Respiratory Physiology & Neurobiology 197 (2014) 15–18

Contents lists available at ScienceDirect

Respiratory Physiology & Neurobiology journal homepage: www.elsevier.com/locate/resphysiol

Circadian cerebrovascular reactivity to CO2 J. Strohm a , J. Duffin b,c,∗ , J.A. Fisher b,c a

Department of Medical and Life Sciences, Hochschule Furtwangen University, Germany Department of Physiology, University of Toronto, Toronto, ON, Canada M5S 1A8 c Department of Anaesthesiology, University of Toronto, and University Health Network, Toronto, ON, Canada b

a r t i c l e

i n f o

Article history: Accepted 10 March 2014 Available online 20 March 2014 Keywords: Carbon dioxide Hypoxia Cerebral blood flow Blood pressure Humans

a b s t r a c t Cerebrovascular reactivity (CVR) assesses the ability of the cerebral vasculature to adjust cerebral blood flow in response to changes in arterial carbon dioxide (CO2 ), and is used as an indicator of cerebrovascular health. A common method of estimating CVR is to measure the increase in blood flow velocity of the middle cerebral artery (MCAv), using trans-cranial Doppler ultrasound, in response to a CO2 stimulus. We used this method to measure the CVR of 10 subjects in the mornings and evenings of two consecutive days. Mean arterial pressure (MAP) was also measured, and CVR was determined solely from tests where MAP remained unchanged in response to CO2 . CVR was measured as the slopes of MCAv responses to a ramp CO2 stimulus fitted with linear regression, and significantly increased from evening to morning each day, with no significant day-to-day differences. We concluded that these measurements of CVR exhibited a circadian rhythm, and were repeatable from one day to the next. © 2014 Elsevier B.V. All rights reserved.

1. Introduction The measurement of cerebrovascular reactivity to CO2 (CVR) using trans-cranial Doppler ultrasound (TCD) to measure the blood flow velocity in the middle cerebral artery (MCAv) is a commonly used method to assess brain vascular reactivity (Sato et al., 2012; Willie et al., 2012). However, only a few studies have examined the possibility that a circadian rhythm exists in humans, and the dayto-day variability of CVR is unknown. TCD measurements of resting MCAv have a circadian rhythm (Conroy et al., 2005), lowest at noon and highest at midnight, and are affected by sleep fragmentation so that both resting MCAv and MCAv while breathing 5% CO2 are less in the morning than the previous evening (Qureshi et al., 1999). Furthermore, measurements of CVR in humans, estimated from MCAv changes induced by breathing 5% CO2 , show an overnight decrease (Cummings et al., 2007) that may be related to the overnight reduction in autoregulation (Ainslie et al., 2007). These findings suggest the existence of a circadian rhythm for CVR. However, the methodology previously used to measure CVR provided an uncertain CO2 stimulus (Fierstra et al., 2013). Therefore we re-examined the hypothesis using a methodology that provides a known and standardised CO2 stimulus, and compared morning

∗ Corresponding author at: Department of Physiology, Medical Sciences Building, 1 King’s College Circle, University of Toronto, Toronto, Ontario, Canada M5S 1A8. Tel.: +1 416 597 1325x291; fax: +1 416 597 1330. E-mail address: j.duffi[email protected] (J. Duffin). http://dx.doi.org/10.1016/j.resp.2014.03.003 1569-9048/© 2014 Elsevier B.V. All rights reserved.

and evening measurements of CVR. Since the CO2 stimulus was accurately reproduced between tests we also examined the reproducibility of the CVR measurement from one day to the next by repeating the CVR measurements on two consecutive days. 2. Methods 2.1. Subjects and ethical approval Ten subjects (4 female), aged between 22 and 36 years participated in this study after approval from the Research Ethics Board at the Toronto General Hospital (University Health Network), and after each gave their written informed consent. These studies therefore conformed to the standards set by the latest revision of the Declaration of Helsinki. Subjects were chosen to avoid any history or symptoms of cardiovascular, cerebrovascular, or respiratory diseases, and were taking no medications other than oral contraceptives. They were instructed to abstain from caffeinated or alcoholic drinks and vigorous exercise for at least 12 h before the CVR measurements were made. 2.2. Apparatus Subjects were instrumented as follows. They breathed through a face mask secured against leaks with adhesive tape (Tegaderm, 3M Health Care, St. Paul, MN, USA). Ventilation was measured with a mass flow sensor (AWM720P1 Airflow, Honeywell; Freeport,

J. Strohm et al. / Respiratory Physiology & Neurobiology 197 (2014) 15–18

220

60

A

40 180

20

140

0

MCAv (%)

-20

MAP (mmHg)

100

MCAv

IL, USA), and beat-by-beat middle cerebral artery flow velocity (MCAv) was measured bilaterally with 2 MHz trans-cranial Doppler ultrasound sampled at 125 Hz (ST3 Spencer Technologies; Seattle, USA). Beat-by-beat mean arterial pressure (MAP) and heart rate (HR) were measured with finger plethysmography sampled at 200 Hz (Nexfin, BMYE; Amsterdam, The Netherlands). Breathed gas was sampled and analysed for the partial pressures of CO2 and O2 (RespirActTM , Thornhill Research Inc., Toronto, Canada) and recorded at 20 Hz. These instruments saved a digitised record of each subject for retrospective analysis as well displaying their measures in real time.

PETCO2 MAP

16

-40 -60

PETCO2 (mmHg)

60

-80

2.3. Protocol

-100 400

700

1000

1300

Time (s) 220

60

B

40 180

20 0

140 MCAv (%)

-20

100

MCAv

PETCO2 MAP

Subjects were seated in a comfortable chair and the following sequence of changes in the end-tidal partial pressures of CO2 (PetCO2 ) were applied while the partial pressure of O2 (PetCO2 ) was maintained at 100 mmHg. PetCO2 was held constant at 40 mmHg for 2 min, followed by a step increase to 50 mmHg held for 5 min, a slow decline to 35 mmHg over 2 min held for 3 min, a slow increase to 50 mmHg over 4 min, and a return to 40 mmHg. The method described by Slessarev et al. (2007) was used to apply this sequence, using a sequential gas delivery (SGD) circuit and a computer controlled gas blender (RespirActTM ; Thornhill Research Inc, Toronto, Canada). Subjects were asked to breathe in time to a metronome at a frequency of 15 b/min and empty the inspiratory bag of the SGD circuit on each breath. This testing protocol was applied in the morning between 8 and 10 AM and in the evening between 5 and 7 PM for two consecutive days.

20 100

-40 MAP (mmHg)

-60

60

-80 PETCO2 (mmHG)

20 800

-100 1400

1100

2.4. Data analysis

Time (s) 50

3. Results The stimulus protocol illustrated in Fig. 1 was well tolerated and completed by all subjects, with only one missed test session, so that a total of 39 test sessions were completed. However, not all of the results obtained from the step and ramp test stimuli provided satisfactory MCAv data for CVR calculations. MAP increases with CO2

C 30

MCAv (%)

For each test, beat-by-beat values of MAP, HR and the 4-s averages of MCAv were time aligned with the breath-by-breath PetCO2 and PetO2 measures and their breath-by-breath values calculated. The average breath-by-breath MAP and MCAv over the 2-minute initial period of the protocol served as baseline measures, and changes in MCAv were calculated as a percent change from this baseline. CVR was calculated from the breath-by-breath MCAv vs. PetCO2 relations for both the step and ramp stimuli as follows: First the MAP vs. PetCO2 relations were examined for changes during the step stimulus, and tests with increasing MAP with CO2 were discarded. CVR was calculated as the slope of a linear regression fit from the start of baseline measurements to the end of the 5-min step increase in PetCO2 . For ramp tests, those with increasing MAP with PetCO2 throughout the ramp were discarded, but if MAP only increased above a threshold PetCO2 , then the ramp portion where MAP remained constant was used for calculating CVR as the slope of a linear regression fit. Statistical comparisons were made using repeated measures analyses of variance (rmANOVA) computed with statistics software (SigmaPlot 12.5, Systat Software, Germany). Where factors were found to be significant, post hoc all pairwise multiple comparisons were made using the Holm–Sidak method. Analysis of the day-to-day changes used the intraclass correlation coefficient (ICC) as an estimate of reliability calculated using an Excel spreadsheet (newstats.org).

10

-10

-30 30

35

40

45

50

PETCO2 (mmHg) Fig. 1. An example of a test session and its analysis. (A) The time course of the recorded variables. Note the rise in MAP during the step stimulus. (B) A close up of the ramp stimulus portion showing the constancy of MAP. (C) Analysis of the ramp response showing the linear fit.

over the course of the experiment occurred frequently during the step stimulus as illustrated in Fig. 1. If CO2 increases activate the sympathetic system and increase MAP then CBF regulation will be affected. We wanted to measure the cerebrovascular reactivity to CO2 alone, and not combined with the effects of pressure autoregulation and sympathetic activation. For example, if MAP increases then at least two more factors are involved in producing the measured MCAv changes in addition to the vasodilatory effects of CO2 . First, pressure autoregulation vasoconstriction, which competes with the CO2 vasodilation, and second, the effect of an increased pressure gradient increasing flow, an effect independent of vessel diameter control. In 40.7% of the step tests MAP increased by more than 10 mmHg. The mean (SD) increase was 15.6 (7.8) mmHg, compared to 5.1

J. Strohm et al. / Respiratory Physiology & Neurobiology 197 (2014) 15–18

17

Table 1 Mean (SD) baseline values of MAP, MCAv and PetCO2 . Morning MCAv

PetCO2

MAP

MCAv

PetCO2

80.7(8.5) 79.6(8.0)

44.2 (14.4) 51.4 (15.5)

40.7 (1.8) 41.3 (0.4)

83.6(8.6) 80.9(6.3)

46.7 (10.7) 44.7 (10.8)

41.1 (0.8) 41.2 (0.6)

5

5

4

4

CVR day 2 (%/mmHg)

CVR (%/mmHg)

Day 1 Day 2

Evening

MAP

*

3

*

2

1

2

1

0

0 morning

evening

evening

morning

day one

0

1

2

3

4

5

CVR day 1 (%/mmHg)

day two

Fig. 2. CVR values determined from the linear regression fits of the MCAv (%) responses to the ramp stimulus. The dots and lines are the individual subject values and the bar and error bars indicate the mean and SD. Asterisks denote significant differences.

(2.2) mmHg. As a result, CVR calculated from the step stimulus measurements were obtained for only 5 subjects, and so further statistical testing of the step responses was abandoned. By contrast, MAP either remained constant throughout the ramp stimulus or increased only when a threshold PetCO2 was exceeded. In the latter case CVR was calculated from the ramp responses to CO2 below the threshold. Testing for day-to-day and morning-to-evening differences in the baseline values of MCAv, PetCO2 and MAP using a 2-way rmANOVA found no differences (Table 1). These CVR values were analysed using a two-way rmANOVA with factors day (1 vs. 2) and time of day (evening vs. morning). Table 1 shows the results; day was not a significant factor but time of day was a significant factor on both day 1 and day 2. Fig. 2 illustrates this finding (Table 2). With differences found between morning and evening CVR values, we examined the test–retest relationship between day 1 and day 2 for each time of day (Fig. 3) and calculated the Pearson correlation coefficient as 0.65. Test reliability was measured with the ICC as 0.68 with limits between 0.39 and 0.84. 4. Discussion 4.1. Overall findings These experiments were designed to provide measurements of CVR derived from both step and ramp CO2 stimuli. However, MAP

Table 2 Mean (SD) CVR for the MCAv (%) responses to the ramp CO2 stimuli. The P values are from the 2-way rmANOVA post hoc all pairwise multiple comparisons using the Holm–Sidak method, with significant values bolded and italicised.

Day 1 Day 2 P

3

Morn

Eve

P

3.30 (0.53) 3.05 (0.58) 0.493

2.65 (0.86) 2.54 (0.72) 0.516

0.005 0.016

Fig. 3. CVR testing reliability plot of day 2 measures vs. day 1 measures, with the line of identity shown as a solid line. Morning test measures are shown as circles and evening test measures are shown as triangles.

increased with CO2 in almost half of the step tests, but remained constant or increased above a threshold CO2 during the ramp tests. With the increase in MAP a confounding factor, CVR could not be determined for the step tests, so that the comparison of CVR measures between days, and between morning and evening, were derived solely from the linear regression fitting of the ramp responses. These experiments showed that a time of day effect on CVR can be detected using a ramp CO2 stimulus and a linear regression fit to the MCAv response. Day-to-day CVR variation was not detected, as long as the measurements were made at the same time of day, demonstrating the repeatability of the testing. We concluded that the time of day differences in CVR was evidence that CVR is subject to a circadian rhythm. 4.2. Comparison with previous findings We found that CVR increased from evening to morning in contrast to the overnight decrease in CVR (mean ± SD, 5.3 ± 0.6 vs. 4.6 ± 1.1%/T) found by Cummings et al. (2007). They used a breath hold technique to elevate CO2 and measured the resulting MCAv response; a technique with several problems. First, the PaCO2 stimulus change was unknown. Although PetCO2 was measured at the termination of the breath hold, the actual PaCO2 stimulus is neither reliably measured nor repeatable between tests (Fierstra et al., 2013). Second, MAP increased during the breath hold (see Fig. 4 in Cummings et al., 2007), so that the increase of MCAv due solely to the vasodilatory influence of CO2 cannot be separated from this confounding factor. We suggest that the magnitude of the increase in MAP with breath holding decreases overnight, and therefore so does the MCAv response to breath holding. The ramp stimulus we employed was implemented using prospective end-tidal targeting with a sequential gas delivery breathing circuit, which has been shown to control PaCO2 equal to PetCO2 (Ito et al., 2008). We therefore suggest that the difference in findings between our study and that of Cummings et al. (2007) is due to the methodology used to measure CVR. The CVR values we measured are similar to those found by others in

18

J. Strohm et al. / Respiratory Physiology & Neurobiology 197 (2014) 15–18

recent studies (mean ± SD 2.9 ± 0.47%/mmHg; Willie et al., 2012 and 3.25 ± 1.0%/mmHg; Skow et al., 2013 and 3.31 ± 0.87; Sato et al., 2012). We are therefore confident that our measures correctly reflect of the overnight change in CVR. If we are correct in our conclusion that CVR experiences a circadian rhythm then this finding has implications for the control of breathing since CVR interacts with the central respiratory chemoreflex (Ainslie and Duffin, 2009). Spengler et al. (2000) noted a circadian rhythm in the ventilatory response to hypercapnia such that the sensitivity increased during the sleep period, a change that would tend to destabilise breathing. However, with CVR increasing from evening to morning, the control of central PCO2 improves and the central respiratory chemoreflex gain decreases, thereby assisting the stabilisation of respiration in healthy subjects during sleep (Ainslie and Duffin, 2009). Indeed, measurements of the overnight changes in the central chemoreflex sensitivity found no change in healthy subjects (Mahamed et al., 2005). With an increase in CVR from evening to morning not only does this change help stabilise breathing but since PCO2 increases during sleep (Ainslie and Duffin, 2009). The increased CBF response to PCO2 may be beneficial by driving metabolite clearance (Xie et al., 2013). 4.3. Reliability of CVR measurement These experiments also enabled an estimation of the test–retest reliability of this CVR measurement technique, which was quantified in terms of the ICC as 0.68. This value is the first reliability estimate for CVR testing using hypercapnia produced by prospective end-tidal PCO2 targeting and sequential gas delivery with TCD measurements of MCAv. Previous estimates using inhalation of 5% CO2 95% O2 and TCD measurements of MCAv found an ICC of 0.43 (Totaro et al., 1999) and between 0.46 and 0.82 (McDonnell et al., 2013) depending on the experimenter. We suggest that the ability of our methodology to provide a repeatable control of the PaCO2 stimulus (Fierstra et al., 2013), avoiding the confounding influence of MAP changes with CO2 , as well as the use of a ramp stimulus pattern accounts for the high ICC we obtained. 4.4. Limitations The circadian time course for resting MCAv is such that it is maximum at midnight and minimum at noon (Conroy et al., 2005). Our CVR measurements, made between 8–10 AM and 5–7 PM, were therefore at times when resting MCAv would be similar, and indeed we found no baseline differences. Therefore, if the circadian rhythm of CVR is phase locked to that of resting MCAv, then the timing of measurements in this study may have produced an underestimation of the circadian changes in CVR. As is well understood, measurements of MCAv alone as a surrogate for blood flow may be incorrect if the vessel diameter changes. While some experiments find that MCA diameter does not change in hypercapnia (Serrador et al., 2000) others found increases with high CO2 tensions (Valdueza et al., 1999). We point out that the increase in MAP during these latter experiments likely confounded the findings. We, as others have done (reviewed in Ainslie and Duffin, 2009), assumed that in the range of CO2 tensions employed in this study the changes in MCAv provided an accurate estimation of the changes in MCA blood flow.

4.5. Conclusions Cerebrovascular reactivity to CO2 has a circadian component; decreasing from morning to evening and increasing overnight. These changes serve to mitigate the effect of circadian changes in the respiratory chemoreflex sensitivity on breathing stability during sleep. Our findings also highlight the importance of conducting repeat experiments at the same time of day. Acknowledgements Supported by University of Toronto Merit award to JAF and a non-directed grant from Thornhill Research Inc. We thank Rosemary Regan and Olivia Sobczyk for their technical advice and assistance. References Ainslie, P.N., Duffin, J., 2009. Integration of cerebrovascular CO2 reactivity and chemoreflex control of breathing: mechanisms of regulation, measurement, and interpretation. Am. J. Physiol. Regul. Integr. Comp. Physiol. 296, R1473–R1495. Ainslie, P.N., Murrell, C., Peebles, K., Swart, M., Skinner, M.A., Williams, M.J., Taylor, R.D., 2007. Early morning impairment in cerebral autoregulation and cerebrovascular CO2 reactivity in healthy humans: relation to endothelial function. Exp. Physiol. 92, 769–777. Conroy, D.A., Spielman, A.J., Scott, R.Q., 2005. Daily rhythm of cerebral blood flow velocity. J. Circadian Rhythms 3, 3. Cummings, K.J., Swart, M., Ainslie, P.N., 2007. Morning attenuation in cerebrovascular CO2 reactivity in healthy humans is associated with a lowered cerebral oxygenation and an augmented ventilatory response to CO2 . J. Appl. Physiol. 102, 1891–1898. Fierstra, J., Sobczyk, O., Battisti-Charbonney, A., Mandell, D.M., Poublanc, J., Crawley, A.P., Mikulis, D.J., Duffin, J., Fisher, J.A., 2013. Measuring cerebrovascular reactivity: what stimulus to use? J. Physiol. 591, 5809–5821. Ito, S., Mardimae, A., Han, J., Duffin, J., Wells, G., Fedorko, L., Minkovich, L., Katznelson, R., Meineri, M., Arenovich, T., Kessler, C., Fisher, J.A., 2008. Non-invasive prospective targeting of arterial PCO2 in subjects at rest. J. Physiol. 586, 3675–3682. Mahamed, S., Hanly, P.J., Gabor, J., Beecroft, J., Duffin, J., 2005. Overnight changes of chemoreflex control in obstructive sleep apnoea patients. Respir. Physiol. Neurobiol. 146, 279–290. McDonnell, M.N., Berry, N.M., Cutting, M.A., Keage, H.A., Buckley, J.D., Howe, P.R., 2013. Transcranial Doppler ultrasound to assess cerebrovascular reactivity: reliability, reproducibility and effect of posture. PeerJ 1, e65. Qureshi, A.I., Christopher Winter, W., Bliwise, D.L., 1999. Sleep fragmentation and morning cerebrovasomotor reactivity to hypercapnia. Am. J. Respir. Crit. Care Med. 160, 1244–1247. Sato, K., Sadamoto, T., Hirasawa, A., Oue, A., Subudhi, A.W., Miyazawa, T., Ogoh, S., 2012. Differential blood flow responses to CO2 in human internal and external carotid and vertebral arteries. J. Physiol. 590, 3277–3290. Serrador, J.M., Picot, P.A., Rutt, B.K., Shoemaker, J.K., Bondar, R.L., 2000. MRI measures of middle cerebral artery diameter in conscious humans during simulated orthostasis. Stroke 31, 1672–1678. Skow, R.J., MacKay, C.M., Tymko, M.M., Willie, C.K., Smith, K.J., Ainslie, P.N., Day, T.A., 2013. Differential cerebrovascular CO2 reactivity in anterior and posterior cerebral circulations. Respir. Physiol. Neurobiol. 189, 76–86. Slessarev, M., Han, J., Mardimae, A., Prisman, E., Preiss, D., Volgyesi, G., Ansel, C., Duffin, J., Fisher, J.A., 2007. Prospective targeting and control of end-tidal CO2 and O2 concentrations. J. Physiol. 581, 1207–1219. Spengler, C.M., Czeisler, C.A., Shea, S.A., 2000. An endogenous circadian rhythm of respiratory control in humans. J. Physiol. 526, 683–694. Totaro, R., Marini, C., Baldassarre, M., Carolei, A., 1999. Cerebrovascular reactivity evaluated by transcranial Doppler: reproducibility of different methods. Cerebrovasc. Dis. 9, 142–145. Valdueza, J.M., Draganski, B., Hoffmann, O., Dirnagl, U., Einhaupl, K.M., 1999. Analysis of CO2 vasomotor reactivity and vessel diameter changes by simultaneous venous and arterial Doppler recordings. Stroke 30, 81–86. Willie, C.K., Macleod, D.B., Shaw, A.D., Smith, K.J., Tzeng, Y.C., Eves, N.D., Ikeda, K., Graham, J., Lewis, N.C., Day, T.A., Ainslie, P.N., 2012. Regional brain blood flow in man during acute changes in arterial blood gases. J. Physiol. 590, 3261–3275. Xie, L., Kang, H., Xu, Q., Chen, M.J., Liao, Y., Thiyagarajan, M., O’Donnell, J., Christensen, D.J., Nicholson, C., Iliff, J.J., Takano, T., Deane, R., Nedergaard, M., 2013. Sleep drives metabolite clearance from the adult brain. Science 342, 373–377.