Article Text

other Versions

Low-intensity daily smoking and mortality risk among Mexican women
  1. Daniela Sarahí Gutiérrez-Torres1,
  2. Marion Brochier2,
  3. Dalia Stern3,
  4. Adrian Cortés-Valencia2,
  5. Juan Eugenio Hernández-Ávila4,
  6. Evangelina Morales-Carmona4,
  7. Tonatiuh Barrientos-Gutierrez2,
  8. Maki Inoue-Choi1,
  9. Martin Lajous2,5,
  10. Neal D Freedman1
  1. 1Division of Cancer Epidemiology and Genetics, National Cancer Institute, Rockville, Maryland, USA
  2. 2Centro de Investigación en Salud Poblacional, Instituto Nacional de Salud Pública, Ciudad de México, Mexico
  3. 3CONAHCyT – Center for Research on Population Health, National Institute of Public Health, Cuernavaca, Morelos, Mexico
  4. 4Centro de Investigación en Evaluación y Encuestas, Instituto Nacional de Salud Pública, Cuernavaca, Mexico
  5. 5Global Health and Population, Harvard T.H. Chan School of Public Health, Boston, Massachusetts, USA
  1. Correspondence to Dr Martin Lajous, Centro de Investigación en Salud Poblacional, Instituto Nacional de Salud Pública, Ciudad de México, 14000, Mexico; mlajous{at}insp.mx

Abstract

Objective To examine the association between low-intensity smoking (10 or less cigarettes per day) and all-cause and cause-specific mortality risk among women who smoke and by age at cessation among women who previously smoked.

Methods In this study, 104 717 female participants of the Mexican Teachers’ Cohort Study were categorised according to self-reported smoking status at baseline (2006/2008) and were followed for mortality through 2019. We estimated HRs and 95% CIs for all-cause and cause-specific mortality using multivariable Cox proportional hazards regression models with age as the underlying time metric.

Results Smoking as few as one to two cigarettes per day was associated with higher mortality risk for all causes (HR: 1.36; 95% CI 1.10 to 1.67) and all cancers (HR: 1.46; 95% CI 1.05 to 2.02), compared with never smoking. Similarly, slightly higher HRs were observed among participants smoking ≥3 cigarettes per day (all causes HR: 1.43; 95% CI 1.19 to 1.70; all cancers HR: 1.48; 95% CI 1.10 to 1.97; cardiovascular disease HR: 1.58; 95% CI 1.09 to 2.28).

Conclusions In this large study of Mexican women, low-intensity smoking was associated with higher mortality risk for all causes and all cancers. Interventions are needed to promote cessation among women who smoke at low-intensity in Mexico, regardless of how few cigarettes they smoke per day.

  • Low/Middle income country
  • Smoking Caused Disease
  • Prevention
  • Priority/special populations

Data availability statement

Data are available on reasonable request. The data that support the findings of this study are available from the corresponding author, ML, on reasonable request. We have adopted a data enclave approach and guidelines are available on our website https://esmaestras.org/available-mtc-data/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Accumulating evidence has shown that smoking 10 or fewer cigarettes per day (low-intensity smoking) increases the risk of death from all-causes. Yet, little evidence exists from low-income and middle-income countries, including in Mexico, where this smoking pattern is common.

WHAT THIS STUDY ADDS

  • In this large study of Mexican women, smoking as little as one to two cigarettes per day was associated with a higher risk of death from all causes and all cancers.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • These findings highlight the need to promote cessation among women who smoke even a small number of cigarettes per day.

Introduction

Despite substantial progress in tobacco control, there are currently 1.3 billion adults who smoke worldwide, 80% of whom live in low-income and middle-income countries (LMICs).1 LMICs are also the most disproportionately affected by the burden of smoking-related morbidity and mortality.2 In addition to the current decline in smoking prevalence,3 a growing proportion of people smoke 10 or less cigarettes/day (often defined as low-intensity smoking) or just on some days (non-daily smoking).4 5

Recent studies conducted in the USA, Europe and Asia have demonstrated that people who report low-intensity and non-daily smoking have higher risk of cancer6 and cardiovascular disease (CVD),7 as well as all-cause and cause-specific mortality than people who report never smoking.8–13 However, few studies have been conducted in LMICs12 14 where the low-intensity15 16 and non-daily patterns are common.17 In Mexico, nationally representative survey data indicate that the overall prevalence of current smoking remained stable between 2009 (15.9%) and 2015 (16.4%).18 19 However, in 2015, 53.7% of people who smoke cigarettes do not do so every day (daily smoking 7.6%, non-daily smoking: 8.8%) and the prevalence of cigarette smoking among males (daily: 11.9%; non-daily: 13.3%) was about three times the prevalence among females (daily: 3.6%; non-daily: 4.6%).19 Among people who smoke daily, the average number of cigarettes/day was 7.7 (males: 8.0; females: 6.8), which is substantially lower than in the USA (14.2 in 2015)4 and other regions.5 Furthermore, 43.0% of participants who smoked in last 30 days reported smoking fewer than five cigarettes/day (men: 42.8%, women: 43.6%).19

Low-intensity and non-daily smoking patterns are also common in Latin American countries such as Venezuela, Colombia, Peru, Ecuador and Chile,15 and among immigrants from this region in the USA.20 Results from the 1997–2014 National Health Interview Surveys show that 53% of the Latino population who reported smoking in the USA smoked five cigarettes/day or less, compared with 22% of the non-Latino population.21 However, there is considerable variation in the mean of cigarettes/day by sex and country of origin (Mexican women: 6.8; South American women: 7.8; Dominican women: 9.5).20

With this study, we aimed to contribute to the evidence of the association between low-intensity smoking and the risk of mortality in LMICs. Using detailed smoking information from the Mexican Teachers’ Cohort (MTC) Study, we estimated all-cause and cause-specific mortality risk by cigarettes/day among women who smoke and by age at cessation among women who previously smoked.

Methods

Study population

The study sample included participants of the MTC Study, a prospective cohort established between 2006 and 2008 that enrolled 115 313 female teachers aged 25 years and older in 12 states of Mexico. A full description of the cohort has been reported previously.22 For this analysis, baseline smoking was defined from questionnaires administered in 2006 (n=27 979 participants) and 2008 (n=106 464 participants). Of those participants who completed both the 2006 and 2008 questionnaires (n=19 130), we used the 2006 data in our main analyses. Our analytical sample included 104 717 women after excluding 9.2% (n=10 596) of participants with missing or inconsistent information on smoking (eg, women who previously smoked with missing data for age at cessation).

Measurements

Cigarette smoking and covariates

Baseline smoking status was defined using the question: Do you currently smoke tobacco? (Yes; No, because I quit; I have never smoked) and participants were classified as those who currently smoke, those who previously smoked and those who never smoked. Participants who smoked daily were asked about the number of cigarettes they smoked per day. Among this group, we created two categories (1–2 and ≥3) based on the distribution of cigarettes/day (median: 3, IQR:1–5). Participants who previously smoked were asked about the age when they quit smoking, and we categorised age at cessation into four groups: ≤29, 30–39, 40–49 and ≥50 years. Participants who quit in the last year (n=826) were classified as part of the current smoking category due to the high probability of relapse during the first year after a quitting attempt.23 Participants who completed the 2008 questionnaire were additionally asked about the age when they started smoking.

The baseline questionnaires additionally asked about other sociodemographic factors such as marriage status, rurality and indigenous ancestry. Participants who reported living with someone or being legally married were considered married. Rurality was defined if participants worked in a community of <2500 inhabitants and those who reported speaking an indigenous language or having parents who did so were considered indigenous. We also included participants’ health service provider for major illness (Mexican Institute for Social Security and Services for State Employees (ISSSTE), Mexican Institute for Social Security (IMSS), private and other public providers). Prevalent comorbidities such as diabetes, cardiac infarction, stroke and cancer were determined by self-report. Having a history of chronic disease was defined when participants answered ‘Yes’ or provided an age, as requested, to at least two of the following questions: ‘Have you been diagnosed by a doctor with any of the following diseases? (Yes/No)’, ‘At what age did you receive the diagnosis?andDid you receive treatment? (Yes/No)’. Smoking status and cigarettes smoked per day were reassessed in the 2011 follow-up questionnaire, which was completed by 86 376 cohort participants.

To evaluate socioeconomic status (SES), we developed a score by adding one point for ownership of a list of up to seven assets (car, telephone, cellphone, microwave, vacuum, computer and access to internet) and dividing the population into tertiles (≤3, 4–5 and ≥6 assets). We chose this index of household assets because it has a moderate correlation (Pearson r=0.52) with a previously developed multidimensional socioeconomic position (SEP) index in this population24 and was as robust as the SEP index for classifying women based on risk factors such as obesity and current cigarette smoking that vary by SES. For the physical activity assessment, participants reported the average number of hours per week spent walking and doing moderate and vigorous physical activity during work and leisure time over the last 12 months (range: 0–>10 hours/week). Then, this time was multiplied by the corresponding metabolic equivalents of task (METs) for each type of activity (mild=2.2, moderate=4.7 and vigorous=6.0) to categorise the population in tertiles of physical activity (≤16.4, 16.5–33.0 and ≥33.1 METs). MTC’s physical activity questionnaire was recently validated against the International Physical Activity Questionnaire, and a moderate correlation was found for reported intensities (range: Pearson r=0.45–0.80) in a subsample of the cohort.25 Additionally, a variable assessing sedentary behaviour was created using the number of hours per week watching television (<7, 7.1–14 and >14 hours) as a proxy.26 Participants’ dietary quality was assessed using the Alternative Healthy Eating Index 2010 (AHEI-2010),27 which ranges from 0 (lowest quality) to 110 (highest quality), and alcohol intake (0, <0.1, 0.1–0.2, >0.2 drinks per day) was derived from the reported frequency of alcoholic beverages consumed in the last year. Finally, secondhand smoke (SHS) exposure in adulthood was defined from the participant’s report of the number of years living with someone who smokes (<1, 5–9, 10–19 and ≥20 years).

Outcome ascertainment

Deaths were identified through employer databases (updated once a year), next-of-kin reports and cross-linkage with three major pension-fund management databases. Deaths were then confirmed through two national mortality registries where the date of death, main and contributing causes of death were collected.28 The National Institute of Statistics, Geography, and Informatics dataset was available from 1 January 2006 to 31 December 2016, and the Ministry of Health Epidemiological and Death Statistics System from 1 January 2006 to 31 December 2019. Causes of death were coded using the 10th revision of the International Classification of Diseases and classified using the following coding: all cancers (C00-C96), CVD (cerebrovascular disease (I60-I69), ischaemic heart disease (I20-I25)) and diabetes (E10-E13). All cohort participants were cross-linked using a probabilistic record linkage algorithm specifically developed for Mexico, by the Public Health Intelligence Unit at the National Institute of Public Health.29 The accuracy of mortality registries for identification of deaths in Mexico was recently assessed and found to be high, with a sensitivity of 87.9% (95% CI 85.3 to 90.6) and a specificity of 99.1% (95% CI 97.9 to 99.7).28 For the main analyses, follow-up time started on the date of completion of the baseline questionnaire (2006 or 2008) and continued through the date of death or 31 December 2019, whichever came first. When different, the start of follow-up time for sensitivity analyses is specified further.

Statistical analyses

Cox proportional hazards regression models were used to calculate HRs and 95% CIs for each mortality outcome (all causes, all cancers and CVD including diabetes), using age as the underlying time metric and stratifying the baseline risk by 5-year age groups.30 For analyses evaluating associations of current smoking, we used people who had never smoked as the reference group, and for the analysis for age at cessation, the reference was people who reported current smoking. All models were adjusted for marital status (married/partnered and single), rurality (yes or no), indigenous language (yes or no), SES (tertiles, ≤3, 4–5 and ≥6 assets), healthcare providers (ISSSTE, IMSS, private and other public provider), physical activity (tertiles, ≤16.4, 16.5–33.0 and ≥33.1 METs), time watching TV (<7, 7.1–14 and >14 hours per week), alcohol consumption (0, 0.1–0.2 and >0.2 drinks per day), AHEI-2010 score (tertiles, 18.5–49.8, 49.9–57.8 and 57.9–97.0) and SHS exposure (<1, 5–9, 10–19 and ≥20 years). For analyses of cigarettes/day among people who smoke, models were additionally adjusted for age at smoking initiation to evaluate if this association was independent of the duration of exposure. Indicator variables were included in the models for covariates with missing information. We did not adjust for comorbidities such as diabetes, CVD or cancer because those conditions are related to mortality and can be related to cigarette smoking. To minimise reverse causality, we conducted two sensitivity analyses: (1) excluding participants with baseline comorbidities (diabetes, cardiac infarction, stroke and cancer) as participants with these comorbidities may have been more likely to quit smoking and (2) excluding deaths that occurred within 2 years of follow-up. As an additional sensitivity analysis, we adjusted for body mass index (BMI).

To evaluate longitudinal changes in cigarette smoking, we compared responses at baseline (2006 and 2008) with responses to the 2011 questionnaire among 82 234 study participants (78.5%) who completed the follow-up questionnaire. For the 2011 questionnaire, participants responded whether they were currently smoking or not and also provided the category that best described the number of cigarettes they smoked per day (1–4, 5–14, 25–34, 35–44 and ≥45). We created a new categorical variable for the number of cigarettes per day to compare both assessments (1–4, 5–14 and ≥15).

As a sensitivity analysis, we modelled smoking as a time varying exposure by incorporating smoking information in 2011. In separate analyses, we restricted to participants who completed the 2011 questionnaire and among this subset compared mortality associations with smoking information collected at baseline and smoking information collected in 2011. Analyses were conducted using SAS V.9.4 software (SAS Institute, Cary, North Carolina, USA). Statistical significance was defined as p<0.05, and all tests were two sided.

Results

At baseline, 79.3% (n=83 053) of participants had never smoked, 9.8% (n=10 253) reported smoking in the past and 10.9% (n=11 411) reported currently smoking. Participants who smoked reported an average of 4.0 cigarettes/day (SD: 4.0). Compared with women who have never smoked, paticipants who smoked daily were less likely to be married, work in a rural community and speak an indigenous language. A higher proportion of women who smoked were in the highest tertile of SES, spent >14 hours per week watching TV, had a higher consumption of alcohol (median: 0.06 drinks/day) and scored in the lowest tertile of AHEI-2010. Approximately 42% (n=4 751) of participants who reported smoked everyday also reported smoking one to two cigarettes/day (mean 1.4, SD: 0.5) and had a median age at initiation of 22 years. Compared with the one to two cigarettes/day group, women who reported smoking ≥3 cigarettes per day (mean: 6.4, SD: 4.4) started smoking earlier (median age of 20 years) and reported more years of exposure to SHS. Among participants who reported smoking in the past, the median age at cessation was 30 years (table 1).

Table 1

Demographic characteristics by baseline smoking status among women in the Mexican Teachers’ Cohort (MTC) Study

Of 82 234 respondents of the 2011 questionnaire, 36.7% (n=3 259) of participants who smoked at baseline reported no longer smoking in 2011. Conversely, 0.4% (n=245) of participants who reported having never smoked and 5.9% (n=473) of participants who reported formerly smoking at baseline reported smoking in 2011. Among those who smoked one to four cigarettes/day at baseline and continued to smoke in 2011, 95.9% (n=2 597) reported smoking one to four cigarettes/day at both timepoints (online supplemental table 1).

Supplemental material

There were 2 001 deaths during a mean (SD) follow-up period of 10.6 (1.5) years, and the two most common causes of death were cancer (n=715) and CVD (n=443). Compared with participants who had never smoked, women who currently smoked had 1.32 times the risk of all-cause mortality (95% CI 1.15 to 1.52) and 1.36 times the risk for both cancer mortality (95% CI 1.09 to 1.71) and CVD mortality (95% CI 1.01 to 1.84) (table 2). Smoking reported on the 2011 follow-up questionnaire had similar associations with mortality as smoking reported on the baseline questionnaire (online supplemental table 2). However, associations strengthened slightly in analyses that incorporated changes in smoking between baseline and 2011. For example, the HR for current smoking with all-cause mortality was 1.41 (95% CI 1.21 to 1.63) (table 2).Additionally, associations were similar after adjustment for BMI (online supplemental table 3)

Table 2

All-cause and cause-specific mortality risks by smoking at baseline (2006–2008) and smoking at the first follow-up assessment (2011)

Regarding the number of cigarettes, smoking as few as one to two cigarettes/day was associated with higher mortality risk for all causes (HR: 1.36; 95% CI 1.10 to 1.67) and all cancers (HR: 1.46; 95% CI 1.05 to 2.02) compared with never smoking. Mortality risks were higher among women who smoked ≥3 cigarettes/day compared with women who had never smoked (all causes HR: 1.43; 95% CI 1.19 to 1.70; all cancers HR: 1.48; 95% CI 1.10 to 1.97; CVD HR: 1.58; 95% CI 1.09 to 2.28). HRs for all-cause, cancer and CVD mortality were somewhat stronger after additional adjustment for age at smoking initiation (table 3), although CIs were wider due to a smaller number of respondents. Associations for age at smoking initiation and mortality are shown in online supplemental table 4.

Table 3

All-cause and cause-specific mortality risks among participants who currently smoked by number of cigarettes/day at baseline

Stronger associations for all-cause mortality were also observed in sensitivity analyses that excluded participants who died within the first 2 years of follow-up (one to two cigarettes/day HR: 1.29, 95% CI 1.03 to 1.63; ≥3 cigarettes/day HR: 1.42, 95% CI 1.17 to 1.72) and after excluding participants who reported a history of diabetes, cardiac infarction, stroke or cancer (one to two cigarettes/day HR:1.37, 95% CI 1.08 to 1.74; ≥3 cigarettes/day HR: 1.55, 95% CI 1.27 to 1.90) (online supplemental table 3).

Compared with women who smoked daily, mortality risks for all causes, all cancers and CVD were lower among women who quit smoking before age of 40 years. Among those who quit before 30 years old, the HRs (95% CI) were 0.71 (0.54 to 0.94), 0.62 (0.38 to 0.99) and 0.74 (0.41 to 1.33), respectively (figure 1).

Figure 1

All-cause and cause-specific mortality risk by age at cessation among women who previously smoked. *Currently smoking includes 826 participants who quit smoking in the last year. aData not reported due to a small number of deaths (n<10). bCardiovascular disease includes cerebrovascular disease, ischaemic heart disease and diabetes. cModel adjusted for marriage status, rurality, indigenous language, SES, healthcare provider, physical activity, TV watching time, alcohol consumption, AHEI index and SHS exposure. Baseline risk was stratified by 5-year age groups. AHEI, alternative healthy eating index; SES, socioeconomic status; SHS, secondhand smoke.

Discussion

In this large prospective cohort study of Mexican women, participants who reported smoking one to two cigarettes/day had a higher risk of mortality from all causes and all cancers than women who never smoked. Slightly higher risks were observed among participants who smoked ≥3 cigarettes/day. We also observed lower mortality risk with smoking cessation, particularly among women who quit at a younger age.

Results from this analysis are consistent with a growing number of studies conducted in the USA,6 8 10 Asia12 13 and Europe,31 showing an association between low-intensity daily smoking with a higher risk of mortality in both women and men. In the NIH-AARP (National Institutes of Health-American Association of Retired Persons) Diet and Health Study,8 women who consistently smoked <1 and 1–10 cigarettes per day had 1.50 (95% CI 0.72 to 3.15) and 1.92 (95% CI 1.63 to 2.27) times the mortality risk than women who have never smoked. In a subsequent analysis, Inoue-Choi et al10 examined the association between low-intensity and non-daily smoking with mortality risk among adult respondents of the 1992–2011 Tobacco Use Supplements to the Current Population Survey (TUS-CPS) in the USA. Participants who reported current non-daily smoking were stratified by whether they previously smoked daily or had never smoked daily (lifelong non-daily smokers). Compared with people who smoke daily, participants who smoked non-daily throughout their lives were younger (<45 years) and more likely to be from racial/ethnic minority groups (non-Hispanic black, Hispanic or other). Similar HRs for all-cause mortality were observed in men and women who reported never smoking every day. However, the risk of all-cause mortality was lower among Hispanic individuals who reported lifelong non-daily smoking (HR: 1.18, 95% CI 0.73 to 1.89) than among non-Hispanic white individuals (HR: 1.81, 95% CI 1.59 to 2.05) and non-Hispanic black individuals (HR: 1.88, 95% CI 1.56 to 2.25), with a higher smoking intensity associated with higher mortality risk. Similar dose–response associations were observed in a pooled analysis of nine cohort studies in Japan where HRs (95% CI) for all-cause mortality in women were 1.28 (1.00 to 1.62) for one to two cigarettes/day, 1.49 (1.34 to 1.66) for three to five cigarettes/day and 1.68 (1.55–1.81) for 6–10 cigarettes/day.13

Although smoking-related deaths represent a substantial proportion of premature deaths in Latin American countries,32 there is a paucity of studies examining the mortality risks of low-intensity smoking. In Mexico, only one study besides ours has evaluated this association. In the Mexico City Prospective Study, established between 1998 and 2004, smoking less than 10 cigarettes/day was associated with higher all-cause mortality risk (RR=1.17, 95% CI 1.10 to 1.25) relative to never smoking, with similar associations observed in men and women.14 It is unclear why our estimates are stronger compared with this study. However, there are some differences between the two cohorts. First, the Mexico City prospective study enrolled men and women from two urban districts. In contrast, the MTC study enrolled only women and the source population came from a more geographically diverse area in Mexico, which may better capture the variability in smoking behaviours that have been reported in national surveys.33 Second, in the Mexico City Prospective Study, the questionnaire was filled by trained nurses during an in-person interview with the participants, while in the MTC study, the smoking assessment was self-reported. However, even though we cannot rule out some errors from recall and social desirability bias, smoking intensity was lower in the MTC than in the Mexico City Prospective Study (mean MTC: 4.0; mean Mexico City Prospective Study: 6.0). In addition, MTC participants were slightly younger (mean age (SD): 43 (7) years) compared with female participants in the Mexico City Prospective Study (mean age (SD): 46 (10) years). For this reason, it will be interesting to examine the effects of smoking in the MTC study population in the future as the cohort ages.

One important strength of our analysis was the availability of repeated smoking assessment during follow-up. We found that 37% of participants who smoked daily at baseline reported no longer smoking in the 2011 questionnaire. Additionally, the magnitude of associations for current versus never smoking strengthened slightly when smoking was treated as a time-varying exposure in our models. We also found that the number of cigarettes/day was consistently low at both the baseline and the 2011 assessment. This observed concordance provides reassurance that associations with mortality in our study likely reflect the effect of long-term low-intensity daily smoking, rather than participants increasing their number of cigarettes/day during follow-up.

We also evaluated the association of smoking cessation with mortality risk. Compared with women who smoked daily, those who quit at younger age reduced substantially their risk for all-cause mortality (≤29 years HR: 0.71, 95% CI 0.54 to 0.94) as well as their risk for cancer mortality (≤29 years HR: 0.62, 95% CI 0.38 to 0.99). Prior studies have observed similar associations between smoking cessation and lower mortality risks.34 In the aforementioned NIH-AARP cohort study,8 compared with people who never smoked, all-cause mortality risk for smoking 1–10 cigarettes/day was lower among participants who quit at 20–29 years (HR: 1.11, 95% CI 1.00 to 1.22), than among those who quit at an older age (≥50 years HR:1.42, 95% CI 1.27 to 1.59). In the TUS-CPS study,10 the lowest all-cause mortality risk was observed among participants who reported more than 10 years since quitting (HR: 1.09, 95% CI 1.04 to 1.15). In the Mexico City Prospective Study, compared with people who never smoked, participants who smoked in the past had half of excess of mortality risk (RR=1.15, 95% CI 1.12 to 1.22) of those who smoked daily (RR=1.29, 95% CI 1.25 to 1.37).14

Additional strengths of our analysis include a prospective design and a large and geographically diverse sample of women. Within the cohort, 90% of participants who smoked daily reported smoking 10 cigarettes/day or less, which is consistent with previous reports from national surveys.18 19 33 The cohort also uses well-validated procedures to linkage mortality data using independent databases. Our study also had several limitations. First, smoking assessment was based on self-reports and is subject to recall bias. However, a strong concordance between cotinine concentrations and self-reported smoking status has been reported among Mexican adolescents35 and among Mexican American adults in the USA.36 Second, our study did not assess non-daily smoking, which is common in Mexico.33 37 Third, our statistical power was limited for the analyses of less-common causes of death. Fourth, the MTC participants are teachers and thus are educated women with job stability. Studies are needed in other socioeconomic groups, including men. Lastly, as in other observational studies, we cannot discard the possibility of residual and unmeasured confounding, although our study benefited from a comprehensive adjustment of other mortality risk factors including alcohol, diet, SHS exposure, BMI, physical activity and SES.

In summary, we observed that smoking as little as one to two cigarettes/day was associated with a higher risk of all-cause and cancer mortality among women in Mexico. Our results highlight the adverse health effects of low-intensity smoking and support efforts to promote cessation among women who smoke at low intensity, including in Latin America.

Data availability statement

Data are available on reasonable request. The data that support the findings of this study are available from the corresponding author, ML, on reasonable request. We have adopted a data enclave approach and guidelines are available on our website https://esmaestras.org/available-mtc-data/.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by Comité de ética en investigación, Instituto Nacional de Salud Pública de México ID 1645. At baseline, response and delivery of questionnaire to study investigators was considered consent by the institutional ethics committee.

Acknowledgments

We recognised study participants and INSP’s research support staff for their time and commitment to advance the health of women in Mexico. We would like to thank the Servicio Profesional Docente at the Mexican Ministry of Education, with special thanks to Victor Sastré, Director of Promotion in Service. We also thank all the staff that has participated in questionnaire distribution from the State Ministries of Education of Baja California, Chiapas, Ciudad de México, Durango, Guanajuato, Hidalgo, Jalisco, Estado de México, Nuevo Léon, Sonora, Veracruz and Yucatan. Finally, we would like to acknowledge ISSSTE, IMSS, ISSSTECali, ISSTELeon, ISSSTESon, ISSTECH, ISSEG and ISSTEY for their continued support of the project.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Contributors Guarantor: ML. Conceptualisation: DSG-T, ML and NDF. Data curation: DS and AC-V. Methodology: TB-G, DSG-T, MI-C and NDF. Formal analysis and writing – original draft: DSG-T. Investigation: JEH-A, EM-C, MB and ML. Funding acquisition and supervision: NDF. Writing – review and editing: all authors.

  • Funding This study was supported by National Cancer Institute, Division of Cancer Epidemiology and Genetics, Intramural Research Program.

  • Disclaimer The comments and opinions expressed in this paper are those of the authors and does not reflect the official policy of the Department of Health and Human Service, National Institutes of Health and National Cancer Institute.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.