Measuring cerebrovascular reactivity in a hemodialysis cohort

Highlights

• •Re-breathing can be used to induce hypercapnia as stimuli for measuring cerebrovascular reactivity.
• •Lower cerebrovascular reactivity is seen in patients with prior stroke and older age.
• •Transcranial doppler ultrasound may be a useful measure for patients at risk for hypoperfusion due to circulatory stress.

Abstract

Introduction Cerebrovascular reactivity (CVR) can inform about cerebral vascular health and provide prognostic information on risk cerebral ischemic injury. Use of transcranial Doppler ultrasound (TCD) is a non-invasive and inexpensive method to measure CVR and often uses a stimulus of increase in arterial partial pressure of carbon dioxide (pCO₂). We evaluate re-breathing and breath-hold to measure CVR in a medically complex hemodialysis cohort who are at risk for cerebral hypoperfusion due to circulatory stress of hemodialysis.

Methods CVR was measured using both a 30 s breath-hold and a re-breathing period. We used percent change in mean flow velocity of the middle cerebral artery, measured with TCD over the change in end-tidal CO₂ to calculate CVR. Paired T-test was used to compare the parameters of CVR and Pearson correlation to evaluate relevant risk factors for lower CVR.

Results 16 participants completed both CVR measurements, with mean (SD) age of 64.2 (11.2) years. CVR measured from each technique was similar 3.4 (2.9) %/mmHg (breath-hold) vs 2.7 (1.6) %/mmHg, (re-breathing) p = 0.37. Older age and history of stroke were associated with lower CVR when measured with re-breathing but not with breath-hold technique.

Conclusions Re-breathing to increase pCO₂ and measure CVR is well-tolerated by a frail older medically complex patient population and may be a way to measure cerebrovascular health.

1. Introduction

Cerebrovascular reactivity (CVR) is the change in cerebral blood flow that accompanies a vasoactive stimulus and assesses the regulatory capacity of the cerebral vasculature as an indicator of cerebrovascular health. Impaired CVR is associated with important outcomes including mortality [1], stroke [2], and cognitive impairment. [3,4] In hemodialysis patients, impaired CVR is a risk factor for cerebral hypoperfusion during hemodialysis; even more so when accounting for the circulatory stress from the characteristic rapid fluid removal and drops in systemic blood pressure that occur during hemodialysis. [5,6] Measurement of CVR can provide information on pathophysiology, disease outcomes, and prognostic information. However, common methods for measuring CVR can be difficult in frail and medically ill patient population such as those with end-stage kidney disease (ESKD) on hemodialysis.

Methods to measure CVR usually involve use of MRI (with most common MRI technique being blood oxygen level dependent MRI) or transcranial Doppler. MRI can provide detailed spatial information on CVR. However, MRI use is restricted for some patients due to implants, and some patient populations including those with significant claustrophobia, heart or lung disease, and ESKD may not be able to tolerate an MRI scan due to inability to lie flat. In our experience using BOLD MRI to measure CVR, 8 out of 18 (44 %) patients with ESKD were unable to complete the MRI due to MRI screen failure or claustrophobia. [7] Thus, a non-MRI based method to measure CVR may be more suitable in certain patient populations.

Transcranial Doppler (TCD) provides an inexpensive, portable, non-invasive method to measure changes in cerebral vascular blood flow velocity with good tolerability. In this communication we describe a method to measure CVR that is tolerable to our frail, medically ill patient population (ESKD on hemodialysis). When using TCD, breath-hold index has often been used to measure CVR given its convenience compared to inspired CO₂, [8] but this index lacks details on the actual change in pCO₂ that is occurring and can vary by individual. We sought to compare re-breathing and breath-hold as stimuli while measuring change in etCO₂ to assess CVR. We hope that our experience can help others who want to measure cerebrovascular health in medically complex and ill patient populations as part of clinical research.

2. Methods

We recruited participants with ESKD treated with HD from four Milwaukee, Wisconsin area community dialysis units. Each participant provided informed written consent to the protocol. Inclusion criteria were age ≥40 years and receiving thrice weekly conventional in-center HD or peritoneal dialysis. Exclusion criteria were non-decisional, cerebral vascular malformations (in the medical record or noted during TCD exam), and use of supplemental oxygen. We collected patient demographics and comorbidity data.

2.1. Cerebrovascular reactivity measurement

We used the Multi-dop T TCD from DWL USA to insonate both left and right middle cerebral arteries through the transtemporal window using 2 MHz probes that were attached to a headframe. The mean flow velocity was recorded continuously in cm/s at a sample rate of 100 Hz and recorded using QL software from Compumedics Germany. Both right and left MCAs were insonated but only one side was used in analysis and the side was chosen on viewing the spectral data with the highest density at the outer most part of the waveform. Simultaneously, we recorded end-tidal CO₂ (etCO₂) using Nonin RespSense Capnograph with sampling done using nasal cannula. Participants were asked to lie on a hospital bed during the measurements, with head of bed at approximately 30°. Participants would breathe normally through their nose for one minute, at the end of one minute we would place a face mask with an attached bag and have them continue to breathe into the facemask, increasing their etCO₂ which we measured with the capnograph. The etCO₂ is a surrogate for pCO₂. We continued re-breathing until a 10 mmHg increase in etCO₂ or a maximum of two minutes. See Fig. 1 for capnograph of re-breathing. The participants also completed a breath-hold as an alternative to re-breathing. They would breathe normally through their nose for one minute, noted as normocapnia. At the end of one minute, we have them pinch their nose and hold their breath for 30 s. At the end of 30 second they were instructed to exhale through their nose and then resume normal breathing. They were also instructed to breathe out even if they could not complete the 30 s so that we would still capture the increase in etCO₂ for their breath hold. See Fig. 1 for capnograph of a breath-hold and capnograph of re-breathing.

Fig. 1.

Capnography of breath hold vs re-breathing strategy

In the breath-hold capnography (Panel A) the etCO2 gets to essentially 0 during the breath-hold as no CO2 should be measured. With re-breathing capnography (Panel B) there is a step wise increase in measured etCO2 during the re-breathing period.

We calculated the baseline MFV as the average of the MFV values during the 1 min of normocapnia (60 values) and the peak MFV as the average of the 3 MFV values at peak etCO2 (either during breath-hold or re-breathing). We used the average of the 3 values to avoid an isolated inaccurate reading but kept it to 3 values to avoid incorporating data that would not be at peak etCO2. The CVR was calculated as the precent change in mean flow velocity (MFV) from baseline to peak divided by the absolute change in etCO₂ from normocapnia to the peak etCO₂ [8].

$$\frac{\frac{MCAvpeak-MCAvbaseline}{MCAvbaseline}*100}{etCO2peak-etCO2normovapnia}=CVRin\%/mmHg$$

We used paired-sample T-test to compare CVR parameters with re-breathing vs breath-hold. We used Pearson correlation between the outcome variables (CVR) and the continuous variables of interest including age and point biserial correlation when the predictor was categorial (diagnosis of diabetes, hypertension, and stroke). Analyses were completed using R version 4.3.3. A p-value < 0.05 was used for statistical significance.

3. Results

There were 16 participants who completed both the re-breathing and breath-hold CVR measurement techniques. The average age was 64.2 (11.2) years, and 93 % were male, see Table 1. There was outlier CVR measurement identified when for a CVR measured using the breath-hold technique (z-score of 2.5). Excluding this outlier there is a significant correlation between re-breathing and breath-hold CVR (R = 0.56, p = 0.03). The change in MFV were similar between the two techniques with a higher change in etCO₂ with re-breathing of 11.3 mmHg vs 9.2 mmHg, p = 0.01, see Table 2. We found that older age was associated with lower CVR, R = −0.68, p < 0.01 for the re-breathing technique but not the breath-hold technique. History of stroke trended with a lower CVR, R = −0.45, p = 0.08 for re-breathing, but no association with breath-hold CVR, see Table 3.

Table 1.

Summary of participant demographics (N = 16).

Characteristics	Mean (SD) or N (%)
Age (years)	64.2 (11.2)
Male	15 (93.8)
BMI	27.4 (5.4)
Primary ESKD Cause (%) * note some people had two primary reasons for ESKD
Diabetes	6 (37.5)
Hypertension	7 (43.8)
Glomerulonephritis	1 (6.3)
Obstruction	1 (6.3)
Hereditary	1 (6.3)
Unknown	2 (12.5)
Comorbidities
Hypertension	14 (87.5)
Smoker	8 (50.0)
Diabetes	7 (43.8)
Congestive heart failure	5 (31.3)
Combined vascular disease	9 (56.4)
Stroke	6 (37.5)
Coronary artery disease	7 (43.8)
Peripheral vascular disease	1 (6.3)

Combined vascular disease: participant had at least one or more comorbidities of stroke, coronary artery disease, peripheral vascular disease.

Table 2.

Cerebrovascular reactivity measures comparing breath-hold and re-breathing technique.

Variable	Breath-hold technique	Re-breathing technique	p-value comparing breath-hold to re-breathing	t-statistic for paired t-test	Cohen's d
CVR (%/mmHg)	3.4 (2.9)	2.7 (1.6)	0.30	1.09	0.32
etCO2 (mmHg)	39.8 (4.4)	40.3 (4.2)	0.44	0.78	0.11
ΔetCO2 (mmHg)	9.2 (3.5)	11.3 (3.3)	0.01	3.07	0.63
MFV (cm/s)	46.5 (13.7)	48.5 (15.9)	0.18	1.41	0.14
ΔMFV (cm/s)	12.0 (7.4)	15.3 (12.9)	0.27	1.15	0.31

CVR = cerebrovascular reactivity, etCO₂ = end-tidal CO2, MFV = mean flow velocity of middle cerebral artery. For Cohen's d < 0.20 is negligible, 0.20–0.49 is small, and 0.50–0.79 is medium.

Table 3.

Summary of Correlation Between Predictor Variable and Cerebrovascular Reactivity Comparing Techniques for Inducing Hypercapnia.

Predictor variable	Re-breathing CVR		Breath-hold CVR
	Correlation coefficient (R)	p-value	Correlation coefficient (R)	p-value
Age (years)	−0.68	<0.01*	−0.45	0.08
Diabetes Mellitus	−0.22	0.42	−0.22	0.41
Hypertension	−0.9	0.75	−0.01	0.76
Coronary artery disease	< 0.01	0.99	−0.28	0.30
Peripheral Vascular disease	−0.25	0.36	−0.20	0.45
Stroke	−0.45	0.08	−0.34	0.20
Smoker	−0.24	0.37	−0.49	0.05

P-values are not adjusted for multiple comparisons.

^⁎p-value < 0.05.

4. Discussion

The ability to measure CVR is important given its association with important outcomes of mortality and stroke, in addition to its relationship with cognition. [9,10] In our ESKD population, the circulatory stress of hemodialysis can lead to cerebral ischemia in those with decreased cerebrovascular health. Thus, measuring CVR can provide important prognostic information in this population. Use of TCD and induced hypercapnia to measure CVR is non-invasive and low cost compared to MRI, but there are concerns about accuracy specifically on the change in pCO₂ that is obtained. [11] We used two methods to increase pCO₂ and found that using re-breathing to measure CVR, allowed for more accurate measurement of the change in etCO₂ and resulted in a CVR measurement that was associated with expected factors of age and history of stroke.

Using a re-breathing technique, we could continuously monitor etCO₂ and had more control over the change in etCO₂ as we were able to stop the re-breathing at a 10 mmHg increase. With the breath-hold technique we are not able to see what the change in etCO₂ until the patient released their breath and had more variation in the resulting change in etCO₂. Further, anecdotally our patients had more trouble completing the breath-hold for 30 ss and ending with an exhalation (rather than immediate inhalation) which is needed to capture the etCO₂ at the end of the breath-hold. Inaccurate capture of etCO₂ at the end of breath hold did lead to an erroneous CVR value in one of our participants due to a measured change of 3.9 mmHg in etCO₂, and a very high outlying CVR calculation of 10.8 %/mmHg compared to the 2.1 %/mmHg CVR we obtained in the same patient when using the re-breathing technique. In our cohort and in another study that evaluated breath-hold compared to CO₂ inhalation [12] the breath-hold CVR measurements were higher and possibly overestimated the actual CVR. We also found that there was less head movement during re-breathing compared to breath-hold. During the breath-hold, participants sometimes moved their head when pinching their nose shut. Head movement may impact the cerebral blood flow velocity readings if the ultrasound probes shift during movement potentially reducing accuracy of the MFV data. We also found that there may be effects of valsalva on the MFV readings when patient would take a deep breath for the breath-hold. During the breath-hold an increase in intrathoracic pressure leads to a slight drop in the MFV reading initially and a sympathetic response may lead to rebound in MFV which may affect the peak MFV reached during hypercapnia from breath-hold. [13]

We note that re-breathing CVR correlated better with expected risk factors of age and history of stroke, indicating more accurate readings. We expect that CVR declines with age [14,15] in part due to increased arterial stiffness from arteriosclerosis. Further, there was a trend between having a history of stroke and lower CVR, which would also be expected given the cerebrovascular disease that is noted in stroke victims. [16]

A significant limitation of our study was the low sample size of only 16 participants. Also, this is a post-hoc analysis that was completed when we refined our CVR measurement technique as part of a larger study. Thus, we did not have set protocols to obtain patient perspective of the techniques. However, we feel that these issues may be encountered by others, so it is important to share the knowledge we gained. Another limitation is we had no data on changes in oxygen saturation during re-breathing, so could not account for change in pO₂ in this study. However, blood flow does not change in brain tissue until pO₂ fall below approximately 50mmHg. [17] This would be unlikely if O₂ saturation remains above 90 %. Adding pulse oximetry to ensure O₂ saturation remains above 90 % would improve the re-breathing methodology.

5. Conclusion

We found that using TCD and re-breathing to increase pCO₂ and measure CVR is well-tolerated by our frail, medically ill, hemodialysis patient population, who are at risk for cerebral ischemic injury from the hemodynamic stress of hemodialysis. Although our cohort is small, we hope that our findings can still aid other investigators when deciding on methodology to measure cerebrovascular health in medically complex frail patient populations, especially those at risk for hypoperfusion due to circulatory stress.

Funding

This study was supported by grant 1R03DK132441 from the NIH/NIDDK.

Disclosures

None.

CRediT authorship contribution statement

Claire Seigworth: Writing – review & editing, Project administration, Investigation. Isabelle Grassl: Investigation, Data curation. Dawn F. Wolfgram: Writing – review & editing, Writing – original draft, Visualization, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Dawn Wolfgram reports financial support was provided by National Institute of Diabetes and Digestive and Kidney Diseases. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.