Skin Barrier Function and Cognition among Older Adults

TO THE EDITOR

Age-associated declines in the skin barrier function result in increased exposure to cutaneous microbes and toxins and lead to chronic, low-grade systemic inflammation (Hu et al., 2017). Chronic inflammation is now thought to be an important mechanism in the onset and progression of dementia, and multiple chronic inflammatory states have been associated with subsequent cognitive impairment (Heppner et al., 2015; Wen et al., 2022). Epidemiologic studies of eczema (atopic dermatitis) support the rationale for investigating skin barrier decline and cognition: eczema is a pathogenic example of skin barrier dysfunction; eczema severity has been associated with cognitive function (Silverberg et al., 2020), and moderate to severe eczema among older adults has been associated with an increased risk of dementia in multiple longitudinal studies (Eriksson et al., 2008; Magyari et al., 2019; Pan et al., 2021). Our objective was to examine the association between skin barrier function, as measured by water loss and tape stripping, and cognition.

Our pilot study used data from the Baltimore Longitudinal Study of Aging (BLSA), an ongoing cohort designed to study aging. Our analyses include 237 of 279 participants who underwent skin barrier measurements in addition to all usual follow-up assessments between January 2019 and March 2020. Participants who had previously received a diagnosis of cognitive impairment (n = 24), were under the age of 50 years (n = 14), or were lacking data on cognitive measures (n = 4) were excluded. All participants provided written informed consent. The BLSA protocol is approved by the National Institutes of Health Central Institutional Review Board.

Participants were asked to avoid caffeine, topical products, and smoking 3 hours before measurement, and all measures were performed in the same examination center 15 to 20 minutes after acclimatization. Transepidermal water loss (TEWL) was measured using the AquaFlux Model AF2000 (Biox Systems, London, United Kingdom), a closed chamber device. After a baseline TEWL measurement on the left volar forearm 10 cm from the wrist, tape stripping was performed with 14-mm D-Squame discs (CuDerm, Dallas, TX). Tape strips were repeatedly applied to the same location, and TEWL measurements were collected after every five strips until TEWL reached four times the baseline. The primary outcome was the area under the curve when TEWL was plotted against the number of tape strippings (Alexander et al., 2018). We also evaluated skin barrier integrity (defined as the total number of tape strips) as a secondary outcome.

Participants underwent cognitive testing at all visits. The cognitive measures have been previously shown to have high construct validity for predicting cognitive decline and dementia, including among BLSA participants. Five cognitive domain scores—verbal memory, attention, executive function, verbal fluency, and visuospatial ability—were computed from z-scores of standardized cognitive tests. All cognitive assessments conducted concurrently with and within 10 years before the skin barrier measures were included.

We used linear mixed-effects models to evaluate the association between epidermal barrier function and each of the cognitive domain scores. We anchored the time of origin at the visits where both skin barrier and cognitive assessments were conducted (time zero), and the time of follow-up for cognitive assessments was coded as years before time zero. The main effect of skin barrier function estimates the cross-sectional association between skin barrier function and cognition, whereas the interaction term between skin barrier function and time estimates the association of skin barrier function with longitudinal cognitive change. We included fixed effects in adjusted models for skin barrier function, sex, age at skin barrier assessment, race, education, room temperature, humidity, season, body mass index, history of eczema, smoking, and time and the two-way interactions between time and all other predictors. All continuous covariates were centered at their medians, and the skin barrier variables were centered at the median and scaled by the interquartile range. Random effects included intercept and time with unstructured covariance. Missing data were treated as missing at random. All analyses were conducted in SAS 9.4 (Cary, NC).

Participants’ average age was 76.8 ± 10.1 years, and the majority were female (57.8%) and self-identified as White (73.0%; Table 1). A mean of 7.3 years of retrospective data with a total of 1,018 cognitive assessments were available for longitudinal analyses.

Table 1.

Participant Characteristics

	Participants (N = 237),
Characteristics	Mean (SD) or n (%)
Demographics
Age at skin measurement (y)	76.8 (10.1)
Range (min–max)	50.8–97.1
Sex
Female	137 (57.8%)
Male	100 (42.2%)
Race/ethnicity
White	173 (73.0%)
Black	50 (21.1%)
Other	14 (5.9%)
Years of education	17.4 (2.3)
Health measures
Body mass index	26.6 (4.6)
Range (min–max)	17.1–43.5
History of eczema, yes (N miss = 8)	18 (7.9%)
Currently smokes, yes	78 (32.9%)
External factors
Room temperature, Fahrenheit	75.5 (2.0)
Range (min–max)	69.6–80.6
Room humidity (SD)	25.8 (8.4)
Range (min–max)	16–50
Skin barrier measurements
Barrier function (area under the curve)	260.4 (99.0)
Range (min–max)	91.9–641.4
Barrier integrity (max strips)	28 (11.0)
Range (min–max)	5–70
Barrier permeability (baseline TEWL)	6.7 (3.1)
Range (min–max)	0.2–19.4
Cognitive measurements
Total assessments per subject	4.3 (1.7)
Range (min–max)	1–11
Follow-up time (y)	7.3 (2.5)
Range (min–max)	0–10.5

Abbreviations: max, maximum; min, minimum; TEWL, transepidermal water loss.

As hypothesized, we found that a higher area under the curve (i.e., worse barrier function) was associated with faster 10-year declines in verbal memory (β = −0.14, P = 0.048; Table 2). Better skin barrier integrity (maximum number of strips) was also associated with less annual decline in verbal memory (β = 0.021, P = 0.013). For all other cognitive domains, similar trends were observed, but associations were not significant. In studies of dementia onset, components of the verbal memory domain are the first to decline and have been linked to early structural changes in the brain (Bigelow et al., 2015; Chetelat and Baron, 2003; Dickerson and Eichenbaum, 2010; Lin et al., 2011).

Table 2.

Fully Adjusted Linear Mixed-Effects Regression Results Showing Associations between Skin Barrier Function and Cognitive Change

	Effect on Annual Cognitive Change	Effect on 5-Year Cognitive Change	Effect on 10-Year Cognitive Change
Cognitive Domain1	β (SE)2	β (SE)2	β (SE)2	P-Value
Verbal memory	−0.014 (0.007)	−0.072 (0.036)	−0.14 (0.071)	0.048
Attention	−0.003 (0.006)	−0.015 (0.030)	−0.030 (0.060)	0.615
Executive function	−0.003 (0.006)	−0.017 (0.031)	−0.034 (0.061)	0.574
Verbal fluency	−0.005 (0.006)	−0.026 (0.028)	−0.053 (0.056)	0.344
Visuospatial ability	−0.010 (0.006)	−0.049 (0.030)	−0.098 (0.061)	0.112

Abbreviations: BMI, body mass index; CDT, Clock Drawing Test; CVLT, California Verbal Learning Test; SE, standard error; TMT-A, Trail Making Test part A; TMT-B, Trail Making Test part B; WAIS-R, Wechsler Adult Intelligence Test – Revised.

¹Verbal memory was assessed using two measures from the CVLT (Immediate Free Recall—sum of five trials—and Long Delay Free Recall). Attention was assessed using the TMT-A and Digit Span Forward subset of the WAIS-R. Executive function was assessed using TMT-B and Digit Span Backward subset of the WAIS-R. Verbal fluency was assessed using category and letter fluency tests. Visuospatial ability was assessed using the Card Rotations Test and two CDTs where participants were asked to draw a clock indicating times of 3:25 and 11:10. Z-scores of individual cognitive measures were calculated using the mean and SD at the time of skin barrier function assessment. The TMT-A and TMT-B (Reitan, 1958) scores were natural log-transformed, z-scored, and then inverted to align directions of trajectories with other measures.

²Models were adjusted for sex, age at measurement, race, education, skin barrier function, room temperature, humidity, season, time of measurement, BMI, history of eczema, and smoking.

In summary, we found that in a population-based cohort with up to 10 years of longitudinal cognitive data, poorer skin barrier function was associated with faster declines in verbal memory performance. Although observed effect sizes were small and associations for other domains of cognitive change were nonsignificant, all measures showed associations in the direction of the hypothesized change.

Our study is limited by the methods for measurement of skin barrier function among older adults. TEWL alone paradoxically improves with age, likely because of progressive skin dehydration with aging. Therefore, we used the area under the curve because it incorporates both TEWL and skin integrity; future studies should explore measures more sensitive to age-associated changes in skin barrier function and repeat them over time. Furthermore, retrospective measures of cognition were analyzed owing to the recent introduction of skin barrier measures. Prospective studies with repeated concurent measurements of skin barrier function and cognition may provide stronger associations. Finally, there may be potential survival bias and competing risks on cognitive decline, and our results may not be broadly generalizable because our participants consisted of predominantly White, well-educated older adults.

Additional research is needed to understand the impact of epidermal barrier function on inflammatory profiles and the relationship with cell senescence, which has been linked to aging and cognitive decline. Because previous studies have shown that application of moisturizers, which improve skin barrier function, decreases inflammatory markers in the blood and may mitigate cognitive decline in small clinical trials (Hu et al., 2017; Ye et al., 2022, 2019), future work should also examine the role of moisturizers in relation to cognition and whether skin barrier restoration could help to reduce inflammation-associated cognitive decline.

AbbreviationS:

BLSA

Baltimore Longitudinal Study of Aging

TEWL

transepidermal water loss

Data availability statement

Data from the BLSA are available on request by proposal submission through the BLSA website (blsa.nih.gov). All requests are reviewed by the BLSA Data Sharing Proposal Review Committee.