Sleep-Disordered Breathing Is Associated With Impaired Odor Identification in Older U.S. Adults

Abstract

Background

Sleep-disordered breathing (SDB) is a common, underdiagnosed condition in older adults with major health consequences, including disrupted central nervous system functioning. Whether SDB may affect sensory function is unclear. We sought to address this question by comparing 2 forms of olfactory testing which measure peripheral and central olfactory processing.

Methods

We assessed SDB (survey-reported snoring frequency, nighttime apneic events, or diagnosis of sleep apnea) in the National Social Life, Health, and Aging Project, a nationally representative sample of older U.S. adults. Odor sensitivity (peripheral) and odor identification (central) were assessed with validated instruments. Logistic regression was used to test the relationship between SDB and olfaction, accounting for relevant covariates, including demographics, cognition, and comorbidity.

Results

Twenty-nine percent of older U.S. adults reported symptoms of SDB (apneic events or nightly snoring). Of these, only 32% had been diagnosed with sleep apnea. Older adults with SDB (those who reported symptoms or have been diagnosed with sleep apnea) were significantly more likely to have impaired odor identification (odds ratio 2.13, 95% confidence interval 1.19–3.83, p = .012) in analyses that accounted for age, gender, race/ethnicity, education, cognition, comorbidities (including depression), and body mass index. Presence of SDB was not associated with impaired odor sensitivity (odds ratio 1.03, 95% confidence interval 0.75–1.43, p = .84).

Conclusion

SDB is highly prevalent but underdiagnosed in older U.S. adults and is associated with impaired odor identification but not odor sensitivity. These data support the concept that SDB affects pathways in the central nervous system which involve chemosensory processing.

Sleep-disordered breathing (SDB) is a category of disorders characterized by abnormal respiratory patterns during sleep. These sleep disturbances, including obstructive sleep apnea, central sleep apnea, and upper airway resistance syndrome, are common in older adults and can lead to daytime sleepiness, depression, and serious systemic comorbidities including hypertension, heart failure, and stroke (1,2). Presence of SDB is an independent risk factor for the development of cognitive impairment (3,4), and increasing severity of SDB increases all-cause mortality risk (1). Despite these serious and well-documented consequences, more than 80% of men and 90% of women with symptoms of moderate to severe SDB remain undiagnosed (5).

Olfactory dysfunction is common among older adults and is increasingly recognized as an early marker for neurological disease (6–10). Olfactory dysfunction is more prevalent in those with SDB (11–13), but the mechanism of the association remains unclear. In particular, prior reports have been mixed as to whether the dysfunction lies in odor sensitivity or odor identification. Odor sensitivity measures the ability to detect presence of an odor and gauges function of the peripheral olfactory apparatus, while odor identification, which requires recognition and naming of common odors, incorporates central cognitive processing and memory (14–16). Patients with SDB may have impaired peripheral odor transduction, possibly due to decreased nasal airflow or olfactory nerve hypoxia (11). However, it is also possible that the cognitive changes caused by SDB interfere with central olfactory processing and recognition. In order to understand why patients with SDB perform worse on olfactory testing, it is necessary to determine whether the impaired performance reflects disruptions in peripheral or central processes in the same individual.

The relationship of SDB with peripheral and central olfactory function has yet to be investigated in a large, representative population study. Here, we use data from a nationally representative sample of U.S. older adults (the demographic with the highest incidence of both SDB and olfactory dysfunction (5,17)): the National Social Life, Health, and Aging Project (NSHAP).

Method

Participants

NSHAP is a longitudinal, nationally representative study of community-dwelling older U.S. adults (18). In 2005–2006, interviewers from the National Opinion Research Center conducted in-home interviews with respondents born between 1920 and 1947, and follow-up interviews with respondents and their co-resident partners were conducted in 2010–2011 (5-year follow-up) and 2015–2016 (10-year follow-up). Interviews included the collection of demographic data and extensive assessment of physical, psychological, and social well-being. Beginning in 2015–2016, a parallel study was also originated on a new cohort of respondents born between 1948 and 1965 and their co-resident partners. Here we analyze data from the 2015–2016 time point, when both extensive olfactory and sleep data were collected by design. We include co-resident partners who are age-eligible (born between 1920 and 1965). Further details regarding NSHAP study design are available elsewhere (18–20). The study was approved by the Institutional Review Boards of the University of Chicago and the National Opinion Research Center, and all respondents provided written informed consent.

Assessment of SDB

Sleep symptoms were assessed in a questionnaire filled out after the completion of in-home interviews and returned by mail (20). Return rate for the leave-behind questionnaire was 85%. Respondents were asked “How often do you snore?” and chose from “Every night,” “Most nights,” “Occasionally,” “Never,” or “I do not know.” Responses of “I do not know” were coded as nonresponse. Respondents were also asked “Has anyone ever told you that you stop breathing or gasp for breath during sleep?” and chose “No” or “Yes.” They were also asked “Has a doctor ever told you that you have sleep apnea?” and chose “No” or “Yes.” The use of self-reported snoring and apneic events has been shown to be useful for measuring SDB in large surveys in which polysomnography for all participants is not feasible (21–24). We classified respondents as having SDB if they reported either snoring every night or most nights, being told that they stop breathing or gasp for breath during sleep, or a previous diagnosis of sleep apnea, similar to prior work (24).

Assessment of Olfaction

Odor sensitivity and odor identification were measured in all respondents who had previously been eligible for olfactory testing in 2010–2011 (2/3 of the original cohort by design). Olfaction was tested using the odor sensitivity and odor identification components of the validated Olfactory Function Field Exam (25,26). Briefly, for odor sensitivity, respondents were presented with triads of Sniffin’ Stick odor pens with 2 odorless pens and 1 containing n-butanol and were asked to identify which pen contained the odor. Respondents were tested with 6 groups with increasing concentrations in the n-butanol-containing pen. Those who could detect the odor at 2 or fewer concentration levels were classified as anosmic. For odor identification, respondents were presented with 5 Sniffin’ Stick odor pens with common smells. For each pen, they were asked to identify the smell from a set of 4 word/picture choices. Refusals were coded as incorrect (refusal rates for the 5 odors were all less than 1%). Those who could only identify one or zero odors were classified as anosmic. Pens were obtained from Burghart Messtechnik (Wedel, Germany) and utilized according to the manufacturer’s instructions.

Additional Covariates

We accounted for potential confounders including age, gender, race/ethnicity, education, cognitive function, body mass index (BMI), depression, and medical comorbidities. Race and ethnicity were provided by self-report using standard National Institutes of Health categories (White/Caucasian, Black/African American, non-Black Hispanic, Other). Due to small numbers, the “Other” category included Asian/Pacific Islander, American Indian/Alaskan Native, and other responses. Education level was measured by the highest degree or certification earned. Cognition was measured with a validated adaptation of the Montreal Cognitive Assessment for survey administration (27–29) and scored from 0 to 20. Depression was measured using the NSHAP Depressive Symptoms Measure (30), an adaptation of the validated Center for Epidemiological Studies-Depression scale. This 11-item measure assesses participants’ depressive symptoms in the last week and is scored from 0 to 22, with higher scores indicating more frequent depressive symptoms. Height and weight were directly measured. Burden of medical comorbidity was quantified using a version of the validated Charlson index (31), adapted for the self-reported medical history collected in NSHAP (32).

Statistical Analysis

Analyses were performed using person-level weights, accounting for differential nonresponse and differential probability of selection. This approach enables conclusions about the population of older U.S. adults, not just the survey sample (19). Design-based standard errors were calculated using the linearization method together with the strata and Primary Sampling Unit indicators provided with the data set. Multivariate logistic regression models were used to evaluate the relationship of SDB with olfactory function, adjusting for age, gender, race/ethnicity, education, BMI, medical comorbidity, depression, and cognition. Results are presented as odds ratios (ORs) and 95% confidence intervals (CIs). Wald tests were used to determine p values. Statistical significance was set at p < .05. All statistical analyses were conducted using Stata Version 15.1 (StataCorp LLC, College Station, TX).

Results

Thirty percent of U.S. older adults had SDB, as defined by snoring every night or most nights, stopping breathing or gasping for air at night, or previous diagnosis of sleep apnea (n = 3877, Table 1). Interestingly, among older adults with self-reported symptoms of SDB (snoring, stopping breathing, or gasping for air), only 32% had been diagnosed with sleep apnea.

Table 1.

Population Characteristics

Age
Mean ± SD (range)	66.7 ± 11.4 (50–95)
Gender
Male	46.2%
Female	53.8%
Race/Ethnicity
White	75.1%
Black	12.1%
Hispanic, non-Black	8.5%
Other	4.3%
Education level
Less than high school	11.0%
High school/equivalent degree	23.3%
Some college/Associate’s degree	36.6%
Bachelor’s degree or more	29.1%
BMI
Mean ± SD (range)	29.5 ± 6.4 (13.8–69.6)
Charlson comorbidity index
Mean ± SD (range)	1.2 ± 1.5 (0–11)
NSHAP Depressive Symptoms Measure (modified CES-D)
Mean ± SD (range)	5.4 ± 4.7 (0–22)
MoCA-SA
Mean ± SD (range)	14.2 ± 3.8 (0–20)
How often do you snore?
Every night	16.7%
Most nights	19.1%
Occasionally	50.3%
Never	13.9%
Has anyone ever told you that you stop breathing or gasp for breath during sleep?
Yes	17.7%
No	82.3%
Has a doctor ever told you that you have sleep apnea?
Yes	13.5%
No	86.5%
Snore every night or most nights OR stop breathing/gasp for breath	29.0%
Snore every night or snore most nights OR stop breathing/gasp for breath OR previously diagnosed with sleep apnea	30.1%

Notes: BMI = body mass index; CES-D = Center for Epidemiological Studies-Depression scale; MoCA-SA = Montreal Cognitive Assessment for survey administration; NSHAP = National Social Life, Health, and Aging Project. Population characteristics of U.S. older adults. Based on a sample of 3877 older adults. Weighted averages and percentages calculated with person-level survey weights.

Older adults with SDB were more than twice as likely to be anosmic based on the odor identification task (OR 2.13, 95% CI 1.19–3.83, p = .012) in analyses accounting for age, gender, race/ethnicity, education, BMI, comorbidities, depression, and cognition (n = 2316, Table 2). Older participants were more likely to be anosmic (OR 1.15, 95% CI 1.09–1.20, p < .001), as were women (OR 0.50, 95% CI 0.27–0.95, p = .036). Participants with higher depressive symptom scores (OR 1.09, 95% CI 1.01–1.17, p = .031) and those with worse cognition scores (OR 0.88, 95% CI 0.79–0.98, p = .024) were also more likely to be anosmic.

Table 2.

Effect of Sleep-Disordered Breathing on Odds of Odor Identification Anosmia

	Odds Ratio (95% Confidence Interval)
Sleep-disordered breathing	2.13 (1.19–3.83)*
Age (per year)	1.15 (1.09–1.20)*
Gender (vs men)
Women	0.50 (0.27–0.95)*
Race/Ethnicity (vs White)
Black	1.06 (0.50–2.25)
Hispanic, non-Black	1.15 (0.44–2.97)
Other	1.09 (0.21–5.51)
Education level (vs <high school)
High school/equivalent degree	1.24 (0.47–3.31)
Some college/Associate’s degree	0.92 (0.33–2.56)
Bachelor’s degree or more	2.33 (0.81–6.70)
BMI	0.97 (0.88–1.06)
Charlson comorbidity index	0.89 (0.70–1.14)
NSHAP depressive Symptoms Measure	1.09 (1.01–1.17)*
MoCA-SA	0.88 (0.79–0.98)*

Notes: BMI = body mass index; MoCA-SA = Montreal Cognitive Assessment for survey administration; NSHAP = National Social Life, Health, and Aging Project. Sleep-disordered breathing is associated with higher likelihood of odor identification anosmia. n = 2316. Logistic regression with survey weights. Sleep-disordered breathing is defined as either snoring every night or most nights, stopping breathing or waking up gasping at night, or previous diagnosis of sleep apnea. Anosmia in odor identification is defined as the ability to detect fewer than 2 of 5 odors.

*p < .05.

We tested the robustness of these results by varying the criteria for SDB. Similar results were obtained for those reporting relevant symptoms alone and not counting those with only a sleep apnea diagnosis (OR 2.15, 95% CI 1.18–3.94, p = .015, Supplementary Table 1). Odds of anosmia in the odor identification task increased when we analyzed only those experiencing the most severe symptoms: those who snore every night, stop breathing, or gasp for air (OR 2.35, 95% CI 1.06–5.23, p = .037, Supplementary Table 2), and only those who stop breathing or gasp for air (OR 2.59, 95% CI 1.16–5.78, p = .022, Supplementary Table 3). When considering only respondents with previous diagnoses of sleep apnea, we found similar odds of odor identification anosmia, but the association fell slightly short of statistical significance (OR 2.04, 95% CI 0.97–4.27, p = .056, Supplementary Table 4).

There was no association between SDB and odor sensitivity (OR 1.03, 95% CI 0.75–1.43, p = .84) when accounting for age, gender, race/ethnicity, education, BMI, comorbidities, depression, and cognition (n = 2315, Table 3). In this analysis, increasing chronological age was the only covariate associated with increased odds of odor sensitivity anosmia (OR 1.03, 95% CI 1.01–1.06, p = .004). Unlike for odor identification anosmia, cognition, gender, and depression were not associated with odor sensitivity anosmia.

Table 3.

Effect of Sleep-Disordered Breathing on Odds of Odor Sensitivity Anosmia

	Odds Ratio (95% Confidence Interval)
Sleep-disordered breathing	1.03 (0.75–1.43)
Age (per year)	1.03 (1.01–1.06)*
Gender (vs men)
Women	0.91 (0.64–1.30)
Race/Ethnicity (vs White)
Black	1.05 (0.60–1.82)
Hispanic, non-Black	1.28 (0.70–2.35)
Other	1.75 (0.83–3.65)
Education level (vs <high school)
High school/equivalent degree	0.77 (0.42–1.39)
Some college/Associate’s degree	0.72 (0.37–1.42)
Bachelor’s degree or more	0.77 (0.39–1.52)
BMI	1.01 (0.98–1.05)
Charlson comorbidity index	1.04 (0.93–1.17)
NSHAP Depressive Symptoms Measure	0.99 (0.96–1.02)
MoCA-SA	0.97 (0.92–1.03)

Notes: BMI = body mass index; MoCA-SA = Montreal Cognitive Assessment for survey administration; NSHAP = National Social Life, Health, and Aging Project. Sleep-disordered breathing is not associated with higher likelihood of odor sensitivity anosmia. n = 2315. Logistic regression with survey weights. Sleep-disordered breathing is defined as either snoring every night or most nights, stopping breathing or waking up gasping at night, or previous diagnosis of sleep apnea. Anosmia in odor sensitivity is defined as the ability to detect odor at fewer than 3 of 6 n-butanol concentrations.

*p < .05.

Discussion

Prior studies have linked olfactory dysfunction with SDB (11–13), but these studies have been inconsistent in distinguishing whether the impairment lies in odor sensitivity, which reflects peripheral olfactory transduction, or odor identification, which reflects more central olfactory processing and cognition. Here, we use a large, nationally representative sample of older adults to clarify this relationship. We find that there is no association between SDB and impaired odor sensitivity. However, older adults with SDB are more than twice as likely to have impaired odor identification. Thus, our findings suggest that SDB impairs olfaction via a central mechanism; the ability of the nose and olfactory nerve to transduce an odor is unaffected, but the ability to recognize the odor, match it to the memory of a smell, and recognize its name or image is disrupted.

The long-term disruption in brain function caused by SDB is well-documented, with the most apparent deficits in the neuropsychological domains of attention, executive function, and verbal memory (33,34), domains which have also been closely linked to odor identification (35–37). This provides a compelling neuropsychological link between SDB and olfaction: SDB impairs function in the very same cognitive domains needed to recognize and name common odors. As expected, we find an association between impaired odor identification and worse cognitive performance. However, it is interesting to note that this relationship does not affect the association between SDB and impaired odor identification, supporting the idea that the olfactory system is sensitive to disruption of pathways that are distinct from these components of cognitive function.

There are several limitations to this study. Due to the logistics of this omnibus national study in which interviewers visited respondents’ homes for only 2–3 hours, it was not feasible to perform polysomnography. Therefore, we are unable to diagnose specific conditions such as obstructive sleep apnea and instead rely on self-reports of symptoms and past diagnoses to classify patients with SDB. This method has been useful in prior large population studies (22–24), and our findings are in line with the expected prevalence of SDB in this population; however, future studies could use polysomnography in large cohorts to clarify this relationship and distinguish between specific conditions such as obstructive versus central sleep apnea. NSHAP did not collect information on daytime sleepiness; however, impaired odor identification has only been associated with 24-hour sleep deprivation and not with subjective sleepiness (38). Additionally, data were not collected on nasal obstruction. While obstruction may affect sleep quality, is it not a major contributor to SDB and is not thought to have a significant independent effect on cognitive function (39). Future work will be necessary to clarify the extent to which disruption in airflow affects cognitive outcomes. Finally, in this study, we use cross-sectional data, so we are unable to determine causality or the timing of olfactory impairment relative to sleep symptom onset.

The olfactory system has been proposed to be a “canary in the coalmine” (40) for human health because olfactory impairment can be an early signal of central nervous system pathology that is otherwise not yet detectable. Indeed, olfactory dysfunction is an early indicator of mild cognitive impairment, cognitive decline, and neurodegeneration (6–8), as well as a predictor for specific diseases including Alzheimer’s disease and Parkinson’s disease (9,10). The fact that SDB causes impairment in the same cognitive domains that are involved in olfactory identification, along with our finding that older adults with SDB are substantially more likely to have dysfunctional odor identification, raises the question of whether olfaction could also be a “canary in the coalmine” for cognitive sequelae of SDB. Future work should evaluate the role of olfactory testing in predicting potentially accelerated cognitive decline in adults with SDB.

We found that 30% of older U.S. adults have SDB, consistent with prior reports that range from 30% to as high as 80% (5,41,42). Here, we find that among those who report current symptoms of SDB, 68% have never been diagnosed with sleep apnea. This substantial underdiagnosis, consistent with prior reports (5), underscores the importance of improved screening and detection of SDB in older adults, especially given its significant long-term sequelae. Polysomnography is expensive and time-consuming, which limits its use despite the large numbers of patients who need this diagnostic tool. Given the significant underdiagnosis of SDB, a reliable method not only to screen but also to recognize early signs of its resultant cognitive impairment could substantially improve the health of older adults.

Funding

This work was supported by the National Institute on Aging and the National Institutes of Health (R01AG043538, R01AG048511, R37AG030481). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conflicts of Interest

None declared.

Author Contributions

Study concept and design: J.K.S., X.Y., J.M.P., and M.K.M. Acquisition and analysis of data: J.K.S., X.Y., and K.E.W. Drafting of the manuscript: J.S. Critical revision of the manuscript for important intellectual content: J.K.S., K.E.W., M.K.M., and J.M.P.