Abstract
Individuals with high-level spinal cord injury (SCI) experience low blood pressure (BP) and cognitive impairments. Such dysfunction may be mediated in part by impaired neurovascular coupling (NVC) (i.e., cerebral blood flow responses to neurologic demand). Ten individuals with SCI > T6 spinal segment, and 10 age- and sex-matched controls were assessed for beat-by-beat BP, as well as middle and posterior cerebral artery blood flow velocity (MCAv, PCAv) in response to a NVC test. Tests were repeated in SCI after 10 mg midodrine (alpha1-agonist). Verbal fluency was measured before and after midodrine in SCI, and in the control group as an index of cognitive function. At rest, mean BP was lower in SCI (70 ± 10 versus 92 ± 14 mm Hg; P<0.05); however, PCAv conductance was higher (0.56 ± 0.13 versus 0.39 ± 0.15 cm/second/mm Hg; P<0.05). Controls exhibited a 20% increase in PCAv during cognition; however, the response in SCI was completely absent (P<0.01). When BP was increased with midodrine, NVC was improved 70% in SCI, which was reflected by a 13% improved cognitive function (P<0.05). Improvements in BP were related to improved cognitive function in those with SCI (r2 = 0.52; P<0.05). Impaired NVC, secondary to low BP, may partially mediate reduced cognitive function in individuals with high-level SCI.
INTRODUCTION
Spinal cord injury (SCI) is a devastating condition often resulting in disturbed motor, sensory, and autonomic function. As a result of disrupted descending autonomic spinal pathways after SCI, sympathetic vasomotor tone is impaired, often resulting in arterial hypotension. 1 Those with injury at or above T6 spinal segment, because of a myriad of potential mechanisms (including decentralization of sympathetically mediated vasomotor tone in the splanchnic region and legs) have more profound autonomic disturbances including resting hypotension and orthostatic hypotension as compared with those with lower level SCI. 2
An abundance of evidence, from both animal and human studies, demonstrated that there is an association between low blood pressure (BP) and a variety of cognitive impairments.3,4 There is well-established evidence that individuals with SCI have a high rate of traumatic brain injury that can result in a significant number of cognitive dysfunctions. 5 However, the high prevalence of cognitive dysfunction (i.e., between 10% and 60%) in SCI6,7 is likely to be at least partially due to widespread hypotension. 8 This notion is further highlighted by a significant positive relationship between systolic BP (SBP) and cognitive function in the SCI population. 9
It has been speculated that cognitive dysfunction in chronic hypotension is due to reduced blood flow perfusion in the cerebrovasculature. 10 Resting cerebral blood flow (CBF), however, has been shown to be similar at rest in SCI as compared with able-bodied (AB) individuals.11,12 Although resting CBF may be preserved after SCI, cerebrovascular reserve (i.e., the ability of the cerebrovascular system to respond to acute increases in metabolic demand, mechanical or neural stimuli) may be affected, resulting in an insufficient CBF response. 13 Supporting this contention, elegant work by Harper and Glass 14 showed that progressive hypotension abolishes the capacity of the cerebrovasculature to dilate or constrict.
Indeed, in individuals with chronic hypotension, it has been reported that there is a blunted increase (i.e., peak response 40% lower) in middle cerebral artery (MCA) mean velocity during a preparatory phase before cognitive testing (i.e., 5 seconds before reaction time). 15 Interestingly, it was demonstrated in the normotensive controls that increases in MCA flow velocity during the preparatory phase were related to reaction time. 15 This protocol for evaluating the MCA flow velocity response to cognition has never been reported, and the ‘cognitive task’ was that of 20 individual simple-reaction time tests, each divided by 50 seconds; likely resulting in limited cerebral activation as compared with more complex processes, 16 such as those found in everyday living.
Neurovascular coupling (NVC; the relationship between local neural activity and subsequent changes in CBF: i.e., cognitive task) is a well-established protocol examining the metabolic component of cerebrovascular reserve, and is considered one of the three fundamental CBF regulatory measures. 17 Impaired NVC, which is associated with stroke and hypertension, disrupts both the delivery of substrates to activated cerebral tissue and the removal of harmful by-products, and is likely to have an integral role in cognitive dysfunction. 18 As cerebral perfusion pressure is a critical component of CBF control, it is possible that NVC mediates the relationship between reduced BP and cognitive function in SCI. 14 Wecht et al 9 partially examined this relationship by showing the MCA flow velocity response to cognition (i.e., Stroop test) was not different between AB and SCI. Although the internal carotid/MCA provides >75% of total CBF, arteries responsible for perfusing the brainstem (i.e., vertebral/posterior cerebral arteries (PCA)) have been shown to be more responsive to various cognitive stimuli.20–22 Considering the different responsiveness of the posterior cerebral vasculature, a global evaluation of NVC incorporating an assessment of both the MCA and PCA is likely to increase sensitivity for detecting changes in cerebrovascular reserve.
There is a lack of studies examining NVC in either those with low BP or SCI. To further shed light on the mechanism of cognitive dysfunction in low BP secondary to SCI, NVC was examined with and without the administration of midodrine hydrochloride, an alpha-1-agonist, to increase BP to AB level in the SCI group. The primary hypothesis was that those with SCI will have an impaired NVC response and cognitive function as compared with AB. Second, it was hypothesized that increasing BP through the administration of midodrine will improve NVC and cognitive function in SCI.
MATERIALS AND METHODS
Ten individuals (seven male, three female) with motor complete SCI participated in this study (C4-T5, AIS A and B; Table 1). Eight participants were <1 year after injury while 2 were >1 year after injury (Table 1). The control group was composed of 10 age- and sex-matched AB individuals. All testing took place at the regional rehabilitation centre, in Vancouver, Canada. Participants with SCI were approached by a research coordinator after being notified by an attending physician that the patient met the inclusion criteria. Control participants were recruited with posters placed around the University of British Columbia campus. Participant characteristics are presented in Table 2.
Individual demographic information for spinal cord injured participants
AIS grade, American Spinal Injury Association Impairment Scale; DOI, duration of injury; education, years of post secondary.
Selected cardiovascular variables for spinal cord injured and able-bodied participants
AB, able-bodied controls; BMI, body mass index; N/A, not applicable; SCI, high-level spinal cord injury; TBI; traumatic brain injury.
All participants were instructed to abstain from exercise and alcohol for 24 hours before testing. No caffeine was permitted on the day of testing. In addition, participants were requested to abstain from all other medications on the day of testing, and to only have a small meal ˜1 hour before testing. All foods ingested were monitored. The majority of participants had a single juice box or water, whereas two had a small yogurt. Those who were smokers or had a history of CVD were excluded from participation. Having a history of severe traumatic brain injury was also an exclusion criterion. One patient had suffered a minor brain injury secondary to falling from his wheelchair. Follow-up imaging was negative however, and cognitive function was not affected by the injury. After this, we chose to include him in the study. All participants provided written informed consent in accord with the Clinical Research Ethics Board at the University of British Columbia, who approved this study.
Testing took place over 2 days, each separated by at least 48 hours, between 10 am-12 pm. A number of these participants have been tested for a study of another purpose, but none of the primary outcome measures overlap. 23 The two testing days were identical except for the administration of 10 mg midodrine on one of the days. In order to remove the learning effect, the order of days (i.e., whether baseline or midodrine trial went first) was randomized.
Protocol
A visual task was employed to activate the occipital lobe (while measuring posterior cerebral artery blood flow velocity (PCAv)), and a verbal fluency task to preferentially activate the left cerebral cortex (while measuring middle cerebral artery blood flow velocity (MCAv)), the order of which was randomized. Assessing NVC using transcranial Doppler has generally been evaluated by measuring the blood flow velocity response in the cerebral artery upstream to the vascular bed being activated by the given task. As a clinical decision to mitigate the amount of time required for this study, we chose to focus on the artery most likely to be stimulated during the two tasks. Previous studies have documented that MCAv is largely unchanged during visual-spatial tasks and, conversely, is significantly elevated during the verbal frequency task.21,24 This latter test results in unremarkable changes in PCAv. Occipital stimulation was achieved by reading from a magazine. 25 Previous work has shown that light input to the occipital cortex provides identical CBF velocity response as to a reading task, suggesting the light input, and not reading, is responsible for the hyperemic response. 26 Three minutes of baseline data, as well as a mock practice were recorded to ensure hemodynamic homeostasis. After this, 10 cycles occurred, each consisting of 30 seconds eyes closed followed by 30 seconds reading. An auditory stimulus provided notification of ‘reading’ and ‘eyes-closed’ periods. The velocity response was averaged for all 10 trials for each participant.
The verbal fluency task was adapted from a previously described methodology. 24 This test had the combined benefit of having been shown to preferentially activate the left MCA, and be a valid and ubiquitous marker of cognition.24,27 A PowerPoint (Microsoft, Redmond, WA, USA) slide show was used consisting of a series of 10 slides presenting a single letter for 30 seconds followed by the words ‘Eyes Closed’ for 30 seconds. The ten letters were M, H, O, I, W, B, T, F, A, S. The letters were separated into two stages (i.e., self-reported word totals versus verified word totals) separated by two minutes, the order of which was randomized. The self-reported word total stage involved 7 cycles, where the participant recorded the number of words they generated on their right hand (those with limited hand motion were instructed to imagine counting on right hand) in response to letters M–T. Immediately after the eyes-closed period began, the participant was instructed to report how many words they had generated. The verified total words section involved letters F, A, S, and consisted of the participants stating each word orally while a tester recorded each word. The two-stage approach was completed for two reasons: (1) to ensure accurate verbal fluency scores and (2) ensure accurate end-tidal partial pressure of carbon dioxide (PETCO2) during cognition (i.e., accurate PETCO2 would not be possible if participants verbally expressed words while have respiratory mask on as expired flow tracing could be spurious, secondary to talking). The hemodynamic response to the seven cycles of the M–T stage was averaged and used for analysis.
On the day of assessment using midodrine, a 10 mg oral dose was administered. This dose was selected as it has been shown to elicit improvements in BP, with no additional side effects than a 5 mg dose, whereas larger doses can lead to more severe/frequent side effects. 28 The set-up for post-midodrine testing was initiated precisely 75 minutes after midodrine administration to test at the time of peak effect, which based on pharmacokinetics is ˜1 hour after oral intake. Each participant was tested in the seated position after 15 minutes of quiet rest. Similar alpha-1 agonists did not alter intracranial vessel diameter. 29 The blood–brain barrier is thought to prevent intravascular catecholamines from binding to adrenergic receptors located in cerebral arterioles. 30
Data Acquisition
For each participant, brachial BP was measured (BpTRU-BPM-100, Coquitlam; VSM Medical, Vancouver, BC, Canada) on the right brachial artery (before and after both the baseline and post-drug procedure). Measurement of BP occurred three times, with the latter two being averaged. The following were sampled at 1,000 Hz using an analog-to-digital converter (Powerlab/16SP ML 795; ADInstruments, CO Springs, CO, USA) interfaced with data acquisition software on a laptop computer (LabChart 7 ADInstruments): non-invasive beat-by-beat BP measurement via finger photoplethysmography (Finometer PRO, Finapres Medicine Systems, Amsterdam, The Netherlands), electrocardiogram (ML 132; ADInstruments), PETCO2(CO2 Analyzer Gold Edition-17515, Ventura, CA, USA), velocity in the left middle cerebral artery (MCAv) or right posterior cerebral artery (PCAv; Doppler-Box, Compumedics DWL, Singen, Germany). These arteries were insonated using a 2 Mhz probe mounted on the temporal bone and a fitted head strap. As described in depth elsewhere, the P1 segment of the PCA was insonated at depths between 60 to 70 mm, whereas the MCA was insonated from 45 to 55 mm. Arteries were confirmed using ipsilateral common carotid artery compression, ensuring an increase in PCA velocity and decrease in MCA velocity.
Cognitive Assessment
To evaluate cognitive assessment from verbal fluency outcome, the following metrics were calculated: (1) average number of words recorded for letter F, A, S (VF,AVE); (2) total number of words generated to letters F, A, and S (VFFAS) and (3) average number of self-reported words (VFSELf).
Data Analysis
All signals were visually inspected for artifacts or noise and corrected by linear interpolation. The CBFv signals were filtered by a low-pass filter with a cut-off frequency of 10 Hz (LabChart 7). All hemodynamic variables were sampled at a (heart) beat-by-beat basis (as detected by the echocardiogram) while PETCO2 was sampled on a breath-by-breath basis (as detected by the peak of the first derivative of the PETCO2 waveform). All signals were then transferred to Excel (Microsoft) where custom designed cubic spline interpolation allowed for re-sampling at 5 Hz. Systolic and diastolic BPs were then recorded as well as peak MCAv/PCAv and minimum MCAv/PCAv. From this mean arterial pressure as (2*diastolic blood pressure + systolic BP)/3 and mean CBF (CBFvmean) as (2*CBFv minimum + CBFv maximum)/3 were calculated. This also allowed for the calculation of cerebrovascular conductance (CVC; CBFvmean/mean arterial pressure MAP). Cerebral pulsatility index was also calculated according to the measure of Gosling, and normalized cerebral pulsatility index for systemic pulsatility index to generate pulsatility ratio; the latter of which is more directly associated with local cerebrovascular responses. 31 All the above metrics were calculated five times per second, and all averaged trials (i.e., 10 trials for the visual task, and seven for the verbal task) for each participant.
In addition to looking at absolute changes during the activation period, 15 to 25 seconds of the eyes-closed period, was used as baseline values to calculate percent changes. Average and peak values from the 30-second activation period are reported. The activation period was binned into six 5-second long averages for analysis (Figures 1 and 2).
Cerebrovascular response to visual-spatial task. Mean arterial pressure (MAP), posterior cerebral artery (PCA) mean flow velocity (vmean), cerebrovascular conductance (CVC), and partial pressure end-tidal carbon dioxide (PETCO2) response during verbal fluency testing. Gray line represents able-bodied individuals, black dashed line represents spinal cord injury (SCI) before midodrine, thick black line represents spinal cord injury after midodrine (SCImido). The black filled rectangle represents the 30-second activation period, whereas the six smaller empty rectangles represent 5-second bins from the activation period. Mean arterial pressure and PETCO2 did not change significantly in response to the visual task. Posterior cerebral artery mean flow velocity and PCAcvc did not significantly increase in SCI. Posterior cerebral artery mean flow velocity and PCAcvc increased significantly during activation in SCI after midodrine; however, this response was still reduced as compared with the significant response in AB. AB, able-bodied control; CBF, cerebral blood flow.
Cerebrovascular response to verbal fluency task. Mean arterial pressure (MAP), middle cerebral artery (MCA), mean flow velocity, cerebrovascular conductance (CVC), and partial pressure end-tidal carbon dioxide (PETCO2) response during verbal fluency testing. Gray line represents able-bodied individuals, black dashed line represents spinal cord injury (SCI) before midodrine, thick black line represents spinal cord injury after midodrine (SCImido). The black filled rectangle represents the 30-second activation period, whereas the six smaller empty rectangles represent 5-second bins from the activation period. Although MCAv, MCAcvc, and PETCO2 significantly changes during the activation period, there were no significant differences between AB, SCI, or SCImido with respect to their responses. AB, able-bodied controls; CBF, cerebral blood flow.
Statistical Analysis
After confirmation of normal distribution, midodrine-free and midodrine values were compared using paired-sample t-tests while midodrine free and midodrine data were compared with the control group using independent sample t-tests. Bivariate correlations were also performed. A P value less than 0.05 was considered significant. Data are reported as mean ± s.d. To date, no study has examined the reliability of transcranial Doppler-based NVC.
RESULTS
Resting hemodynamic variables are presented in Table 3. Briefly, MAP, systolic BP, and diastolic blood pressure increased with midodrine administration while HR was reduced. Midodrine increased BP in the SCI group (SCImido) up to values found in AB controls (Table 3). Both MCA and PCA mean flow velocity were similar between AB and SCI at baseline.
Seated resting hemodynamic variables for spinal cord injured individuals with and without midodrine and able-bodied participants
AB, able-bodied controls; DBP, diastolic blood pressure; HR, heart rate;
Different from AB (P<0.05);
different from SCI (P<0.05),
different from SCImido (P = 0.06).
Cerebral Blood Flow Velocity Response
Able-bodied versus spinal cord injured—visual task. In AB, both MAP and PETCO2 were unchanged during the visual task (Figure 1). In contrast, in AB, PCA mean flow velocity and PCAcvc were elevated (all P<0.01). In addition, PCA pulsatility ratio decreased (1.70 ± 0.66 versus 1.58 ± 0.56 au; P = 0.006) during the visual task.
In those with SCI without midodrine, there were no changes in MAP or PETCO2 during the visual task. Also, although PCA mean flow velocity did not increase (38.2 ± 8.7 versus 39 ± 8.0 cm/second; P = 0.51) in the SCI, PCA pulsatility ratio was reduced (1.43 ± 0.66 versus 1.35 ± 0.63, P = 0.003 au; Figure 1). No other calculated parameter changed during the visual task in SCI without midodrine.
As compared with AB, those with SCI reported similar average changes in MAP and PETCO2, but reduced PCA mean flow velocity and PCAcvc response (P = 0.009, P = 0.012). Peak changes in PCA mean flow velocity (P = 0.018) and cvc (P = 0.03) were also mitigated in SCI as compared with AB (Figure 1).
Able-bodied versus spinal cord injured-verbal fluency task. Able-bodied individuals had an increase in MCA mean flow velocity in response to cognition (60.5 ± 12 versus 63.4 ± 12.7; P = 0.013), but no change in MAP or MCAcvc. Similar to SCI, PETCO2 and MCA pulsatility ratio were reduced during the activation (34.5 ± 2.7 versus 33.6 ± 3.4 mm Hg; P = 0.016; 1.61 ± 0.41 versus 1.53 ± 0.17 au; P = 0.04). In SCI without midodrine, MCA mean flow velocity increased over the VF task (58.2 ± 12.9 versus 60 ± 12.7 cm/second; P = 0.004) in SCI. Average PETCO2 also decreased during cognition in SCI (33.3 ± 3.5 versus 32.5 ± 3.7 mm Hg; P = 0.016). No change in MAP or MCAcvc occurred in response to cognition (Figure 2), although MCA pulsatility ratio decreased (1.73 ± 0.43 versus 1.67 ± 0.42 au; P = 0.003). No difference between SCI and AB occurred during the verbal fluency (Figure 2).
Effect of Midodrine in Spinal Cord Injection
Visual task. With midodrine in SCI, PCA mean flow velocity and PCAcvc increased during the visual task, which predominately increases neuronal activity in the occipital cortex (36 ± 13 versus 39 ± 15 cm/second; P = 0.001; 0.44 ± 0.16 versus 0.48 ± 0.18 cm/second · mm Hg−1 P = 0.001). PCA pulsatility ratio was also reduced in SCI with midodrine during the visual task (1.83 ± 0.60 versus 1.78 ± 0.59 au; P = 0.036; Figure 1). Also, PETCO2 decreased during the visual task (34.5 ± 2.7 versus 33.8 ± 3.1 mm Hg; P = 0.032).
Changes in PCA mean flow velocity over the activation period were 70% greater with midodrine administration in SCI as compared with without midodrine (P = 0.03), whereas the PCA pulsatility ratio response trended to be reduced (P = 0.088). The peak change in PCAv also trended to be greater with midodrine (P = 0.066). Changes in PETCO2 and MAP in response to visual activity were not different with midodrine as compared with SCI. Posterior cerebral artery mean flow velocity average response over activation period was lower with midodrine as compared with AB in response to visual activation (P = 0.06), and from 10 to 20 seconds of the activation period specifically (Figure 1).
Verbal fluency task. There was no change in the NVC response to verbal fluency in SCI with midodrine (i.e., a task that predominantly increases frontal and temporal lobe neuronal activity), although the peak change in MCA pulsatility ratio tended to occur later with midodrine (9.5 ± 9.3 versus 16.8 ± 9.9 seconds; P = 0.06).
Cognitive Function
Midodrine administration resulted in improved VFAVE (8.4 ± 3.6 versus 9.5 ± 2.9 words; P = 0.04), and VFFAS (25.4 ± 11 versus 28.6 ± 8.8 words; P = 0.045) but not VFSELF (P = 0.465) in the SCI group. Without midodrine, SCI had lower VFAVE (8.4 ± 3.6 versus 11.3 ± 2.1 words; P = 0.023), VFSELF (10.0 ± 3.8 versus 13.0 ± 3.2 words; P = 0.038), and a trend in VFFAS (P = 0.08) as compared with AB. All verbal fluency scores were highly correlated (VFFAS versus VFAVE, r = 0.94; VFAVE versus VFSELF, r = 0.88; VFFAS versus VFSELF, r = 0.81; all P<0.001). Those with the greatest increases in resting BP, as a result of midodrine, reported the largest improvements in cognitive function (Figure 3). Similarly, those who had the largest reduction in resting MCA conductance reported the largest improvements in cognitive function (Figure 3).
The relationship between changes systemic and cerebral parameters and changes in verbal fluency (VF) scores in those with spinal cord injury (SCI). Individuals with larger increases in mean arterial pressure (MAP) also had larger increases in VF. Also, those with greater reductions in resting conductance in the middle cerebral artery (MCAcvc) had larger increases in VF. Symbols represent individual participants.
DISCUSSION
This study examined the NVC response in the middle and posterior cerebral arteries in SCI with and without midodrine, as well as in matched AB control group. The primary findings are as follows: (1) NVC of the PCA (visual test) is severely impaired in SCI as compared with AB, but is improved by increasing BP with midodrine up to AB levels; (2) MCA NVC (verbal fluency test) is similar in SCI to AB; (3) One aspect of cognition, verbal fluency, improves after increasing BP in SCI; (4) those with the greatest increases in BP, and reductions in MCA conductance, had the greatest improvements in verbal fluency.
This is the first study to examine CBFv of the PCA in those with SCI. In a seated position, although PCAv was maintained, PCAcvc was elevated. An increase in CVC indicates increased conductance at the pial level, to maintain CBF in the posterior region. 32 Reductions in blood flow in the posterior region of the brain may cause an interruption of the blood supply to the medulla oblongata, which contains autonomic control centers, and discrete regions responsible for consciousness. 33 An increase in PCAcvc to preferentially maintain perfusion of the cardiovascular and respiratory centers has been highlighted in response to head-up tilt in AB,22,34 and an exaggerated adaptation of this response may explain the widely reported ‘remarkable tolerance to orthostatic hypotension’ in the SCI population. 35
With regard to the PCA NVC response, the blunted effect in SCI is likely secondary to reduced cerebral perfusion pressure. In support of this contention, increasing BP to AB levels resulted in a normalization of PCAcvc and improvement in the PCA NVC response (Figure 1). The NVC response may have a minimum threshold for cerebral perfusion pressure, below which NVC does not occur. A similar minimum perfusion threshold relationship was demonstrated in an early animal study showing the hyperemic response to increased PaCO2 is reduced and eventually absent as BP decreases. 14 More recently, in humans, attenuated cerebrovascular reactivity to hypercapnia during pharmacologically induced hypotension has been shown in two studies using ganglionic blockade 36 and sympathetic blockade. 37 This study extends these findings, showing that the metabolic response (i.e., NVC) is also largely dependent on the extent of perfusion pressure.
In this study, increasing BP in SCI only partially improved the NVC-PCA response; however, suggesting other CBF regulatory mechanisms may be dysfunctional in SCI. Attenuated NVC, even after midodrine, could be due to a variety of factors outside of low BP, including reduced nitric oxide availability, 38 glucose intolerance, 39 and dyslipidemia. 40
This study clearly showed that NVC of the MCA is preserved to AB levels in those with SCI. These findings appear to be in opposition to that reported by Jegede et al (2010) 9 ; however, differences may be explained by a combination of (a) a dissimilar cognitive task (i.e., verbal fluency versus Stroop test), (b) a more stringent procedure (i.e., seven versus three cycles averaged), or (c) shorter activation period averaged (30 versus 45 seconds that we used in our study). It is unclear from the Jegede et al (2010) 9 study if an eyes-closed rest period preceded cerebral activation, or if verbal responses to cognition influenced PETCO2, as the latter was not measured. The current data show that the MCAv response to verbal fluency appears to be unaffected by varying BP in SCI. This study reported similar resting MCA mean flow velocity, suggesting preserved static cerebral autoregulation in SCI (a premise that has been widely supported).11,12,41 The present study also showed similar resting MCAcvc in SCI, suggesting that resting cerebrovascular tone was maintained. Preservation of resting cerebrovascular tone may permit a suitable CBF response to metabolic stimulation. 13 As NVC of the MCA was well maintained at baseline, increasing perfusion pressure through midodrine exerted little effect. When Duschek and colleagues 42 treated 25 chronic hypotensives (90% female) with an ˜63% greater dose of midodrine (i.e., 0.4 mg/kg) than the present study, they reported increased seated MCAv, and a greater MCAv hyperemic response during preparation to cued reaction time. These authors, however, did not compare to a normotensive control group. In light of this, it is unclear if seated MCAv or MCA NVC is lower in normotensive individuals and improved in chronic hypotensives with midodrine, or if the exaggerated effect on MCAv is due to the high dosage employed. Such higher doses resulted in hypertension for 1 to 3 hours and related adverse reactions (e.g., persistent supine hypertension, as well as marked paresthesias and pruritus). 28 The current study provides evidence that resting MCAv and NVC of the MCA is similar in those with SCI as compared with AB, and is not altered by increasing BP.
The hypotensive SCI group in the present study reported significantly reduced cognitive function at baseline, which is consistent with the literature.6,42 This study also found improved cognition (i.e., improved verbal fluency) after increasing BP with 10 mg midodrine. Further, the present data show that hypotensive individuals with SCI who have the largest improvements in BP also have the largest improvements in cognition (Figure 3). Several cross sectional studies have reported reduced cognitive function in hypotensive AB individuals, 42 and one recent study showed that hypotensive SCI had reduced memory and a trend for slowed attention and processing speed as compared with normotensive SCI. 9 This study is the first prospective work to show that acute increases in BP can improve aspects of cognitive function in those with SCI; thereby establishing a direct relationship similar to that shown in AB.
Clinical Implications
The finding that NVC is reduced in those with SCI links cognitive decline to reduced cerebrovascular reserve. Improving BP and cerebrovascular reserve through midodrine administration served to improve orthostatic hypotension, enhance NVC, and consequently improve cognitive function. This study provides direct clinical evidence that 10 mg midodrine administration is an efficacious pharmaceutical therapy for the cerebrovascular dysfunction and reduced cognitive function after SCI. This study highlights the potential for an inverted-U shape with respect to BP and cerebral function where in addition to hypertension, low BP leads to cerebrovascular and cognitive decline. As too high a BP damages the microvasculature, low BP may impair cerebral metabolic status and cerebrovascular responsiveness to cognition.
Limitations
As is the case when using transcranial Doppler for the assessment of CBF, an assumption is made that arterial diameter is maintained constant to accurately reflect changes in flow. Indeed, it has been shown that administration of similar alpha-1 agonists have not led to reductions in intracranial vessel diameter, 29 allowing the present study to fairly assume maintained MCA and PCA diameter with midodrine. However, vessels downstream to the MCA and PCA, the blood–brain barrier usually prevents intravascular catecholamines from binding to adrenergic receptors located in cerebral arterioles. 30 As recently reviewed, evidence showing a close relationship between low resting BP/orthostatic hypotension and cognitive dysfunction is mounting. 42 However, when manipulating BP using pharmacological interventions, the true effect of BP on cognition has the potential to be confounded by unknown mechanisms influencing cognition. Midodrine is a selective alpha1 receptor agonist. In contrast to alpha2, alpha1 does not have a role in cortical activation directly. 43 However, we acknowledge that in the intact animal or human, redundant mechanisms operate to govern the behavior of complex system (i.e., to maintain hemostasis). For example, a number of dilator factors, such as prostanoids, nitric oxide, and histamine can counteract constrictor effects of noradrenaline in the cerebral circulation.44,45 Whether midodrine-induced elevations in BP may lead to secondary additional release of vasoactive substances that could improve cognitive function is unknown. Only four of the 10 individuals who were administered midodrine demonstrated improvements in cognitive performance, suggesting that the effect may be limited to less than half SCI individuals. Alternatively, as BP increases were related to increased cognition, it may be that larger increases in BP are required in some individuals. This was not a masked, placebo-controlled trial. Although subjects were randomized with respect to the order they had NVC evaluated (to mitigate any learning effect of VF tasks), participants were aware of the medications being provided. However, the participants were masked to the purpose of the study.
CONCLUSIONS
This study shows that NVC, specifically that of the PCA, is impaired in those with SCI but can be partially normalized after BP increases with 10 mg of midodrine administration. This dose also resulted in an improvement in cognitive function, which was directly related to improvements in BP. Finally, a relationship between cerebrovascular resistance and cognition has been shown that has not been previously reported. This study highlights the inverted-U shape describing the relationship between BP and cerebral function; where in addition to hypertension, low BP leads to cerebrovascular and cognitive decline. Future studies should examine other CBF regulatory metrics in low BP SCI before and after acutely increasing BP to elucidate whether dysfunction is due to reduced perfusion pressure or cerebrovascular dysfunction. Furthermore, the chronic effects of low BP on cerebral function and sophisticated measures of cognition in those with SCI need to be explored. The cognitive effects of other pressor drugs, such as nitro-L-arginine methyl ester (inhibits nitric oxide synthase), should also be examined. Greater clinical focus on low BP management needs to occur in the SCI population.
Footnotes
The authors declare no conflict of interest.
ACKNOWLEDGMENTS
The authors acknowledge the individuals who participated in this study. The authors acknowledge the statistical consultation of Vanessa Noonan and Gina Zhong.