ASSOCIATIONS OF BIOAVAILABLE SERUM TESTOSTERONE WITH COGNITIVE FUNCTION IN OLDER MEN: RESULTS FROM THE NATIONAL HEALTH AND NUTRITION EXAMINATION SURVEY

Panagiotis Giannos, MSc, et al 1Department of Life Sciences, Faculty of Natural Sciences, Imperial College London, London, UK. 2Society of Meta-research and Biomedical Innovation, London, UK. 3Department of Musculoskeletal Biology, Institute of Life Course and Medical Sciences, University of Liverpool, Liverpool, UK. 4Department of Geriatrics, Donald W. Reynolds Institute on Aging, Center for Translational Research in Aging and Longevity, etc

 Journals of Gerontology: MEDICAL SCIENCES, 2023, Vol. 78, No. 1

Ronald Peters MD Commentary

Safe bioidentical hormone replacement is a prominent tool for health promotion which is the focus for integrative or functional medical practice.  Memory loss is commn in the older years of life and can be reversed.

Abstract

 Background: Age-associated cognitive decline may be influenced by testosterone status. However, studies evaluating the impact of bioavailable testosterone, the active, free testosterone, on cognitive function are scarce. Our study determined the relationship between calculated bioavailable testosterone and cognitive performance in older men.

 Methods: We used data from the U.S. National Health and Nutrition Examination Survey (NHANES) between 2013 and 2014. This study consisted of 208 men aged ≥60 years. Bioavailable serum testosterone was calculated based on the total serum testosterone, sex hormone–binding globulin, and albumin levels, whereas cognitive performance was assessed through the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) Word List Learning Test (WLLT), Word List Recall Test (WLRT), and Intrusion Word Count Test (WLLT-IC and WLRT-IC), the Animal Fluency Test (AFT), and the Digit Symbol Substitution Test (DSST). Multiple linear regression analyses were performed upon adjustment for age, ethnicity, socioeconomic status, education level, medical history, body mass index, energy, alcohol intake, physical activity levels, and sleep duration.

 Results: A significant positive association between bioavailable testosterone and DSST (β: 0.049, p = .002) score was detected, with no signs of a plateau effect. No significant associations with CERAD WLLT (p = .132), WLRT (p = .643), WLLT-IC (p = .979), and WLRT-IC (p = .387), and AFT (p = .057) were observed.

 Conclusion: Calculated bioavailable testosterone presented a significant positive association with processing speed, sustained attention, and working memory in older men above 60 years of age. Further research is warranted to elucidate the impact of the inevitable age-related decline in testosterone on cognitive function in older men.

Age-related cognitive decline can be a precursor of dementia, which currently remains a major public health challenge (1). An aging brain is characterized by a reduction in gray and white matter volume and is associated with decreased processing, attention, memory, and executive function (2). In the United States, there are ~5.1 million people living with dementia (3), with an estimated financial impact projected at >$9 trillion worldwide by 2050 (4). Currently, there are multiple risk factors associated with cognitive decline throughout the lifespan, including genetic predisposition, malnutrition, physical inactivity, and androgen deficiency (5,6), factors that represent an important focus for future research in an attempt to treat and/or manage a cognitive decline in older age.

Evidence regarding the influence of the primarily male hormone testosterone on cognitive function is conflicting (7–13), while both total and bioavailable levels of testosterone consistently decline with aging (14). Specifically, previous studies have observed associations between low serum testosterone levels and decreased cognitive performance (7,8) or dementia severity (9), whereas other studies have neither observed associations (10,13) nor negative correlations between these variables (11,12). These findings have led to uncertainty as to whether serum testosterone is linked with risk of cognitive decline in men during aging.

One of the challenges in interpreting published literature is that studies variably report total testosterone or bioavailable testosterone. The majority of testosterone within the human body is tightly bound (~60% of total testosterone) to sex hormone–binding globulin (SHBG) and to a lesser extent (~38%) bound to albumin. Only a small fraction (~2%) of total testosterone is unbound or “free” and thus “biologically active” and available at the tissue level (15). SHBG may vary according to a wide variety of factors including nutritional state, weight, androgen levels, intercurrent illness, and age (16,17). The subsequent influence this has on testosterone binding and thus activity makes bioavailable testosterone a more preferable parameter than total testosterone in isolation. Indeed, as total testosterone typically reduces with age, while SHBG typically increases (18), the overall effect of aging will have a greater effect on bioavailable, active testosterone. Therefore, the aim of the current cross-sectional study was to examine the association of bioavailable testosterone with cognitive function in older men.

Method

Study Design and Participants

We retrieved publicly available data from participants aged ≥60 years from one survey cycle in the National Health and Nutrition Examination Survey (NHANES): 2013–2014. NHANES is conducted by the Centers for Disease Control and Prevention (CDC) and the National Center for Health Statistics (NCHS) to monitor health in the U.S. population. A cutoff age of ≥60 years old was used based on data availability for total serum testosterone, SHBG, and albumin concentrations. Exclusion criteria included no recorded data for total serum testosterone, SHBG, and albumin concentrations, and incomplete data for the cognitive assessments or missing demographic data. The NHANES protocol was approved by the NCHS Research Ethics Review Board, while all participants provided written informed consent.

Bioavailable Testosterone Assessment

Bioavailable serum testosterone was computed according to the Vermeulen methodology (19) using measured concentrations of total serum testosterone, SHBG, and albumin. Total serum testosterone from overnight fasted samples was estimated through isotope

dilution–liquid chromatography–tandem mass spectrometry (ID–LC–MS/MS) method. SHBG was reacted with immune antibodies and chemoluminescence measurements of microparticles and measured by a photomultiplier tube. Albumin concentration was assessed using the DcX800 method by means of a biochromatic digital endpoint methodology with Bromcresol Purple.

Cognitive Assessment

Cognitive function was evaluated using a variety of tests including the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) Word List Learning Test (WLLT), Word List Recall Test (WLRT), and Intrusion Word Count Test (WLLT-IC and WLRT-IC),

the Animal Fluency Test (AFT), and the Digit Symbol Substitution Test (DSST). The assessments were administered by trained, qualified personnel at the end of the in-person private interview at the mobile examination centers. The full details of the cognitive function interviews have been presented elsewhere (https://www.cdc.gov/nchs/nhanes/index.htm).

The CERAD WLLT, WLLT-IC, WLRT, and WLRT-IC examine the immediate and delayed learning ability for novel verbal information and consist of 3 progressive learning trials followed by a delayed recall challenge with a range of scores between 0 and 10.

The AFT assesses executive function by evaluating categorical verbal fluency with scores ranging from 3 to 39. The DSST comprises a performance challenge from the Wechsler Adult Intelligence Scale- III, which assays processing speed, sustained attention, and working memory, and is scored between 0 and 105. Higher test scores depict better cognitive performance. Participants without a response for any of the tests were excluded.

Covariates

Age (years), ethnicity (race), socioeconomic status (family income to poverty ratio [FIPR]), education level (school qualification), medical history (memory–cognitive function loss and stroke), body mass index (BMI; kg/m2), daily sleep duration (hours of sleep spent at

night on weekdays or workdays) and physical activity (minutes spent doing moderate-intensity sports, fitness, or recreational activities), daily energy intake (kcal), and alcohol intake (g) were considered as covariates. All covariates were identified as potential confounders in the relationship between bioavailable testosterone and cognitive performance. Participants with a current medical prescription of aromatase inhibitors and glucocorticoids were excluded, considering that they may alter endogenous testosterone levels (20).

Age groups consisted of participants with ≥60 years of age and classified into 60–69, 70–79, and ≥80 years of age. Ethnic groups comprised of Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, non-Hispanic Asian, and other (multi) race. Social economic status was categorized as low–middle (FIPR < 1) and middle–high (FIPR ≥ 1). Education level was defined as no high school degree, at most a high school degree or a college degree at minimum. Medical history based on loss of cognitive-memory function

or stroke was categorized as Yes/No responses in terms of past incidence reported by a doctor or other health professional. BMI was defined as a participant’s weight in kilograms divided by the square of height in meters. A physical activity of <150 min/wk was considered low-moderate and ≥150 min/wk was considered moderate high. Energy and alcohol intake were calculated as averages of the 24-hour dietary recall and categorized into low, moderate, and high. A BMI of <18 kg/m2 was considered low, 18–24.9 kg/m2 moderate and ≥25 kg/m2 high. An energy intake of <2 000 kcal was considered as low, 2 000–3 000 kcal moderate, and >3 000 kcal high. Alcohol intake < 15 g was considered low, 15–30 g moderate, and >30 g high. Sleep duration of ≤6 hours was classified as low, 7–9 hours as moderate, and >9 hours as high.

Statistical Analysis

Multiple linear regression analyses were performed to examine the association between bioavailable and total testosterone and cognitive function (overall and test-specific cognitive performance) with the adjustment of all covariates. A restricted cubic spline was employed to model the nonlinear and dose–response relationship between calculated bioavailable testosterone and cognitive function using 3 knots after covariate adjustments. Statistical significance was established as p < .05. Statistical analysis was performed using IBM SPSS statistics software (Version 28.0, IBM Corp., Armonk, NY).

Results

 Characteristics of Study Participants

Data for cognitive function and calculated bioavailable testosterone were available for a total of 208 participants (Figure 1). Background information (ie, sociodemographic status, anthropometrics, and nutritional characteristics) of all participants are reported in Table 1.

The study population had a mean age of 69.4 (} 0.5) years and consisted of mostly non-Hispanic White (53%) of high socioeconomic status (69%) with a college degree at minimum (67%). Sleep duration was moderate (63%) among participants and physical activity was high (55.3%). Energy intake was below the typical recommended consumption (51%) or within recommendations (42%). Alcohol intake was low (78%), and BMI was high (71%). The mean calculated bioavailable testosterone was 185.7 (} 4.0) ng/dL. The average score for CERAD WLLT was 19.3 (} 0.3) of 30, 6.1 (} 0.2) of 10 for CERAD WLRT, 0.6 (} 0.1) of 12 for CERAD WLLT-IC, 0.4 (} 0.1) of 10 for CERAD WLRT-IC, 18.3 (0.4) of 40 for the AFT, and 49.0 (} 1.0) of 100 for the DSST.

Figure 1. Flow chart of the screening process for the selection of eligible participants in the National Health and Nutrition Examination Survey

(NHANES).

FOR ENTIRE ARTICLE