Relationship Between Cotinine-Verified Smoking Status and Incidence of Hypertension in 74,743 Korean Adults

Abstract

Background: The relationship between chronic smoking and hypertension (HTN) is inconclusive in previous studies, which were mainly based on self-reported smoking status. The aim of this study was to evaluate the association of cotinine-verified smoking status with incident HTN.

Methods and Results: A total of 74,743 participants (43,104 men; age 38±5.4 years) were included in the study, with a mean follow-up period of 29 months. Individuals were divided into 4 groups on the basis of their cotinine-verified smoking status at baseline and at follow-up (never-smoking, new-smoking, former-smoking, and sustained-smoking). The incidence rate of HTN in the never-smoking, new-smoking, former-smoking, and sustained-smoking groups was 8.2%, 7.6%, 10.1%, and 8.7% for men and 1.8%, 2.5%, 1.5%, and 2.2% for women, respectively. In a multivariate Cox-hazard regression analysis adjusted for the variables with a univariate relationship, new-smoking and sustained-smoking had decreased relative risks (RRs) for incident HTN compared with never-smoking (RR [95% CI], 0.75 [0.58, 0.96] for new-smoking and 0.82 [0.74, 0.90] for sustained-smoking). Cotinine-verified current smoking at baseline was also inversely associated with incident HTN compared with cotinine-verified never-smoking at baseline (0.91 [0.84, 0.98]). These results remained significant only in men, although there was no sex interaction.

Conclusions: This longitudinal study showed that cotinine-verified new-smoking and sustained-smoking decreased the risk for incident HTN, especially in men, compared with never-smoking.

Cigarette smoke has various detrimental effects on the cardiovascular system.¹ However, the relationship between chronic smoking and hypertension (HTN) remains inconclusive. Previous cross-sectional studies have shown diverse results in their assessments of the association between chronic smoking and HTN.^2–7 We reported that cotinine-verified current smoking was inversely associated with HTN and systolic and diastolic blood pressure (BP), especially in men.⁸ However, because of their cross-sectional design, the previous studies could not determine a causal relationship.

Editorial p 1510

Several longitudinal studies have assessed the relationship between chronic smoking and incident HTN but have also shown inconsistent results.^9–13 Previous studies could not exclude a misclassification error of smoking status because the studies were based on participants’ responses to self-reported questionnaires regarding smoking status. They also only assessed the effect of baseline smoking status on incident HTN; studies evaluating the effect of cotinine-verified smoking on incident HTN are lacking. Cotinine-verified assessment of smoking status has been used to validate self-reported smoking status because participants’ responses on self-reported questionnaires might underestimate the actual prevalence of cigarette smoking.^14,15 Therefore, the aim of the present study was to assess whether a change of cotinine-verified smoking status at baseline and at follow-up would affect incident HTN, and to determine the relationship between cotinine-verified smoking status at baseline and incident HTN in a large Korean population. In addition, this study was done to evaluate if there is the difference of the results in subgroup analyses according to sex.

Methods

Study Population

The Kangbuk Samsung Health Study (KSHS) is a retrospective cohort study of adult Korean men and women who underwent a comprehensive annual or biennial examination at the Total Healthcare Centers of the Kangbuk Samsung Hospital from 2002 to 2011, and the Kangbuk Samsung Cohort Study (KSCS) is a prospective cohort study of the same individuals since 2012 and is ongoing. Among 82,016 individuals who had urinary cotinine measurement data and data for BP and HTN and who visited between 2011 and 2012 at baseline and were followed at 2014, we included 80,777 individuals and excluded 1,239 individuals with missing data for medical history, antihypertensive medication, or urinary cotinine measurements at follow-up. Of those, 74,743 individuals without HTN at baseline (43,104 men; mean age±standard deviation, 38±5.4 years) were finally included in this study, with a mean follow-up period of 29 months (12–44 months). This study was approved by the Institutional Review Board of Kangbuk Samsung Hospital (IRB No: 2016-04-072) and all participants provided written informed consent.

BP Measurements

The procedures of the anthropometric and laboratory measurements were the same as described in our previous study.^8,16 BP was measured 3 times by trained nurses using an automated oscillometric device (53000-E2, Welch Allyn, NY, USA) while subjects were seated with the arm supported at heart level and the 3 readings were averaged to obtain the systolic BP (SBP) and diastolic BP (DBP); 1 BP measurement was recorded in individuals enrolled in 2011. HTN was defined as high BP (SBP ≥140 mmHg and/or DBP ≥90 mmHg) or current use of antihypertensive medication.

Definition of Self-Reported and Cotinine-Verified Smoking Status

Urinary cotinine concentration was measured using the DRI Cotinine Assay (Microgenics Corp., Fremont, CA, USA) with a modular P800 (Roche Diagnostics, Tokyo, Japan) after stopping smoking at least 10 h. Cotinine-verified current smoking was defined as a urinary cotinine level ≥50 ng/mL. The sensitivity and specificity values of the cotinine cutoff level of 50 ng/mL for distinguishing current-smoking from never-smoking were 84.8% and 98.2%, respectively.

Individuals were divided into 4 groups on the basis of cotinine-verified smoking status at baseline and at follow-up: never-smoking as cotinine level <50 ng/mL at baseline and <50 ng/mL at follow-up; new-smoking, with cotinine level <50 ng/mL at baseline and ≥50 ng/mL at follow-up; former-smoking as cotinine level was ≥50 ng/mL at baseline and <50 ng/mL at follow-up; and sustained-smoking, with cotinine level ≥50 ng/mL at baseline and ≥50 ng/mL at follow-up.

Assessment of self-reported smoking status was the same as described in our previous study.¹⁶ Individuals were also divided into 4 categories according to self-reported smoking status at baseline and at follow-up: never-smoking, as no smoking at baseline and follow-up; new-smoking, as no smoking at baseline but currently smoking at follow-up; former-smoking, as smoking at baseline but quitting smoking by follow-up; and sustained-smoking, as smoking continuously at baseline and follow-up.

Unobserved smoker was classified as cotinine-verified current-smoking among the self-reported never-smokers.

Statistical Analysis

Data are expressed as the mean±standard deviation or as the median [interquartile range] for continuous variables; the 2 exceptions were cotinine and daily alcohol consumption, which are expressed as mean±standard error of mean. Data for categorical variables are expressed as percentages (%). Among the variables, serum triglycerides and high-sensitivity C-reactive protein (hsCRP) were log-transformed for analysis, and the urinary cotinine level was entered as log (cotinine+1) to diminish heteroskedasticity. The characteristics of the 4 cotinine-verified smoking groups were compared using ANOVA’s t-test, Chi-square test, or Fisher’s exact test. Posthoc comparisons between the 4r groups were performed using Scheffe’s multiple comparison. Comparisons of the characteristics between individuals with and without incident HTN were assessed with Student’s t-test or the Chi-square test. Multivariate Cox-hazard regression analyses were performed to evaluate the effects of self-reported and cotinine-verified smoking status on incident HTN. A multivariate model (model 1) was adjusted for age, sex, body mass index, frequency of vigorous exercise (≥5 times/week), frequency of alcohol consumption (≥3 times/week), and presence/absence of diabetes. Model 2 was adjusted for variables in model 1 and the variables with a univariate relationship (P<0.05), comprising blood urea nitrogen (BUN), creatinine, uric acid, total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein (LDL) cholesterol, triglycerides, and hsCRP. PASW version 18 (SPSS, Chicago, IL, USA) was used to perform statistical analyses, and P<0.05 was considered statistically significant.

Results

Among the 74,743 participants, the overall incidence of HTN was 5.7%, with rates of 8.6% for men and 2.8% for women. The respective incidence rates of HTN among the 4 groups according to the change of cotinine-verified smoking status from baseline to follow-up were 4.7%, 7.0%, 9.5%, and 8.3% in the never-smoking, new-smoking, former-smoking, and sustained-smoking groups (P<0.001); 8.2%, 7.6%, 10.1%, and 8.7% in these same groups among men (P<0.001) respectively; 1.8%, 2.5%, 1.5%, and 2.2% in these same groups among women (P=0.611), respectively (Figure).

Figure.

Incidence of new hypertension according to cotinine-verified smoking status. *Chi-square test for men and Fisher’s exact test for women were used. P<0.001 for men, P=0.611 for women.

The differences in the baseline characteristics of the 4 groups according to cotinine-verified smoking status in the overall population were significant for all variables shown in Table 1. For men, there were significant differences for all variables among the 4 groups (Table S1). For women, there were significant differences among the 4 groups in all variables except baseline SBP and DBP, total cholesterol, HDL-C, creatinine, hsCRP, frequency of vigorous exercise, and prevalence of diabetes (Table S2).

Table 1. Baseline Characteristics of the 4 Cotinine-Verified Smoking Status Groups

	Never-smoking (n=56,094)	New-smoking (n=1,207)	Former-smoking (n=5,699)	Sustained-smoking (n=11,743)	P value
Age, years	38.0±5.5*,‡	38.9±5.3*,§	37.9±5.4§,¶	39.0±5.2‡,¶	<0.001
Sex (male), n (%)	25,685 (45.8)	1,047 (86.7)	5,307 (93.1)	11,065 (94.2)	<0.001
Weight, kg	63.3±12.0*,†,‡	72.0±10.8*	72.5±10.9†	72.5±11.0‡	<0.001
Body mass index, kg/m2	22.7±3.1*,†,‡	24.4±3.0*	24.3±3.0†	24.3±3.1‡	<0.001
Waist circumference, cm	79.9±9.0*,†,‡	85.6±8.3*	85.6±8.4†	85.6±8.4‡	<0.001
Systolic BP, mmHg	106.7±11.9*,†,‡	111.7±10.7*	112.4±10.6†,¶	111.8±10.4‡,¶	<0.001
Diastolic BP, mmHg	68.2±8.8*,†,‡	71.5±8.2*	72.2±8.3†	71.9±8.2‡	<0.001
Total cholesterol, mmol/L	4.97±0.86*,†,‡	5.24±0.92*	5.22±0.91†	5.19±0.90‡	<0.001
Triglycerides, mmol/L	0.94 [0.69, 1.37]*,†,‡	1.30 [0.90, 1.88]*,§,[]	1.38 [0.98, 2.01]†,§,¶	1.42 [0.99, 2.06]‡,[],¶	<0.001
LDL-C, mmol/L	3.05±0.80*,†,‡	3.35±0.85*	3.31±0.83†	3.30±0.83‡	<0.001
HDL-C, mmol/L	1.54±0.38*,†,‡	1.40±0.36*,§,[]	1.36±0.34†,§,¶	1.33±0.33‡,[],¶	<0.001
Glucose, mmol/L	5.22±0.64*,†,‡	5.43±0.94*	5.40±0.81†	5.42±0.88‡	<0.001
BUN, mmol/L	4.34±1.08*,†,‡	4.53±1.08*	4.47±1.03†	4.51±1.05‡	<0.001
Creatinine, μmol/L	74.2±16.4*,†,‡	84.1±13.2*	81.0±11.7†	85.1±11.6‡	<0.001
Uric acid, μmol/L	298.6±84.8*,†,‡	353.2±72.4*	363.9±75.1†	354.9±75.0‡	<0.001
hsCRP, mg/L	0.4 [0.2, 0.8]*,†,‡	0.5 [0.3, 1.0]*	0.5 [0.3, 1.0]†,¶	0.5 [0.3, 1.1]‡,¶	<0.001
Cotinine, ng/mL	0.3±0.0*,†,‡	2.8±0.3*,§,[]	837.3±10.0†,§,¶	1,265.8±8.2‡,[],¶	<0.001
Self-reported smoking status, n (%)					<0.001
Never-/former-smoking	45,429/46,648 (97.4)	764/1,116 (68.5)	689/5,386 (12.8)	784/10,940 (7.2)
Current-smoking	1,219/46,648 (2.6)	352/1,116 (31.5)	4,697/5,386 (87.2)	10,156/10,940 (92.8)
Daily alcohol consumption, g/day	10.6±0.1*,†,‡	23.1±0.8*	21.2±0.3†,¶	23.3±0.2‡,¶	<0.001
Alcohol consumption (≥3 times/week), n (%)	4,955/50,941 (5.7)	251/1,166 (21.5)	1,177/5,584 (21.1)	2,780/11,411 (24.4)	<0.001
Vigorous exercise (≥5 times/week), n (%)	2,206/54,146 (4.1)	49/1,180 (4.2)	141/5,565 (2.5)	348/11,435 (3.0)	<0.001
Metabolic syndrome, n (%)	2,729/43,380 (6.3)	132/937 (14.1)	573/4,059 (14.1)	1,340/8,722 (15.4)	<0.001
Diabetes mellitus, n (%)	1,008/56,083 (1.8)	57/1,207 (4.7)	193/5,698 (3.4)	495/11,742 (4.2)	<0.001

Data are shown as mean±standard deviation, median [interquartile range] or percentage, except for cotinine and daily alcohol consumption, which are expressed as mean±standard error of mean. Triglycerides, hsCRP, cotinine, and daily alcohol consumption were log-transformed for this analysis. P-values based on one-way ANOVA or Chi-square test. Statistically significant (P<0.05) by Scheffe’s multiple comparison: *never-smoking vs. new-smoking, ^†never-smoking vs. former-smoking, ^‡never-smoking vs. sustained-smoking, ^§new-smoking vs. former-smoking, ^[]new-smoking vs. sustained-smoking, ^¶former-smoking vs. sustained-smoking. BP, blood pressure; BUN, blood urea nitrogen; HDL-C, high-density lipoprotein cholesterol; hsCRP, high-sensitivity C-reactive protein; LDL-C, low-density lipoprotein cholesterol.

Differences in the characteristics between the 2 groups with and without incident HTN in the overall population and in men were significant for all variables (Tables 2,S3). In contrast, there were significant differences between the 2 groups in women in all variables except for BUN, urinary cotinine level, and frequency of alcohol drinking and vigorous exercise (Table S3).

Table 2. Baseline Characteristics of Individuals With and Without Incident Hypertension

	Hypertension (−) (n=70,486)	Hypertension (+) (n=4,257)	P value
Age, years	38.1±5.4	39.6±5.8	<0.001
Sex (male), n (%)	39,410 (55.9)	3,694 (86.8)	<0.001
Weight, kg	65.0±12.2	74.6±12.1	<0.001
Body mass index, kg/m2	23.0±3.1	25.3±3.3	<0.001
Waist circumference, cm	80.9±9.1	88.0±8.7	<0.001
Systolic BP, mmHg	107.1±11.3	122.1±9.1	<0.001
Diastolic BP, mmHg	68.5±8.5	79.9±6.8	<0.001
Total cholesterol, mmol/L	5.01±0.87	5.34±0.91	<0.001
Triglycerides, mmol/L	1.02 [0.72, 1.50]	1.47 [1.03, 2.11]	<0.001
LDL-C, mmol/L	3.10±0.81	3.43±0.83	<0.001
HDL-C, mmol/L	1.50±0.38	1.35±0.35	<0.001
Glucose, mmol/L	5.25±0.68	5.55±0.99	<0.001
BUN, mmol/L	4.37±1.07	4.56±1.08	<0.001
Creatinine, μmol/L	76.4±16.1	84.9±13.7	<0.001
Uric acid, μmol/L	309.6±85.1	364.7±85.7	<0.001
hsCRP, mg/L	0.4 [0.2, 0.8]	0.6 [0.3, 1.2]	<0.001
Cotinine, ng/mL	254.6±2.4	402.8±11.3	<0.001
Self-reported smoking status, n (%)			0.003
Never-/former-smoking	45,316/60,289 (75.2)	2,350/3,801 (61.8)
Current-smoking	14,973/60,289 (24.8)	1,451/3,801 (38.2)
Daily alcohol consumption, g/day	13.2±0.1	22.7±0.4	<0.001
Alcohol consumption (≥3 times/week), n (%)	8,235/65,038 (12.7)	928/4,064 (22.8)	<0.001
Vigorous exercise (≥5 times/week), n (%)	2,566/68,211 (3.8)	178/4,115 (4.3)	0.066
Metabolic syndrome, n (%)	3,904/53,984 (7.2)	870/3,114 (27.9)	<0.001
Diabetes mellitus, n (%)	1,505/70,474 (2.1)	248/4,256 (5.8)	<0.001

Data are shown as mean±standard deviation, median [interquartile range] or percentage, except for cotinine and daily alcohol consumption, which are expressed as mean±standard error of mean. Triglycerides, hsCRP, cotinine, and daily alcohol amount were log-transformed for this analysis. P-values based on Student’s t-test or the Chi-square test. Abbreviations as in Table 1.

In the multivariate Cox-hazard regression analyses of the overall population, cotinine-verified current smoking at baseline significantly decreased the risk for incident HTN in models 1 and 2 (relative risk (RR) [95% confidence interval (CI)], 0.93 [0.87, 1.00] in model 1 and 0.91 [0.84, 0.98] in model 2) compared with cotinine-verified never-smoking at baseline, but self-reported former-smoking and current smoking at baseline did not (Table 3). The results for the 4 smoking groups according to the change of cotinine-verified smoking status from baseline to follow-up showed that for new-smoking and sustained-smoking there were significantly decreased RRs for incident HTN in model 1 (0.74 [0.60, 0.93] and 0.88 [0.81, 0.95], respectively) compared with never-smoking. The RRs in model 2 remained significant for new-smoking and sustained-smoking (0.75 [0.58, 0.96] and 0.82 [0.74, 0.90], respectively). In the 4 smoking groups according to the change of self-reported smoking status from baseline to follow-up, only new-smoking was significantly associated with a lower risk for incident HTN in models 1 and 2 (0.83 [0.70, 0.99] in model 1 and 0.80 [0.66, 0.97] in model 2).

Table 3. Multivariate Cox-Hazard Regression Analyses for the Effects of Cotinine-Verified Smoking Groups on Incidence of Hypertension in the Overall Population

	Model 1	Model 2
At baseline
Cotinine-verified current-smoking†	0.93 [0.87, 1.00]*	0.91 [0.84, 0.98]*
At baseline
Former-smoking‡	1.01 [0.92, 1.11]	1.04 [0.94, 1.16]
Current-smoking‡	0.94 [0.87, 1.02]	0.92 [0.84, 1.01]
At baseline
Urinary cotinine level§	0.99 [0.98, 1.00]*	0.99 [0.98, 1.00]*
4 groups according to change of cotinine-verified smoking status between baseline and follow-up
New-smoking[]	0.74 [0.60, 0.93]**	0.75 [0.58, 0.96]*
Former-smoking[]	1.01 [0.91, 1.11]	1.08 [0.96, 1.21]
Sustained-smoking[]	0.88 [0.81, 0.95]**	0.82 [0.74, 0.90]**
4 groups according to change of self-reported smoking status between baseline and follow-up
New-smoking#	0.83 [0.70, 0.99]*	0.80 [0.66, 0.97]*
Former-smoking#	0.98 [0.84, 1.15]	1.01 [0.84, 1.21]
Sustained-smoking#	0.91 [0.77, 1.09]	0.97 [0.80, 1.18]

Values are expressed as relative risk [95% confidence interval]. *P<0.05, **P<0.01. ^†Reference group was “cotinine-verified never-smoking” at baseline; ^‡reference group was “self-reported never-smoking” at baseline; ^§log-transformed for the analysis; ^[]reference group was “cotinine-verified never-smoking”. New-smoking defined as cotinine level <50 ng/mL at baseline and ≥50 ng/mL at follow-up; Former-smoking as cotinine level ≥50 ng/mL at baseline and <50 ng/mL at follow-up; Sustained-smoking as cotinine level ≥50 ng/mL at baseline and at follow-up. ^#Reference group was “self-reported never-smoking”. New-smoking was defined as no smoking at baseline and currently smoking at follow-up; Former-smoking as smoking at baseline but no smoking at follow-up; Sustained-smoking as smoking at both baseline and follow-up. Model 1 adjusted for age, sex, body mass index, vigorous exercise (≥5 times/week), alcohol consumption (≥3 times/week), and presence/absence of diabetes. Model 2 adjusted for variables in model 1, waist circumference, BUN, creatinine, uric acid, total cholesterol, HDL-C, LDL-C, triglycerides, and hsCRP. Abbreviations as in Table 1.

There was no sex interaction for the association between smoking and incident HTN (P for interaction=0.60), but subgroup analyses according to sex showed different results. The results of multivariate Cox-hazard regression analyses for men were in perfect accord with those for the overall population (Table 4). Meanwhile, for women, none of the aforementioned smoking-related variables was associated with incident HTN (Table 4).

Table 4. Multivariate Cox-Hazard Regression Analyses for the Effects of Cotinine-Verified Smoking Groups on Incidence of Hypertension According to Sex

	Model 1		Model 2
	Men	Women	Men	Women
At baseline
Cotinine-verified current-smoking†	0.93 [0.87, 0.99]*	1.21 [0.76, 1.89]	0.90 [0.83, 0.98]*	1.15 [0.67, 1.98]
At baseline
Former-smoking‡	1.02 [0.93, 1.11]	1.37 [0.81, 2.32]	1.06 [0.96, 1.18]	1.18 [0.62, 2.25]
Current-smoking‡	0.94 [0.86, 1.02]	1.72 [0.96, 3.08]	0.92 [0.84, 1.02]	1.18 [0.55, 2.52]
At baseline
Urinary cotinine level§	0.99 [0.98, 1.00]*	1.03 [0.96, 1.10]	0.99 [0.97, 1.00]*	1.01 [0.93, 1.11]
4 groups according to change of cotinine-verified-smoking status between baseline and follow-up
New-smoking[]	0.73 [0.58, 0.92]**	1.07 [0.40, 2.88]	0.72 [0.56, 0.94]*	1.38 [0.50, 3.78]
Former-smoking[]	1.00 [0.91, 1.10]	1.34 [0.60, 3.00]	1.06 [0.95, 1.20]	1.79 [0.79, 4.03]
Sustained-smoking[]	0.87 [0.81, 0.94]**	1.15 [0.67, 1.97]	0.82 [0.74, 0.90]**	0.92 [0.45, 1.86]
4 groups according to change of self-reported-smoking status between baseline and follow-up
New-smoking#	0.82 [0.69, 0.98]*	1.44 [0.54, 3.88]	0.78 [0.64, 0.95]*	1.80 [0.57, 5.66]
Former-smoking#	0.99 [0.84, 1.16]	1.39 [0.76, 2.57]	1.03 [0.85, 1.24]	1.22 [0.57, 2.65]
Sustained-smoking#	0.93 [0.78, 1.11]	2.33 [0.33, 16.64]	0.99 [0.81, 1.21]	2.69 [0.38, 19.35]

Values are expressed as relative risk [95% confidence interval]. *P<0.05, **P<0.01. ^†Reference group was “cotinine-verified never-smoking” at baseline; ^‡reference group was “self-reported never-smoking” at baseline; ^§log-transformed for the analysis; ^[]reference group was “cotinine-verified never-smoking”. New-smoking defined as cotinine level <50 ng/mL at baseline and ≥50 ng/mL at follow-up; Former-smoking as cotinine level ≥50 ng/mL at baseline and <50 ng/mL at follow-up; Sustained-smoking as cotinine level ≥50 ng/mL at baseline and at follow-up.^# Reference group was “self-reported never-smoking”. New-smoking was defined as no smoking at baseline and currently smoking at follow-up; Former-smoking as smoking at baseline but no smoking at follow-up; Sustained-smoking as smoking at both baseline and follow-up. Model 1 adjusted for age, body mass index, vigorous exercise (≥5 times/week), alcohol consumption (≥3 times/week), and presence/absence of diabetes. Model 2 adjusted for variables in model 1, waist circumference, BUN, creatinine, uric acid, total cholesterol, HDL-C, LDL-C, triglycerides, and hsCRP. Abbreviations as in Table 1.

Discussion

This study showed that (1) cotinine-verified current smoking at baseline significantly decreased the risk for incident HTN; (2) cotinine-verified new-smoking and sustained-smoking according to the change of cotinine-verified smoking status from baseline to follow-up were significantly associated with decreased risk for incident HTN, especially in men; and (3) the effects of cotinine-verified smoking status on incident HTN were different from those of self-reported smoking status.

Several studies that have evaluated the acute effect of cigarette smoke on BP have shown consistent results that cigarette smoking temporarily increased BP.^17–19 The putative mechanism explaining these results is the effect of nicotine on sympathetic nervous system activation. However, prior experimental studies have also reported a tolerance to the effect of nicotine on BP and a decrease of BP through vasodilating mediators and relaxation of vascular smooth muscle.^20–24 Previous epidemiologic studies that have assessed the effect of chronic cigarette smoking on BP or HTN reflect the various biological effects of smoking on BP and have shown discordant results. In particular, most cross-sectional studies based on self-reported surveys have demonstrated that current smoking either was or was not inversely associated with BP.^2–4 The results of cotinine-verified cross-sectional studies have also been similar to those studies of self-reported smoking status.^5–7 Recently, we reported that cotinine-verified smoking was negatively associated with HTN independent of covariates, especially in men.⁸ However, the cross-sectional design of the previous studies prevented clarification of a causal relationship between chronic smoking and HTN.

To our knowledge, 5 longitudinal observational studies have assessed the association between chronic smoking and incident HTN or BP change,^9–13 and have shown inconsistent results. A study involving 4,549 American Indians from the Strong Heart Study reported that current smoking was not significantly associated with incident HTN at a 4-year follow-up visit but was inversely related to both SBP and DBP.⁹ However, that study did not perform comparison analyses according to sex. Another 3-year follow-up study of 4,489 Japanese men and women found a difference according to sex; current smoking in men significantly reduced the incidence of HTN (odds ratio (OR) [95% CI], 0.66 [0.48, 0.91]), but the same trend was not true for women (1.65 [0.62, 4.38]).¹⁰ Another 5-year follow-up study conducted in 2,107 Japanese men also reported that chronic smoking reduced changes in BP and the incidence of HTN.¹¹ The findings in the men in our study showed an inverse association between current smoking at baseline and incident HTN, which was consistent with the previous studies.^10,11 However, a 14-year longitudinal study involving 8,251 men revealed that smoking was independently associated with the onset of systolic HTN, but not diastolic HTN.¹² A previous study of 28,236 women in the Women’s Health Study revealed that smoking was modestly associated with an increased risk of incident HTN for 9.8 years,¹³ which was not in accordance with our result in women. Some differences in the results of these studies could be attributed to several common limitations. First, previous studies were based on participants’ responses to self-reported questionnaires. Therefore, misclassification errors of smoking status could have been present. Second, the studies only considered the participants’ baseline smoking status and did not reassess their smoking status at follow-up visit. In particular, the lack of information on any change in smoking status at follow-up in long-term follow-up studies over 5 years would be an important confounding factor in assessing the association between smoking status and incidence of HTN. Third, even after adjusting for various covariates, residual confounding variables might remain.

In contrast with prior studies, one distinct aspect of our study is that it is the first and largest longitudinal study to perform a simultaneous assessment of cotinine-verified smoking status in addition to a survey of self-reported smoking status and also to include a change of smoking status from baseline to follow-up. As a result, cotinine-verified current smoking at baseline was associated with low risk for incident HTN but self-reported current smoking at baseline was not. The reason of this difference between self-reported and cotinine-verified results cannot be explained definitively; however, one reason could be misclassification error of smoking status, as the prevalence of self-reported never-smoking among cotinine-verified current-smoking was 1.8%, whereas that of self-reported current-smoking in the cotinine-verified never-smoking group was 9.6%. In addition, 22.4% of cotinine-verified never-smoking group were self-reported former smokers (data not shown). Another remarkable finding in the present study was that either cotinine-verified new-smoking or cotinine-verified sustained-smoking according to the change of smoking status from baseline to follow-up showed a decreased risk for incident HTN compared with cotinine-verified never-smoking. However, the results according to self-reported smoking status revealed that only new-smoking was associated with low risk for incident HTN. In particular, the results for new-smoking could be attributed to the significant decrease in body weight observed at follow-up in the new-smokers compared with the other 3 groups (0.53 kg, −0.45 kg, 1.55 kg, and 0.44 kg in the never, new, former, and sustained smokers, respectively, P<0.001). Furthermore, the incidence of HTN in sustained smokers with weight gain (≥2 kg) during the follow-up period was higher than in those with weight reduction (≥2 kg) (9.7% vs. 8.5%, P=0.009). Mean SBP and DBP in the sustained smokers with weight gain also increased during the follow-up period (1.14 mmHg and 2.58 mmHg, respectively), whereas mean SBP and DBP in those with weight loss did not (−3.13 mmHg and −0.38 mmHg, respectively).

In the subgroup analyses, only the results in men were consistent with the above results in the overall population. This difference according to sex could be related to the fact that the rate of sustained-smoking in women was extremely low (0.6%); consequently, the CI of the risk for incident HTN in women was relatively wider than that in men. Furthermore, the lack of a difference in the risk for incident HTN between new-smoking and never-smoking in women might be attributed to most female new-smokers smoked less than 2 cigarettes per day, whereas almost half of the male new-smokers smoked more than 5 cigarettes per day (data not shown).

In the present study, unobserved smoking at follow-up as well as at baseline showed a decreased risk for incident HTN (Table S4). These results are consistent with those of our cross-sectional study.⁸ As described previously, unobserved smoking includes secondhand smoke (SHS), environmental tobacco smoke exposure, and under-reporting of exposure to cigarette smoke.⁸ Previous studies have shown discordant results for the association between SHS and HTN.^8,25–27 However, those studies included different populations and were mostly cross-sectional in design. Accordingly, further prospective studies with longer durations are needed to confirm the present findings.

Study Limitations

First, our population might not represent the entire Korean population because participants in our study were mostly middle-aged adults residing in urban areas. Second, in 2012 and 2014, BP measurements were obtained 3 times, and the average of the 3 values was used. However, in 2011, the BP was measured only once, which may have produced a misclassification bias. When we examined the results obtained from individuals enrolled in 2012 (n=41,689), we found a similarity to the main results in the overall population (data not shown). Third, the present study had a relatively short follow-up of 29 months (12–44 months). However, a follow-up study with longer duration would be possible because the KSHS is ongoing. Fourth, because the urinary cotinine cutoff level for cotinine-verified current smoking is not uniformly established, a level of difference per studies exists. However, recent results in our cross-sectional study showed that the degree of agreement between self-reported and cotinine-verified smoking status was almost perfect (k=0.870).⁸ Last, potential confounders remain, such as SHS exposure, genetic variations in nicotine metabolism, dietary factors, and other lifestyle changes including alcohol drinking and exercise during follow-up.

Nonetheless, the present study was worthwhile because it is the first and largest observational study to evaluate the relationship between cotinine-verified smoking status and incident HTN with a comprehensive set of data including smoking status at both baseline and follow-up.

Conclusions

In this longitudinal study, the effects of cotinine-verified smoking status on incident HTN were different from those of self-reported smoking status. Cotinine-verified current smoking at baseline was inversely associated with incident HTN compared with cotinine-verified never-smoking at baseline. Furthermore, either cotinine-verified new-smoking or sustained-smoking decreased the risk for incident HTN compared with cotinine-verified never-smoking; these results were distinct only in men. However, chronic cigarette smoking should not be overlooked because this behavior leads to various deleterious effects on the cardiovascular system. Further sex-specific studies with a longer follow-up period are required to determine the long-term effects of smoking on HTN.

Disclosures

The authors have no conflicts of interest regarding the content of this article.

Funding / Support / Grant

None.

Acknowledgments

None.

Supplementary Files

Supplementary File 1

Table S1. Baseline characteristics of the 4 cotinine-verified-smoking status groups in men

Table S2. Baseline characteristics of the 4 cotinine-verified-smoking status groups in women

Table S3. Baseline characteristics of individuals with and without incident hypertension in men and women

Table S4. Multivariate cox-hazard regression analyses for the effects of unobserved smoking on incidence of hypertension in the overall population

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-17-1188

References

• 1.   Piepoli MF, Hoes AW, Agewall S, Albus C, Brotons C, Catapano AL, et al. 2016 European Guidelines on cardiovascular disease prevention in clinical practice: The Sixth Joint Task Force of the European Society of Cardiology and Other Societies on Cardiovascular Disease Prevention in Clinical Practice (constituted by representatives of 10 societies and by invited experts) Developed with the special contribution of the European Association for Cardiovascular Prevention & Rehabilitation (EACPR). Eur Heart J 2016; 37: 2315–2381.
• 2.   Thut AB, Blizzard L, Schmidt MD, Luc PH, Granger RH, Dwyer T. The association between smoking and hypertension in a population-based sample of Vietnamese men. J Hypertens 2010; 28: 245–250.
• 3.   Imamura H, Uchida K, Kobata D. Relationship of cigarette smoking with blood pressure, serum lipids and lipoproteins in young Japanese women. Clin Exp Pharmacol Physiol 2000; 27: 364–369.
• 4.   Arkwright PD, Beilin LJ, Rouse I, Armstrong BK, Vandongen R. Effects of alcohol use and other aspects of lifestyle on blood pressure levels and prevalence of hypertension in a working population. Circulation 1982; 66: 60–66.
• 5.   Benowitz NL, Sharp DS. Inverse relation between serum cotinine concentration and blood pressure in cigarette smokers. Circulation 1989; 80: 1309–1312.
• 6.   Istvan JA, Lee WW, Buist AS, Connett JE. Relation of salivary cotinine to blood pressure in middle-aged cigarette smokers. Am Heart J 1999; 137: 928–931.
• 7.   Kang JH, Song YM. Association between cotinine-verified smoking status and metabolic syndrome: Analyses of Korean National Health and Nutrition Examination Surveys 2008–2010. Metab Syndr Rel Disord 2015; 13: 140–148.
• 8.   Kim BJ, Han JM, Kang JG, Kim BS, Kang JH. Association between cotinine-verified smoking status and hypertension in 167,868 Korean adults. Blood Press 2017; 26: 303–310.
• 9.   Wang W, Lee ET, Fabsitz RR, Devereux R, Best L, Welty TK, et al. A longitudinal study of hypertension risk factors and their relation to cardiovascular disease: The Strong Heart Study. Hypertension 2006; 47: 403–409.
• 10.   Nagahama K, Inoue T, Iseki K, Touma T, Kinjo K, Ohya Y, et al. Hyperuricemia as a predictor of hypertension in a screened cohort in Okinawa, Japan. Hypertens Res 2004; 27: 835–841.
• 11.   Okubo Y, SuwazonoY, Kobayashi E, Nogawa K. An association between smoking habits and blood pressure in normotensive Japanese men: A 5-year follow-up study. Drug Alcohol Depend 2004; 73: 167–174.
• 12.   Dochi M, Sakata K, Oism M, Tanaka K, Kobayashi E, Suwazono Y. Smoking as an independent risk factor for hypertension: A 14-year longitudinal study in male Japanese workers. Tohoku J Exp Med 2009; 217: 37–43.
• 13.   Bowman TS, Gaziano JM, Buring JE, Sesso HD. A prospective study of cigarette smoking and risk of incident hypertension in women. J Am Coll Cardiol 2007; 50: 2085–2092.
• 14.   Benowitz NL. Cotinine as a biomarker of environmental tobacco smoke exposure. Epidemiol Rev 1996; 18: 188–204.
• 15.   Coultas DB, Howard CA, Peake GT, Skipper BJ, Samet JM. Discrepancies between self-reported and validated cigarette smoking in a community survey of New Mexico Hispanics. Am Rev Respir Dis 1988; 137: 810–814.
• 16.   Kim BJ, Han JM, Kang JG, Rhee EJ, Kim BS, Kang JH. Relationship of cotinine-verified and self-reported smoking status with metabolic syndrome in 116,094 Korean adults. J Clin Lipidol 2017; 11: 638–645.
• 17.   Failla M, Grappiolo A, Carugo S, Calchera I, Giannattasio C, Mancia G. Effects of cigarette smoking on carotid and radial artery distensibility. J Hypertens 1997; 15: 1659–1664.
• 18.   Groppelli A, Giorgi DM, Omboni S, Parati G, Mancia G. Persistent blood pressure increase induced by heavy smoking. J Hypertens 1992; 10: 495–499.
• 19.   Rhee M, Na S, Kim Y, Lee M, Kim H. Acute effects of cigarette smoking on arterial stiffness and blood pressure in male smokers with hypertension. Am J Hypertens 2007; 20: 637–641.
• 20.   Hexum TD, Russett LR. Plasma encephalin-like peptide response to chronic nicotine infusion in guinea pig. Brain Res 1987; 406: 370–372.
• 21.   Holtz J, Sommer O, Bassenge E. Development of specific tolerance to nicotine infusions in dogs on chronic nicotine treatment. Klin Wochenschr 1984; 62(Suppl II): S51–S57.
• 22.   Larose P, Ong H, Januszewicz P, Cantin M, DuSouich P. Characterization of the effect of nicotine on vasopressin and atrial natriuretic factor in the rabbit. J Pharmacol Exp Ther 1988; 244: 1093–1097.
• 23.   Kim KS, Borzelleca JF, Bowman ER, McKennis H Jr. Effects of some nicotine metabolites and related compounds on isolated smooth muscle. J Pharmacol Exp Ther 1968; 161: 59–69.
• 24.   Dominiak P, Fuchs G, von Toth S, Grobecker H. Effects of nicotine and its major metabolites on blood pressure in anaesthetized rat. Klin Wochenschr 1985; 63: 90–92.
• 25.   Wu L, Yang S, He Y, Liu M, Wang Y, Wang J, et al. Association between passive smoking and hypertension in Chinese non-smoking elderly women. Hypertens Res 2017; 40: 399–404.
• 26.   Alshaarawy O, Xiao J, Shankar A. Association of serum cotinine levels and hypertension in never smokers. Hypertension 2013; 61: 304–308.
• 27.   Tabara Y, Takahashi Y, Setoh K, Muro S, Kawaguchi T, Terao C, et al. Increased aortic wave reflection and smaller pulse pressure amplification in smokers and passive smokers confirmed by urinary cotinine levels: The NAGAHAMA Study. Int J Cardiol 2013; 168: 2673–2677.