UK Biobank population

The UK Biobank is a prospective cohort for the investigation, prevention, diagnosis, and treatment of chronic diseases, such as cardiovascular (CV) diseases in adults. A total of 502478 Britons from the UK National Health Service Register were included between 2006 and 2010, across 22 UK cities. The cohort was phenotyped and genotyped from participants who responded to a questionnaire and had a computer-assisted interview, from their physical and functional measures, and who provided blood, urine, and saliva samples. Data included socioeconomic, behavior and lifestyle, mental health battery, clinical diagnoses and therapies, genetics, imaging, and physiological biomarkers from blood and urine samples. The cohort protocol can be found in the literature¹².

Study population

In all, 499549 volunteers of the UK Biobank who responded to the questionnaire on smoking status were recruited. We excluded 28990 participants with previous CV events from the analyses due to the interaction between tobacco smoking and CV disorders; 179646 participants were excluded for missing data, and we thus analyzed 290913 individuals in this study (Figure 1).

Figure 1

Flowchart of participant selection

Blood pressure measurement

Systolic (SBP) and diastolic blood pressure (DBP) were measured twice at the assessment center by the use of an automated BP device (Omron 705 IT electronic blood pressure monitor; OMRON Healthcare Europe B.V. Kruisweg 577 2132 NA Hoofddorp) or manually by the use of a sphygmomanometer with an inflatable cuff in association with a stethoscope if the blood pressure device failed to measure the BP or if the largest inflatable cuff of the device did not fit around the individual’s arm.

The participant sat in a chair for all the measures. The measures were carried out by nurses trained in measuring BP. Available multiple measurements for one participant were averaged. The Omron 705 IT BP monitor satisfied the Association for the Advancement of Medical Instrumentation SP10 standard and was validated by the British Hypertension Society protocol, with an overall ‘A’ grade for both SBP and DBP¹³. Nevertheless, automated devices measure higher BP in comparison to manual sphygmomanometers, thus, we adjusted both SBP and DBP that were measured using the automated device using the algorithms¹⁴:

SBP=3.3171+0.92019×SBP (mmHg)+6.02468×sex coefficient

and

DBP=14.5647+0.80929×DBP (mmHg)+2.01089×sex coefficient

where the sex coefficient =1 for males and 0 for females.

Covariates

Diabetes status was defined on either receiving anti-diabetic medication or diabetes diagnosed by a doctor or a fasting glucose concentration ≥7mmol/L. Dyslipidemia was defined as having a fasting plasma total-cholesterol or triglycerides level of ≥6.61 mmol/L (255 mg/dL) or >1.7 mmol/L (150 mg/dL), respectively, or having statins medication. Medications were characterized by the question: ‘Do you regularly take any of the following medications?’.

Hypertension was defined as SBP of at least 140 mmHg and/or DBP of at least 90 mmHg, according to guidelines by the European Society of Cardiology, and/or antihypertensive drug used, or hypertension diagnosed by a doctor. CV diseases were defined by heart attack, angina, and stroke, as diagnosed by a doctor, and reported in questionnaires. Body mass index (BMI) was calculated as weight (kg) divided by height-squared (m²) and categorized as: high >30, moderate 25–30, and low <25 kg/m²). Biological parameters were detailed in the UK Biobank protocol. Education level was defined in three categories: high (college or university degree); intermediate (A/AS levels or equivalent, O levels/GCSEs or equivalent, other profession qualification, e.g. nursing, teaching etc.); and low (none of the aforementioned). Yearly income level (in £) was defined as: high, >52000; moderate, 18000–51999; and low, <18000.

Smoking status

Participants were categorized by self-report, as ‘current’, ‘past’ or ‘never’ smokers. Current tobacco smokers were defined as participants who responded ‘yes, on most or all days’ or ‘yes, only occasionally’ to the question: ‘Do you smoke tobacco now?’. Smoking pack-years were calculated for individuals who have ever smoked. Smoking pack-years were calculated as the average number of packs smoked per day multiplied by the total number of years of smoking in lifetime. The general definition of a pack-year is the number of cigarettes smoked per day, divided by twenty, multiplied by the number of years of smoking. In the UK Biobank, the number of years of smoking is calculated by subtracting the age of starting smoking from the age smoking was stopped (or age at inclusion for current smokers), using the equation:

Pack-years = Number of cigarettes per day/20×(age stopped smoking - age started smoking)

For current smokers, the participants had to respond to: ‘About how many cigarettes do you smoke on average per day?’; and for past smokers: ‘About how many cigarettes did you smoke on average per day?’. Participants who responded ‘never smoked’ were allocated zero for both smoking pack-years and cigarettes per day.

Alcohol consumption

Although the alcohol questionnaire has not been formally validated, several studies have shown expected associations with alcohol¹⁵. For alcohol drinker status, participants had to responded for their alcohol status: ‘current’, ‘past’, or ‘never’. Then, participants self-reported the number of alcohol units (10 mL of pure ethanol) consumed, in ‘units per week’ or ‘units per month’ (for less frequent drinkers), across numerous beverage categories (red wine, white wine/champagne, beer/cider, spirits, fortified wine, or other). The UK Biobank assessment defined units of alcohol as: a pint or can of beer/lager/cider=two units; a 25 mL single shot of spirits=one unit; and a standard glass of wine (175 mL) =two units. The number of weekly units was computed by summing all the units consumed in all categories in a week. When reported monthly, the intake was converted to units per week by dividing by 4.3. The number of weekly units was divided by 7 to determine units per day. Participants who responded ‘past’ or ‘never’ were allocated zero for daily alcohol consumption according to the UK Biobank.

Statistical analysis

Characteristics of the study population were described as mean with standard deviation (SD) for continuous variables. Categorical variables were described as number and percentage. Statistical analyses were stratified by gender since hypertension differs between men and women¹⁶ and a difference in tobacco consumption between gender was observed¹⁷. Comparisons between all groups of smoking status were performed using ANOVA tests.

Association between smoking status, smoking pack-years or cigarettes per day with alcohol status or alcohol consumption per day, and blood pressure levels (SBP and DBP), were examined with linear regression models, computing regression coefficients (B) with standard error (SE), adjusted for Model 1: antihypertensive medication + age; Model 2: model 1 + BMI; and Model 3: model 2 + diabetes, dyslipidemia, education level, and income level.

Associations between smoking status and alcohol consumption with hypertension prevalence were examined with logistic regression models with odds ratio (OR) and 95% confidence interval (CI), adjusted for Model 1: antihypertensive medication + age; Model 2: model 1 + BMI; and Model 3: model 2 + diabetes, dyslipidemia, education level and income level. Interactions were examined by including simultaneous alcohol consumption per day and smoking pack-years or cigarettes per day and their interaction term. Relationships between smoking and alcohol consumption with SBP, DBP, and hypertension were investigated in each subgroup, i.e. current, past, or never smokers. To investigate the synergistic effects between smoking pack-years/cigarettes per day and alcohol consumption on blood pressure (SBP and DBP) in current smokers, the differences in correlation were assessed using Steiger’s Z test between the adjusted individuals and combined models. Statistics were performed using SAS software (version 9.4; SAS Institute, Carry, NC). A p<0.05 was considered statistically significant.

Tobacco smoking and hypertension

Several social factors and individual behaviors can display BP levels among current smokers¹⁸. However, studies have reported that smoking increases BP⁶. High level of nicotine activates the sympathetic nervous system leading to a release of epinephrine, norepinephrine and vasopressin hormones¹⁹. Nevertheless, the chronic effect of tobacco smoking remains unclear. Several studies showed that current smokers had lower BP levels compared to non-smokers³. However, epidemiological studies showed a dose-dependent effect of smoking on BP^2,6, even if a meta-analysis highlighted no causal association between BP and smoking heaviness in current smokers⁴. Moreover, former smokers were higher hypertensive than never smokers and the risk of hypertension increased with the number and duration of cigarettes smoked²⁰. Nevertheless, there is no consensus regarding the role of chronic tobacco smoking on BP. Tobacco smoking has chemical toxicants which can have detrimental effects and damage¹⁸. Some findings showed that chronic smokers presented high BP values²¹. More so, since chronic smokers had higher SBP than those hypertensive because of old age²². Chronic smoking enhanced several pathways such as oxidative stress, alteration of nitric oxide (NO) and bioavailability, endothelial dysfunction, and then increased BP^7,20. The effect of a chronic tobacco smoking on BP can be explained by the damage caused by nicotine and carbon monoxide, two main compounds of tobacco²³. Nicotine leads to a vasoconstriction and vasoparalytic effects. In parallel, carbon monoxide affects the arterial wall and leads to irreversible damage on arteries leading to increased BP. Former smokers presented a decrease in BP only if carbon monoxide did not already affect the arterial wall²⁴. Chronic carbon monoxide exposure, as in chronic smokers, presented irreversible alterations of blood vessels²⁵. However, to date, few reports have documented and shown an association between smoking and onset of hypertension. This relationship should be mainly documented with clear evidence⁴.

Alcohol and hypertension

Consistent with previous studies, our findings highlight that alcohol consumption is significantly associated with increased BP²⁶. However, the association between hypertension and alcohol consumption remains unclear in women²⁷. A meta-analysis study assessed the presence of a gender-specific relationship between alcohol consumption and hypertension⁹. In our study, we found different thresholds for alcohol consumption and hypertension determination: 3.19 units/day for men, and 2.71 units/day for women. These findings are consistent with a recent meta-analysis showing that alcohol consumption increased the risk of hypertension in men for consumption of more than 1 to 2 drinks/day (when considering one red wine drink=2 units) while heavy consumptions were significant in both genders²⁸. Thus, a gender dose-response relationship was observed between alcohol consumption and hypertension. One of the possible explanations could be the many drinking occasions with an average of alcohol consumption among men than women (Table 1). The frequency of alcohol consumption presented different effects on BP²⁹. Previous studies have shown that the consumption amount of alcohol was associated with high BP and that the reduction in alcohol intake lowered BP in a dose-dependent response⁶. Alcohol consumption may be responsible for vasoconstriction of blood vessels, increased heart rate, activation of the sympathetic nervous system and loss in magnesium⁶.

Synergistic effects of tobacco smoking and alcohol consumption on hypertension

Drinking and smoking behaviors generally occur together³⁰. Alcohol consumption can affect the relationship between smoking and BP, whereas the relationship between alcohol consumption and BP did vary by smoking status³¹. Thus, the synergistic effects remained unclear. The fact that both alcohol consumption and tobacco smoking can interact with the sympathetic nervous system could explain a synergistic effect. However, only one study has shown that the combine reduction in alcohol consumption and tobacco smoking was associated with reduction in hypertension³². Moreover, very few studies have focused on this possible interaction to highlight this possible synergistic effect among current smokers^6,31. Our study showed an added effect of alcohol consumption on smoking pack-years and cigarettes per day in current smokers. The alcohol consumption effect could be reinforced by the neurochemical action of nicotine, explaining the added value of alcohol in smokers³³. However, we found no added effect of tobacco consumption on alcohol consumption. Alcohol is known to be highly associated with abdominal obesity and thus increased risk of obesity³⁴ whereas tobacco smoking was correlated with lower BMI³⁵. As BMI, after age, was one of the main factors of increased risk of hypertension³⁶, the absence of added effects of tobacco use on alcohol consumption observed could be explained by these inverse interactions with BMI.

Strengths and limitations

The main strength of this study is the very large sample size of the cohort. The cross-sectional observational design limits the relationship of causality. Reverse causation cannot be ruled out. The UK Biobank study showed a low response rate of 5.5% and possible volunteers bias may be involved. Nevertheless, given the large sample size and high internal validity, these are unlikely to affect the reported associations. In addition, the study cohort consisted of middle-aged European participants, so our findings may not be generalized to other age groups and ethnic populations. In addition, the UK Biobank used standardized protocols to collect anthropometric data including BP measurements; this ensures replication of data collection for all volunteers regardless of when, where and by whom they are performed and adds validity to our results. However, our study presents some limitations. Socioeconomic data were collected by self-reporting. Medical history and comorbidities have been collected by self-reporting and physician verification during medical examination in health centers. The cross-sectional design of the study may represent a limitation since reverse causation cannot be excluded. Smoking pack-years, cigarettes per day and alcohol consumption were self-reported by questionnaire. Moreover, periods of quitting smoking have not been included in calculating smoking pack-years, due to a lack of information about the duration period of stop smoking. Due to the adjustment for several factors which are causal pathways, a collider bias should be considered for the interpretation of the results observed.