both the absolute and relative scales to examine trends in childhood obesity disparities over time. Data from 7066 children, ages 2 to 18 years, in the National Health and Nutrition Examination Surveys were used to explore trends in racial/ethnic and SES disparities in pediatric obesity from 2001 to 2010 using a set of different disparity metrics. Racial/ethnic and SES-related disparities in pediatric obesity did not change significantly from 2001 to 2010 and remain significant. Disparities in obesity have not improved during the past decade. The use of different disparity metrics may lead to different conclusions with respect to how disparities have changed over time, highlighting the need to evaluate disparities using a variety of metrics.

Purpose: In Utah, the prevalence of unhealthy male risk behaviors are lower than in most other male populations, whereas women experience higher mortality risk because of higher fertility rates. Therefore, we hypothesize that the Utah sex differential in mortality would be small and less than in Sweden and Denmark. Methods: Life tables from Utah, Denmark, and Sweden were used to calculate cohort life expectancies for men and women born in 1850-1910. Results: The sex difference in cohort life expectancy was similar or larger in Utah when compared with Denmark and Sweden. The change over time in the sex differences in cohort life expectancy was approximately 2 years smaller for active Mormons in Utah than for other groups suggesting lifestyle as an important component for the overall change seen in cohort life expectancy. Sex differences in cohort life expectancy at the age of 50 years were similar for individuals actively affiliated with the Church of Jesus Christ of Latter-day Saints and for Denmark and Sweden. Conclusions: The hypothesis that a smaller sex difference in cohort life expectancies in Utah would be detected in relation to Denmark and Sweden was not supported. In Utah, the male-female differences in life expectancy remain substantial pointing toward biological mechanisms or other unmeasured risk factors.

Previous reports have described favorable changes in the relationship between systolic blood pressure and age in recent birth cohorts. The obesity epidemic might threaten that pattern. To update analyses of differences between birth cohorts in the relationship between systolic blood pressure and age and to determine whether increases in obesity have had adverse effects. We examined the systolic blood pressure distributions across birth cohorts born between 1890 and 1990 in 68,070 participants, aged 18-74 years, in the National Health (and Nutrition) Examination Surveys between 1960 and 2008. We postulated that age-adjusted 10th, 25th, 50th, 75th, and 90th percentiles of systolic blood pressure had decreased in more recent versus earlier cohorts, and that this pattern had slowed or reversed recently due, at least in part, to obesity. After adjusting for gender, race, age and age(2), the 10th, 25th, 50th, 75th, and 90th percentiles of systolic blood pressure were 1.1, 1.4, 1.9, 2.5, and 3.4 mmHg lower for each decade more recently born (all P < .0001). Quadratic terms for birth cohort were positive and significant (P < .001) across all percentiles, consistent with a decelerating cohort effect. Mediation of this deceleration was observed for body mass index ranging from 20.4% to 44.3% (P < .01 at all percentiles). More recent cohorts born in the United States between 1890 and 1990 have had smaller increases in systolic blood pressure with aging. At any age, their systolic blood pressure distributions are shifted lower relative to earlier cohorts. Decreases of 1.9 mmHg in the median systolic blood pressure per decade translates into 11.4-13.3 mmHg over 6-7 decades, a shift that would contribute importantly to lower rates of cardiovascular diseases. These favorable changes are slowing, perhaps owing, at least in part, to the obesity epidemic.

Disparities in psychological distress across socioeconomic status and sex persist throughout adulthood as cohorts age. In this study, we investigate the extent to which this persistence represents either (i) a single set of individuals who at the start of adulthood show distress that is chronic and long lasting or (ii) different sets of individuals that have a staggered onset of short-term distress throughout adulthood. We use path analysis on data from the National Child Development Study, a longitudinal cohort study that assessed psychological distress at ages 23, 33, and 42 years. About 80% of distress disparities at age 42 result from chronic distress that was present in a single set of individuals at least 19 years earlier at the beginning of adulthood. These results support a targeted approach to the reduction of distress disparities that focuses on young adults with high levels of distress and seeks to improve their long-term mental health.

To examine the association of leisure time physical activity, walking pace and resting heart rate with disease-specific mortality in a prospective cohort study by reporting updated analyses of an earlier report we produced with the British epidemiologist, Jerry N. Morris (1910-2009). In the original Whitehall study, 19,019 male, nonindustrial, London-based government employees, aged from 40 to 69 years in 1967 and 1970, participated in a medical examination during which data on leisure time physical activity (N = 6715), self-rated walking pace (N = 6729), and resting heart rate (N = 1183) were collected. Cox proportional hazards analyses were used to estimate hazard ratios for the relation between these exposures and disease-specific mortality. In models adjusted for a range of covariates including socioeconomic status, smoking, and obesity, high resting heart rate was associated with a modestly elevated rate of mortality from all causes (hazard ratio; 95% confidence interval: tertile 3 vs. tertile 1: 1.17; 0.99, 1.37 p[trend]: 0.07) and respiratory disease (1.69; 1.04, 2.76 p[trend]: 0.03). Of the two markers of physical activity, walking pace was inversely related to mortality ascribed to all causes (slow vs. high walking pace 1.71; 1.53, 1.91 p[trend]: <0.001]), coronary heart disease (2.03; 1.68, 2.47 p[trend]: <0.001), and total cancers (1.25; 0.98, 1.59 p[trend]: 0.04). The corresponding associations for leisure time activity were typically weaker. For other mortality endpoints-respiratory disease (walking pace: 1.96; 1.48, 2.60 p[trend]: <0.001]), hematopoietic cancer (walking pace: 1.36; 0.52, 3.51 p[trend]: 0.03), stomach cancer (inactive versus active leisure time: 1.53; 0.88, 2.64 p[trend]: 0.04), and rectal cancer (walking pace: 4.85; 1.70, 13.8 p[trend]: 0.007)-individual activity indices revealed effects, but not both. Higher levels of physical activity indexed by the various markers herein appeared to confer protection against a range of mortality outcomes.

In 1980 we examined 707 67-year-old men, 656 of whom had no previous myocardial infarction. During 8 years of follow-up, 70 (10.7%) of the 656 men developed a first myocardial infarction or died from coronary heart disease (CHD). The incidence of CHD increased 1.6-fold from the lowest to the highest quintile of cholesterol levels, 2.7-fold from the lowest to the highest quintile of triglyceride levels, and 2.2-fold among those with diabetes. Blood pressure, smoking habits, and two measurements of obesity (body mass index and waist circumference) were not significantly related to the incidence of CHD. In multivariate analysis, serum triglyceride levels and blood glucose concentration remained as significant risk factors for CHD. This may reflect that disturbances in glucose and triglyceride metabolism (as part of a metabolic syndrome?) are more important CHD risk factors in older than in younger men.

Using data on children hospitalized with measles in Copenhagen from 1915 to 1925, I found that secondary cases (infected at home) exposed to two or more index cases had a higher case fatality rate than did children exposed to a single index case (relative risk (RR) = 1.90; 95% confidence interval (CI): 1.12 to 3.22). Compared with surviving cases, fatal secondary cases had a shorter interval between their own rashes and rash in the index case (P < .02), suggesting a shorter period of incubation for severe cases. Secondary cases infected by a severe case of measles had higher mortality (RR = 3.87; 95% CI: 1.65 to 9.08) than did secondary cases infected by an index case without pneumonia. These observations suggest that differences in patterns of exposure, possibly due to the dose of infection, may be important for understanding variation in measles mortality.

PURPOSE: To evaluate risk for all-cause and cause-specific mortality in a large, primarily female (73%) cohort of radiologic technologists. METHODS: The study consists of 145,915 radiation technologists, certified in the American Registry of Radiologic Technologists (1926–1982) and followed through 1997. Causes of death were obtained from death certificates or, more recently, through NDI Plus. Standardized Mortality Ratios (SMR) were computed and tests of homogeneity were performed to detect differences in mortality among causes. Poisson models were used to estimate risks using an internal comparison group. RESULTS: Significantly low SMRs were observed for all causes (0.76), all cancers (0.82), and diseases of circulatory system (0.69). Compared to U.S. women, the risk for breast cancer mortality bordered around unity (SMR 1.01, 95% CI 0.94–1.09). However, relative to all other cancers, breast cancer mortality was significantly increased (RSMR 1.24, p < 0.01). Elevated risk for breast cancer was associated with certification before 1940 (SMR 1.55, 95% CI 1.24–1.91), and duration of certification of 20-29 (SMR 1.21, 95% CI 1.06–1.37) and 30+ years (SMR 1.77, 95% CI 1.54–2.02). A 35% increase in leukemia risk was evident for women certified for a duration of 20–29 years and a 36% increase among women certified for 30+ years. Poisson analysis revealed a significant increase in breast cancer risk with increasing number of years certified among women first certified before 1940 (p < 0.001) and during 1940–49 (p = 0.05) compared to women first certified in 1950 or later. CONCLUSIONS: Preliminary findings of this study suggest increased breast cancer risk associated with occupational radiation exposures prior to 1950 and with long-term cumulative exposures. However, potential confounding by reproductive and other risk factors needs to be evaluated.

To determine trends in incidence and survival between 1935 and 1991 and to evaluate risk factors for ovarian cancer among Olmsted County, Minnesota women. All newly diagnosed cases of ovarian cancer among Olmsted County women in 1975-1991 were identified using the medical records linkage system of the Rochester Epidemiology Project. In order to assess trends, incidence rates in the subset of Rochester women were compared with Rochester rates for 1935-1974. Survival was evaluated by the Kaplan-Meier product-limit method. A case-control analysis of risk factors compared Olmsted County women with invasive epithelial ovarian cancer and an age-matched group of women from the community by logistic regression. Altogether, 129 Olmsted County women were newly diagnosed with ovarian cancer in 1975-1991. The age-adjusted (to 1970 United States whites) incidence rate was 22.5 per 100,000 person-years. Median survival from initial diagnosis was 3.7 years. Compared to an equal number of controls, the 103 women with invasive epithelial disease were more likely to be nulliparous (odds ratio [OR] 1.9; 95% CI 0.95-3.9) but less likely to have a history of thyroid disease (OR 0.4; 95% CI 0.2-0.8), hypertension (OR 0.4; 95% CI 0.1-0.9) or nonsteroidal estrogen use (OR 0.5; 95% CI 0.2-0.9). Prior hysterectomy (OR 0.5; 95% CI 0.2-0.9) and unilateral oophorectomy (OR 0.2; 95% CI 0.04-0.7) were also associated with reduced risk. The incidence of ovarian cancer in this community in 1975-1991 was little changed from rates 20 years earlier. There has been some improvement in survival from ovarian cancer in this population compared to 1935-1974, but still less than 50% survive for 5 years. Prior hysterectomy and unilateral oophorectomy appear protective for ovarian cancer.

Recent reports of increasing incidence, especially in men, led us to update through 1989 an earlier study of colorectal cancer incidence in Rochester that covered the period 1940 through 1979. The combined data reflected cancer trends in the community over half a century. Data resources of the Rochester Epidemiology Project were used to identify new cases of colon and rectal cancer among Rochester residents. Incidence rates were estimated using decennial census data, and 95% confidence intervals were based on the Poisson distribution. The GLIM statistical package was used to evaluate trends over time. Age-adjusted (US white 1970) incidence rates of colorectal cancer for men were 53.7, 61.3, 53.7, 54.2, and 52.5 per 100,000 person-years, respectively, for the decades 1940 to 1949, 1950 to 1959, 1960 to 1969, 1970 to 1979, and 1980 to 1989. Comparable rates for women were 42.7, 49.3, 42.9, 40.7, and 40.9 per 100,000 person-years. No statistically significant changes were seen in the incidence of colon and rectum cancer for men or women. There was no consistent trend in tumor stage at diagnosis and the mean size of the initial lesion did not change with time. The incidence of colorectal cancer has not changed over the past 50 years in this community where case ascertainment has been consistent and complete.

Between 1945 to 1949 and 1975 to 1979, the average annual incidence of stroke declined by 45%, from 209 per 100,000 population to 115 per 100,000. For 1980 to 1984, the incidence rate of stroke was 17% higher than that for 1975 to 1979. This pattern--a decline followed by a recent increase--was also evident for cerebral infarction and intracerebral hemorrhage, but rates for subarachnoid hemorrhage remained stable throughout the period of study. The onset of the decline in incidence rates coincided with the introduction of effective antihypertensive therapy, but stabilized and increased rates were associated with continuing improvement in the control of hypertension. The stabilization and increase in the incidence rates of stroke coincided with the introduction of computed tomography (CT), which appeared to increase the detection of less severe strokes. The 30-day mortality for cerebral infarction decreased from 24% in 1945 to 1949 to 12% in 1980 to 1984, while that for intracerebral hemorrhage declined from 91% to 48% and that for subarachnoid hemorrhage from 64 to 42%. The dramatic decrease in mortality for intracerebral hemorrhage occurred during the 10 years when CT was first introduced and was due to the identification of small hemorrhages. These hemorrhages would have been classified as cerebral infarcts in the pre-CT era. Improved management of the secondary complications of cerebral infarction and subarachnoid hemorrhage may explain some of the improved survival for these two stroke subtypes.

To identify the life course model that best describes the association between life course socioeconomic position (SEP) and cardiovascular (CVD) risk factors (ie, body mass index [BMI], systolic and diastolic blood pressure, total cholesterol, low-density lipoprotein, high-density lipoprotein, triglycerides, and glycated hemoglobin) and explore BMI across the life course as mediators of the relationship. The Medical Research Council National Survey of Health and Development was used to compare partial F-tests of simpler nested life course SEP models corresponding to critical period, accumulation, and social mobility models with a saturated model. Then, the chosen life course model for each CVD risk factor was adjusted for BMI at age 53 and lifetime BMI (ages 4, 26, 43, and 53 years). Among women, SEP was generally associated with CVD risk factors in a cumulative manner, whereas childhood critical period was the prominent model for men. When the best-fitting SEP models were used, we found that adjustment for BMI at age 53 reduced associations for all outcomes in both genders. Further adjustment for lifetime BMI (4, 26, 43, and 53 years) did not substantially alter most associations (except for triglycerides). SEP at different points across life influences CVD risk factors differently in men and women.

The purpose of this paper is to determine the age-specific relationships between risk factors at age 40 through 75 years and ischemic heart disease (IHD), and to determine the effects of aging on these relationships in a cohort of 3983 Canadian males. The Manitoba Follow-Up Study is the prospective investigation of cardiovascular disease as it develops in a cohort of 3983 young men. Over a period of 45 years, from 1948 to 1993, 1094 study members (27%) developed clinical evidence of IHD. Blood pressure, body weight, smoking, and presence of diabetes mellitus have been recorded at regular intervals throughout the follow-up period. Using measurements from examinations every 5 years between ages 40 and 75 years, age-specific Cox proportional hazard models were fit to relate these risk factors to IHD. The adjusted relative risk of IHD for systolic blood pressure, diastolic blood pressure and smoking were found to significantly (p < 0.001) decline with advancing age. The adjusted relative risk for body mass index and presence of diabetes mellitus for ischemic heart disease did not vary with age (p > 0.05). After age 65 years, these risk factors were of little value for the prediction of IHD. The relative risk and statistical significance of blood pressure and smoking, as risk factors for IHD, decline with age.

PURPOSE: A unique opportunity for epidemiological studies of cancer and other health effects of radiation exposures exists around the Semipalatinsk Nuclear Test Site in Kazakhstan. The present study is the first analysis of leukemia risk among the residents of downwind settlements exposed to radioactive fallout from atmospheric nuclear weapon tests (1949–1963) and followed up from 1960 to 1998. METHODS: Within the cohort of 10,000 exposed subjects a case-control study was nested, including 22 leukaemia cases (except chronic lymphoid leukemia) and 132 controls individually (1:6 ratio) matched by birth year and sex. Leukemia deaths were identified by death certificates and diagnoses were verified by hospital records. The individual dose including internal and external exposure assessment was estimated according to the residency and age at exposure. All odds ratios were adjusted for ethnicity (Russian or Kazakh) as an independent variable. RESULTS: The median dose of exposure for all subjects was 0.89 Sv ranging from 0.01 to 5.71 Sv. A nearly two-fold increased risk of leukemia was found (OR = 1.91; 95% CI = 0.38 to 9.67) for persons exposed to doses of >2.0 Sv as compared to those exposed to <0.5 Sv, but no increase in risk with the dose was found for those exposed to doses lower than 2 Sv. Detailed evaluation of dose-response showed an excess relative risk for leukemia of 10% per 1 Sv of additional exposure. CONCLUSIONS: Our findings suggest that there is an increased risk of leukemia among those exposed to >2 SV as compared to those exposed to <0.5 Sv, but this could have been a chance finding due to the small number of cases and low statistical power.

To analyze the components of the favorable trends in gastric cancer in Europe. From official certified deaths from gastric cancer and population estimates for 42 countries of the European geographical region, during the period 1950 to 2007, age-standardized death rates (World Standard Population) were computed, and an age-period-cohort analysis was performed. Central and Northern countries with lower rates in the 2005 to 2007 period, such as France (5.28 and 1.93/100,000, men and women respectively) and Sweden (4.49 and 2.21/100,000), had descending period and cohort effects that decreased steeply from the earliest cohorts until those born in the 1940s, to then stabilize. Former nonmarket economy countries had mortality rates greater than 20/100,000 men and 10/100,000 women, and displayed a later start in the cohort effect fall, which continued in the younger cohorts. Mortality remained high in some countries of Southern and Eastern Europe. The decrease in gastric cancer mortality was observed in both cohort and period effects but was larger in the cohorts, suggesting that the downward trends are likely to persist in countries with higher rates. In a few Western countries with very low rates an asymptote appears to have been reached for cohorts born after the 1940s, particularly in women.

Current measures of successful quitting are insensitive to changes induced by tobacco control activities. We evaluated whether changes in the incidence of successful quitting, a new measure of cessation, can inform policy makers how population subgroups responded. Smokers from National Health Interview Surveys (NHIS) (1965 through 1992, n = 140,199) were used to determine the number of current smokers eligible to quit at the beginning of each year from 1950 through 1990. Incidence of quitting, computed for different demographic subgroups, was the ratio of those newly successfully quit each year to those eligible to quit. Overall, incidence increased over fivefold, from < 1% in 1950 to a still low 5% in 1990. When the health risks of smoking were first disseminated, middle-aged men had the highest quitting incidence. Gender differences in younger smokers occurred following the beginning of the public health campaign of the mid 1960s, as the dangers of smoking to the fetus were documented. Younger adult smokers appeared to increase quitting markedly in the 1970s, around the beginning of the nonsmokers' rights movement. Quitting patterns in middle-aged African Americans were similar to whites, although at much reduced levels. Younger African Americans had low quitting incidence until 1989. Incidence differed by educational attainment; regardless of age, during the 1970s and 1980s, those with some college increased their quitting incidence markedly. Incidence of quitting is a sensitive indicator of relatively short-term changes in successful quitting in population subgroups and should facilitate evaluation efforts.

The objective of this study was to evaluate mortality rates from ischemic heart disease among Icelanders during the period of 1951 to 1985. In some developed countries, the number of deaths from ischemic heart disease declined markedly in this time period, and it is interesting to study whether the same has occurred in Iceland. The study was based on information obtained from the Statistical Bureau of Iceland, which keeps records of deaths based on death certificates as well as other population records. Nonparametric tests were used to correlate death rates and calendar years. Rates per 100,000 were calculated and plotted. The results indicated that the mortality rates from ischemic heart disease among Icelanders have not yet peaked.

To evaluate the partial effects of vaccination against equine Japanese encephalitis (JE) and characterize other prognostic factors based on previous outbreak records in Japan from 1953 to 1960. Individual case records, which included demographic information, vaccination history, and clinical information (dates of onset, recovery and death, and symptoms), were investigated. The relations between two outcomes, JE death and symptomatic period, and other variables were examined. Of a total reported 803 cases during the observation period, 453 (56.5%) were diagnosed with either serological, histopathological, or epizootiological methods. Vaccination (adjusted odds ratio=0.77, 95% confidence interval: 0.61, 0.97) and an older age (adjusted odds ratio=0.83, 95% confidence interval: 0.71, 0.96) significantly reduced the risk of JE death. The symptomatic period was also significantly shortened with vaccination (p<0.001). The risk of JE death was lowered and the symptomatic period of survivors shortened with inactivated JE vaccination. These findings demonstrate the partial effects of vaccination in reducing the burden of this disease.

Trends in mortality rates for liver cirrhosis between 1955 and 1990 have been analyzed for 38 countries (two from North America, six from Latin America, five from Asia, 23 from Europe, and Australia and New Zealand) on the basis of official death certification data derived from the World Health Organization database. Chile and Mexico had exceedingly high rates (around 60/100,000 males and 15/100,000 females in the late 1980s), while in Canada, the United States, and Latin American countries that provided data, cirrhosis death rates were between 5 and 17/100,000 males and 3 and 5/100,000 females over the same calendar period. The pattern of trends was, however, similar in all American countries, with some increase between the 1950s and the 1970s, and declines thereafter. A similar trend was observed in Japanese males, whose rate was 13.6 in 1990. Conversely, cirrhosis mortality declined steadily from 8.0 to 4.6 in Japanese females. Appreciable downward trends were observed in Hong Kong and Singapore, whereas mortality increased in Thailand. In Europe, in the late 1950s, the highest rates were registered in Portugal (33.6/100,000 males and 14.6/100,000 females), followed by France (31.8/100,000 males and 14.1/100,000 females), Austria, Italy, Spain, and Germany. Most of these countries, however, after some further rise up to the 1970s, showed reversal of the trends over most recent years. Thus, in the late 1980s or early 1990s, only Austria, Italy, and Portugal had cirrhosis mortality around 30/100,000 males and 10/100,000 females. Britain, Ireland, and Nordic countries started from much lower values (2 to 4/100,000 males), but showed some, although discontinuous, upward trend.(ABSTRACT TRUNCATED AT 250 WORDS)

High childhood IQ test scores have been associated with increased alcohol dependency and use in adult life, but the relationship between childhood IQ and illegal drug use in later life is unclear. Participants were 6713 members of the 1958 National Child Development Survey whose IQ was assessed at 11 years and had their lifetime illegal drug use measured at 42 years of age. In analyses adjusted for a range of covariates, a 1 SD (15-point) increase in IQ scores was associated with an increased risk of illegal drug use in women: ever using cannabis (odds ratio [OR], 1.30; 95% confidence interval [95% CI], 1.16-1.45), cocaine (OR, 1.66; 95% CI, 1.21-2.27), amphetamines (OR, 1.50; 95% CI, 1.22-1.83), amyl nitrate (OR, 1.79; 95% CI, 1.30-2.46) and "magic mushrooms" (OR, 1.52; 95% CI, 1.18-1.98). Associations were of lower magnitude in men. In this cohort, high childhood IQ was related to illegal drug use in adulthood.

The aim of the study is to determine period and birth-cohort effects in the early initiation of phencyclidine (PCP) use in drug users in New York City (NYC). We analyzed data collected from two surveys of street-recruited drug users in NYC. We used survival analysis and proportional hazards modeling to assess period and birth-cohort effects on risk for early initiation of PCP use. Of 787 participants, 292 (37.1%) had used PCP by the age of 23 years. Before 1987, there was a greater risk for initiation of PCP use through the age of 23 years (hazard ratio [HR] = 34.77; 95% confidence interval [CI], 21.45-56.36). Proportional hazards modeling showed that those born in the 1971 to 1975 birth cohort compared with those born in 1976 to 1980 had a lower risk for initiation of PCP use through age 23 years (HR = 0.58; 95% CI, 0.37-0.91). Other significant predictors of PCP use by age 23 included white race and having been in a juvenile detention center. There are period and birth-cohort differences in the likelihood of early initiation of PCP use. Changes in drug culture and social norms may influence the likelihood of initiation of PCP use. This may have implications for interventions aimed at slowing the nationwide increase in use of PCP.

Serum total cholesterol (> or = 6.7 mmol/L) measured in 1960 in the Charleston Heart Study cohort was found to be a risk for mortality from coronary heart disease during the period of 1960 to 1988 in white men (relative risk [RR] 1.5; 95% confidence interval [CI]: 1.1, 2.2), white women (RR 1.7; 95% CI: 1.1, 2.7), and black women (RR 1.6; 95% CI: .9, 2.9) after age, systolic blood pressure, smoking status, education level, obesity, and diabetes were considered. For black men, the relative risk was .96 (95% CI, .39, 2.39). Only among white women was the relative risk (RR 2.4; 95% CI, 1.2, 4.5) increased among those in the older ages (55 to 74) in 1960. The evidence for cholesterol as a risk factor for coronary disease mortality in black men is inconclusive and requires further study.

We sought to investigate the relationship between regular daily smoking in adolescence and in adulthood, the onset age of suicidal behavior, and the repetitiveness of suicide attempts by the age of 31 years. Data from the Northern Finland 1966 Birth Cohort (NFBC 1966, n = 7995) were linked with the National Finnish Hospital Discharge registers (FHDR). Smoking habits of the cohort members were assessed at the ages of 14 and 31 years. After adjusting for hospital-treated psychiatric disorders and several sociodemographic characteristics, women with prolonged smoking from 14 to 31 years especially were at increased hazard (adjusted hazard ratio, 6.67; 95% confidence interval, 3.06-14.52) for having their first suicide attempt at younger age compared with infrequent smokers/nonsmokers. Smoking habits were not associated with the repetitiveness of suicide attempts in either gender. This study confirms the association between smoking and suicidality. Further studies are needed to investigate the neurobiological basis of this association.

To analyze patterns and trends in gastric cancer mortality in the European Union (EU) over the period 1968-1992, paying special attention to changes associated with birth cohort. Poisson log-linear models were used to quantify geographic differences and relative annual changes. To assess trends associated with birth cohort, invariant parameters from sex-specific age-period-chohort models (net drift and curvature), for each country, were used to choose a restricted slope range for cohort effect. Gastric cancer mortality declined throughout the EU. The male-to-female ratio stood at around 2 in all countries, yet showed a slight rise over time. Portugal reported the highest age-adjusted rates for men and women (45.63 and 23.31 per 100,000 person-years, respectively). The rate ratio between two extreme countries (Portugal/Denmark) exceeded 3. Quantitative intercountry differences were found in trend slopes, with a decrease of 5% per annum in Finland. Risk of dying associated with birth cohort decreased over successive generations. Small local rises in risk, in almost all countries among generations born around the 1940s, support the importance of diet early in life in the etiology of gastric cancer. Despite the substantial decline in gastric cancer mortality witnessed in the EU, stress must be accounted for the wide differences among countries and the smaller decline in the youngest generations, particularly among women. This latter finding suggests a possible stabilization or even a rise in the rates in future, rendering it important for these trends to be monitored over the next few years.

Skin cancer incidence data from West Germany are available only for the territory of the Federal State of Saarland. We examined time trends in melanocytic and non-melanocytic skin cancer mortality for the territory of West Germany including a population of about 66 million people. We analyzed the melanocytic and non-melanocytic skin cancer mortality data (1968-99) from West Germany including West Berlin. We calculated age-specific and age-standardized mortality rates and used Poisson regression to estimate underlying age, cohort, and period effect. The estimated percent annual increase of the skin melanoma mortality rate was 1.0% (95% CI, 0.7-1.3) among men and 0.5% (95% CI, 0.2-0.7) among women. This increase is mainly due to a rate increase in people aged 60 years or more. The skin melanoma mortality trend was best explained by age-, cohort-, and period effects. The risks increased in each successive birth cohort born between 1890 and roughly 1935. Thereafter, the risks declined through the most recent birth cohort born in 1975. Skin melanoma mortality in West Germany showed an increase from 1968 through 1999 in people aged 60 years or more. The favorable mortality decline by birth cohort in the most recent birth cohort is an important indicator of a likely decline in mortality over the coming years.

This study examined racial variations in CHD (coronary heart disease) mortality rates (1968-1992) of residents aged 35-84 in the state economic areas (SEAs) surrounding the ARIC (Atherosclerosis Risk in Communities) study. The quarter century of CHD mortality rates are discussed in relation to racial and gender differences in baseline risk factors measured in the ARIC cohort and to the incidence of hospitalized myocardial infarction and case fatality rates obtained from the community surveillance component of the ARIC study between 1987 and 1994, inclusive. Five-year average annual, gender- and age-specific CHD mortality rates were compared across race groups, based on National Vital Statistics data for state economic areas. Five-year average annual CHD mortality declined 2.6% for white men and women and 1.6% and 2.2% for black men and women, respectively. The black-white mortality rate ratio increased over time for men and women. The black-white mortality age crossover (higher black than white mortality in young men, lower black than white mortality at older ages) had disappeared by the end of the observation. CHD mortality was markedly greater in black than white women at all ages and time periods. The black disadvantage in CHD mortality was increasingly greater in the ARIC SEAs than in the United States as a whole. Persistent and increasing racial disparities in CHD mortality occurred in the ARIC SEAs concurrently with racial differences in risk factors, the incidence of myocardial infarction, and case fatality rates.

Previous studies suggest that the risk of coronary heart disease (CHD) may be related to educational level. In the present study, we looked at factors in childhood and adolescence as potential confounders of the relationship between educational level and risk of CHD. We also examined hypothesized mediation by socioeconomic factors in adulthood. The 1969 conscription cohort consisting of 49,321 Swedish men born in 1949 through 1951, who conscripted for military service in 1969/70, provided information on potential confounders. This was linked with register-based information on childhood social circumstances, education level, occupational class, income and job control in adulthood, and follow-up information on CHD during 1991 through 2007. The relative risk of CHD increased with lower educational level. Among men with the lowest educational level, the relative risk of CHD was 1.81 (95% confidence interval, 1.60-2.05) compared with the highest educated group of men. Adjustment for childhood socioeconomic circumstances, and cognitive ability and behavior-related factors measured in late adolescence, attenuated the association considerably. Additional adjustment for socioeconomic position, income, and job control in adulthood did not attenuate the association further. It may be that educational level and risk of CHD are associated importantly owing to confounding from factors in childhood and adolescence.

To estimate occupation-specific risk of seminomas and nonseminoma subtypes of testicular cancer among Swedish men gainfully employed in 1970 over the period 1971-1989. Age-period standardized incidence ratios were computed in a dataset linking cancer diagnoses from the Swedish national cancer register to occupational and demographical data obtained in the census in 1970. Log-linear Poisson models were fitted, allowing for geographical area and town size. Taking occupational sector as a proxy for socioeconomic status, occupational risks were recalculated using intra-sector analyses, where the reference group comprised other occupations in the same sector only. Risk estimators per occupation were also computed for men reporting the same occupation in 1960 and 1970, a more specifically exposed group. Seminomas and nonseminomas showed a substantial geographical variation. The association between germ-cell testicular tumors and high socioeconomic groups was found mainly for nonseminomas. Positive associations with particular occupations were more evident for seminomas, for which railway stationmasters, metal annealers and temperers, precision toolmakers, watchmakers, construction smiths, and typographers and lithographers exhibited a risk excess. Concrete and construction worker was the only occupation consistently associated with nonseminomas. Among the many occupations studied, our results corroborate the previously reported increased risk among metal workers, specifically related with seminomatous tumors in this study. Our results confirm the geographical and socioeconomical differences in the incidence of testicular tumors. These factors should be accounted for in occupational studies. The different pattern of occupations related with seminomas and nonseminomas support the need to study these tumors separately.

This article presents trends in stroke mortality and incidence in Finland among people aged 25 to 74 years. Between 1971 and 1980, stroke mortality declined steeply: 4.1% per year among men and 5.5% per year among women. Between 1981 and 1991 the decline was smaller; about 2.2% per year in men and 2.8% per year in women. The North Karelia stroke register showed that stroke mortality declined in men from 155 per 100,000 per year in 1972 to 1973, to 87 per 100,000 per year in 1982 to 1983, and in women from 114 to 44 per 100,000 per year. A slight decline in mortality was observed during the 1980s in men, but not in women. The incidence of stroke also declined in North Karelia during the 1970s, from 328 to 248 per 100,000 per year in men, and from 230 to 141 per 100,000 in women. In the FINMONICA stroke register, the average rate of decline in incidence of stroke between 1983 and 1989 was 1.7% per year in men and 1.8% per year in women. Declines in incidence and mortality from subarachnoid hemorrhage were observed in both men and women; nevertheless it was the decline in cerebral infarction that accounted for most of the changes since about 80% of all strokes are cerebral infarctions. In conclusion, despite steep falls in stroke mortality and incidence in the 1970s, stroke mortality is still high in Finland compared with other nations. During the 1980s, the decline in stroke mortality was less and incidence leveled off until it resumed from 1987 to 1989.(ABSTRACT TRUNCATED AT 250 WORDS)