Association between Cigarette Smoking and Erectile Tumescence: The Mediating Role of Heart Rate Variability

Abstract

Cigarette smoking deleteriously affects erectile function, and conversely, quitting smoking improves erectile hemodynamics. Underlying mechanisms by which smoking (or reduction of smoking frequency) may affect erectile physiology are not well understood. This study examined the mediating role of heart rate variability (HRV; a marker of sympathovagal balance) among a sample of male chronic smokers from the United States. Sixty-two healthy men (M_age = 38.27 years; SD = 10.62) were assessed at baseline (while smoking regularly), at mid-treatment (while using a nicotine patch), and at follow-up, four weeks after patch discontinuation. Cigarette use, frequency-domain parameters of HRV (low frequency [LF], high frequency [HF], LF/HF ratio), and physiological sexual arousal responses (via penile plethysmography) were assessed at each visit. Results were consistent with mediation, in that greater reductions in cigarette use from baseline to follow-up were associated with longitudinal increases in LF, which in turn showed positive relations with across-time changes in erectile tumescence. Neither HF or LF/HF ratio mediated the relationship between smoking and erection. In conclusion, HRV mediated the inverse relationship between reductions in smoking and enhancements in erectile tumescence. Results underscore the possibility that cigarette use may deleteriously affect erectile function peripherally, in part, by disrupting cardiac autonomic function.

Cigarette smoking is an independent risk factor for erectile dysfunction (ED), and results of large cross-sectional^1–3 and longitudinal⁴ epidemiological studies show that smokers are approximately 1.5 to 2 times as likely as nonsmokers to experience ED. Acute affects of smoking/nicotine intake on sexual health in men have also been demonstrated in several laboratory-based studies. Specifically, smoking has been shown to acutely attenuate sexual arousal responses in habitual smokers,⁵ and ingestion of isolated nicotine (e.g., nicotine gum) has been shown to acutely reduce physiological sexual arousal in nonsmokers.^6,7 This relationship also holds in the reverse direction. That is, longitudinal studies have shown that quitting smoking promotes improvements in erectile function,⁸ even as soon as 24 hours post-cessation.^9,10

Mechanisms by which smoking (or reduction of smoking rate) may affect erectile physiology are not well understood. Smoking may exert its deleterious affects on sexual arousal in a number of ways: (i) centrally, by eliciting dose-dependent neurotransmitter and neuroendocrine effects;^11,12 (ii) peripherally, by acting as a sympathetic nervous system (SNS) agonist;¹³ (iii) or at the biochemical level by disrupting nitric oxide synthesis, responsible for endothelium-dependent smooth muscle relaxation.¹⁴ Complex interactions among these pathways may also exist.¹⁵

Despite emerging literature describing the beneficial effects of smoking cessation on sexual health, no data exist regarding potential mediators of this salutatory effect. One potential mechanism by which smoking cessation may affect erectile function is via changes in heart rate variability (HRV). HRV is a measure of the balance between parasympathetic and sympathetic maneuvers, and reflects the degree of variability from mean heart rate across time. While in a resting state, the heart beats naturally in irregular intervals, due to parasympathetic dominance, which decelerates heart rate and increases HRV. Conversely, SNS activation accelerates heart rate and decreases HRV. HRV is therefore an indicator of cardiac health; elevations in HRV represent healthy cardiac function, whereas reductions in HRV, leaves the heart vulnerable to arrhythmia and sudden death.¹⁶ It has been shown that smoking cessation improves a number of autonomic nervous system indices, including HRV.^17,18 Additionally, HRV has been implicated in erectile function. Specifically, individuals with erectile impairments have been shown to display abnormal cardiac autonomic regulation (i.e., sympathetic hyperactivity).¹⁹

As a first step to begin to identify potential mechanisms underlying tobacco’s effects on penile hemodynamics, this study examined the mediating role of HRV among a sample of male chronic smokers from the United States who were enrolled in a smoking cessation intervention. HRV was chosen as a potential mediator, given that this parameter is a marker of sympathovagal balance, which is a chief underlying mechanism of erectile response (i.e., parasympathetic dominance is pro-erectile, whereas sympathetic dominance is anti-erectile).²⁰ It was hypothesized that changes in HRV (from baseline to post-cessation follow-up) would mediate the relationship between changes in smoking frequency and changes in erectile hemodynamics. More specifically stated, it was hypothesized that decreases in cigarette use over time would be associated with increases in HRV, which in turn would be associated with temporal enhancements in erectile responses.

Subject and Methods

Participants

All data presented herein were taken from a nonrandomized intervention study examining the association between smoking cessation and sexual health indices in men.²¹ Participants were adult male smokers who were motivated and committed to quit. Men were recruited between 2008 and 2010 via community and online advertisements. Inclusionary criteria included: (i) being an adult between the ages of 23 and 60; (ii) smoking at least 15 cigarettes per day for a minimum of 5 consecutive years; (iii) no self-reported sexual dysfunction prior to smoking onset; (iv) being sexually active; and (v) reporting a heterosexual sexual identity. Exclusion criteria were as follows: (i) use of medications or report of medical conditions known to affect sexual or vascular functioning; (ii) medical conditions that could make nicotine administration unsafe (e.g., history of myocardial infarction, stroke, heart arrhythmias, angina, uncontrolled hypotension or hypertension); (iii) history of severe drug or alcohol abuse during the past 12 months (≥ 16 points on the Alcohol Use Disorders Identification Test (AUDIT)²² and ≥ 6 on the Drug Abuse Screening Test (DAST-10)²³); (iv) a body mass index (BMI) < 18.5 kg/m² or > 35 kg/m²; and (v) self-report of a sexually transmitted infection. A total of 65 men who were enrolled in the parent study; however, for the current report, 3 individuals reporting cardiovascular disease (CVD) were removed, resulting in a final sample size of 62 participants.

HRV

Cardiac autonomic function was assessed using a three-channel electrocardiograph (ECG). Signals were collected at a rate of 80 samples/second, low-pass filtered (to 0.5 Hz), digitized (40 Hz), and recorded using a Model MP100WS data acquisition unit (BIOPAC Systems, Inc., Santa Barbara, CA, USA) and the software package AcqKnowledge III, Version 3.73 (BIOPAC Systems, Inc.). Inter-beat intervals were derived, and artifacts were then identified and removed manually. Fast Fourier Transform, which derived the spectral distribution, was conducted automatically using Kubios HRV Analysis Software (Biosignal Analysis and Medical Imaging Group, University of Kuopio, Kuopio, Finland). This resulted in the calculation of low-frequency (LF) power (.04 – .15 Hz), high-frequency (HF) power (.15 – .40 Hz), and the ratio of these two indices (LF/HF), reflecting sympathovagal balance.²⁴

Erectile Tumescence

Erectile hemodynamics were assessed via penile plethysmography using a mercury-in-rubber strain gauge (Hokanson, Inc., Bellevue, WA, USA) to capture continuous changes in penile circumference. The signal was sampled at a rate of 80 samples/second, bandpass filtered (to 0.5 Hz), digitized (40 Hz), and recorded using a Model MP100WS data acquisition unit (BIOPAC Systems, Inc.) and the software package AcqKnowledge III, Version 3.73 (BIOPAC Systems, Inc.). The primary index of physiological sexual arousal was within-session percent change from the unaroused state (i.e., penile flaccidity) to maximum arousal (largest circumferential measurement obtained) in response to erotic stimuli.

Procedure

A detailed description of procedures are described in the parent study.²¹ In brief, eligible participants were scheduled for three separate laboratory visits: at baseline while smoking regularly, four weeks later while using a high-dose nicotine transdermal patch, and four weeks after completing the 8-week patch intervention (follow-up). Participants were allowed to enter the laboratory at their preferred nicotine levels; however, they were not allowed to smoke during the testing session. On average, participants reported consuming 7.19 cigarettes (SD = 4.79; range = 1 – 20) from awakening that day until presenting to the laboratory for the experimental session, and reported an average duration of 29.50 minutes (SD = 18.59; range = 5 – 90) since smoking their last cigarette. During each visit, self-report data were collected (socio-demographic information, medical history, substance use per the AUDIT and DAST, erectile function per the International Index of erectile Function (IIEF),²⁵ and smoking characteristics, including nicotine dependence assessed via the Fagerstrom Test for Nicotine Dependence (FTND)²⁶), as well as anthropometric (height and weight), cardiovascular (systolic and diastolic blood pressures), electrocardiographic (HRV), and physiological sexual arousal data. Electrocardiographic (ECG) and sexual arousal data were collected simultaneously while participants viewed stimuli designed to facilitate sexual arousal responses. Specifically, participants were randomized to view one of three 11-minute audiovisual sequences that consisted of an initial 3-minute non-sexual segment (one of three documentary film presentations), immediately followed by an 8-minute erotic film presentation. For all film sequences, the erotic film segments depicted a heterosexual couple engaging in consensual petting (1 minute), oral sex (3 minutes), and vaginal intercourse (4 minutes). Pilot testing established that the erotic films induced equivalent genital and self-reported increases in sexual arousal. The film segments (neutral, erotic) within each film sequence were always presented in the same order, and the three film sequences were counterbalanced across participants. For the purposes of the current study, only ECG data from the initial 3-minute baseline period were examined. At the completion of the laboratory visit, participants were given nicotine patches and they were asked to start the 8-week nicotine replacement therapy the following morning. The procedures of visits 2 and 3 were identical to the first session. The protocol was approved by the university Institutional Review Board.

Statistical Analysis

Pack years were calculated by multiplying the number of packs of cigarettes smoked per day by the number of years smoked. Missing HRV, smoking frequency, and sexual arousal values were imputed using full information maximum likelihood estimation,²⁷ which produces more accurate parameter estimates compared to per protocol analyses or last observation carried forward techniques.²⁸ Change scores were calculated separately for smoking frequency, HRV, and erectile tumescence by subtracting values obtained at follow-up from values obtained at baseline. Tests of simple indirect effects were employed with an SPSS (SPSS Inc., Chicago, IL, USA) application developed by Preacher and Hayes.²⁹ This macro provides a test of the indirect effect (i.e., changes in smoking frequency on changes in erectile tumescence though changes in HRV, with age, pack years, and drinking severity as covariates (see Figure 1), using a bootstrapping procedure (n = 10,000 resamples), which generated a sampling distribution for ab. The macro also incorporates traditional tests of mediation using the Baron and Kenny approach.³⁰ The simple indirect effects approach is a more sensitive and powerful test of mediation³¹ and is not dependent on the normality assumption underlying the Sobel³² test and the causal steps approach to mediation proposed by Baron and Kenny.³⁰ To assess indirect effects, 95% confidence intervals were generated for the parameter estimates. These parameter estimates were considered statistically significant if the confidence intervals did not include zero. Squared semipartial correlations (sr²) were reported for the both the direct and indirect effects, to illustrate the reduction in proportion of explained unique variance.

Figure 1.

Changes in heart rate variability (HRV) mediating associations between changes in smoking frequency and changes in erectile tumescence. Change scores were derived for each variable by subtracting values obtained at follow-up from values obtained at baseline. The dashed line represents the direct effect controlling for HRV.

Results

Participant Characteristics

Please see Table 1 for sociodemographic and smoking characteristics. The total sample (N = 62) ranged in age from 23 to 58 years with a mean age of 38.27 years (SD = 10.62). Eight participants reported taking medications at the time of enrollment (mood stabilizer: n = 3, antidepressant: n = 1; anticonvulsants, n = 1; benzodiazepine: n = 1, serotonin modulator: n = 1, H₂ blocker: n = 1; antihistamine, n = 2). No participants reported a history of myocardial infarction (MI) or current CVD, and none were taking cardiac/cardiovascular medications. Regarding participant flow, 17 participants dropped out after the initial visit, and an additional 12 after the second visit, resulting in a total of 33 participants who completed the follow-up session. Additional details pertaining to participant flow using CONSORT guidelines can be found in the parent study.⁸

Table 1.

Baseline Characteristics of the Participant Sample

Characteristic	M	SD	n	%
Age (years)	38.27	10.62
Education (years)a	14.78	2.21
Ethnicity
White			55	88.7
Black/African-American			2	3.2
Latino/a			2	3.2
Asian			2	3.2
Other			1	1.6
Marital status
Single			32	51.6
Married/cohabiting			21	33.9
Divorced			9	14.5
Substance use measures
Alcohol useb	5.08	3.41
Drug usec	.63	.85
Smoking measures
Pack years	20.54	15.27
Smoking frequency (cigarettes/day)	20.95	7.31
Nicotine dependence leveld,e	5.44	1.96
IIEF – EF scoref	26.61	5.09
EDg			18	29.0
BMI (kg/m2)	25.79	4.35
Cardiac function
Resting HR (bpm)	79.77	12.92
Systolic BP (mm Hg)	129.35	16.36
Diastolic BP (mm Hg)	84.89	13.27

Note. N = 62. BMI = body mass index; BP = blood pressure; bpm = beats per minute; ED = erectile dysfunction; EF = erectile function; IIEF = International Index of Erectile Function²⁵; mm Hg = millimeters mercury; HR = heart rate; M = mean; SD = standard deviation.

^aData missing for 2 participants.

^bAssessed with the Alcohol Use Disorders Identification Test.²² Possible score range from 0 to 40, with higher scores reflecting increasing levels of problematic drinking.

^cAssessed with the Drug Abuse Screening Test, 10-item.²³ Possible score range from 0 to 10, with higher scores indicating greater substance abuse severity.

^dData missing for 1 participant.

^eAs per the Fagerström Test of Nicotine Dependence.²⁶ Possible score range from 0 to 10, with higher scores indicating greater dependency to nicotine.

^fIIEF – EF score is based on a scale from 1 to 30, with 1 indicating the most severe erectile difficulties and 30 indicating the least severe (or no difficulty).

^gAccording to the IIEF erectile functioning threshold score of 25.

Associations between Smoking, HRV, and Erectile Tumescence Characteristics

Zero-order correlations among study variables are presented in Table 2. Age was positively associated with pack years, r = .78, p < .001, and negatively correlated with drinking severity, r = −.36, p < .01. Pack years was negatively correlated with erectile tumescence, r = −.25, p < .05; both cigarette use and alcohol use were positively associated with one another, r = .43, p < .001. Changes in cigarette consumption were negatively associated with changes in erectile tumescence, r = −.25, p < .05, changes in LF, r = −.35, p < .01, and changes in HF, r = −.28, p < .05. Furthermore, changes in LF were associated with changes in erectile tumescence, r = .34, p < .01. Finally, both LF and HF were highly correlated with one another, r = .82, p < .001.

Table 2.

Zero-order Correlations among Study Variables

Variable	1	2	3	4	5	6	7	8	M	SD
1 Age (years)	-	.78***	−.36**	.01	−.05	.17	−.22	−.07	38.27	10.62
2 Pack years		-	−.17	.15	−.04	.11	−.17	−.25*	20.54	15.27
3 Baseline drinking severitya			-	.43***	−.11	−.23	.26*	.17	5.08	3.41
4 Δ Cigarette use frequency (cigs/day)				-	−.35**	−.28*	.17	−.25*	−8.06	33.16
5 Δ LF power (msec2)					-	.82***	−.03	.34**	−76.00	1085.49
6 Δ HF power (msec2)						-	−.32**	.17	−127.31	565.58
7 Δ LF/HF ratio (msec2)							-	.08	−.05	2.48
8 Δ Erectile tumescence (mm)								-	−4.83	6.79

Note. N = 62. Change (Δ) represents a variable’s value at follow-up subtracted from the value obtained at baseline. HF = high-frequency power; LF = low-frequency power; LF/HF = low-frequency to high-frequency ratio. M = mean; SD = standard deviation.

^aAssessed with the Alcohol Use Disorders Identification Test.²²

^*p < .05,

^**p < .01,

^***p < .001

Mediation Analyses

After entering age, pack years, and baseline drinking severity as covariates, changes in cigarette frequency inversely predicted changes in LF (B = −12.37, SE = 4.60, p = .01), and changes in LF positively predicted changes in erectile tumesence (B = .001, SE = .001, P = .02) (See Table 3). Furthermore, the indirect effect of changes in cigarette consumption through changes in LF was significant, as indicated by 95% confidence intervals (CIs) not containing a value of zero (CI = −.067, −.001). When the effect of changes in cigarette use on erectile tumesence was controlled, the direct effect was reduced (although still statistically significant) (B = −.09, SE = .02, P = .001, sr² = .14 to B = −.07, SE = .02, P = .01, sr² = .05), which was suggestive of partial mediation.³⁰ Both HF power and the LF/HF ratio did not mediate the relationship between changes in cigarette use and erectile tumescence.

Table 3.

Mediation Analyses

Indirect effects
	M	SE	LL 95% CI	UL 95% CI
Bootstrap results	−.019	.015	−.067	−.001
Causal steps approach30
	B	SE	t	sr2
Step 1: Path c	−.09	.02	−3.67***	.14
Step 2: Path a	−12.37	4.60	−2.69**	.11
Step 3: Path b	.001	.001	2.32*	.12
Step 4: Path c′	−.07	.02	−2.81**	.05
Partial effects of covariate on erectile function
	B	SE	t
Age	−.37	.11	−3.39***
Pack years	.36	.07	4.85***
Baseline drinking severity	.53	.24	2.19*

Note. Bootstrap sample size = 10,000. B = unstandardized regression coefficient; CI = confidence interval; LL = lower limit; M = mean; SE = standard error; UL = upper limit. Model summary for dependent variable model: R² = .43, Adj R² = .38, F(5,56) = 8.56, P < .001.

^*p < .05,

^**p < .01,

^***p < .001

To strengthen the interpretation of the above mediation analyses, we conducted an additional analysis reversing the mediator and dependent variable.^33,34 Results were not consistent with mediation in this direction. Specifically, the indirect effect of changes in cigarette use through changes in erectile tumescence was nonsignificant, as indicated by 95% confidence intervals (CIs) containing a value of zero (CI = −14.04, .43).

Discussion

The current study examined the mediating role of HRV on the association between changes in cigarette use and changes in erectile tumescence among a sample of male chronic smokers without CVD from the United States who were enrolled in a smoking cessation intervention. Results were consistent with a causal mediation chain, indicating that changes in LF power (a frequency-domain parameter of HRV) mediated the inverse relationship between changes in cigarette frequency and changes in erectile tumescence. Alternatively stated, greater reductions in cigarette use (from baseline to follow-up) were associated with increased HRV over time, which in turn showed positive relations with longitudinal changes in physiological sexual arousal responses. The formal test of this indirect effect²⁹ and the traditional causal steps approach³⁰ were both significant.

Taken together, results suggest that reductions in cigarette smoking were associated with decreases in sympathetic tone and enhanced autonomic cardiac function. This finding is in line with results of prior studies showing increases in frequency-domain measures of HRV as a result of smoking cessation.^17,18,35,36 Additionally, results point toward the potential underlying role of cardiac autonomic function in the physiological sexual arousal process. That enhancement in HRV promoted improvements in erectile tumescence adds to the scarce, yet emerging, body of literature delineating the possible relationship between HRV and erectile function.¹⁹ Although this study highlights the potential mediating role of the peripheral nervous system on the relationship between smoking and erectile response, it by no means rules out the possibility that other mediating pathways are concomitantly implicated (e.g., neurotransmitter and neuroendocrine effects under central nervous system control, ^11,12 and/or activity at the biochemical level¹⁴).

Strengths of the current study include: (i) longitudinal (across a 12-week time frame) examination of the interplay between smoking, HRV, and erectile hemodynamics; (ii) examination of a variety of frequency-domain HRV parameters; (iii) exclusion of individuals with a history of MI, CVD, and cardiovascular medication use, all of which are conditions that have been associated with cardiac autonomic imbalance;^37,38 and (iv) use of rigorous statistical methods employing both traditional and state-of-the-art approaches to mediation. Despite this investigation’s novel findings, several study limitations warrant mention. First, smoking was assessed via self-report rather than via biochemical verification, and therefore, the validity of this dependent measure cannot be determined with absolute certainty. However, participants provided saliva samples at all experimental visits and they were spuriously told that these would be assayed for nicotine content. This “bogus pipeline” technique has been shown to produce accurate and reliable estimates of smoking.³⁹ The second limitation was with regard to the generalizability of the participant sample; African-American individuals were under-sampled compared to the general population of the United States ⁴⁰. Furthermore, although a notable proportion of men reported ED (29%), the mean age - and consequently the mean smoking duration - of the current sample was lower compared to other studies examining relationships between smoking, sexual function, and cardiac function. As such, it is unclear if results herein can be extrapolated to typical male patients presenting with ED who are characteristically older, have lower IIEF scores, and have longer smoking histories.

In conclusion, this is the first study to examine the mediating role of HRV on the relation between cigarette use and erectile tumescence among male chronic smokers without a history of MI or CVD. Results indicated that greater reductions in cigarette use from baseline to follow-up were associated with increases in HRV, which in turn showed positive relations with across-time changes in erectile tumescence. Given the salutatory effects of reducing smoking frequency on penile hemodynamics, results underscore the possibility that cigarette use may deleteriously affect erectile function peripherally, in part, by disrupting cardiac autonomic function. Further research examining other underlying biological mechanisms is necessary to establish a causal link between these variables.