Author + information
- Received July 11, 1997
- Revision received November 12, 1997
- Accepted December 17, 1997
- Published online March 15, 1998.
- Mario Ciruzzi, MDA,*,
- Palmira Pramparo, MDA,
- Osvaldo Esteban, MDA,
- Jorge Rozlosnik, MDA,
- Jorge Tartaglione, MDA,
- Blanca Abecasis, MDA,
- Jorge César, MDA,
- Jose De Rosa, MDA,
- Carlos Paterno, MDA,
- Herman Schargrodsky, MDA,
- for the Argentine FRICAS Investigators1
- ↵*Dr. Mario Ciruzzi, Sarmiento 3760 3B (1190), Buenos Aires, Argentina.
Objectives. We sought to study the relation between passive smoking at home and the risk of acute myocardial infarction (AMI).
Background. Previous epidemiologic studies have linked environmental tobacco smoke to an increased risk of coronary heart disease, but the evidence to support this view is not strong enough. To study this issue further, we analyzed the data from a case-control study conducted in Argentina between 1991 and 1994.
Methods. Case patients included 336 never-smokers with AMI. Control patients were 446 never-smokers admitted to the same network of hospitals with a wide spectrum of acute disorders unrelated to smoking or to known or suspected risk factors for AMI. Data on the smoking habits of the participants’ close relatives (spouse and children) were collected by trained interviewers using a structured questionnaire.
Results. Compared with subjects whose relatives had never smoked, the multivariate odds ratios for passive smokers, according to the smoking status of their relatives, were 1.68 (95% confidence interval [CI] 1.20 to 2.37) for one or more relatives who smoked; 1.59 (95% CI 0.85 to 2.96) for a spouse who smoked; 1.24 (95% CI 0.61 to 2.52) for a spouse who smoked 1 to 20 cigarettes/day; 4.03 (95% CI 0.99 to 16.32) for a spouse who smoked >20 cigarettes/day; and 1.80 (95% CI 1.20 to 2.68) for one or more children who smoked. There was a significant interaction between passive smoking and hypercholesterolemia (≥240 mg/dl), hypertension, diabetes and family history of MI.
Conclusions. In never-smokers, passive smoking at home appeared to be associated with the risk of AMI, and ∼14% of cases in men and 18% of cases in women in this Argentinian cohort are attributable to passive smoking.
The adverse effects of environmental tobacco smoke have received wide attention in recent years. It is derived from two sources: 15% from mainstream smoke, which is the smoke inhaled by the smoker, and 85% from sidestream smoke, which is the smoke emitted into the surrounding air from the burning cigarette between puffs . Qualitatively, both types of smoke contain similar components. The difference is in the greater amount of carbon monoxide, benzopyrene, ammonia and other carcinogenic compounds that sidestream smoke contains in relation to mainstream smoke. The principal mechanisms by which environmental tobacco smoke affects the cardiovascular system include the reduced ability of the blood to deliver oxygen to the myocardium and increased platelet activity [2, 3].
The clinical effect of these smoke-associated problems has been consistently studied in several epidemiologic studies. In healthy nonsmokers, environmental tobacco smoke has been linked to lung cancer and other diseases of respiratory function [1, 4, 5]. Moreover, in children of smoking parents, several studies have consistently demonstrated an increased frequency of respiratory problems . In reference to cardiovascular disease, which is the principal cause of death in industrialized countries, an increasing number of studies have shown a positive relation between passive smoking and the risk of coronary heart disease [6–17], but there is still uncertainty regarding this possible association. Although some studies reported relative risks obtained without appropriate control for the confounding effects of other major risk factors, other studies reported weak relations between passive smoking and cardiovascular disease and still others found relative risk that was not statistically significant. Moreover, the interaction between passive smoking and other risk factors for coronary heart disease has not been investigated.
To provide further information on this topic, this report presents findings from a large case-control study of myocardial infarction (MI) conducted in Argentina, whose design enabled us to control for several other major risk factors for coronary heart disease.
The present study is derived from the Factores de Riesgo Coronario en América del Sur (FRICAS) study, an ongoing case-control investigation that began in November 1991 of the risk factors of MI and that is based on a network of medical centers in 10 countries of South America. The Argentine branch finished the recruitment of patients in August 1994, including 1,060 cases with acute MI (AMI) and 1,071 controls, matched for age, gender and medical center. The present analysis is based on data collected from this Argentinian cohort and includes 336 nonsmoker cases and 446 nonsmoker control subjects who said they had never smoked.
A structured questionnaire was used by trained interviewers to obtain information on education, habits such alcohol, coffee and mate (infusion native of South America) consumption, frequency of consumption of a few selected dietary items, a few indicators of physical activity, self-reported weight and height, history of hypertension and family history of MI. The presence of diabetes was recorded if the subject had been diagnosed or treated as a diabetic by a doctor before admission to the medical center. Social class was stratified according the ownership of a house or car: low strata with no house or car; medium strata with a house or car; and high strata with a house and car. Participants were asked about the smoking status (nonsmoker, current smoker) of their close relatives (spouse and children). The number of cigarettes was assessed only for the spouse. No information was collected on the duration of smoking and the exposure to tobacco smoke on the job. On average, <4% of case and control patients refused to be interviewed. Furthermore, blood samples were taken by venipuncture as soon as possible after admission to the hospital, both for cases and controls to obtain total serum cholesterol levels.
1.1 Case Patients
Case patients included were 336 never-smokers who had been admitted for a first episode of AMI to 35 coronary care units located in Buenos Aires (capital and suburbs) and nine provinces of Argentina. Those who had a history of ischemic heart disease, rheumatic valvular disease, cardiomyopathy or cardiac surgery were excluded. They were eligible if they met the standard World Health Organization criteria for AMI, including pathologic Q waves with evolution, or any two of the following: a typical history of chest pain for at least 30 min, electrocardiographic (ECG) changes with evolution or elevated cardiac enzyme levels . The median age of the cases was 66 years (range 27 to 81).
1.2 Control Patients
Control patients were never-smokers identified in the same centers as the cases, but they were admitted for acute conditions not related to known or suspected risk factors for AMI. All those admitted for cardiocerebrovascular, chronic or neoplastic conditions or with a history of cardiovascular disease were excluded from the control group. A total of 446 subjects were interviewed; 45% had traumatic conditions, 30% had surgical conditions and 25% had other miscellaneous illnesses such as acute infections, dental disorders and disorders of the ear, nose, skin and throat. The median age of the controls was 65 years (range 24 to 83).
1.3 Data Analysis
Odds ratios (as estimators of relative risks) of AMI, together with their 95% confidence intervals, were first derived from data stratified according to gender and age in decades, using the Mantel-Haenszel procedure [19, 20]. To account for a number of potential confounding factors, unconditional multiple logistic regression analysis was used. Included in the regression equation were terms for gender, age (years), years of education, social class, body mass index (Quetelet’s index [kg/m2]), cholesterolemia (mg/dl), history of hypertension and diabetes and family history of AMI . The Mantel extension chi-square test was used to evaluate trends across strata of increasing dose .
Attributable risk proportions were computed by means of the method described by Bruzzi et al. , which provides a summary attributable risk for multiple factors after allowance for confounding variables. The method requires the knowledge of the distribution of exposure to the risk factors only among case patients, provided that they are representative of the whole cohort with the disease, and of the odds ratio associated with the exposure. Statistical analyses were performed using Statistica/W statistical software .
A single-blind test–retest was calculated by two interviewers using the kappa statistic . Eighty-two subjects’ relatives (43 case, 39 control patients) had a telephone interview by a third interviewer 1 to 4 weeks after their first interview. The global K index for the 82 patients was 0.69 (0.65 for cases, 0.71 for controls).
In the overall case-control study data set (1,071 case, 1,060 control patients), the prevalence of smokers was 43% for case patients and 28% for control patients, and the multivariate odds ratios of current smokers in relation to nonsmokers was 2.24 (95% confidence interval [CI] 1.83 to 2.75).
The distribution of case and control patients according to gender, age, social status, exercise, education, body mass index, history of diabetes and hypertension, hypercholesterolemia (≥240 mg/dl) and family history of MI is shown in Table 1. More case patients were >75 years. Control patients reported more physical exercise and tended to be more educated than case patients. Case patients more frequently reported a history of diabetes, hypertension and hypercholesterolemia and a family history of MI.
Table 2shows the distribution of case and control patients according to the smoking status of their relatives. At least one relative smoked in 39% of case patients and in 26.2% of control patients; 6.9% of case patients and 6.7% of control patients reported a spouse who smoked. A spouse who smoked >20 cigarettes/day was present in 2.4% of case patients compared with only 0.7% of control patients. Moreover 25.9% of case patients and 15.7% of control patients reported one or more children who smoked.
The odds ratios for AMI according to the smoking status of the subjects’ relatives are reported in Table 3. The risk increased to 68% (multivariate odds ratio [OR] 1.68, 95% CI 1.20 to 2.37) in nonsmoking subjects with at least one relative who smoked. The risk increased to 59% in those with a spouse who smoked. There was a strong positive association between the number of cigarettes smoked per day by the spouse and the risk of AMI for the patient, with estimates nearly fourfold for patients who had a spouse who smoked heavily (≥20 cigarettes/day). With one or more children who smoked, the OR was nearly twofold for the patient (OR 1.80).
The risk of a close relative who smoked increased the risk in men to 89% and in women to 54% (Table 3). A smoking spouse increased the risk to 92% in men and to 50% in women, and one or more children who smoked increased the risk more than twofold in men (OR 2.30) and 41% in women.
The relation between passive smoking and AMI is further examined in Table 4according to gender, age and other major covariates. The risk estimates for passive smoking were above the unity in almost all the strata considered. The association was stronger in those who were referred before diabetes and in those with a high social status. Likewise, the risk estimates were particularly high in those with low and high levels of education. In subjects <55 years old, the OR was under the unity (OR 0.84).
2.1 Association of Passive Smoking with Other Coronary Risk Factors
To investigate the interaction between passive smoking and known risk factors, the OR for each of these was evaluated, both alone and in combination with the presence of passive smoking (Table 5). The effect on OR was multiplicative for diabetes and hypercholesterolemia. In those subjects with diabetes, the OR increased from 1.74 in those with nonsmoking relatives to more than fivefold (OR 5.26) in those with smoking relatives. Similarly, in those subjects with hypercholesterolemia (≥240 mg/dl), the OR increased from 2.03 to fourfold (OR 4.01) in the absence and presence of passive smoking, respectively. An additive model was found between the relation of passive smoking and family history of MI and hypertension. In those who were referred because of a family history of MI, the OR increased from 2.57 in those not exposed to environmental tobacco smoke to more than fourfold (OR 4.08) in those exposed to environmental tobacco smoke, and in those who reported a history of hypertension, the OR increased from 1.92 to 3.28.
This study confirms that passive smoking is an indicator of risk of AMI. After allowance for other potential confounding variables, the ORs remained elevated. The presence of one or more close relatives who smoked increased the risk to ∼89% in men and 54% in women.
Others studies reported an elevated risk of coronary heart disease in passive smokers. Wells studied the magnitude of the risk of coronary heart disease in 12 published epidemiologic studies; the pooled relative risks of nonfatal coronary events and of dying of heart disease were 1.3 (95% CI 1.4 to 1.6) and 1.2 (95% CI 1.1 to 1.4), respectively. Only two of these studies, which had controlled for confounding variables that included the principal known and suspected risk factors for coronary heart disease, showed significant statistically ORs. Hole et al. , in a prospective study in Scotland, found a relative risk of mortality from ischemic heart disease of 2.01 (95% CI 1.21 to 3.35). In Georgia, a prospective study of up to 20 years of follow-up of 513 women >40 years old, showed an increased risk of mortality from coronary heart disease (OR 1.59, 95% CI 0.99 to 2.57) .
Two recent case-control studies examined the effect of passive smoking on cardiovascular disease. La Vecchia et al. reported a multivariate relative risk of AMI of 1.21 (95% CI 0.57 to 2.52) for those subjects married to a current smoker in relation to those with a spouse who never smoked. In China, He et al. assessed data from nonsmoking females: 59 with coronary heart disease and 126 control subjects. After adjustment for age, hypertension, type A personality, cholesterol and high density lipoprotein cholesterol level, the ORs for passive smoke from husband and co-workers were 1.24 (95% CI 0.56 to 2.72) and 1.85 (95% CI 0.86 to 4.00), respectively. Data from two other large, prospective cohort studies also support the evidence of an association between environmental tobacco smoke and coronary heart disease. Steenland et al. found that after adjustment for other coronary risk factors, passive smoking increased the risk of coronary heart disease to 22% in men and to 10% in women. Kawachi et al. examined the effect of environmental tobacco smoke in 32,046 nonsmoking women over a period up to 10 years. The relative risk of total coronary heart disease was 1.58 among women reporting occasional exposure and 1.91 among those reporting regular exposure.
Thus, the present results are in agreement with most findings of case-control and cohort studies, confirming that passive smoking may play an important role in the ischemic manifestation of coronary artery disease in different populations with different risk factor exposure characteristics. Likewise, these findings confirm that the association was appreciably stronger in both genders and in various strata of selected covariables.
The finding of a dose response with the number of cigarettes smoked per day by the spouse adds weight to the possibility of environmental tobacco smoke being causally related to the development of AMI. This linear direct trend relation that we found confirms the observation from several studies, including the Multiple Risk Factor Intervention Trial , a Scottish study , a Chinese study and the Nurses’ Health Study . Others studies, however, found no linear relation, including a British study , a prospective study from Maryland and the American Cancer Society cohort study . The inclusion of too few patients in the British study, or inadequate control for changes in the amount of cigarettes smoked during the follow-up period of cohort studies, may well have flattened or distorted the slope of the relation.
In this study, the combined exposure of passive smoking with other coronary risk factors raised the AMI risk. These interactions strongly suggest a synergism between environmental tobacco smoke and other coronary risk factors: to an additive model when passive smoking was associated with hypertension or a family history of MI and to a multiplicative model when the association was with diabetes or hypercholesterolemia. These samples of synergy were not analyzed in other studies of passive smoking and coronary heart disease and suggest important direction for further investigations.
3.1 Strengths and Limitations of the Study
Among the strengths of this study are the almost complete participation, with only 4% of subjects refusing to participate, and the comparability between recruitment areas of case and control patients. Using a case-control design, we were able to collect information on risk factors for AMI from a large sample in a short period, controlling for major, identified potential confounding factors.
Short-term exposure to environmental tobacco smoke results in increased platelet sensitivity and a decreased ability of the heart to receive and process oxygen . A case-control study design may be particularly appropriate to investigate the passive smoking–AMI relation, if this reflects a short-term effect on risk. In fact, it may provide more accurate information on passive smoking during a short time before occurrence of AMI than cohort investigations, which generally obtain information on smoking exposure years or even decades before disease onset.
Most studies that linked environmental tobacco smoke at home with coronary heart disease obtained information on the spouse’s smoking habits [6–8, 12, 14, 15]and others reported exposure with someone else at home [9–11, 13, 17]without identifying the subject. None obtained data on other close-contact relatives’ habits. In our study, we observed that the risk of AMI was increased to 80% in those with a child who smoked. These results are of particular interest and may offer more reassuring evidence than data previously available in terms of more general epidemiologic inference.
A potential limitation of this study is that patients with AMI had to survive long enough to be interviewed and therefore may not have been representative of all patients with AMI. Nonetheless, it seems unlikely that the exclusion of patients who died before or soon after admission to the hospital for a first AMI would have affected our findings, because the relation of passive smoking to fatal disease appears quite similar to its relation to nonfatal illness. The association between passive smoking and AMI was observed in other prospective studies of both fatal and nonfatal AMI.
Another limitation of the present study design was that information was based only on self-reporting. However, such information bias is likely to be similar for both case and control patients and should dilute the strength of the association to the same extent.
Although this study shows that passive smoking is associated with an increased risk of AMI, the OR was lower for younger subjects; however, the wide confidence interval (95% CI 0.34 to 2.04) does not allow any definitive conclusion on this issue.
These results indicate that passive smoking at home increase the risk of AMI. If the observed association is real, in terms of population attributable risk, >14% of first episodes of AMI in men who never smoked and 18% in women who never smoked in Argentina could be avoided by preventing passive smoking at home. These results underline the importance of new public health policy against environmental tobacco smoke, to make a real impact on the control of coronary heart disease in various areas of the world, especially in developing nations in which tobacco consumption is increasing.
We thank Charles Hennekens, MD, for the revision of the manuscript.
A.1 Argentine Factores de Riesgo Coronario en América del Sur (FRICAS) Investigators
Steering Committee:H. Schargrodsky, M. Ciruzzi, P. Pramparo, J. Rozlosnik.
Coordinating Committee:B. Abecasis, C. Brenner, J. César, H. Delmonte, O. Esteban, B. Labonia, H. Montagna, C. Paterno, V. Rudich, S. Soifer, J. Tartaglione
Participating Centers: Buenos Aires (Capital)—Alemán Hospital: D. Siskos; Argerich Hospital: S. Centeno; Churruca Hospital: D. Galván, S. Cherkerdemian; Fernández Hospital: C. Nejamsky, D. Rigou; Israelita Hospital: A. Kiezelstein, S. D’agostino, A. Bronstein; Italiano Hospital: V. Rudich, R. Oliveri; Pirovano Hospital: I. Luluaga, H. Zylberstejn, M. Fortunato, P. Soria, J. Lázzari; Zubizarreta Hospital: C. Brenner, Y. Plotquin; Anchorena Medical Center: E. Mele, L. Quintana; Guemes Medical Center: R. Ahuad; Hacienda Medical Center: M. Haquim; Mater Dei Medical Center: R. Calvino, O. Iavicoli; Méndez Medical Center: A. Monetti, L. Eda, B. Kogan; Trinidad Medical Center: M. Festa, B. Fromen; Clı́nica Del Sol Medical Center: E. Esparza Iraola, C. Taquini. Buenos Aires (Provinces)—La Matanza Hospital: C. Rodrigo; San Isidro Hospital: D. Romero Matos; Posadas Hospital: B. Abecasis; San Juan de Dios Hospital: B. Abecasis. Bahı́a Blanca (City):O. Camou, D. Solı́s. Coronel Suárez (City):A. Caccavo. La Plata (City):R. De Marco, P. Pardo. Tres Lomas (City):A. Alfonso. Corrientes:J. Vaccaro, Entre Rı́os (Concordia):J. De la Cruz Ojeda. Jujuy (San Salvador De Jujuy):R. Peleteiro, G. Bustamante. Misiones (Posadas):S. Castillo. Neuquén (San Martı́n De Los Andes):G. Pichel. Rio Negro (Viedma):R. Coniglio. Santa Fe (Rosario):D. Piskorz, R. Grisolia, C. Girino, M. Mancini. Tucumán (San Miguel De Tucumán):J. De Rosa, J. Waisman.
- acute myocardial infarction
- confidence interval
- Factores de Riesgo Coronario en América del Sur
- myocardial infarction
- odds ratio
- Received July 11, 1997.
- Revision received November 12, 1997.
- Accepted December 17, 1997.
- The American College of Cardiology
- Wells AJ
- Taylor AE,
- Johnson DC,
- Kazemi H
- Hirayama T
- United States Environmental Protection Agency
- Garland C,
- Barret-Connor E,
- Suarez L,
- Criqui M,
- Wingard D
- Svendsen KH,
- Kuller LH,
- Martin MJ,
- Ockene JK
- Helsing KJ,
- Sandler DP,
- Comstock GW,
- Chee E
- Hole DJ,
- Gillis CR,
- Chopra C,
- Hawthorne VM
- He Y,
- Lam TH,
- Li TS,
- et al.
- Steenland K,
- Thun M,
- Lally C,
- Heath C Jr.
- Kawachi I,
- Colditz GA,
- Speizer FE,
- et al.
- ↵Ischaemic Heart Disease Registers. Report of the Fifth Working Group. Copenhagen, Denmark: World Health Organization, 1971.
- Mantel N,
- Haenszel W
- Breslow NE,
- Day NE
- Bruzzi P,
- Green SB,
- Byar DP,
- Brinton LA,
- Schairer C
- ↵Statistica/W, sw 40340486J6D45; Statsoft 1994.