Impact of age and cyclooxygenase inhibition on the hemodynamic response to acute cognitive challenges

Abstract

Structural and functional changes in the cerebral vasculature occur with advancing age, which may lead to impaired neurovascular coupling (NVC) and cognitive decline. Cyclooxygenase (COX) inhibition abolishes age-related differences in cerebrovascular reactivity, but it is unclear if COX inhibition impacts NVC. The purpose of this study was to examine the influence of aging on NVC before and after COX inhibition. Twenty-three young (age = 25 ± 4 yr) and 21 older (age = 64 ± 5 yr) adults completed two levels of difficulty of the Stroop and n-back tests before and after COX inhibition. Middle cerebral artery blood velocity (MCAv) was measured using transcranial Doppler ultrasound and mean arterial blood pressure (MAP) was measured using a finger cuff. Hemodynamic variables were measured at rest and in response to cognitive challenges. During the Stroop test, older adults demonstrated a greater increase in MCAv (young: 2.2 ± 6.8% vs. older: 5.9 ± 5.8%; P = 0.030) and MAP (young: 2.0 ± 4.9% vs. older: 4.8 ± 4.9%; P = 0.036) compared with young adults. There were no age-related differences during the n-back test. COX inhibition reduced MCAv by 30% in young and 26% in older adults (P < 0.001 for both). During COX inhibition, there were no age-related differences in the percent change in MCAv or MAP in response to the cognitive tests. Our results show that older adults require greater increases in MCAv and MAP during a test of executive function compared with young adults and that any age-related differences in NVC were abolished during COX inhibition. Collectively, this suggests that aging is associated with greater NVC necessary to accomplish a cognitive task.

INTRODUCTION

Cognitive aging manifests as a decrease in brain functional capacity in processing speed, memory, and executive function (1). Age-related structural changes in the brain, which contribute to cognitive aging, include decreases in white matter (2), gray matter (3), and total brain volume (4). Specifically, aging is associated with a decrease in gray matter volume of the prefrontal cortex, an area associated with executive function and cognitive performance (5, 6). Identifying risk factors associated with accelerated cognitive decline during midlife may help prevent or delay the development of clinical cognitive impairment and/or dementia later in life. It is speculated that vascular dysfunction precedes cognitive impairment (7), as hypoperfusion has been linked to brain atrophy and elevated white matter hyperintensities with age (8). Over the past decade, initiatives have been launched to investigate vascular contributions to cognitive impairment and dementia and establish measurable risk factors identifiable in middle-aged adults (9). Arterial stiffness is one such measurable risk factor that increases with advancing age (10, 11). Elevated arterial stiffness is associated with lower cerebral perfusion (12), higher white matter hyperintensity volume (13–15), and a decline in cognitive function (16, 17).

Neurovascular coupling (NVC) is defined as the relationship between neuronal activity and subsequent changes in blood flow, ensuring adequate delivery of oxygen to match metabolic demand (18). However, with advancing age and the development of pathologies such as hypertension or stroke, impairments in NVC may occur (18–20). Although reductions in NVC are multifaceted, structural and functional changes to the vasculature likely play a critical role in the development of cognitive decline (19, 21). Impaired NVC with advancing age suggests that changes in NVC may be an early marker of risk for cognitive decline. NVC can be assessed using acute cognitive challenges that induce a neuronal metabolic stimulus and require augmentation of blood flow. Previous studies evaluating hemodynamics during acute cognitive challenges have shown increases in middle cerebral artery blood velocity (MCAv) (22–25), with no change in cerebral pulsatility index (PI) (22). However, older adults exhibit greater cerebral PI during acute cognitive challenges compared with young adults, suggesting an augmented cerebral PI response with advancing age (25). Although measuring NVC in response to a cognitive challenge may represent the cerebrovascular response of day-to-day tasks, the effects of aging and acute reductions in cerebral blood velocity remain unclear.

Vasodilating prostaglandins have been established as a potential mechanism underlying the increase in cerebral blood velocity in response to hypercapnia (e.g., cerebrovascular reactivity) (26, 27), but the role of prostaglandins in NVC is unknown. Cyclooxygenase (COX) is the enzyme responsible for producing vasodilating prostaglandins such as prostacyclin (PGI₂). Previous research has reported no effect of the common COX inhibitor aspirin on cognitive function (28) or cognitive test scores in older adults (29). However, indomethacin (INDO) is a COX inhibitor that lowers MCAv at rest in humans (30–32). Because aging is associated with a reduction in prostaglandin-mediated vasodilation (33, 34), the combination of older age and COX inhibition may alter NVC or the systemic hemodynamics. Our previous work has shown that COX inhibition with INDO increased central blood pressure and aortic augmentation index in older adults (35). However, the effects of COX inhibition on NVC and the systemic hemodynamic response to a cognitive challenge are unclear. Therefore, the impact of acute INDO administration could be important for understanding COX-mediated mechanisms, as well as the effect of COX inhibitors, which are frequently used to manage pain.

The purpose of this study was to evaluate age-related differences in NVC and the hemodynamic response to acute cognitive challenges before and after administration of a COX inhibitor in young and older adults. We hypothesized that older adults would exhibit an augmented NVC and hemodynamic response to acute cognitive challenges compared with young adults and hypothesized that this difference would be abolished during COX inhibition. Because of the increase in arterial stiffness with age (10, 11) and the association between increased arterial stiffness and a decline in cognitive function (16, 17), a secondary aim was to investigate associations between baseline arterial stiffness and NVC or hemodynamic responses to acute cognitive challenges.

METHODS

Participants

Twenty-three young (18–35 yr) and 21 older (55–75 yr) adults participated in this study. All participants had a body mass index (BMI) between 18 and 29 kg/m². Participants were nonsmoking and excluded if they presented with a history or evidence of cardiovascular disease, stroke, hypertension, diabetes, hepatic disease, renal disease, hematological disease, neurovascular disease, peripheral vascular disease, or other chronic diseases as determined by a health history questionnaire. In addition, participants were excluded if they were currently taking any vasoactive medications or were prescribed nonsteroidal anti-inflammatory drugs (NSAIDs). Young women were studied during the early follicular phase of their menstrual cycle or the placebo pill phase of oral contraceptives (n = 8 young women). Older women were postmenopausal for >1 year and were not currently taking hormone replacement therapy. All study procedures were approved by the University of Wisconsin-Madison Institutional Review Board and performed according to the Declaration of Helsinki by obtaining written informed consent from each participant. This study was registered under ClinicalTrials.gov (NCT02653638).

Experimental Procedures

Participants arrived at the Bruno Balke Biodynamics Laboratory at the University of Wisconsin-Madison following an overnight fast having refrained from consuming caffeine or alcohol and performing strenuous exercise 24 h before the study. In addition, participants were asked to withhold over-the-counter medication on the day of the study and avoid the use of NSAIDs for 5 days leading up to the study visit. Upon arrival to the laboratory, height and weight were obtained using a standard scale and stadiometer. All study procedures were performed in a dim, temperature-controlled room kept between 22°C and 24°C. A timeline of the study protocol is shown in Fig. 1.

Figure 1.

Study day timeline. Mean arterial pressure, heart rate, and middle cerebral artery blood velocity were continuously measured throughout the protocol. Participants completed two levels of difficulty of the Stroop color and word test. Each level consisted of a 2-min baseline period, a 3-min cognitive test, and a 2-min recovery period followed by a brief washout period. Participants then completed two levels of difficulty of the n-back working memory test following the same protocol. Indomethacin was administered followed by a 90-min wash-in period. Then, the cognitive testing protocol was repeated.

Cardiovascular Measurements

Following 10 min of supine rest, arterial blood pressure was measured in triplicate using a brachial blood pressure cuff (CardioCap, Datex Ohmeda, GE Healthcare, Fairfield, CT), and the average of the three measurements was recorded. Heart rate (HR) was measured using a 3-lead electrocardiogram (CardioCap, Datex Ohmeda, GE Healthcare, Fairfield, CT). Beat-to-beat mean arterial blood pressure (MAP) was continuously measured using a noninvasive finger cuff (Finometer, Finapres Medical Systems, Amsterdam, The Netherlands). Central arterial stiffness (cfPWV) was assessed using applanation tonometry of the carotid and femoral sites (Sphygmocor, AtcorMedical, Sydney, NSW, Australia) as described previously (36). Briefly, participants rested for 15 min before three trials of cfPWV recorded by measuring the pulse at the carotid and the femoral artery. The total transit distance was calculated by adding the distance between the carotid artery and suprasternal notch to the distance between the suprasternal notch and femoral artery. cfPWV was calculated using the intersection tangent foot-to-foot algorithm. The average of three measurements of cfPWV is reported.

Cerebrovascular Measurements

For cerebrovascular assessments, blood velocity of the right middle cerebral artery (MCA) was measured using a 2-MHz transcranial Doppler ultrasound (TCD) (Spencer Technologies, Seattle, WA). Optimal MCAv signals were obtained according to techniques originally developed by Aaslid et al. (37) and using established guidelines (38). On an initial screening day, the MCA was imaged in each participant, and the depth, anatomical location, probe angle, and mean flow were recorded. On the experimental study day, the ultrasound probe was secured using an adjustable headband to maintain the position and angle of the probe throughout the protocol. Before each block of the cognitive challenges, optimal signal quality was confirmed. End-tidal CO₂ (ETCO₂) was measured at rest using a nasal cannula (CardioCap, Datex Ohmeda, GE Healthcare, Fairfield, CT).

Cognitive Challenges to Measure NVC

NVC was assessed by measuring the MCAv, MAP, and HR response to two cognitive challenges, the Stroop color and word test (Stroop), a test of executive function (39), and the n-back working memory test (n-back), a test of working memory (40). All participants completed two levels of difficulty of the Stroop test as described previously (41, 42). Following the Stroop test, participants completed two levels of difficulty of the n-back test. A laptop containing custom versions of the Stroop and n-back tests was positioned in front of the participant. The computer screen was adjusted to minimize head movement while performing the tests to minimize signal loss of the MCA. Prior to each cognitive test, the directions were read aloud to the participant by the researcher. All participants were familiarized with the cognitive challenges and allowed to practice a standardized set before measurement to ensure understanding. Baseline cardiovascular and cerebrovascular measurements were recorded for 2 min before the test. Next, each cognitive test was administered for 3 min followed by a 2-min recovery period with a brief washout period between tests. The protocol was repeated for each subsequent test (Fig. 1). Participants were asked to verbally provide answers and were instructed to complete the test to the best of their ability.

Cyclooxygenase Inhibition

Following the completion of the first block of cognitive testing, the COX inhibitor INDO was administered as previously described (32). Briefly, INDO was administered orally at 1.2 mg/kg of body weight with 10 mL of Maalox to reduce gastric irritation. After INDO was administered, participants remained in a semirecumbent position during a 90-min wash-in period. During the drug wash-in period, participants were permitted to read or quietly rest, and physiological variables were constantly monitored by the study team. Participants refrained from consuming anything other than small sips of water and did not engage in strenuous cognitive activity. After 10 min of rest in the supine position, baseline data were collected, and an identical block of cognitive tests was completed following the 90-min wash-in period.

Data Analysis and Statistics

All measurements were recorded using Windaq Data Acquisition Software at 250 Hz (DATAQ Instruments, Akron, OH) and stored offline for analysis. The measurements included: MCAv, cerebral PI, MAP, pulse pressure (PP), HR, and cerebrovascular conductance index (CVCi). Cerebral PI of the MCA was calculated as (systolic MCAv – diastolic MCAv)/mean MCAv. CVCi was calculated as MCAv/MAP. Baseline cardiovascular and cerebrovascular measurements were analyzed during the middle 60 s of the 2-min baseline period. The response to each cognitive test was analyzed using the average value obtained during the first 60 s of the 3-min cognitive test. Raw values for all cardiovascular measurements and the percent change from baseline are reported. The responses to each level of difficulty of both the Stroop and n-back tests were similar and highly correlated within participants (Stroop: average r = 0.67, n-back: average r = 0.67). For analysis, the response to each level of difficulty of the cognitive challenge was averaged such that NVC and hemodynamic responses to level 1 and level 2 of the Stroop or n-back tests are reported. Statistical analyses were completed using RStudio (The R Foundation, Vienna, Austria). Normality was assessed using Shapiro–Wilk tests. In the case of a non-normal distribution, Wilcoxon rank-sum tests and Wilcoxon signed-rank tests were used to evaluate age-related differences between or within groups for primary outcome variables. Participant demographics and baseline cardiovascular and cerebrovascular variables were compared between young and older adults using independent-samples t tests (two tailed). For our primary hypothesis, age-related differences in NVC and hemodynamic responses to cognitive challenges were compared using independent-samples t tests (one tailed). During INDO, age-related differences in NVC and hemodynamic responses to cognitive challenges were compared using independent-samples t tests (one tailed). Within-group differences in all baseline variables and NVC and hemodynamic responses to cognitive challenges before and during INDO were compared using dependent-samples t tests (two tailed). Effect sizes for all comparisons were calculated using Cohen’s d. As a secondary aim, associations between cfPWV obtained at baseline and NVC and hemodynamic responses to each cognitive test were analyzed using Pearson correlation coefficients. Statistical significance was set a priori at P < 0.05.

RESULTS

There were no age-related differences in BMI or HR (Table 1). At rest, older adults had greater diastolic blood pressure and MAP, but values were still in the normotensive range (Table 1). As expected, older adults had higher cfPWV at rest compared with young adults (Table 1). There were no age-related differences in ETCO₂ values measured at rest during the control condition (young: 40 ± 5 mmHg vs. older: 41 ± 4 mmHg; P = 0.284, d = 0.33) and after INDO administration (young: 42 ± 2 mmHg vs. older: 41 ± 3 mmHg; P = 0.089, d = 0.53)

Table 1.

Participant characteristics and hemodynamic variables at rest

	Young Adults	Older Adults	P Value (Effect Size)
n	23	21
Sex, M/F	10/13	11/10
Age, yr	25 ± 4	64 ± 5	<0.001 (8.61)
Weight, kg	68 ± 10	69 ± 13	0.915 (0.03)
Height, cm	171 ± 8	172 ± 8	0.960 (0.02)
Body mass index, kg/m2	23 ± 2	23 ± 3	0.854 (0.06)
Systolic blood pressure, mmHg	119 ± 9	122 ± 9	0.345 (0.29)
Diastolic blood pressure, mmHg	69 ± 7	75 ± 6	0.007 (0.87)
Mean arterial pressure, mmHg	86 ± 7	91 ± 7	0.020 (0.73)
Heart rate, beats/min	55 ± 8	55 ± 7	0.940 (0.02)
cfPWV, cm/s	5.9 ± 0.9	7.8 ± 1.3	<0.001 (1.63)

Values are expressed as means ± SD. cfPWV, carotid-femoral pulse wave velocity. P value in bold indicates a significant difference between young and older adults (independent-samples t test).

Responses to the Stroop Color and Word Test

At baseline, before the start of the Stroop test, there were no differences in MAP, PP, HR, or cerebral PI between groups (Table 2). MCAv and CVCi were significantly lower in older adults compared with young adults (Table 2). In response to the Stroop test, there were no age-related differences in MAP, PP, HR, or cerebral PI values (Table 2). MCAv and CVCi values were significantly lower in older adults compared with young adults in response to the Stroop test (Table 2). Because there were age-related differences in some of the baseline values, when MAP and MCAv were expressed as a percent change from baseline, the changes in MAP (young: 2.0 ± 4.9% vs. older: 4.8 ± 4.9%; P = 0.036, d = 0.56) and MCAv (young: 2.2 ± 6.8% vs. older: 5.9 ± 5.8%; P = 0.030, d = 0.61) in response to the Stroop test were significantly greater in older adults compared with young adults (Fig. 2). However, there were no age-related differences in the percent change in PP (young: 3.5 ± 6.8% vs. older: 2.2 ± 7.2%; P = 0.280, d = 0.17), HR (young: 13 ± 10% vs. older: 14 ± 8.5%; P = 0.301, d = 0.16), CVCi (young: 0.3 ± 6.5% vs. older: 1.4 ± 5.8%; P = 0.276, d = 0.19), or cerebral PI (young: −0.3 ± 8.9% vs. older: −2.0 ± 5.2%; P = 0.215, d = 0.23).

Table 2.

Cerebrovascular and hemodynamic variables during the Stroop color and word test

Stroop Color and Word Test	Young Adults	Older Adults	P Value (Effect Size)
MAP, mmHg
Control
Baseline	90 ± 13	94 ± 11	0.152 (0.27)
Stroop	92 ± 12	98 ± 13	0.051 (0.50)
Indomethacin
Baseline	91 ± 12	96 ± 13	0.121 (0.36)
Stroop	92 ± 13	98 ± 15	0.093 (0.41)
PP, mmHg
Control
Baseline	74 ± 16	73 ± 12	0.445 (0.04)
Stroop	76 ± 18	75 ± 11	0.415 (0.06)
Indomethacin
Baseline	73 ± 14	76 ± 11	0.208 (0.25)
Stroop	74 ± 14	76 ± 13	0.379 (0.09)
HR, beats/min
Control
Baseline	60 ± 7	59 ± 8	0.231 (0.22)
Stroop	68 ± 10	67 ± 10	0.387 (0.09)
Indomethacin
Baseline	55 ± 8	55 ± 8	0.468 (0.02)
Stroop	60 ± 9	62 ± 9	0.269 (0.19)
MCAv, cm/s
Control
Baseline	60 ± 13	50 ± 16	0.017 (0.66)
Stroop	63 ± 14	53 ± 17	0.027 (0.60)
Indomethacin
Baseline	42 ± 8*	37 ± 9*	0.016 (0.67)
Stroop	44 ± 9*	38 ± 9*	0.027 (0.60)
CVCi, cm/s/mmHg
Control
Baseline	0.68 ± 0.17	0.54 ± 0.18	0.007 (0.78)
Stroop	0.70 ± 0.19	0.55 ± 0.18	0.007 (0.77)
Indomethacin
Baseline	0.47 ± 0.11*	0.39 ± 0.12*	0.012 (0.70)
Stroop	0.48 ± 0.12*	0.40 ± 0.12*	0.015 (0.68)
Cerebral PI, AU
Control
Baseline	0.71 ± 0.12	0.76 ± 0.10	0.054 (0.50)
Stroop	0.70 ± 0.12	0.74 ± 0.11	0.126 (0.35)
Indomethacin
Baseline	0.86 ± 0.16*	0.89 ± 0.13*	0.223 (0.23)
Stroop	0.87 ± 0.17*	0.89 ± 0.14*	0.416 (0.08)

Values are expressed as means ± SD. Cerebral PI, cerebral pulsatility index; CVCi, cerebrovascular conductance index; HR, heart rate; MAP, mean arterial blood pressure; MCAv, middle cerebral artery blood velocity; PP, pulse pressure. P value in bold indicates a significant difference between young and older adults at baseline or in response to the Stroop color and word test (independent-samples t test or Wilcoxon rank-sum test). *P < 0.05 significantly different vs. control at baseline or in response to the Stroop color and word test (dependent-samples t test or Wilcoxon signed-rank test).

Figure 2.

Percent change (% change) in cardiovascular and cerebrovascular variables in response to the Stroop color and word test before (control) and after (indomethacin) administration of indomethacin. Values are expressed as means % change ± SD, and individual data points are provided. A: percent change in mean arterial blood pressure (MAP). B: percent change in middle cerebral artery blood velocity (MCAv). P value indicates a significant difference between young (n = 23; 10 M, 13 F) and older adults (n = 21; 11 M, 10 F) (independent-samples t test or Wilcoxon rank-sum test) or within groups across conditions (dependent-samples t test or Wilcoxon signed-rank test) in response to the Stroop color and word test.

Responses to the n-Back Working Memory Test

At baseline, before the start of the n-back test, there were no age-related differences in MAP, PP, or HR (Table 3). MCAv and CVCi were significantly lower, whereas cerebral PI was significantly higher in older adults compared with young adults (Table 3). In response to the n-back test, there were no age-related differences in PP or HR values (Table 3); however, MAP and cerebral PI values were significantly greater and both MCAv and CVCi values were significantly lower in older adults compared with young adults (Table 3). When these values were expressed as a percent change from baseline, the change in HR (young: 14 ± 8.7% vs. older: 9.0 ± 8.1%; P = 0.019, d = 0.65) was greater in young adults compared with older adults in response to the n-back test. In response to the n-back test, there were no age-related differences in the percent change in MAP (young: 3.5 ± 4.4% vs. older: 3.2 ± 5.2%; P = 0.435, d = 0.06), PP (young: 3.8 ± 5.6% vs. older: 2.2 ± 9.5%; P = 0.255, d = 0.20), MCAv (young: 6.2 ± 5.0% vs. older: 7.4 ± 8.0%; P = 0.272, d = 0.21), CVCi (young: 2.9 ± 5.8% vs. older: 4.5 ± 7.1%; P = 0.208, d = 0.24), or cerebral PI (young: −4.0 ± 5.9% vs. older: −4.3 ± 6.1%; P = 0.437, d = 0.06) (Fig. 3).

Table 3.

Cerebrovascular and hemodynamic variables during the n-back working memory test

n-Back Working Memory Test	Young Adults	Older Adults	P Value (Effect Size)
MAP, mmHg
Control
Baseline	91 ± 13	97 ± 13	0.054 (0.49)
n-Back	94 ± 12	100 ± 12	0.046 (0.52)
Indomethacin
Baseline	92 ± 13	98 ± 14	0.066 (0.46)
n-Back	93 ± 13	100 ± 15	0.044 (0.52)
PP, mmHg
Control
Baseline	74 ± 16	77 ± 14	0.289 (0.17)
n-Back	77 ± 16	78 ± 11	0.431 (0.05)
Indomethacin
Baseline	74 ± 14	76 ± 13	0.304 (0.16)
n-Back	75 ± 14	77 ± 13	0.284 (0.17)
HR, beats/min
Control
Baseline	59 ± 7	60 ± 8	0.292 (0.17)
n-Back	67 ± 10	66 ± 11	0.282 (0.18)
Indomethacin
Baseline	56 ± 8	56 ± 8	0.494 (0.01)
n-Back	59 ± 8	60 ± 8	0.453 (0.04)
MCAv, cm/s
Control
Baseline	60 ± 14	51 ± 16	0.022 (0.63)
n-Back	64 ± 14	54 ± 16	0.022 (0.63)
Indomethacin
Baseline	44 ± 9*	38 ± 8*	0.013 (0.70)
n-Back	46 ± 9*	40 ± 9*	0.018 (0.65)
CVCi, cm/s/mmHg
Control
Baseline	0.67 ± 0.19	0.52 ± 0.17	0.004 (0.85)
n-Back	0.69 ± 0.19	0.55 ± 0.17	0.006 (0.80)
Indomethacin
Baseline	0.49 ± 0.12*	0.40 ± 0.12*	0.006 (0.79)
n-Back	0.50 ± 0.12*	0.41 ± 0.12*	0.007 (0.78)
Cerebral PI, AU
Control
Baseline	0.71 ± 0.12	0.78 ± 0.12	0.030 (0.58)
n-Back	0.67 ± 0.10	0.74 ± 0.10	0.016 (0.67)
Indomethacin
Baseline	0.85 ± 0.15*	0.90 ± 0.13*	0.147 (0.32)
n-Back	0.84 ± 0.16*	0.88 ± 0.13*	0.202 (0.26)

Values are expressed as means ± SD. Cerebral PI, cerebral pulsatility index; CVCi, cerebrovascular conductance index; HR, heart rate; MAP, mean arterial blood pressure; MCAv, middle cerebral artery blood velocity; PP, pulse pressure. P value in bold indicates a significant difference between young and older adults at baseline or in response to the n-back working memory test (independent-samples t test or Wilcoxon rank-sum test). *P < 0.05 significantly different vs. control at baseline or in response to the n-back working memory test (dependent-samples t test or Wilcoxon signed-rank test).

Figure 3.

Percent change (% change) in cardiovascular and cerebrovascular variables in response to the n-back working memory test before (control) and after (indomethacin) administration of indomethacin. Values are expressed as means % change ± SD, and individual data points are provided. A: percent change in mean arterial blood pressure (MAP). B: percent change in middle cerebral artery blood velocity (MCAv). P value indicates a significant difference between young (n = 23; 10 M, 13 F) and older adults (n = 21; 11 M, 10 F) (independent-samples t test or Wilcoxon rank-sum test) or within groups across conditions (dependent-samples t test or Wilcoxon signed-rank test) in response to the n-back working memory test.

Effects of Indomethacin on Baseline Hemodynamics

The effects of INDO administration are shown in Table 2. In young adults, administration of INDO had no effect on MAP (P = 0.682, d = 0.08) or PP (P = 0.533, d = 0.09) but resulted in a decrease in HR (P < 0.001, d = 0.63), MCAv (P < 0.001, d = 1.6), and CVCi (P < 0.001, d = 1.4) compared with control (Table 2). In addition, INDO resulted in an increase in cerebral PI in young adults compared with control (P < 0.001, d = 1.1; Table 2). Similarly, in older adults, administration of INDO had no effect on MAP (P = 0.326, d = 0.17) or PP (P = 0.332, d = 0.20) but resulted in a decrease in HR (P = 0.021, d = 0.43), MCAv (P < 0.001, d = 1.1), and CVCi (P < 0.001, d = 1.0) compared with control (Table 2). INDO resulted in an increase in cerebral PI in older adults compared with control (P < 0.001, d = 1.1; Table 2).

Effects of Indomethacin on the Response to Cognitive Testing

During INDO, before the start of the Stroop test, there were no age-related differences in baseline MAP, PP, HR, or cerebral PI (Table 2); however, MCAv and CVCi were significantly lower in older adults compared with young adults (Table 2). In response to the Stroop test during INDO, there were no age-related differences in MAP, PP, HR, or cerebral PI values (Table 2); however, MCAv and CVCi values were significantly lower in older adults compared with young adults (Table 2). There were no age-related differences in the percent change in MAP (young: 0.3 ± 3.6% vs. older: 2.1 ± 6.4%; P = 0.124, d = 0.38), HR (young: 9.4 ± 11% vs. older: 12 ± 13%; P = 0.207, d = 0.25), MCAv (young: 2.0 ± 6.4% vs. older: 4.4 ± 5.4%; P = 0.095, d = 0.42), CVCi (young: 2.1 ± 6.6% vs. older: 2.8 ± 5.9%; P = 0.371, d = 0.12), or cerebral PI (young: 2.1 ± 1.3% vs. older: −0.6 ± 1.0%; P = 0.068, d = 0.44) in response to the Stroop test (Fig. 2). During INDO, young adults had a greater percent change in PP (young: 4.0 ± 6.4% vs. older: −0.1 ± 7.2%; P = 0.027, d = 0.60) in response to the Stroop test compared with older adults. During INDO, young adults had a similar percent change in MAP (control: 2.0 ± 4.9% vs. INDO: 0.3 ± 3.6%; P = 0.069, d = 0.46) and MCAv (control: 2.2 ± 6.8% vs. INDO: 2.0 ± 6.41.3%; P = 0.465, d = 0.02) in response to the Stoop test compared with control (Fig. 2). Older adults had a significantly lower percent change in MAP (control: 4.8 ± 4.9% vs. INDO: 2.1 ± 6.4%; P = 0.020, d = 0.48) and a similar percent change in MCAv (control: 5.9 ± 5.8% vs. INDO: 4.4 ± 5.4%; P = 0.173, d = 0.28) in response to the Stroop test compared with control (Fig. 2).

During INDO, before the start of the n-back test, there were no age-related differences in baseline MAP, PP, HR, or cerebral PI (Table 3); however, MCAv and CVCi were significantly lower in older adults compared with young adults (Table 3). In response to the n-back test during INDO, there were no age-related differences in PP, HR, or cerebral PI values (Table 3); however, MAP was significantly greater, whereas both MCAv and CVCi values remained significantly lower in older adults compared with young adults (Table 3). There were no age-related differences in the percent change in MAP (young: 1.0 ± 2.8% vs. older: 1.7 ± 4.1%; P = 0.263, d = 0.25), PP (young: 1.6 ± 4.7% vs. older: 1.8 ± 6.2%; P = 0.463, d = 0.04), HR (young: 7.2 ± 7.9% vs. older: 7.7 ± 9.0%; P = 0.416, d = 0.06), MCAv (young: 3.6 ± 3.8% vs. older: 4.5 ± 4.4%; P = 0.231, d = 0.22), CVCi (young: 2.6 ± 4.6% vs. older: 2.8 ± 4.2%; P = 0.440, d = 0.02), or cerebral PI (young: −0.7 ± 4.3% vs. older: −1.8 ± 3.6%; P = 0.191, d = 0.28) (Fig. 3). During INDO, young adults had a significantly lower percent change in MAP (control: 3.5 ± 4.4% vs. INDO: 1.0 ± 2.8%; P = 0.009, d = 0.67) and MCAv (control: 6.2 ± 5.0% vs. INDO: 3.6 ± 3.8%; P = 0.013, d = 0.59) in response to the n-back test compared with control (Fig. 3). There were no differences in the percent change in MAP (control: 3.2 ± 5.2% vs. INDO: 1.7 ± 4.1%; P = 0.101, d = 0.27) or MCAv (control: 7.4 ± 8.0% vs. INDO: 4.5 ± 4.4%; P = 0.070, d = 0.48) in response to the n-back test during INDO in older adults compared with control (Fig. 3).

Associations between Arterial Stiffness and NVC or Hemodynamic Responses to Cognitive Testing

There was no association between cfPWV and the percent change in MAP in response to the Stroop test in young (Fig. 4A) or older adults (Fig. 4B). In addition, there was no association between cfPWV and the percent change in MAP in response to the n-back test in young (r = 0.33, P = 0.120) or older (r = 0.04, P = 0.856) adults. There were no associations between cfPWV and the percent change in HR or MCAv in response to the Stroop or n-back test in young or older adults. During INDO, there was a significant association between baseline cfPWV (which was measured during control condition) and the percent change in MAP in response to the Stroop test in young (Fig. 4C) and older adults (Fig. 4D). During INDO, there was no association between cfPWV and the percent change in MAP in response to the n-back test in young (r = 0.17, P = 0.451) or older (r = 0.06, P = 0.806) adults. There were no associations between cfPWV and the percent change in HR or MCAv in response to the Stroop or n-back test in young or older adults during INDO.

Figure 4.

Associations between percent change in mean arterial blood pressure (% change MAP) in response to the Stroop color and word test and carotid-femoral pulse wave velocity (cfPWV) at rest. A: young adults (n = 23; 10 M, 13 F) (control); B: older adults (n = 21; 11 M, 10 F) (control); C: young adults (n = 23; 10 M, 13 F) (indomethacin); D: older adults (n = 21; 11 M, 10 F) (indomethacin). P value indicates a significant association between cfPWV and the MAP response to the Stroop color and word test (Pearson correlation).

DISCUSSION

The primary aim of this study was to evaluate age-related differences in NVC and the hemodynamic response to acute cognitive challenges before and after administration of a COX inhibitor. The main findings of the present study are that older adults have an augmented NVC and hemodynamic response to the Stroop test compared with young adults; however, no age-related differences were observed in response to the n-back test. These results indicate that older adults have an augmented NVC and hemodynamic response to a test of executive function compared with young adults. COX inhibition reduced cerebral blood velocity by 30% in young and 26% in older adults; however, there were no age-related differences in NVC and hemodynamic responses to the Stroop or n-back tests during INDO. Taken together, the results of the present study suggest that older adults require a greater increase in MAP and MCAv to accomplish a cognitive test assessing executive function.

Previous studies have reported NVC and hemodynamic changes in response to cognitive challenges in young and older adults (22, 24, 25). The results of the present study build upon existing literature in this area by investigating, for the first time to our knowledge, age-related differences in NVC and hemodynamics before and after COX inhibition. In contrast to work conducted by Heffernan et al. (25), we observed age-related differences in MCAv values at rest and in response to both cognitive challenges, which may be attributed to the young adults being 10 yr younger (25 ± 4 yr in our study vs. 35 ± 10 yr in the study by Heffernan et al.). Furthermore, we report no age-related differences in cerebral PI in response to the Stroop test, whereas Heffernan et al. (25) reported greater cerebral PI in response to the Stroop test in older adults compared with young adults. However, in the present study, older adults demonstrated a greater cerebral PI response to the n-back test compared with young adults, suggesting that the modality of the cognitive challenge may impact the results. Specifically, older adults demonstrated a larger increase in MAP and MCAv only in response to the Stroop test compared with young adults during the control condition.

With advancing age, differential activation patterns of brain regions have been observed during the Stroop (43), n-back (44, 45), and other tests of working memory (46), possibly due to a change in strategy to accomplish the task (47). Furthermore, cognitive performance on such tasks declines with age (44, 45, 48, 49). Together, these age-related differences suggest that older adults activate different areas of the brain, and possibly to a greater degree, to accomplish the same cognitive challenge compared with young adults. Indeed, older adults demonstrate a greater percent change in blood flow in response to the Stroop test measured using functional magnetic resonance imaging (fMRI) than young adults (43). In addition to NVC and hemodynamic differences observed in this study, the type of cognitive activity presented by each test stimulates different regions of the brain as the Stroop test is a measure of executive function, whereas the n-back test is a measure of working memory. Previous works evaluating changes in regional cerebral blood flow during the Stroop test using positron emission tomography have shown activation of anterior regions of the brain (50, 51), such as the anterior cingulate cortex (52). In addition, compared with young adults, older adults activate different areas of the brain during the Stroop test as measured by fMRI, suggesting a shift in strategy to accomplish the task (43). Completion of the n-back test results in activation of several areas of the brain including the premotor cortices, prefrontal cortices, and posterior parietal cortices (40). Similarly, age-related differences in brain activation patterns have been observed during the n-back and other tests of working memory (44–46, 53–55). Specifically, older adults demonstrate an inability to augment prefrontal cortex activity as the difficulty of the task increases when compared with young adults (44, 54). These differences may be explained by diminished functional coupling between regions of the brain (53) and problems with sustained attention during progressive increases in difficulty of the n-back test (56). Taken together, the shift in processing strategy and differential activation patterns of areas of the brain with advancing age may explain the differences in MAP and MCAv responses observed between cognitive challenges in this study.

The use of NSAIDs and COX inhibitors has increased over time (57), and the existing literature on the relationship between COX inhibition and cardiovascular risk is conflicting (58). Previous studies suggest that some COX inhibitors increase cardiovascular risk (59), whereas others suggest no elevated cardiovascular risk (60, 61). Aspirin use is not associated with changes in cognitive function (28, 29); however, given that a previous study has shown that INDO administration resulted in an impairment in cognitive performance (62), it is important to consider the acute effects on NVC. INDO inhibits the COX enzyme, which blocks the production of PGI₂. PGI₂ is a powerful vasodilator that interacts with circulating platelets throughout the vasculature; however, the vasodilatory effects of PGI₂ are attenuated in normal aging (34). Previous work from our laboratory has shown that age-related differences in MCAv and cerebrovascular reactivity to hypercapnia were abolished with INDO, suggesting that the role of vasodilating prostaglandins to augment cerebral blood flow decreases with age (32). The present study extends the existing literature on the effects of COX inhibition by investigating age-related differences in the response to cognitive challenges. During the control condition, we observed age-related differences in NVC and the hemodynamic response to the Stroop test, but no differences in response to the n-back test. Consistent with our previous work, COX inhibition decreased resting MCAv and CVCi in both young and older adults by 26%–30% and 28%–31%, respectively (32). Here, we show that COX inhibition abolishes age-related differences in NVC and the hemodynamic response to the Stroop test. When comparing young and older adults, COX inhibition did not significantly impact NVC or the hemodynamic response to the n-back test. Yet, in young adults, the percent change in MAP and MCAv in response to the n-back test was lower during INDO compared with the control condition. This suggests that young adults may rely more on vasodilating prostaglandins compared with older adults to complete a test of working memory. Taken together, COX inhibition abolished age-related differences in NVC and the hemodynamic response to a test of executive function. These findings add to our previous research demonstrating that age-related differences in cerebrovascular reactivity are abolished following COX inhibition (32).

We also sought to investigate associations between arterial stiffness and NVC or hemodynamic responses to cognitive challenges before and during COX inhibition. There were no significant associations between cfPWV and NVC or hemodynamic responses to either cognitive challenge during the control condition. Yet, there was a positive association between baseline cfPWV and the MAP response to the Stroop test during COX inhibition in both young and older adults. This positive association was not apparent in response to the n-back test in either group during COX inhibition. These data could suggest that individuals with higher arterial stiffness may rely more on increases in MAP to augment cerebral blood velocity during the Stroop test. It is important to note that these associations between arterial stiffness and the MAP response to the Stroop test were only observed during COX inhibition.

Although we did not observe age-related differences in resting MAP during COX inhibition, we have previously shown that COX inhibition augmented central blood pressure and increased aortic augmentation index in older adults (35). The associations between cfPWV and the MAP response to the Stroop test during COX inhibition could be a compensatory response to the effect of INDO on MCAv. Increases in MAP may be necessary to elevate cerebral blood flow during neuronal activation. However, the ability to augment cerebral blood flow via vasodilation of the microvessels may be influenced by arterial stiffness and other vasodilators affecting the endothelium including nitric oxide (NO) and prostaglandins (63). During COX inhibition, MCAv is lower, yet the percent increase in MCAv in response to the Stroop test was similar to the control condition, suggesting a different mechanism to ensure adequate NVC during INDO. The finding that age-related differences in the MAP response to the Stroop test were abolished following COX inhibition may partially explain the positive correlation between cfPWV and the MAP response to the Stroop test during INDO. This observation suggests that adults with higher arterial stiffness augment MAP to a greater degree to ensure adequate NVC during a test of executive function when prostaglandin production is inhibited. Furthermore, as arterial stiffness increases with age, greater reliance on blood pressure to maintain homeostatic cerebral blood flow may be one possible mechanism explaining the progressive increase in blood pressure with age (64).

Methodological Limitations

There were several strengths of the study. First, evaluating NVC and changes in hemodynamics using the average response of two levels of difficulty of two distinct cognitive tests. Although the two different cognitive challenges evoke differential activation patterns in the brain, the average NVC and hemodynamic response were highly correlated, which indicates that these responses are not driven by outliers. Second, beat-by-beat measurement of both MCAv and MAP to evaluate changes during cognitive challenges may provide a more accurate measurement compared with measurements taken periodically during the test (e.g., every 30 s, 1 min, or sporadically during the cognitive challenge), which previous studies have used. Third, to our knowledge, this is the first study to evaluate the effects of COX inhibition on NVC and hemodynamics in response to acute cognitive tests, which has implications for aging humans.

There were methodological limitations of the study, which include the use of TCD to measure MCAv, as this assumes that the diameter of the MCA is relatively constant as long as the insonation angle and depth were consistent. Previous studies evaluating changes in MCA diameter using MRI have revealed conflicting evidence suggesting the MCA diameter may change during a vasoactive stimulus (65–68). To our knowledge, there have been no studies evaluating changes in MCA diameter during an acute cognitive challenge. Furthermore, previous studies evaluating the hemodynamic response to cognitive challenges have used TCD to measure changes in MCAv (22, 24, 25, 69, 70). COX inhibition may affect MCA diameter, as INDO has been shown to decrease cerebral blood flow and relax pial and cerebral arteries in animal models (71, 72). In humans, INDO administration decreases MCAv and increases cerebral PI (30), which may suggest a vasoconstrictor response. Another limitation is that we did not score the cognitive challenges, although participants were not informed of this until after the completion of the study. Participants were encouraged to complete the cognitive challenge to the best of their ability. As the purpose of this study was to evaluate changes in hemodynamics in response to acute cognitive challenges, evaluating cognitive performance was not included because of the inability to accurately quantify cognitive function and the validity of these test scores while participants are fully instrumented lying in the supine position. Moreover, participants were instructed to provide answers to the cognitive challenges verbally and quietly to minimize jaw displacement. In addition, the quality of the signal was closely monitored, and the Doppler probe was secured with a headband to minimize changes in depth, angle, and position of the probe. Due to the verbal responses, accurate ETCO₂ could not be measured during cognitive tests; therefore, only ETCO₂ at rest is included.

Perspectives and Significance

Older adults demonstrated an augmented NVC and hemodynamic response to a test of executive function compared with young adults. COX inhibition decreased resting MCAv and CVCi in both young and older adults. However, during INDO, there were no age-related differences in NVC and hemodynamic responses to a test of executive function or working memory. There was a positive association between baseline cfPWV and the magnitude of the pressor response to the Stroop test in young and older adults during INDO. Collectively, these results suggest that older adults may need to augment MCAv and MAP to a greater degree when engaging in a test of executive function. In addition, individuals with higher central arterial stiffness, regardless of age, may rely more on augmenting blood pressure to ensure adequate cerebral perfusion during COX inhibition, which reduces cerebral blood velocity by ∼26%–30%. Understanding the effects of acute cognitive challenges on the vasculature and subsequent changes in NVC and hemodynamics may be important contributors to cognitive decline with advancing age. Further investigation into age-related differences in NVC and cerebral blood flow regulation and the role of COX during physiological stressors is warranted to understand the potential long-term effects on cerebrovascular health.

GRANTS

This study was funded by the National Institutes of Health Grants HL118154 (to J.N.B.) and HL007936 (to K.B.M.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.N.B. conceived and designed research; A.G.P., K.B.M., A.T.C., N.A.E., A.J.H., A.E.C., M.W.E., and J.N.B. performed experiments; A.G.P. analyzed data; A.G.P. and J.N.B. interpreted results of experiments; A.G.P. prepared figures; A.G.P. and J.N.B. drafted manuscript; A.G.P., K.B.M., A.T.C., N.A.E., A.J.H., A.E.C., M.W.E., and J.N.B. edited and revised manuscript; A.G.P., K.B.M., A.T.C., N.A.E., A.J.H., A.E.C., M.W.E., and J.N.B. approved final version of manuscript.