Methods

Study subjects

The study was conducted in the period 2004–2007 in three regional hospitals in Hong Kong (the Prince of Wales Hospital, Alice Ho Miu Ling Nethersole Hospital and Northern District Hospital). These are the major regional hospitals managed by Hospital Authority New Territories East Cluster, serving Shatin, Tai Po, Northern New Territories, Sai Kung and the outlying islands of Hong Kong.

All the subjects of this study were female never-smokers. IHD cases were recruited from the cardiac wards or coronary care unit during the study period. These included patients with angina pectoris who were confirmed to have coronary stenosis of 70% or more in at least one coronary artery; patients with acute coronary syndrome who had chest pain, typical electrocardiogram changes and elevation of cardiac enzyme in the form of creatine kinase-MB or troponin T; and patients who had a clinical diagnosis of IHD and received either percutaneous coronary intervention or coronary artery bypass grafting surgery. Diagnosis of IHD was based on International Classification of Diseases 10th revision (ICD-10) codes I20 to I25. Controls were defined as patients attending the same hospitals during the same period of IHD cases without any clinical symptom, sign or suspicion of cardiovascular disease in their medical history. In the case and the control groups, individuals who fulfilled the criteria of “non-consent” (3%) and “conditions related to cigarette use” (5%) were excluded, such as chronic obstructive pulmonary disease, asthma, bronchitis and lung cancer.

Passive smoking exposure

The subjects were interviewed in person, using a standardised questionnaire, regarding their demographics (gender, age, education level), lifestyle (smoking habit, alcohol intake, physical activity), family IHD history and duration and intensity of exposure to household passive smoking. The interviewers were blinded to the case/control status of the subjects. The habit of smoking and alcohol intake were determined by asking “Do you have the habit of smoking/alcohol intake?”. Physical activity was assessed with the question “Do you exercise routinely (aerobic exercise at least 3 times per week, for at least 30 min every time)?” while family IHD history was determined with the question “Do you have any family members who got IHD before 55 years old?” Passive smoking exposure at home was determined by asking “Do (have) any of your family members smoke (smoked) in your household? If yes, answer how many years has (did) each person smoked; how many packs of cigarettes per day does (did) each person smoke; and how many hours per day are (were) you exposed to smoking?” Other medically related risk factors for IHD, such as hypertension, diabetes, hypercholesterolaemia, obesity, gout, stroke, menopause and hormonal exposure, were confirmed by reviewing medical records from the hospital computer system.

The scope of passive smoking was household exposure. The source of exposure included those from parents, spouse, children, siblings, relatives and others at home. We evaluated the dose metrics of passive smoking exposure in terms of “pack years” (1 pack year = 365 packs, or the equivalent of 1 pack/day for 1 year) and “hour years” (1 hour year = 365 h, or the equivalent of 1 h/day for 1 year), which indicated the cumulative amount and duration of exposure.11 We also inquired about passive smoking exposure at work, however nearly all subjects were housewives who did not work outside home and a few were long-retired persons who may have had significantly poor recall. Therefore, occupational exposure to passive smoking was negligible in the current study.

Statistical analysis

All data were analysed using statistical software (SPSS 13.0; SPSS, Chicago, Illinois, USA). Differences between the two groups were identified by independent Student t test for parametric continuous variables and Mann–Whitney U test for non-parametric variables. The Pearson χ² test or Fisher exact probability test was used when appropriate for categorical variables. A p value <0.05 was considered as statistically significant.

ORs and 95% CIs were computed by unconditional logistic regression, adjusted for covariates such as age, education level, history of hypertension, diabetes mellitus, hypercholesterolaemia, stroke, gout, family history of IHD, physical activities, alcohol intake and oestrogen use. “Never exposed” was defined as the reference category with OR = 1.

Results

Clinical characteristics

A total of 314 female patients who had been diagnosed with IHD and 319 controls were enrolled in this study. The clinical characteristics of patients with IHD and controls are shown in table 1. Patients with IHD were slightly older than control subjects (mean (SD) 72.3 (8.5) vs 68.7 (12.2) years, p<0.001), and had a higher prevalence of hypertension (74.2%), diabetes mellitus (43.0%) and hypercholesterolaemia (55.1%) than controls (p<0.001). In addition, 12.1% of cases had a family history of IHD (p = 0.037), and a smaller proportion of IHD cases were physically active (40.1% vs 48.6%, p = 0.032). Other factors did not differ between the IHD and control groups.

Exposure to passive smoking

Table 1 shows that more cases with IHD than controls (30.3% vs 20.4%, p = 0.004) were exposed to household passive smoking. After being adjusted for age, education, hypertension, diabetes mellitus, hypercholesterolaemia, gout, history of stroke, family history of IHD, physical inactivity, alcohol intake and oestrogen use, non-smokers ever exposed to household passive smoking were associated with a significantly higher risk of IHD (OR 1.52, 95% CI 1.01 to 2.27, p = 0.043) when compared to those who were never exposed (table 2).

Subjects exposed to an average of <1 pack of cigarettes per day from their family members had an OR for IHD of 1.14 (95% CI 0.58 to 2.23), whereas subjects exposed to an average of ⩾1 pack per day had an increased OR of 1.69 (95% CI 1.07 to 2.68, p = 0.025). The risk of IHD was already high for duration of exposure of <5 years, with an OR of 1.26 (95% CI 0.18 to 8.77). However, a much higher risk for IHD was present for a longer duration of exposure for ⩾5 years, with an OR of 1.52 (95% CI 1.01 to 2.29, p = 0.043). An OR of 1.28 (95% CI 0.76 to 2.15) was found for passive smoking exposure of <4 h per day, while a higher OR of 1.82 (95% CI 1.05 to 3.15, p = 0.032) was found for exposure of ⩾4 h per day. Pack years of exposure displayed an increasing OR of 1.44 (95% CI 0.38 to 5.48, p = 0.591) and 1.53 (95% CI 1.01 to 2.32, p = 0.043) with an increasing exposure level of <5 pack years and ⩾5 pack years, respectively. Similarly, the risk of IHD increased by the cumulative exposure duration, with ORs of 1.22 (95% CI 0.59 to 2.52, p = 0.587) and 1.61 (95% CI 1.03 to 2.52, p = 0.037) at the exposure level of <20 hour years and ⩾20 hour years, respectively. There was a significant dose-response association of all the exposure measures to risk of IHD (p<0.01 for trend) (table 2).

Discussion

The current study reported the association between household passive smoking and presence of IHD among never-smoke female subjects. Since most study subjects were housewives who did not work and hence did not receive substantial exposure outside the home, there was little exposure misclassification due to unquantified workplace exposures. A significant association persisted even after being adjusted for other known risk factors and a dose-response pattern was also found for the cumulative amount and duration of exposure to passive smoking.

Over the past two decades, many epidemiological studies and reviews attempted to explore the effect of passive smoking on the risk of IHD. Hazardous effect of passive smoking exposure was more obvious in population studies that contained both genders.12 13 Only a few studies focused on household passive smoking exposure among female never-smokers. In 1988 to 1989, within the framework of the Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico (GISSI)-2 study, a case-control study was conducted to investigate 113 patients with first-ever acute myocardial infarction (AMI) and 225 controls for the exposure to spousal smoking. The association between passive household smoking and IHD in females, however, was not established.14 He et al conducted a case-control study that enrolled 59 female patients with IHD and 126 controls in China (Xi’an) from 1989 to 1992. The crude OR and adjusted OR for passive smoking from husbands’ cigarette smoke were 2.12 (95% CI 1.06 to 4.25) and 1.24 (95% CI 0.56 to 2.72), respectively.7 Although the risk of passive smoking and IHD was statistically insignificant, a significant dose-response relationship between the ORs for IHD and spousal cigarette consumption, expressed as the cumulative quantity and duration of passive smoking, was observed from this study. Another case-control study conducted in Argentina between 1991 and 1994 indicated that the multivariate ORs of AMI for female passive smokers were 1.24 to 4.03, according to the different smoking status of their relatives.15 These studies either had small sample sizes or demonstrated only an increased risk for high level of exposure, though they suggested a potential link between passive household smoking and IHD in female subjects. In contrast, the current study established that passive household smoking contributed a 1.5-fold increased risk of IHD in female never-smoke subjects compared to those without exposure in a relatively larger population. Furthermore, the risk was also higher for female subjects exposed to a “high-dose”, which was 1.7-fold, 1.8-fold and 1.5-fold for ⩾1 pack of cigarettes per day, ⩾4 h of exposure per day and ⩾5 years of exposure, respectively.

There were potential limitations in this study. Selection and information bias are the main sources of systematic errors in case-control studies. Since the study subjects were selected from hospitals, they were not completely representative of community residents. Recruited subjects were limited to those who could respond to questionnaires. It is possible that patients and controls who did not respond to questionnaires were associated with passive smoking exposure in a different way. Although we strictly followed the inclusion criteria during subject selection, insignificant misclassification might still exist since a small percentage of asymptomatic patients with IHD may have been wrongly assigned as controls. Thus, the ORs of IHD reported from our study would be underestimated.16

We tried to avoid information bias through accurate and detailed data acquisition from the medical records via the hospital computer system, in particular the onset of cardiovascular risk factors investigated. We also avoided interviewer bias by blinding interviewers to the case-control status. However, recall bias may still be an important issue in terms of our reliance on self-report, because cases were more likely to recall their passive smoking exposure than controls. Studies based on self-report seem to underestimate the risks of exposure to passive smoking.16 Potential measures to avoid recall bias in assessment of passive smoking in future studies include test–retest design and biochemical markers test of salivary and urinary cotinine assay.17 18 19 Moreover, large prospective studies will be helpful to confirm a causal relation between household passive smoking and IHD in female subjects.

In conclusion, our findings add to the accumulating epidemiological evidence linking household passive smoking exposure with a higher risk of IHD. Current data have indicated that 50% of non-smokers in mainland China and 48% of non-smokers in Hong Kong are exposed to passive smoking.20 21 Considering the high prevalence of passive smoking exposure, the consequences of passive smoking with regard to cardiovascular disease burden underscore the public health policy that all persons, and especially women, should be fully protected from tobacco smoke.