The United Arab Emirates (UAE), an Arabian Gulf state, has a population of 3.1 million, of whom only a small minority are citizens of the country, the remainder being expatriate workers and their families. The country has undergone rapid development over the last 30 years following the discovery of oil and the formation of the country from 7 emirates. A modern infrastructure has been established, and residents have undergone significant changes in lifestyle that parallel the rapid development of their country, including transition from deficiency diseases and undernutrition to degenerative diseases associated with overnutrition. These changes may lead to a rise in risk factors for chronic diseases and changes in reproductive factors in women. Establishing prevalences for such factors is essential for public-health planning (1), and planning must be done separately for men and women in Muslim countries, such as UAE, because strict sexual segregation is followed for most health programmes.
In developing countries, especially Muslim countries, population-prevalence studies are difficult, particularly for women, because of lack of access to a representative sample of community-dwelling people and reticence on the part of subjects to participate. These barriers can be overcome by using community survey techniques developed specifically for this purpose (2). We used such techniques to provide a representative sample of community-dwelling adult female citizens in Al Ain city, UAE, a small desert oasis city with about 300,000 people located in an agricultural district.
Only four studies have assessed the prevalence of risk factors for chronic diseases or reproductive factors among female citizens of UAE. Female citizens represented over half of the members of three communities studied in 1989-1990 for prevalence of obesity, diabetes, and hypertension (3) but other risk factors for chronic diseases and reproductive factors were not studied. This study showed a prevalence of obesity of 28% among women surveyed, a rate that prompts concern and identifies a need for further study. The UAE Ministry of Health performed two community surveys on child and family health in 1987 and 1995, which included some information on reproductive factors in women but did not look at lifestyle factors (4,5). In addition, one survey of postmenopausal women has been reported (6), but this survey included all residents, not just citizens, and did not look at lifestyle factors.
Since none of these studies provided adequate information to guide public-health planning for female citizens of UAE, the present study was conceived. The aim of this study was to establish the prevalence of various health-related lifestyle and reproductive factors among adult female citizens in Al Ain and to investigate their association with personal characteristics to guide health promotion and disease-prevention planning for UAE women.
MATERIALS AND METHODS
Approval for the study was obtained from the Research Ethics Committee of the Faculty of Medicine and Health Sciences, UAE University, UAE.
A random sample of community-dwelling female UAE citizens living in the urban area of Al Ain medical district was surveyed during September 2000-August 2001. The stratified sampling technique was chosen to allow for an efficient data-collection process. Certain areas of the city are set aside for citizens' residences, and non-citizens are restricted from living in these areas. Seventy-five such areas were identified, and 16 were selected randomly. The best available map of each area selected was used for identifying all streets in the area, and a predetermined number of streets were randomly selected based on the size of the area. Every third house on each selected street was visited up to a maximum of 10 houses. All female UAE citizens aged over 19 years in each household visited were eligible for the study. Since no published demographic data are available in which female citizens are specifically identified, the actual sampling frame of female citizens was unknown. Therefore, sampling ratios were chosen based on the investigators' previous experience with community surveys in this population to recruit approximately 500 subjects for the study. This number was calculated to be sufficient to produce 95% confidence limits for prevalence of less than 10% even if as many as 20% of responses were missing.
Following informed consent, a trained healthcare worker interviewed each eligible subject in Arabic using a questionnaire in Arabic. Items covered included demographic data, reproductive history, physical activity, tobacco use, perceived health status (good, fair, poor), and sun exposure. Subjects were asked to list all medications, dietary supplements, and vitamin preparations currently being taken. They were also asked to list all chronic illnesses that they believed they had and also were specifically asked about certain common diseases.
Menopause was defined as cessation of menstrual periods for more than six months in a woman aged [greater than or equal to] 45 years in the absence of known pregnancy. Women who met this definition were defined as postmenopausal and the remainder as premenopausal. Skin exposure to sunlight was determined from the subject's report of the average length of time in minutes per day she had spent in the courtyard of her house in the previous six weeks. In this cultural group, this is the place where the majority of sunlight exposure occurs because, in all other outdoor situations, women generally wear garments completely covering their body. Since sunlight exposure of over 15 minutes per day is considered the minimum necessary to produce adequate vitamin D (7), subjects were divided into those who met this requirement and those who did not.
Physical activity was estimated as the average over the last six weeks based on a series of questions validated for the assessment of customary activity in the elderly (8). These questions were chosen because female UAE citizens are rarely involved in sports activities. Five indicators of current activity were asked: time spent walking outdoors, walking speed, time spent standing indoors, frequency of muscle loading activity, such as climbing stairs or carrying a load, and time spent in productive or leisure activities, both indoors and outdoors. Four levels of intensity were defined for each indicator and were assigned a score of 1 to 4 with 1 being least active and 4 being most active. A composite score was derived from the above by summing the score for the five indicators. Missing values for physical activity were substituted with the median score for the subject's age group for each activity. The resulting score was subsequently divided into quartiles for analysis. The activities reported by women were also compared with the levels recommended by an expert panel as necessary to maintain health (9).
Standing height in centimetres and weight in kilograms were recorded for all subjects using an electronic weigh scale and a meter stick. Body mass index (BMI) was calculated from these two measurements. Using internationally-recognized definitions, underweight was defined as a BMI of <18.5 kg/[m.sup.2], overweight as BMI of 25-29.9 kg/[m.sup.2], and obesity as BMI of [greater than or equal to] 30 kg/[m.sup.2] (10). Body fat content was measured in non-pregnant women by bioelectric impedance (Biodynamics Body Composition Analyzer, Model 310e, Biodynamics Corporation, Seattle, Washington, USA) following the manufacturer's instructions (11). This technique accurately estimates fat-free mass and total body water. From this and body weight, fat mass and percentage of body fat were estimated using the manufacturer's proprietary formulae. Obesity based on bioelectric impedance was defined as [greater than or equal to] 35% of body fat as recommended in the User's Guide supplied by the manufacturer (11), although [greater than or equal to] 40% of body fat was also examined as possibly more appropriate for women aged about 40 years (12).
Data processing and analysis
Data were processed using SPSS version 10.0 and Epi Info 2000. The results of analyses which were adjusted for the selection of more than one subject in the same family using the complex sample tables in Epi Info 2000 software did not differ from the results of analyses which ignored the sampling strategy. Therefore, the results of the simpler analyses are presented. Central tendency was expressed as mean and standard deviation (SD), and correlation was expressed as r. Chi-square test was used for ascertaining the significance of differences in proportions between two or more groups. Probability of <0.05 by 2-tailed test was considered to be statistically significant. For chronic diseases risk factors found to be highly prevalent, prevalence was adjusted for other associated exposure variables (p<0.2 on univariate analysis) by means of multiple logistic regression analysis.
In total, 183 households, which included 535 eligible women, were surveyed. Very few (5%) households approached refused to participate. No eligible woman in a participating household refused to participate in the survey, although a number of them refused or were unable to answer individual questions.
Table 1 shows the characteristics of the study subjects. Seventy-nine percent were citizens by birth rather than the marriage, thereby representing the indigenous population. Health status was generally good with only 25% reporting a chronic illness and 84% reporting sufficient exercise to meet expert recommendations. However, the majority was either overweight or obese (Table 2). Table 2 also shows that the majority of subjects was in the youngest age group, indicative of the age structure of this population. The mean age was 34.3 (SD 14.7) years. Age and education were highly significantly (p<0.001) associated in this rapidly developing population, with 84% of those aged over 40 years having no education and many of those aged less than 40 years (42%) having post-secondary education.
The prevalence of overweight and obesity, as defined by BMI, in this population was 27% and 35% respectively (Table 2). Table 2 also indicates the major relationships between obesity, as defined by BMI, and other lifestyle and reproductive characteristics. BMI was chosen to define the various categories of weight in this table to allow for international comparisons. Body fat levels as measured by bioelectric impedance were similarly distributed in younger women. However, although prevalence of obesity as defined by BMI declined in women in the older age groups (Table 2.), in all age groups over 40 years had similar high prevalence rates of elevated body fat, whether 35% or 40% of fat was used as a cut-off. As expected, percentage of body fat was significantly correlated with BMI (r=0.67, p<0.001) but elevated body fat (whether 35% or 40% was used as a cut-off) identified a somewhat different group of the population as obese than did BMI of [greater than or equal to] 30 (Table 3). Most women classified as obese by percentage of body fat but non-obese by BMI were aged over 40 years. Pregnant women had a weight distribution similar to non-pregnant women, and removing them from analysis of other factors made no difference in the findings.
Since obesity was so prevalent, univariate and multivariate logistic regression analyses were performed to identify other exposure variables significantly associated with it. Body fat percentage of [greater than or equal to] 35 was chosen to define obesity for this analysis because it appeared to be the most accurate definition of obesity in these women (12). Univariate logistic regression identified the following variables as significantly associated (p<0.2) with obesity: education, age, physical activity score, health status (good/not good), presence of chronic illness, menopausal status, and number of pregnancies. Of these variables, education, menopausal status, and number of pregnancies were too highly correlated with age (coefficient >0.6 by Pearson or Spearman correlation) to be entered in multivariate analysis with age. The remaining variables, such as age, physical activity score, health status, and presence of chronic illness, were entered into multivariate logistic regression analysis. Only age was a significant predictor of obesity in this analysis (p=0.000, odds ratio 1.05 with 95% CI 1.04-1.07, n=436). Use of BMI of [greater than or equal to] 30 or body fat percentage of [greater than or equal to] 40 as a definition of obesity led to similar results in multivariate logistic regression analysis.
Those who were more physically active were significantly more likely to be better educated (p<0.001) and were also more likely to be younger (p<0.001) with all of those in the oldest age group falling into the lowest quartile for physical activity. Taking multivitamins was the most common in the 30-39-year age group (52%) and the least common in the 60-69-year age group (15%) (p<0.01). This activity was not significantly associated with level of education but was highly associated with pregnancy (p<0.001) with 74% of pregnant women taking multivitamins, whereas only 38% of non-pregnant women took multivitamins. Sunlight exposure of [greater than or equal to] 15 minutes per day was significantly associated with both least (86% prevalence) and best educated (76% prevalence) groups and was also associated with reporting higher exercise levels (p<0.01 for both). There was no statistically significant association between sun exposure and age or menopausal status.
The mean age, at menarche, of these women was 13.2 (SD 1.51) years with the median being 13 years and the range being 9 to 20 years. Of the non-pregnant, premenopausal women, 26% reported irregular menses and 23% reported current or past use of oral contraceptive pills. In women who had reached menopause, mean fertility was 5.2 children with a median of 7 children and a range of 0 to 12 child(ren). There was no significant association between the number of children and education because few postmenopausal women had received any education at all.
Pregnant women were similar to non-pregnant women. The only two areas of significant difference were a higher likelihood of taking multivitamins (p<0.001) and fewer reports of chronic illnesses (p=0.001) in pregnant women. They had a tendency to report better health but this did not reach statistical significance (p=0.09). No pregnant woman reported smoking. Postmenopausal women were significantly different from premenopausal women (Table 4). They were more likely to report poor health (p<0.001) and chronic illnesses (p<0.001), get less physical activity (p<0.001), have a higher percentage of body fat (p<0.01), and less likely to take multivitamins (p<0.001). It was not possible to record age at menopause in these women as most had poor recollection of the timing of their menopause.
The most important finding of this study is the high prevalence (35%) of obesity, as defined by BMI, among the subjects. Since the study population was a random sample of female UAE citizens residing in Al Ain city, this prevalence can be extrapolated at least to the female population of this city and possibly to the urban female population of UAE. This high prevalence of obesity represents a significant predictor of high rates of cardiovascular diseases, diabetes, osteoarthritis, and cancer in this population in the future and demands public-health action. Therefore, health services for these women should emphasize the prevention and treatment of obesity. Prevention programmes should begin in childhood since 21% of women are already obese by the third decade. Treatment programmes should be developed to appeal to women with limited education since the highest prevalence of this problem occurs among those with no education or primary education. Development of both prevention and treatment programmes would require further study of women's knowledge, attitudes, and practices in several areas, including physical activity, diet, and body image.
The prevalence of obesity among this population was higher than that reported for women in Europe (12), the USA (13), and developing countries (14), if age group-specific rates were considered, and the only groups reported to have comparable rates were U.S. black women (13), Kuwaiti women, and some Pacific Islanders (14). It was also higher than the 28% prevalence among females reported in a population-based survey of the same age groups in the same general area of UAE in 1989-1990 (3) but the rate (21%) in 20-29-year olds was lower than that reported in 1996 in a study of female students (30%) of UAE university (16). Unfortunately, the definition of obesity in the latter study was broader than that used in this study so direct comparisons are impossible. The prevalence of overweight and obesity among the 20-29-year old women was much higher (47%) than that reported in a survey of female UAE university student citizens aged 18-30 years in 1993 (28.4%) (15). An increasing prevalence of obesity over time has been reported in many other countries (12,13).
The prevalence of obesity, as defined by BMI, declined among older and postmenopausal women, while the prevalence of high body fat content (defined as either [greater than or equal to] 35% or [greater than or equal to] 40%) did not decline. This supports other evidence that BMI tends to underestimate fatness as age increases, especially among women (12, 17,18). This may be because of the decline in muscle and bone mass with age. Since bioelectric impedance is felt to be a more accurate measure of body fatness than BMI, all women aged over 30 years in this population should be considered at high risk of obesity.
The change with age in prevalence of obesity, as defined by BMI, is similar to that observed among U.S. women (13). The peak prevalence among U.S. women, however, is reached in the 7th decade, while the peak prevalence among the Al Ain women is reached in the 5th decade. The association between age and obesity was so strong that the risk of obesity (defined as [greater than or equal to] 35% of fat), as determined by multivariate logistic regression, increased by 5% with every one year increase in age.
The associations between obesity and lower education and lower physical activity have been reported elsewhere (12).
Other than the high prevalence of obesity, the study subjects were generally healthy. They had a very low prevalence of smoking, and this does not appear to be changing over time as it is similar to the rate reported in the 1995 Family Health Survey (4). Most women reported good health and did not report significant chronic diseases, reported sufficient activity to meet levels recommended by an expert panel (9), and got sufficient sunlight exposure despite their customary dress. Most pregnant women were also in good health and took multivitamin supplements, possibly because these were prescribed free of charge to citizens at the primary healthcare centres. Most younger women were well-educated and physically active. Unfortunately, menopausal women did not share such good health and reported significant deficiencies in lifestyle and health status. Therefore, specific health-promotion programmes aimed at this group of women are recommended.
Since, in this population, age and education were highly correlated and also probably correlated with other unmeasured variables, such as socioeconomic status and cultural and social heritage in this rapidly developing society, it is not possible to elucidate the reasons for associations between risk factors, such as obesity and exposures, such as education and age. They are, however, real and useful for public-health planning of prevention and treatment programmes.
The mean age (13.2 years) at menarche in this population was significantly (p<0.05) higher than that reported in Britain (12.9 years) and the U.S.A. (12.9 years for whites and 12.1 years for blacks) (19,20). It was, however, significantly (p<0.05) younger than 13.8 years reported in a survey of all UAE women (both citizens and non-citizens) done in 1996-1997 (6). The rate of oral contraceptive use was lower than that reported in a recent UAE survey of married women (4). This is to be expected as many women in this survey were unmarried. Fertility of postmenopausal women was high (5.2) but lower than that reported in the UAE Family Health Survey for urban women aged over 40 years (4). This survey was conducted in 1995, and the fertility may have fallen in the intervening six years.
This study has several limitations. It was a cross-sectional survey and, as such, can only identify associations but cannot ascribe causality. In addition, many variables collected were self-reported by women and subject to reporting error. However, weight, height, and bioelectric impedance were measured, making the identification of obesity not subject to this weakness. The subjects were randomly selected from the citizen population of Al Ain city. Thus, findings can only be extrapolated to this population. Despite these limitations, the high prevalence of obesity is very likely to be a valid finding and warrants public-health action.
The study was conducted with the financial support of a United Arab Emirates University Year 2000 research grant and carried out in Al Ain, United Arab Emirates. We wish to acknowledge the contribution to the study by the female UAE citizens of Al Ain, without whose willing cooperation the study would not have been possible. We would also like to acknowledge the dedication of nurses who collected data.
(1.) Preface. In: Halperin W, Baker EL. Public health surveillance. New York: Van Nostrand Reinhold, 1992:xv-xvi.
(2.) Bennett S, Woods T, Liyanage WM, Smith DL. A simplified general method for cluster-sample surveys of health in developing countries. World Health Stat Q 1991;44:98-106.
(3.) El Mugamer IT, Ali Zayat AS, Hossain MM, Pugh RNH. Diabetes, obesity and hypertension in urban and rural people of Bedouin origin in the United Arab Emirates. J Trop Med Hyg 1995;98:407-15.
(4.) Fikri M, Farid SM. United Arab Emirates family health survey 1995. Abu Dhabi: Ministry of Health, 2000.
(5.) United Arab Emirates. Ministry of Health. Reproductive patterns and child survival in the United Arab Emirates. Abu Dhabi: Ministry of Health, 1996.
(6.) Rizk DEE, Bener A, Ezimokhai M, Hassan MY, Micallef R. The age and symptomatology of natural menopause among United Arab Emirates women. Maturitas 1998;29:198-202.
(7.) Specker BL, Valanis B, Hertzberg V, Edwards N, Tsang RC. Sunshine exposure and serum 25-hydroxyvitamin D concentrations in exclusively breast-fed infants. J Pediatr 1985;107:372-6.
(8.) Dallosso HM, Morgan K, Bassey EJ, Ebrahim SB, Fantem PH, Arie THD. Levels of customary physical activity among the old and the very old living at home. J Epidemiol Community Health 1988;42:121-7.
(9.) Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C et al. Physical activity and public health: a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. J Am Med Assoc 1995;273:402-7.
(10.) NHLBI Obesity Education Initiative Expert Panel on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults. Identification, evaluation, and treatment of overweight and obesity in adults: the practical guide. Bethesda, MD: National Institutes of Health, 2000:10. (NIH publication no. 00-4084).
(11.) Biodynamics. Biodynamics model 310e body composition analyzer: users guide. 6th ed. Seattle: Biodynamics Corporation.
(12.) Seidel JC, Flegal KM. Assessing obesity: classification and epidemiology. Br Med Bull 1997;53:238-52.
(13.) Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and trends in obesity among US adults, 1999-2000. J Am Med Assoc 2002;288:1723-7.
(14.) Martorell R, Khan LK, Hughes ML, Grummer-Strawn LM. Obesity in women from developing countries. Eur J Clin Nutr 2000;54:247-52.
(15.) Musaiger AO, Radwan HM. Social and dietary factors associated with obesity in university female students in United Arab Emirates. J R Soc Health 1995;115:96-9.
(16.) Amine EK, Samy M. Obesity among female university students in the United Arab Emirates. J R Soc Health 1996;116:91-6.
(17.) Gallagher D, Visser M, Sepulveda D, Pierson RN, Harris T, Heymsfield SB. How useful is body mass index for comparison of body fatness across age, sex, and ethnic groups? Am J Epidemiol 1996;143:228-39.
(18.) Roubenoff R, Dallal GE, Wilson PWF. Predicting body fatness: the body mass index vs estimation by bioelectrical impedance. Am J Public Health 1995;85:726-8.
(19.) Whincup PH, Gilg JA, Odoki K, Taylor SJC, Cook DG. Age of menarche in contemporary British teenagers: survey of girls born between 1982 and 1986. Br Med J 2001;322:1095-6.
(20.) Herman-Giddens ME, Slora EJ, Wasserman RC, Bourdony CJ, Bhapkar MV, Koch GG et al. Secondary sexual characteristics and menses in young girls seen in office practice: a study from the pediatric research in office settings network. Pediatrics 1997;99:505-12.
Anne O. Carter , Hussain F. Saadi , Richard L. Reed , and Earl V. Dunn 
 School of Clinical Medicine and Research (Formerly with Department of Community Medicine, United Arab Emirates University), University of the West Indies, Cave Hill, Barbados,  Department of Internal Medicine, and  Department of Family Medicine, Faculty of Medicine and Health Sciences, Al Ain, United Arab Emirates University, United Arab Emirates
Correspondence and reprint requests should be addressed to:
Dr. Anne O. Carter
School of Clinical Medicine and Research
University of the West Indies
Cave Hill, Barbados
Email: firstname.lastname@example.org or email@example.com
Table 1. Characteristics of study subjects Proportion No. of (%) (95% subjects confidence with valid Characteristics interval) responses Citizenship UAE by birth 79 (75-82) 421 UAE by marriage 21 (18-25) 111 Menopausal status Postmenopausal 17 (14-20) 87 Premenopausal 84 (81-87) 439 Pregnancy status Pregnant 7 (5-10) 39 Not pregnant 93 (90-95) 489 Cigarette smoking Current smoker 2 (1-4) 8 Current non-smoker 98 (96-99) 485 Use of oral contraceptives Ever used 21 (20-27) 109 Never used 79 (75-82) 403 Perceived health status Good 73 (69-77) 385 Fair 22 (19-26) 118 Poor 5 (3-7) 26 Chronic illness Present 28 (24-32) 149 Absent 71 (67-75) 382 Percentage of body fat [greater than or equal to] 35 45 (40-50) 200 <35 55 (50-60) 247 [greater than or equal to] 40 21 (17-25) 94 <40 79 (75-83) 353 Sunlight exposure [greater than or equal to] 15 minutes/day 76 (72-79) 395 <15 minutes/day 24 (20-28) 122 Take vitamin supplements Yes 41 (37-45) 205 No 59 (55-63) 298 Take vitamin supplements (pregnant women) Yes 74 (58-85) 28 No 26 (15-42) 10 Physical activity Above minimal recommendations (9) 84 (81-87) 444 Below minimal recommendations (9) 16 (12-18) 85 Table 2. Proportions (expressed as %) of population and various subgroups of population within BMI-defined weight categories Proportion Proportion with ideal Proportion underweight weight BMI overweight Characteristics BMI <18.5 18.5-24.9 BMI 25-29.9 Total population (95% confidence 8 31 27 interval) (6-11) (27-35) (23-31) Age (years) * 20-29 12 41 26 30-39 3 15 33 40-49 3 14 23 50-59 2 26 28 60-69 4 30 30 70-79 4 38 33 Education * None 5 24 27 Primary 4 16 29 Secondary 12 29 23 Post-secondary 7 45 30 Physical activity score quartile * First 11 25 27 Second 6 24 26 Third 9 35 24 Fourth 5 36 33 Menstrual periods Regular 7 32 28 Irregular 12 29 21 Menopausal status Postmenopausal 4 31 35 Premenopausal 8 31 26 Perceived health status * Good 5 31 29 Fair 10 29 22 Poor 27 27 27 Chronic illness present * Yes 6 22 26 No 8 34 28 Proportion obese BMI No. (%) of [greater than subjects or equal with valid Characteristics to] 30 responses Total population (95% confidence 35 521 interval) (31-39) (100) Age (years) * 20-29 21 266 (51) 30-39 49 88 (17) 40-49 60 73 (14) 50-59 44 43 (8) 60-69 37 27 (5) 70-79 25 24 (5) Education * None 45 164 (32) Primary 51 55 (11) Secondary 36 145 (28) Post-secondary 17 148 (29) Physical activity score quartile * First 37 131 (25 Second 44 127 (25) Third 31 128 (25) Fourth 26 130 (25) Menstrual periods Regular 34 315 (74) Irregular 38 112 (26) Menopausal status Postmenopausal 31 85 (17) Premenopausal 35 427 (83) Perceived health status * Good 34 373 (72) Fair 38 116 (23) Poor 19 25 (5) Chronic illness present * Yes 47 144 (28) No 30 373 (72) * The association between BMI category and this characteristics is significant (p<0.05) by chi-square or Fisher's exact test if numbers were small BMI = Body mass index Table 3. Proportions (expressed as %) of subgroups of population defined by body fat content within BMI-defined weight categories Proportion Proportion % body fat by Proportion with ideal overweight bioelectric underweight weight BMI BMI impedance BMI <18.5 18.5-24.9 25-29.9 <35 11 45 36 [greater than or equal to] 35 2 12 18 <40 8 35 32 [greater than or equal to] 40 2 10 13 No. (%) of % body fat by Proportion subjects bioelectric obese with valid impedance BMI >30 responses <35 9 241 (55) [greater than or equal to] 35 68 197 (45) <40 25 344 (79) [greater than or equal to] 40 75 94 (21) BMI = Body mass index Table 4. Proportions (expressed as %) of premenopausal and postmenopausal women with various characteristics Proportion of Proportion of postmenopausal premenopausal women with women with Characteristics characteristics characteristics Citizenship By birth 82 79 By marriage 18 21 Perceived health Status * Good 37 80 Fair 44 17 Poor 19 2 Chronic illness * Present 61 21 Absent 39 79 Percentage of body fat * <35 34 60 [greater than or equal to] 35 66 40 <40 60 83 [greater than or equal to] 40 40 17 Sunlight exposure (minutes/day) [greater than or equal to] 15 81 75 <15 19 25 Vitamin supplements * Yes 21 44 No 79 56 Physical activity quartile * First 60 18 Second 17 26 Third 17 27 Fourth 6 29 No. (%) of subjects with valid Characteristics responses Citizenship By birth 415 (79) By marriage 108 (21) Perceived health Status * Good 381 (73) Fair 113 (22) Poor 26 (5) Chronic illness * Present 143 (27) Absent 379 (73) Percentage of body fat * <35 245 (56) [greater than or equal to] 35 194 (44) <40 347 (79) [greater than or equal to] 40 92 (21) Sunlight exposure (minutes/day) [greater than or equal to] 15 388 (76) <15 121 (24) Vitamin supplements * Yes 202 (41) No 293 (59) Physical activity quartile * First 130 (25) Second 130 (25) Third 130 (25) Fourth 130 (25) * The association between menopausal status and this characteristics is significant (p<0.05) by chi-square or Fisher's exact test if numbers were small…