586
Exp Physiol 99.3 (2014) pp 586–598
Research Paper
Assessment of dynamic cerebral autoregulation and cerebrovascular CO2 reactivity in ageing by measurements of cerebral blood flow and cortical oxygenation Madelijn H. Oudegeest-Sander1,2,∗ , Arenda H. E. A. van Beek1,∗ , Karin Abbink1,2 , Marcel G. M. Olde Rikkert1 , Maria T. E. Hopman2 and Jurgen A. H. R. Claassen1 Departments of 1 Geriatric Medicine and 2 Physiology, Radboud University Medical Center, Nijmegen, The Netherlands
Experimental Physiology
New Findings r What is the central question in this study? It is unknown to what extent increasing age influences the dynamic adaptations of cerebral blood flow velocity and cortical oxygenation in response to changes in blood pressure (cerebral autoregulation) and to changes in carbon dioxide (cerebrovascular CO2 reactivity). r What is the main finding and its importance? We have shown that ageing up to 86 years is associated with an overall preservation of dynamic cerebral autoregulation and cerebrovascular CO2 reactivity, leading to a sufficiency of cerebral cortical oxygenation during daily life activities, despite the decrease in absolute cerebral blood flow velocity and increase in cerebrovascular resistance with advancing age.
With ageing, cerebral blood flow velocity (CBFV) decreases; however, to what extent dynamic cerebral autoregulation and cerebrovascular CO2 reactivity are influenced by ageing is unknown. The aim was to examine the dynamic responses of CBFV and cortical oxygenation to changes in blood pressure (BP) and arterial CO2 across different ages. Fifty-eight participants in three age groups were included, as follows: young (n = 20, 24 ± 2 years old), elderly (n = 20, 66 ± 1 years old), and older elderly (n = 18, 78 ± 3 years old). The CBFV was measured using transcranial Doppler ultrasound, simultaneously with oxyhaemoglobin (O2 Hb) using near-infrared spectroscopy and beat-to-beat BP measurements using Finapres. Postural manoeuvres were performed to induce haemodynamic fluctuations. Cerebrovascular CO2 reactivity was tested with hyperventilation and CO2 inhalation. With age, CBFV decreased (young 59 ± 12 cm s−1 , elderly 48 ± 7 cm s−1 and older elderly 42 ± 9 cm s−1 , P < 0.05) and cerebrovascular resistance increased (1.46 ± 0.58, 1.81 ± 0.36 and 1.98 ± 0.52 mmHg cm−1 s−1 , respectively, P < 0.05). Normalized gain (autoregulatory damping) increased with age for BP–CBFV (0.88 ± 0.18, 1.31 ± 0.30 and 1.06 ± 0.34, respectively, P < 0.05) and CBFV–O2 Hb (0.10 ± 0.09, 0.12 ± 0.04 and 0.17 ± 0.08, respectively, P < 0.05) during the repeated sit–stand manoeuvre at 0.05 Hz. Even though the absolute changes in CBFV and cerebrovascular resistance index during the cerebrovascular CO2 reactivity were higher in the young group, the percentage changes in CBFV, cerebrovascular resistance index and O2 Hb were similar in all age groups. In conclusion, there was no decline in dynamic cerebral autoregulation and cerebrovascular CO2 reactivity with increasing age up to 86 years. Despite the decrease in cerebral blood flow
∗
M. H. Oudegeest-Sander and A. H. E. A. van Beek contributed equally to this work.
DOI: 10.1113/expphysiol.2013.076455
C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
Exp Physiol 99.3 (2014) pp 586–598
Cerebral perfusion and cortical oxygenation in ageing
587
velocity and increase in cerebrovascular resistance with advancing age, CBFV and cortical oxygenation were not compromised in these elderly humans during manoeuvres that mimic daily life activities. (Received 27 September 2013; accepted after revision 16 December 2013; first published online 20 December 2013) Corresponding author J. A. H. R. Claassen: Department of Geriatric Medicine (925), Radboud University Medical Center, PO Box 9101, 6500 HB Nijmegen, The Netherlands. Email: [email protected]
Introduction Physiological ageing is associated with well-recognized changes in the systemic and cerebral vasculature (Ferrari et al. 2003; Lakatta & Levy, 2003; O’Rourke & Hashimoto, 2007). Systemic vascular changes lead to an increased risk for cardiovascular events, while cerebrovascular adaptations enhance the risk of (ischaemic) stroke. Furthermore, resting cerebral blood flow (CBF) is decreased in older individuals (Lipsitz et al. 2000; Sorond et al. 2005). Whether this decrease in CBF reflects a decrease in cerebral demand or a true diminution in blood flow remains unknown; likewise, whether ageing leads to diminished cerebrovascular function remains unclear. Cerebrovascular function can be studied using two physiological phenomena, namely dynamic cerebral autoregulation and cerebrovascular reactivity to carbon dioxide. Dynamic cerebral autoregulation reflects the ability of the brain to stabilize CBF despite rapid daily life changes in blood pressure (Paulson et al. 1990; Panerai, 1998; van Beek et al. 2008). Dynamic cerebral autoregulation remains stable up to the age of 73 years (Lipsitz et al. 2000; Sorond et al. 2005; van Beek et al. 2008) but has not been estimated in healthy elderly subjects above this age. Most studies examined the effects of small blood pressure changes (Diehl et al. 1991, 1998) instead of larger fluctuations in blood pressure during postural changes, which may provide a stronger challenge for dynamic cerebral autoregulation and may therefore be more sensitive to detect changes in cerebrovascular function (Claassen et al. 2009). Cerebrovascular reactivity to changes in CO2 , the second phenomenon that indicates cerebrovascular function, declines significantly with increasing age up to 90 years (Bakker et al. 2004). Unfortunately, most studies investigated either the dynamic cerebral autoregulation (Narayanan et al. 2001; Sorond et al. 2005; Claassen et al. 2009; van Beek et al. 2010a) or the cerebrovascular CO2 reactivity (Ringelstein et al. 1988; Bakker et al. 2004; Claassen et al. 2007; Dineen et al. 2011; van Beek et al. 2011), but the studies did not examine both aspects of the cerebrovascular system simultaneously. Dynamic cerebral autoregulation and cerebrovascular CO2 reactivity are both directly related to the oxygenation of the brain tissue and therefore to human brain function (van Beek et al. 2010b, 2012). The majority of studies C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
of dynamic cerebral autoregulation and cerebrovascular CO2 reactivity have used transcranial Doppler ultrasound (TCD) to obtain continuous measurements of cerebral blood flow velocity (CBFV), together with continuous blood pressure measurements (Finapres). Transcranial Doppler ultrasound does not provide information regarding the oxygenation of the brain. The addition of near-infrared spectroscopy (NIRS) can be used to measure concentrations of oxygenated and deoxygenated haemoglobin and provide estimates of cortical brain tissue oxygenation (Reinhard et al. 2006). Thus, by using Finapres and transcranial Doppler sonography concomitantly with near-infrared spectroscopy, it is possible to determine two different but dependent dynamic relationships, namely the relationship between blood pressure and cerebral blood flow on the one hand and the relationship between cerebral blood flow and cerebral cortical oxygenation on the other (van Beek et al. 2010b, 2012). This study aimed to investigate effects of age on dynamic cerebral autoregulation and cerebrovascular CO2 reactivity in order to examine cerebrovascular function by investigating the relationship between blood pressure, cerebral blood flow and cerebral cortical oxygenation in a group of young, elderly and older elderly subjects. The combination of examining dynamic cerebral autoregulation and cerebrovascular CO2 reactivity using both TCD and NIRS in this population is, to our knowledge, unique. We hypothesized that the cerebrovascular function, and thereby the cortical oxygenation, are diminished in healthy elderly aged 74 years and older compared with young subjects and with elderly subjects below the age of 74 years. To investigate this hypothesis, beat-to-beat blood pressure, cerebral blood flow and cortical oxygenation were measured simultaneously to investigate cerebrovascular haemodynamics during changes in blood pressure and changes in arterial CO2 .
Methods Ethical approval
The study was approved by the Medical Ethics Committee, and was performed according to the Declaration of
588
M. H. Oudegeest-Sander and others
Helsinki and Good Clinical Practice guidelines. All participants gave written informed consent. Participants
Fifty-eight participants in the following three different age groups were included: a young group (young, mean age 24 ± 2 years, range 21–28 years), an older group (elderly, mean age 66 ± 1 years, range 65–69 years) and the oldest group (older elderly, mean age 78 ± 3 years, range 74–86 years). Participants were recruited at the Departments of Physiology and Geriatric Medicine, Radboud University Medical Center, Nijmegen, The Netherlands. All participants were extensively screened. Participants with hypertension (>160/90 mmHg), diabetes mellitus reported in the medical history (fasting glucose >6.9 mmol l−1 ), hypercholesterolaemia (total cholesterol >7.5 mmol l−1 ), body mass index >32.5 kg m−2 , compromised cognition (Mini Mental State Examination 0.4 in this study (Zhang et al. 1998; Claassen et al. 2009; van Beek et al. 2010a). Repeated sit–stand manoeuvre at 0.05 Hz. The repeated sit–stand manoeuvre causes oscillations that are much larger than the spontaneous oscillations that occur during the baseline sitting rest position; therefore, the dynamic relationship can be assessed with greater coherence. Transfer function analysis was performed in the very low frequency range, 0.02–0.07 Hz. The dynamic relationships between blood pressure as input and CBFV as output (MAP–CBFV) and CBFV–O2 Hb were determined. Data were also included only if coherence was >0.4. We have previously shown that upstream oscillations in CBFV induced by changes in blood pressure make an important contribution to the downstream brain tissue level oscillations in O2 Hb (van Beek et al. 2012). It is our experience that oscillations induced at a frequency of 0.05 Hz result in clear blood pressure and CBFV oscillations with good coherence, whereas oscillations induced at a frequency of 0.1 Hz result in more noise and thereby loss of reliable data (van Beek et al. 2010a, 2012). Furthermore, in accordance with the high-pass filter, dynamic cerebral autoregulation is mainly active in the low and very low frequency regions. Single squat–stand manoeuvre. For the single squat–stand manoeuvre we analysed two time periods: (1) squatting; and (2) maximal decline after standing up from squatting. For this, we used 30 s of steady-state measurements in the squatting position (1) and the period of five consecutive seconds in which MAP and CBFV had decreased the most upon standing (2). Both absolute and relative changes in MAP and CBFV were calculated. Cerebrovascular CO2 reactivity. Regarding the analysis of the cerebrovascular CO2 reactivity, absolute and relative changes in heart rate, MAP, CBFV, CVRi and O2 Hb were calculated. The full range of cerebrovascular
Exp Physiol 99.3 (2014) pp 586–598
CO2 reactivity was calculated (Ringelstein et al. 1988), which represents the percentage change in CBFV from hypocapnia (hyperventilation) to hypercapnia (inhalation of 7% CO2 ) relative to baseline CBF = {[CBFVmax (7% CO2 ) − CBFVmin (hyperventilation)]/CBFVrest }/100%. We have also calculated the change in CBFV between hypocapnia and hypercapnia per millimetre of mercury change in EtCO2 between hypocapnia and hypercapnia (Bailey et al. 2013). Statistical analysis
Data are presented as means ± SD. Comparisons between the different age groups were performed using one-way ANOVA with Bonferroni correction for multiple comparisons. We checked the data with the Hochberg method because of sample size differences (especially with the single squat–stand manoeuvre) and with the Games–Howell method because of different group variances. As these three methods did not show relevant differences, we mention only the results from the one-way ANOVA with Bonferroni correction. Statistical significance was set at a P value of 0.4 were used. There were no relevant differences in coherence results between age groups. The older elderly had a slightly lower coherence in the LF compared with the young group (P = 0.001), but mean coherence remained well above 0.4. Values for phase were very similar across age groups, but young subjects had a higher mean phase in the VLF (where cerebral autoregulation is most active); this difference reached statistical significance when compared with younger elderly subjects (P = 0.023). However, VLF phase did not differ between younger and older elderly groups. Table 2 lists absolute gain, which is sensitive to individual differences in CVFV or blood pressure, and normalized gain, which is not affected by absolute CBFV or blood pressure. Absolute gain was marginally higher in the LF compared with the elderly (P = 0.05). Normalized gain revealed no differences between age groups in the VLF and LF range where cerebral autoregulation is active; however, normalized gain was higher in the HF in the older elderly compared with the young group (P = 0.006). Repeated sit–stand manoeuvre at 0.05 Hz
The results from the repeated sit–stand manoeuvre are presented in Table 3. Two subjects (one in each elderly group) lost the bilateral TCD signal, but the remaining unilateral TCD signal was of sufficient quality to be used. The CBFV was higher in the young group compared with both elderly groups (P < 0.001), and the CVRi was lower in the young group compared with the older elderly group (P = 0.008). Below, we will briefly summarize these transfer function parameters (see Table 3). The induced oscillations led, as expected, to higher power spectrum density between baseline and the repeated C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
sit–stand manoeuvre (P < 0.001) for the three age groups (Tables 2 and 3). As with spontaneous oscillations, power spectrum density for blood pressure and CBFV appeared smaller in the older elderly group, but this was statistically significant only for CBFV (P < 0.05). Due to inclusion of data with coherence >0.4, some participants were excluded from the transfer function analysis of MAP–CBFV (n = 4) and CBFV–O2 Hb (n = 15) as indicated in Table 3. For the blood pressure–CBFV relationship, similar to the transfer function analysis for spontaneous oscillations, there were no differences between age groups for coherence and phase during induced VLF oscillations at 0.05 Hz. However, gain and normalized gain were higher in the elderly group compared with the older elderly and for normalized gain also compared with the young group (Table 3). The phase shift between CBFV and O2 Hb was lower in the elderly compared with the older elderly group (P = 0.003), and the normalized gain was marginally higher in the older elderly compared with the young group (P = 0.05). When performing a non-parametric Spearman’s ρ linear regression using all three groups combined (n = 58) with transfer function analysis metrics and age, phase CBFV–O2 Hb (r = −0.320, P = 0.034), normalized gain MAP–CBFV (r = 0.276, P = 0.041) and normalized gain CBFV–O2 Hb (r = 0.454, P = 0.002) were statistically significant, indicating a small decrease in phase (CBFV–O2 Hb) and increase in gain (MAP–CBFV and CBFV–O2 Hb) with advancing age. Single squat–stand manoeuvre
Table 4 and Fig. 1 present the results from the single squat–stand manoeuvre. Nine older elderly sujects
592
M. H. Oudegeest-Sander and others
Table 2. Transfer function analysis of mean arterial pressure–cerebral blood flow velocity was used to examine the dynamic cerebral autoregulation during spontaneous blood pressure oscillations in the sitting rest position Young
Elderly
Older elderly
15 20 20
15 17 14
13 15 17
Coherence VLF 0.53 ± 0.06 LF 0.77 ± 0.11∗b HF 0.72 ± 0.13
0.55 ± 0.09 0.68 ± 0.15 0.77 ± 0.13
0.52 ± 0.08 0.62 ± 0.09 0.70 ± 0.15
Phase (rad) VLF 0.99 ± 0.36∗a LF 0.49 ± 0.21 HF 0.07 ± 0.15
0.69 ± 0.28 0.47 ± 0.23 0.11 ± 0.11
0.74 ± 0.24 0.51 ± 0.20 –0.01 ± 0.14
Gain (cm s−1 mmHg−1 ) VLF 0.60 ± 0.15 LF 0.83 ± 0.18†a HF 0.91 ± 0.25
0.52 ± 0.15 0.68 ± 0.20 0.84 ± 0.19
0.50 ± 0.15 0.70 ± 0.15 0.85 ± 0.20
Normalized gain (%cm s−1 %mmHg−1 ) VLF 0.77 ± 0.29 0.93 ± 0.17 LF 1.16 ± 0.31 1.20 ± 0.31 1.50 ± 0.24 HF 1.24 ± 0.32∗b
0.90 ± 0.23 1.43 ± 0.54 1.65 ± 0.50
Power spectrum density for BP (mmHg2 Hz−1 ) VLF 8.16 ± 4.75 10.34 ± 6.54∗b LF 4.22 ± 3.26∗ab 2.28 ± 1.81 HF 0.62 ± 0.41 0.74 ± 0.70
4.64 ± 2.24 0.80 ± 0.50 0.37 ± 0.39
Coherence >0.4 VLF (n) LF (n) HF (n)
Power spectrum density for CBFV [(cm s−1 )2 Hz−1 ] 6.36 ± 4.32∗b 2.79 ± 2.10 VLF 6.41 ± 3.53∗b LF 3.54 ± 2.80∗ab 1.66 ± 1.23 0.54 ± 0.29 0.56 ± 0.47 0.28 ± 0.16 HF 0.73 ± 0.57∗b Abbreviations: BP, blood pressure; CBFV, cerebral blood flow velocity; HF, high frequency; LF, low frequency; and VLF, very low frequency. ∗ P < 0.05 compared with the elderlya or older elderlyb ; and † P = 0.05 compared with the elderlya .
were included in this analysis, because not all older elderly subjects could perform this manoeuvre owing to comorbidity, e.g. osteoarthritis of the hip or knee. The squat–stand manoeuvre induced large changes in blood pressure and CBFV, but overall there were no differences in relative changes in blood pressure and CBFV between age groups. Heart rate increased significantly more in the young group compared with both elderly groups (P < 0.001), but the absolute and percentage changes in MAP were similar in all groups. The CBFV was lower and CVRi was higher with increasing age during both squat and stand positions (P < 0.01). The fall in absolute CBFV upon standing was largest in young subjects (P = 0.013), but relative changes did not differ between age groups. The absolute decrease in CVRi with standing
Exp Physiol 99.3 (2014) pp 586–598
was larger in the elderly compared with the young group (P = 0.029). Cerebrovascular CO2 reactivity
Table 5 presents the results from measurements during hyperventilation and CO2 inhalation. In the older elderly group, 17 participants successfully completed this manoeuvre. Hyperventilation and CO2 inhalation led to significant changes in EtCO2 in all groups (P < 0.01), resulting in substantial changes in MAP, CBFV, CVRi and O2 Hb. Absolute CBFV was highest and CVRi lowest in the youngest age group throughout the measurements. Absolute changes in CBFV and CVRi during hyperventilation and CO2 inhalation were also highest in the young group. However, relative changes expressed as percentage changes from baseline were similar for all ages (Table 5). Discussion The aim of the present study was to examine the influence of different ages on dynamic cerebral autoregulation and cerebrovascular CO2 reactivity in combination with the assessment of cerebral cortical oxygenation. The main findings are that in individuals above the age of 70 years (age 74–86 years in this study), dynamic cerebral autoregulation and cerebrovascular CO2 reactivity retain normal function. Cerebral blood flow and cortical oxygenation were not compromised in these elderly subjects during manoeuvres that mimic daily life activities and during changes in arterial CO2 . Cerebral autoregulation
An intact dynamic cerebral autoregulation is characterized by a high-pass filter with decreasing phase and increasing gain with increasing frequency domains (Giller, 1990; Zhang et al. 1998). For spontaneous oscillations during baseline sitting, the transfer function parameters all followed this pattern in all three age groups, and all older elderly subjects adhered to the high-pass filter model. Moreover, the hallmark of dynamic cerebral autoregulation, i.e. the necessary adaptation of CVRi upon a decrease in blood pressure (e.g. when standing after squatting) in order to maintain sufficient CBF, was intact in older individuals. The present study showed that dynamic cerebral autoregulation in this oldest age group was comparable to that of a group of young subjects and to a group of older adults below the age of 70 years. Moreover, the efficiency of the dynamic cerebral autoregulatory response, as reflected by the phase of the transfer function analysis during the repeated sit–stand manoeuvre, was similar in the three age groups, indicating that the response C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
Exp Physiol 99.3 (2014) pp 586–598
593
Cerebral perfusion and cortical oxygenation in ageing
Table 3. Transfer function analyses of mean arterial blood pressure–cerebral blood flow velocity (MAP–CBFV) and cerebral blood flow velocity–oxygenated haemoglobin (CBFV–O2 Hb) were used to examine the dynamic cerebral autoregulation and cortical oxygenation during induced blood pressure oscillations using the repeated sit–stand manoeuvre at 0.05 Hz Young (n = 20) Haemodynamics MAP (mmHg) Heart rate (beats min−1 )
89 ± 18 79 ± 8
Elderly (n = 20)
Older elderly (n = 18)
91 ± 15 75 ± 10
88 ± 16 76 ± 9
Cerebral perfusion CBFV (cm s−1 ) CVRi (mmHg cm−1 s−1 )
59.7 ± 9.9∗ab 1.54 ± 0.47∗b
48.1 ± 8.3 1.91 ± 0.40
45.2 ± 9.3 2.06 ± 0.63
Dynamic cerebral autoregulation MAP–CBFV Coherence Phase (rad) Gain (cm s−1 mmHg−1 ) Normalized gain (%cm s−1 %mmHg−1 )
n = 19 0.63 ± 0.06 0.73 ± 0.33 0.62 ± 0.17 0.88 ± 0.18∗a
n = 19 0.67 ± 0.08 0.61 ± 0.21 0.70 ± 0.16∗b 1.31 ± 0.30∗b
n = 16 0.63 ± 0.11 0.80 ± 0.28 0.55 ± 0.20 1.06 ± 0.34
CBFV–O2 Hb Coherence Phase (rad) Gain (cm s−1 mmHg−1 ) Normalized gain (%cm s−1 %mmHg−1 )
n = 13 0.54 ± 0.09 −0.64 ± 0.31 0.07 ± 0.03 0.10 ± 0.09†b
n = 16 0.56 ± 0.08 −0.41 ± 0.49∗b 0.06 ± 0.02 0.12 ± 0.04
n = 14 0.50 ± 0.07 −0.89 ± 0.25 0.08 ± 0.02 0.17 ± 0.08
50.30 ± 22.27 27.25 ± 14.34∗b 0.23 ± 0.17
62.77 ± 33.25 34.67 ± 17.69∗b 029 ± 0.25
42.90 ± 18.20 15.02 ± 6.03 0.17 ± 0.09
Power spectrum density Blood pressure (mmHg2 Hz−1 ) CBFV [(cm s−1 )2 Hz−1 ] O2 Hb [(μmol l−1 )2 Hz−1 ]
Abbreviations: CBFV, cerebral blood flow velocity; CVRi, cerebrovascular resistance index; MAP, mean arterial pressure; and O2 Hb, oxygenated haemoglobin. ∗P < 0.05 compared with the elderlya or older elderlyb ; and † P = 0.05 compared with the older elderlyb .
time to restore the CBFV after blood pressure changes was comparable between the groups. There were important differences in baseline cerebral haemodynamics between the three age groups. The CBFV decreased and CVRi increased with advancing age. Given that there were no differences in blood pressure between the different age groups, the increased CVRi with ageing could be due either to reduced cerebral blood flow or to changes in the characteristics of the blood vessels. It is known that arterial walls become thicker and stiffer and vascular elasticity diminishes with ageing, thereby increasing vascular resistance (Lakatta & Levy, 2003; Monahan, 2007; O’Rourke & Hashimoto, 2007). Indeed, Gommer et al. (2012) found that higher peripheral vascular resistance correlated with higher CVRi. In the present study, we did not measure absolute cerebral blood flow (in millilitres per minute per 100 g of brain tissue) and we are therefore unable to differentiate between these two explanations for the increased CVRi. The transfer function parameters gain and normalized gain, in contrast, are also thought to be affected by vascular stiffness. Despite the increase in CVRi with age, there was no consistent difference in (normalized) gain with increasing age, because gain was not higher in the older elderly compared with the younger elderly subjects. Nevertheless, linear C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
regression did show a significant increase for normalized gain MAP–CBFV and CBFV–O2 Hb with age. Taken together, we postulate that dynamic cerebral autoregulation is not compromised by the effects of physiological ageing up to the age of 86 years.
Cerebrovascular CO2 reactivity
The relative change in CBFV during changes in EtCO2 was similar in the three age groups. In other words, the ability to vasoconstrict or vasodilate in response to changes in arterial CO2 is preserved with ageing. This finding is in contrast with literature showing that young subjects have a greater cerebrovascular CO2 reactivity compared with older subjects (Bakker et al. 2004). Cerebrovascular CO2 reactivity can be affected by structural vascular properties as well as vascular endothelial function (Diehl et al. 1991; Seals et al. 2011). Vascular endothelium is an important regulatory organ that modulates the tone of the underlying vascular smooth muscle cells and maintains cardiovascular homeostasis by contributing to vasoconstriction and vasodilatation (Glasser et al. 1996). It is known that the endothelial function is reduced with ageing (Seals et al. 2011). Despite these potential vascular
594
M. H. Oudegeest-Sander and others
Exp Physiol 99.3 (2014) pp 586–598
Table 4. The single squat–stand manoeuvre was used to examine the cerebral perfusion and cortical oxygenation during an orthostatic hypotension challenge test in healthy subjects, mimicking daily life activities Young (n = 20)
Elderly (n = 20)
Older elderly (n = 9)
Squat MAP (mmHg) Heart rate (beats min−1 ) CBFV (cm s−1 ) CVRi (mmHg cm−1 s−1 )
87 75 64.4 1.40
± ± ± ±
16 10 11.0∗ab 0.44∗ab
96 72 50.5 1.99
± ± ± ±
16 12∗b 9.4 0.46
102 83 46.0 2.29
± ± ± ±
14 8 8.4 0.60
Stand MAP (mmHg) Heart rate (beats min−1 ) CBFV (cm s−1 ) CVRi (mmHg cm−1 s−1 )
58 104 47.7 1.29
± ± ± ±
15 10∗ab 10.1∗ab 0.58∗b
61 84 39.2 1.62
± ± ± ±
10 13 6.7 0.28∗b
69 88 33.3 2.11
± ± ± ±
13 8 5.5 0.54
−29 −34 29 40 −16.7 −26.2 −0.11 −2.22
± ± ± ± ± ± ± ±
7 8 10∗ab 18∗ab 5.2∗a 7.5 0.20∗a 1.04
−35 −36 11 16 −11.3 −21.5 −0.37 −2.72
± ± ± ± ± ± ± ±
14 10 6 10 6.6 10.7 0.35 1.42
−33 −33 5 6 −12.7 −27.2 −0.18 −2.98
± ± ± ± ± ± ± ±
9 7 3 4 4.2 6.0 0.39 1.06
Haemodynamic response MAP (mmHg) %MAP (mmHg) Heart rate (beats min−1 ) %Heart rate (beats min−1 ) CBFV (cm s−1 ) %CBFV (cm s−1 ) CVRi (mmHg cm−1 s−1 ) O2 Hb (μmol l−1 )
Abbreviations: CBFV, cerebral blood flow velocity; CVRi, cerebrovascular resistance index; MAP, mean arterial pressure; O2 Hb, oxygenated haemoglobin; , standing minus squatting; and %, percentage change between standing and squatting compared with the squat values. ∗P < 0.05 compared with the elderlya or older elderlyb .
changes with ageing, the percentage changes in CBFV and O2 Hb in response to changes in CO2 were similar for all groups.
Cortical oxygenation
The dynamic relationship between oscillations in CBFV and cortical O2 Hb may reflect a regulatory mechanism to maintain brain tissue oxygenation (Rowley et al. 2007; Iadecola et al. 2009). The relationship between blood pressure, CBFV and O2 Hb lends support to the notion that upstream oscillations in CBFV induced by changes in blood pressure make an important contribution to the downstream brain tissue level oscillations in O2 Hb (van Beek et al. 2012). It must be acknowledged that NIRS also measures O2 Hb in venous blood as well as in the arterial blood supply. However, the NIRS method has been validated to measure changes at the cerebral cortical tissue level, and the dynamic changes in O2 Hb have been related to changes in the arterial and arteriolar component, with little change in the venous compartment, which remains relatively constant (Smielewski et al. 1995; Brady et al. 2007). Near-infrared spectroscopy can be influenced by blood in the extracranial tissue (skin); however, the interoptode distance of 5.0 cm that we used has been shown to measure intracranial cerebral haemodynamic
changes in O2 Hb and HHb with negligible influence of skin blood flow (Reinhard et al. 2006). In the present study, repeated sit–stand manoeuvres induced strong blood pressure oscillations, which resulted in strong CBFV oscillations that led to O2 Hb oscillations with sufficient coherence to perform transfer function analysis of the relationship between CBFV and O2 Hb. The phase between CBFV and O2 Hb was negative in all three groups, indicating that O2 Hb oscillations follow those of CBFV with a time delay caused by the transit time of the blood. The gain of the transfer function analysis between CBFV and O2 Hb might reflect damping of the CBFV fluctuations (perhaps by distensible arterioles or capillaries), resulting in smaller or larger oscillations in cortical oxygenation. Gain may, however, also be influenced by metabolic reserve and its effect on oxygenated haemoglobin, or by a reduced diffusion of oxygen, leading to enhanced oscillations in O2 Hb in response to changes in CBFV (van Beek et al. 2012). With increasing age, the phase lag between CBFV and O2 Hb became larger, whereas the normalized gain became higher. This is in line with results from patients with Alzheimer’s disease, a disease in which cerebral microvasculature is affected. In the Alzheimer’s disease patients, the phase lag was larger and gain higher between CBFV and O2 Hb at 0.05 Hz compared with age-matched control subjects (van Beek et al. 2012). It is possible that the microvasculature in older individuals becomes C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
Exp Physiol 99.3 (2014) pp 586–598
Cerebral perfusion and cortical oxygenation in ageing
compromised compared with young individuals. We speculate that increased microvascular tortuosity (which leads to increased vessel length and thereby transit time, as has been demonstrated in arterial spin labelling studies) could explain the larger phase shift and that reduced distensibility and concomitant reduced damping could explain larger gain.
595
The single squat–stand manoeuvre elicited a reduction in O2 Hb in the frontal cortex, which was similar in all groups. This challenging manoeuvre, which is comparable with activities in daily life, such as standing up from a chair or putting on a sock or shoe, does not seem to compromise oxygenation of the frontal cortex in healthy older individuals. This would confirm our observation of preserved dynamic cerebral autoregulation based on TCD in these healthy subjects. These compensatory mechanisms may, however, fail in the event of more severe stress (for example, during septic shock) or with diseases that compromise cerebral macro- and microvasculature further, such as Alzheimer’s disease. Limitations
Figure 1. Blood pressure (A), cerebral blood flow velocity (CBFV; B) and cortical oxygenation (oxygenated haemoglobin, O2 Hb; C) responses of the three groups during the single squat–stand manoeuvre The vertical dashed line indicates the moment of standing after 1 min of squatting. The continuous black line represents the young, the continuous grey line the elderly, and the dotted grey line the older elderly group. C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
In order to induce larger haemodynamic oscillations, a repeated sit–stand manoeuvre was applied. The induced large blood pressure oscillations during the repeated sit–stand manoeuvre provide strong and physiologically relevant haemodynamic perturbations and lead to an improved estimation of the transfer function parameters and higher power spectrum density, thereby increasing the reliability of the assessment (Claassen et al. 2009; van Beek et al. 2010a). However, the manoeuvre could be seen as aerobic exercise and may therefore affect our measurements. However, Ogoh et al. (2005) found that mild- to moderate-intensity exercise did not impair dynamic cerebral autoregulation. It is possible that the older elderly subjects experienced the repeated sit–stand manoeuvre as mild exercise; however, compared with baseline values their heart rates did not increase more than in the other two groups, nor did the EtCO2 change significantly compared with baseline. Thus, confounding effects of aerobic exercise during the repeated sit–stand manoeuvre are unlikely. Cerebral blood flow velocity was measured in the middle cerebral artery while NIRS was obtained from the frontal cortex, which receives blood supply from the anterior cerebral artery and the middle cerebral artery (van Beek et al. 2012). We assumed that relative changes of CBFV would be similar between the middle and anterior cerebral arteries in response to blood pressure changes and that the supplying anterior cerebral artery would be likely to react in a similar manner to the middle cerebral artery. The cut-off point of coherence of
Exp Physiol 99.3 (2014) pp 586–598
Research Paper
Assessment of dynamic cerebral autoregulation and cerebrovascular CO2 reactivity in ageing by measurements of cerebral blood flow and cortical oxygenation Madelijn H. Oudegeest-Sander1,2,∗ , Arenda H. E. A. van Beek1,∗ , Karin Abbink1,2 , Marcel G. M. Olde Rikkert1 , Maria T. E. Hopman2 and Jurgen A. H. R. Claassen1 Departments of 1 Geriatric Medicine and 2 Physiology, Radboud University Medical Center, Nijmegen, The Netherlands
Experimental Physiology
New Findings r What is the central question in this study? It is unknown to what extent increasing age influences the dynamic adaptations of cerebral blood flow velocity and cortical oxygenation in response to changes in blood pressure (cerebral autoregulation) and to changes in carbon dioxide (cerebrovascular CO2 reactivity). r What is the main finding and its importance? We have shown that ageing up to 86 years is associated with an overall preservation of dynamic cerebral autoregulation and cerebrovascular CO2 reactivity, leading to a sufficiency of cerebral cortical oxygenation during daily life activities, despite the decrease in absolute cerebral blood flow velocity and increase in cerebrovascular resistance with advancing age.
With ageing, cerebral blood flow velocity (CBFV) decreases; however, to what extent dynamic cerebral autoregulation and cerebrovascular CO2 reactivity are influenced by ageing is unknown. The aim was to examine the dynamic responses of CBFV and cortical oxygenation to changes in blood pressure (BP) and arterial CO2 across different ages. Fifty-eight participants in three age groups were included, as follows: young (n = 20, 24 ± 2 years old), elderly (n = 20, 66 ± 1 years old), and older elderly (n = 18, 78 ± 3 years old). The CBFV was measured using transcranial Doppler ultrasound, simultaneously with oxyhaemoglobin (O2 Hb) using near-infrared spectroscopy and beat-to-beat BP measurements using Finapres. Postural manoeuvres were performed to induce haemodynamic fluctuations. Cerebrovascular CO2 reactivity was tested with hyperventilation and CO2 inhalation. With age, CBFV decreased (young 59 ± 12 cm s−1 , elderly 48 ± 7 cm s−1 and older elderly 42 ± 9 cm s−1 , P < 0.05) and cerebrovascular resistance increased (1.46 ± 0.58, 1.81 ± 0.36 and 1.98 ± 0.52 mmHg cm−1 s−1 , respectively, P < 0.05). Normalized gain (autoregulatory damping) increased with age for BP–CBFV (0.88 ± 0.18, 1.31 ± 0.30 and 1.06 ± 0.34, respectively, P < 0.05) and CBFV–O2 Hb (0.10 ± 0.09, 0.12 ± 0.04 and 0.17 ± 0.08, respectively, P < 0.05) during the repeated sit–stand manoeuvre at 0.05 Hz. Even though the absolute changes in CBFV and cerebrovascular resistance index during the cerebrovascular CO2 reactivity were higher in the young group, the percentage changes in CBFV, cerebrovascular resistance index and O2 Hb were similar in all age groups. In conclusion, there was no decline in dynamic cerebral autoregulation and cerebrovascular CO2 reactivity with increasing age up to 86 years. Despite the decrease in cerebral blood flow
∗
M. H. Oudegeest-Sander and A. H. E. A. van Beek contributed equally to this work.
DOI: 10.1113/expphysiol.2013.076455
C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
Exp Physiol 99.3 (2014) pp 586–598
Cerebral perfusion and cortical oxygenation in ageing
587
velocity and increase in cerebrovascular resistance with advancing age, CBFV and cortical oxygenation were not compromised in these elderly humans during manoeuvres that mimic daily life activities. (Received 27 September 2013; accepted after revision 16 December 2013; first published online 20 December 2013) Corresponding author J. A. H. R. Claassen: Department of Geriatric Medicine (925), Radboud University Medical Center, PO Box 9101, 6500 HB Nijmegen, The Netherlands. Email: [email protected]
Introduction Physiological ageing is associated with well-recognized changes in the systemic and cerebral vasculature (Ferrari et al. 2003; Lakatta & Levy, 2003; O’Rourke & Hashimoto, 2007). Systemic vascular changes lead to an increased risk for cardiovascular events, while cerebrovascular adaptations enhance the risk of (ischaemic) stroke. Furthermore, resting cerebral blood flow (CBF) is decreased in older individuals (Lipsitz et al. 2000; Sorond et al. 2005). Whether this decrease in CBF reflects a decrease in cerebral demand or a true diminution in blood flow remains unknown; likewise, whether ageing leads to diminished cerebrovascular function remains unclear. Cerebrovascular function can be studied using two physiological phenomena, namely dynamic cerebral autoregulation and cerebrovascular reactivity to carbon dioxide. Dynamic cerebral autoregulation reflects the ability of the brain to stabilize CBF despite rapid daily life changes in blood pressure (Paulson et al. 1990; Panerai, 1998; van Beek et al. 2008). Dynamic cerebral autoregulation remains stable up to the age of 73 years (Lipsitz et al. 2000; Sorond et al. 2005; van Beek et al. 2008) but has not been estimated in healthy elderly subjects above this age. Most studies examined the effects of small blood pressure changes (Diehl et al. 1991, 1998) instead of larger fluctuations in blood pressure during postural changes, which may provide a stronger challenge for dynamic cerebral autoregulation and may therefore be more sensitive to detect changes in cerebrovascular function (Claassen et al. 2009). Cerebrovascular reactivity to changes in CO2 , the second phenomenon that indicates cerebrovascular function, declines significantly with increasing age up to 90 years (Bakker et al. 2004). Unfortunately, most studies investigated either the dynamic cerebral autoregulation (Narayanan et al. 2001; Sorond et al. 2005; Claassen et al. 2009; van Beek et al. 2010a) or the cerebrovascular CO2 reactivity (Ringelstein et al. 1988; Bakker et al. 2004; Claassen et al. 2007; Dineen et al. 2011; van Beek et al. 2011), but the studies did not examine both aspects of the cerebrovascular system simultaneously. Dynamic cerebral autoregulation and cerebrovascular CO2 reactivity are both directly related to the oxygenation of the brain tissue and therefore to human brain function (van Beek et al. 2010b, 2012). The majority of studies C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
of dynamic cerebral autoregulation and cerebrovascular CO2 reactivity have used transcranial Doppler ultrasound (TCD) to obtain continuous measurements of cerebral blood flow velocity (CBFV), together with continuous blood pressure measurements (Finapres). Transcranial Doppler ultrasound does not provide information regarding the oxygenation of the brain. The addition of near-infrared spectroscopy (NIRS) can be used to measure concentrations of oxygenated and deoxygenated haemoglobin and provide estimates of cortical brain tissue oxygenation (Reinhard et al. 2006). Thus, by using Finapres and transcranial Doppler sonography concomitantly with near-infrared spectroscopy, it is possible to determine two different but dependent dynamic relationships, namely the relationship between blood pressure and cerebral blood flow on the one hand and the relationship between cerebral blood flow and cerebral cortical oxygenation on the other (van Beek et al. 2010b, 2012). This study aimed to investigate effects of age on dynamic cerebral autoregulation and cerebrovascular CO2 reactivity in order to examine cerebrovascular function by investigating the relationship between blood pressure, cerebral blood flow and cerebral cortical oxygenation in a group of young, elderly and older elderly subjects. The combination of examining dynamic cerebral autoregulation and cerebrovascular CO2 reactivity using both TCD and NIRS in this population is, to our knowledge, unique. We hypothesized that the cerebrovascular function, and thereby the cortical oxygenation, are diminished in healthy elderly aged 74 years and older compared with young subjects and with elderly subjects below the age of 74 years. To investigate this hypothesis, beat-to-beat blood pressure, cerebral blood flow and cortical oxygenation were measured simultaneously to investigate cerebrovascular haemodynamics during changes in blood pressure and changes in arterial CO2 .
Methods Ethical approval
The study was approved by the Medical Ethics Committee, and was performed according to the Declaration of
588
M. H. Oudegeest-Sander and others
Helsinki and Good Clinical Practice guidelines. All participants gave written informed consent. Participants
Fifty-eight participants in the following three different age groups were included: a young group (young, mean age 24 ± 2 years, range 21–28 years), an older group (elderly, mean age 66 ± 1 years, range 65–69 years) and the oldest group (older elderly, mean age 78 ± 3 years, range 74–86 years). Participants were recruited at the Departments of Physiology and Geriatric Medicine, Radboud University Medical Center, Nijmegen, The Netherlands. All participants were extensively screened. Participants with hypertension (>160/90 mmHg), diabetes mellitus reported in the medical history (fasting glucose >6.9 mmol l−1 ), hypercholesterolaemia (total cholesterol >7.5 mmol l−1 ), body mass index >32.5 kg m−2 , compromised cognition (Mini Mental State Examination 0.4 in this study (Zhang et al. 1998; Claassen et al. 2009; van Beek et al. 2010a). Repeated sit–stand manoeuvre at 0.05 Hz. The repeated sit–stand manoeuvre causes oscillations that are much larger than the spontaneous oscillations that occur during the baseline sitting rest position; therefore, the dynamic relationship can be assessed with greater coherence. Transfer function analysis was performed in the very low frequency range, 0.02–0.07 Hz. The dynamic relationships between blood pressure as input and CBFV as output (MAP–CBFV) and CBFV–O2 Hb were determined. Data were also included only if coherence was >0.4. We have previously shown that upstream oscillations in CBFV induced by changes in blood pressure make an important contribution to the downstream brain tissue level oscillations in O2 Hb (van Beek et al. 2012). It is our experience that oscillations induced at a frequency of 0.05 Hz result in clear blood pressure and CBFV oscillations with good coherence, whereas oscillations induced at a frequency of 0.1 Hz result in more noise and thereby loss of reliable data (van Beek et al. 2010a, 2012). Furthermore, in accordance with the high-pass filter, dynamic cerebral autoregulation is mainly active in the low and very low frequency regions. Single squat–stand manoeuvre. For the single squat–stand manoeuvre we analysed two time periods: (1) squatting; and (2) maximal decline after standing up from squatting. For this, we used 30 s of steady-state measurements in the squatting position (1) and the period of five consecutive seconds in which MAP and CBFV had decreased the most upon standing (2). Both absolute and relative changes in MAP and CBFV were calculated. Cerebrovascular CO2 reactivity. Regarding the analysis of the cerebrovascular CO2 reactivity, absolute and relative changes in heart rate, MAP, CBFV, CVRi and O2 Hb were calculated. The full range of cerebrovascular
Exp Physiol 99.3 (2014) pp 586–598
CO2 reactivity was calculated (Ringelstein et al. 1988), which represents the percentage change in CBFV from hypocapnia (hyperventilation) to hypercapnia (inhalation of 7% CO2 ) relative to baseline CBF = {[CBFVmax (7% CO2 ) − CBFVmin (hyperventilation)]/CBFVrest }/100%. We have also calculated the change in CBFV between hypocapnia and hypercapnia per millimetre of mercury change in EtCO2 between hypocapnia and hypercapnia (Bailey et al. 2013). Statistical analysis
Data are presented as means ± SD. Comparisons between the different age groups were performed using one-way ANOVA with Bonferroni correction for multiple comparisons. We checked the data with the Hochberg method because of sample size differences (especially with the single squat–stand manoeuvre) and with the Games–Howell method because of different group variances. As these three methods did not show relevant differences, we mention only the results from the one-way ANOVA with Bonferroni correction. Statistical significance was set at a P value of 0.4 were used. There were no relevant differences in coherence results between age groups. The older elderly had a slightly lower coherence in the LF compared with the young group (P = 0.001), but mean coherence remained well above 0.4. Values for phase were very similar across age groups, but young subjects had a higher mean phase in the VLF (where cerebral autoregulation is most active); this difference reached statistical significance when compared with younger elderly subjects (P = 0.023). However, VLF phase did not differ between younger and older elderly groups. Table 2 lists absolute gain, which is sensitive to individual differences in CVFV or blood pressure, and normalized gain, which is not affected by absolute CBFV or blood pressure. Absolute gain was marginally higher in the LF compared with the elderly (P = 0.05). Normalized gain revealed no differences between age groups in the VLF and LF range where cerebral autoregulation is active; however, normalized gain was higher in the HF in the older elderly compared with the young group (P = 0.006). Repeated sit–stand manoeuvre at 0.05 Hz
The results from the repeated sit–stand manoeuvre are presented in Table 3. Two subjects (one in each elderly group) lost the bilateral TCD signal, but the remaining unilateral TCD signal was of sufficient quality to be used. The CBFV was higher in the young group compared with both elderly groups (P < 0.001), and the CVRi was lower in the young group compared with the older elderly group (P = 0.008). Below, we will briefly summarize these transfer function parameters (see Table 3). The induced oscillations led, as expected, to higher power spectrum density between baseline and the repeated C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
sit–stand manoeuvre (P < 0.001) for the three age groups (Tables 2 and 3). As with spontaneous oscillations, power spectrum density for blood pressure and CBFV appeared smaller in the older elderly group, but this was statistically significant only for CBFV (P < 0.05). Due to inclusion of data with coherence >0.4, some participants were excluded from the transfer function analysis of MAP–CBFV (n = 4) and CBFV–O2 Hb (n = 15) as indicated in Table 3. For the blood pressure–CBFV relationship, similar to the transfer function analysis for spontaneous oscillations, there were no differences between age groups for coherence and phase during induced VLF oscillations at 0.05 Hz. However, gain and normalized gain were higher in the elderly group compared with the older elderly and for normalized gain also compared with the young group (Table 3). The phase shift between CBFV and O2 Hb was lower in the elderly compared with the older elderly group (P = 0.003), and the normalized gain was marginally higher in the older elderly compared with the young group (P = 0.05). When performing a non-parametric Spearman’s ρ linear regression using all three groups combined (n = 58) with transfer function analysis metrics and age, phase CBFV–O2 Hb (r = −0.320, P = 0.034), normalized gain MAP–CBFV (r = 0.276, P = 0.041) and normalized gain CBFV–O2 Hb (r = 0.454, P = 0.002) were statistically significant, indicating a small decrease in phase (CBFV–O2 Hb) and increase in gain (MAP–CBFV and CBFV–O2 Hb) with advancing age. Single squat–stand manoeuvre
Table 4 and Fig. 1 present the results from the single squat–stand manoeuvre. Nine older elderly sujects
592
M. H. Oudegeest-Sander and others
Table 2. Transfer function analysis of mean arterial pressure–cerebral blood flow velocity was used to examine the dynamic cerebral autoregulation during spontaneous blood pressure oscillations in the sitting rest position Young
Elderly
Older elderly
15 20 20
15 17 14
13 15 17
Coherence VLF 0.53 ± 0.06 LF 0.77 ± 0.11∗b HF 0.72 ± 0.13
0.55 ± 0.09 0.68 ± 0.15 0.77 ± 0.13
0.52 ± 0.08 0.62 ± 0.09 0.70 ± 0.15
Phase (rad) VLF 0.99 ± 0.36∗a LF 0.49 ± 0.21 HF 0.07 ± 0.15
0.69 ± 0.28 0.47 ± 0.23 0.11 ± 0.11
0.74 ± 0.24 0.51 ± 0.20 –0.01 ± 0.14
Gain (cm s−1 mmHg−1 ) VLF 0.60 ± 0.15 LF 0.83 ± 0.18†a HF 0.91 ± 0.25
0.52 ± 0.15 0.68 ± 0.20 0.84 ± 0.19
0.50 ± 0.15 0.70 ± 0.15 0.85 ± 0.20
Normalized gain (%cm s−1 %mmHg−1 ) VLF 0.77 ± 0.29 0.93 ± 0.17 LF 1.16 ± 0.31 1.20 ± 0.31 1.50 ± 0.24 HF 1.24 ± 0.32∗b
0.90 ± 0.23 1.43 ± 0.54 1.65 ± 0.50
Power spectrum density for BP (mmHg2 Hz−1 ) VLF 8.16 ± 4.75 10.34 ± 6.54∗b LF 4.22 ± 3.26∗ab 2.28 ± 1.81 HF 0.62 ± 0.41 0.74 ± 0.70
4.64 ± 2.24 0.80 ± 0.50 0.37 ± 0.39
Coherence >0.4 VLF (n) LF (n) HF (n)
Power spectrum density for CBFV [(cm s−1 )2 Hz−1 ] 6.36 ± 4.32∗b 2.79 ± 2.10 VLF 6.41 ± 3.53∗b LF 3.54 ± 2.80∗ab 1.66 ± 1.23 0.54 ± 0.29 0.56 ± 0.47 0.28 ± 0.16 HF 0.73 ± 0.57∗b Abbreviations: BP, blood pressure; CBFV, cerebral blood flow velocity; HF, high frequency; LF, low frequency; and VLF, very low frequency. ∗ P < 0.05 compared with the elderlya or older elderlyb ; and † P = 0.05 compared with the elderlya .
were included in this analysis, because not all older elderly subjects could perform this manoeuvre owing to comorbidity, e.g. osteoarthritis of the hip or knee. The squat–stand manoeuvre induced large changes in blood pressure and CBFV, but overall there were no differences in relative changes in blood pressure and CBFV between age groups. Heart rate increased significantly more in the young group compared with both elderly groups (P < 0.001), but the absolute and percentage changes in MAP were similar in all groups. The CBFV was lower and CVRi was higher with increasing age during both squat and stand positions (P < 0.01). The fall in absolute CBFV upon standing was largest in young subjects (P = 0.013), but relative changes did not differ between age groups. The absolute decrease in CVRi with standing
Exp Physiol 99.3 (2014) pp 586–598
was larger in the elderly compared with the young group (P = 0.029). Cerebrovascular CO2 reactivity
Table 5 presents the results from measurements during hyperventilation and CO2 inhalation. In the older elderly group, 17 participants successfully completed this manoeuvre. Hyperventilation and CO2 inhalation led to significant changes in EtCO2 in all groups (P < 0.01), resulting in substantial changes in MAP, CBFV, CVRi and O2 Hb. Absolute CBFV was highest and CVRi lowest in the youngest age group throughout the measurements. Absolute changes in CBFV and CVRi during hyperventilation and CO2 inhalation were also highest in the young group. However, relative changes expressed as percentage changes from baseline were similar for all ages (Table 5). Discussion The aim of the present study was to examine the influence of different ages on dynamic cerebral autoregulation and cerebrovascular CO2 reactivity in combination with the assessment of cerebral cortical oxygenation. The main findings are that in individuals above the age of 70 years (age 74–86 years in this study), dynamic cerebral autoregulation and cerebrovascular CO2 reactivity retain normal function. Cerebral blood flow and cortical oxygenation were not compromised in these elderly subjects during manoeuvres that mimic daily life activities and during changes in arterial CO2 . Cerebral autoregulation
An intact dynamic cerebral autoregulation is characterized by a high-pass filter with decreasing phase and increasing gain with increasing frequency domains (Giller, 1990; Zhang et al. 1998). For spontaneous oscillations during baseline sitting, the transfer function parameters all followed this pattern in all three age groups, and all older elderly subjects adhered to the high-pass filter model. Moreover, the hallmark of dynamic cerebral autoregulation, i.e. the necessary adaptation of CVRi upon a decrease in blood pressure (e.g. when standing after squatting) in order to maintain sufficient CBF, was intact in older individuals. The present study showed that dynamic cerebral autoregulation in this oldest age group was comparable to that of a group of young subjects and to a group of older adults below the age of 70 years. Moreover, the efficiency of the dynamic cerebral autoregulatory response, as reflected by the phase of the transfer function analysis during the repeated sit–stand manoeuvre, was similar in the three age groups, indicating that the response C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
Exp Physiol 99.3 (2014) pp 586–598
593
Cerebral perfusion and cortical oxygenation in ageing
Table 3. Transfer function analyses of mean arterial blood pressure–cerebral blood flow velocity (MAP–CBFV) and cerebral blood flow velocity–oxygenated haemoglobin (CBFV–O2 Hb) were used to examine the dynamic cerebral autoregulation and cortical oxygenation during induced blood pressure oscillations using the repeated sit–stand manoeuvre at 0.05 Hz Young (n = 20) Haemodynamics MAP (mmHg) Heart rate (beats min−1 )
89 ± 18 79 ± 8
Elderly (n = 20)
Older elderly (n = 18)
91 ± 15 75 ± 10
88 ± 16 76 ± 9
Cerebral perfusion CBFV (cm s−1 ) CVRi (mmHg cm−1 s−1 )
59.7 ± 9.9∗ab 1.54 ± 0.47∗b
48.1 ± 8.3 1.91 ± 0.40
45.2 ± 9.3 2.06 ± 0.63
Dynamic cerebral autoregulation MAP–CBFV Coherence Phase (rad) Gain (cm s−1 mmHg−1 ) Normalized gain (%cm s−1 %mmHg−1 )
n = 19 0.63 ± 0.06 0.73 ± 0.33 0.62 ± 0.17 0.88 ± 0.18∗a
n = 19 0.67 ± 0.08 0.61 ± 0.21 0.70 ± 0.16∗b 1.31 ± 0.30∗b
n = 16 0.63 ± 0.11 0.80 ± 0.28 0.55 ± 0.20 1.06 ± 0.34
CBFV–O2 Hb Coherence Phase (rad) Gain (cm s−1 mmHg−1 ) Normalized gain (%cm s−1 %mmHg−1 )
n = 13 0.54 ± 0.09 −0.64 ± 0.31 0.07 ± 0.03 0.10 ± 0.09†b
n = 16 0.56 ± 0.08 −0.41 ± 0.49∗b 0.06 ± 0.02 0.12 ± 0.04
n = 14 0.50 ± 0.07 −0.89 ± 0.25 0.08 ± 0.02 0.17 ± 0.08
50.30 ± 22.27 27.25 ± 14.34∗b 0.23 ± 0.17
62.77 ± 33.25 34.67 ± 17.69∗b 029 ± 0.25
42.90 ± 18.20 15.02 ± 6.03 0.17 ± 0.09
Power spectrum density Blood pressure (mmHg2 Hz−1 ) CBFV [(cm s−1 )2 Hz−1 ] O2 Hb [(μmol l−1 )2 Hz−1 ]
Abbreviations: CBFV, cerebral blood flow velocity; CVRi, cerebrovascular resistance index; MAP, mean arterial pressure; and O2 Hb, oxygenated haemoglobin. ∗P < 0.05 compared with the elderlya or older elderlyb ; and † P = 0.05 compared with the older elderlyb .
time to restore the CBFV after blood pressure changes was comparable between the groups. There were important differences in baseline cerebral haemodynamics between the three age groups. The CBFV decreased and CVRi increased with advancing age. Given that there were no differences in blood pressure between the different age groups, the increased CVRi with ageing could be due either to reduced cerebral blood flow or to changes in the characteristics of the blood vessels. It is known that arterial walls become thicker and stiffer and vascular elasticity diminishes with ageing, thereby increasing vascular resistance (Lakatta & Levy, 2003; Monahan, 2007; O’Rourke & Hashimoto, 2007). Indeed, Gommer et al. (2012) found that higher peripheral vascular resistance correlated with higher CVRi. In the present study, we did not measure absolute cerebral blood flow (in millilitres per minute per 100 g of brain tissue) and we are therefore unable to differentiate between these two explanations for the increased CVRi. The transfer function parameters gain and normalized gain, in contrast, are also thought to be affected by vascular stiffness. Despite the increase in CVRi with age, there was no consistent difference in (normalized) gain with increasing age, because gain was not higher in the older elderly compared with the younger elderly subjects. Nevertheless, linear C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
regression did show a significant increase for normalized gain MAP–CBFV and CBFV–O2 Hb with age. Taken together, we postulate that dynamic cerebral autoregulation is not compromised by the effects of physiological ageing up to the age of 86 years.
Cerebrovascular CO2 reactivity
The relative change in CBFV during changes in EtCO2 was similar in the three age groups. In other words, the ability to vasoconstrict or vasodilate in response to changes in arterial CO2 is preserved with ageing. This finding is in contrast with literature showing that young subjects have a greater cerebrovascular CO2 reactivity compared with older subjects (Bakker et al. 2004). Cerebrovascular CO2 reactivity can be affected by structural vascular properties as well as vascular endothelial function (Diehl et al. 1991; Seals et al. 2011). Vascular endothelium is an important regulatory organ that modulates the tone of the underlying vascular smooth muscle cells and maintains cardiovascular homeostasis by contributing to vasoconstriction and vasodilatation (Glasser et al. 1996). It is known that the endothelial function is reduced with ageing (Seals et al. 2011). Despite these potential vascular
594
M. H. Oudegeest-Sander and others
Exp Physiol 99.3 (2014) pp 586–598
Table 4. The single squat–stand manoeuvre was used to examine the cerebral perfusion and cortical oxygenation during an orthostatic hypotension challenge test in healthy subjects, mimicking daily life activities Young (n = 20)
Elderly (n = 20)
Older elderly (n = 9)
Squat MAP (mmHg) Heart rate (beats min−1 ) CBFV (cm s−1 ) CVRi (mmHg cm−1 s−1 )
87 75 64.4 1.40
± ± ± ±
16 10 11.0∗ab 0.44∗ab
96 72 50.5 1.99
± ± ± ±
16 12∗b 9.4 0.46
102 83 46.0 2.29
± ± ± ±
14 8 8.4 0.60
Stand MAP (mmHg) Heart rate (beats min−1 ) CBFV (cm s−1 ) CVRi (mmHg cm−1 s−1 )
58 104 47.7 1.29
± ± ± ±
15 10∗ab 10.1∗ab 0.58∗b
61 84 39.2 1.62
± ± ± ±
10 13 6.7 0.28∗b
69 88 33.3 2.11
± ± ± ±
13 8 5.5 0.54
−29 −34 29 40 −16.7 −26.2 −0.11 −2.22
± ± ± ± ± ± ± ±
7 8 10∗ab 18∗ab 5.2∗a 7.5 0.20∗a 1.04
−35 −36 11 16 −11.3 −21.5 −0.37 −2.72
± ± ± ± ± ± ± ±
14 10 6 10 6.6 10.7 0.35 1.42
−33 −33 5 6 −12.7 −27.2 −0.18 −2.98
± ± ± ± ± ± ± ±
9 7 3 4 4.2 6.0 0.39 1.06
Haemodynamic response MAP (mmHg) %MAP (mmHg) Heart rate (beats min−1 ) %Heart rate (beats min−1 ) CBFV (cm s−1 ) %CBFV (cm s−1 ) CVRi (mmHg cm−1 s−1 ) O2 Hb (μmol l−1 )
Abbreviations: CBFV, cerebral blood flow velocity; CVRi, cerebrovascular resistance index; MAP, mean arterial pressure; O2 Hb, oxygenated haemoglobin; , standing minus squatting; and %, percentage change between standing and squatting compared with the squat values. ∗P < 0.05 compared with the elderlya or older elderlyb .
changes with ageing, the percentage changes in CBFV and O2 Hb in response to changes in CO2 were similar for all groups.
Cortical oxygenation
The dynamic relationship between oscillations in CBFV and cortical O2 Hb may reflect a regulatory mechanism to maintain brain tissue oxygenation (Rowley et al. 2007; Iadecola et al. 2009). The relationship between blood pressure, CBFV and O2 Hb lends support to the notion that upstream oscillations in CBFV induced by changes in blood pressure make an important contribution to the downstream brain tissue level oscillations in O2 Hb (van Beek et al. 2012). It must be acknowledged that NIRS also measures O2 Hb in venous blood as well as in the arterial blood supply. However, the NIRS method has been validated to measure changes at the cerebral cortical tissue level, and the dynamic changes in O2 Hb have been related to changes in the arterial and arteriolar component, with little change in the venous compartment, which remains relatively constant (Smielewski et al. 1995; Brady et al. 2007). Near-infrared spectroscopy can be influenced by blood in the extracranial tissue (skin); however, the interoptode distance of 5.0 cm that we used has been shown to measure intracranial cerebral haemodynamic
changes in O2 Hb and HHb with negligible influence of skin blood flow (Reinhard et al. 2006). In the present study, repeated sit–stand manoeuvres induced strong blood pressure oscillations, which resulted in strong CBFV oscillations that led to O2 Hb oscillations with sufficient coherence to perform transfer function analysis of the relationship between CBFV and O2 Hb. The phase between CBFV and O2 Hb was negative in all three groups, indicating that O2 Hb oscillations follow those of CBFV with a time delay caused by the transit time of the blood. The gain of the transfer function analysis between CBFV and O2 Hb might reflect damping of the CBFV fluctuations (perhaps by distensible arterioles or capillaries), resulting in smaller or larger oscillations in cortical oxygenation. Gain may, however, also be influenced by metabolic reserve and its effect on oxygenated haemoglobin, or by a reduced diffusion of oxygen, leading to enhanced oscillations in O2 Hb in response to changes in CBFV (van Beek et al. 2012). With increasing age, the phase lag between CBFV and O2 Hb became larger, whereas the normalized gain became higher. This is in line with results from patients with Alzheimer’s disease, a disease in which cerebral microvasculature is affected. In the Alzheimer’s disease patients, the phase lag was larger and gain higher between CBFV and O2 Hb at 0.05 Hz compared with age-matched control subjects (van Beek et al. 2012). It is possible that the microvasculature in older individuals becomes C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
Exp Physiol 99.3 (2014) pp 586–598
Cerebral perfusion and cortical oxygenation in ageing
compromised compared with young individuals. We speculate that increased microvascular tortuosity (which leads to increased vessel length and thereby transit time, as has been demonstrated in arterial spin labelling studies) could explain the larger phase shift and that reduced distensibility and concomitant reduced damping could explain larger gain.
595
The single squat–stand manoeuvre elicited a reduction in O2 Hb in the frontal cortex, which was similar in all groups. This challenging manoeuvre, which is comparable with activities in daily life, such as standing up from a chair or putting on a sock or shoe, does not seem to compromise oxygenation of the frontal cortex in healthy older individuals. This would confirm our observation of preserved dynamic cerebral autoregulation based on TCD in these healthy subjects. These compensatory mechanisms may, however, fail in the event of more severe stress (for example, during septic shock) or with diseases that compromise cerebral macro- and microvasculature further, such as Alzheimer’s disease. Limitations
Figure 1. Blood pressure (A), cerebral blood flow velocity (CBFV; B) and cortical oxygenation (oxygenated haemoglobin, O2 Hb; C) responses of the three groups during the single squat–stand manoeuvre The vertical dashed line indicates the moment of standing after 1 min of squatting. The continuous black line represents the young, the continuous grey line the elderly, and the dotted grey line the older elderly group. C 2013 The Authors. Experimental Physiology C 2013 The Physiological Society
In order to induce larger haemodynamic oscillations, a repeated sit–stand manoeuvre was applied. The induced large blood pressure oscillations during the repeated sit–stand manoeuvre provide strong and physiologically relevant haemodynamic perturbations and lead to an improved estimation of the transfer function parameters and higher power spectrum density, thereby increasing the reliability of the assessment (Claassen et al. 2009; van Beek et al. 2010a). However, the manoeuvre could be seen as aerobic exercise and may therefore affect our measurements. However, Ogoh et al. (2005) found that mild- to moderate-intensity exercise did not impair dynamic cerebral autoregulation. It is possible that the older elderly subjects experienced the repeated sit–stand manoeuvre as mild exercise; however, compared with baseline values their heart rates did not increase more than in the other two groups, nor did the EtCO2 change significantly compared with baseline. Thus, confounding effects of aerobic exercise during the repeated sit–stand manoeuvre are unlikely. Cerebral blood flow velocity was measured in the middle cerebral artery while NIRS was obtained from the frontal cortex, which receives blood supply from the anterior cerebral artery and the middle cerebral artery (van Beek et al. 2012). We assumed that relative changes of CBFV would be similar between the middle and anterior cerebral arteries in response to blood pressure changes and that the supplying anterior cerebral artery would be likely to react in a similar manner to the middle cerebral artery. The cut-off point of coherence of
