香京julia种子在线播放

A comprehensive search of studies investigating the association between smoking and BC was carried out before February 2022 in electronic databases of PubMed, Web of science, EBSCO, and the Cochrane Library. The complete retrieval formula that was used to identify the related studies includes: (“breast cancer” OR “breast neoplasms” OR “BC”) AND (“smoking” OR “tobacco smoke pollution” OR “tobacco use” OR “tobacco products” OR “active smoking” OR “passive smoking” OR “secondhand smoking” OR “tobacco”). The reference lists of retrieved studies and conference records were also reviewed for potentially inclusive studies. When referring to duplicate literature, the original article was included if the study was published as an abstract or an original article. Also, if a study was continuously updated and reported, only the most recent or comprehensive articles were included. This meta-analysis was conducted according to the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) guidelines (41). The population, intervention, comparison, outcome, and setting (PICOS) criteria were used to describe the research question. Participants in this study were people who had not previously been diagnosed with BC, the intervention was exposure to tobacco environments, including active and passive smoking, the comparison was a non-smoker, the outcome was the incidence of BC, and the setting was observational research. This meta-analysis’s prospero registration number was CRD42022322699.

An eligible criterion was formulated. The specific criteria were as follows. Inclusion criteria: (1) all included studies are observational studies. (2) The main exposure of study was smoking including active and passive smoking, and the outcome was BC risk. (3) All studies included available data which reported the relationship between smoking and BC. Exclusion criteria: (1) the study was conducted on BC population and used mortality or recovery rate as the outcome. (2) The study was published in duplicate. (3) The study was not published in English.

A jointly agreed data collection form was used to extract all data. Information was extracted as follows: the author’s name, year of publication, study type, age, exposure assessment, number of participants, number of BC cases, number of smokers, number of non-smokers, variables adjusted in the statistical analyses, and outcomes. To ensure the objectivity and accuracy of the data, two researchers independently extracted data from each study. Disagreements were resolved by consensus or consultation with a third researcher.

The quality of each included study was evaluated by the Newcastle-Ottawa Quality Assessment Scale (NOS) checklist, a tool used for quality assessment of non-randomized studies. NOS checklist is composed of eight items classified into three aspects, including selection, comparability, and outcome. The maximum scores of this checklist were nine, and scores between seven and nine were identified to be of higher study quality.

The primary objective was to explore the relationship between smoking and the incidence of BC. Secondary objectives were to explore the relationship between the incidence of BC and smoking subgroups (e.g. smoking pattern, smoking time, smoking frequency, smoking place, smoking cessation time, age of starting smoking), the relationship between smoking and BC in different populations (e.g. fertility status, menopausal status, race), and the association between smoking and different BC subtypes (e.g. estrogen receptor-positive (ER+) BC, estrogen receptor-negative (ER-) BC). The results after adjusting for relevant confounding factors were used consistently for the processing of relevant data from the included articles.

The Stata software version 12 (StataCorp, College Station, Texas, USA) was used to analyze the data. The confidence interval (CI) of odd ratio (OR) was set at 95% to examine the relationship between smoking and BC risk. Heterogeneity of included studies was tested by Q statistic and I² statistic to quantitatively assess inconsistency. For statistical results, values of p<0.10 and I²>50% were considered to be representative of having statistically significant heterogeneity. Based on the heterogeneity of smoking intensity, smoking duration, race, BC subtype, etc. in different studies, in order to improve the reliability of the results, the random effects model was uniformly used in this study. When more than ten studies were included, sensitivity analysis and publication bias test were performed to evaluate the stability and reliability of their results. Publication bias was evaluated by the Begg’s test. Results with P-values less than 0.05 were considered to be statistically significant.

A total of 19,746 relevant articles were identified based on retrieval formula described in the methods section by initial search in PubMed, EBSCO, Web of Science, and Cochrane Library database. No additional records were identified through other sources. A total of 8,463 duplicate articles were deleted, and 11,283 articles were excluded due to the title or abstract. The remaining 932 articles were reviewed through full-text. Among them, 855 articles were eliminated because of being non-observational study (n=339), duplicate publication (n=218), not exploring the risk of BC (n=176), no relevant results reported (n=85), and not published in English (n=37). Eventually, 77 articles (13, 32–34, 42–115) composed of 2,326,987 participants were selected for this meta-analysis. The detailed search and study selection process was shown in Figure 1 .

Of the 77 included studies, 24 were cohort studies (2,138,338 participants and 55,703 BC cases), 53 were case-control studies (188,649 controls and 58,859 BC cases). The participants in the two studies included men and women, and the rest were women. All studies were published between 1988 and 2022, with follow-up periods ranging from 6 to 24.6 years. Regarding age at recruitment, eight studies did not set the upper age limit, four studies did not set a lower age limit and four studies did not report the requirement for age. Among them, 30 studies were conducted in America, 24 were in Asia, 22 were in Europe, and 1 was in Oceania. Fifty-six studies investigated the association between active smoking and BC risk, 39 investigated the association between passive smoking and BC risk. The number of smokers (included active and passive smokers) was 1,326,603 in cohort studies and 108,175 in case-control studies. In order to collect data and evaluate relevant exposure factors, 59 studies chose questionnaire, 9 studies chose interview, and 9 studies chose questionnaire combined with interviews. In addition, the adjustment of potential confounding factors varied in different studies. Most of the adjustment parameters were age, body mass index (BMI), family history of BC, total energy intake, alcohol consumption, number of births, and physical activity. The characteristics of the included studies were shown in Table 1 and Supplementary Table 1 .

Fifty-six studies recorded data about active smoking in female population that was inducing BC. Studies had shown that women who actively smoked had a significantly higher incidence of BC than those who had never actively smoked (OR=1.15, 95% CI=1.11-1.20, p<0.001, I^{2 =} 54.9%). Among them, current active smoking (OR=1.12, 95% CI=1.08-1.16, p=0.007, I^{2 =} 40.1%) and former active smoking (OR=1.09, 95% CI=1.06-1.12, p<0.001, I^{2 =} 33.3%) had a significantly increase on the incidence of BC, but current active smoking increased the incidence of BC more than former active smoking. In other words, active smoking is a risk factor for women, and the population who is still active smoking is under more risk than the population who quit smoking after active smoking. In addition, cohort studies (OR=1.13, p<0.001) and case-control studies (OR=1.19, p<0.001) had consistently concluded that active smoking increases the risk of BC in women. The detailed data was contained in Table 2 .

The correlation between smoking and BC is affected by menopausal status. Related data were available in 23 studies with premenopausal BC and 25 with postmenopausal BC. The analysis showed that active smoking increases the incidence of premenopausal BC (OR=1.24, 95% CI=1.17-1.32, p<0.001, I^{2 =} 6.2%), but had no effect on postmenopausal BC (OR=1.03, 95% CI=0.97-1.10, p=0.314, I^{2 =} 30.8%) with slight heterogeneity. The detailed data was contained in Table 2 .

Years were used to measure smoking duration in this study. The related data were divided into ‘<20 years group’, ‘20-30 years group’, ‘30-40 years group’, and ‘>40 years group’ according to the most studies. The results showed that women who smoked for less than 20 years (OR=1.06, p<0.001), 20-30 years (OR=1.15, p<0.001), 30-40 years (OR=1.15, p<0.001), and more than 40 years (OR=1.22, p<0.001) had a higher incidence of BC than those without smoking history. The incidence of BC was positively correlated with smoking duration. The detailed data was contained in Table 2 .

Cigarettes per day were used to measure smoking intensity in this study. The data is grouped by 10 cigarettes per day, 20 cigarettes per day, and 30 cigarettes per day. Subgroup analysis showed smoking which less than 10 cigarettes per day (OR=1.06, p=0.001), between 10-20 cigarettes per day (OR=1.19, p<0.001), between 20-30 cigarettes per day (OR=1.16, p<0.001), and more than 30 cigarettes per day (OR=1.18, p=0.001) increased the incidence of BC with statistical significance. The incidence of BC increased with the increase of smoking intensity. The detailed data was contained in Table 2 .

Pack-years were used to simultaneously assess smoking duration and smoking intensity. Pack-years were defined as the product of the number of cigarettes smoked per day and the number of years of smoking. According to the grouping criteria of the included studies, this study divided the relevant data into ‘<10 pack-years group’, ‘10-20 pack-years group’, ‘20-40 pack-years group’, and ‘>40 pack-years group’. The analysis showed that women who smoke with less than 10 pack-years (OR=1.05, p=0.005), 10-20 pack-years (OR=1.11, p<0.001), 20-40 pack-years (OR=1.21, p<0.001), and >40 pack-years (OR=1.17, p<0.001) had a higher incidence of BC than those who had never smoked. The detailed data was contained in Table 2 .

In this study, smoking initiation age was divided into ‘<16 years group’, ‘17-19 years group’, and ‘>20 years group’. The results suggested that active smoking, regardless of the age at which smoking started is younger than 16 years old (OR=1.11, 95% CI=1.07-1.15), between 17-19 years old (OR=1.06, 95% CI=1.12-1.20), or older than 20 years old (OR=1.08, 95% CI=1.04-1.11), would significantly increase the incidence of BC in women with slight heterogeneity. The detailed data was contained in Table 2 .

Years of quitting smoking were used to measure the effect of smoking cessation in the participants. Data were grouped by 10- and 20-year cessation years. Subgroup analysis showed that previous smoking history remained a risk factor for BC among women who had quit smoking for less than 20 years. Among them, the harm of previous smoking history to women who quit smoking for less than 10 years (OR=1.27, 95% CI=1.15-41, p<0.001) is significantly greater than that to those who quit smoking for 10-20 years (OR=1.05, 95% CI=1.00-1.09, p=0.046). With increased time to quit smoking comes a reduction in the harm caused by previous smoking history. Previous smoking history was no longer an observable risk factor for BC in women who had quit smoking for more than 20 years (OR=1.01, 95% CI=0.97-1.06, p=0.552). The detailed data was contained in Table 2 .

Six studies explored the association between active smoking and BC in different fertility statuses. The analysis showed that active smoking can increase the risk of BC in the multiparous population (OR=1.13, 95% CI=1.07-1.20, p<0.001), but had no effect on BC in the nulliparous population (OR=1.05, 95% CI=0.92-1.20, p=0.432) without heterogeneity. The detailed data was contained in Table 2 .

Regarding the relationship between active smoking and BC risk before/after the first birth, 24 studies contained data before the first birth and 22 studies contained data after the first birth. The results of the analysis showed that active smoking significantly increased the incidence of BC, regardless of whether the mother was smoking before the first birth (OR=1.22, 95% CI=1.17-1.27, p<0.001) or smoking after the first birth (OR=1.08, 95% CI=1.04-1.12, p<0.001), with slight heterogeneity. Furthermore, active smoking before the first birth had a greater impact on inducing BC than active smoking after the first birth. The detailed data was contained in Table 2 .

Among those who actively smoked before the first birth, data were grouped by 5 years of smoking. Subgroup analysis showed that active smoking before the first birth increased the risk of BC whether the duration of smoking less than 5 years (OR=1.06, p=0.023) or more than 5 years (OR=1.24, p<0.001). There was a positive correlation between the smoking duration before the first birth and the risk of BC. Among those who have actively smoked after the first born, data were grouped by 10 years of smoking. Subgroup analysis showed that active smoking after the first birth had no effect on BC whether the duration of smoking less than 10 years (OR=1.00, p=0.922) or more than 10 years (OR=1.06, p=0.077). However, with the increase of smoking duration, active smoking had a tendency to harm the female population after the first birth by inducing BC. The detailed data was contained in Table 2 .

Six studies examined the association between active smoking and BC subtypes. The results showed that active smoking increased the incidence of ER+ BC (OR=1.13, 95% CI=1.08-1.18, p<0.001), but had no effect on ER- BC (OR=1.08, 95% CI=0.97-1.19, p=0.155), without heterogeneity. The detailed data was contained in Table 2 .

This study grouped data by 10-year active smoking aimed to investigate the correlation between different smoking duration and BC subtype. The analysis showed that active smoking for less than 10 years did not increase the incidence of BC, regardless of whether it was ER+ BC (OR=0.99, p=0.870) or ER- BC (OR=1.02, p=0.699). Active smoking for more than 10 years had no effect on ER- BC (OR=1.08, p=0.105), but could increase the incidence of ER+ BC (OR=1.14, p=0.007). The detailed data was contained in Table 2 .

This study investigated the effect of smoking on BC subtypes at different smoking intensities by grouping data at 10 cigarettes per day boundaries. Subgroup analysis showed that smoking less than 10 cigarettes per day (OR=1.08, p=0.041) and more than 10 cigarettes per day (OR=1.18, p=0.002) could increase the risk of ER+ BC, and the risk was positively related to smoking intensity. For ER- BC, smoking less than 10 cigarettes per day had not been discovered as being effective (OR=0.97, p=0.603), However, smoking more than 10 cigarettes per day could increase the risk of suffering from ER- BC (OR=1.18, p=0.049). The results suggested that the occurrence of ER+ BC was more likely to be affected by active smoking than ER- BC. The detailed data was contained in Table 2 .

Thirty-nine studies documented BC risk data from passive smoking in women. The analysis showed that the risk of BC was significantly higher among women who passively smoked than those without passive smoking episode (OR=1.17, 95% CI=1.09-1.24, p<0.001, I^{2 =} 59.2%). Among them, current passive smoking had a significant effect on BC (OR=1.31, 95% CI=1.08-1.60, p=0.007, I^{2 =} 27.6%), but such history had no effect on BC (OR=1.18, 95% CI=0.97-1.43, p=0.107, I^{2 =} 42.5%). This suggests that passive smoking, especially current passive smoking would increase the risk of BC. Furthermore, cohort studies (OR=1.08, 95% CI=1.03-1.13) and case-control studies (OR=1.15, 95% CI=1.14-1.39) had consistently concluded that passive smoking increases the risk of BC in women. The detailed data was shown in Table 3 .

Eleven studies included data on the relationship between passive smoking and BC in different menopausal states. The analysis showed that passive smoking increased the risk of premenopausal BC (OR=1.29, 95% CI=1.13-1.49, p<0.001, I^{2 =} 37.3%), but had no effect on the incidence of postmenopausal BC (OR=1.13, 95% CI=0.93-1.36, p=0.218, I^{2 =} 73.5%). The detailed data was contained in Table 3 .

Regarding the relationship of passive smoking and BC in different exposure places, 11 studies had data on home exposure, 11 studies had data on work exposure, and 5 studies had data on both home and work exposure. Subgroup analysis showed no relationship between passive smoking and BC incidence in different passive smoking exposure settings. However, passive smoking exposure at work (OR=1.09, p=0.051) and exposure at both home and work (OR=1.40, p=0.051) had a trend of harm to female population. The detailed data was contained in Table 3 .

In terms of the association between passive smoking and BC at different exposure ages, 16 studies had data on exposure in childhood, 15 studies had data on exposure in adult, and 8 studies had data on exposure in children and adult. Subgroup analyses showed that passive smoking increased BC risk regardless of exposure to childhood (OR=1.15, p=0.002), adult (OR=1.21, p=0.014), or both childhood and adult (OR=1.49, p=0.003). Among them, the increased risk of BC in those with simultaneous exposure in childhood and adult was significantly greater than that in those only with a single age group. The detailed data was contained in Table 3 .

Years were used to measure the duration of passive smoking exposure in this study. The relevant data were divided into ‘less than 10 years group’, ‘10-20 years group’, ‘20-30 years group’, and ‘more than 30 years group’, in the way most studies were segmented. This study showed that passive smoking which duration was less than 10 years in female population had no effect on BC (OR=0.99, p=0.876), while passive smoking exposure for 10-20 years (OR=1.13, p=0.011), 20-30 years (OR=1.38, p<0.001) and more than 30 years (OR=1.35, p=0.004) had a significant impact on the incidence of BC, compared to women who had never smoked. In all, increased incidence was positively correlated with longer duration of passive smoking exposure. The detailed data was contained in Table 3 .

The NOS checklist was adopted to objectively evaluate the quality of included observational studies in this meta-study. 95.83% of the cohort studies were of high quality (NOS score >7), while 94.33% case-control studies were of high quality (NOS score >7). The quality ratings of cohort and case-control studies were listed in Supplementary Tables 2 and 3 .

Publication bias was evaluated by the Begg’s test. The results of Begg’s test indicated the absence of publication bias among included articles (p>0.05). Sensitivity analysis was used to assess whether the individual studies affected the overall results or not. The results indicated that the analysis was relatively stable.