Exposure to Cadmium, Lead, and Tobacco Smoke and the 10-Year Cumulative Incidence of Olfactory Impairment

Key Points

Question

What environmental exposures are associated with the 10-year cumulative incidence of olfactory impairment?

Findings

In this longitudinal cohort study including 2312 individuals, the overall incidence of olfactory impairment was low, but rates increased with age. Participants with higher levels of blood cadmium or high tobacco smoke exposure, as a current smoker or from environmental tobacco smoke, were at an increased risk of developing olfactory impairment.

Meaning

Modifiable environmental exposures may contribute to olfactory impairment that occurs with aging.

Abstract

Importance

Olfactory impairment is common in older adults. Identification of modifiable risk factors for olfactory impairment at midlife has the potential to reduce the burden of olfactory impairment at older ages.

Objective

To determine the 10-year cumulative incidence of olfactory impairment and evaluate potentially modifiable risk factors for impairment including exposure to cadmium, lead, and tobacco smoke.

Design, Setting, and Participants

Data from the Beaver Dam Offspring Study, a longitudinal cohort study of sensory health and aging in a general population, were available from examinations at baseline (2005-2008), 5 years (2010-2013), and 10 (2015-2017) years. A total of 2312 participants without olfactory impairment at baseline and with olfaction data available at the 5- and/or 10-year examination were included. The present study was conducted from April 28, 2020, to January 8, 2021.

Main Outcomes and Measures

Olfactory impairment was measured by the San Diego Odor Identification Test. Cox discrete-time proportional hazards analyses were used to model associations between baseline covariates, including blood cadmium and lead levels and tobacco smoke exposure, and the 10-year cumulative incidence of olfactory impairment.

Results

Of the 2312 participants, 1269 (54.9%) were women; mean age was 49 years (range, 22-84 years) at baseline. The 10-year cumulative incidence of olfactory impairment was 4.6% (95% CI, 3.7%-5.6%) and increased with age. Because of high collinearity, cadmium and tobacco smoke exposure were modeled separately. In a multivariable adjusted model, higher blood cadmium level (hazard ratio [HR], 1.70; 95% CI, 1.05-2.74) was associated with the 10-year cumulative incidence of olfactory impairment. Substituting tobacco smoke exposure for cadmium in the model, high exposure to tobacco smoke as a current smoker (HR, 2.94; 95% CI, 1.63-5.29, smoker vs no exposure) or from environmental tobacco smoke (HR, 2.65; 95% CI, 1.24-5.63, high vs no exposure) was also associated with an increased risk for developing olfactory impairment. Blood lead levels were not associated with olfactory impairment.

Conclusions and Relevance

Results of this longitudinal cohort study suggest that modifiable environmental exposures may contribute to olfactory impairment that occurs with aging. Identification of modifiable risk factors for olfactory impairment may lead to prevention strategies that have the potential to reduce the burden of olfactory impairment at older ages.

This cohort study examines the incidence of olfactory impairment in adults following long-term exposure to potential risk factors.

Introduction

Impairments in olfactory function may affect safety, nutrition, and quality of life and are associated with depression and increased risk for cognitive impairment, neurodegenerative diseases, and mortality.^{1,2,3,4,5,6,7,8,9} Olfactory impairment is common in older people, with up to 40% of those aged 70 years and older experiencing impairment.¹⁰ Previous population studies have identified some modifiable risk factors, including atherosclerosis, cerebrovascular disease, cardiovascular-related risk factors (exercise, alcohol, and smoking), medications, nasal conditions, and head injury, that may contribute to age-related olfactory impairments in middle-aged and older adults.^{11,12,13,14,15} Gradual declines in olfactory function with age could also be the result of exposure to environmental toxins, but few studies in the general population have evaluated these exposures as potential risk factors for olfactory impairment.

Cadmium and lead are known neurotoxins that accumulate in the body.^16,17 These heavy metals are pervasive in the environment primarily as the result of anthropogenic activities.^16,17,18 Cadmium has been associated with olfactory dysfunction in occupational studies in which exposure levels are high, but less is known about its effects on olfactory function in people without occupational exposure.^19,20,21 Lead exposure is considered a potential cause for olfactory dysfunction, but previous studies have been few and results inconsistent.^{22,23,24,25,26} Cadmium exposure in the nonsmoking population primarily occurs through the ingestion of foods that have taken up cadmium from contaminated soil (root and green leafy vegetables, grains).^17,18 Tobacco leaves also readily take up cadmium from the soil, and inhalation of tobacco smoke is the primary source of exposure in cigarette smokers.^17,18,27 Because the human body absorbs more than twice as much cadmium through inhalation vs ingestion, smokers typically have higher cadmium levels than nonsmokers.^17,18,27 In the general population, lead exposure most commonly occurs through the ingestion of contaminated food or water or through hobbies.¹⁶ Although lead is also present in tobacco smoke, smoking is not considered a primary source of exposure.^16,28

In addition to cadmium and lead, tobacco smoke contains thousands of chemical compounds, many of which are associated with morbidity and mortality in humans.²⁸ Smoking has been associated with olfactory dysfunction in some, but not all, studies, although few longitudinal studies have been conducted.^{10,11,12,15,29,30,31,32,33} Absent from most previous studies has been the evaluation of the effects of environmental tobacco smoke (ETS) (ie, secondhand smoke) on olfactory function. Exposure to tobacco smoke, either as a smoker or through ETS, is associated with cardiovascular disease, stroke, cancer, and premature death.³⁴

The purpose of this cohort study was to determine the 10-year cumulative incidence of olfactory impairment in the Beaver Dam Offspring Study (BOSS), a general population cohort of primarily middle-aged adults, and evaluate potential risk factors associated with the development of olfactory impairment including cadmium, lead, and tobacco smoke exposure.

Methods

Participants

Data are from participants in the BOSS (2004-present), a longitudinal study of sensory health and aging in the adult offspring of the Epidemiology of Hearing Loss Study (1993-2020), a population-based study in Beaver Dam, Wisconsin.^35,36 Beaver Dam is a city of more than 15 000 residents located in south central Wisconsin. Participation in the BOSS required having had a parent in the Epidemiology of Hearing Loss Study but did not require residing in Beaver Dam. At baseline, 75% of BOSS participants lived within 75 miles of Beaver Dam, 7% lived elsewhere in Wisconsin, and 18% resided in another state or country; 70% reported their home was in a city/town and 30% resided in rural areas. Self-identified race and ethnicity data were collected at baseline as required by the National Institutes of Health and 98% of the BOSS cohort self-identified as non-Hispanic White; the cohort reflects the race/ethnicity of Beaver Dam when the parent population (Epidemiology of Hearing Loss Study) was identified in 1987-1988.³⁵ BOSS participants were first examined in 2005-2008 with follow-up examinations at 5 (2010-2013) and 10 (2015-2017) years.^15,37,38 Of the 3298 baseline participants, 2848 (86.4%) had olfactory data: 109 participants with and 2739 participants without olfactory impairment.²⁹ Written informed consent was obtained from all participants before each examination and approval for the present research, conducted from April 28, 2020, to January 8, 2021, was obtained from the University of Wisconsin Health Sciences Institutional Review Board. Participants received financial compensation for time and travel expenses. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.

Olfaction Measure

The 8-item San Diego Odor Identification Test (SDOIT) was used to measure olfactory function at all 3 examinations using the same standardized protocol; detailed methods of the SDOIT have been reported.^15,29 The SDOIT has been shown to have good reliability.³⁹ The odorants used in the SDOIT are coffee, peanut butter, chocolate, bubble gum, baby powder, Play-Doh, mustard, and cinnamon. Odorants were presented in a random order with a 45-second lag between presentations to minimize adaptation. If an odor was not identified correctly at the first presentation, the participant was given the correct name and the odorant was presented again later in the test sequence. A picture array with the odorants plus 12 distracters was available during the test to aid with identification. The SDOIT score is the number of odors (0-8) correctly identified after 2 trials, and olfactory impairment was defined as identifying fewer than 6 of 8 odorants correctly.^10,29

Neurotoxins

Whole blood samples obtained at baseline were stored at −80 °C. Samples were assayed in 2016 for cadmium and lead at the Wisconsin State Laboratory of Hygiene using inductively coupled plasma mass spectrometry. Quality control samples were within plus or minus 10% of the target and 10% of samples were duplicated.^38,40 Because cadmium and lead data were highly skewed, data were analyzed categorically as quintiles (Q), with high levels (Q5) compared with lower levels (Q1-4) and as a doubling of levels, using a log base 2 transformation.

Tobacco Smoke Exposure

Smoking history and ETS exposure were obtained by interview; the ETS questionnaire was previously validated against blood cotinine levels⁴¹ (eTable 1 in the Supplement). A categorical tobacco smoke exposure variable was created by first classifying participants as current smokers or nonsmokers and then classifying nonsmokers’ environmental tobacco smoke exposure as none, low, moderate, or high based on their reported hours of exposure at home, work, or in social settings.⁴¹ Data were analyzed with each level of tobacco smoke exposure (low, moderate, high, and current smoker) compared with nonsmokers with no ETS exposure (none).

Other Potential Risk Factors

Metabolic (total and high-density lipoprotein cholesterol, hemoglobin A_1c) and inflammatory (white blood cell count, high-sensitivity C-reactive protein, interleukin-6, soluble intercellular adhesion molecule-1, and soluble vascular cell adhesion molecule-1) markers were measured in baseline blood samples.^37,38 Diabetes was defined as a hemoglobin A_1c level greater than or equal to 6.5% (to convert to proportion of total hemoglobin, multiply by 0.01) or a physician diagnosis of diabetes or suspected diabetes with current treatment. Interleukin-6, soluble intercellular adhesion molecule-1, and soluble vascular cell adhesion molecule-1 were analyzed as tertiles (tertile 3 vs tertiles 1 and 2) and high-sensitivity C-reactive protein was divided into 3 risk groups (<1, 1-3, and >3 mg/L) based on established clinical cut points.

Vascular health measures were obtained at baseline using standardized protocols. Height (centimeters) and weight (kilograms) were measured and body mass index (calculated as weight in kilograms divided by height in meters squared) was determined. Obesity was defined as a body mass index greater than 30. Blood pressure was measured, and hypertension was defined as a systolic blood pressure greater than or equal to 140 mm Hg or diastolic blood pressure greater than or equal to 90 mm Hg at examination or self-reported physician-diagnosed hypertension and use of antihypertensive medication. Carotid artery ultrasonographic scans were obtained for intima media thickness and plaque assessment.^14,15,37

Medical history, behavioral, and demographic factors were obtained by questionnaire.³⁷ Cardiovascular disease (CVD) was classified as present if the participant reported a history of any 1 of 11 physician-diagnosed conditions or procedures.⁸ Other relevant factors included medications used in the past month (statins, antihypertensives, and oral and nasal corticosteroids), educational level (years completed), occupation, alcohol intake (usual consumption of beer, wine, or liquor in the past year converted to grams of ethanol per week), history of heavy alcohol consumption (4 or more drinks/d), exercise (number of times per week long enough to work up a sweat) and history of head injuries (concussion, broken nose, and skull fracture).^12,13,15 Available nasal health information included history of nasal polyps, deviated septum, and allergies; self-report of sinus problems, a cold, and physician-diagnosed sinus infection in the week before examination; and nasal congestion on the day of examination.¹³

Statistical Analysis

Analyses were conducted with SAS, version 9.4 (SAS Institute Inc). Kaplan-Meier survival estimates were used for calculating cumulative incidence of olfactory impairment over the 10-year period. Cox discrete-time proportional hazards analyses were used to model associations between baseline covariates and 10-year cumulative incidence of olfactory impairment. Covariates were first evaluated in models including age and sex and those with a significance level less than .05 were included in the full multivariable model. Because smoking is a primary source of cadmium exposure, cadmium and tobacco smoke exposure were modeled separately. Sex-specific analyses were conducted and interactions between sex and other covariates were tested in the overall models.

Some variables were further evaluated as time-varying covariates (carotid intima media thickness, plaque, hypertension, CVD, and head injury) by lagging the values 5 years. For participants with 5 years of follow-up data regarding incident status, baseline status of the risk factor was used. For participants with a full 10 years of follow-up data regarding incident status, 5-year status of the risk factor was used. To account for potential effects of nasal congestion at follow-up, the final multivariable models were adjusted for participant self-report of a cold, sinus problems, or a physician-diagnosed sinus infection in the week before or nasal congestion on the day of the examination for the phase (5- or 10-year) when they became an incident case or were censored.

Results

Of the 2739 baseline participants at risk for developing an olfactory impairment, 2312 (84.4%) had follow-up olfactory data and were included in these analyses. The 427 participants (15.6%) without follow-up data were slightly younger (mean age, 47.3 vs 48.8 years) and more likely to be current smokers (24.2% vs 16.6%) than those with follow-up, but there were no meaningful differences between groups regarding sex, years of education, or baseline SDOIT score. Of the 2312 included participants, 1269 (54.9%) were women and 1043 (45.1%) were men, and the mean age was 49 years (range, 22-84 years). Baseline cadmium and lead levels were low, but higher in smokers than nonsmokers in all ETS categories (Table 1). In Q5 of cadmium, 62.9% of the participants were smokers and 30.1% were nonsmokers with no or low ETS exposure.

Table 1. Baseline Cadmium and Lead Levels in the Beaver Dam Offspring Study.

Variable	Participants, No. (%)	Mean (CV)a
		Cadmium, μg/L	Lead, μg/dL
Total	2124	0.27 (1.09)	1.27 (0.63)
Baseline tobacco smoke exposure
None	964 (45.4)	0.22 (0.82)	1.09 (0.59)
Low	499 (23.5)	0.20 (0.80)	1.33 (0.60)
Moderate	176 (8.3)	0.20 (0.75)	1.49 (0.63)
High	137 (6.5)	0.23 (0.76)	1.32 (0.74)
Smoker	348 (16.4)	0.87 (0.88)	1.63 (0.59)
Incident olfactory impairment
No	2040 (96.0)	0.26 (1.08)	1.26 (0.63)
Yes	84 (4.0)	0.33 (1.43)	1.57 (0.54)

Abbreviation: CV, coefficient of variation.

SI conversion factors: To convert cadmium to nanomoles per liter, multiply by 8.896; lead to micromoles per liter, multiply by 0.0483.

^aGeometric mean.

The overall 10-year cumulative incidence of olfactory impairment was 4.6% (95% CI, 3.7%-5.6%) (Table 2). The risk of developing olfactory impairment increased with age (hazard ratio [HR], 1.61; 95% CI, 1.45-1.79 per 5-year increase, adjusted for sex) and the risk of impairment for men was higher than women (HR, 1.34; 95% CI, 0.89, 2.03 for men vs women adjusted for age), although the imprecision of the estimate prevents definitive conclusions regarding the role of sex.

Table 2. Ten-year Cumulative Incidence of Olfactory Impairment in the Beaver Dam Offspring Study.

Baseline age, y	Women		Men		Total (%)
	No. at risk (%)	Cumulative incidence, % (95% CI)	No. at risk (%)	Cumulative incidence, % (95% CI)	No. at risk (%)	Cumulative incidence, % (95% CI)
<45	450 (35.5)	0.5 (0.1-2.0)	355 (34.0)	2.1 (0.9-4.6)	805 (34.8)	1.2 (0.6-2.4)
45-54	478 (37.7)	2.9 (1.6-5.0)	407 (39.0)	4.0 (2.4-6.6)	885 (38.3)	3.4 (2.3-4.9)
55-64	261 (20.6)	6.7 (4.1-10.9)	226 (21.7)	8.5 (5.4-13.4)	487 (21.1)	7.5 (5.4-10.5)
65-84	80 (6.3)	22.7 (14.3-35.0)	55 (5.3)	20.9 (11.6-35.9)	135 (5.8)	22.3 (15.5-31.4)
21-84	1269	4.1 (3.0-5.4)	1043	5.2 (3.9-6.9)	2312	4.6 (3.7-5.6)

In age- and sex-adjusted models, the 10-year cumulative incidence of olfactory impairment was associated with higher blood cadmium level (Q5 vs Q1-4: HR, 1.89; 95% CI, 1.19-3.02) and high tobacco smoke exposure, either as a current smoker vs none (HR, 2.64; 95% CI, 1.49-4.69) or from high-level ETS exposure vs none (HR, 2.66; 95% CI, 1.30-5.41) (Table 3). A doubling of cadmium level was associated with a 22% increase (HR, 1.22; 95% CI, 1.03-1.45) in the risk for developing olfactory impairment. In analyses excluding participants who were current smokers, the cadmium estimate for the incidence of olfactory impairment was attenuated (Q5 vs Q1-4: HR, 1.32; 95% CI, 0.65-2.68). Blood lead level was not associated with olfactory impairment. A history of head injury, analyzed as a time-varying covariate, was associated with an increased risk for 10-year cumulative incidence of olfactory impairment (HR, 1.93; 95% CI, 1.11-3.38), whereas obesity was associated with a reduced risk (HR, 0.65; 95% CI, 0.43-0.99). Interactions with sex were present for antihypertensive medication and CVD. Use of antihypertensive medication was associated with olfactory impairment in women (HR, 2.25; 95% CI, 1.21-4.16) but not men (HR, 0.74; 95% CI, 0.40-1.37) as was time-lagged CVD, which was also associated with olfactory impairment in women (HR, 2.52; 95% CI, 1.21-5.24) but not men (HR, 0.76; 95% CI, 0.31-1.86). There was no association between a history of nasal polyps (HR, 0.90; 95% CI, 0.29-2.86) or use of nasal corticosteroids (HR, 0.56; 95% CI, 0.18-1.77) and the cumulative incidence of olfactory impairment. There were too few users of oral corticosteroids to allow analyses.

Table 3. Potential Baseline Risk Factors and 10-Year Cumulative Incidence of Olfactory Impairment in the Beaver Dam Offspring Study.

Baseline	Available data, No.	Incident olfactory impairmenta
		No. (%)		Age-sex adjusted HR (95% CI)
		No	Yes
Categorical risk factors
Women	1269	1225 (55.2)	44 (48.4)	1 [Reference]
Men	1043	996 (44.8)	47 (51.7)	1.34 (0.89-2.03)b
Educational level, ≥16 y	2302	760 (34.4)	23 (25.6)	0.76 (0.47-1.23)
Occupation (not professional or managerial)	2238	1458 (67.8)	57 (65.5)	0.91 (0.59-1.42)
Cadmium level (Q5 ≥ 0.53 μg/L vs Q1-4)	2124	390 (19.1)	25 (29.8)	1.89 (1.19-3.02)
Lead level (Q5 ≥ 2.07 μg/dL vs Q1-4)	2124	386 (18.9)	25 (29.8)	1.28 (0.80-2.07)
Tobacco smoke exposure
None	1055	1020 (45.9)	35 (38.5)	1 [Reference]
Low	539	522 (23.5)	17 (18.7)	1.07 (0.59-1.93)
Moderate	190	181 (8.2)	9 (9.9)	1.25 (0.60-2.61)
High	144	134 (6.0)	10 (11.0)	2.66 (1.30-5.41)
Current smoker	384	364 (16.4)	20 (22.0)	2.64 (1.49-4.69)
Deviated septum	2282	179 (8.2)	14 (15.6)	1.95 (1.10-3.46)
Allergies	2305	865 (39.1)	40 (44.0)	1.46 (0.96-2.22)
Head injury	2312	619 (27.9)	29 (31.9)	1.93 (1.11-3.38)c
Obesity, BMI >30	2291	991 (45.1)	39 (42.9)	0.65 (0.43-0.99)
Exercise ≥1/wk vs none	2310	1373 (61.9)	52 (57.1)	0.99 (0.65-1.51)
History heavy alcohol used	2310	396 (17.9)	20 (22.0)	1.26 (0.75-2.10)
Alcohol use, g/wk past year
None	238	221 (10.0)	17 (18.7)	1.64 (0.92-2.94)
0-14	946	911 (41.0)	35 (38.5)	1 [Reference]
>14-74	563	548 (24.7)	15 (16.5)	0.80 (0.43-1.47)
>74	564	540 (24.3)	24 (26.4)	1.17 (0.68-2.02)
Depressive symptoms	2284	315 (14.4)	11 (12.2)	0.80 (0.43-1.52)
Hypertension	2310	771 (34.8)	49 (53.9)	1.11 (0.72-1.73)
Diabetes	2303	121 (5.5)	7 (7.8)	0.79 (0.36-1.73)
Carotid plaque	2274	478 (21.9)	35 (38.5)	1.06 (0.65-1.73)
Statin medication	2311	323 (14.6)	24 (26.4)	1.06 (0.67-1.66)
Antihypertensive medication	2311
Women		249 (20.3)	24 (54.6)	2.25 (1.21-4.16)b
Men		280 (28.1)	17 (36.2)	0.74 (0.40-1.37)b
hsCRP (>3.0 mg/L)	2204	478 (22.6)	18 (20.2)	0.78 (0.46-1.32)
IL-6 (T3 ≥ 2.28 pg/mL)	2186	671 (32.0)	39 (44.3)	1.26 (0.82-1.94)
sICAM-1 (T3 ≥ 238.5 ng/mL)	2192	681 (32.4)	37 (42.1)	1.32 (0.86-2.02)
sVCAM-1 (T3 ≥ 635.4 ng/mL)	2204	681 (32.2)	38 (42.7)	1.02 (0.65-1.60)
TNF-α (T3 ≥ 0.61 pg/mL)	2204	687 (32.5)	40 (44.9)	1.45 (0.95-2.20)
Continuous risk factors, mean (SD)
Age, y, per 5 y	2312	48.4 (9.5)	57.7 (9.8)	1.61 (1.45-1.79)e
HDL-C, per 10 mg/dL	2256	49.9 (14.7)	50.5 (14.2)	1.09 (0.94-1.25)
Non-HDL-C, per 30 mg/dL	2256	153.8 (39.1)	147.8 (31.7)	0.89 (0.76-1.05)
WBC, per103/μL	2256	7.3 (1.9)	7.3 (2.0)	1.06 (0.95-1.18)
Mean IMT, per 0.1 mm	2277	0.65 (0.13)	0.71 (0.17)	0.98 (0.86-1.13)

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); HDL-C, high-density lipoprotein cholesterol; HR, hazard ratio; hsCRP, high-sensitivity C-reactive protein; IL-6, interleukin-6; IMT, carotid artery intimal medial thickness; Q, quintile; sICAM-1, soluble intercellular adhesion molecule-1; sVCAM-1, soluble vascular cell adhesion molecule; T, tertile; TNF, tumor necrosis factor; WBC, white blood cell count.

SI conversion factors: To convert cadmium to nanomoles per liter, multiply by 8.896; lead to micromoles per liter, multiply by 0.0483; non-HDL-C or HDL-C to millimoles per liter, multiply by 0.0259; and WBC to ×10⁹ per liter, multiply by 0.001.

^aTotal of 2221 participants without incident olfactory impairment and 91 with incident olfactory impairment.

^bAdjusted for age.

^cTime-lagged variable.

^dHeavy alcohol consumption is defined as 4 or more drinks per day.

^eAdjusted for sex.

Owing to collinearity, cadmium and tobacco smoke exposure were modeled separately. Higher blood cadmium level (Q5 vs Q1-4: HR, 1.70; 95% CI, 1.05-2.74) remained associated with an increased risk for developing olfactory impairment in a multivariable model adjusted for age, CVD, antihypertensive medication, head injury, obesity, and deviated septum (Table 4, model 1). Substituting tobacco smoke exposure for cadmium in the model, high ETS exposure vs none (HR, 2.65; 95% CI, 1.24-5.63) and being a current smoker vs none (HR, 2.94; 95% CI, 1.63-5.29) were associated with an increased risk for olfactory impairment in the multivariable adjusted model (Table 4, model 2). Adjusting for nasal congestion at follow-up did not change associations (eTable 2 in the Supplement).

Table 4. Multivariable Models for 10-Year Cumulative Incidence of Olfactory Impairment in the Beaver Dam Offspring Study.

Variable	HR (95% CI)
	Model 1, with cadmium (n = 2082)	Model 2, with tobacco smoke exposure (n = 2261)
Age, per 5 y	1.62 (1.43-1.83)	1.69 (1.49-1.91)
Cadmium, Q5 vs Q1-4	1.70 (1.05-2.74)	NA
Tobacco smoke exposure
None	NA	1 [Reference]
Low	NA	1.24 (0.67-2.31)
Moderate	NA	1.43 (0.67-3.02)
High	NA	2.65 (1.24-5.63)
Current smoker	NA	2.94 (1.63-5.29)
Cardiovascular diseasea
Women	2.14 (1.03-4.47)	2.28 (1.10-4.75)
Men	0.62 (0.23-1.65)	0.78 (0.32-1.90)
Antihypertensive medication
Women	2.14 (1.11-4.10)	2.23 (1.17-4.22)
Men	0.80 (0.40-1.60)	0.84 (0.43-1.64)
Head injurya	1.81 (1.002-3.25)	1.82 (1.02-3.26)
Obesity	0.65 (0.41-1.02)	0.62 (0.39-0.97)
Deviated septum	1.66 (0.90-3.07)	1.74 (0.95-3.19)

Abbreviations: CVD: cardiovascular disease; HR, hazard ratio; NA, not applicable; Q, quintile.

^aTime-lagged variable.

Discussion

In this prospective 10-year cohort study, higher baseline blood cadmium level and exposure to tobacco smoke, either as a current smoker or from high ETS exposure, were associated with an increased risk for developing olfactory impairment. These results suggest that modifiable environmental exposures may contribute to olfactory impairment that occurs with aging. Overall, the 10-year cumulative incidence of olfactory impairment was low, but the incidence increased with age with rates doubling with each successive decade. Among those older than 65 years, more than 1 in 5 participants developed an olfactory impairment in 10 years. Identification of modifiable risk factors for olfactory impairment at midlife has the potential to reduce the burden of olfactory impairment at older ages.

To our knowledge, this is the first longitudinal cohort study to report an association between blood cadmium concentration at levels found in the general population and the development of olfactory impairment. Previous occupational studies have reported associations between blood cadmium levels and olfactory dysfunction but at levels 10 to 100 times higher than present in the BOSS cohort.^19,21 Blood cadmium levels in the BOSS were slightly lower than those reported in National Health and Nutrition Examination Surveys (NHANES), 1999-2010.⁴² A recent study in NHANES, 2011-2014 found no association between blood cadmium level and measured olfactory dysfunction, but the NHANES study was cross-sectional compared with the long-term prospective design of our study.⁴³

Environmental exposure to cadmium has been associated with bone, kidney, and cardiovascular diseases; cancer; mortality; and, previously in the BOSS cohort, an increased risk for developing hearing and vision impairments.^{17,18,38,40,44,45} On a cellular level, cadmium has been associated with apoptosis of neurons, oxidative stress, impaired neurogenesis, endocrine disruption, and epigenetic effects.^45,46 It is believed that inhaled cadmium is taken up by the olfactory epithelium and transported along the olfactory sensory neurons to the olfactory bulb in the brain, where it accumulates.^46,47,48

Current smoking and high ETS exposure in nonsmokers were also associated with an increased risk for developing olfactory impairment. Tobacco smoke is thought to affect olfactory function by damaging the olfactory epithelium and increasing apoptosis of olfactory sensory neurons.^49,50 Effects may be reversible, though, as olfactory function in smokers has been reported to improve after smoking cessation.^51,52 Exposure to tobacco smoke is also associated with atherosclerosis, which has been associated with a decline in olfactory function in this cohort.^15,28,34 It has been shown that people exposed to ETS are exposed to the same toxic chemicals smokers are, albeit at lower doses, and are at risk for the same adverse health conditions.³⁴ In the present study, the estimate of risk for developing olfactory impairment among those with high ETS exposure was only slightly less than that of current smokers. Our results emphasize the importance of including ETS exposure, in addition to smoking, as a potential risk factor in future studies of olfaction.

Previous studies of smoking and olfaction have had mixed results, but have varied in design, prevalence of smoking, and participant ages.^{10,11,12,15,29,30,31,32,33} In the present study of a middle-aged cohort with 10 years of follow-up, participants were just entering the phase of life during which health effects from long-term smoking may start to manifest and loss to follow-up due to smoking-related mortality was still low; these factors may have improved our ability to detect an association. In addition, previous studies of smoking and olfactory dysfunction may have been confounded by misclassifying those with high ETS exposure with nonsmokers, which could have attenuated potential associations.

This study cannot disentangle the effects of cadmium from those of tobacco smoke on olfactory function. In the general US population, tobacco smoke is a major source of cadmium exposure, making it difficult to determine whether either exposure has independent associations with olfactory function.^17,27,45 The mean cadmium level in current smokers was 4 times higher than in nonsmokers in this study. Many adverse health effects associated with smoking are also associated with cadmium exposure.^17,27,28 Therefore, cadmium could be one of the mediators of the association between tobacco smoke exposure and olfactory impairment. However, cadmium is only one of thousands of components of tobacco smoke.²⁸ Nonsmokers with high ETS exposure also had an increased risk for olfactory impairment despite having lower cadmium levels, suggesting that other components in tobacco smoke may also affect olfactory function. Although the cadmium estimate for olfactory impairment was attenuated in analyses excluding current smokers, we cannot determine whether this attenuation was attributable to the absence of an association in nonsmokers or reduced power to detect an association because of the small number of nonsmoking participants in the highest quintile. The route of cadmium exposure may also be an important factor. Inhaled cadmium has direct contact with the olfactory pathway, which could potentially have more adverse effects on olfactory function than ingested cadmium.^46,47,48

Although baseline lead levels were slightly higher among those who developed olfactory impairment, lead was not associated with olfactory impairment in adjusted models. A recent cohort study reported a marginal association between blood lead level and olfactory dysfunction, but at concentrations higher than in the BOSS cohort.²³ Previous occupational studies of lead exposure and olfactory dysfunction have been mostly negative.^22,24,25

In addition to environmental exposures, a history of a head injury, a well-established cause of olfactory dysfunction and previously associated with olfactory decline in this cohort, was associated with an increased risk for developing olfactory impairment.^15,53 The association of obesity with a reduced risk of olfactory impairment is also consistent with previous results in this cohort.¹⁵ Two cross-sectional studies have reported similar findings^33,54 and a recent prospective cohort study in older adults reported poor olfaction associated with weight loss.⁵⁵

A history of CVD and use of antihypertensive medications were associated with an increased risk for olfactory impairment but in women only. Atherosclerosis and hypertension have previously been associated with olfactory decline and impairment in middle-aged adults.^14,15 The association with antihypertensive medications could be related to effects of the medication or the condition for which they were prescribed.⁵⁴ Why these associations were seen only in women is unclear, but there may be differential effects of vascular disease by sex or it may be easier to detect an association in women because CVD is less common. Inflammatory and metabolic markers were not associated with olfactory impairment in this study.

Strengths and Limitations

The strengths of this study include the longitudinal design with 10 years of follow-up in a well-characterized cohort with extensive covariate data. Cadmium, lead, and olfactory function were measured, and a standardized questionnaire was used to quantify ETS exposure.⁴¹

Limitations of this study include the lack of an odor threshold test that could be more sensitive to small changes in olfactory function than odor identification. The overall incidence of olfactory impairment was low in this relatively young and healthy cohort, which may have limited our ability to detect some associations. In addition, we did not have data on the participants’ history of chronic rhinosinusitis or sinus surgery and were not able to adjust for these conditions in analyses. However, use of nasal corticosteroids or a history of nasal polyps at baseline or nasal congestion at follow-up were not associated with an increased risk of developing olfactory impairment. Participants were primarily non-Hispanic White, which could limit the generalizability of these findings to other populations. Although urinary cadmium levels are considered a better measure of cumulative exposure, the effect of circulating blood cadmium was the exposure of interest for this prospective study of risk factors for developing olfactory impairment.^42,56

Conclusions

In this cohort study we found modifiable risk factors, including exposures to cadmium and tobacco smoke, associated with an increased risk of developing olfactory impairment. Although the 10-year incidence of olfactory impairment was low in early midlife, rates increased in later midlife. Promotion of smoking cessation, which would reduce exposures to cadmium and ETS as well, may be beneficial for olfactory health with aging.

Supplement.

eTable 1. Environmental Tobacco Smoke Questionnaire

eTable 2. Multivariable Models for 10-Year Cumulative Incidence of Olfactory Impairment in the Beaver Dam Offspring Study, Adjusting for Nasal Congestion at Follow-up