Alcohol consumption and cancer incidence in women: interaction with smoking, body mass index and menopausal hormone therapy

The Million Women Study is a population-based prospective study [39]. In median year 1998 (IQR 1998–1999), women who were invited for National Health Service (NHS) breast cancer screening at 66 screening centres in England and Scotland, were invited to join the Million Women Study by completing a postal questionnaire about socio-demographic, anthropometric, health and lifestyle factors, including how much wine, beer, and spirits they drank on average each week. Subsequent surveys have been sent to participants every 3–5 years since recruitment. The study was approved by the East of England-Cambridge South Research Ethics Committee (97/5/001) and all participants provided written consent for follow-up through medical records. Study questionnaires and data access policies are available on the study website (www.​millionwomenstud​y.​org).

All participants were followed by electronic record linkage to routinely collected NHS data on cancer registrations, deaths and emigrations. This enabled virtually complete follow-up (only 1.4% of the cohort have been lost to follow-up) [39]. Cancer diagnoses were coded to the 10^th revision of the World Health Organisation’s International Classification of Diseases. We examined the 21 most common cancers: oral cavity and pharynx (C00-C14), larynx (C32), stomach (C16), colorectum (C18-C20), liver (C22), pancreas (C25), lung (C34), malignant melanoma (C43), breast (C50), cervix (C53), endometrium (C54), ovary (C56), renal cell carcinoma (C64), bladder (C67), brain (C71), thyroid (C73), non-Hodgkin’s lymphoma (C82-C85), multiple myeloma (C90), and leukaemia (C91-C93, C95) and cancer of the oesophagus (C15), which was investigated separately by subtype, given the known differences in aetiology. Classification of oesophageal cancers by subtype was based on the following ICD-O morphology codes: oesophageal squamous cell carcinoma (OSCC): 8050–8086, and oesophageal adenocarcinoma (OAC): 8140–8145, 8190–8231, 8260–8263, 8310, 8315, 8480–8490, 8570–8575.

The baseline for analyses of alcohol intake in relation to cancer risk was the recruitment questionnaire in median year 1998. Of the 1,364,268 women eligible for the analysis, 40,640 were excluded because they had a cancer (other than non-melanoma skin cancer, C44) registered before recruitment, 10,400 were excluded because they had missing information on alcohol consumption and, given the close correlation between alcohol consumption and smoking, a further 74,597 were excluded because their smoking status at baseline was not known. A further 438,056 women were excluded because they reported drinking no alcohol or less than 1 drink per week. This exclusion was done to account for the fact that most of the women who reported drinking less than one drink per week were ex-drinkers, [40] who may have stopped drinking due to poor health and could have different characteristics from drinkers, which might be difficult to measure but could be associated with the development of cancers [41]. Given that we were interested in the potential interaction of alcohol with MHT use, the analysis sample was further restricted to postmenopausal women (defined as those who reported that they had experienced natural menopause or had undergone bilateral oophorectomy). Women who were premenopausal, perimenopausal, or of unknown menopausal status at recruitment were assumed to be postmenopausal after they reached the age of 55 years and were included in follow-up from age 55 years, because 96% of women in this cohort with a known age at natural menopause were postmenopausal by that age. Therefore only 5454 women (who had not previously reported natural menopause or bilateral oophorectomy) were excluded either because they were diagnosed with cancer prior to 55 years old or because they exited the cohort before they reached 55 years old.

After exclusions, 795,121 women remained, for whom person-years were calculated from the date that the recruitment questionnaire was completed to the earliest of the following: first registration with cancer, death, emigration, or end of follow-up on 31^st December 2018. For analyses relating to cervical and endometrial cancers (n = 600,036), women were only included in the analysis if they had not reported a hysterectomy prior to recruitment. For analyses relating to ovarian cancer (n = 699,491), women were only included if they had not reported a bilateral oophorectomy prior to recruitment. For analyses relating to breast cancer (n = 756,462), a small proportion of women (4.9%) were excluded if they returned their recruitment questionnaire by post after their breast cancer screening appointment to ensure their answers were not influenced by any information gained at their screening appointment.

Women were grouped into four categories according to the total number of alcoholic drinks (i.e. 1 glass of wine, half a pint of beer/lager/cider, or 1 measure of spirits) reported at baseline: 1–2, 3–6, 7–14, and ≥ 15 drinks per week. Women who reported drinking 1–2 drinks per week were taken as the reference category. Cox proportional hazard models were used, with time since recruitment as the underlying time variable, to estimate the hazard ratios (henceforth referred to as relative risks [RRs]) and their 95% confidence intervals (CIs) for the various cancer sites. Analyses were stratified by year of birth and year of recruitment. Analyses were adjusted for the following variables recorded at recruitment: five regions of residence (London and Southeast, Southwest, Midlands, Northern England, Scotland), deprivation quintile (according to the Townsend index which is a score incorporating census area data for employment, car ownership, home ownership and household overcrowding [42]), educational qualifications (tertiary, secondary, technical, none), smoking (never, past, current < 5, 5–9, 10–14, 15–19, 20–24, ≥ 25 cigarettes per day), oral contraceptive use (never, ever), menopausal hormone therapy (MHT) use (never, past, current), BMI (< 20.0, 20.0–22.4, 22.5–24.9, 25–27.4, 27.5–29.9, 30.0–32.4, 32.5–34.9, ≥ 35 kg/m²), and strenuous physical activity (defined as exercise that is enough to cause sweating or a fast heart beat; never/rarely/ < once, once, 2–7 times per week). In order to include the same women in all analyses, a separate category was created for the small number of women with missing data for each adjustment variable (< 5% for each variable). For analyses relating to breast, ovarian, cervical and endometrial cancers, additional confounders were also included: age at menarche (< 12, 12–13, ≥ 14 years-old) and a combined variable of age at first birth (< 25, ≥ 25 years) and parity (nulliparous, 1-2, ≥ 3); and in breast cancer analyses, adjustment was also made for family history of breast cancer (yes, no). Potential violations of the proportional hazards assumption were assessed by examination of findings separately by 5 year follow-up periods. When comparing more than one group, group-specific 95% confidence intervals were calculated to allow direct comparison between any two groups [43].

The associations between alcohol and cancer incidence were further investigated in subgroups of women according to smoking status (never, past, current), BMI (< 25, 25–29, ≥ 30 kg/m²) and use of MHT (never, past, current) at baseline. Interactions between alcohol and the other factors were tested using a likelihood ratio test, and we adjusted for false discovery rate (FDR) using the Benjamini–Hochberg method, and an FDR threshold of 0.05 was set for significance. In analyses of interactions, all upper aerodigestive cancers (defined as cancers of the oral cavity, pharynx and larynx and OSCC) were combined and their interaction with smoking status at four levels (never, past, current < 10 cigarettes/day, current ≥ 10 cigarettes/day) was also calculated and presented graphically. For hormone-related cancers (breast, cervical, endometrial and ovarian), further analyses restricted the alcohol-BMI interaction to never users of MHT and restricted the alcohol-MHT interaction to women of normal BMI (20–24.9 kg/m²).

In order to assess the likely impact of reducing alcohol intake among women who reported drinking at least 1 drink per week, we estimated the proportion of aerodigestive, breast, colorectal, and pancreatic cancers, attributable to drinking more than 1–2 drinks per week, based on the distribution of women, and estimated relative risks by alcohol (and in the case of aerodigestive cancers by alcohol and smoking).

Stata version 17 (StataCorp, College Station, TX, USA) was used for all analyses. All statistical tests were two-sided.

It has been suggested that tobacco and alcohol could have a synergistic effect on upper aerodigestive cancers because alcohol is thought to enhance the solubility, or the mucosal penetration, of tobacco carcinogens and thereby worsen the effect of smoking [46] but the prospective evidence has been uncertain. Interactions between alcohol and smoking with regard to specific upper aerodigestive sites, such as oesophageal squamous cell carcinoma, have been reported previously [10, 12‐14] but most were not statistically significant when formally tested [12‐14]. The prospective evidence is even more scarce for interactions between alcohol and smoking for cancers of the oral cavity, pharynx or larynx, with a statistically significant interaction between alcohol and smoking (p heterogeneity = 0.03) reported in the Netherlands Cohort Study for all oral cavity, pharynx and larynx cancers combined [11] and a non-significant interaction reported for oral cavity and pharynx cancer in the Japan Public Health Center cohort study [16]. Our findings clearly show that the majority of the alcohol-related excess risk of the upper aero-digestive cancers (oral cavity, pharynx, larynx and oesophageal squamous cell carcinoma) is in heavy smokers, with some excess risk in light smokers and past smokers and very little excess risk in never smokers, supporting the hypothesis that alcohol primarily acts by exacerbating the adverse effects of smoking [1]. In a previous analysis of the Million Women Study [9], the alcohol-related excess risk of upper aerodigestive cancer was only seen in current smokers, with little or no effect in never or past smokers. It is plausible that because of longer follow-up and the accrual of more cases in the analysis reported here, the small but significant alcohol-related excess risks for past smokers has emerged.

Interactions between alcohol and smoking in relation to other cancers have been investigated in several prospective studies, [17‐25, 30] although significant interactions for only two cancer sites (thyroid and renal cell carcinoma) have been reported, to our knowledge [26‐28]. However, the evidence for an interaction between alcohol and smoking on the risk of thyroid cancer is inconsistent, with one study reporting an alcohol-associated lower risk only in never smokers, [27] and the other, in a much larger study with 89,527 thyroid cases, reporting a lower risk in current smokers [28]. We found no evidence of an interaction between alcohol and smoking in relation to thyroid cancer, despite having a sufficiently large number of thyroid cases (n = 883). For all the other cancers, our findings, which are based on 100,000 more cancers than the largest previous study, did not suggest any substantial interaction between alcohol and smoking on cancer risk, other than for upper aerodigestive cancers, at least within the range of intakes observed in these middle-aged women.

It has also been suggested that alcohol and obesity may act synergistically on the risk of certain cancers due to the principal alcohol metabolite, acetaldehyde, worsening the chronic inflammatory state related to obesity [8]. There have been reports from some case–control studies of higher alcohol-associated risks of colorectal and liver cancer, in heavier individuals [7, 8], but little evidence from prospective studies. This hypothesis is not supported by our findings nor by the evidence from other prospective studies on other cancers including oesophageal squamous cell carcinoma, renal cell carcinoma and malignant melanoma [10, 24, 26]. A pooled analysis of 14 cohort studies has reported a significant interaction for pancreatic cancer, with higher alcohol-associated risks in normal weight compared with overweight and obese participants, but it was only seen in men [47].

A further hypothesis has been proposed for the effect of alcohol and BMI on hormonal cancers. Since higher alcohol intake has been related to higher levels of circulating sex hormones [4, 5], the effect of alcohol on these cancers might be expected to be proportionately less in heavier women as they tend to already have higher circulating levels of sex hormones, since adipose tissue is the main site of oestrogen production in postmenopausal women [38]. There is some evidence from prospective studies of an effect of alcohol on risks of breast and endometrial cancer being confined to lean women, [35, 38] but other studies have not found such differences in risk by BMI [3, 19, 25, 29, 36]. In these data, there was no evidence of an elevated alcohol-associated risk of hormonal cancers in lean women, even after restricting our analyses to non-MHT users, which is in line with the findings from several smaller prospective studies [3, 19, 25, 29].

Like greater adiposity, MHT use is associated with higher levels of circulating oestrogens and so if alcohol acts on female cancers through increasing oestrogen, one might expect alcohol-associated risks to be proportionately greater in never users of MHT. However, several studies have suggested that MHT use may, in fact, increase the adverse effect of alcohol on the risk of these cancers [32, 33, 38]. We, along with other studies, [3, 20, 25, 29, 34, 37], did not find any evidence of interaction between alcohol and MHT on the risk of breast, endometrial, cervical or ovarian cancer, even after restricting our analyses to women of normal BMI.

The strengths of this study lie in its prospective design, the minimal loss to follow-up for cancers, and the large number of cases, all of which enabled us to thoroughly investigate interactions with other important lifestyle factors that could modify the associations between alcohol and cancer. In addition, the analysis was conducted only in alcohol drinkers, and those who reported drinking no, or only very little, alcohol were excluded. Using a reference group of women who drank 1–2 drinks per week means that the associations are less susceptible to bias, since those most likely to have quit drinking because of illness are not included in the reference group. Although we did not test formally for departures in log-linearity of associations of alcohol intake with cancer risk, the patterns of risk across categories of intake were broadly consistent with an assumption of log-linearity. A limitation of the study is that reporting of alcohol intake can be subject to measurement error, and women may change their intakes over time, although use of repeated measures of alcohol intake for estimation of dose response effects should have helped mitigate the latter type of bias. We were only able to study the effect of alcohol and its interactions in women and not in men. Men tend to drink more alcohol than women, which might lead to differences in attributable risk between men and women, but there is little evidence to suggest that trends in risk per drink differ materially by sex, as a comprehensive meta-analysis found no significant difference in risk by sex, except for one cancer [48]. It was not possible to investigate the effects of alcohol associated with all subtypes of the many cancers examined, but reports from this cohort have been published elsewhere investigating the association between alcohol consumption and subtypes of haematological malignancies [49].