Tecumseh Community Health Study Research Articles

Estimation of maximal oxygen consumption and heart rate recovery using the Tecumseh sub-maximal step test and their relationship to cardiovascular risk factors

BackgroundMaximum aerobic capacity (VO2max) is associated with lower cardiovascular and total mortality. Step tests can be used to provide an estimate of (VO2max) in epidemiological or home-based studies. We compared different methods of estimation of VO2max and heart rate recovery and evaluated the relationship of these estimates with cardiovascular risk factors. MethodsData were analysed from 2286 participants in the Tecumseh Community Health Study (>16 y and <70 y) who performed a step test. VO2max was estimated from heart rate using three methods and the results compared. The magnitude of heart rate recovery (HRR) and the time constant of recovery based on different time intervals post-exercise were also estimated. ResultsEstimated VO2max showed good or poor agreement depending on the method used. VO2max correlated inversely with systolic blood pressure (SBP), body mass index (BMI), total cholesterol, blood glucose following a 100 g oral load (PLG) and Framingham risk score. In a multivariable model age sex, cigarette smoking, SBP, BMI and PLG were significantly inversely associated with VO2max. Correlations with risk factors were strongest for HRR measured over the first 30 s of recovery. Only the time constant calculated from the 3 min post-exercise period correlated significantly with risk factors. ConclusionsThe Tecumseh step test can be used to provide estimates of VO2max and heart rate recovery. Estimated VO2max was inversely associated with higher systolic BP, higher BMI and worse glucose tolerance. Measurements of HRR over the first 30 s and the time constant calculate from the first 3 min of recovery correlate most closely with risk factors.

BMI and Headache Among Women: Results From 11 Epidemiologic Datasets

To evaluate the association between BMI (kg/m(2)) and headaches among women. Cross-sectional analysis of 11 datasets identified after searching for all large publicly available epidemiologic cohort study datasets containing relevant variables. Datasets included National Health Interview Survey (NHIS): 1997-2003, the first National Health Examination and Nutrition Survey, Alameda County Health Study (ACHS), Tecumseh Community Health Study (TCHS), and Women's Health Initiative (WHI). The women (220,370 in total) were aged 18 years or older and had reported their headache or migraine status. Using nonlinear regression techniques and models adjusted for age, race, and smoking, we found that increased BMI was generally associated with increased likelihood of headache or severe headache among women. A BMI of approximately 20 was associated with the lowest risk of headache. Relative to a BMI of 20, mild obesity (BMI of 30) was associated with roughly a 35% increase in the odds for experiencing headache whereas severe obesity (BMI of 40) was associated with roughly an 80% increase in odds. Results were essentially unchanged when models were further adjusted for socioeconomic variables, alcohol consumption, and hypertension. Being diagnosed with migraine showed no association with BMI. Among US women, a BMI of approximately 20 (about the 5th percentile) was associated with the lowest likelihood of headache. Consistently across studies, obese women had significantly increased risk for headache. By contrast, the risk for diagnosed migraine headache per se was not obviously related to BMI. The direction of causation and mechanisms of action remain to be determined.

The Gender Gap in Coronary Heart Disease Mortality: Is There a Difference between Blacks and Whites?

The gender difference (gender gap) in mortality due to coronary heart disease (CHD) decreases with age. This relationship has not been well characterized in diverse populations. To examine the gender gap in CHD mortality across age groups and to compare the age dependency of the gender gap between blacks and whites, we conducted a prospective cohort study combining data from 9 U.S. epidemiological studies (Atherosclerosis Risk in Communities Study [ARIC], Charleston Heart Study, Evans County Study, Framingham Heart Study [original and offspring cohorts], National Health Examination Follow-up Study [NHEFS], Rancho Bernardo Study, San Antonio Heart Study, and Tecumseh Community Health Study). Baseline examinations were performed between 1958 and 1990 (depending on the study), and mean follow-up was 13.7 years in general communities in several U.S. geographic areas. We included 39,614 subjects >30 years and free of cardiovascular disease (CVD) at baseline (18% blacks, 37% men). Completion of follow-up was >97% for all studies. As the main outcome measures, age-specific CHD mortality rates and male/female CHD mortality hazard ratios were calculated using Cox hazards regression. During 542,605 person-years of follow-up, 2,812 CHD deaths were observed (18% in blacks, 46% in men). At age 45, white men were at a 6-fold increased risk compared with white women (95% confidence interval [95% CI] 4.6-7.9), whereas black men had a 2-fold increased risk of fatal CHD compared with black women (1.4-3.6). At age 95, men and women were at equal risk in both whites (0.9-1.4) and blacks (0.7-1.6). The difference in the age dependency of the gender gap between blacks and whites was significant (p < 0.0001). The gender difference in CHD mortality was more pronounced in whites than in blacks at younger ages. This discrepancy was not explained by adjustment for CHD risk factors and suggests that other factors may be responsible for the ethnic variation in the gender gap.

Relation between cardiovascular risk factors at adult age, and physical activity during youth and adulthood: the Leuven Longitudinal Study on Lifestyle, Fitness and Health.

This study investigated the relationship between sports participation/physical activity during youth (13 - 18 years of age) and adulthood (30 - 40 years of age), and cardiovascular risk factors (body fat and fat distribution, blood pressure, lipoprotein levels and cardiorespiratory fitness) at 40 years of age. Subjects were 166 Flemish males from "The Leuven Longitudinal Study on Lifestyle, Fitness and Health". Physical activity was assessed by means of a sports participation inventory and the Tecumseh community Health Study Questionnaire. In addition to correlation and multiple stepwise regression analyses, different groups (at risk, not at risk) were contrasted on sports participation/physical activity parameters using ANOVA. Long-term exposure during adulthood to daily physical activity was slightly related to a low/high risk profile for waist circumference, percent body fatness, triglycerides and peak VO(2). Sports participation during adolescence was not related to levels of cardiovascular risk factors at 40 years of age.

Doubly labelled water validation of three physical activity questionnaires.

This study investigated the validity of the Baecke Questionnaire, the Five City Project Questionnaire, and the Tecumseh Community Health Study Questionnaire in 19 Flemish males, using correlation and multiple stepwise regression analyses. The three questionnaires are commonly used physical activity questionnaires in epidemiological studies. The physical activity level (PAL) as measured with the doubly labelled water method was used as the criterion. The Baecke total activity index showed the highest correlation coefficient with PAL (r = 0.69, p < 0.001). Also the sweat index from the Five City Project Questionnaire, and total daily energy expenditure from the Tecumseh Community Health Study Questionnaire showed significant associations with PAL, respectively 0.57 (p < 0.05) and 0.64 (p < 0.01). Multiple stepwise regression analyses supported the findings from the correlation study. The largest individual contribution in PAL was from the activity index (45%) for the Baecke Questionnaire, the sweat index (29%) for the Five City Project Questionnaire, and total daily energy expenditure (38%) for the Tecumseh Community Health Study Questionnaire. In conclusion, the questionnaires, and certainly the Baecke Questionnaire, can provide valid data about physical activity. Therefore they are useful in epidemiological studies.

Weight loss increases and fat loss decreases all-cause mortality rate: results from two independent cohort studies.

In epidemiological studies, weight loss is usually associated with increased mortality rate. Contrarily, among obese people, weight loss reduces other risk factors for disease and death. We hypothesised that this paradox could exist because weight is used as an implicit adiposity index. No study has considered the independent effects of weight loss and fat loss on mortality rate. We studied mortality rate as a function of weight loss and fat loss. Analysis of 'time to death' in two prospective population-based cohort studies, the Tecumseh Community Health Study (1890 subjects; 321 deaths within 16y of follow-up) and the Framingham Heart Study (2731 subjects; 507 deaths within 8y of follow-up), in which weight and fat (via skinfolds) loss were assessable. In both studies, regardless of the statistical approach, weight loss was associated with an increased, and fat loss with a decreased, mortality rate (P < 0.05). Each standard deviation (s.d.) of weight loss (4.6 kg in Tecumseh, 6.7 kg in Framingham) was estimated to increase the hazard rate by 29% (95% confidence interval CI), (14%, 47%, respectively) and 39% (95% CI, 25%, 54% respectively), in the two samples. Contrarily, each s.d. of fat loss (10.0 mm in Tecumseh, 4.8 mm in Framingham) was estimated to reduce the hazard rate 15% (95% CI, 4%, 25%) and 17% (95% CI, 8%, 25%) in Tecumseh and Framingham, respectively. Generalisability of these results to severely (that is, body mass index BMI) > or = 34) obese individuals is unclear. Among individuals that are not severely obese, weight loss is associated with increased mortality rate and fat loss with decreased mortality rate.

Cancer incidence in the rural community of Tecumseh, Michigan. A pattern of increased lymphopoietic neoplasms

The Tecumseh Community Health Study (TCHS), initiated in 1959 at the University of Michigan School of Public Health, has provided a resource for long term prospective studies of the incidence of cancer in the setting of a midwestern rural farming community. A survey of total and site specific cancer incidence among 7016 male and female adult residents from 1959-1987 was conducted, and the observed number compared with the expected number, based on the age-sex-race-calendar period-and site-specific cancer incidence rates reported by the Connecticut tumor registry. Based on the results of this survey, a hypothesis was advanced concerning the potential risks of exposures to insecticides and herbicides. This was pursued by analyzing for each county in Michigan the comparative annual number of acres and percentage of acreage treated with agricultural chemicals in 1978 and for the period 1982-1987. Finally, because of the availability of information on lifestyle risk factors that had been collected in the 1960s on all participants, a nested case-control study was implemented. The standardized incidence ratio (SIR), based on the observed number divided by the expected number of invasive cancer cases (all sites combined [excluding nonmelanoma skin cancer]), was not significant in females (SIR, 0.95; 95% confidence interval [CI], 0.86-1.05) nor males (SIR, 0.91; 95% CI, 0.81-1.01). A significant increase was demonstrated for males and females combined in the incidence of lymphopoietic neoplasms, namely non-Hodgkin's lymphoma, Hodgkin's disease, and chronic lymphocytic leukemia; the combined SIR was 1.40 (95% CI, 1.03-1.86; P = 0.03). In the Department of Commerce surveys of counties in Michigan, the Tecumseh area (Lenawee County) was ranked highest with respect to the average annual number of acres (240,000 or more) and the percent of acreage (40%) sprayed with herbicides and insecticides. A comparison of temporal trends for cancer incidence since 1973 was reviewed for the Wayne-Oakland-Macomb tricounty area, in which the survey estimated that less than 80,000 acres per year, or less than 8% of acreage, were treated with pesticides. The comparison of the Tecumseh cohort, for all sites combined, was not significantly different from expectation in females (SIR, 1.01; 95% CI, 0.89-1.14), and was decreased by more than 10% in males (SIR, 0.88; 95% CI, 0.77-1.00). However, the SIR for non-Hodgkin's lymphoma in females was significantly elevated (SIR, 1.92; 95% CI, 1.07-3.11, P = 0.02); the trend for increased risks of lymphomas and leukemias was also evident in males. In the nested case-control study, based on risk factor information documented prior to diagnosis, the relative risk of a family history of lymphoma, leukemia, or multiple myeloma was significantly increased among patients with lymphoproliferative neoplasms (odds ratio, 3.81; 95% CI, 1.49-9.79; P = 0.005). This prospective epidemiologic study conducted in a rural farming community in Michigan has found significant increases in the incidence dence of lymphoproliferative cancers. A plausible hypothesis, consistent with the preliminary findings, is that the reported cancer pattern is an expression of risk resulting from sustained environmental exposures to agricultural chemicals, perhaps in conjunction with familial or genetic factors.

Obesity as a Risk Factor for Osteoarthritis of the Hand and Wrist: A Prospective Study

The Tecumseh Community Health Study provides an opportunity to investigate the role of obesity in the etiology of osteoarthritis. This longitudinal study, conducted in Tecumseh, Michigan, began in 1962 with baseline examinations of clinical, biochemical, and radiologic characteristics. A 1985 reexamination of the cohort characterized osteoarthritis status in 1,276 participants, 588 males and 688 females, who were aged 50-74 years at this follow-up. Baseline obesity, as measured by an index of relative weight, was found to be significantly associated with the 23-year incidence of osteoarthritis of the hands among subjects disease free at baseline. Greater baseline relative weight was also associated with greater subsequent severity of osteoarthritis of the hands. The difference between baseline and follow-up weight values was not significantly associated with the incidence of osteoarthritis of the hands. Furthermore, there was no evidence that development of osteoarthritis subsequently led to increased incidence of obesity.

Epidemiology and genetics of osteoarthritis

Osteoarthritis is the most common of the arthropathies. Previous population-based epidemiologic studies have estimated the prevalence of osteoarthritis by age, gender, race, and site in many populations. Cross-sectional and retrospective data from these prevalence surveys have been analyzed to identify possible risk factors for the development of osteoarthritis at specific sites. Data from prospective population-based epidemiologic studies including the Framingham Osteoarthritis Study, the Tecumseh Community Health Study, the National Health and Nutrition Examination Survey I Epidemiologic Followup Study, and the Baltimore Longitudinal Study of Aging are being analyzed to estimate the incidence of osteoarthritis by age, gender, and site, and the prognosis of subjects with osteoarthritis at specific sites in terms of morbidity and mortality, as well as to confirm the association of risk factors with the development and progression of osteoarthritis by site. Additionally, the application of techniques of molecular biology to the study of the genetics of familial osteoarthritis has identified hereditary defects in type II collagen that appear to predispose to the early development of osteoarthritis in affected family members. Results of these studies highlight opportunities for the primary and secondary prevention of osteoarthritis.

Alcohol consumption and the risk of breast cancer: A report from the Tecumseh Community Health Study

The relationship between prior alcohol consumption and the risk of breast cancer was studied in 1954 women in the Tecumseh Community Health Study (TCHS) who entered the cohort in 1959–1960 and were followed potentially for 28 years. The mean alcohol consumption at baseline was 0.89 (SD 2.2) oz/week for premenopausal women and 0.85 (SD 2.2) oz/week for postmenopausal women. Only 25% of the cohort consumed more than 0.5 oz of ethanol/week or about 1.6 g/day. The adjusted relative risks (RRs) for breast cancer associated with the use of ethanol vs never drinking were 0.93 (95% CI, 0.40 – 2.18) for ex-drinkers, 1.08 (95% CI, 0.64 – 1.82) for 0-< 1 drink/day, 1.23 (95% CI, 0.49 – 3.10) for 1-<2 drinks/day and 1.12 (95% CI, 0.25 – 5.01) for ⪰2 drinks/day. There were only 37 subjects in the group at the highest level of consumption (⪰2 drinks/day). There was no significant interaction between alcohol and the period of onset of breast cancer (premenopausal or postmenopausal). In the TCHS, alcohol consumption generally at levels not exceeding 2 drinks/day, was not significantly associated with an increased risk of breast cancer. Although we have found little excess risk associated with alcohol consumption, the wide confidence intervals summarized above are not inconsistent with previously published reports that have suggested a modest positive association.

Progression of osteoarthritis of the hand and metacarpal bone loss. A twenty‐year followup of incident cases

We examined the prospective relationship between metacarpal bone mass and osteoarthritis (OA) of the hand, using incidence data from the historical cohort in the Tecumseh Community Health Study (Tecumseh, MI). Women were examined for radiographic evidence of OA and for bone mass twice, 20-23 years apart (1962-1965 and 1985; 683 subjects with an age range of 55-74 in 1985). Two measures of OA were evaluated: the highest score assigned to any of the 32 wrist/hand joints, and the sum of scores for all wrist/hand joints. After adjustment for age, women who were classified as having OA (by either measure of OA) in 1985 were more likely to have more cortical area at baseline, which indicates greater bone mass. Women who developed OA in the 23-year period were more likely to experience a significantly greater widening of the medullary cavity over time, an indicator of increased bone resorption. Women with increasing levels of OA involvement also had an increased likelihood of greater cortical area loss. We conclude that women who later developed OA were more likely to have higher baseline bone mass than women who did not develop OA, but these women also had a greater likelihood of bone loss over time.

A commingling analysis of obesity in the Tecumseh community health study.

A commingling analysis of the distributions of obesity as quantified by Quetelet's index, triceps skinfold, suprailiac skinfold, umbilical skinfold, subscapular skinfold, a trunk-extremity ratio of skinfold measurements and Quetelet-adjusted trunk-extremity ratio was performed in a population of 9,226 individuals living in Tecumseh, Michigan, during 1962-1965. For all but the (non-Quetelet-adjusted) trunk-extremity ratio, two or three distributions were needed to fit the data adequately even after performing a power transformation to eliminate skewness. Although this evidence is compatible with a major gene effect, segregation analyses of these data using the unified, mixed model are necessary to resolve whether the commingling is due to a major gene.

Heterogeneity among populations for familial aggregation of blood pressure.

A simple model of familial aggregation was applied to age-sex corrected systolic (SBP) and diastolic (DBP) blood pressure data from two separate studies: The Québec Family Study and the Tecumseh Community Health Study. Examination of the familial correlations suggested heterogeneity across studies in all DBP correlations, but for SBP only spouse correlations were higher in Québec. Fitting the path model to the data revealed that the only source of heterogeneity for SBP was in spouse resemblance (u), which was significant in Québec but could be dropped from the model in Tecumseh. For SBP, no further heterogeneity was detected, with the vertical transmission being compatible with (but not evidence for) purely polygenic inheritance (i.e., τf = τm = 1/2). Familiality estimates (t2 ) for SBP were quite high (nearly 40%), and extra sibling resemblance was significant. For DBP, heterogeneity was detected in all parameters of the model except for spouse resemblance. Common sibship environmental effects (b) were of little importance in Tecumseh but were highly significant in Québec. Vertical transmission from mothers (τm ) could be fixed at 1/2 in Tecumseh but was only about half as large in Québec, and familiality estimates were higher in Québec (nearly 60%) than in Tecumseh (about 25%). One source of the heterogeneity for SBP was due to cohabitation, which positively impacted on sibling correlations. The source of the sample heterogeneity in DBP may be through environmental factors which have a smaller impact in Tecumseh than in Québec, although genetic heterogeneity could not be ruled out. Differences in the study designs, as well as differences in the two populations, could account for the heterogeneity detected.

Familial Transmission of Alcohol Use: IV. A Seventeen-Year Follow-up on the Relationships between Parent and Adult Offspring Alcohol Use; Tecumseh, Michigan

The sample for this study comprised 190 three member sets of father, mother, and adult son or daughter who were first interviewed in 1960 as part of Cardiovascular Examination I of the Tecumseh Community Health Study. At that time, alcohol data were collected from self-reports for both parents and offspring. In 1977 alcohol data were again collected for the adult offspring only, from self-reports. In 1960 offspring's drinking was significantly related to both fathers' and mothers' drinking and remained so in 1977 despite shifts over time in the proportions of offspring across five alcohol use categories. The majority of offspring did not fall into their parents' category, but the mode by and large reflected the parent category. The majority of offspring of non-drinkers did drink, but at low and not high volume; the majority of offspring of high volume drinkers drank at lower levels than their parents, and were not likely to be life-long abstainers, although they may have become past drinkers. Two non-drinking parents reinforced non-drinking in their adult offspring and two drinking parents reinforced the early onset and persistence of high volume drinking. Among offspring who changed their drinking level between 1960 and 1977, those who had drunk less than their parents were the most likely to have increased their drinking, and those who had drunk more than their parents were the most likely to have decreased their drinking. Patterns observed in this study may be viewed as intrafamilial sources of drinking stability transmitted across generations in a single community.

Health and Other Characteristics of Employed Women and Homemakers in Tecumseh, 1959-1978

Using data collected in 1959-60 and 1978-79 in the Tecumseh Community Health Study, characteristics of 866 women aged 20-44 years at baseline were analyzed according to employment status at baseline and follow-up (employed/employed, homemaker/employed, employed/homemaker and homemaker/homemaker). To avoid the bias of including unhealthy women with homemakers, disabled and retired women were excluded. Women employed at both exams more frequently reported current alcohol consumption than other women, and the association persisted after controlling for covariables. Smoking habits, rates of reported pregnancy outcomes and conditions, and rates of gynecological problems did not differ significantly among employment groups.

Health and Other Characteristics of Employed Women and Homemakers in Tecumseh, 1959-1978

Using data collected in 1959-60 and 1978-79 in the Tecumseh Community Health Study, characteristics of 866 women aged 20-44 years at baseline were analyzed according to employment status, at baseline and follow-up (employed/employed, homemaker/employed, employed/homemaker, and homemaker/homemaker). To avoid the bias of including unhealthy women with homemakers, disabled and retired women were excluded. There were few significant differences in health among the groups. Prevalence rates of respiratory conditions were slightly higher at follow-up for women employed at both exams, while prevalence rates of cardiovascular conditions were slightly higher for homemakers at follow-up. The rate of reporting heart attacks was significantly higher for homemakers at both exams, although rates of diagnosed coronary heart disease were similar among employment groups.

Characteristics of individuals and long term reproducibility of dietary reports: The Tecumseh diet methodology study

Food frequency reports in 1967–1969 were compared to frequency reports of the same foods asked retrospectively in 1982–1983 about 1967–1969 for 1184 respondents aged 45–64 years in the Tecumseh Community Health Study. The kappa statistic for concordance of the retrospective and baseline reports was used as a summary measure of the individual's ability to reproduce his or her earlier diet report. Reproducibility was estimated for total diet, represented by 83 foods, and for 9 subsets of foods of epidemiologic interest. In bivariate and multivariate analyses, reproducibility was strongly related to stability of diet; those whose diets changed least over the 15-year period had greatest diet reproducibility. Greater total diet reproducibility was also found among men with higher education, among women of < 110% desirable weight reporting no special diet and among women reporting no medications. Consistent with current models of memory, the retrospective report of diet was strongly related to the current report of diet. Agreement between the retrospective and baseline diet reports was greater than agreement between the current and baseline diet reports. This indicates that, as a proxy for past diet, the retrospective report of diet is superior to the current report. Similar relationships were found for the 9 subset of foods.

Commingling analysis of blood pressure in the Tecumseh community health study.

A commingling analysis of the distribution of systolic and diastolic blood pressure was performed on measurements from 9,226 individuals living in Tecumseh, Michigan during 1962-1965. Even after correcting for skewness in the distributions, multiple normal distributions fit the data significantly better than a single normal one, suggesting some level of commingling. The exact nature of the components differed depending on the phenotype, and the number of components entertained, suggesting a possibly complex etiology, but with some definite qualitative effects. Segregation analysis under the mixed model is needed to address the nature of blood pressure in this community more definitively.

Factors associated with pain and osteoarthritis in the Tecumseh Community Health Study

Previous longitudinal studies have shown no associations between increasing amount of radiographic hand osteoarthritis (OA) and levels of hand pain/disability. In this longitudinal study, we aimed to study whether radiographic hand OA was related to pain/disability in cross-sectional and longitudinal settings focusing on joint-specific analyses.We included 190 patients (173 women, mean (standard deviation, SD) age 61.5 (5.7) years) from the Oslo hand OA cohort, of whom 112 had 7-year follow-up data. Finger joints were scored for radiographic OA according to the Kellgren–Lawrence scale and Osteoarthritis Research Society International (OARSI) atlas. Pain and function were assessed by clinical examination (joint tenderness), grip strength and the Australian/Canadian (AUSCAN) questionnaire. Associations between radiographic hand OA and tenderness in the same joint were examined by logistic regression analyses with Generalized Estimating Equations, whereas associations between overall amount of radiographic OA and hand pain/disability were assessed by linear regression (adjusted for age and sex).A dose-dependent association was found between the severity of radiographic OA and tenderness in the same joint. Joints that progressed into severe radiographic OA during follow-up had the highest odds of developing tenderness (OR = 11, 95% confidence interval (CI) 4.0–33). Incident erosions seemed to be the most important individual feature associated with incident tenderness (OR = 6.2, 95% CI 3.2–12). Weak associations were found between the amount of radiographic hand OA and overall hand pain/disability.Radiographic hand OA is associated with tenderness in the same joint, and erosive development strongly predicts future joint tenderness independent of other radiographic features.

Prevalence of radiologically defined osteoarthritis in the finger and wrist joints of adult residents of Tecumseh, Michigan, 1962-65

Radiographs of the fingers and wrists of adult participants in the Tecumseh Community Health Study in 1962–1965 were examined for signs of osteoarthritis (OA). The severity of OA for each of 32 joints of the fingers and wrists was recorded for each individual. Attention was restricted to the 3035 participants who were 32 years of age or older and for whom a diagnosis of OA was available for each of 32 joints. Joint-specific prevalence rates of OA increased sharply with age for both sexes, and at the older ages, the prevalence rates for most joints were higher for females. Older individuals with OA also had a greater number of affected joints, with females having a greater number of affected joints than males. Of those individuals aged 44 years or younger, only 6.2% had one or more joints affected with OA. The percentages were 21.6 and 42.0% for those aged 45–59 years and 60 or more years, respectively. The distal interphalangeal (DIP) joints were the most frequently affected joints in all age categories for both sexes and OA in the proximal interphalangeal (PIP) joints was positively associated with OA in the DIP joints. However, controlling for the number of affected DIP joints, the PIP joints of older subjects were more likely to exhibit OA than the PIP joints of younger subjects. Though there is an association between OA in the DIP and PIP joints, there was only a small, nonsignificant association (OR= 1.24, 95% CI=0.83, 1.84) between disease in the DIP and PIP joints of the same finger.