Changes in tobacco-related morbidity and mortality in French women: worrying trends

Abstract

Background

The high prevalence of smoking among French women since the 1970s has been reflected over the past decade by a strong impact on the health of women. This paper describes age and gender differences in France of the impact of smoking on morbidity and mortality trends since the 2000s.

Methods

Smoking prevalence trends were based on estimates from national surveys from 1974 to 2017. Lung cancer incidence were estimated from 2002–12 cancer registry data. Morbidity data for chronic obstructive pulmonary disease (COPD) exacerbation and myocardial infarction were assessed through hospital admissions data, 2002–15. For each disease, number of deaths between 2000 and 2014 came from the national database on medical causes of death. The tobacco-attributable mortality (all causes) was obtained using a population-attributable fraction methodology.

Results

The incidence of lung cancer and COPD increased by 72% and 100%, respectively, among women between 2002 and 2015. For myocardial infarction before the age of 65, the incidence increased by 50% between 2002 and 2015 in women vs. 16% in men and the highest increase was observed in women of 45–64-year-olds. Mortality from lung cancer and COPD increased by 71% and 3%, respectively, among women. The estimated number of women who died as a result of smoking has more than doubled between 2000 and 2014 (7% vs. 3% of all deaths).

Conclusions

The increase in the prevalence of smoking among women has a major impact on the morbidity and mortality of tobacco-related diseases in women and will continue to increase for a number of years.

Introduction

In industrialized countries, smoking generates a considerable health burden and is the leading cause of preventable death.¹ Women have been an important target of the tobacco industry, which has used all marketing tools for decades to encourage more women to smoke.² By using a glamorous and sophisticated image through fashion, ads, films and TV, the tobacco industry has worked hard to surround cigarettes with a positive social image and market that image to women. It also adapted to this new market by developing products specifically designed for women, including flavoured and ‘slim’ cigarettes. Harnessing societal pressure to stay thin, advertisements presented cigarettes as a way for women to control their weight.³ Finally, ads also focused on women’s independence, encouraging women to smoke as a symbol of their equality with men. In France and elsewhere, the result of this shift in behaviours has been an increase in tobacco use among women over the past 50 years. In 2017, 24% of women aged 18–75 reported smoking every day compared to 30% of men. This trend has remained relatively stable since the 2000s.^4,5

Research has shown that the same quantity of tobacco results in a greater risk of heart attack and chronic obstructive pulmonary disease (COPD) in women than in men.^6,7 However, studies have shown conflicting results with respect to lung cancer regarding this point.^8–10

The aim of this article is to describe smoking-related morbidity and mortality trends in France since the early 2000s for both men and women in relation to changes in the prevalence of tobacco use and according to age categories. The morbidity and mortality rates for three pathologies highly attributable to long-term exposure to smoking (lung cancer, COPD) and to shorter-term exposure to smoking (myocardial infarction).¹¹ Finally, recent trends in smoking-related deaths for all pathologies causally linked to smoking were estimated.

Methods

Prevalence of tobacco use

In France, the prevalence of daily smoking was measured by general quota-sampling population surveys between 1974 and 1991 and cross-sectional surveys, called Health Barometers (‘Baromètre santé’), were then carried out from 1992 to 2017. These surveys were carried out over the phone using random samples of the population living in mainland France and speaking French.¹²

Morbidity of the main tobacco-related pathologies

Myocardial infarction- and COPD-related morbidities were studied by examining hospitalizations for myocardial infarction or for COPD exacerbation. Data were extracted from the French Hospitalization Activity Database (‘Programme de médicalisation des systèmes d’information–Médecine, chirurgie obstétrique’, hospital discharge databases PMSI) on hospital admissions. For each year of the study period (2002–15), patients over the age of 35, living in France and hospitalized at least once during the year for a myocardial infarction or COPD exacerbation were selected. Hospital stays were identified with a primary diagnosis (PD) of myocardial infarction with the codes I21, I22 and I23 in accordance with the 10th revision of the International Classification of Diseases (ICD-10). For COPD exacerbations, hospital stays corresponding to one of the following criteria were studied: 1/PD of COPD with an acute lower respiratory infection (J44.0) or with an acute, unspecified episode (J44.1); 2/PD of acute respiratory failure (J96.0) or a lower respiratory infection (J09-J18, J20-J22) with a significant associated diagnosis (SAD) of COPD (J44) or emphysema (J43); 3/PD of COPD (J44) or emphysema (J43) with a SAD of an acute COPD exacerbation (J44.0–J44.1), lower respiratory infection (J10–J18, J20–J22), or acute respiratory failure (J96.0).

Lung cancer incidence rate for 2002–12 were estimated in France using the French network of cancer registries (Francim) databases, for which classification of neoplasm was based on the third edition of the International Classification of Diseases for Oncology (ICD-O-3).

Mortality of the main pathologies associated with tobacco use

Mortality data for the three pathologies studied were extracted from the national database of the French Epidemiological Centre on Medical Causes of Death (Inserm-CépiDc), which codes deaths in accordance with the ICD-10. Deaths occurring between 2000 and 2014 in people aged over 35, living in France and identified with an ICD-10 code for myocardial infarction (I21-I23), lung cancer (C33–C34), or COPD (J40–J44) were selected.

Number of smoking-related mortalities (all causes)

The number of tobacco-related deaths was based on a calculation of attributable deaths, which measures the proportion of all cases of a pathology within a population that could have been prevented if people had not used tobacco. We used the estimate method developed by Peto et al.¹³ and modified by Parkin,¹⁴ which combines mortality data, lung cancer death rates and the relative risk of death associated with smoking¹³ (Supplementary Material). The malignant and non-malignant pathologies associated with tobacco consumption are taken from a report issued by the World Health Organization regarding tobacco-related mortality.¹⁵ Regarding cancer risk, we used the list updated by the International Agency for Research on Cancer,¹⁶ which includes two new sites (colorectal cancer and ovarian mucinous carcinoma). For all of these pathologies, the mortality data (coded with ICD 10) from 2000 to 2014 were taken from the national Inserm-CépiDc database on causes of death. The raw number of deaths was corrected to account for deaths for which the certificate mentioned an ill-defined or unspecified cause (R96–R99 codes) or an ill-defined or unspecified site of cancer (C76–C80, C97 codes).¹⁷ Finally, for each tobacco-related pathology, the relative risk of death for smokers compared to non-smokers was primarily taken from analyses from the Cancer Prevention Study II (CPS-II), a very large American cohort of over one million people that was monitored from 1982 to 2006 (Supplementary table S1).

Statistical analyses

For each year, crude rates of each pathology were calculated, and stratified by sex. Age-standardized rates were estimated using standardization on the census of the 2010 European population. The average annual change in the crude rates by age group were calculated for each sex using Poisson regression model. The proportion of tobacco-attributable deaths are presented by sex and for the under 65-year-old age group (premature deaths) and the 65-and-over age group. The average annual variation of the proportion of tobacco-attributable deaths was estimated using a log-binomial model that was adjusted for the year, age group, sex and that took into account interactions between sex and age as well as the year and age.

Compliance with ethical standards

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent: Not applicable.

Results

Prevalence of daily smoking

Since the 1970s, discrepancies in the prevalence of daily smoking between men and women were observed in France (Supplementary figure S1). In men, daily smoking decreased from 60% in the 1970s to less than 40% in the early 2000s, then stabilized. In women, on the contrary, daily smoking rose in the 1970s and stayed at around 30% until the early 2000s. After dropping, smoking among women increased again during the second half of the 2000s to reach 30% in 2010. Remaining unchanging for a few years, daily smoking dropped in 2017 in both men and women (Supplementary figure S1).

The relative stability of the prevalence of daily smoking in women in the past few years is the result of disparate changes among the various age groups (figure 1). The prevalence of daily smoking increased among women aged 45–54, from 21.5% in 2000–30.8% in 2017. Among women aged 55–64, the decrease observed in 2017 occurred after the prevalence doubled between 2000 (11.0%) and 2016 (21.1%). This increase corresponds to the arrival of women born between 1950 and 1960 in these age groups; these women were among the first generations to widely adopt smoking habit in the 1970s. The prevalence of daily smoking decreased among 35- to 44-year-old women between 2000 and 2017.

Figure 1

Prevalence of daily smoking according to age and sex in people aged 35–75, France, 2000–17

Tobacco-related morbidity

In France in 2015, there were 65 311 and 79 158 patients hospitalized for myocardial infarction and a COPD exacerbations, respectively. In 2012, 39 378 new cases of lung cancer were reported (table 1). Nearly one-third of the patients were women (31% for myocardial infarction, 37% for COPD exacerbations and 28% for lung cancer). The incidence rate for lung cancer among women was 5.7 per 10 000 in 2012. The sex ratio (M/W) of the standardized rate was the lowest for COPD exacerbations (2.3) and highest for lung cancer (3.1).

Table 1

Morbidity (2015) and mortality (2014) numbers and rates for myocardial infarction, COPD and lung cancer in France

	Myocardial infarction			COPD			Lung cancera
	Men	Women	Total	Men	Women	Total	Men	Women	Total
Morbidity
Number,bn	44 822	20 489	65 311	49 344	29 814	79 158	28 152	11 226	39 378
Under 65, %	57.7	30.5	39.6	22.9	25.3	23.8	43.8	49.1	45.3
Median age, years	65.0	79.0	68.0	75.0	77.0	76.0	66.0	65.0	67.0
Average age, years	65.5	74.9	68.4	73.6	74.5	73.9	66.8	66.0	66.6
Crude rate/10 000	25.7	10.4	17.5	28.2	15.1	21.3	16.4	5.7	10.7
Standardized rate/10 000	26.5	9.0	17.0	30.6	13.2	20.4	17.2	5.6	10.8
Mortality
Number, n	8544	6049	14 593	5399	3090	8489	22 482	8887	31 369
Under 65, %	26.9	8.3	19.2	9.8	7.3	8.9	35.3	39.2	36.4
Median age, years	76.0	86.0	81.0	82.0	86.0	84.0	67.0	67.0	67.0
Average age, years	73.7	83.3	77.7	79.9	83.7	81.3	69.2	69.0	69.1
Crude rate/100 000	49.3	30.8	39.5	31.2	15.8	23.0	129.6	45.3	84.9
Standardized rate/100 000	54.0	23.3	36.6	37.0	12.0	21.3	137.6	42.6	84.1

a

The lung cancer incidence are estimated for 2012.

b

Number of patients’ hospitalized for myocardial infarction and COPD exacerbation, estimated number of incident cases for lung cancer.

Between 2002 and 2012, standardized incidence rate for lung cancer increased by 72% among women while its remained stable among men (Supplementary figure S2). With respect to patients’ hospitalized for a COPD exacerbation, the rate doubled for women between 2002 and 2015 but only increased by 30% for men. Among women, mean annual percent change in rates have increased significantly in all age groups for COPD with the highest increase in women aged 55–64 (+9.4%/year) (figure 2). For lung cancer, the increase was significant in women aged 45- to 54-year-old with the highest increase in women aged 55–64 (+8.3%/year).

Figure 2

Average annual change in tobacco-related morbidity (patients hospitalized for a myocardial infarction or for a COPD exacerbation in France between 2002 and 2015 and incidence of lung cancer between 2002 and 2012)

The rate of patients hospitalized for myocardial infarction remained stable between 2002 and 2015 (Supplementary figure S2). This overall stability masked changes that took place in each age group (figure 2). Among women aged 35–65 who were hospitalized for myocardial infarction, the rate increased by over 50% compared to 16% among men. This increase was the greatest among women between the ages of 45 and 54 with an average annual variation of 4.8%. After 65 years, the rate of patients’ hospitalized for myocardial infarction decreased for both men and women.

Tobacco-related mortality

In 2014, around 31 369 deaths due to lung cancer, 14 593 deaths due to myocardial infarction and 8489 deaths due to COPD were recorded (table 1). Women accounted for 27% of lung cancer deaths compared to 36% for COPD and 41% for myocardial infarction. Between 2000 and 2014, the lung cancer mortality rate increased by over 71% among women but decreased by 15% among men (data not shown). This increase was especially dramatic among women aged 55–64 and was high for all other age groups above 45 years old (Supplementary figure S3). Regarding COPD, the mortality rate generally increased by 3% in women between 2000 and 2014 and decreased by 21% in men. The rise in the mortality rate was especially significant among women between the ages of 45 and 64 (Supplementary figure S3). Unlike the two previous pathologies, the mortality rate for myocardial infarction decreased among men and women between 2000 and 2014. However, among the youngest age groups (35–54 years old), the drop in the mortality rate was two times greater among men than women (Supplementary figure S3).

Tobacco-attributable deaths

In 2014, the number of tobacco-attributable deaths was estimated at 54 623 for men (19.9% of all deaths) and 19 022 for women (7.0% of all deaths). Supplementary tables S3 and S4 provides more information regarding the number and the proportion of tobacco-attributable deaths for each pathology and by age. The changes in the number of tobacco-attributable deaths between 2000 and 2014 differed among men and women (Supplementary figure S4). The percentage of tobacco-attributable deaths among women increased with an average annual growth rate of 6.4% (95% CI = 5.7–7.1) between 2000 and 2014 compared to a 0.8% drop for men (95% CI = 0.7–1.0), regardless of age group (figure 3). As a result, the number of women who died in 2014 due to smoking was two times higher than the number estimated in 2000 (8027 deaths, i.e. 3.1% of all deaths).

Figure 3

Temporal trends of the fraction of deaths attributable to smoking by sex and age from 2000 to 2014 in mainland France

Discussion

A high prevalence of smoking is observed since the 1970s in French women which leads to major consequences on the morbidity and mortality rates of tobacco-related pathologies. Indeed, while the number of lung cancer cases and patients hospitalized for myocardial infarction or COPD exacerbations as well as the number of these diseases-related deaths is less than among men, the upward trend of these pathologies among women is extremely worrying, especially for those aged 45–64.

Increased tobacco-related morbidity and mortality among women

Lung cancer, COPD and myocardial infarction have long been considered to be exclusively male pathologies. Given their troubling increase in women over the past decade, these pathologies are some of the leading causes of female mortality. In the near future, if the trend is following, lung cancer is expected to outpace breast cancer to become the number-one cause cancer related deaths among women.¹⁸

Smoking remains the primary risk factor for lung cancer, COPD and myocardial infarction, especially among young people.^19,20 The rise in tobacco use is therefore enough to account for the temporal trends described in this article. With respect to myocardial infarction, while tobacco remains the main factor, especially in people under the age of 65, other risk factors could be involved in the changes observed with this pathology in people over the age of 65.

The stable rates of tobacco use among women since the early 2000s is due to two contrasting shifts. Women between the ages of 18 and 35 are smoking less, while women smoking has risen sharply among women aged 45 and more. This increase reflected a generational effect caused when women born in the 1950s entered this age group. For these generations, smoking became commonplace in the 1960s and 1970s.⁴

The consumption levels of cigarettes are still very high in France, especially compared to other countries where the prevalence of smoking among women topped out at 14% as in the United States and Great Britain in 2015.^21,22 The decrease of smoking prevalence in young women (18–34 years) could be a positive signal. If this decrease persists, it may have a beneficial effect on the incidence of myocardial infarction in the relatively short term.²³ Nevertheless, COPD and lung cancer appear after cumulative exposure to cigarette smoke amounting to several pack-years. The burden related to these pathologies among women should therefore continue to grow in the coming years, even if the prevalence of smoking decreased. This phenomenon should increase even more due to the ageing population and the improved survival rates of patients suffering from other tobacco-related pathologies such as myocardial infarction.

Similar changes in the morbidity and mortality of lung cancer have been described in many other European countries where the prevalence of smoking among women has increased like in France.²⁴

In some countries that implemented anti-tobacco policies several years ago, such as the United Kingdom, where the prevalence of tobacco smoking has dropped recently, the incidence of lung cancer among women is only slightly up and female mortality rates have stabilized.²⁵ In the United States, where the prevalence of smoking has been decreasing for the past twenty years, the incidence and mortality rate of lung cancer has dropped, including among women.²⁵ The number of hospitalizations for COPD have stopped increasing in the US since the late 1990s for both men and women.²⁶ Nevertheless, while the COPD-related mortality rate has been dropping for men since 2000, this rate has continued to increase for men, though at a much slower pace.²⁷

Changes in smoking prevention policy since the 1990s

In France, regulations have limited the marketing strategies used by the tobacco industry. The Evin law in 1991 prohibited the advertisement of tobacco products and have been reinforced to prohibit smoking in indoor public places and workplaces in 2007 and 2008. The 2016 Health law mandated plain cigarette packaging starting on 1 January 2017 and prohibited the use of flavouring and brand names that suggest that tobacco has a beneficial effect.

While the relevance of gendered cessation campaigns has not been shown, it seems that offering female smokers specific help to quit smoking could be a useful solution. Indeed, some recent studies suggest it might be relevant to differentiate between help provided to men and women on an individual level.²⁸ Increasingly, it appears that any approach geared towards helping women stop smoking must account for the differences in tobacco addiction processes between men and women,^29–31 which result in varying levels of sensitivity to existing treatment.³² This avenue has yet to be further investigated but could encourage differentiated treatment methods in clinical practice.

Strengths and limitations

The prevalence of tobacco use was based on self-reported survey data and may therefore be subject to response biases. The long-term trend in tobacco consumption has been estimated from cross-sectional surveys based on different methodologies. The survey design is stable and comparable since 2000’s. However, the marked change in smoking observed before 2000, somewhat robust regardless of the survey series, suggests that this method effect have little impact on the observed overall trend. It should be noted that the method to estimate numbers of deaths attributable to smoking did not use these measures on tobacco consumption. A notional prevalence was estimated, calibrated on lung cancer, a major marker of the smoking epidemic in the population. This approach made it possible to indirectly take into account the history of smoking at the population level (prevalence, cumulative exposure, duration of smoking, time since stopping, etc.). For morbidity, the use of hospital discharge database represents another limitation given that only patients hospitalized for a myocardial infarction and for a COPD exacerbation were counted, thereby underestimating the real burden of these diseases. Indeed, depending on the study, the annual rate of COPD exacerbations varies from 0.5–3.5 per patient and the annual rate of hospitalization ranges from 0.09–2.4 per patient.³³

Conclusions

The increase in the prevalence of tobacco use since the 1970s and the changes in morbidity and mortality markers among women is highly a warning, especially in women over 45 years. The high level of tobacco use in women suggests the burden of these pathologies will continue to increase for a number of years. While general smoking prevention campaigns should still be pursued to denormalize the use of tobacco products, extra funding could be used to develop and implement ad campaigns designed to segment the target populations and maximize their impact.

Conflicts of interest: None declared.

Key points

• Tobacco use have dramatically increase in women since the 1970s.

• A dramatic increase in myocardial infarction, lung cancer and COPD exacerbation morbidity was observed in France.

• The number of women who died in 2014 due to tobacco use was two times higher than the number estimated in 2000.

• The current high level of tobacco use in women suggests the burden of these diseases will continue to increase for a number of years.

• The differences in tobacco addiction processes between men and women argue for differential treatment methods in clinical practice to help women in smoking cessation.