Tecumseh Study Research Articles

Bayesian model choice for epidemic models with two levels of mixing

This paper considers the problem of choosing between competing models for infectious disease final outcome data in a population that is partitioned into households. The epidemic models are stochastic individual-based transmission models of the susceptible-infective-removed type. The main focus is on various algorithms for the estimation of Bayes factors, of which a path sampling-based algorithm is seen to give the best results. We also explore theoretical properties in the case where the within-model prior distributions become increasingly uninformative, which show the need for caution when using Bayes factors as a model choice tool. A suitable form of deviance information criterion is also considered for comparison. The theory and methods are illustrated with both artificial data, and influenza data from the Tecumseh study of illness.

Acute Bloody Diarrhea: A Medical Emergency for Patients of All Ages

Acute bloody diarrhea should be considered a medical emergency. Its causes are frequently serious or actionable or both and are usually identified. However, acute bloody diarrhea as a stand-alone clinical presentation has received little scholarly attention in the past several decades. Although the range of possible causes of acute bloody diarrhea is broad, infectious considerations are paramount and should always be prioritized in the evaluation of such patients. History, examination, and laboratory testing should be focused on minimizing time to diagnosis (and, by extension, to implementing appropriate therapy). Strategically chosen tests and imaging, avoidance of extraneous diagnostic pursuits, and provision of supportive care while awaiting diagnostic clarity are central to the adroit management of patients with acute bloody diarrhea. Diagnostic considerations differ somewhat between adults and children but have many elements in common, including the need for vigilance in detecting Escherichia coli O157:H7 infection. In this review, we discuss diagnostic approaches (emphasizing the importance of rapid, accurate, and thorough microbiologic investigation) and measures that can be taken to support patients while awaiting information that determines the cause of their disease. These topics are discussed in the context of the medical care that is available to children and adults with bloody diarrhea in most institutions in developed nations.

Office and out of office heart rate measurements: which clinical value?

Office and out of office heart rate measurements: which clinical value?

Prevention Conference VI: Diabetes and Cardiovascular disease: Writing Group I: epidemiology.

In patients with diabetes, a high priority must be given to modification of the major risk factors for cardiovascular disease (CVD). There is growing evidence that control of these risk factors will reduce the likelihood of developing CVD and its complications in patients with diabetes.1 In clinical management of patients with diabetes, attention must be given to the following risk factors: smoking, hypertension, prothrombotic state, low-density lipoprotein (LDL) cholesterol and diabetic dyslipidemia, hyperglycemia, overweight/obesity, physical inactivity, and adverse nutrition. Specific considerations of Writing Group IV will be reviewed. They will be discussed in light of current recommendations for management of risk factors in diabetes as presented by the American Diabetes Association (ADA), the American Heart Association (AHA), and the national education programs sponsored by the National Heart, Lung, and Blood Institute (NHLBI).2–11⇓⇓⇓⇓⇓⇓⇓⇓⇓ These recommendations are summarized in the Table. View this table: Table 1107395. Goals for Risk Factor Management in Patients With Diabetes In addition to being a cause of many forms of cancer and chronic lung disease, cigarette smoking is a major cardiovascular risk factor. When a smoking patient also has diabetes, this patient is doubly at risk for CVD. Thus, every effort must be made to convince patients with diabetes who smoke to give up the smoking habit. This need is strongly reinforced by a position statement from the ADA.10 Elevated blood pressure is a major independent risk factor for multiple cardiovascular end points: coronary heart disease (CHD), stroke, chronic renal failure, and heart failure.11 Patients with diabetes have an increased prevalence of hypertension.12 Multiple factors undoubtedly contribute to hypertension in patients with diabetes, eg, obesity, insulin resistance, hyperinsulinemia, and renal disease. Systolic hypertension appears to be the main blood pressure–related risk factor in patients with diabetes.13 …

A “gender blind” relationship of lean body mass and blood pressure in the Tecumseh study

Background: Body size correlates positively with blood pressure (BP) but there is controversy about the roles of obesity versus muscularity in this relationship. Methods: We examined the BP relationship with overweight, lean body mass (LBM), and muscle performance in 231 adolescents (17.25 ± 3.07 years, 123 males). The skinfold thickness (SKINT) was used to measure overweight, as this was a growing population. Results: Maximal foot torque, a measure of muscle strength, correlated strongly (r = 0.51, P < .001) to LBM attesting to the validity of the calculated LBM. Anthropometric measurements were available also in 944 adults (29.9 ± 5.5 years, 461 men). Correlations of LBM to systolic (adolescents r = 0.52, adults r = 0.19, both P < .001) and diastolic (adolescents r = 0.47, adults r = 0.20, both P < .001) BP were highly significant. SKINT also correlated significantly to systolic and diastolic BP in adolescents and in adults, respectively. In both genders and populations an increasing SKINT was associated with a similar increase in BP, but this effect was superimposed on an average 10 mm Hg between-gender BP difference. The LBM in both groups and genders related to the BP in an identical fashion; the men were on the high and the women on the low end of the same BP/LBM correlation line. Thus, the amount of LBM erased categoric BP differences between the genders. Conclusions: The gender-related BP differences appear to reflect the inherent gender differences in muscle bulk.

Home Blood Pressure as a Predictor of Future Blood Pressure Stability in Borderline Hypertension: The Tecumseh Study

We evaluated time-related blood pressure trends in the Tecumseh study participants, none of whom received antihypertensive treatment. At baseline the blood pressures were measured in the field clinic and by self measurement at home (twice daily for 7 days). After a mean of 3.2 ± 0.42 years, the clinic and home pressure readings were repeated. Nine hundred forty-six subjects had clinic and home blood pressure readings at baseline. Of these 735 (380 men, 355 women; average age, 32 years) also completed the second examination. Blood pressure, morphometric data, and biochemical measures at the first examination were used as predictors of future clinic blood pressures. Five hundred ninety-six subjects were normotensive on both examinations (81%). Of 79 subjects (10.7%) with clinic hypertension (>140 mg Hg systolic or 90 mm Hg diastolic) at baseline, 38 remained hypertensive (“sustained hypertension”) and 41 became normotensive (“transient hypertension”) after 3 years. Another 60 normotensives at baseline (10.4%) became hypertensive on second examination (“de novo hypertensives”; incidence; 8.1%). The home blood pressure readings on both examinations were reproducible. The three hypertensive groups had elevated home blood pressure, were overweight, had dyslipidemia, and higher insulin values. Only the home blood pressure proved predictive of subsequent blood pressure trends. A home blood pressure of 128 and 83 mm Hg or higher detected “sustained” hypertension with a 48% sensitivity and 93% specificity. Readings of 120 and 80 mm Hg or lower predicted future normotension with a 45% sensitivity and a 91% specificity. We conclude that self determination of the blood pressure at home is useful in the management of borderline hypertension. An algorithm for the management of these patients is proposed.

Relationship of Tachycardia With High Blood Pressure and Metabolic Abnormalities

Faster resting heart rate has been shown to be associated with a higher risk of developing hypertension and a greater incidence of cardiovascular morbidity and mortality. The aim of this study was to investigate the distribution of heart rate and its relationship with blood pressure and other cardiovascular risk factors in three populations. One European general population (Belgian study), one North American general population (Tecumseh study), and one European hypertensive population (HARVEST trial) were studied. Within each population, mixture analysis was used to investigate whether a mixture of two normal distributions explained the variance in heart rate better than a single distribution. In the men of all populations, mixture analysis identified a larger subpopulation of subjects with normal heart rate and a smaller one with fast heart rate. The subgroups with tachycardia had higher blood pressure and lipid levels than those with normal heart rate. In the populations in which they were measured, fasting insulin and postload glucose were also higher in the men with faster heart rate. A subgroup with tachycardia could also be singled out among the women from Tecumseh, but no relation between heart rate and blood pressure could be found. These findings show that in Western societies, high heart rate pertains to a distinct subgroup of subjects, who are more frequently men and exhibit the characteristic features of the insulin resistance syndrome. Sympathetic overactivity is likely to be the mechanism underlying this clinical condition.

Trends in Left Ventricular Function over Three Years in the Tecumseh Study

Amerena J, Nesbitt S, Krause L, Grant E, Lu H, Julius S. Trends in left ventricular function over three years in the Tecumseh study.The relationship between blood pressure and left ventricular diastolic function was examined in participants of the Tecumseh Blood Pressure study. When subjects were divided into three blood pressure groups according to blood pressure levels 3 years apart, it was found that subjects who were had sustained “hypertension” at both time points (SH) had a decreased early/late diastolic filling rate (E/A ratio) compared to subjects who were hypertensive at only one of the timepoints (OH) arid those who were consistently normotensive (NN) (1.71 ± 0.02 NN vs 1.55 ± 0.05 OH, (p < 0.0001) vs 1.56 ± 0.07 SH, (p < 0.04)). This relative order was maintained when the second estimation of diastolic filling was performed 3 years later, but the E/A ratio had decreased significantly in all groups (1.54 ± 0.01 NN vs 1.45 ± 0.03 OH (p < 0.01) vs 1.37 ± 0.06 SH (p < 0.006)), consistent with an age-related reduction in diastolic filling. Heart rates were significantly higher in the hypertensive groups initially (63.5 NN vs 66.2 OH (p < 0.03) vs SH 67.3 (p < 0.01)) and increased in all groups over time, with the largest increase in the SH group (64.9 2 0.04 NN vs 67.8 ± 1.02 OH (p < 0.0001) vs 70.9 ± 1.6 SH (p < 0.01)). Stroke volume index changed in all groups over time, with the increase greatest in the NN group and least in the SH groups the reverse of this pattern was seen for changes in heart rate. All subjects gained weight over the 3 years of the study so that these unexpected changes in stroke volume index and heart rate could be a consequence of a weight gain-related increase in the sympathetic tone, although we have no direct evidence that this is the case. Should this be so, the smaller increase in stroke volume in conjunction with the larger increase in the heart rate in the sustained hypertensive group may reflect the effects of mild blood pressure elevation in producing a reduction in left ventricular diastolic function associated with a decrease in the inotropic responsiveness of the heart to enhanced sympathetic tone.

Higher risk of cardiovascular mortality among lean hypertensive individuals in Tecumseh, Michigan.

A cohort of 2181 men and women, aged 40 to 79 years, without evidence of coronary heart disease or cancer at entry to the Tecumseh Study was evaluated. Subjects were defined as lean if their Metropolitan Life Insurance table relative weight was < 110 (n = 584) and as obese if their relative weight was > or = 120 (n = 1024). There were 688 subjects with hypertension at study entry (systolic blood pressure > or = 160, diastolic blood pressure > or = 95, or treated). The 29-year relative risk (RR) of mortality from ischemic heart disease (IHD) or cardiovascular disease (CVD) associated with systolic blood pressure level was significant for both lean and obese subjects. Among hypertensive subjects, the RR of fatal IHD for lean versus obese hypertensive subjects was 1.87 (95% confidence interval, 1.21 to 2.88) and the RR of fatal CVD was 1.56 (95% confidence interval, 1.10 to 2.20) using a Cox proportional-hazards model to adjust for the independent effects of age and traditional CVD risk factors. The findings are consistent with other studies in men showing lean hypertensive subjects to be at greater risk of IHD or CVD mortality than obese hypertensive subjects. A similar finding is now observed in women. Associations do not prove causality or dictate management. Nevertheless, the unexplained higher mortality in lean versus obese hypertensive subjects has now been reported with sufficient frequency to suggest that the association is real (if unexplained). Determining the reasons for this association may improve targeted prevention and treatment strategies.

Effect of home blood pressure and gender on estimates of the familial aggregation of blood pressure. The Tecumseh Blood Pressure Study.

Blood pressure (BP) readings from a single clinic visit are often used in population studies investigating the genetic basis of BP. We examined first-degree relatives in the Tecumseh Blood Pressure study to compare heritability estimates of BP readings obtained in the clinic-office setting (the average of two seated readings) with self-reported home BP readings (the average of 14 readings) taken over a 1-week period. The hypothesis tested was that repeated BP readings obtained in the home over the 1-week period would have fewer artifacts (i.e., environmentally induced variability in BP) and thus would better estimate the true "basal" BP that, in turn, would improve heritability estimates. We and others assume that the true basal BP level is heritable. We therefore expected that this "true" BP, by reducing BP variability of offspring, would show a stronger between-sibling correlation and that it would correlate better to parental BP as measured in a clinic setting. Correlation coefficients were calculated between siblings in the present Tecumseh study using self-reported home BP and clinic BP readings. Among 380 siblings (average age, 31.4 years), correlation coefficients for the home readings were of the same magnitude as for office readings (home, r = 0.23, p less than 0.01; office, r = 0.24, p less than 0.01). When offspring clinic BP readings were compared with archived BP data on parents, the correlation between offspring clinic and parental clinic BP readings was stronger (r = 0.24, p less than 0.05) than the correlation of offspring home BP readings to parental clinic BP readings (r = 0.17, p less than 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)

Tecumseh study of illness. XIII. Influenza infection and disease, 1976-1981.

The occurrence of influenza was followed in Tecumseh, Michigan during the five year period 1976-1981 by identifying onset of acute respiratory illness and by virus isolation and serology. Type B outbreaks were observed in 1976-1977 and 1979-1980, type A (H3N2) in 1977-1978 and 1980-1981, and type A (H1N1) viruses in 1977-1978, 1978-1979, and 1980-1981. Evidence of low level circulation of viruses in the year preceding an outbreak was not obtained. Age-specific isolation rates from specimens collected by the community physicians differed from age-specific isolation rates from specimens collected from the surveillance, suggesting the operation of a selection mechanism in the former. Symptoms associated with virus isolation were strongly influenced by age. Within age groups, several variables, especially median duration, indicated type A (H3N2) had produced the most severe illnesses, type A (H1N1) the mildest, with type B intermediate. Age-specific infection rates determined by serology for the 1976-1977 and 1977-1978 influenza seasons confirmed the consistently high rates for type A (H3N2) in children with some fall-off with increasing age. Type A (H1N1) rates peaked in children aged 5-19 years and type B in children aged 5-14 years. This may be related in part to insensitivity of the hemagglutination inhibition test in those under age 5 years. Infection with type A (H1N1) was detected at low frequency in adults. Pathogenicity was calculated based on the serologic data. It was estimated for all ages combined that, at a minimum, type A (H3N2) infection produced febrile illness in 25% of cases and type B infection produced respiratory illness in 34% of cases.

MORTALITY FROM CORONARY HEART DISEASE IN THE TECUMSEH STUDY

The coronary heart disease mortality of participants in the Tecumseh study was examined with particular emphasis on the roles of diabetes and glucose tolerance as risk factors. The cohort consisted of 921 men and 937 women aged 40 years and older who did not have evident coronary heart disease at entry to the study during the period 1959-1965 and whose outcome was determined in the period 1977-1979. Previously diagnosed diabetes was a statistically significant risk factor for coronary heart disease mortality in both sexes even after controlling for systolic blood pressure, serum cholesterol, relative weight, and cigarette smoking. High blood glucose score in nondiabetics was associated with excess coronary heart disease mortality after controlling for other risk factors, but the magnitude of this effect was substantially below that of diabetes. The predictive power of most risk factors except age itself decreased among progressively older segments of the population.

The Tecumseh study. XII. Enteric agents in the community, 1976-1981.

Enteric illnesses in residents of Tecumseh, Michigan, were studied from 1976 to 1981. The frequency of illness among adults and children increased each year in the late autumn, usually in November. This peak of illness preceded the main period of appearance of the rotaviruses, which were less regularly associated with another peak of illness. Rotaviruses were identified in 3.8% of all stool specimens collected; in specimens from children under two years of age, the annual rate of rotavirus identification was 10.4%. All rotaviruses were identified during the period from late December to early April. Bacterial pathogens were isolated from 3.3% of stool specimens, with no concentration in any age group or season. However, most enteric illnesses in the community were not associated with recognized pathogens. Testing of blood specimens collected in 1976-1978 by complement fixation confirmed the seasonal pattern of rotavirus activity. Rotavirus infections were documented in all age groups, even in older adults, and were associated with symptomatic illness.

Serologic responses to nonprevalent influenza A viruses during intercyclic periods.

Sera collected at different times in the course of the Tecumseh Study of Respiratory Illness have been tested for the presence of antibodies to a variety of nonprevalent human and animal type A influenza viruses. In sera collected during the 1966-1968 period of H2N2 prevalence, rises in antibody titer were detected against the H3 antigen. The frequency of these rises in titer increased gradually and significantly as the time of emergence of the H3N2 virus approached. Occasional rises in Hsw1 antibody titer and presence of such antibody in sera obtained from young children were also detected in specimens collected at that time. In pre-vaccination 1976 serum collection, the prevalence of antibody against Hsw1N2 was significantly higher in certain age groups than that found in 1966-1968. During the 1976-1977 period, evidence for increase in H1N1 antibody frequency was found in nonvaccinated children, again before the emergence of the H1N1 strain. Possible interpretations of the results are discussed.

Blood glucose and risk of coronary heart disease.

Health status of 1877 Tecumseh Study subjects aged 35-64 years was ascertained in 1977. They represented 77% of the persons in this age range who were apparently healthy and had participated in comprehensive examinations of nearly the entire population of the community in 1959-1960 and 1962-1965. Subjects who developed coronary heart disease had a significantly higher mean blood glucose concentration than other members of the cohort, even after exclusion of diabetics. Similarly, when examined as single variables, age, sex, serum cholesterol, systolic blood pressure, number of cigarettes smoked per day, and relative weight were significantly related to incidence of coronary events. In the multiple logistic function, however, age, cigarette smoking, blood pressure, and blood glucose were the only significant variables. In a two-way interaction model, glucose and cholesterol were a highly predictive pair. After exclusion of diagnosed diabetics, glucose by itself or in interaction with other variables was not significant in the multiple logistic functions.

Alterations in Pulmonary Function Following Respiratory Viral Infection

Respiratory viral illness is a major cause of morbidity in both adults and children. This report focuses on both the acute and chronic effects on respiratory function of these ubiquitous infections. Infant airways are particularly vulnerable due to the relatively low conductance in immature peripheral airways. Bronchiolitis, caused predominantly by respiratory syncytial virus, is the most important of these viral illnesses and is emerging as a major risk factor for the subsequent development of obstructive airway disease in adults, possibly by interference with normal alveolar proliferation. The basic pathogenic mechanism involved in adult respiratory viral infection is bronchial hyperreactivity, presumably secondary to epithelial damage and resultant sensitization of rapidly adapting airway receptors. In addition, there may be virus-related alterations in the autonomic and humoral regulation of airway tone. Viral infections may alter the effects of common air pollutants on respiratory function.

Serum cholesterol, diet, and the decline in coronary heart disease mortality

There has been a recent and dramatic decline in coronary heart disease death rates in the United States. The possible contribution of secular trends in serum cholesterol to this decline was examined by comparison of data from five epidemiologic studies: the Framingham Study, the Tecumseh Study, the Health Examination Survey, the Health and Nutrition Examination Survey, and the Lipid Research Clinics Program Prevalence Study. It appears that in the 1960s there was a modest decline in serum cholesterol levels (approximately 5 mg/dl in middle-aged adults). Although it is impossible to exclude methodologic variation as the main explanation for the observed trends, the contribution of dietary changes to this decline was also examined, by comparison of data from the Diet-Heart Study and the Lipid Research Clinics study. The decline in serum cholesterol was found to be consistent with the observed changes in dietary fat intake in the 1960s. When the potential impact of the decline in cholesterol levels on coronary heart disease death rates was assessed by use of a Framingham multivariate logistic risk function, it was found that a decline in cholesterol of 5 mg/dl would predict a 4.3% decline in 6-year coronary heart disease death rates in middle-aged men.

Daily nutritional intake and serum lipid levels. The Tecumseh study

To determine the influence of diet on serum cholesterol and triglyceride levels among adults, 24-hr dietary recall interviews were conducted among 957 men and 1,082 women resident in the community of Tecumseh, Michigan. Trained interviewers obtained detailed description of all foods consumed during 24 hr before venipuncture for lipid determination. Using a list of nutritional composition of 2,706 foods prepared from standard references, nutritionists determined quantities of all nutrients common to the American diet which were consumed by each participant according to a 24-hr diet record. For analysis, men and women were grouped into lower middle, and upper tertiles according to serum cholesterol and triglyceride levels. The mean daily consumption of each dietary component was virtually identical in all tertiles for men and women but differed between sexes. Cholesterol and triglyceride levels were unrelated to quality, quantity, or proportions of fat, carbohydrate, or protein consumed in the 24-hr recall period.

Independence of serum lipid levels and dietary habits. The Tecumseh study

Independence of serum lipid levels and dietary habits. The Tecumseh study

Blood pressures and electrocardiographic findings in the South Australian Aborigines.

Blood pressure measurements were recorded in 522 adults and 141 10-19 year-old full and part blood Aborigines in five communities. The means for systolic and diastolic blood pressures at 40 years were close to those reported for Europeans, although below this age, values tended to be lower, and above 40 years tended to be higher than those reported in the Tecumseh study. Hypertension, as defined by the Princeton criteria, was present in 29%, more often in the men (1-6 to 1-0), and eight subjects satisfied the criteria for hypertensive heart disease (HHD). 522 electrocardiograms were recorded on adult subjects at five Aboriginal communities and classified according to categories of the Minnesota code. Of the 210 abnormalities observed, minor T wave inversions and minor S-T segment depression were the most commonly encountered, and were more frequent in female subjects. Q wave changes typical of myocardial infarction was found in 5% of the tracings and occurred mainly in older men. If hypertension and certain ECG codes are assumed to be "risk factors" for the development of clinical ischaemic heart disease (IHD), the urbanized Aboriginal had a higher prevalence compared with Caucasian subjects of the Busselton study. "Probable" and "suspect" ECG changes of IHD, although mainly in the older subjects, were found to be associated with hyperglycaemia, as recognised in Western society. It is postulated that urbanized Aborigines are prone to cardiovascular degenerative disease to a similar or possibly larger extent than Caucasians.