Sodium nitroprusside dilates cerebral vessels and enhances internal carotid artery flow in young men

Abstract

Key points

• Sodium nitroprusside lowers blood pressure by vasodilatation but is reported to reduce cerebral blood flow.

• In healthy young men sodium nitroprusside reduced blood pressure, total peripheral resistance, and arterial CO₂ tension and yet cerebral blood flow was maintained, with an increase in internal carotid artery blood flow and cerebrovascular conductance.

• Sodium nitroprusside induces both systemic and cerebral vasodilatation affecting internal carotid artery more than vertebral artery flow.

Abstract

Cerebral autoregulation maintains cerebral blood flow (CBF) despite marked changes in mean arterial pressure (MAP). Sodium nitroprusside (SNP) reduces blood pressure by vasodilatation but is reported to lower CBF, probably by a reduction in its perfusion pressure. We evaluated the influence of SNP on CBF and aimed for a 20% and then 40% reduction in MAP, while keeping MAP ≥ 50 mmHg, to challenge cerebral autoregulation. In 19 healthy men (age 24 ± 4 years; mean ± SD) duplex ultrasound determined right internal carotid (ICA) and vertebral artery (VA) blood flow. The SNP reduced MAP (from 83 ± 8 to 69 ± 8 and 58 ± 4 mmHg; both P < 0.0001), total peripheral resistance, and arterial CO₂ tension ($P_{\mathrm{aC}{\mathrm{O}}_2}$; 41 ± 3 vs. 39 ± 3 and 37 ± 4 mmHg; both P < 0.01). Yet ICA flow increased with the moderate reduction in MAP but returned to the baseline value with the large reduction in MAP (336 ± 66 vs. 365 ± 69; P = 0.013 and 349 ± 82 ml min^–1; n.s.), while VA flow (114 ± 34 vs. 112 ± 38 and 110 ± 42 ml min^–1; both n.s.) and CBF ((ICA + VA flow) × 2; 899 ± 135 vs. 962 ± 127 and 918 ± 197 ml min^–1; both n.s.) were maintained with increased cerebrovascular conductance. In conclusion, CBF is maintained during SNP‐induced reduction in MAP despite reduced $P_{\mathrm{aC}{\mathrm{O}}_2}$ and the results indicate that SNP dilates cerebral vessels and increases ICA flow.

Key points

• Sodium nitroprusside lowers blood pressure by vasodilatation but is reported to reduce cerebral blood flow.

• In healthy young men sodium nitroprusside reduced blood pressure, total peripheral resistance, and arterial CO₂ tension and yet cerebral blood flow was maintained, with an increase in internal carotid artery blood flow and cerebrovascular conductance.

• Sodium nitroprusside induces both systemic and cerebral vasodilatation affecting internal carotid artery more than vertebral artery flow.

Introduction

Cerebral blood flow (CBF) is thought to be maintained by autoregulation when mean arterial pressure (MAP) ranges from approximately 60 to 150 mmHg (Lassen, 1959) following modulation of pial and large cerebral artery tone (Willie et al. 2014; Lewis et al. 2015). Yet CBF may be affected by changes in MAP within these limits, particularly by reduction in MAP (Numan et al. 2014; Willie et al. 2014). Further, the lower limit of cerebral autoregulation varies between individuals and may depend on the intervention applied to reduce MAP (Drummond, 1997). Thus, during orthostatic stress CBF decreases with a minimal reduction in MAP (Sato et al. 2012a) while, on the other hand, CBF may increase in response to vasodilators (Joshi et al. 2002; Hussain et al. 2009; Perko et al. 2011) and cerebral oxygenation is preserved even when MAP is reduced to about 40 mmHg during anaesthesia (Nissen et al. 2009).

Regarding the influence of sodium nitroprusside (SNP) on CBF, traditional assessment of CBF by ¹³³Xe clearance has led to contradictory conclusions (Brown et al. 1977; Henriksen & Paulson, 1982; Henriksen et al. 1982). The conflicting reports on the effect of SNP on CBF may relate to a concomitant increase in ventilation that reduces arterial CO₂ tension ($P_{\mathrm{aC}{\mathrm{O}}_2}$) and thereby CBF (Kety & Schmidt, 1948), albeit SNP and hypotension may also affect cerebral CO₂ reactivity (Harper & Glass, 1965; Lavi et al. 2003). Even when corrected for changes in $P_{\mathrm{aC}{\mathrm{O}}_2}$, a SNP‐induced reduction in MAP of approximately 20% has been reported to reduce CBF (Brown et al. 1977; Henriksen et al. 1982) or have no effect (Henriksen & Paulson, 1982), while a 40% reduction in MAP reduced CBF by about 15% (Henriksen & Paulson, 1982).

The brain is perfused by the internal carotid (ICA) and vertebral (VA) arteries. The brainstem is important for regulation of blood pressure and vertebro‐basilar hypoperfusion may be important in regard to development of (pre)syncopal symptoms (Shin et al. 1999). Considering that changes in flow may be more pronounced in ICA than in VA, as demonstrated during orthostatic stress (Sato et al. 2012a; Ogoh et al. 2015), and CO₂ reactivity may be different in the two vessels (Deegan et al. 2010, Sato et al. 2012b; Willie et al. 2012), we evaluated the influence of SNP on CBF by determining both ICA and VA flow, as well as middle cerebral artery mean blood velocity (MCA V _mean). Administration of SNP aimed for a reduction in MAP by 20% and then 40%, while keeping MAP ≥ 50 mmHg, to challenge the conventionally accepted lower limit of cerebral autoregulation. In order to account for the potential influence of hypocapnic vasoconstriction, we evaluated CO₂ reactivity in ICA and VA by hyperventilation at rest. CBF is reported as measured and also with a ‘correction’ for eventual changes in $P_{\mathrm{aC}{\mathrm{O}}_2}$, accepting that CO₂ reactivity may be affected during SNP infusion. We hypothesised that a SNP‐induced reduction in MAP affects CBF, with a larger reduction in ICA flow than VA flow, and that CO₂ reactivity is larger for ICA than for VA.

Methods

Subjects and ethical approval

Twenty healthy men were recruited after giving verbal and written informed consent as approved by the ethical committee of the Copenhagen Region (H‐17017045), in accordance with the Declaration of Helsinki, and the study was registered as a clinical trial (https://clinicaltrials.gov/ct2/show/NCT03317652?cond=NCT03317652&rank=1; Clinical Trials ID NCT03317652; October, 2017). Inclusion criteria were men aged 18–35 years and exclusion criteria included alcohol intake ≥ 420 g week^–1; body mass index < 18 or > 25 kg m^–2; smoking; heart, lung, liver, kidney, or metabolic disease that require medication; anti‐hypertensive medication and medication that may affect CBF; neurologic disease considered to affect CBF including epilepsy and sclerosis; > 15% obstruction of the ICA (Saam et al. 2008); haemoglobin < 6 mM; vitamin B12 deficiency; Leber's hereditary optic neuropathy; use of monoamine oxidase inhibitors; and intake of sildenafil or vardenafil for 24 h and tadalafil for 48 h.

Measurements

A radial (n = 3) or brachial artery catheter (n = 16) was inserted for evaluation of blood gas variables and connected to a transducer (Edwards Life Sciences, Irvine, CA, USA) placed at heart level, and a monitor (Patient Monitor M1166A model 66s, Hewlett Packard, CA, USA) reported MAP and heart rate. Invasive arterial pressure monitoring allowed for evaluation of stroke volume, cardiac output (CO), and total peripheral resistance by modified pulse contour analysis (Finometer, Finapres Medical Systems, Amsterdam, The Netherlands) which is correlated to evaluation by thermodilution (Harms et al. 1999; Ameloot et al. 2013). A line was placed in a cubital vein for infusion of SNP (Cavafix 20G 32 cm, Braun, Melsungen, Germany).

Right ICA and VA flow and MCA V _mean were determined. Duplex ultrasound evaluated ICA at least 1.5 cm distal to its bifurcation and VA between the transverse processes C2–5 at 10–12 MHz (Logiq E, GE Medical System, Jiangsu, China) in a longitudinal view with the head turned left. At a stable angle ≤ 60 deg, pulsed wave Doppler determined the angle‐corrected time averaged maximum velocity that corresponds to two times the mean velocity (Li et al. 1993) and edge‐detection software evaluated vessel diameter (Brachial Analyser for Research v. 6, Medical Imaging Applications LLC, Coralville, IA, USA). The ICA and VA flows were calculated as:

$$\begin{array}{ccc}\mathrm{Blood}\mathrm{flow}\hfill & =\hfill & {\displaystyle 60\times \pi \times \left(\frac{Diameter}{2}\right)2\times }\hfill \\ & & {\displaystyle \frac{Timeaveragedmaximumbloodvelocity}{2}.}\hfill \end{array}$$

(Thomas et al. 2015), with:

$$\mathrm{CBF}=\left(\mathrm{ICA}\mathrm{flow}+\mathrm{VA}\mathrm{flow}\right)\times 2.$$

To limit the influence of ventilation on CBF, two recordings over 15–20 s were conducted at each MAP and the mean reported with evaluation of the two arteries in random order.

Frictional stress on the ICA and VA caused by blood velocity was estimated as shear rate using the following formula (Parker et al. 2009):

$$\begin{array}{cc}\mathrm{Shear}\mathrm{rate}=\hfill & \\ {\displaystyle \frac{4\times \mathrm{Time}\mathrm{averaged}\mathrm{maximum}\mathrm{blood}\mathrm{velocity}}{\mathrm{Diameter}}.}\hfill \end{array}$$

Transcranial Doppler evaluated MCA V _mean using a 2 MHz probe (Multidop X; DWL, Sipplingen, Germany) stabilised by a headband. Two pairs of electrodes were placed at the right side of the neck and in the upper left mid‐axillary line to determine changes in extracellular and total water admittance by current at 3 and 100 kHz (C‐guard, DanMeter, Odense, Denmark). The low frequency current crosses the cell membrane only with difficulty and we report the difference in electrical admittance between the two frequencies, which reflects the intracellular compartment admittance and thereby changes in central blood volume (Cai et al. 2000).

Central haemodynamics and MCA V _mean were recorded at 100 Hz (Powerlab 16/35 and LabChart 7, ADInstruments, Bella Vista, Australia) and thoracic admittance every 15th second and saved on a PC. Values were averaged over 2 min and arterial blood sampled in pre‐heparinised syringes (QS50, Radiometer, Copenhagen, Denmark) for analysis of blood gas variables (ABL 725, Radiometer).

Infusion protocol

Subjects presented after at least 4 h fast and had abstained from alcohol and caffeine for 12 h and rigorous exercise for 24 h. The subjects were supine in a room kept at 22⁰C with baseline evaluation at least 30 min after instrumentation. Cerebral CO₂ reactivity was evaluated by hyperventilation for 8 min to provoke a 5–9 mmHg reduction in $P_{\mathrm{aC}{\mathrm{O}}_2}$.

Subsequently, the subjects rested for at least 15 min before administration of SNP in 5% dextrose (Nitropress Hospira, Pfizer, IL, USA) covered by aluminium foil, and infused using a pump (Perfusor Space, B. Braun, Melsungen, Germany). The aim was to achieve a 20% and then a 40% reduction in MAP, while keeping MAP ≥ 50 mmHg and infusion rate ≤ 10 μg kg^–1 min^–1, with the administration of SNP stopped if nausea or dizziness appeared (Henriksen & Paulson, 1982; Van Lieshout et al. 2003). When the targeted MAP was reached, MAP was maintained for 2 min before measurement and the infusion rate reduced over 5–10 min to avoid rebound hypertension (Bünemann et al. 1991).

Statistics

A power calculation indicated that 14 subjects were needed to detect a 15% change in CBF, corrected for $P_{\mathrm{aC}{\mathrm{O}}_2}$, resulting from a SNP‐induced reduction in MAP by 40% (Henriksen & Paulson, 1982), assuming a SD of 18% (Lewis et al. 2015) with a 5% significance level and a power of 80%. The primary outcome was change in CBF resulting from a SNP‐induced 40% reduction in MAP. Secondary outcomes were change in CBF resulting from a 20% reduction in MAP, change in ICA flow compared to VA flow, and CO₂ reactivity for the two arteries.

Vascular conductance was flow divided by MAP and an index was calculated similarly for MCA V _mean. The CO₂ reactivity for ICA and VA was calculated as:

$$\mathrm{C}{\mathrm{O}}_2\mathrm{reactivity}=\frac{100\times \mathrm{Change}\mathrm{in}\mathrm{blood}\mathrm{flow}}{\mathrm{Change}\mathrm{in}\times \mathrm{Baseline}\mathrm{blood}\mathrm{flow}},$$

and was calculated similarly for MCA V _mean. The ICA and VA flow and thus CBF during SNP infusion was also ‘corrected’ for change in $P_{\mathrm{aC}{\mathrm{O}}_2}$ from baseline by the determined CO₂ reactivity:

$$\begin{array}{ccc}\hfill \mathrm{Corrected}\mathrm{blood}\mathrm{flow}& =& \mathrm{Blood}\mathrm{flow}\times \hfill \\ & & \left({\displaystyle 1-\frac{\mathrm{C}{\mathrm{O}}_2\mathrm{reactivity}\times \mathrm{Change}\mathrm{in}}{100}}\right),\hfill \end{array}$$

and similarly for MCA V _mean, with the reservation that CO₂ reactivity may be affected by SNP. We estimated the relative change in MCA diameter in response to SNP by reordering the equation of blood flow, assuming similar changes in ICA and MCA flow, as ICA probably supplies MCA:

$$\begin{array}{ccc}\hfill \mathrm{MCA}\mathrm{blood}\mathrm{flow}& =& \pi \times \mathrm{MCA}{\mathrm{V}}_{\mathrm{mean}}\times \hfill \\ & & {\displaystyle {\left(\frac{\mathrm{MCA}\mathrm{diameter}}{2}\right)}^2,}\hfill \end{array}$$

whereby:

$$\begin{array}{cc}\mathrm{Change}\mathrm{in}\mathrm{MCA}\mathrm{diameter}=\hfill & \\ {\displaystyle \sqrt{\frac{\mathrm{Change}\mathrm{in}\mathrm{ICA}\mathrm{blood}\mathrm{flow}}{\mathrm{Change}\mathrm{in}\mathrm{MCA}{V}_{\mathrm{mean}}}}.}\hfill \end{array}$$

A Student's paired t test was used to compare variables at baseline and during hyperventilation and to evaluate change in flow for ICA vs. VA during infusion of SNP, and the CO₂ reactivity of the two vessels. Change by SNP infusion was evaluated by a repeated measures mixed model, fitted by restricted maximum likelihood in a structured covariance model with intervention (baseline and low and high SNP infusion rate) as a fixed effect and the subject variable as a random effect, in order to take into account the correlation of data within subjects (Proc mixed; SAS 9.4, SAS Institute, Cary, NC, USA). Arterial lactate was analysed after logarithmic transformation to obtain a normal distribution. Evaluation of changes in arterial O₂ saturation was by Friedman's test. Change in the estimated MCA diameter was by a one‐sample t test. Pearson correlation coefficient was used to evaluate the relation between diameter and shear rate for ICA and VA and between changes in MAP and CO and changes in shear rate. Values are presented as means ± SD or median (IQR) for non‐normally distributed data, and changes as means (95% CI). Statistical significance was set at P < 0.05.

Results

The study was not conducted for one subject as we could not place the arterial catheter. Subjects in the study were aged 24 ± 4 years, with a mean height of 182 ± 5 cm, and a mean weight of 75 ± 9 kg (body mass index 22 ± 2 kg m^–2). For one subject, ICA flow was not evaluated following the moderate reduction in MAP, and three subjects became ill during the large reduction in MAP, at which point the infusion of SNP was stopped, the subjects’ legs were elevated and data obtained at that time was not included in the analysis.

Cerebral CO₂ reactivity

Cerebral haemodynamic and arterial gas variables at baseline and during hyperventilation are presented in Table 1. The CO₂ reactivity was similar for ICA and VA (2.7 ± 1.3 vs. 3.0 ± 1.7% mmHg^–1; P = 0.5393) and thus the value for CBF was 2.8 ± 1.0% mmHg^–1. For MCA V _mean it was 2.1 ± 1.5% mmHg^–1.

Table 1.

Cerebral haemodynamic and arterial gas variables at baseline and during hyperventilation

	Baseline (n = 19)	Hyperventilation (n = 19)	P value
ICA TAVMAX (cm s–1]	47 ± 9	38 ± 8*	< 0.0001
ICA diameter (mm)	5.5 ± 0.5	5.5 ± 0.6	0.794
ICA blood flow (ml min–1)	336 ± 66	270 ± 61*	< 0.0001
ICA conductance (ml min–1 mmHg–1)	4.1 ± 0.9	3.2 ± 0.7*	< 0.0001
ICA shear rate (s–1)	347 ± 87	286 ± 77*	0.0002
VA TAVMAX (cm s–1)	29 ± 6	23 ± 5*	< 0.0001
VA diameter (mm)	4.0 ± 0.4	3.9 ± 0.5*	0.0403
VA blood flow (ml min–1)	114 ± 34	88 ± 29*	< 0.0001
VA conductance (ml min–1 mmHg–1)	1.4 ± 0.4	1.0 ± 0.4*	< 0.0001
VA shear rate (s–1)	293 ± 64	240 ± 56*	< 0.0001
CBF (ml min–1)	899 ± 135	717 ± 127*	< 0.0001
CBF conductance (ml min–1 mmHg–1)	10.8 ± 1.7	8.5 ± 1.4*	< 0.0001
MCA mean blood velocity (cm s–1)	69 ± 16	57 ± 16*	< 0.0001
MCA conductance index (cm s–1 mmHg–1)	0.8 ± 0.2	0.7 ± 0.2*	< 0.0001
Arterial CO2 tension (mmHg)	41 ± 3	34 ± 4*	< 0.0001
Arterial O2 tension (mmHg)	106 ± 8	121 ± 13*	< 0.0001

ICA, internal carotid artery; TAVMAX, time averaged maximum blood velocity; VA, vertebral artery; CBF, cerebral blood flow; MCA, middle cerebral artery. Values are means ± SD. P values are from Student's paired t test. ^* P < 0.05 vs. baseline.

Infusion of sodium nitroprusside

Haemodynamics and arterial variables

Infusion of SNP (at 2.0 (1.5–2.5) and 8.6 (6.0–9.4) μg kg^–1 min^–1) reduced MAP by 17 ± 5% and 31 ± 6% (both P < 0.0001; Fig. 1; Table 2), with an accumulated dose of 14 (10–23) μg kg^–1 over 10 (9–14) min and 71 (54–79) μg kg^–1 over 20 (17–24) min and with no significant difference between values derived from the radial and brachial arteries.

Figure 1.

Mean arterial pressure and arterial CO₂ tension at baseline and during infusion of sodium nitroprusside (SNP)

A and B, mean arterial pressure (A) and arterial CO₂ tension (B) at baseline (n = 19) and during low (n = 19) and high (n = 16) SNP infusion rate. Values are individual data and means ± SD. P values represent the overall effect as estimated by a repeated measure mixed model. ^* P < 0.05 vs. baseline. ^† P < 0.05 for low vs. high SNP infusion rate.

Table 2.

Central haemodynamic and arterial variables at baseline and during infusion of sodium nitroprusside (SNP)

	Baseline (n = 19)	SNP low rate (n = 19)	SNP high rate (n = 16)	P value
Mean arterial pressure (mmHg)	83 ± 8	69 ± 8*	58 ± 4* †	< 0.0001
Heart rate (min–1)	55 ± 9	82 ± 14*	94 ± 16* †	< 0.0001
Stroke volume (ml)	96 ± 9	93 ± 16	91 ± 18	0.1538
Cardiac output (l min–1)	5.3 ± 1.0	7.4 ± 1.5*	8.4 ± 2.0* †	< 0.0001
TPR (mmHg min l–1)	17 ± 4	10 ± 2*	7 ± 1* †	< 0.0001
Thoracic admittance index (S × 104)	102 ± 22	98 ± 20*	97 ± 20*	0.0042
Arterial CO2 tension (mmHg)	41 ± 3	39 ± 3*	37 ± 4* †	< 0.0001
Arterial O2 tension (mmHg)	106 ± 8	117 ± 14*	126 ± 12* †	< 0.0001
Arterial O2 saturation (%)	98 (98–98)	99 (98–99)*	99 (99–99)* †	0.0003
Haemoglobin concentration (mM)	8.8 ± 0.7	8.9 ± 0.7*	8.9 ± 0.7*	0.0002
Arterial O2 content (mM)	8.6 ± 0.7	8.8 ± 0.7*	8.8 ± 0.7*	< 0.0001
Arterial glucose (mM)	5.3 ± 0.2	5.6 ± 0.4*	6.1 ± 0.5* †	< 0.0001
Arterial lactate (mM)	0.5 (0.4–0.6)	0.7 (0.6–0.8)*	0.6 (0.5–0.7)* †	0.0001

TPR, total peripheral resistance. Values are means ± SD or median (IQR). P values represent the overall effect as estimated by a repeated measure mixed model. ^* P < 0.05 vs. baseline. ^† P < 0.05 for low vs. high SNP infusion rate.

Heart rate and CO increased, while total peripheral resistance was reduced with increasing SNP infusion, whereas stroke volume was maintained despite a reduction in thoracic electrical admittance (Table 2). The $P_{\mathrm{aC}{\mathrm{O}}_2}$ decreased with the infusion rate (by 2.3 mmHg (95% CI: 0.7–3.8; P = 0.0051) and 4.5 mmHg (95% CI: 2.9–6.1; P < 0.0001)), while arterial O₂ tension and saturation, haemoglobin, O₂ content, glucose, and lactate increased marginally.

Cerebral haemodynamics

The SNP infusion did not affect ICA blood velocity significantly but its diameter increased at the low infusion rate and also tended to increase at the high rate (P = 0.0525; Table 3). In contrast, VA blood velocity decreased, but its diameter also increased. Thus, ICA flow increased at the low rate (by 27 ml min^–1; 95% CI: 6–48; P = 0.013), but returned to the baseline value at the high rate, while there was no significant change in VA flow and CBF (by 52 ml min^–1 (95% CI: –1 to 106; P = 0.0563) and 24 ml min^–1 (95% CI: –32 to 80; P = 0.3937); Fig. 2). Hence, the change in ICA flow was larger than that for VA at the low rate (9 ± 13% vs. –1 ± 20%; P = 0.0321), but similar at the high rate (5 ± 15% vs. –3 ± 26%; P = 0.2382), and both ICA and VA conductance increased. Shear rate for ICA was unaffected by SNP, whereas it decreased for VA. For both ICA and VA, shear rate was unrelated to change in diameter at both infusion rates. Further, changes in MAP and CO were unrelated to change in shear rate for both vessels.

Table 3.

Cerebral haemodynamic variables at baseline and during infusion of sodium nitroprusside (SNP) without and with ‘correction’ for changes in arterial CO₂ tension by the determined CO₂ reactivity

	Baseline (n = 19)	SNP low rate (n = 19)	SNP high rate (n = 16)	P value
Without ‘correction’ for changes in arterial CO2 tension
ICA TAVMAX (cm s–1)	47 ± 9	47 ± 10	48 ± 12	0.9823
ICA diameter (mm)	5.5 ± 0.5	5.8 ± 0.5*	5.6 ± 0.3	0.0018
ICA blood flow (ml min–1)	336 ± 66	365 ± 69*	349 ± 82	0.0417
ICA conductance (ml min–1 mmHg–1)	4.0 ± 0.8	5.3 ± 0.8*	6.1 ± 1.3* †	< 0.0001
ICA shear rate (s–1)	347 ± 87	332 ± 90	348 ± 94	0.4222
VA TAVMAX (cm s–1)	29 ± 6	25 ± 5*	26 ± 8*	0.0017
VA diameter (mm)	4.0 ± 0.4	4.3 ± 0.6*	4.3 ± 0.6*	0.0002
VA blood flow (ml min–1)	114 ± 34	112 ± 38	110 ± 42	0.7963
VA conductance (ml min–1 mmHg–1)	1.4 ± 0.4	1.6 ± 0.6*	1.9 ± 0.7* †	< 0.0001
VA shear rate (s–1)	293 ± 64	242 ± 52*	247 ± 95*	0.0001
CBF (ml min–1)	899 ± 135	962 ± 127	918 ± 197	0.157
CBF conductance (ml min–1 mmHg–1)	10.8 ± 1.7	14.1 ± 2.3*	15.9 ± 3.1* †	< 0.0001
MCA mean blood velocity (cm s–1)	69 ± 16	59 ± 14*	52 ± 14* †	< 0.0001
MCA conductance index (cm s–1 mmHg–1)	0.82 ± 0.17	0.86 ± 0.18	0.91 ± 0.23* †	0.0014
With ‘correction’ for changes in arterial CO2 tension
ICA blood flow (ml min–1)	336 ± 66	388 ± 74*	382 ± 78*	< 0.0001
ICA conductance (ml min–1 mmHg–1)	4.0 ± 0.8	5.7 ± 1.2*	6.6 ± 1.2* †	< 0.0001
VA blood flow (ml min–1)	114 ± 34	119 ± 40	125 ± 46	0.2608
VA conductance (ml min–1 mmHg–1)	1.4 ± 0.4	1.7 ± 0.6*	2.2 ± 0.8* †	< 0.0001
CBF (ml min–1)	899 ± 135	1023 ± 131*	1013 ± 183*	0.0002
CBF conductance (ml min–1 mmHg–1)	10.8 ± 1.7	15.0 ± 2.5*	17.6 ± 2.8* †	< 0.0001
MCA mean blood velocity (cm s–1)	69 ± 16	61 ± 15*	56 ± 15* †	< 0.0001
MCA conductance index (cm s–1 mmHg–1)	0.82 ± 0.17	0.89 ± 0.19*	0.98 ± 0.27* †	< 0.0001

ICA, internal carotid artery; TAVMAX, time averaged maximum blood velocity; VA, vertebral artery; CBF, cerebral blood flow; MCA, middle cerebral artery. n = 18 for ICA and CBF at the low SNP rate. Values are means ± SD. P values represent the overall effect as estimated by a repeated measure mixed model. ^* P < 0.05 vs. baseline. ^† P < 0.05 for low vs. high SNP infusion rate.

Figure 2.

Cerebral haemodynamic variables at baseline and during infusion of sodium nitroprusside (SNP) without ‘correction’ for changes in arterial CO₂ tension

A–D, internal carotid artery blood flow (A), vertebral artery blood flow (B), cerebral blood flow (C), and cerebral blood flow conductance (D) at baseline (n = 19) and during low (n = 18; n = 19 for VA flow) and high (n = 16) SNP infusion rate. Values are individual data and means ± SD. P values represent the overall effect as estimated by a repeated measure mixed model. ^* P < 0.05 vs. baseline. ^† P < 0.05 for low vs. high SNP infusion rate.

In contrast to ICA flow, which probably supplies the MCA, MCA V _mean decreased with infusion rate (by 14 ± 7% and 25 ± 10%; both P < 0.0001) and the difference from changes in ICA flow was significant at both infusion rates (by 23 ± 14% and 29 ± 14%; both P < 0.0001), while MCA V _mean conductance index decreased at the high infusion rate. Thus, the results suggest that MCA diameter increased at both infusion rates (by 12 ± 7% and 18 ± 8%; both P < 0.0001). At the time (pre)syncopal symptoms developed in three subjects, as identified by nausea and dizziness, MAP was reduced by 20%, 42%, and 26%, CBF by 27%, 8%, and 16%, and $P_{\mathrm{aC}{\mathrm{O}}_2}$ by 7, 9, and 9 mmHg, while CO was increased by 4% and 66% or reduced by 8%.

Correction for change in $P_{\mathrm{aC}{\mathrm{O}}_2}$

After ‘correction’ using the determined CO₂ reactivity at rest, infusion of SNP at the two rates increased ICA flow similarly (by 50 ml min^–1 (95% CI: 27–72; P < 0.0001) and 50 ml min^–1 (95% CI: 26–73; P = 0.0001), respectively), whereas VA flow was unaffected (Table 3). Thus, the change in ICA flow was larger than for VA at the low infusion rate (16 ± 14% vs. 5 ± 21%; P = 0.0234), but similar at the high infusion rate (15 ± 15% vs. 10 ± 28%; P = 0.4617) and conductance increased for both arteries. Thereby CBF increased similarly at the two rates (by 112 ml min^–1 (95% CI: 55–170; P = 0.0004) and 120 ml min^–1 (95% CI: 60–180; P = 0.0003)).

In contrast, MCA V _mean decreased with increasing SNP rate (by 10 ± 9% and 19 ± 10%; both P < 0.0001), while its conductance index increased. Thus change in MCA V_mean was still different from that of ICA flow (by 26 ± 15% and 35 ± 15%; both P < 0.0001) suggesting that MCA diameter increased (by 14 ± 8% and 20 ± 8%; both P < 0.0001).

Discussion

The focus of the study was on whether CBF is influenced by a sodium nitroprusside (SNP)‐induced reduction in MAP that challenges the lower limit of cerebral autoregulation. In healthy young men, a SNP‐induced reduction in MAP did not affect VA flow and CBF, although $P_{\mathrm{aC}{\mathrm{O}}_2}$ decreased and ICA flow increased with a moderate reduction in MAP. Further, when taking the reduction in $P_{\mathrm{aC}{\mathrm{O}}_2}$ into account, SNP increased CBF, albeit CO₂ reactivity may be attenuated by SNP. SNP lowers blood pressure by vasodilatation, as total peripheral resistance decreased despite a reduced central blood volume as indicated by thoracic electrical admittance. Here we demonstrate that SNP dilates large cerebral vessels and increases ICA flow.

In apparent contrast to its effect on CBF, SNP reduced MCA V _mean, as reported also by Lucas et al. (2010). We take that observation to indicate that SNP dilates large cerebral arteries, as demonstrated for both ICA and VA and confirming the results of Liu et al. (2013) for ICA. Thus, MCA V _mean underestimates changes in CBF in response to SNP and here the results suggest that SNP increased MCA diameter by approximately 15%, as ICA flow increased while MCA V _mean was reduced. Indeed, MCA diameter is reported to increase by about 18% in response to nitroglycerin (White et al. (2000).

A concern is that three subjects became ill in response to administration of SNP and presented a reduction in CBF by 8–27%, similar to the reduction in CBF and symptoms of incipient fainting following a SNP‐induced 40% reduction in MAP reported by Henriksen & Paulson (1982). Thus, reduction in MAP should be kept within 20–30%, although Lucas et al. (2010) reported no (pre)syncopal symptoms when MAP was reduced by 40% in response to SNP in healthy young men.

Vasodilators and cerebral blood flow

The lower limit of cerebral autoregulation varies between individuals and appears to depend on whether the intervention implies cerebral vasodilatation (Drummond, 1997). The CBF increases in response to vasodilatation by, for example, L‐arginine (Perko et al. 2011), adenosine (Hussain et al. 2009), and verapamil (Joshi et al. 2002), while nitroglycerin may (Bednarczyk et al. 2002) or may not affect CBF (White et al. 2000). SNP dilates pial vessels (Auer 1978) and we take the results to illustrate the influence of nitric oxide‐mediated vasodilatation on CBF. A reduction in CO may limit CBF (Meng et al. 2015), but as CO increased with administration of SNP there was no indication for a central limitation to control of CBF. Changes in MAP within the conventional limits of cerebral autoregulation may affect CBF (Numan et al. 2014; Willie et al. 2014) and we interpret the results of maintained CBF despite reductions in MAP and $P_{\mathrm{aC}{\mathrm{O}}_2}$ to indicate that SNP reduces the lower limit of cerebral autoregulation.

SNP and ¹³³Xe clearance‐determined CBF

Taken together, there seems to be a distinct difference in the effect of SNP on CBF when evaluated by ¹³³Xe clearance or by duplex ultrasound. With ¹³³Xe clearance, SNP reduces or does not influence CBF (Brown et al. 1977; Henriksen & Paulson, 1982; Henriksen et al. 1982), while with duplex ultrasound evaluation of ICA and VA flow, the results indicate an increase in CBF. Some considerations seem relevant in regard to the ¹³³Xe clearance method. First, ¹³³Xe clearance determines not only CBF but also scalp blood flow, although during exercise increased skin blood flow appears not to be significant in evaluation of CBF (Thomas et al. 1989). More importantly, MCA V _mean and VA flow analysis seems to provide an explanation for the discrepancy between results obtained when recording of CBF by ¹³³Xe clearance or duplex ultrasound during administration of SNP. Cerebral vasodilatation in response to SNP is so large that not only MCA V _mean but also VA blood velocity decreases and thus transient time for blood decreases and clearance of ¹³³Xe is slowed, and this is interpreted as a reduction in CBF.

Regional distribution of flow

Vasomotor control appears to differ for ICA and VA (Sato et al. 2012a; Willie et al. 2012; Ogoh et al. 2015), which may be related to more prominent sympathetic innervation of arteries originating from the ICA than those supplied by VA (Edvinsson & Owman, 1977). Sympathetic activity appears to limit surges in CBF during transient hypertension and hypercapnia but may have only a modest effect during hypotension and hypocapnia (Cassaglia et al. 2008; ter Laan et al. 2013; Brassard et al. 2017). As evaluated by administration of L‐arginine, transcranial Doppler‐determined endothelial function seems more important for the posterior than for the middle cerebral artery (Perko et al. 2011). Yet the effect on CBF by SNP appeared more prominent for ICA than for VA.

Cerebral CO₂ reactivity

Cerebral CO₂ reactivity for ICA has been found to be larger (Sato et al. 2012b), similar (Deegan et al. 2010), or lower than for VA (Willie et al. 2012, Lewis et al. 2015) and in this study it was similar for the two vessels. We took changes in $P_{\mathrm{aC}{\mathrm{O}}_2}$ in response to SNP into account, but in another study cerebral CO₂ reactivity was attenuated when MAP was reduced (Harper & Glass, 1965). In anaesthetised patients, a SNP‐induced reduction in MAP by some 30% approximately halved cerebral CO₂ reactivity (Matta et al. 1995), while in healthy young subjects a minor reduction in MAP by SNP reduced CO₂ reactivity by about 15% (Lavi et al. 2003). Thus, ‘correction’ using the CO₂ reactivity determined at rest may overestimate CBF during SNP infusion. The $P_{\mathrm{aC}{\mathrm{O}}_2}$ was reduced in response to SNP, possibly by increased ventilation mediated by unloading of the arterial baroreflex (Stewart et al. 2011). Evaluation of CO₂ reactivity for MCA V _mean may underestimate changes in CBF, as a reduction in $P_{\mathrm{aC}{\mathrm{O}}_2}$ by approximately 13 mmHg reduces MCA diameter by about 5% (Coverdale et al. 2014), while a reduction by about 10 mmHg seems not to affect its diameter (Verbree et al. 2014).

Limitations

In this study, evaluation of cerebral haemodynamics was on the right side and we can only assume the responses on the left side would be similar, as shown during administration of nitroglycerin (Bednarczyk et al. 2002) and L‐arginine (Pretnar‐Oblak et al. 2006) and in response to hypoxia (Mikhail Kellawan et al. 2017). We did not evaluate whether SNP affects CBF when MAP is preserved, but as CO increased with administration of SNP there was no indication for a central limitation to CBF.

The study could have clamped $P_{\mathrm{aC}{\mathrm{O}}_2}$ as SNP probably affects cerebral CO₂ reactivity, and the evaluation encompassed only Caucasian men. We aimed for a 40% reduction in MAP, maintaining MAP ≥ 50 mmHg, but only attained a 31% reduction in MAP because some subjects demonstrated a low MAP at baseline, developed (pre)syncopal symptoms, or because the maximal SNP infusion rate was reached. The arterial O₂ tension was not controlled but we take the minor increase induced by SNP to have limited effect on CBF (Ainslie & Duffin, 2009).

Conclusion

Sodium nitroprusside induces systemic and cerebral vasodilatation and increased cerebrovascular conductance. The CBF was maintained in response to SNP‐induced reduction in MAP despite reduced $P_{\mathrm{aC}{\mathrm{O}}_2}$ and the results indicate differences in regional regulation of CBF as SNP increased ICA but not VA flow. Future studies could evaluate the effect of SNP on CBF when $P_{\mathrm{aC}{\mathrm{O}}_2}$ is maintained and whether SNP affects CBF when MAP is preserved.

Additional information

Competing interests

The authors declare that they have no conflict of interest.

Author contributions

N.D.O.: conception and design of the study, acquisition, analysis, and interpretation of data and writing the first draft of the manuscript. M.F.: acquisition of data and revision of the manuscript. N.H.S.: conception and design of the study, interpretation of data, and revision of the manuscript. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This study was supported by the University of Copenhagen.

Biography

Niels D. Olesen is doing his PhD under the supervision of Dr Niels H. Secher. He is studying the influence of changes in blood pressure and anaesthesia on cerebral blood flow. Outside of the lab he enjoys spending time with his family and reading. In the future, he plans to become an anaesthesiologist and continue his scientific pursuits.