Carotid Intima Media Thickness, Atherosclerosis, and 5-Year Decline in Odor Identification: The Beaver Dam Offspring Study

Abstract

Background.

The purpose of this study was to determine if subclinical markers of atherosclerosis are associated with a decline in olfactory function.

Methods.

The San Diego Odor Identification Test was administered to 2,302 participants (age 21–84 years) at the baseline (2005–2008) and 5-year follow-up (2010–2013) examinations of the Beaver Dam Offspring Study. A decline in odor identification was defined as a decrease in San Diego Odor Identification Test score of 2 or more (range 0–8) from Beaver Dam Offspring Study 1 to Beaver Dam Offspring Study 2. Carotid intima media thickness and plaque, blood pressure, pulse wave velocity, and body mass index were measured and other risk factor data were obtained by interview.

Results.

Overall 3.2% of participants had a decline in San Diego Odor Identification Test score at 5 years. In age- and sex-adjusted models, mean intima media thickness (odds ratio = 1.17, 95% CI = 1.01, 1.34, per 0.1 mm) and number of sites (range 0–6) with carotid artery plaque (odds ratio = 1.35, 95% CI = 1.11, 1.65, per site) at baseline were associated with an increased risk for decline. Plaque score (odds ratio = 1.24, 95% CI = 1.01, 1.53) remained a significant independent predictor of olfactory decline in a model that included age, sex, hypertension, body mass index, alcohol, and smoking.

Conclusions.

Subclinical atherosclerosis was associated with an increased risk for olfactory decline indicating that atherosclerosis may be one of the risk factors for the decline in olfactory function seen with aging. Strategies to improve vascular health may also benefit olfactory health.

Although olfactory impairment is common in older adults, many are unaware of their dysfunction suggesting some decline in olfactory function occurs slowly over time (1,2). Gradual declines in olfactory function with age may have different etiologies than sudden impairments that are associated with specific events (trauma or infection). A slow progression of subclinical decline in olfactory function would be similar to the long trajectory recognized in the development of cardiovascular disease (CVD), which begins as vascular changes and atherosclerosis early in adulthood and manifests as clinical disease years later. In CVD, the risk factor profile in midlife is believed to an important predictor of healthy aging and longevity (3); this may be true for olfactory impairment as well. Olfactory impairment and CVD have some risk factors in common such as smoking, diabetes mellitus, and exercise. Smoking, an important risk factor for atherosclerosis (4,5), has been associated with olfactory dysfunction in some studies (1,2,6,7), but not all (8). People with diabetes mellitus have an increased risk of developing atherosclerosis (9) and CVD (10), and some studies have found people with diabetes have more olfactory dysfunction, though again, findings have not been consistent (1,2,8,11). Weinstock and coauthors (11) found within a diabetes cohort that those with macrovascular disease (coronary artery or peripheral vascular) had impaired olfaction compared with those without macrovascular disease. In population studies, older adults in the Epidemiology of Hearing Loss Study who were current smokers or had a history of stroke were more likely to have olfactory impairment in cross-sectional analyses (1). Longitudinal analyses found the use of lipid-lowering medication and regular exercise were associated with a reduced risk of developing olfactory impairment at 5 years (12) and exercise was also associated with a reduced risk at 10 years (13). In a younger cohort, the Beaver Dam Offspring Study (BOSS), thicker carotid artery intima media thickness (IMT), a subclinical marker of generalized atherosclerosis, was associated with prevalent olfactory impairment (2). Together, these findings suggest that atherosclerosis may be associated with olfactory dysfunction that occurs with aging. However, to establish temporal sequence, longitudinal data are needed from a younger cohort that has not yet entered the period of greatest risk for olfactory impairment. The purpose of this study was to determine if subclinical markers of atherosclerosis are associated with a decline in olfactory function at 5 years in a primarily midlife cohort of adults participating in the longitudinal BOSS.

Methods

The BOSS is a longitudinal cohort study of the adult children of participants in the ongoing, population-based Epidemiology of Hearing Loss Study (1993–present) (14–16). There were 3,296 participants (ages 21–84 years) in the baseline BOSS (2005–2008), and 2,792 (84.7%) of them, plus an additional 80 people who were unable to participate in the baseline phase, participated in the 5-year follow-up phase (2010–2013; 21–88 years) (17). There were 2,838 participants who had olfactory data at baseline and of those, 2,302 (81%) completed the olfactory testing at follow-up and had data available for these analyses.

Informed consent was obtained from all participants prior to each examination and approval for this research was obtained from the Health Sciences Institutional Review Board of the University of Wisconsin. Examination and questionnaire data were obtained by trained and certified examiners following similar standardized protocols at each examination.

Olfaction Measure

The San Diego Odor Identification Test (SDOIT) was used to measure olfaction at both examinations (2). The SDOIT is a standardized test with good test–retest reliability; detailed SDOIT methods have been previously reported (1,2,18,19). The SDOIT consists of eight common odorants (coffee, chocolate, peanut butter, Play doh, baby powder, bubble gum, cinnamon, and mustard) presented in a random order and a manner that prevents visual cues, with a 45-second lag between odorant presentations to minimize adaptation. To aid in identification, a picture array with the odorants plus 12 distracters is available to participants during the test. If an odor is not identified, the participant is given the correct name, and it is presented later in the test sequence to allow for learning of unfamiliar odors. The SDOIT score is the number of odors (0–8) correctly identified after two trials. Olfactory impairment was defined as identifying fewer than six of the eight odorants correctly (1,2). There were 2,222 participants without an olfactory impairment at baseline who were at risk for developing olfactory impairment at follow-up. Olfactory decline was defined as a decrease of 2 or more in SDOIT score from baseline to the 5-year follow-up examination (18). Of the 2,302 with olfactory data, 15 were removed from the olfactory decline analyses because their baseline SDOIT score was 0 or 1 and could not decline by 2 or more, leaving 2,287 participants.

Atherosclerosis Measures

High resolution B-mode carotid artery ultrasound images were obtained (AU4, Esaote North America Inc., Indianapolis, IN) at baseline on the right and left sides using a modified Atherosclerosis Risk in Communities scanning protocol (20). Focused views of the distal common and proximal internal carotid arteries and the bifurcation were obtained. The IMT of the near and far walls was measured in 1.0cm segments of the common, internal, and bifurcation following a modified Atherosclerosis Risk in Communities reading protocol (21); the mean of the 12 walls was calculated for the IMT. Plaque was determined to be present if acoustic shadowing was associated with at least one of the following characteristics: a change in wall shape (protrusion into the lumen), a change in wall texture, an IMT > 1.5mm, or in the absence of acoustic shadowing at least two of the characteristics were present (22). Plaque score was defined as the number of sites (0–6) with plaque present. Mean IMT and plaque score were analyzed separately due to the interrelatedness of the measures.

Other Vascular Measures

Height (cm) and weight (kg) were measured and body mass index (BMI; kg/m²) calculated. Blood pressure was measured after the participant had been sitting for 5 minutes using an automated blood pressure machine (Dinamap, GE Health Systems, Milwaukee, WI). Hypertension was defined as a measured systolic blood pressure ≥140mm Hg or diastolic blood pressure ≥90mm Hg at examination or self-reported physician diagnosed high blood pressure and use of antihypertensive medication. Nonfasting total and high-density lipoprotein cholesterol levels were obtained and non–high density lipoprotein cholesterol was calculated (total minus high-density lipoprotein cholesterol). Pulse wave velocity (arterial stiffness) was measured on the right side from the carotid artery to the femoral artery and from the carotid artery to the radial artery on a subset of participants (n = 1,417) using the Complior System (Alam Medical, France), which has been previously validated (23).

Additional Covariates

Medical history, demographic, and behavioral factors were obtained by questionnaire at both examinations. Available nasal health information included a history of nasal polyps, deviated septum, allergies, sinus infections, problems or a cold in the week before examination, nasal congestion on the day of examination, use of nasal steroids, and history of head injuries (concussion, broken nose, or skull fracture). Relevant demographic and behavioral factors included education (years completed), smoking history (current, past, or never), alcohol intake (usual consumption of beer, wine, or liquor in the past year converted to grams per week), and exercise (number of times per week exercise long enough to work up a sweat). CVD was classified as present if the participant reported a history of physician diagnosed angina, myocardial infarction, or stroke.

To adjust for incident nasal congestion at follow-up, participants with self-reported upper respiratory or sinus infections or problems in the week before, or a stuffy nose on the day of the 5-year follow-up examination, among those with none of these conditions reported at baseline, were classified as having new nasal congestion. Participants who reported that they experienced a head injury (concussion, broken nose, or skull fracture) between the baseline and follow-up examinations were classified as having a new head injury.

Statistical Analyses

All analyses were performed using SAS, version 9.2 (SAS Institute, Inc., Cary, NC). The overall age-stratified 5-year incidence of olfactory impairment and the occurrence of olfactory decline were determined by computing the binomial proportions and 95% confidence limits. Because age and sex have been important determinants of olfactory impairment in previous studies (1,2), IMT, plaque score, other cardiovascular measures, and olfaction-related covariates were evaluated for their association with olfactory decline in individual logistic regression models adjusted for age and sex. Covariates that were significantly associated with olfactory decline in the age- and sex-adjusted models were assessed in multivariable models with IMT and plaque score.

Results

The mean age at baseline was 49 years and 1,244 (54%) were women and 1,058 (46%) were men. At baseline, the mean SDOIT score was 7.6 (range: 0–8; standard deviation = 1.0) and 88% of the cohort scored 7 or 8 at both examinations. Most participants (67%) had more than 12 years of education. Less than 4% of the cohort had a history of CVD, 24% had one or more sites with carotid plaque, 36% had hypertension, 45% had a BMI of 30kg/m² or higher, and 16% were current smokers (Supplementary Table 1).

Five-Year Incidence of Olfactory Impairment

The 5-year incidence of olfactory impairment was 1.7% (95% confidence interval [CI] = 1.2, 2.3) overall and increased significantly with age (odds ratio [OR] = 1.64, 95% CI = 1.40, 1.93) but did not differ by sex (OR = 1.71, 95% CI = 0.88, 3.29, adjusted for age). The 5-year incidence rate was highest (7.7%) among those 65 years and older and no one under the age of 35 years developed an olfactory impairment in 5 years (Table 1).

Table 1.

The 5-Year Incidence of Olfactory Impairment and Olfactory Decline by Age

Incidence of Olfactory Impairment				Olfactory Decline
Age (y)	N at Risk	# Cases	% (95% CI)	N at Risk	# Cases	% (95% CI)
21–34	130	0	0	131	0	0
35–44	637	3	0.5 (0.1, 1.4)	641	5	0.8 (0.3, 1.8)
45–54	850	11	1.3 (0.7, 2.3)	867	23	2.7 (1.7, 4.0)
55–64	475	14	3.0 (1.6, 4.9)	502	27	5.4 (3.6, 7.7)
65–84	130	10	7.7 (3.8, 13.7)	146	17	11.6 (6.9, 18.0)
All	2222	38	1.7 (1.2, 2.3)	2287	72	3.2 (2.5, 4.0)

Note: CI = confidence interval.

Olfactory Decline

The overall rate of olfactory decline was 3.2% (n = 72) and the risk for decline increased with age (OR = 1.59, 95% CI = 1.41, 1.79, for every 5-year increase, adjusted for sex), but there was no difference in risk of decline between men and women (OR = 1.33, 95% CI = 0.83, 2.16, adjusted for age).

Risk factors for olfactory decline.

In age- and sex-adjusted logistic regression models, carotid mean IMT (OR = 1.17, 95% CI = 1.01, 1.34, per 0.1mm) and plaque score (OR = 1.35, 95% CI = 1.11, 1.65, per site) were associated with an increased risk for decline in SDOIT score at follow-up (Supplementary Table 1). Other cardiovascular risk factors associated with an increased risk of olfactory decline were current smoking (OR = 2.22, 95% CI = 1.15, 4.28; vs never smoked) and hypertension (OR = 1.78, 95% CI = 1.05, 2.99), whereas BMI (OR = 0.95, 95% CI = 0.91, 0.99, per kg/m²) was associated with a slight decreased risk of decline. Two levels of weekly alcohol consumption were borderline significant. Those who reported no alcohol consumption in the last year at baseline had an increased risk (OR = 1.89, 95% CI = 1.00, 3.56) of olfactory decline, whereas those who reported consuming 15–74g per week had a reduced risk (OR = 0.44, 95% CI = 0.19, 1.02) compared with those consuming 14g or less a week. History of a new head injury between the baseline and 5-year follow-up (OR = 2.85, 95% CI = 1.15, 7.03) and new nasal congestion at follow-up (OR = 1.83, 95% CI = 1.03, 3.23) were also associated with an increased risk of decline in SDOIT score (Supplementary Table 1). Exercise, non–high density lipoprotein cholesterol, statin use, pulse wave velocity, history of CVD, and diabetes mellitus were not significantly associated with decline in SDOIT score. There were too few participants with nasal steroid use to support analyses with olfactory decline.

Mean IMT (OR = 1.16, 95% CI = 1.01, 1.34, per 0.1mm) remained a significant independent predictor of decline in SDOIT score in a multivariable model including age, sex, hypertension, BMI, and alcohol consumption. However, when smoking was included in this model, IMT was slightly attenuated (IMT OR = 1.13, 95% CI = 0.98, 1.31, per 0.1mm), and smoking moderately attenuated (OR = 1.78, 95% CI = 0.90, 3.55, current vs never), and neither were statistically significant (Table 2).

Table 2.

Multivariable Models for 5-Year Olfactory Decline

Risk Factors	IMT Model OR (95% CI)	Plaque Model OR (95% CI)
IMT, per 0.1 mm	1.13 (0.98, 1.31)
Plaque score (0–6), per site		1.24 (1.01, 1.53)
Age	1.42 (1.23, 1.65)	1.38 (1.19, 1.60)
Sex	1.22 (0.71, 2.09)	1.20 (0.71, 2.07)
Hypertension	2.07 (1.18, 3.63)	1.94 (1.09, 3.45)
BMI (kg/m 2)	0.93 (0.88, 0.97)	0.94 (0.89, 0.99)
Alcohol intake (g/wk) (vs 0–14)
None	1.78 (0.92, 3.42)	2.00 (1.02, 3.91)
>14 and <75	0.40 (0.17, 0.95)	0.46 (0.19, 1.10)
75+	0.87 (0.45, 1.68)	0.95 (0.49, 1.89)
Smoking (vs never)
Current	1.78 (0.90, 3.55)	1.68 (0.84, 3.38)
Past	1.51 (0.85, 2.68)	1.46 (0.81, 2.63)

Notes: BMI = body mass index; CI = confidence interval; IMT = intima media thickness; OR = odds ratio.

Plaque score (OR = 1.24, 95% CI = 1.01, 1.53, per site) remained a significant independent predictor of olfactory decline in a multivariable model that included age, sex, hypertension, BMI, alcohol, and smoking, though smoking (OR = 1.68, 95% CI = 0.84, 3.38, current vs never) was attenuated and not statistically significant (Table 2). New head injury and new nasal congestion were not significant in the fully adjusted models for IMT or plaque score and were not retained in the models.

Discussion

In this study, we found carotid IMT and plaque were associated with a decline in odor identification score at 5 years indicating that atherosclerosis may be one of the risk factors for the decline in olfactory function seen with aging. The presence of plaque is indicative of atherosclerosis and the number of sites with plaque remained an independent predictor of olfactory decline in the full multivariable model that included other cardiovascular risk factors. The 5-year risk for olfactory decline increased 24% for each additional carotid site with plaque. Whereas some of the risk factors for atherosclerosis have been shown previously to be associated with olfaction, we believe this is the first longitudinal cohort study to show directly that atherosclerosis may be associated with olfactory dysfunction.

Mean IMT was significantly associated with olfactory decline in a multivariable model with age, sex, hypertension, BMI, and alcohol but unlike plaque, was slightly attenuated and no longer statistically significant when smoking was included in the model. Although intima media thickening and plaque are related, intima media thickening may be an early stage of atherosclerosis or, at low levels, an adaptive response to hemodynamic changes that are not related to atherosclerotic thickening (24). It is possible that in this relatively young cohort, plaque may be a more definitive marker of atherosclerotic changes than IMT. However, smoking was also attenuated and not statistically significant in models with IMT or plaque indicating atherosclerosis may be in the pathway between smoking and olfactory decline.

It has been reported that the association between smoking and olfactory function might be the result of the effect of the chemicals present in cigarette smoke on the nasal epithelium (25). However, the current study suggests smoking may also affect olfaction through its effects on the vascular system. Frye and coauthors (7) showed pack-years of smoking was significantly related to olfactory score and years since cessation of smoking was related to improvement in score. Longer exposure to smoking has been found to increase the risk for atherosclerosis; studies have found that current smokers have a higher risk of plaque and a dose–response association was seen with pack-years and IMT, whereas past smokers had no association with plaque and IMT tended to be thinner among those with longer cessation of smoking (26,27).

The inclusion of other cardiovascular risk factors in the multivariable model with IMT or plaque score is debatable. Hypertension, smoking, BMI, and alcohol consumption have all been associated with thicker IMT and the progression of atherosclerosis (4,5,28,29). Due to the causal nature of the associations between these risk factors and IMT, atherosclerosis and each other, including all these factors in the same models, may be inappropriate because IMT and plaque may be in the causal pathway. Though, with the exception of smoking, the other CVD risk factors all retained significance in the models with IMT and plaque score, which may indicate that they have associations with olfactory decline that are independent of atherosclerosis.

Our findings are consistent with a recent study of diabetes and olfaction that also found hypertension to be associated with lower olfactory function (30). Hypertension, through its widespread effects on the vascular system, may have effects on the olfactory system as well. Weekly alcohol consumption had a J-shaped association with olfactory decline similar to its association with carotid IMT (29). Abstainers may have health-related reasons for not drinking or not drink because of a past history of alcohol abuse and these conditions also may have affected their olfactory function (31,32). BMI had an inverse relationship with olfactory decline; however, the average BMI in this cohort is high (about 30kg/m²), the effect size was small, and the difference in mean BMI between decliners and nondecliners was less than 1kg/m².

The mechanism for the association between atherosclerosis and olfactory decline is unknown, but changes in the vascular walls of the carotid artery are used as a marker of generalized atherosclerosis. Subclinical atherosclerosis may cause subtle changes in blood flow that cause ischemia or reduce the efficiency of removal of byproducts of cellular metabolism. A recent study in middle-aged adults reported increased IMT, at levels considered to be in the “normal range,” was associated with diminished cerebral neuronal viability (33). In the Framingham Offspring Cohort (mean age 58 years), carotid IMT and atherosclerosis were associated with subclinical brain magnetic resonance imaging indices of ischemia (34). If similar changes are occurring in the vasculature of the peripheral or central olfactory structures, this could possibly affect olfactory function.

A report of a head injury since the baseline exam and nasal infection or congestion at follow-up, well-known risk factors for olfactory impairment (35,36), were both associated with olfactory decline in age- and sex-adjusted models but were not significant in the full multivariable models. Arterial stiffness, diabetes mellitus, and a history of CVD were not associated with olfactory decline in this study, but the prevalence of these conditions at baseline was low in this cohort. Exercise and statin use, which were associated with a reduced risk of developing olfactory impairment in the Epidemiology of Hearing Loss Study, were not associated with olfactory decline in this study (12,13). Exercise was assessed similarly in both studies, but we may have been underpowered to detect modest associations in this cohort due to the younger age and low rate of olfactory decline. A similar pattern has been noted for physical activity and IMT where the impact of physical activity increased with age (37). Statins have been associated with slower progression and regression of IMT (38) suggesting their use might be beneficial for olfaction as seen in the Epidemiology of Hearing Loss Study (12). However, in this younger cohort, it is possible the elevated, though nonsignificant, statin estimate may be a marker of an increased CVD risk profile and not a drug effect. Overall, the 5-year incidence of olfactory impairment was very low in this cohort though the risk for impairment increased with age as expected. The mean age of the cohort was 49 years at baseline, so much of the cohort has not yet reached the age of greatest risk for olfactory impairment.

There are some limitations of this study that should be noted. Olfactory threshold levels were not obtained in this study, which may have provided a more precise estimate of olfactory function. Olfactory decline was based on testing status at the baseline and 5-year follow-up examinations and may not reflect the participant’s long-term olfactory status as olfactory function may fluctuate. However, this was a longitudinal study in a large well-defined cohort with good retention, a standardized odor identification test with good reliability was used to measure olfaction at both examinations (18) and several vascular risk factors were measured using established methods.

Conclusion

In this study, atherosclerosis was associated with a decline in olfactory score in 5 years. These longitudinal findings add support to the previous cross-sectional report from the BOSS cohort that found mean IMT associated with the prevalence of olfactory impairment (2). Other modifiable cardiovascular risk factors were also associated with olfactory decline. These findings suggest that vascular health in midlife may be an important determinant of olfactory health later in life and some olfactory decline seen with aging may be preventable. This cohort is relatively young and only a small percentage had decline in their olfactory function in 5 years, additional longer term follow-up is needed to determine if these associations are seen with olfactory impairment as the cohort ages and enters the high risk period.

Supplementary Material

Supplementary material can be found at: http://biomedgerontology.oxfordjournals.org/

Funding

The project described was supported by R01AG021917 (K.J.C.) from the National Institute on Aging , National Eye Institute, and National Institute on Deafness and Other Communication Disorders and an unrestricted grant from Research to Prevent Blindness. The content is solely the responsibility of the authors and does not necessarily reflect the official views of the National Institute on Aging or the National Institutes of Health.