Health Reports
Risk factors for hypertension in Canada

by Alexander A. Leung, Tracey Bushnik, Deirdre Hennessy, Finlay A. McAlister, and Douglas G. Manuel

Release date: February 20, 2019

Hypertension (or high blood pressure) is a leading modifiable risk factor for cardiovascular disease and accounts for more than 10% of the population-attributable fraction (PAF) for mortality worldwide.Note 1Note 2 Hypertension affects almost 1 in 4 Canadian adults,Note 2Note 3 and the lifetime incidence of developing high blood pressure is estimated to be 90%.Note 4.

Over the last several decades, impressive gains have been made in improving hypertension detection, treatment, and control,Note 5Note 6Note 7Note 8Note 9Note 10 but comparatively little work has been done in promoting primary prevention, a subject of growing interest. Encouragingly, accumulating evidence suggests that healthy lifestyle factors (such as a healthy diet and increased physical activity) contribute to lowering blood pressureNote 11Note 12Note 13Note 14 and that managing these risks can offset, at least to some extent, genetic predisposition towards hypertension and the development of subsequent cardiovascular sequelae.Note 15 Further, community-based approaches can be effective in reducing blood pressure along with other cardiovascular risk factors in the population.Note 16Note 17

The possible impact of an aging population,Note 18 worsening levels of obesity,Note 19 sedentary lifestyles,Note 20 and high sodium consumptionNote 21 on the burden of hypertension calls for a better understanding of the major risk factors associated with hypertension. Previous studies have commonly reported older age,Note 22Note 23Note 24Note 25Note 26Note 27Note 28 female sex,Note 22Note 24Note 27 increased body mass index or waist circumference,Note 22Note 23Note 24Note 25Note 26Note 27Note 28 and family history of hypertension or premature cardiovascular diseaseNote 22Note 24Note 25Note 27 to be significant predictors of high blood pressure. However, many existing prediction models were limited to people of white ethnicityNote 22Note 23Note 24Note 28 and individuals without diabetes.Note 22Note 23Note 24Note 25 As a result, their generalizability is limited. Moreover, few models have examined the association of nutrition, physical activity, or lifestyle with hypertension, and none have been derived in Canada. Quantifying the major risk factors for hypertension by means of high-quality measured data in Canada is important for informing health policy and targeted interventions nationally. To this end, using data from the first four cycles of the Canadian Health Measures Survey (CHMS), this study examined the major risk factors for hypertensive status among Canadians aged 20 to 79 years, and employed a composite risk score to predict hypertension in women and men across a wide range of ages.

Data and methods

Data source

The data are from the first (2007 to 2009), second (2009 to 2011), third (2012 to 2013), and fourth (2014 to 2015) cycles of the Canadian Health Measures Survey (CHMS). The CHMS is an ongoing survey designed to provide comprehensive direct health measures at the national level, and it collects information from community-dwelling individuals. Full-time members of the Canadian Armed Forces and people living on reserves or in other Aboriginal settlements, in institutions and in some remote regions are excluded (collectively representing approximately 3% of the Canadian population).Note 29 The CHMS involves an in-person household interview and a subsequent visit to a mobile examination centre (MEC). The household interview gathers general demographic and socioeconomic data and detailed health, nutrition and lifestyle information. At the MEC, direct physical measurements are taken, including collection of blood and urine samples. Information about medication use is obtained during the household interview and also at the MEC. CHMS participants receive an accelerometer to wear for one week to monitor activity levels. Detailed information about the CHMS is available online.Note 30Note 31Note 32Note 33Note 34

The four cycles combined had a total of 13,533 respondents aged 20 to 79. The combined non-response rate for 20- to 79-year-olds in the four cycles was 52%. For the present study, 108 subjects were excluded because of pregnancy, and another 18 were excluded because of missing values for their systolic blood pressure (SBP) and diastolic blood pressure (DBP). The resulting analytical sample size was 13,407.

Measures

Blood pressure. SBP and DBP were measured with the BpTRUTM BPM-300 device (BpTRU Medical Devices Ltd., Coquitlam, British Columbia) at the MEC. The BpTRUTM is an automated electronic monitor that has been validated and recommended for use by Hypertension Canada.Note 35Note 36 Although the BpTRUTM SBP and DBP readings have been found to be slightly lower than conventional manual blood pressure readings, the BpTRUTM readings may estimate blood pressure status more accurately.Note 37 Following a five-minute rest period, six measurements were taken at one-minute intervals for each participant while unattended, and the last five measurements were averaged together to determine the average SBP and DBP levels.Note 38 During the home interview, 56 respondents aged 20 to 79 who could not visit the MEC had their blood pressure measured with the BpTRUTM BPM-100 device.

Medication use. Current medications were recorded during the household and clinic interviews, and these were assigned to codes from the Anatomical Therapeutic Chemical (ATC) classification system, corresponding to beta blockers, agents acting on the renin-angiotensin system, thiazide diuretics, calcium channel antagonists, and other antihypertensive agents.Note 39

Diabetes. Respondents were categorized as having diabetes if their measured serum glycated hemoglobin A1c was 6.5% or higher, if they used glucose-lowering medication (ATC codes in category A10), or if they reported a diagnosis of diabetes from a health care provider.Note 40

Hypertension. Respondents were considered hypertensive if their mean SBP was 140 mm Hg or higher or mean DBP was 90 mm Hg or higher, or if they had been taking an antihypertensive medication in the month prior to the measurement being taken. Hypertensive respondents were considered controlled if they were taking antihypertensive medication and had mean SBP lower than 140 mm Hg and mean DBP lower than 90 mm Hg.

Covariates

Sociodemographic characteristics. Age was categorized into bands (20 to 39, 40 to 59, 60 to 69, and 70 to 79 years). Marital status was defined as “married or living common law,” vs. not. Highest level of education was defined as “less than secondary school graduation,” vs. “secondary school graduation or higher.” Household income (n=2,720 cases were imputed) was adjusted for household size and was categorized according to the lowest household income quintile, vs. above the lowest quintile. White or non-white ethnicity was based on respondents’ answer to which racial or cultural group they belonged.

Personal and family history. Having a regular medical doctor was categorized as “yes” or “no.” Smoking status was classified as “smoking daily or occasionally,” vs. “not smoking.” Cardiovascular disease was defined by self-reported heart disease, heart attack, and/or stroke. Family history of high blood pressure was categorized as “yes,” “no,” or “not known” (respondents with missing information [n=887] were assigned to the category “not known”). Family history of early cardiovascular disease was defined as an immediate family member being diagnosed with heart disease or stroke before the age of 60.

Risk factors. Risk factors were selected a priori according to a predefined protocol and informed by clinical reasoning. Exercise was categorized as less than 150 minutes per week of moderate-to-vigorous physical activity, vs. 150 minutes or more per week. This variable was based on four valid days of accelerometry data (and respondents with less than four valid days of data [n=3,063] were assigned to the category “not known”).

Fruit and vegetable consumption was categorized as fewer than five times per day, vs. five or more times per day (an indicator of diet quality).Note 41 This was derived from the sum of the frequency of daily consumption of the following: 100% fruit juices; fruit; tomatoes or tomato sauce (excluding tomato paste, ketchup, and pizza sauce); lettuce or green leafy salad; potatoes (including baked, boiled, and mashed potatoes, and potato salad, but excluding sweet potatoes); spinach, mustard greens or collards (excluding kale); and “other” types of vegetables (not mentioned here).

Respondents were overweight or obese if they had a body mass index (BMI) based on measured height and weight of 25.0 kg/m2 or more. The presence of diabetes was determined as described above. Chronic kidney disease (CKD) was defined as an estimated glomerular filtration rate of less than 60 mL/min/1.73 m2.Note 42 Cholesterol was categorized according to non-fasting non-high-density lipoprotein (HDL) cholesterol of 4.3 mmol/L or higher, vs. lower than 4.3 mmol/L (calculated by subtracting participants’ HDL blood measure of cholesterol from their blood measure of total cholesterol).Note 43 These factors were summed into a risk score with values ranging from 0 (no risk factors were present) to 6 (all six risk factors were present) to determine whether the component risk factors were additive.

Data were missing for some analytical variables: marital status (n=8), education (n=139), having a regular doctor (n=4), smoking (n=41), fruit and vegetable consumption (n=66), overweight or obesity (n=48), CKD (n=194), non-HDL cholesterol (n=192), risk score (n=342), and systolic blood pressure (n=8).

Statistical analysis

Descriptive statistics were used to examine the characteristics of the study population and the prevalence of hypertension. Two sets of logistic regression models were run separately for women and men, and relative risks (RRs) were estimated. Model 1 estimated the association between the risk factors and hypertension, adjusting for other covariates. Model 2 estimated the association between the risk score and hypertension, adjusting for other covariates. Model 1 was rerun six times, including an interaction term between age group and one risk factor at a time, to test whether age modified the risk factor’s association with hypertension. Model 2 was also rerun, including an interaction term between age group and the risk score. To account for the survey’s complex sampling design, all analyses were weighted on the basis of the combined survey weight from cycles 1, 2, 3 and 4 of the CHMS.Note 34 Replicate weights generated by Statistics Canada were used to perform the variance estimation (95% confidence interval [CI]) and significance testing.Note 30Note 31Note 32Note 33 The data were analyzed with SAS 9.3 and SUDAAN 11.0 (46 denominator degrees of freedom in the SUDAAN procedure statements were used).

To estimate the proportion of cases of hypertension due to a given risk factor, population-attributable fractions (PAFs) were calculated according to a modification of Levin’s formula:Note 44 PAF = Pe (RRe-1) / [1 + Pe (RRe-1)], where Pe is the prevalence of the exposure in the population (e.g., proportion of patients with diabetes) and RRe is the relative risk of hypertension due to that exposure. For this study, the prevalence estimates were taken from Table 1, and the RRes were based on adjusted RRs from Table 2. For example, the PAF attributable to diabetes among men was calculated from the prevalence of 9.4% (Table 1) and an adjusted RR of 1.68 (Table 2) to produce the estimate of 6.0% from [0.094 (1.68-1)] / [1 + 0.094 (1.68-1)].

Two sensitivity analyses were conducted. First, the average BpTRUTM blood pressure readings were adjusted on the basis of the following correction factors: adjusted BpTRUTM SBP=11.4+(0.93 × BpTRUTM SBP) and adjusted BpTRUTM DBP=15.6+(0.83 × BpTRUTM DBP).37 These adjusted values were then used to define hypertension with the 140 mm Hg SBP and 90 mm Hg DBP thresholds. This allowed for comparability between the present study and others based on manual blood pressure readings. Second, the SBP threshold of 130 mm Hg and the DBP threshold of 80 mm Hg were applied to the SBP and DBP components of the definition of hypertension, for them to be consistent with the recent blood pressure clinical guidelines from the American College of Cardiology and American Heart Association.Note 45 Under this revised definition, overall prevalence and the association between the risk factors or risk score and hypertension were examined.

Results

The population prevalence of hypertension among Canadians aged 20 to 79 was found to be significantly higher for men (24.5%, 95% CI: 22.7% to 26.4%) than for women (21.5%, 95% CI: 19.8% to 23.2%). Men were slightly younger, on average, than women, and were more likely to be smokers, have cardiovascular disease, eat fruits and vegetables fewer than five times per day, be overweight or obese, have diabetes, and have non-HDL cholesterol at a level of 4.3 mmol/L or higher (Table 1). However, women were more likely than men to engage in less than 150 minutes per week of moderate-to-vigorous physical activity, have a family history of high blood pressure or early cardiovascular disease, or have CKD. Men had, on average, a higher average risk score (2.3, 95% CI: 2.2 to 2.3) than women (2.1, 95% CI: 2.0 to 2.1).

After covariate adjustment, five out of six of the candidate risk factors were significantly associated with an increased risk of hypertension among women aged 20 to 79 (Table 2): being less physically active (RR, 1.26, 95% CI: 1.05 to 1.51), eating fruits and vegetables fewer than five times per day (RR, 1.15, 95% CI: 1.00 to 1.34), being overweight or obese (RR, 1.57; 95% CI: 1.35 to 1.83), having diabetes (RR, 2.25, 95% CI: 1.92 to 2.65), and having CKD (RR, 1.49, 95% CI: 1.21 to 1.83). Among men aged 20 to 79, being less physically active (RR, 1.19, 95% CI: 1.01 to 1.39), eating fruits and vegetables fewer than five times per day (RR, 1.18, 95% CI: 1.02 to 1.36), being overweight or obese (RR, 1.45, 95% CI: 1.18 to 1.78), and having diabetes (RR, 1.68, 95% CI: 1.41 to 2.01) were significantly associated with an increased risk of hypertension.

The proportion of cases of hypertension in the population attributable to each risk factor (PAF) was then estimated (Table 2). For both men and women, being overweight or obese was the leading risk factor for having hypertension, this factor contributing to 24% of all cases. Eating fruits and vegetables fewer than five times per day accounted for 9% of cases of hypertension in women and 12% in men. The fraction of hypertension attributable to diabetes was similar in women (8%) and men (6%). In contrast, being less physically active contributed to more cases of hypertension in women, at 12%, than in men, at 7%. In women, around 3% of cases of hypertension were attributable to CKD. Less than 1% of hypertension could be attributed to CKD in men and to elevated non-HDL cholesterol levels in either sex.

The strength of association for many risk factors varied according to age group. Among those aged 70 to 79, none of the individual risk factors were associated with hypertension in men, and only diabetes was significantly associated with hypertension in women (Appendix Table 1). Furthermore, being overweight or obese was a risk factor for women predominantly at ages 40 to 69, whereas having CKD was a risk factor largely at ages 20 to 39. For men, consuming fruits and vegetables fewer than five times per day was a risk factor at ages 40 to 59, while being less active and being overweight or obese were risk factors primarily at ages 60 to 69. Having non-HDL cholesterol at a level of 4.3 mmol/L or higher was not associated with increased risk of hypertension for adult men as a whole; however, it was associated with a significantly reduced risk of hypertension among men aged 60 to 69.

When the risk factors were summed to create a risk score, the prevalence of hypertension increased significantly (p<0.05) with each unit increase in the score for women and men aged 20 to 79 (Figure 1).

The risk score was also associated with systolic blood pressure (SBP) levels. Among the non-hypertensive population, mean SBP rose significantly with each unit increase in the risk score (Figure 2). This gradient was not evident among those with a diagnosis of hypertension, mainly because most of these individuals were treated with anti-hypertensive medications (results not shown). In fact, none of the risk factors in the score were associated with hypertension control rates (Appendix Table 2).

Sensitivity analysis

Applying the correction factors to the BpTRUTM SBP and DBP values had almost no effect on the estimate of the prevalence of hypertension (women: 22%, 95% CI: 20% to 23%; men: 25%, 95% CI: 23% to 27%) when compared with the results presented in Table 1. Similarly, there was no impact on the association between the risk factors or risk score and hypertension (results not shown). Lowering the SBP threshold to 130 mm Hg and the DBP threshold to 80 mm Hg to define hypertension resulted in an absolute 9% increase in the prevalence of hypertension among women (31%, 95% CI: 28% to 34%) and an absolute 17% increase in prevalence among men (41%, 95% CI: 39% to 44%). In terms of risk factors, non-HDL cholesterol at a level of 4.3 mmol/L or higher became significantly associated with the risk of hypertension for both women and men, while eating fruits and vegetables fewer than five times per day became only marginally significant for men (Appendix Table 3). Although the predicted prevalence of hypertension according to risk score was higher overall, a positive gradient similar to that found in Figure 1 was still present.

Discussion

In this study, nearly one-quarter of Canadians aged 20 to 79 were found to be hypertensive. Engaging in less than 150 minutes per week of moderate-to-vigorous physical activity, eating fruits and vegetables fewer than five times per day, being overweight or obese, having diabetes, and having CKD were independently associated with higher blood pressure levels and an increased risk of hypertension. When these risk factors were combined into a composite risk score (where the presence of each risk factor represented one point), a strong positive gradient was present for predicting hypertension. It is important to note that being overweight or obese, consuming fruits and vegetables less often, being less active, and having diabetes were the leading risk factors contributing to the greatest attributable fraction of hypertension cases in the population. The fact that these risk factors were not associated with hypertension control rates reflects the very high rates of awareness and antihypertensive therapy use among patients with hypertension in this population.

These findings are consistent with, and extend, those of other reports, relating healthy lifestyle and behaviours to better blood pressure control.Note 11Note 12Note 13 Participation in moderate-intensity physical activity,Note 46Note 47 consumption of a diet rich in fruits and vegetables,Note 48Note 49Note 50 and maintenance of a healthy body weightNote 51Note 52Note 53 have all been described as lowering blood pressure. In addition, counselling interventions promoting healthy behaviours (such as physical activity, healthful diet and weight management) are effective in reducing blood pressure in individuals who do not have hypertension or known cardiovascular risk factors.Note 14 However, determining the independent effects of the individual components of a healthy lifestyle on blood pressure is challenging, as these risk factors are highly interrelated. Many positive effects may be due to described physiological mechanisms.Note 11Note 12 Additional benefits may be realized through a greater tendency to adhere to medical advice and treatment among individuals who engage in healthy lifestyles. In the present study, physical activity was a stronger risk factor for hypertension among women, whereas eating fruits and vegetables was more important for men. The observed risk of hypertension in younger people is noteworthy because targeted interventions for these individuals may be particularly effective.

The associations between high blood pressure and having diabetes, having CKD, and exhibiting elevated non-HDL cholesterol are complex. Diabetes and hypertension commonly coexist. More than half of patients with diabetes also have hypertension,Note 54 and people with elevated blood pressure are nearly 2.5 times more likely to develop diabetes.Note 55 Indeed, diabetes was the strongest risk factor associated with high blood pressure in the present analysis, conferring a twofold increased risk of hypertension overall, and the greatest risk was observed in younger individuals, even after accounting for possible differences in blood pressure thresholds (i.e., 140/90 mm Hg or higher vs. 130/80 mm Hg or higher). In contrast, CKD was associated with a greater RR of hypertension among women, independent of diabetes, whereas this was not the case for men. The differences in risk between men and women may, in part, reflect the complex relationship between CKD and hypertension. For instance, the prevalence of high blood pressure among individuals with CKD varies according to the etiology of renal dysfunction,Note 56 between ethnicities,Note 57Note 58 and according to various socioeconomic factors.Note 58 Finally, elevated non-HDL cholesterol by itself was not significantly associated with high blood pressure. However, when incorporated into the risk score, it incrementally increased the risk of hypertension along with other risk factors. The importance of isolated hypercholesterolemia in hypertension prediction is uncertain and warrants further investigation.

Risk factors for hypertension are often modifiable. Most prevention efforts to date have focused on individual-level interventions promoting physical activity,Note 46Note 47 healthy diet,Note 48Note 49Note 50 and weight loss.Note 52Note 53 However, achieving clinically meaningful reductions in blood pressure may require large lifestyle changes,Note 14 which may be beyond what an average person is able to independently sustain. Accordingly, public policy promoting the requisite conditions for healthy living (e.g., by supporting walkability through the built environment; improving access to, and affordability of, fruits and vegetables; modifying food preparation to reduce dietary salt) is important.Note 16Note 17

There are many strengths to this study. This is the first population-based study conducted in Canada examining preventable risks associated with hypertension. This study is also the first to quantify the PAF of preventable risks for hypertension. In addition to examining individual risk factors for high blood pressure, this study further demonstrates a strong additive effect associated with cumulative risk factors. This suggests that many cases of hypertension in Canada are largely preventable. Data were drawn from a nationally representative sample of Canadians whose blood pressure was measured with an automated device in accordance with a standardized technique. As well, all other clinical data used in the analysis were prospectively collected in keeping with systematic methodologies.

This study has a number of limitations. First, although strong associations were observed between a number of risk factors and hypertension, the data were cross-sectional, and temporal relationships could not be established. Second, the presence of certain comorbidities, use of medications, and dietary intake were self-reported and not independently verified, and therefore open to misclassification. Third, it is likely that some of the risk factors contributed to high blood pressure through cumulative exposure. A single measurement of these factors merely provides a crude estimation of true lifetime exposure and may not be representative of day-to-day variation (e.g., in diet or physical activity). Fourth, although being overweight or obese was associated with nearly one-quarter of attributed cases of hypertension, it should be acknowledged that many of the identified risk factors were strongly interrelated (e.g., body mass index, diet and physical activity) and difficult to separate. Consequently, successful policy and interventions will likely need to account for the joint effects of multiple risk factors. Fifth, accelerometry data were used to estimate physical activity, and around one-quarter of respondents had less than four valid days of data available. Missing data were handled by dummy coding, and differential information bias could not be excluded. Sixth, dietary sodium intake and alcohol consumption were not accounted for in this study because of difficulties in accurately quantifying exposure. Finally, for certain exposures with low prevalence, such as having CKD, stratified analyses according to sex and age bands were not possible.

Conclusion

This study found that physical inactivity, a diet low in fruits and vegetables, being overweight or obese, the presence of diabetes, and the presence of CKD were strong risk factors for high blood pressure, and the risk of hypertension increased linearly with each additional exposure. Many of these risk factors are modifiable. Therefore, these findings may be important for health policy and clinical practice. Further study is needed to determine whether hypertension can be delayed or even prevented with early interventions targeting these risk factors.

Appendix

References
Date modified: