Tobacco Documents Online

922 B14Z ~ZZ67 92Z ARTA STROKE 97 (C)AMER HEART A$SOC TX Socioeconomic Level, Sedentary Lifestyle, and Wine Consumption as Possible Explanations for Geographic Distribution of Cerebrovascular Disease Mortality in Spain Fernando Rodriguez Artalejo, MD, PhD; Pilar Guallar-Castill6n, MD; Juan Luis Gutirrrez-Fisac, MD, PhD; Jos6 Ram6n Banegas, MD, PhD; Juan del Rey Calero, MD, PhD Background and Purpose The geographic distribution of cerebrovascular disease (CVD) mortality in Spain spans a wide range, from provinces where mortality is low (70/100 000) and close to that of the United States and other Anglo-Saxon countries, to others where mortality is high (180/100 000) and more akin to that of Portugal and the Mediterranean and central European countries. This report seeks to identify the socioeconomic and lifestyle factors that most contribute to the geographic pattern of CVD mortality in Spain. Methods We performed a study using data collected at a provincial level. Mortality data were taken from official vital statistics, and data on risk factors were obtained from surveys of representative large Spanish population samples. Correla- tion and multiple linear regression analyses were performed on standardized CVD mortality ratios and potential determinants of mortality for the period 1989 to 1993. Results CVD mortality, unemployment and illiteracy rates, blond cigarette smoking, and sedentary lifestyle proved sub- stantially higher in the south and east (Mediterranean coast) of Spain. Saturated fatty acid intake and wine consumption were both lower in these regions, however. Illiteracy, wine consump- tion, sedentary lifestyle, high blood pressure, blond cigarette smoking, prevalence of diabetes, and body mass index >30 explained 59% of the variation in CVD mortality. Only illiter- acy, sedentary lifestyle, and wine consumption registered a statistically significant relationship (P<.05) with CVD mortal- ity. Whereas lower consumption of wine showed a negative association with CVD mortality, higher consumption revealed a positive association. Conclusions Socioeconomic level, as measured by illiteracy, sedentary lifestyle, and wine consumption, may partly explain the higher CVD mortality registered for regions situated in the south and east of Spain. (Stroke. 1996;27:922-928.) Key Words • epidemiology • geography • mortality • Spain In Spain, as with many developed countries,~,2 CVD mortality has progressively declined from 1901 to the present. Nevertheless, CVD was the leading cause of death in Spain in 1993, registering a mortality rate of 107/100 000 inhabitants,3 placing Spain in a middle-ranking position among developed countries. Within the country, however, the geographic distribution of CVD mortality spans a wide range, from provinces where mortality is low (70/100 000) and close to that of the United States and other Anglo-Saxon countries, to others where mortality is high (180/100 000) and more akin to that of Portugal and the Mediterranean and central European countries.4 This variation in CVD mortality represents an important potential for preven- Received November 19, 1996; final revision received January 21, 1997; accepted February 14, 1997. From the Department of Preventive Medicine and Public Health, University of the Basque Country (F.R.A.), Vitoria; De- partment of Preventive Medicine and Public Health, Autonomous University of Madrid (F.R.A., P.G.C., J.L.G.-F., J.R.B., J. del R.C.); Directorate General for Public Health, Ministry of Health (J.LG-F.); and Cardiovascular Epidemiology Unit, National Cen- ter for Epidemiology, Institute of Public Health "Carlos III," Ministry of Health (J.R.B.), Madrid, Spain. Correspondence to Dr Fernando Rodriguez Artalejo, Departa- mento de Medicina Preventiva y Salud P~blica, Universidad Au- t6noma de Madrid, Avda, Axzobispo Morcillo s/n, 28029 Madrid, Spain. © 1997 American Heart Association, Inc. tion provided that the modifiable factors responsible can be identified. This report therefore examines the provin- cial distribution of CVD mortality in Spain plus known socioeconomic and lifestyle risk factors to identify those that most greatly contribute to the geographic pattern of mortality due to this disease. Subjects and Methods The following information was obtained for the 50 provinces of Spain: CVD mortality (ICD-9 codes 430 to 438) data in persons aged 45 to 79 years were gathered from Spanish vital statistics,~ and SMRs were computed at a provincial level for the period 1989 to 1993.6 National mortality by sex and 5-year age groups was used as standard. Information on consumption of foodstuffs, nutrients, and tobacco was taken from the 1980 to 1981 Househokt Budget Survey (Encuesta de Presupuestos Familiares), conducted by the National Statistics Office (Instituto Nacional de Estadis- tica) and National Nutrition Institute (Instituto Nacional de Nutrici6n), and based on a representative Spanish population sample of 25 000 families.7 The survey estimated food and tobacco intake on the basis of the amounts purchased by the families surveyed. Only food consumed at home was included. Food quantities were converted into nutrients by application of standard food composition tables. Salt intake was obtained from a similar survey performed in 1991.8 Data on the preva- lence of hypercholesterolemia; high blood pressure, sedentary lifestyle, diabetes, and BMI >-30 were taken from the 1993 National Health Survey (Encuesta Nacional de Salud de Es- pafia).9 This was an interview-based survey performed by the

behavior demonstrates the association between knowl- edge and practice, and our results are consistent with this literature. A final limitation is that we did not ask the patients to provide detailed information about the magnitude of their stroke risk but asked respondents only whether they were "at risk for stroke." Many health behavior models postulate that one component of the decision to adopt stroke prevention strategies is the perceived prob- ability of the adverse outcome. Some of the patients .responding affirmatively to the question about risk for stroke might still have underestimated the magnitude of this risk and/or might have placed this risk below their "thresholds for action." In conclusion, research on health behavior indicates that patients who are aware of their increased risk for stroke are more likely to begin stroke prevention regi- mens and are more likely to achieve better compliance with these regimens once they begin. Unfortunately, many high-risk patients, including over one half of persons with minor stroke and one third of persons with TIA, are unaware of their increased risk for stroke. Making patients better aware of their increased risk is a first step toward improving stroke prevention practice, which in turn is a step toward reducing the community burden of stroke. Healthcare providers play a crucial role in communicating information about stroke risk. Acknowledgments This work was performed as part of the Stroke Prevention Patient Outcomes Research Team (PORT) and was funded Samsa et al Stroke Risk Knowledge 921 through contract 282-91-0028 from the US Agency for Health Care Policy and Research. We would like to thank Joe Lipscomb, PhD, John Paul, PhD, Pat Venus, PhD, Morris Weinberger, PhD, and David Witter, BA, for their help in the design and execution of the study and to thank Annette Jurgelski for her editorial assistance. References 1. Kreuter MW, Strecher VJ. Changing inaccurate perceptions of health risk: results from a randomized trial. Health PsychoL 1995; 14:56-63. 2. Janz NK, Becker MH. The health belief model: a decade later. Health Educ Q. I984;11:1-47. 3. Tversky A, Kahneman D. Evidential impact of base rates. In: Kahneman D, Slovic P, Tversk-y A, eds. Judgements Under Uncer-. tain~y: Heuristics and Biases. New York, NY: Cambridge University Press; 1982. 4. Weinstein ND. Testing four competing theories of health- protective behavior. Health PsychoL 1993;12:324-333. 5. Matchar DB, McCrory DC, Barnett HJM, Feussner JR. Medical treatment for stroke prevention. Ann Intern Med. 1994;121:54-55. 6. Fried LP, Borhani N, Enright P, Furberg CD, Gardin JM, Kronmal RA, Kuller LH, Manolio TA, Mittelmark MB, Newman A, O'Lea~ DH, Psaty B, Rautaharju P, Tracy RP, Weiler PG. The Cardiovas- cular Health Study: design and rationale. Ann Epidemiol. 1991;1: 263-276. 7. McHomey CA, Ware JE, Raczel AE. The MOS 36-item health survey (SF-36): psychometric and clinical tests of validity in mea- suring physical and mental health outcomes. Med Care. 1993;31: 247-263. 8. Mahoney FI, Barthel DW. Functional evaluation: the Barthel index. Md State Med J. 1965;14:61-65. 9. Burnam MA, Wells KB, Leake B, Landsverk J. Development of a brief screening instrument for detecting depressive disorders. Med Care. 1988;26:775-789. 10. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic Research: Principles and Quantitative Methods. Belmont, Calif: Lifetime Learning Publications; 1982. 11. Samsa G, Matchar DB, Bonito A, Duncan P, Lipscomb J, Enarson C, Witter D, Venus P, Paul J, Weinberger M. Quality of life with stroke: results from a nationally-diverse survey of persons at increased risk for stroke. J Gen Intern Med. 1996;11s1:60. Abstract.

Selected Abbreviations and Acronyms BMI = body mass index CVD = cerebrovascular disease ICD-9 = International Classification of Diseases, 9th Revision SMR = standardized mortality ratio Ministry of Health, covering a representative 21 000-person sample of the noninstitutionalized Spanish population older than 16 years. Information on hypercholesterolemia, high blood pressure, and diabetes was obtained by the following question: "Has your doctor told you that you are currently suffering from one of the following chronic conditions: high blood pressure, high cholesterol, or diabetes (high sugar)?" Information on sedentary lifestyle was obtained by the ques- tion: "What type of physical exercise do you do in your leisure time? Tell me which of these possibilities best describes the major part of your leisure-time activity: (a) I do no exercise at all. I spend most of my leisure time in a sedentary fashion (reading, watching television, going to the cinema, etc); (b) I do some occasional physical or sports activity (walking or riding a bicycle, gardening, easy gymnastics, etc); (c) I do regular physical activity several times a month (tennis, gymnastics, running, swimming, etc); or (d) I do physical training several times a week." Individuals who gave option (a) as their answer were classified as sedentary. BMI (weight [kilograms] divided by height [meters] squared) was calculated on the basis of self-reported information on height and weight, given in re- sponse to the following questions: "Could you tell me how much you weigh, without shoes and clothes on?" and "Could you tell me how tall you are without your shoes on?" Finally, information on illiteracy among the segment of the population older than 45 years was taken from the 1981 Population Census,~o and unemployment rates were obtained from the National Statistics Office's 1981 Active Population Survey (Encuesta de Poblaci6n Activa).it Pearson correlation coefficients were computed, and a mul- tiple linear regression analysis of the SMR was performed. Correlation and regression analyses were weighted by the proportion of CVD deaths for each province over total CVD deaths in Spain. Statistical analysis was performed with the use of the SAS software package.~2 $ Results CVD mortality in Spain showed a coefficient of vari- ation of 31.52%, with 24 of the country's 50 provinces registering an SMR significantly different (P<.05) from the national mean (SMR=100). CVD mortality was higher in provinces lying in the south and east (Medi- terranean coast) of the country (Fig 1). Wine consumption proved to be the risk factor for CVD with greatest provincial variability (Table 1). Al- cohol intake; consumption of beef, legumes, and blond cigarettes; and the percentage of illiteracy among the segment of the population older than 45 years registered coefficients of variation in excess of 40%. Whereas intake of saturated fatty acids and wine were lower in the southern provinces, blond cigarette smoking, sedentary lifestyle, unemployment, and illiteracy among those older than 45 years proved to be relatively higher. Prevalence of high blood pressure showed no defined geographic pattern (Figs 1 and 2). Blond cigarette smoking, prevalence of sedentary lifestyle, BMI ->30, and percentage of unemployment and illiteracy among the segment of the population older than 45 years correlated significantly (P<.05) and posi- Artalejo et al Geography of Stroke Mortality in Spain 923 tively with CVD mortality, while consumption of milk, meat, beef, eggs, wine, and saturated fats showed a negative correlation (Table 1). Salt intake, prevalence of hypercholesterolemia and high blood pressure, and the remaining variables registered correlations that failed to attain statistical significance. To ascertain the independent contribution of the above factors to CVD mortality, a multiple linear regres- sion model was constructed in which the dependent variable was the SMR for CVD. It was decided that saturated fats would not be included in the model, in the absence of unequivocal evidence in the literature of their association with CVD mortality.13 On the same grounds and by reason of their high correlation with saturated fats (Table 2), consumption of milk, meat, beef, and eggs were likewise not included in the model. The jobless rate was also excluded because it showed a marked correla- tion with illiteracy among the segment of the population older than 45 years (Table 2) and because there is evidence in the literature to show that its association with cardiovascular disease mortality is of a lower mag- nitude than that of educatibn34 Prevalence of high blood pressure and diabetes were included, however, because both are important risk factors for CVD,~3,15-17 despite the fact that they failed to attain statistical significance in the crude analysis. Taken together, the variables in the model explained 59% of the variation in CVD mortality across Spain (Table 3). Illiteracy in the segment of the population older than 45 years was the sole variable to show a significant and independent association with CVD mor- tality in Spain. The correlation coefficient for two vari- ables, ie, blond cigarette smoking and BMI -->30, regis- tered a sign contrary to that observed in the crude analysis, owing to both the high correlation between these variables and illiteracy among those older than 45 years and the stronger association between the latter variable and CVD mortality. Hence, and since part of the possible effect of socioeconomic status (as gauged by illiteracy) on CVD mortality is probably mediated by other variables, it was decided that a new model should be constructed excluding illiteracy (Table 4). The vari- ables in the model jointly served to explain 40% of the variation in CVD mortality in Spain. Only consumption of wine and sedentary lifestyle registered a statistically significant relationship (P<.05) with CVD mortality in Spain. Whereas lower consumption of wine showed a negative association with CVD mortality, higher con- sumption (quadratic term in the model) revealed a positive association. Those provinces with a greater prevalence of a sedentary lifestyle registered a higher level of CVD mortality. Discussion This study suggests that socioeconomic level, as ap- proximated by the percentage of illiteracy among the segment of the population older than 45 years, is the variable that best explains the geographic distribution of CVD mortality in Spain. The lower wine consumption and more accentuated sedentary lifestyle of the southern and eastern regions of Spain may also explain the higher CVD mortality in these regions of the country. It is known that CVD mortality exhibits a considerable variation according to social class (the higher the class, the lower the CVD mortality).17 This association has

924 Stroke Vol 28, No 5 May 1997 CEREBROVASCULAR DISEASE MORTALITY SATURATED FATTY-ACID INTAKE BLOND CIGARETTE SMOKING SELF-REPORTED HIGH BLOOD PRESSURE pack/person/year percentage FIG 1. CVD mortality (1989-1993), saturated fatty acid intake (1980-1981), blond cigarette smoking (1980-1981), and prevalence of self-reported high blood pressure (1993) in Spain. Quintiles of provincial distribution are shown. The upper section of each map represents the north of Spain. also been observed at an ecological level, both in the international literature18-2° and in individual studies un- dertaken in Spain.21.= The association could be medi- ated by genetic, physiological, behavioral, or environ- mental variables,t7 Our results suggest that the effect of socioeconomic status on geographic, distribution of CVD mortality could in part be related to consumption of wine and sedentary lifestyle. In Spain, there is a positive dose-response relationship between the frequency of moderate alcohol consumption and socioeconomic sta- tus as measured by the level of education attained, while the dose-response relationship is negative for excessive consumption of alcohol and level of education at- tained.2~ Furthermore, the degree of physical leisure activity tends to be greater among persons at the highest socioeconomic level in Spain24 as well as in other countries.~ The relationship between alcohol consumption and CVD is not clear in the literature on the subject.13 In part, this is due to the fact that many cohort and case-control studies have included lifetime and newer teetotalers in the same group, when the effect of alcohol may be different between the two categories. Ecological studies published to date have failed to furnish consis- tent results. St Leger et alz6 were thus unable to detect a linear association between alcohol consumption and CVD mortality in 18 (mostly European) countries. In contrast, Ueshima et alz7 found quite a strong positive association on using comparable data from 46 prefec- tures in Japan and adjusting for salt intake and a number of socioeconomic factors. Similarly, Sasaki et al~ re- ported a positive association between alcohol consump- tion and CVD mortality in a correlational population- based study embracing 17 countries. Although the relationship between physical activity and risk of CVD has not been extensively studied, the information avail- able is fairly consistent and points to an inverse and significant relationship between the two variables,t~ Nevertheless, assessment of the results of this study calls for a certain measure of prudence, inasmuch as

Artalejo et al Geography of Stroke Mortality in Spain TABLE 1. Correlations Between CVD Mortality (1989-1993) and Socioeconomic and Lifestyle Risk Factors in 50 Provinces of Spain 925 Coefficient of Maximum Minimum Correlation Variable Units Mean Variation, %* Value Value Coefficien~ />~ CVD SMR 100 31.52 144.00 58.00 Milk dL/person/d 3.38 27.00 5,56 0.90 -.40 .003 Dairy products g/person/d 39.68 36.16 109.33 19.55 .22 .110 Meat g/person/d 181.39 16.36 263.18 87.65 -.35 .011 Beef g/person/d 28.53 56.22 79.60 3.84 -.58 ,000 Pork g/person/d 28.17 37.98 71.65 8.03 -.07 .626 Chicken g/person/d 61.11 19.91 82.99 25.26 .27 .051 Fish g/person/d 71.29 27.26 160.83 46.67 -.13 ,360 Sausage meats g/person/d 33.64 25.77 51.94 13.51 .01 .963 Eggs g/person/d 44.50 17.33 67.53 33.45 -.42 .001 Fruit g/person/d 279.39 12.31 324.18 173.09 -.15 .267 Greens g/person/d 396.07 30.53 960.44 241.78 -.07 .599 Legumes g/person/d 24.09 41.92 52.60 8.31 -.03 ,828 Alcohol dL/person/d 1.73 50.86 5.48 0.51 -.13 .353 Wine dL/person/d 1,18 77.11 4.91 0.14 -.26 .061 Salt g/person/d 7.06 70.64 14.46 2.90 -.04 .772 Oils dL/person/d 0.77 14.28 1.04 0.50 .07 .611 Saturated fats g/person/d 36.38 12.20 53.80 29.20 -.44 .001 Monounsaturated fats g/person/d 60.85 12.45 84.60 38,50 - .08 .552 Polyunsaturated fats g/person/d 21.17 19.17 32.60 14.40 -.12 .382 Lipids g/person/d 145.18 11.16 199.13 107.45 -.20 .152 Proteins g/person/d 97,17 10.69 140,76 84.44 -.26 .064 Kilocalodes kcal/person/d 3069.00 12.15 4722.87 2620.41 -.09 .524 Black cigarettes PacWperson/y 56.76 13.09 76.79 3.21 -.10 .452 Blond cigarettes PacPJperson/y 18.49 41.75 32.63 2.84 .27 .057 Hypercholesterolemia % 8.05 26.70 17,80 1.70 .02 .836 Diabetes % 4.24 38.44 4.24 1.63 .24 .082 High blood pressure % 11.15 28.34 21.50 4.00 .24 .088 Sedentary lifestyle % 56.00 14.92 84.40 33.40 .41 .002 Illiteracy (>45 y) % 14.20 51.19 29.55 2.45 .75 .000 BMI =>30 % 9.01 33.85 16.00 2.90 .31 .027 Smokers % 31.93 9,33 38.80 23.30 .09 .496 Unemployment % 1.4.26 34.78 23.70 4.40 ,31 .021 *SD/mean. ~'Pearson correlation coefficient between CVD mortality and its dsk factors. :~Stafistical significance (two-sided test) of the Pearson coefficient. neither the study design nor data can be considered optimum. First, overall CVD mortality data were used, without any distinction as to whether deaths were ische- mic or hemorrhagic in origin, when the risk factors may be partially different between the two.13,~9 Indeed, some studies specifically suggest that the dose-response rela- tionship between alcohol consumption and CVD mor- tality would be positive and continuous in the case of hemorrhagic origin yet d-shaped in the case of ischemic origin.3° It should be noted, however, that most CVD deaths in Spain are ischemic in origin.1,4 Second, this was an ecological study. Our results, which are consistent with those from other countries such as the United States,31 suggest that distribution of known risk factors for CVD--prevalence of high blood pressure and smoking, in particular--does not explain the spatial distribution of the disease in Spain. However, this does not mean that high blood pressure and smok- ing are entirely devoid of influence on the risk of CVD among the Spanish population. This influence depends on the level of analysis)2 There is ample evidence that high blood pressure and smoking raise the risk of CVD at an individual level.13a7 This report in no way seeks to extend its inferences to levels of aggregation beyond that of the provincial geography of Spain. Third, data on socioeconomic and lifestyle variables were drawn from a number of years with the ensuing possibility that, in several cases, such data may not allow for an adequate induction time for their effect on CVD. However, there is evidence of a degree of temporal stability in the data, since the south of Spain has traditionally encompassed less socially favored regions, and changes in lifestyle habits (eg, diet, physical activity) tend to require long periods of time. Fourth, some data on lifestyle habits and biological variables (high blood pressure, hypercholesterolemia, diabetes, weight, and height) were self-reported and not accompanied by objective measurements. Several vari- ables might have been underassessed.3~-3~ Nevertheless, there was no evidence that the error might be different across provinces. Finally, no account was taken of the existence of interprovincial differences in the availability and acces- sibility of high-technology healthcare services in Spain.36 Advances made during recent years in the treatment of CVD have been modest, so that any contribution thereof to geographic distribution of CVD would necessarily be small. However, an ecological study performed in dis- tricts of Catalonia (northwestern Spain) has shown a

926 Stroke Vol 28, No 5 May 1997 WINE CONSUMPTION UNEMPLOYMENT dlfperson/day percentage ILLITERACY SEDENTARY LIFESTYLE percentage FIG 2. Wine consumption (1980-1981), percentage of unemployment (1981), percentage of illiteracy among segment of the population older than 45 years (1981), and percentage of self-reported sedentary lifestyle (1993) in Spain. Quintiles of provincial distribution are shown. The upper section of each map represents the north of Spain. TABLE 2. Correlations Between Principal Risk Factors for CVD in Spain Saturated Blond Sedentary BMI Illiteracy Milk Meat Beef Eggs Wine Fats Cigarettes Diabetes HT Lifestyle _>30 Unemployment (>45 y) " Milk 1.00" Meat 0.21 1.00" Beef 0.52* 0.18 1.00" Eggs 0.62* 0.13 0.27 1.00" Wine 0.39* 0.51" 0.57* 0.08 1.00" Saturated fats 0.71" 0.43* 0.50* 0.55* 0.59* 1.00" Blond cigarettes 0.40* -0.48* -0.32* -0.30* -0.64* -0.54* 1.00" Diabetes -0.08 -0.30* -0.17 -0.17 -0.18 -0.05 0.26 HT 0.02 0.03 -0.20 -0.12 0.06 0.16 -0.07 Sedentary lifestyle 0.00 -0.16 -0.28 -0.13 -0.02 0.02 -0.02 BMI :>30 -0.02 -0.14 -0.17 0.04 0.13 0.11 0.00 Unemployment -0.24 -0.65* -0.30* -0.08 -0.71" -0.54* 0.62* Illiteracy (>45 y) -0.33* -0.45* -0.66* -0.33* -0.33* -0.27* 0.33* 1.00" 0.30* 1.00" 0.100.20 1.00" 0.34* 0.16 0.22 1.00" 0.08 -0.20 0.08 -~0.00 0.48* 0.27* 0.48* 0.52* 1.00" 0.41" 1.00" HT indicates hypertension. *P<.05.

TABLE 3. Multiple Linear Regression Analysis o~ SMR for CVD in Spain in Model Including the Variable Percentage of Illiteracy in Those >45 y Partial Standardized Correlation Variable /3 /~* Coefficient Illiteracy (>45 y) 2.20 0,71 .57 .001 HT 0.14 0.02 .03 .861 Sedentary lifestyle 0.22 0,08 .11 .487 Wine -8.80 -0,36 -.12 .432 Wine~ 1.97 0.37 .14 .373 Blond cigarettes -0.03 -0.01 -.00 .950 Diabetes 0.46 0.03 " ,05 .774 BMI >30 -0.96 -0.13 -.15 .334 R~=.59 HT indicates hypertension. */3 (Sx/Sy), where S is standard deviation, x the independent variable, and y the SMR for CVD. 1"Two-sided test. * negative relationship between degree of blood pressure control and CVD mortality.37 From the standpoint of CVD prevention, our findings are relevant for a number of reasons: first, they stress the importance of reducing socioeconomic inequalities in the distribution of risk factors, such as excessive con- sumption of wine and sedentary lifestyle; second, the same factors responsible for the geographic distribution of CVD are those responsible for that of ischemic heart disease. In an earlier study, we observed that ischemic heart disease mortality was likewise higher in the south and east of Spain and that socioeconomic level and wine consumption were associated with a higher ischemic heart disease mortality in regions along the Mediterra- nean seaboard.38 There is therefore a reasonable likeli- hood that the same prevention policies might serve to reduce mortality from both CVD and ischemic heart disease, which are the first and second leading causes of death in Spain, respectively. Acknowledgments This study was supported in part by a research grant from Merck Sharp and Dohme. We thank Michael Benedict for translating this report into English. TABLE 4. Multiple Linear Regression Analysis of the SMR for CVD in Spain in Model Excluding the Variable Percentage of Illiteracy in Those >45 y Partial Standardized Correlation Variable /~ /~* Coefficient P~ HT 0.88 0.12 .14 .350 Sedentary lifestyle 0.76 0.28 .33 .031 Wine -27.11 -1.10 -.32 .034 Wine2 4.78 0.88 .28 .065 Blond cigarettes 0.13 0.05 .04 .794 Diabetes 0.08 0.01 .00 .968 BMI >30" 1.13 0.15 .17 .275 R2=.40 HT indicates hypertension. */3 (S~/Sy), where S is standard deviation, x the independent variable, and y the SMR for CVD. 1"Two-sided test. Artalejo et al Geography of Stroke Mortality in Spain 927 References 1. Barrado-Lanzarote M J, de Pedro-Cuesta J, Almazdn Isla J. Stroke mortality in Spain 1901-1986. Neuroepidemiology. 1993;12:148-157. 2. Thorn TJ, Epstein FH, Feldman J, Leaverton PE, Wolz M. Total Mortality and Mortality From Heart Diseasg Cancer and Stroke From 1950 to 1987 in 27 Countries. Bethesda, Md: National Heart, Lung, and Blood Institute; 1992. Publication NIH 92-3088. 3. Instituto Nacional de Estadfstica. Defunciones Segtin la Causa de Muerte. 1993. Madrid, Spain: Instituto Nacional de Estadistica; 1996. 4. Barrado-Lanzarote M J, Almazfin-lsla J, Medrano-Albero MJ, de Pedro-Cuesta J. Spatial distribution of stroke mortality in Spain, 1975-1986. Neuroepidemiology. 1995;14:165-173. 5. Instituto Nacional de Estadfstica. Movimiento Natural de la Poblacirn Espahola: Defunciones Segt~n la Causa de Muerte: Ahos 1989-1993. Madrid, Spain: Instituto Nacional de Estadistica; 1992-1996. 6. Breslow NE, Day NE. Rates and rate standardization. In: Breslow NE, Day NE. Statistical Methods in Cancer Research: The Design and Analysis of Cohort Studies. Lyon, France: International Agency for Research on Cancer; 1987:48-81. 7. Instituto Nacional de Estadfstica. Encuesta de Presupuestos Familiares 1980-1981: Tomo V, I"y 2aparte. Madrid, Spain: lnstituto Nacional de Estadistica; 1985. 8. Instituto Nacional de Estadistica. Encuesta de Presupuestos Familiares 1990-1991: Volumen IL Madrid, Spain: Instituto Nacional de Estadfstica; 1994. 9. Ministerio de Sanidad y Consumo. Encuesta Nacional de Salud de Espa~qa 1993. Madrid, Spain: Ministerio de Sanidad y Consumo; 1995. 10. Instituto Nacional de Estadfstica. Censo de Poblaci6n de 1981; Tomo I, Volumen I: Resultados Nacionales: Caracteristicas de la Poblaci6n. Madrid, Spain: Instituto National de Estadistiea; 1985. 11. Instituto National de Estadfstiea. Encuesta de Poblaci6n Activa: A~o 1981. Madrid, Spain: Instituto Nacional de Estadistica; 1981. 12. SAS/STAT Guide for Personal Computers, Version 6.03. Cary, NC: SAS Institute; 1988. 13. Bronner LL, Kanter DS, Manson JE. Primary prevention of stroke. N Engl J Med. 1995;333:1392-1400. 14. Kaplan GA, Keil JE. Socioeconomic factors and cardiovascular disease: a review of the literature. Circulation. 1993;88:1973-1997. 15. MacMahoa S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, Godwin J, Dyer A, Stamler J. Blood pressure, stroke, and coronary heart disease, part 1: prolonged differences in blood pressure: prospective observational studies corrected for the regression dilution bias. Lancet. 1990;335:765-774. 16. Prospective Studies Collaboration. Cholesterol, diastolic blood pressure, and stroke: 13000 strokes in 450000 people in 45 pro- spective cohorts. Lancet. 1995;346:1647-1653. 17. Marmot MG, Poulter NR. Primary prevention of stroke. Lancet. 1992;339:344-347. 18. Siegel PZ~ Deeb LC, Wolfe LE, Wilcox D, Marks JS. Stroke mor- tality and its socioeconomic, racial and behavioral correlates in Florida. Public Health Rep. 1993;108:454-458. 19. Nayha S. Geographical variations in cardiovascular mortality in Finland, 1961-1985. Scand J Soc Med. 1989;40(suppl):1-48. 20. Mackenbach JP, Kunst AE, Looman CW. Cultural and economic determinants of geographical mortality patterns in the Netherlands. J Epidemiol Community Health. 1991;45:231-237. 21. Regidor E, de Marco S, Gutirrrez-Fisac JL, Rodrfguez C. Dife- rencias socioecon6micas en mortalidad en ocho provincias espafiolas. Med Clin (Barc). 1996;106:285-289. 22. Regidor E, Guti~rrez-Fisac JL, Rodrfguez C. Increased socio- economic differences in mortality in eight Spanish provinces. Soc Sci Med. 1995;41:801-807. 23. Regidor E, Gutirrrez-Fisac JL, Rodriguez C. Diferencias y Desigualdades en Salud en Espaha. Madrid, Spain: Dfaz de Santos; 1994. 24. Regidor E, Rodrfguez C, Gutirrrez-Fisac JL. Indicadores de Salud: Tercera Evaluaci6n en Espa~a del Programa Regional Europeo Salud para Todos. Madrid, Spain: Ministerio de Sanidad y Consumo; 1995:227-233. 25. Ford ES, Merritt RK, Health GW, Power KE, Washburn RA, Kriska A, Halle G. Physical activity behaviors in lower and higher socio- - economic status populations.Am J EpidemioL 1991;133:1246-1256. " 26. St. Leger AS, Cochrane ~ Moore F. Factors associated with cardiac mortality in developed countries with particular reference to the consumption of wine. Lancet. 1979;1:1017-1020.

928 Stroke Vol 28, No 5 May 1997 27. Ueshima H, Ohsaka T, Asakura S. Regional differences in stroke mortality and alcohol consumption in Japan. Stroke. 1986;17:19-24. 28. Sasaki S, Zhang XH, Kesteloot H. Dietary sodium, potassium, saturated fat, alcohol, and stroke mortality. Stroke. 1995;26:783-789. 29. Thrift AG, Donnan GA, McNeil J.f. Epidemiology of intracerebral hemorrhage. Epidemiol Rev. 1995;17:361-381. 30. Camargo CA Jr. Moderate alcohol consumption and stroke: the epidemiological evidence. Stroke. 1989;20:1611-1626. 31. Lanska DJ, Kuller LH. The geography of stroke mortality in the United States and the concept of a stroke belt. Stroke. 1995;26: 1145-1149. 32. Rose G. Sick individuals and sick populations. Int J Epidemiol. 1985;14:32-38. 33. Bowlin S J, Morrill BD, Nafziger AN, Jenkins PL, L~wis C, Pearson TA. Validity of cardiovascular disease risk factors assessed by telephone survey: the Behavioral Risk Factor Survey. J Clin Epi- demioL 1993;46:561-571. 34. Nieto Garcia FJ, Bush 2r-L, Kely PM. Body mass definitions of obesity: sensitivity and specificity using self-reported weight and height. Epidemiolog),. 1990;1:146-152. 35. Quiles Izquierdo J, Vioque J. Validez de los datos antropom6trieos declarados para la determinaei6n de la prevalencia de obesidad. Med Clin (Barc). 1996;106:725-729. 36. Aracil Fo Banegas JR, del Llano J, del Liano M, de la Mata I, Gol J, Gon~lez J, Villar F. Sistema Grdfico de Informaci6n Sanitaria en Espaha. Madrid, Spain: MSD; 1996. 37. Treserras R, Serra-Majem L, Canela J, Armario P, Par'dell H, Rue M, Salleras L Ecological association between hypertension and stroke in Catalonia (Spain): development and use of an ecological regression model. J Hum Hypertens. 1990;4:300-302. 38. Rodrfguez Artalejo F, Banegas JR, Garcfa Colmenero C, de! Rey Calero J. Lower consumption of wine and fish as a possible expla- nation for higher ischemic heart disease mortality in Spain's Med- iterranean region, lnt l Epidemiology. 1996;25:1196-1201. 0 4~ 0