Associations between Chronic Cigarette Smoking and Taste Function: Results from the 2013–2014 National Health and Nutrition Examination Survey

Abstract

We identified associations between cigarette-smoking and taste function in the U.S. NHANES 2013–2014. Adults ≥40 years (n=2849, nearly half former or current smokers) rated whole-mouth and tongue-tip bitter (1mM quinine) and salt (1M NaCl, 0.32 M NaCl) intensities and reported smoking history (pack years, PY), dependence (time to first cigarette, TTFC) and menthol/non-menthol use. Perceived intensity on the tongue-tip averaged just below moderate for quinine and moderate to strong for 1M NaCl. Current chronic smokers (≥20 PY) reported lower bitter and salty intensities on the tongue-tip (β: −2.0, 95% CI: −3.7 to −0.4 and β: −3.6, 95% CI: −6.9 to −0.3, respectively) than never smokers. Similarly, compared to never smokers, dependent current smokers (TTFC≤30 minutes) and dependent chronic smokers (≥20 PY, TTFC≤30 minutes) rated less bitter (β: −2.0, 95% CI: −4.0 to 0.1 and β: −2.9, 95% CI: −4.5 to −1.3, respectively) and salty (β: −5.3, 95% CI: −9.3 to −1.4 and β: −4.7, 95% CI: −8.6 to −0.7, respectively) intensities on the tongue-tip. Depressed tongue-tip intensity in dependent smokers (with/without chronicity) versus never smokers was significant in younger (40–65 years), but not older (>65 years) adults. Former smokers, non-chronic/less dependent smokers, and menthol smokers were more likely to report elevated whole-mouth quinine and 1 M NaCl intensities. Tongue-tip and whole-mouth taste intensity concordance varied between smokers and never smokers—current dependent smokers were more likely to rate tongue-tip quinine and NaCl lower than their respective whole-mouth tastants (OR: 1.8, 95% CI: 1.0 to 3.1 and OR: 1.8, 95% CI: 1.1 to 2.8, respectively). In summary, these U.S. nationally-representative data show that current smoking with chronicity and/or dependence associates with lower tongue-tip intensity for bitter and salty stimuli. Smokers with greater exposure to nicotine and/or dependence showed greater risk of taste alterations, with implications for diet- and smoking-related health outcomes.

1. Introduction

Cigarette smoking is a leading and preventable cause of morbidity and mortality worldwide. In 2019, 34.1 million adults (14.0%) in the United States (U.S.) reported being current cigarette smokers [1]. Compared to nonsmokers, smokers are at increased risk of several chronic conditions, including cancer, respiratory disease, cardiovascular disease, diabetes, and central obesity [2–4]. Beyond smoking, smokers report multiple unhealthy behaviors that are known to aggravate chronic disease risk, including poor diet [5, 6], elevated intakes of energy dense foods [7], and excessive alcohol consumption [8]. While several factors presumably influence dietary habits and alcohol intake in smokers, smoking-related alterations in taste could potentially be one of the contributing factors. Here, we explore the association between smoking and taste perception in a nationally representative US sample.

Taste perception involves binding of a tastant to a taste receptor in the oral cavity and initiation of the taste transduction pathway, which relays taste signals to the central nervous system through multiple nerves. Damage anywhere along this pathway can alter taste perception. Several cranial nerves interact to maintain taste function throughout the oral cavity [9], including the chorda tympani branch of cranial nerve VII (CN VII) innervating the tongue tip, the glossopharyngeal nerve (CN IX) innervating the back and posterior side of the tongue, and the superior laryngeal branch of the vagus nerve (CN X) innervating taste from the throat [10]. Because of redundancies in the peripheral innervation of the taste system, complete (ageusia) or severe reductions (hypogeusia) of taste are rare, and any regional taste loss within the whole mouth typically goes unnoticed during eating and ingestive behaviors. However, regional taste loss on the tongue-tip (as a result of viral damage or other environmental insults to the chorda tympani branch of CN VII) can disinhibit inputs from regions innervated by other taste nerves, causing intensified whole-mouth or tactile sensations, or phantom taste sensations (dysgeusia) [10]. Prior work suggests that diminished taste intensity on the tongue-tip – either assessed directly or relative to whole mouth sensation – can associate with important dietary and health outcomes. For example, depressed bitterness on the tongue tip has been associated with lower diet quality [11], while lower bitter taste sensations on the tongue tip relative to whole-mouth sensation has been associated with elevated central adiposity [12]. Given the varying implications of smoking insult on taste and health, studies that comprehensively evaluate regional and whole mouth taste intensity among individuals with varying degrees of smoking exposure are warranted.

Smoking exposure is generally described in pack years, a cumulative measure of number of packs of cigarettes smoked per day multiplied by the number of years the person has smoked. However, time to first cigarette (TTFC) is a simple indicator of nicotine dependence that appears to provide added information on smoking exposure and associated behaviors and outcomes [13]. Based on a narrative review [13], a shorter TTFC shows strong associations with smoking biomarkers independent of smoking frequency, and is also linked to less likelihood of smoking cessation and greater likelihood of relapse after quitting. Shorter TTFC may also be a proxy for how deeply cigarettes are inhaled and the intensity of cigarette exposure [14]. In a previous analysis of the 2011–2014 NHANES, we found that chronic smokers alone and with high nicotine dependence had greater odds of self-reported olfactory alteration compared to never smokers [15]. To our knowledge, no study to date has examined the additive effect of chronic smoking and nicotine dependence in relation to taste function.

Smoking may exert both direct and indirect negative impacts on taste function, and the severity of the impact likely depends on the level of smoking exposure. Greater exposure to nicotine and cigarette smoke have been shown to decrease the number of taste-associated anatomical structures (taste buds) in animals [16] and humans [17, 18]. Smoking is also related to multiple risk factors of taste dysfunction. For example, compared to non-smokers, both regular smokers and highly dependent smokers report greater rates of respiratory infections [19–21], which in turn are associated with increased risk of taste impairments [22]. Chronic smoking can also lower salivary flow rates and cause xerostomia [23, 24], with downstream effects on taste and oral sensation.

Despite the presence of plausible mechanisms by which smoking may alter taste sensations, the overall literature on the association between smoking and taste function is highly mixed, and more critically, assessment of taste function in multiple oral regions is largely lacking, with a few exceptions. Several studies suggest that chronic smokers have higher thresholds (i.e., lower sensitivity) for certain tastants relative to nonsmokers [18, 25–27]; similarly, regional testing using electrogustometry suggests smokers have higher thresholds [28–30]. Intervention studies assessing smoking and taste function indicate smoking cessation from 2 to 9 weeks associates with improvements in taste thresholds [30] and perceived bitterness [31], suggesting that not only does smoking damage taste, but also that former smokers may see improvements in depressed taste alterations over time. However, other studies find no significant differences in taste function between smokers and nonsmokers [32–34], while some others have instead reported elevated taste perception in smokers [35, 36]. For example, in a laboratory-based study with whole mouth and regional testing, younger smokers did not show impaired taste abilities, and instead showed heighten intensities from 1M NaCl [35]. However, this interpretation is somewhat complicated by the fact that concentrated salt elicits both taste and chemesthestic sensations. Still, a separate crowd sourced study with an age diverse sample also reported higher bitter taste intensities among smokers relative to non-smokers when using a bitterant delivered via filter paper discs [36].

The heterogeneity in findings between cigarette smoking and taste may be related to several factors, including a wide variety of methods used to both classify smoking exposure and measure taste function, as well as genetic variability in the samples with respect to TAS2R38, a bitter taste receptor gene. Individuals who exhibit the taster phenotype (ability to taste heightened bitterness from compounds such as propylthiouracil and phenylthiocarbamide) as a result of specific TAS2R38 polymorphisms, may be less likely to initiate and continue smoking [37, 38]. Findings in the literature as to whether taster status facilitates smoking chronicity and dependence are nonetheless mixed, and allude to potential interaction by race/ethnicity [39–41]. In a large U.S. population-based study, for example, taster status was an important risk factor for nicotine addiction in African American but not Caucasian adults [40]. In contrast, a study of nearly 10,000 adults from various U.S. cohort studies found that taster status was related to smoking status in European-Americans and not African Americans [41]. Taster status [35, 42] has also been associated with the use of menthol cigarettes which softens the harsh bitter/irritation sensations from nicotine with minty/cooling sensations [43], thereby facilitating smoking dependence [44]. Preferred use of mentholated cigarette products among certain racial/ethnic groups as well as among those with taster status thus may further confound the association between measures of cigarette smoking and taste function-[35, 42, 45]. The potential complexity of these multiple interactions highlight the need for studies that can comprehensively examine multiple aspects of smoking status and taste function using well-validated measures in diverse populations.

In 2013 and 2014, a brief and well-validated Taste and Smell Protocol (CSX) was standardized and included in the National Health and Nutrition Examination Survey (NHANES) in the United States [46]. The NHANES is an annual health survey and examination administered to a nationally representative sample of civilian, noninstitutionalized U.S. population. Critical to our question here, it includes both a thorough validated measure of tobacco use, and functional assessment of taste. As a part of the taste assessment, participants reported tongue-tip and whole-mouth intensities from quinine (bitter) and NaCl (salty) solutions using the generalized labeled magnitude scale (gLMS), the current gold standard in suprathreshold psychophysical testing.

Here we examine associations between measures of cigarette smoking and taste function in a nationally representative sample of U.S. adults 40 years and older who participated in the NHANES 2013–2014 Taste and Smell Protocol. The large, multifaceted NHANES dataset allowed us to characterize smoking status by chronicity and nicotine dependence while accounting for menthol vs non-menthol cigarette smoking, as well as statistically controlling for a variety of potential confounders, including sociodemographic factors, anthropometrics, and other risk factors of taste and olfactory dysfunction [35, 42, 46–49].

2. Material and methods

2.1. Study participants

The National Center for Health Statistics conducts NHANES each year to examine the health and nutritional status of a nationally representative sample of civilian, noninstitutionalized children and adults in the U.S. using a stratified multistage, probability cluster design. In the NHANES protocol, trained health technicians conduct in-person health interviews at home as well as physical examinations and laboratory analyses in the Mobile Examination Center (MEC), modular trailers that are deployed to local community sites. In the 2013–2014 NHANES cycle, a taste and smell protocol was administered to all eligible participants ≥40 years of age [46]. The NHANES Taste and Smell Protocol consisted of an interviewer administered chemosensory questionnaire (CSQ) in the home as well as a taste and smell examination (CSX) completed in the MEC.

The current study analyzed data from the NHANES 2013–2014, including 2849 participants who completed the NHANES taste exam (CSX) and answered all questions related to smoking exposures (SMQ). The National Center for Health Statistics Research Ethics Review Board approved the study procedures (continuation protocol #2011–17), and all NHANES participants provided informed consent [50].

2.2. Taste outcomes

In the NHANES taste examination (the CSX protocol), participants rated intensities of two tastants (1mM quinine, and 1 M NaCl) applied to the tongue-tip and three tastants (1 M NaCl, 0.32 M NaCl, and 1mM quinine) tasted with a whole mouth sip and spit procedure [51]. This specific procedure has moderate-to-good test-retest reliability over 2 weeks [52]. Participants used the gLMS to rate perceived taste intensities; the gLMS provides ratio level data similar to magnitude estimation, but is easier to use, and is thought to enhance the validity of intensity comparisons across individuals [53, 54]. The gLMS is scaled from 0 (“no sensation”) to 100 (“strongest sensation of any kind”) with empirically derived intermediate labels at 1.4 (“barely detectable”), 6 (“weak”), 17 (“moderate”), 34 (“strong”) and 53 (“very strong”). Prior to taste testing, participants were oriented to the gLMS and were asked to rate the brightness of three light stimuli (moderate, dim, and high) from an LED light box as a cross-modal standard. Participants who correctly ranked the lights (in terms of rated intensity) were eligible for the taste examination.

Taste intensity ratings.

For the tongue tip testing, trained health technicians applied a 1 mM quinine stimulus to the tongue tip followed by a 1 M NaCl stimulus using a saturated cotton swab. For the whole mouth testing, 10 mL aliquots of 1 mM quinine, 1 M NaCl, and 0.32 M NaCl solutions were sampled in the mouth for 3 seconds, and then expectorated. After tasting each stimulus, the participant rated its intensity using a gLMS scale and rinsed with water until any lingering taste was removed, before sampling the next tastant. All five rated intensities (2 tongue tip and 3 whole mouth) were analyzed separately as continuous variables.

Tongue Tip vs. Whole Mouth Taste Categories.

Participants were classified into groups based on their taste intensity ratings on the tongue tip relative to the whole mouth. Tertiles for each tastant (quinine, NaCl) on the tongue tip and whole mouth were grouped as low (1^st tertile), moderate (2^nd tertile), and high (3^rd tertile). Based on the concordance between these groups for tongue tip and whole mouth, participants were classified into three groups: tongue-tip rating concordant with the whole-mouth, tongue tip rating high relative to the whole mouth, or tongue tip rating low relative to the whole mouth. Separate tongue tip vs. whole mouth groupings were created for quinine and NaCl.

2.3. Smoking exposures

The NHANES Smoking Cigarette Use Questionnaire (SMQ) provides information related to current and former smoking, including age that participants started smoking, number of cigarettes smoked per day, as well as time to first cigarette (TTFC) after waking and use of mentholated vs. non-mentholated cigarettes among current smokers [55]. Supplementary Table S1 provides definitions for each smoking exposure. We classified participants as current smokers, former smokers, and never smokers (defined as less than 100 cigarettes smoked in their lifetime). Serum levels of cotinine were measured in the MEC as a biomarker of smoking exposure [56].

Smoking history.

Cigarette packs smoked per year was calculated by multiplying the packs smoked per day by number of years smoked. Pack years was categorized as 0 pack years (never smokers), <20 pack years, or ≥20 pack years (indicating chronic smoking, which is known to be associated with an increase in mortality [57]). Chronic smokers were stratified by smoking status (current or former smoker).

Estimated nicotine dependence.

Current smokers indicated how soon after waking they smoked their first cigarette. Previous research finds that TTFC is a good proxy indicator of nicotine dependence [58, 59], with smokers with more dependence (TTFC less than 30 minutes of waking) showing poor health outcomes [60]. Independent of the number of cigarettes smoked, shorter TTFC associates with increased nicotine and toxin exposure, which may reflect greater extraction of nicotine and smoke per cigarette via deeper inhalation [13, 14]. In our sample, dependence was categorized as never smoker, less dependent smoker (TTFC >30 minutes), or more dependent smoker (TTFC ≤30 minutes).

Smoking history and estimated nicotine dependence.

To capture heavy smoking exposure and to determine the additive effect, if any, of smoking history and nicotine dependence, smoking status was categorized further into: i) current smokers with ≥20 pack years and TTFC ≤30 minutes (chronic more dependent smokers); ii) current smokers with <20 pack years or TTFC >30 minutes (non-chronic less dependent smokers); iii) former smokers; and iv) never smokers.

2.5. Sociodemographic, anthropometric, and taste-related characteristics

Sociodemographic and anthropometric variables were selected from NHANES interview questionnaires (DEMO, ALQ, HSQ, CSQ) and physical examination measures (BMX). Self-reported sociodemographic variables included age (years), gender (male, female), race (Mexican American, other Hispanic, non-Hispanic white, non-Hispanic black, non-Hispanic Asian, or non-Hispanic other/including multiracial), marital status (married/living with partner, or not married), education (high school or less, or more than high school), and ratio of family income to poverty categorized in 4 bins (≤1.1, 1.1 – 2.0, 2.0 – 4.2, >4.2) [61]; for example, a ratio of 2 indicates a household income double the poverty salary level as adjusted family size. Trained health technicians measured height, weight, and waist circumference in the MEC. Body mass index (kg/m²) was calculated from measured height and weight and categorized as underweight (<18.5 kg/m²), normal weight (18.5 – 24.9 kg/m²), overweight (25.0 – 29.9 kg/m²), or obese (≥30.0 kg/m²). Waist circumference >102 cm in males and >88 cm in females was considered elevated [62].

History of alcohol consumption, which has been associated with self-reported taste alterations [48] and chronic smoking [8], was assessed using the NHANES Alcohol Use Questionnaire (ALQ) [63]. Participants who responded ‘yes’ to having ever drank 4 (for females) or 5 (for males) alcoholic beverages or more almost every day for a time in their life were considered to have a history of heavy drinking. Self-reported health on a categorical scale was dichotomized as excellent/very good/good versus fair/poor. Covariates related to smell and taste disorder were selected from the chemosensory questionnaire (CSQ) [51]. Experiencing persistent, unwanted tastes in the mouth, xerostomia (“dry mouth”), or sinonasal symptoms (persistent cold or flu >1 month and/or frequent nasal congestion from allergies) in the last 12 months were dichotomized as yes or no. Participants also reported history of tonsillectomy (yes/no) and history of 3 or more ear infections (yes/no). Consistent with Glennon and colleagues [15], we calculated a chemosensory risk score for each participant based on self-reported history of five conditions that have been associated with risk of taste alterations. The chemosensory risk score captured the number of ‘yes’ responses to experiencing: i) persistent cold or flu >1 month, ii) frequent nasal congestion from allergies, iii) xerostomia, iv) history of tonsillectomy, or v) history of serious head/face injury. The risk score ranged from 0 to 5, with higher scores indicating higher risk of taste alterations, and was based on prior literature reporting these conditions as risk factors of taste and olfactory dysfunction [12, 15, 48].

2.6. Statistical analysis

All statistical analyses were conducted using Statistical Analysis Software (SAS) version 9.4 (SAS Institute, Cary, NC). Given the complex sampling design of NHANES, descriptive statistics and regression analyses were weighted to account for unequal probabilities of selection due to oversampling of certain populations, non-coverage, non-response, and post-stratification to match the Census Bureau population counts [61]. A two-sided p-value of <0.05 without adjustment for multiple comparisons was considered statistically significant for all analyses.

Participant characteristics were described overall and by smoking history and estimated nicotine dependence. Differences in characteristics by smoking history and nicotine dependence were tested using chi-square tests (for categorical variables) or one-way analysis of variance (for continuous variables). Linear regression models were used to estimate associations between smoking exposures and continuous taste intensity rating outcomes, with results reported as estimates of adjusted mean differences between the smoking group and the referent group. Survey-weighted logistic regression models were used to evaluate associations between smoking exposures and tongue tip versus whole mouth taste categories. Adjusted odds ratios (aOR) and 95% CI were estimated relative to the concordant category. For both linear and logistic regression analyses, separate models were conducted for each smoking exposure variable, and never smokers were the reference group for each smoking classification. All regression analyses were adjusted for covariates identified a priori based on previous literature [46, 48], including age (years), sex, race/ethnicity, waist circumference (cm), education, marital status, income-to-poverty ratio, and chemosensory risk score.

With respect to linear regression analyses with continuous taste intensity rating outcomes, we conducted additional sensitivity and exploratory analyses to further evaluate confounding and effect modification by factors such as age, race/ethnicity, menthol exposure etc. In sensitivity analyses, we ran linear regression models that additional adjusted for history of alcohol consumption (heavy drinkers vs not) due to the observed associations with taste alterations [48] and chronic smoking [8]. We also included models with gLMS intensity rating for the middle light stimulus from the LED light box as a covariate to adjust for individual variability in scale usage [12, 64]. To examine effect modification by age, all linear regression analyses were stratified by age group (40–65 years and >65 years). We tested for an interaction between smoking exposures and age group by including an interaction term in the regression models (smoking history by age group, dependency by age group, and smoking status by age group).

In exploratory analyses, we stratified the primary linear regression models by three groups of race/ethnicity, namely non-Hispanic white; non-Hispanic black; and, to account for the smaller sample sizes of the remaining groups, other races, which included Mexican American, other Hispanic, non-Hispanic Asian, and other race – including multiracial. Additionally, we further stratified current smokers in each smoking exposure category by mentholation status, and re-ran the primary linear regression models, to examine whether the direction and magnitude of the associations vary by menthol exposure.

3. Results

The analytical sample includes 2849 participants aged 40 and older who completed measures for smoking exposures and taste outcomes. Table 1 includes characteristics of participants overall and by smoking history and estimated nicotine dependence. Nearly half of participants reported being former or current smokers with a mean smoking history was 23.68 ± 1.04 pack years. Across the smoking exposure groups, current chronic smokers (≥20 PY) who were more dependent (TTFC ≤30 min) had the highest serum cotinine levels. Current chronic smokers who were more dependent were less likely to be menthol users and reported more cigarettes smoked/day in the last 30 days compared to non-chronic (<20 PY) less-dependent (TTFC >30 min) smokers. Chronic more dependent smoking was associated with being non-Hispanic White, and not married, as well as having a lower education level, lower income-to-poverty ratio (i.e., an income closer to the poverty line), and a history of heavy drinking. Non-chronic less dependent smoking was only associated with being non-Hispanic Black. Former smoking was associated with being older and male.

Table 1.

Characteristics of participants’ ≥40 years overall and stratified by smoking history and estimated nicotine dependence in NHANES 2013 – 2014

	Overall	Current chronic more dependent smokera	Current non-chronic less-dependent smokerb	Former smoker	Never smoker	pc
N	2849	168	289	798	1594
Age (years) d	57.55 ± 0.28	53.81 ± 0.94	52.56 ± 0.87	61.51 ± 0.42	56.85 ± 0.44	<.0001
40–65 years	2003 (74.30)	142 (88.51)	254 (90.67)	451 (62.20)	1156 (75.95)	<.0001
>65 years	846 (25.70)	26 (11.49)	35 (9.33)	347 (37.80)	438 (24.05)
Gender
Male	1366 (48.23)	95 (52.95)	138 (42.76)	482 (57.15)	651 (44.10)	<.0001
Female	1483 (51.77)	73 (47.05)	151 (57.24)	316 (42.85)	943 (55.90)
Race
Mexican American	367 (6.53)	6 (1.22)	32 (6.06)	97 (5.68)	232 (7.66)	<.0001
Other Hispanic	228 (3.88)	10 (2.77)	14 (2.45)	68 (3.66)	136 (4.38)
Non-Hispanic White	1324 (73.22)	115 (84.19)	138 (71.12)	420 (78.93)	651 (69.40)
Non-Hispanic Black	568 (10.06)	28 (6.99)	89 (17.30)	132 (6.97)	319 (10.73)
Non-Hispanic Asian	309 (4.48)	2 (0.51)	11 (1.69)	69 (3.18)	227 (6.08)
Other Race - Including Multi-Racial	53 (1.83)	7 (4.33)	5 (1.39)	12 (1.58)	29 (1.74)
Marital status
Married/living as married	1831 (68.94)	83 (51.81)	149 (54.75)	510 (68.11)	1089 (73.82)	<.0001
Not married	1017 (31.06)	85 (48.19)	140 (45.25)	288 (31.89)	504 (26.18)
Education
High school or less	1226 (35.43)	107 (65.15)	165 (48.55)	349 (35.88)	605 (29.46)	<.0001
More than high school	1622 (64.57)	61 (34.85)	124 (51.45)	449 (64.12)	988 (70.54)
Income-to-poverty ratio d	3.16 ± 0.13	2.07 ± 0.24	2.34 ± 0.20	3.20 ± 0.10	3.43 ± 0.12	<.0001
≤1.1	571 (13.82)	67 (33.34)	98 (25.47)	155 (12.69)	251 (9.98)	<.0001
1.1–2.0	562 (17.20)	39 (23.47)	68 (22.83)	168 (16.97)	287 (15.56)
2.0–4.2	790 (31.76)	38 (33.07)	76 (34.89)	230 (33.50)	446 (30.14)
>4.2	710 (37.21)	14 (10.12)	29 (16.81)	203 (36.83)	464 (44.32)
Body mass index (kg/m 2 ) d	29.55 ± 0.20	28.57 ± 0.76	28.55 ± 0.55	30.27 ± 0.28	29.47 ± 0.29	<.0001
<18.5	59 (1.74)	5 (3.27)	16 (4.83)	12 (1.31)	26 (1.24)	0.031
18.5–24.9	708 (23.79)	55 (28.31)	79 (25.31)	163 (19.44)	411 (25.20)
25.0–29.9	953 (34.41)	56 (30.66)	92 (36.14)	285 (33.81)	520 (34.85)
≥30.0	1129 (40.07)	52 (37.76)	102 (33.71)	338 (45.45)	637 (38.71)
Waist circumference (cm) d	101.85 ± 0.36	102.49 ± 1.74	99.09 ± 1.55	104.83 ± 0.59	100.75 ± 0.69	<.0001
High riske	1693 (60.94)	93 (60.37)	160 (58.20)	500 (66.22)	940 (58.79)	0.078
Self-rated health
Excellent/very good/good	1994 (79.71)	90 (60.26)	176 (66.34)	547 (77.73)	1181 (85.19)	<.0001
Fair/poor	693 (20.29)	66 (39.74)	98 (33.66)	217 (22.27)	312 (14.81)
Taste and smell risk factors
Persistent taste in mouth in last 12 months	175 (5.41)	15 (8.50)	23 (8.19)	41 (5.57)	96 (4.48)	0.201
Sinonasal symptoms in the last 12 monthsf	934 (33.46)	60 (36.21)	109 (41.22)	270 (34.64)	495 (31.20)	0.017
Xerostomia in last 12 months	424 (13.48)	33 (17.78)	48 (15.65)	125 (15.80)	218 (11.42)	0.023
History of serious head/face injury	650 (27.02)	64 (37.61)	94 (39.88)	207 (29.13)	285 (22.49)	<.0001
History of tonsillectomy	780 (32.74)	47 (30.27)	71 (27.60)	261 (38.02)	401 (31.24)	0.003
History of frequent ear infections	557 (22.80)	44 (32.13)	64 (24.70)	143 (20.52)	306 (22.54)	0.091
Chemosensory risk score g	1.10 ± 0.03	1.29 ± 0.06	1.27 ± 0.12	1.22 ± 0.02	0.99 ± 0.04	<.0001
Smoking/alcohol exposures
Smoking history (pack years)d,h	23.68 ± 1.04	39.06 ± 1.62	15.32 ± 0.99	22.92 ± 1.31	-	-
Cigarettes smoked/day in last 30 days	14.50 ± 0.71	21.43 ± 0.88	9.85 ± 0.43	-	-	-
Menthol smokersi	127 (24.50)	30 (15.46)	97 (29.82)	-	-	-
Days smoked in last 30 days	29.39 ± 0.14	29.92 ± 0.06	29.03 ± 0.27	-	-	-
Serum cotinine (ng/mL)	58.84 ± 7.40	318.43 ± 9.43	239.77 ± 13.01	24.34 ± 6.12	15.27 ± 3.69	<.0001
Heavy drinkersj	427 (16.22)	57 (40.06)	82 (29.93)	185 (24.12)	103 (7.17)	<.0001
NHANES taste measure ratings d,k
1 mM quinine tongue tip	14.03 ± 0.42	11.21 ± 1.01	14.75 ± 0.83	13.72 ± 0.70	14.40 ± 0.50	<.0001
1M NaCl tongue tip	26.10 ± 0.34	22.77 ± 1.44	27.53 ± 1.15	26.56 ± 0.96	26.00 ± 0.38	<.0001
1 mM quinine whole mouth	51.97 ± 0.54	53.70 ± 3.03	58.94 ± 2.27	52.14 ± 1.09	50.47 ± 0.65	<.0001
1 M NaCl whole mouth	51.61 ± 0.70	48.92 ± 2.90	57.21 ± 1.41	51.35 ± 1.29	51.08 ± 0.85	0.006
0.32 M NaCl whole mouth	32.74 ± 0.46	30.85 ± 1.50	37.83 ± 1.56	32.15 ± 0.98	32.37 ± 0.54	0.270

^aPack years ≥20 years and time to first cigarette ≤30 min

^bPack years <20 years and time to first cigarette >30 min

^cSignificance determined by survey-weighted one-way analysis of variance and χ² test

^dData is presented as weighted mean ± standard error

^eDefined as waist circumference >102 cm in males and >88cm in females

^fExperiencing in the last 12 months: persistent cold/flu >1 month or frequent nasal congestion from allergies

^gCalculated by ‘yes’ responses to: persistent cold/flu >1 month, frequent nasal congestion from allergies, xerostomia in last 12 months, history of tonsillectomy, or history of serious head/face injury

^hFor current and former smokers: packs smoked per day multiplied by years smoked

ⁱFor current smokers with a cigarette menthol indication (n=402)

^jDefined as responding ‘yes’ to having ever drank 4/5 alcoholic beverages every day compared to participants who responded ‘no’ to: having ever drank 4/5 alcoholic beverages/day, having at least 12 alcoholic beverages over last 12 months, or having at least 12 alcoholic beverages/lifetime

^kGeneral Labeled Magnitude Scale intensity values range from 0: “nothing” to 100: “strongest sensation of any kind”; intermediate labels include 1.4: “barely detectable”, 6: “weak”, 17: “moderate”, 34: “strong”, and 53: “very strong”

With respect to health measures, current smokers were more likely to rate their health as fair/poor (versus excellent/very good/good) and to report several taste and smell risk factors, including sinonasal symptoms and xerostomia in the past 12 months, history of serious head/facial injury, and frequent ear infections. Former smokers had a greater BMI and central adiposity (waist circumference in cm) and were more likely to report history of tonsillectomy.

With respect to taste outcomes, the mean gLMS ratings were just below “moderate” for the 1 mM quinine tongue tip test (14.03 ± 0.42), between “moderate” and “strong” for the 1 M NaCl tongue tip test (26.10 ± 0.34), and near “very strong” for both 1 mM quinine whole mouth test (51.97 ± 0.54) and 1 M NaCl whole mouth test (51.61 ± 0.70). Chronic more dependent smoking was associated with lower intensity ratings for quinine and NaCl tongue tip. Non-chronic less dependent smoking was associated with greater whole mouth intensity ratings for both quinine and 1 M sodium chloride.

Table 2 shows associations between smoking and taste intensity ratings. In general, current chronic smoking with more nicotine dependence was associated with lower taste intensities (primarily tongue tip) whereas less dependent, non-chronic smoking was associated with greater whole mouth intensities relative to never smokers. In adjusted models, current chronic smokers reported significantly lower quinine and NaCl taste intensities on the tongue tip compared to never smokers; similarly, current more nicotine dependent smokers reported significantly depressed NaCl intensity and marginally depressed quinine intensity on the tongue tip relative to never smokers. Further, current chronic, more dependent smokers (≥20 pack years and TTFC ≤30 minutes) showed a reduction of quinine and NaCl tongue tip intensities compared to never smokers. In summary, across multiple smoking measures, smokers generally reported depressed taste function on the tongue tip.

Table 2.

Adjusted multivariable linear regression models of associations between smoking exposures and taste intensity ratings in NHANES 2013 – 2014

		1 mM Quinine tongue tip	1 M NaCl tongue tip	1 mM Quinine whole mouth	1 M NaCl whole mouth	0.32 M NaCl whole mouth
	N	βa (95% CI)	βa (95% CI)	βa (95% CI)	βa (95% CI)	βa (95% CI)
Smoking history b
Current chronic smoker	218	−2.05 (−3.66, −0.43)	−3.58 (−6.90, −0.27)	4.36 (−1.87, 10.60)	−0.37 (−7.14, 6.39)	0.22 (−2.04, 2.48)
Former chronic smoker	269	0.29 (−1.56, 2.13)	0.69 (−2.71, 4.08)	1.92 (−1.35, 5.19)	−1.37 (−5.30, 2.55)	−1.06 (−4.74, 2.62)
Non-chronic smoker	768	0.10 (−1.35, 1.55)	0.01 (−1.96, 1.97)	4.03 (2.00, 6.05)	1.92 (−0.97, 4.81)	0.28 (−2.14, 2.70)
Never smoker	1594	Reference	Reference	Reference	Reference	Reference
Nicotine dependence c
More dependent smoker	273	−1.96 (−3.98, 0.06)	−5.35 (−9.34, −1.36)	3.18 (−1.89, 8.25)	−1.59 (−5.61, 2.43)	−2.97 (−6.22, 0.28)
Less-dependent smoker	184	0.35 (−3.39, 4.08)	1.15 (−3.33, 5.63)	8.03 (2.35, 13.71)	5.13 (0.28, 9.98)	5.15 (−0.91, 11.22)
Never smoker	1594	Reference	Reference	Reference	Reference	Reference
Smoking status d
Current chronic more dependent smoker	168	−2.90 (−4.47, −1.34)	−4.66 (−8.58, −0.74)	1.64 (−5.51, 8.80)	−3.74 (−10.22, 2.73)	−3.78 (−7.08, −0.48)
Current non-chronic less-dependent smoker	289	0.69 (−2.27, 3.65)	−0.58 (−4.24, 3.09)	6.81 (1.70, 11.92)	4.09 (0.29, 7.89)	3.61 (−0.72, 7.94)
Former smoker	798	0.00 (−1.40, 1.41)	0.40 (−1.67, 2.47)	3.03 (0.67, 5.38)	0.71 (−2.76, 4.18)	−0.44 (−2.61, 1.74)
Never smoker	1594	Reference	Reference	Reference	Reference	Reference

^aAdjusted for age (years), gender, race/ethnicity, waist circumference (cm), education, marital status, income-to-poverty ratio, and chemosensory risk score

^bChronic smoker: ≥20 pack years (PY); Non-chronic smoker: <20 PY (non-chronic includes both current and former smokers)

^cMore dependent smoker: time to first cigarette (TTFC) ≤30 min; Less-dependent smoker: TTFC >30 min (nicotine dependence only includes current smokers)

^dCurrent chronic more dependent smoker: ≥20 PY and TTFC ≤30 min, Current non-chronic less-dependent smoker: <20 PY or TTFC >30 min

Bold values are statistically (p<0.05) significant

Interestingly, non-chronic and/or less dependent smoking and former smoking were significantly associated with elevated intensity for whole mouth quinine. Likewise, less dependent smoking was also significantly associated with elevated intensity for whole mouth 1 M NaCl. Current chronic more dependent smoking, on the other hand, was associated with a reduction of 0.32 M NaCl whole mouth intensity compared to never smokers, but, when regression analyses were adjusted for history of alcohol consumption, this association was no longer significant. In contrast, all of the other associations above remained significant when controlling for alcohol consumption history. Further, all of these associations remained significant when models were additionally adjusted for moderate light gLMS rating, suggesting effects were not an artifact of differential scale usage across adults.

When current smokers were stratified by menthol cigarette use, the associations between smoking exposures and taste outcomes generally remained consistent in the non-menthol smokers (Supplementary Table S2). Associations between menthol smoking and tongue tip taste outcomes were not statistically significant presumably due to small sample sizes but generally followed the same direction as non-menthol users. For whole mouth taste perception, menthol smokers reported elevated whole mouth quinine intensity compared with never smokers, irrespective of chronicity and dependence status. Current chronic menthol smokers also reported greater whole-mouth intensities from 1 M NaCl compared to never smokers. When stratifying the analyses race/ethnicity, the direction of the associations between smoking exposures and taste intensities on the tongue tip were the same across racial/ethnic groups, yet statistically insignificant among non-Hispanic blacks (data not shown). However, associations between smoking exposures and whole mouth taste intensities only remained significant in non-Hispanic whites, and the directions of the associations were consistent with our primary findings described above.

Significant interactions between smoking exposures and age group were observed for NaCl and quinine tongue tip intensities (all p for interactions<0.05). Table 3 presents associations between smoking and taste intensity stratified by age group (40 to 65 years old versus >65 years old). Stratified analyses show associations between smoking and taste intensity were significant among younger (40 to 65 years) but not older (>65 years) adults. The intensity of quinine on the tongue tip was significantly depressed in current chronic more dependent younger smokers compared with younger never smokers. Similarly, the intensity of NaCl on the tongue tip was significantly depressed in current more dependent younger smokers relative to younger never smokers; the association was marginally significant in current chronic, more dependent younger smokers. Compared to younger never smokers, quinine whole mouth taste intensity was elevated among younger adults who were non-chronic and/or less dependent smokers, and 1 M NaCl intensity was elevated among current non-chronic less dependent younger smokers. There were no significant taste differences between smokers and never smokers for the weaker concentration of NaCl (0.32 M) in either age group.

Table 3.

Adjusted multivariable linear regression models of associations between smoking exposures and taste intensity ratings among participants 40 to 65 years or >65 years in NHANES 2013 – 2014

			Smoking historya		Nicotine dependenceb		Smoking statusc
	Never smoker	Current chronic smoker	Former chronic smoker	Non-chronic smoker	More dependent smoker	Less-dependent smoker	Current chronic more dependent smoker	Current non-chronic less-dependent smoker	Former smoker
	βd (95% CI)	βd (95% CI)	βd (95% CI)	βd (95% CI)	βd (95% CI)	βd (95% CI)	βd (95% CI)	βd (95% CI)	βd (95% CI)
1 mM Quinine tongue tip
40–65 years	Ref	−1.55 (−3.80, 0.70)	0.03 (−2.86, 2.92)	0.48 (−1.56, 2.52)	−1.40 (−3.62, 0.82)	1.32 (−2.88, 5.52)	−2.40 (−4.60, −0.20)	1.39 (−1.88, 4.66)	0.04 (−1.62, 1.69)
>65 years	Ref	−1.82 (−6.01, 2.37)	0.31 (−4.01, 4.63)	−0.49 (−3.85, 2.87)	−2.64 (−6.78, 1.51)	−3.87 (−9.34, 1.60)	−2.00 (−6.32, 2.31)	−3.11 (−8.12, 1.89)	0.04 (−3.45, 3.53)
1 M NaCl tongue tip
40–65 years	Ref	−3.24 (−6.82, 0.35)	−0.21 (−4.72, 4.30)	0.12 (−2.52, 2.76)	−4.91 (−9.06, −0.75)	1.91 (−3.05, 6.88)	−4.13 (−8.43, 0.17)	−0.36 (−4.19, 3.47)	0.23 (−2.78, 3.24)
>65 years	Ref	−1.68 (−9.04, 5.69)	2.21 (−1.22, 5.64)	1.12 (−1.56, 3.79)	−3.89 (−10.45, 2.67)	−0.47 (−11.65, 10.71)	−2.90 (−10.30, 4.51)	0.63 (−10.77, 12.04)	1.61 (−1.17, 4.39)
1 mM Quinine whole mouth
40–65 years	Ref	5.63 (−1.15, 12.41)	3.88 (−1.24, 9.00)	4.08 (1.83, 6.33)	4.34 (−1.20, 9.87)	9.31 (3.00, 15.62)	2.60 (−5.31, 10.51)	7.20 (1.98, 12.42)	3.54 (0.20, 6.88)
>65 years	Ref	−1.33 (−8.21, 5.55)	−0.78 (−5.47, 3.92)	3.25 (−1.15, 7.65)	−2.64 (−12.76, 7.49)	1.18 (−9.67, 12.04)	−2.65 (−13.37, 8.08)	2.18 (−8.62, 12.97)	1.49 (−2.53, 5.50)
1 M NaCl whole mouth
40–65 years	Ref	0.38 (−7.39, 8.16)	−0.50 (−7.57, 6.56)	1.62 (−1.84, 5.08)	−1.04 (−5.87, 3.79)	5.95 (−0.04, 11.94)	−3.47 (−10.58, 3.65)	4.46 (0.35, 8.57)	0.72 (−4.25, 5.69)
>65 years	Ref	−2.93 (−10.89, 5.02)	−2.36 (−7.79, 3.08)	2.76 (−1.85, 7.38)	−1.29 (−11.48, 8.89)	0.12 (−10.20, 10.43)	−2.58 (−11.48, 6.31)	0.34 (−9.64, 10.32)	0.48 (−3.46, 4.42)
0.32 M NaCl whole mouth
40–65 years	Ref	1.29 (−1.85, 4.43)	−1.00 (−6.85, 4.85)	0.32 (−3.26, 3.89)	−2.10 (−6.32, 2.12)	6.73 (−0.18, 13.64)	−3.06 (−7.24, 1.11)	4.12 (−0.52, 8.76)	−0.51 (−4.29, 3.26)
>65 years	Ref	−3.21 (−8.40, 1.97)	−1.63 (−4.24, 0.97)	0.20 (−4.31, 4.72)	−5.41 (−11.97, 1.15)	−2.33 (−8.65, 3.98)	−5.37 (−12.85, 2.12)	−0.68 (−7.20, 5.83)	−0.51 (−3.61, 2.59)

^aChronic smoker: ≥20 pack years (PY); Non-chronic smoker: <20 PY

^bMore dependent smoker: time to first cigarette (TTFC) ≤30 min; Less-dependent smoker: TTFC >30 min

^cCurrent chronic more dependent smoker: ≥20 PY and TTFC ≤30 min, Current non-chronic less-dependent smoker: <20 PY or TTFC >30 min

^dAdjusted for age (years), gender, race/ethnicity, waist circumference, education, marital status, income-to-poverty ratio, and chemosensory risk score

Bold values are statistically (p<0.05) significant

Associations between smoking and taste function as a measure of concordance of tongue tip versus whole mouth taste intensities are displayed in Table 4. In general, smokers had greater odds of reporting lower tongue tip intensities relative to whole mouth intensities for quinine and NaCl compared to never smokers, consistent with results described above. These associations were significant in more dependent smokers, former chronic smokers, non-chronic smokers, and former smokers for quinine, and in current chronic and/or more dependent smokers for NaCl. Former smokers also had greater odds of reporting whole mouth intensity lower than tongue tip intensity for quinine compared to never smokers.

Table 4.

Adjusted multivariable logistic regression models of associations between smoking exposures and taste function by concordance of taste intensity ratings

	Quinine tongue tip rating relative to quinine whole mouth ratinga		NaCl tongue tip rating relative to 1 M NaCl whole mouth ratinga
	Quinine whole mouth low relative to quinine tongue tip	Quinine tongue tip low relative to quinine whole mouth	NaCl whole mouth low relative to NaCl tongue tip	NaCl tongue tip low relative to NaCl whole mouth
	aORb (95% CI)	aORb (95% CI)	aORb (95% CI)	aORb (95% CI)
N	781	837	738	694
Smoking history c
Current chronic smoker	0.98 (0.51, 1.91)	1.79 (0.69, 4.67)	0.94 (0.52, 1.71)	1.75 (1.10, 2.77)
Former chronic smoker	1.29 (0.74, 2.24)	1.42 (1.01, 2.01)	1.03 (0.65, 1.64)	0.91 (0.59, 1.41)
Non-chronic smoker	1.15 (0.82, 1.61)	1.48 (1.20, 1.83)	0.95 (0.70, 1.29)	1.25 (0.95, 1.66)
Never smoker	Reference	Reference	Reference	Reference
Nicotine dependence d
More dependent smoker	0.94 (0.62, 1.44)	1.78 (1.03, 3.05)	0.97 (0.65, 1.44)	1.77 (1.11, 2.82)
Less-dependent smoker	0.62 (0.36, 1.06)	1.26 (0.72, 2.20)	0.82 (0.53, 1.27)	1.25 (0.80, 1.94)
Never smoker	Reference	Reference	Reference	Reference
Smoking status e
Current chronic more dependent smoker	1.07 (0.60, 1.91)	2.04 (0.87, 4.77)	0.94 (0.53, 1.67)	1.61 (1.04, 2.50)
Current non-chronic less-dependent smoker	0.66 (0.42, 1.03)	1.29 (0.87, 1.93)	0.87 (0.61, 1.25)	1.51 (0.93, 2.43)
Former smoker	1.36 (1.01, 1.83)	1.61 (1.25, 2.08)	1.05 (0.85, 1.30)	1.05 (0.83, 1.32)
Never smoker	Reference	Reference	Reference	Reference

^aResults are relative to concordant category (quinine tongue tip rating concordant with quinine whole mouth rating, n=1231; NaCl tongue tip rating concordant with NaCl whole mouth rating, n=1417; quinine tongue tip concordant with NaCl tongue tip, n=1504)

^bAdjusted for age (years), waist circumference, gender, race/ethnicity, marital status, education, income poverty ratio, and chemosensory risk score

^cChronic smoker: ≥20 pack years (PY); Non-chronic smoker: <20 PY

^dMore dependent smoker: time to first cigarette (TTFC) ≤30 min; Less-dependent smoker: TTFC >30 min

^eCurrent chronic more dependent smoker: ≥20 PY and TTFC ≤30 min, Current non-chronic less-dependent smoker: <20 PY or TTFC >30 min

Bold values indicate statistical (p<0.05) significance

4. Discussion

Using nationally representative data from the U.S., we examined associations between smoking and taste function in adults aged 40 years and over while controlling for multiple demographic and health-related characteristics. Compared to never smokers, current smokers who were characterized as more dependent (i.e., needing to smoking within 30 minutes of waking) and/or chronic (≥20 pack years; equivalent of smoking a pack daily for more than 20 years) reported lower bitter (quinine) and salty (NaCl) sensations on the tongue tip. These smokers were also more likely to report lower intensities on the tongue-tip relative to whole mouth for both quinine and NaCl. Former smoking as well as current smoking with less dependency and/or chronicity was generally associated with elevated whole mouth taste intensities, except for weak salt (0.32 M NaCl), which was depressed among current chronic, highly dependent smokers. These associations between smoking and taste were apparent among younger adults (ages 40 to 65 years old) but not among older adults (>65 years old).

Our findings are generally consistent with previous research that supports the association between current smoking and depressed taste function on the anterior tongue. Studies that have assessed taste function by measuring taste thresholds via electrogustometry at multiple regions of the tongue [28–30] and with tastant solutions as stimuli [26] or using taste discrimination methods at suprathreshold concentrations [65] found that smokers had lower taste sensitivity or higher risk of taste impairment than nonsmokers. The present study adds to the literature by using the standardized NHANES taste examination protocol administered to a large sample of participants, which offers several unique advantages. First, it assesses perceived taste intensity regionally and in the whole mouth via standardized procedures [51]. Concentrated quinine and NaCl are drawn across the tongue-tip to stimulate taste from the chorda tympani branch of cranial nerve VII; these tastants, along with a lower concentration of NaCl, are also sampled orally in order to stimulate the whole mouth. Second, the scaling on gLMS is generalized to apply to sensations of any kind, allowing better comparisons of intensity ratings across tastants and participants [53]. Before beginning the taste examination, the technician reviewed the gLMS scale with participants using a standardized script [51] and assured that they can correctly order light standards from an LED lightbox from lowest to highest using the gLMS. Lastly, it allowed us to adjust for individual variability in gLMS rating using a cross-modal standard [52, 64]. The inclusion of a cross modal standard (light) outside the domain of interest (here, taste) ensures that any differences observed are not merely an artifact of differential scale usage. Here, we analyzed the taste variables as individual measures separately on the tongue tip and with the whole mouth, as well as combined into categories of tongue tip relative to whole mouth taste function. The latter categories are intended to capture the basic physiology of taste nerve interactions and alterations [10]. That is, previous experimental data have shown that reductions in taste on the tongue tip result in increases of taste from other cranial nerves to maintain, increase or even cause a phantom taste sensation from the whole mouth [10]. Severe insult to the taste system can result in depression of taste from multiple nerves and reduce whole mouth taste perception. Thus, categorizing taste on the tongue tip relative to whole mouth reflects the interplay between the cranial nerves that mediate oral taste perception.

Previous studies generally compared current or former smokers to nonsmokers, and some assessed dose responses of smoking exposure using either pack years or nicotine dependence. A recent study used data from the NHANES 2013–2014 to analyze associations between smoking status and taste perception by categorizing participants as never, former, or current smokers [66]. This study found that current smokers reported increased whole mouth bitter ratings compared to never smokers, but smoking status was not associated with tongue tip bitterness or NaCl taste intensity ratings. Here, we used validated questions to characterize smoking status by smoking history and nicotine dependence, both individually and combined. Smoking history is a more common measure in smoking studies, capturing how much is smoked (e.g., number of packs) over time (e.g., number of years). The history of smoking is different from the dependency of smoking [67], which reflects psychological and physiological addiction. Some individuals can smoke for several years and not be dependent [68] whereas others can be dependent without chronic smoking exposure. Time to first cigarette is the best single item measure of dependence [58] for associations with behavioral and health outcomes, including cancer, respiratory problems, risk of cardiovascular disease, and depression; in some cases, it can even capture associations independent of simple chronicity measures [67]. Highly dependent smokers may inhale cigarette smoke more deeply, resulting in greater nicotine extraction per cigarette [13, 69]. The current study found that current smokers with chronicity and more dependence reported lower tongue tip intensity compared to never smokers. Combined, current smokers with chronicity and more dependency had depressed tongue tip intensity; however, the beta (β) estimates did not suggest a synergistic effect. Rather, either chronicity or dependency in current smokers appeared solely sufficient to depress taste function on the tongue tip.

While the NHANES smoking questionnaire only measured TTFC in current smokers, number of pack years was captured in current and former smokers, allowing us to stratify chronicity by smoking status. Although smoking chronicity in current smokers was significantly associated with depressed taste function on the tongue tip, there was no association in former chronic smokers. According to smoking cessation studies, individuals who stop smoking may have improvements in their sense of taste. Previous research shows that the negative impact of smoking on taste can be lessened with smoking cessation from 2 to 9 weeks as shown with improvements in taste threshold by electrogustometry [30] and perceived bitterness [31]. In the current study, former smokers and those without chronicity/dependency did not report depressed taste intensities. Instead, non-chronic/less dependent current smokers and former smokers tended to report elevated whole mouth intensities. A recent crowd-sourced study also reported that current smokers experienced elevated taste intensity from concentrated propylthiouracil-impregnated discs compared to never smokers, but the study did not analyze other measures of smoking (e.g., smoking history or nicotine dependence) and there was no association in former smokers [36]. The exact mechanisms underlying elevated taste intensity of the whole mouth in light, non-dependent smokers are not fully understood. Smokers with less exposure to nicotine’s analgesic effects may be more sensitive to the irritant property of NaCl compared to never smokers as suggested in our laboratory-based study of smokers [35]. Taster status may also be implicated, as tasters who perceive oral stimuli as more intense are less likely to be heavy smokers [70, 71]. Our findings have similarities with research on alcohol and taste, which shows that those who drink less alcohol report greater intensity from bitter taste [72, 73], and suggest that smokers with elevated taste may have a sensory hindrance to larger consumption of cigarettes or, alternatively, that they may select menthol to attenuate tobacco bitterness [42]. When stratifying the analyses of current smokers by menthol cigarette use, menthol users were found to report elevated whole mouth quinine compared to never smokers. These results are consistent with prior work that found that menthol smokers report greater bitterness from propylthiouracil than non-menthol smokers [35].

Cigarettes and combustible tobacco smoke contain thousands of chemicals that may affect taste function either at the receptor or cellular level. Consistent with other studies [74], the present study found that chronic and dependent smoking was associated with smoking more cigarettes per day and having higher cotinine levels, suggesting greater nicotine exposure compared to non-chronic less dependent smokers. The question of interest is if depressed taste is associated with a more acute response to nicotine or a more chronic response to cigarette smoking. Nicotine is an irritant as well as an analgesic [75]. In animal models, cigarette smoke stimulates transient receptor potential (TRP) channels, including TRPA1, which is involved in nicotine-induced irritation [76] and TRPV1 receptors, which may connect nicotine with depressed behavioral responses [77]. Menthol, which produces a cooling sensation to reduce the harshness of cigarette smoke [43] also activates TRPA1 [78]. Single nucleotide polymorphisms (SNPs) on Mas-related G-protein coupled receptor and TRP genes, including TRPA1, have been shown to correlate with preference for menthol cigarette smoking, particularly among heavy smokers [79, 80]. In human studies, repeated nicotine exposure on the tongue causes a short-term desensitization, with variability across participants. At lower concentrations, nicotine does not cross-desensitize with capsaicin [81] or piperine [82], but does cross-desensitize at high concentrations of nicotine [83], likely through peripheral action to reduce the chemical excitability of intraoral nociceptors. Concentrated NaCl on the tongue tip is an irritant [84], which is sensed by the amiloride-sensitive sodium channel and the second sodium-permeable channel, a TRPV1 channel [85]. Behavioral studies show that NaCl and capsaicin are mediated by similar receptors [84] and can show similar sensitization through repeated application and desensitization through application and then removal. Nicotine desensitization could lower the perceived intensity of NaCl, however, the effects would wear off within 10 minutes [75]. Accordingly, we reasoned that it is very unlikely that an acute response to nicotine exposure resulted in short term desensitization that can explain our study findings. That is, although there were no restrictions about smoking outside of the MECs, participants completed multiple pre-examination tasks before the taste portion of the taste and smell examination [51], thereby placing testing outside the possible window of acute nicotine desensitization even if the individual had smoked immediately before entering the MEC.

Smoking behaviors and their associations with taste phenotypes are complex and may be related to genetic variations by race/ethnicity. For example, SNPs in taste receptor type 2 (TAS) genes, which encode a family of bitter receptors, have been associated with nicotine dependence in African American but not European American smokers [40]. In addition, taste responsiveness has been shown to differ significantly by race/ethnicity. Prior research in nonsmokers showed that African American and Hispanic participants rated taste sensations, including quinine and NaCl, higher than non-Hispanic whites [86]. Our analysis found that elevated whole mouth taste intensities reported by less dependent and former smokers were only evident in non-Hispanic whites. Prior research shows that bitter tasters may be less likely to be current smokers in European American populations but not African American populations [41]. These results may help explain our findings as heightened bitter sensations from cigarettes may contribute to less smoking or smoking cessation in this population. We did not find that associations between chronic and more dependent smoking and tongue tip taste intensities varied by race/ethnicity, which suggests that taste function in this region may have been altered by greater nicotine and smoking exposure.

There are multiple mechanisms by which greater nicotine exposure, either from chronic smoking or nicotine dependence, could alter taste sensations. Directly, smoking may damage regions of the tongue. In animal models, greater exposure to nicotine significantly reduces the size of the fungiform papillae [16]. In human studies, smokers have a lower mean density of papillae compared to nonsmokers [17, 28, 29]. Other studies found differences between smokers and nonsmokers in morphology and vascularization of filiform papillae [28, 87]. Indirectly, nicotine exposure may alter the quality of saliva [88, 89] and lower salivary flow rates [23], causing xerostomia [24]. Xerostomia can impair transport of tastes to receptors and increases risk of taste impairment (as reviewed [90]) and oral pain [91]. Consistent with present findings, a prior study found that adults who were chronic smokers reported elevated intensity from concentrated NaCl [35] at levels where NaCl is an oral irritant [92]. These smokers did not have patterns of taste function that suggested taste impairment or loss but rather a sensitivity to irritation in the mouth.

In the present study, the smoking related effects on taste were more apparent in younger adults. That is, when associations were stratified by age, they remained significant in younger smokers (40–65 years) but not older smokers. The lack of cigarette smoking effects on taste in older adults could result from complex interplay between genetics, environmental exposures, and aging. In comparing older and younger adults, older adults have lower taste perception, especially for bitter taste [93–96]. As one ages the likelihood of exposures to factors that impair bitter taste perception increase. Retention of bitter taste ability with aging may represent a genetic as well as environmental effect. There are intriguing data that support associations between longevity and bitter receptor genes polymorphisms, such that individuals who live longest taste greater bitterness [97, 98] and, among some individuals, there may be some protection for those who have genetic ability to taste greater bitterness to avoid excessive smoking [41]. The longer life among individuals who taste greater bitterness could reflect a more fit immune system [99], healthier ingestive behaviors toward alcohol [72, 73], and overall healthier diet [11], to name a few. Chronic smoking decreases life expectancy [100]. Chronic smokers who have genetic ability to taste great bitterness, select menthol cigarettes to balance unpleasant bitterness of the cigarettes [35, 42], yet have other conditions that associate with depression of bitterness including poor oral health [101] and sinus/respiratory infections [19, 20, 102]. There are also data to support reduction in ability to taste salt in older adults [103] especially if exposures damage both oral taste and trigeminal sensations in response to concentrated NaCl [10]. It is worth noting that most of the older subjects (>65 years) in the NHANES dataset reported being former or never smokers, with very few current dependent smokers. Thus, our analytical sample may have been underpowered to detect significant associations between smoking exposures and taste outcomes in older adults and should be further explored with larger sample sizes.

Smoking is associated with several sociodemographic and clinical characteristics that may increase risk of poor health. Smokers, particularly those with more dependence, are more likely to be of lower socio-economic status and less skilled labor [104], which may indicate less access to healthcare (including dental care) and greater risk of exposures in workplaces or living environments. More dependent smokers also have poorer health outcomes than nonsmokers and lower rates of smoking cessation compared to less dependence smokers [105]. Current smokers in the NHANES dataset were significantly more likely to have a high school or less education, lower income-to-poverty income, report fair/poor health, as well as have a greater chemosensory risk score (including sinonasal symptoms, xerostomia, history of tonsillectomy, and history of serious head/face injury). These characteristics may distort the relationship between smoking and taste, yet few studies assessing smoking and taste function control for these confounding factors. Although we cannot rule out residual confounding from environmental factors and chemosensory related conditions (e.g., employment/housing conditions, medications), we adjusted for several known confounding variables in our multivariable analysis (e.g., education status, income-to-poverty ratio, and chemosensory risk score), and our analyses support an independent relationship between smoking and taste even after controlling for these covariates.

The findings from this study have important clinical and public health relevance. First, our findings suggest that smoking exposures alter taste perception, which may have important clinical implications. Chronic and more dependent smokers who have altered taste may be at greater risk of unhealthy dietary behaviors, such as consuming salty, sweet, and high-fat foods and beverage items, which may further increase their risk of chronic diseases [106]. Second, we showed that TTFC, independent of chronicity, was related to lowered tongue tip taste, indicating that nicotine dependent smokers may have high nicotine exposure. Our findings also implicate taste as a barrier to nicotine consumption as elevated whole mouth taste perception was associated with non-chronic, less dependent, and former smoking. Finally, we found that menthol smokers had heightened taste sensations, particularly to bitterness, which is found in regular cigarette smoke. These findings support the recent announcement by the FDA to ban menthol from cigarettes as menthol masks unpleasant flavors and the harshness of cigarette smoke and may facilitate tobacco addiction, particularly in bitter tasters [107].

There are several strengths to our study. First, we used data from the NHANES Taste and Smell Protocol collected in 2013 and 2014 to analyze smoking and taste in a large, nationally representative sample of adults using standardized and validated measures of taste function. Second, using the NHANES smoking questionnaire, we analyzed different aspects of smoking, including chronicity and dependency. Third, we were able to control for several known confounding variables in our regression models to examine the association between measures of smoking and taste. However, we cannot rule out residual confounding from variables that were not included in analyses and may be risk factors for taste function. Other limitations to this study include the cross-sectional study design, which limits the ability to infer causality between smoking and taste function. Data presented here were collected prior to the explosion of eCigarette use (i.e., vaping) so these results cannot speak to potential effects of alternative nicotine delivery systems on chemosensation [108, 109]. Also, this study relied on self-reported history to determine smoking and alcohol exposures, as well as sociodemographic and taste-related characteristics. Although a significant interaction between age and smoking was found, we were not able to examine the interaction for taste function categories due to sample size limitations in the older age group. Finally, the taste protocol did not include other taste qualities such as sweet and sour. Previous research finds that whole-mouth quinine may be a good indicator of overall taste function [52], although it is known to activate only some of the TAS2Rs and may not fully capture the overall bitter taste perception [110, 111].

5. Conclusion

In NHANES 2013–2014 analysis of participants ≥40 years of age, measures of smoking were associated with taste function. Our study found that tongue tip bitter and salt tastes were perceived as less intense in current smokers with greater cigarette exposure compared to never smokers, particularly in younger smokers, suggesting that smoking may systematically alter tongue tip taste function. We also found evidence that smokers with less exposure and former smokers had elevated whole mouth taste, but more work is needed to elucidate the mechanisms involved. Understanding how smoking impacts taste perception is important as taste alterations may have implications for diet, alcohol intake and smoking behaviors, as well as associated health outcomes.

Highlights.

• High cigarette exposure associated with low taste perception on the tongue-tip.

• Former, light, or menthol smokers reported greater whole mouth taste perception.

• Patterns of smoking and taste associations were most evident in younger smokers.

• Both history and dependency measures are needed to clarify associations with taste.