Cerebral blood flow in experimental ischemia assessed by 19F magnetic resonance spectroscopy in cats A Brunetti, G Nagashima, A Bizzi, DJ DesPres and JR Alger Stroke 1990, 21:1439-1444 Stroke is published by the American Heart Association. 7272 Greenville Avenue, Dallas, TX 72514 Copyright © 1990 American Heart Association. All rights reserved. Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the World Wide Web at: http://stroke.ahajournals.org/content/21/10/1439
Subscriptions: Information about subscribing to Stroke is online at http://stroke.ahajournals.org//subscriptions/ Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Fax: 410-528-8550. E-mail:
[email protected] Reprints: Information about reprints can be found online at http://www.lww.com/reprints
Downloaded from http://stroke.ahajournals.org/ by guest on July 10, 2011
1439
Cerebral Blood Flow in Experimental Ischemia Assessed by 19F Magnetic Resonance Spectroscopy in Cats A. Brunetti, MD, G. Nagashima, MD, A. Bizzi, MD, D.J. DesPres, and J.R. Alger, PhD We evaluated a "F magnetic resonance spectroscopic technique that detects Freon-23 washout as a means of measuring cerebral blood flow in halothane-anesthetized adult cats during and after transient cerebral ischemia produced by vascular occlusion. The experiments were performed to test the ability of this recently developed method to detect postischemic flow deficits. Results were consistent with postischemic hypoperfusion. The method also proved valuable for measuring small residual flow during vascular occlusion. Our experiments indicate that this method provides simple, rapid, and repeatable flow measurements that can augment magnetic resonance examinations of cerebral metabolic parameters in the study of ischemia. (Stroke 1990;21:1439-1444)
M
agnetic resonance techniques take advantage of the ability of specific nonradioactive nuclei, such as hydrogen-1, phosphorus-31, and fluorine-19, to produce radiofrequency signals when placed in a strong magneticfield.Clinical magnetic resonance imaging (MRI) uses these nuclear properties to permit imaging of water in soft tissues. In a related technique, magnetic resonance spectroscopy (MRS), signals from 1H, 31P, or 19F nuclei, which are constituents of molecules dissolved in the tissue fluid, are obtained and analyzed to yield information about chemical structure from the signal frequency and concentration from the signal amplitude. The *H and 31P MRS techniques are being used with increasing frequency to measure the cerebral metabolic status in ischemic injury models. The interpretation of such metabolic measurements would be facilitated by simultaneous cerebral blood flow (CBF) determinations. However, with a few exceptions,1-3 simultaneous CBF and MRS measurements have not been attempted because most conventional CBF techniques are difficult to use within an MRS From the Neuroimaging Section (A.Br., A. Bi., J.R.A.), Medical Neurology Branch, and the Laboratory of Neuropathology and Neuroanatomical Sciences (G.N.), National Institute of Neurological Disorders and Stroke; and the Biomedical Engineering and Instrumentation Program (D.J.D.), National Center for Research Resources, National Institutes of Health, Bethesda, Md. Supported by the Division of Intramural Research, the National Institute of Neurological Disorders and Stroke, and the National Institutes of Health. Address for correspondence: Jeffry R. Alger, PhD, Neuroimaging Section, NINDS, National Institutes of Health, Building 10, Room 1C-451, Bethesda, MD 20892. Received November 28, 1989; accepted June 22, 1990.
magnet. A solution is to use an MRS method for measurement of CBF, such as the recently proposed Freon method,4"6 wherein 19F MRS is used to determine the indicator's concentration during clearance. This method is rapid, repeatable, does not require moving the subject, and permits volume-averaged CBF measurements to be made over the same tissue volume that produces the metabolite MRS signals. This paper describes the first experiments performed to determine the sensitivity of the 19F-MRS method for the detection of flow disturbances during and after transient experimental ischemia. Materials and Methods
The indicator-dilution method7"9 served as the basis for the technique used here, with Freon-23 gas (CHF3) the physiologically inert, lipid-soluble indicator.4"6-10 It was administered with the inspired gas for a period sufficient to saturate the tissue. Then, it was abruptly discontinued, and 19F-MRS was used to obtain the washout function. Zierler11 has reviewed quantitative approaches to blood flow determination appropriate for analogous methods. Based on the assumptions that 1) the arterial Freon concentration drops abruptly to zero at time zero, 2) Freon recirculation is insignificant, and 3) the region of brain detected by MRS is composed of a single homogeneous compartment, the washout function is given by the following equation: Cb(t)=Cb(0) exp{-kt}
(1) In equation 1, t is time, Q,(t) and Q,(0) are, respectively, the relative brain concentration at time t and at time zero, and k is the experimentally determined
Downloaded from http://stroke.ahajournals.org/ by guest on July 10, 2011
1440
Stroke
Vol 21, No 10, October 1990
FREON - 23
•t
• 679
(all.)
-679
FREQUENCY (Hz)
B.
1.0
2.0
3.0
4.0
5.0 8.0 time (min)
7.0
SO
clearance constant. Within these limits, the CBF measured by the Freon-MRS technique, CBF-FR, can be determined because k is related to the blood flow per unit volume by the expression CBF-FR=kA
(2)
where A is the tissue to venous blood partition coefficient. We performed experiments at 4.7 T on adult cats that were pump-ventilated (30 breaths/min, 30 ml/ breath) with 1% halothane and 30% O2 in N2O. The animal preparation and the global cerebral ischemia methodology have been described previously.12 Brain Freon saturation was effected by 15 minutes of inhalation of a 50% Freon-23 gas mixture. Thus, during the period of Freon inhalation, the N2O concentration was reduced from 70% to 20%. Mean arterial pressure and electroencephalogram were unchanged during Freon inhalation. Measurements were made in three nonischemic control animals and in seven animals that were subjected to 10 minutes of global cerebral ischemia (ischemia group). Flow data were obtained 40-60 minutes before, during, and after the ischemic period. Determinations were made during ischemia by saturating the brain through Freon inhalation, allowing washout to proceed for 1-2 minutes, and then producing ischemia. Magnetic resonance spectroscopic data were obtained with a GE CSI instrument (GE NMR Instruments, Fremont, Calif.) via a home-built 17mm-diameter surface coil placed against the skull after retraction of scalp and extracranial muscles.
9.0
FIGURE 1. Panel A: Representative magnetic resonance spectral data obtained from a Freon-23 washout experiment. Each spectrum obtained in 32 seconds; signals of halothane and Freon-23, a doublet (81 Hz splitting), are labeled. Freon administration began 12 minutes before data collection began. Freon was discontinued during the sixth spectrum. Panel B: Semilogarithmic plot of signal intensities shown in panel A. Vertical line denotes time Freon inhalation was discontinued.
10.0
Thus, the detected signal will have contributions from cortical tissue, subcortical tissue, and bone marrow. We collected 19F spectra with a time resolution of 32 seconds using a repetition time of 1.0 second and a pulse width that was found to produce the maximum signal-to-noise ratio per unit time. Under these conditions, the 19F-MRS signal intensity (or area) is directly proportional to the concentration of Freon-23 dissolved in the tissue adjacent to the coil. The clearance constant, k, was determined by linear regression analysis of the first 4-6 minutes of the semilogarithmic 19F-MRS signal decay curve. Knowledge of relaxation times and pulse widths is not needed for k determination because these parameters act to attenuate each signal amplitude determination made during the washout by an equal fraction. The CBF-FR values were computed from Equation 2 using 0.9 for the venous blood to brain partition coefficient.1013 Results Figure 1 illustrates the method. Spectral data showing the Freon-23 doublet and the somewhat broader signal of halothane (used as anesthetic) appear in Figure 1A. The halothane signal's reproducibility within the scatter due to noise demonstrates the absence of systematic drifts in signal intensities. The Freon signal appears stable in the early spectra because the tissue was already Freonsaturated during collection of these data. The Freon washout is reported by the progressive decreases in the Freon signal intensity in spectra obtained after
Downloaded from http://stroke.ahajournals.org/ by guest on July 10, 2011
Brunetti et al
we discontinued Freon administration. A semilogarithmic plot of the Freon-23 signal intensity versus time (Figure IB) illustrates that the initial phase of the washout fits a single exponential function. The slope gives the clearance constant, k. Linear regression analysis of the first 5 minutes of this washout yields a k value of 0.37+0.01 (i?2=1.00). The corresponding CBF-FR is therefore 33 cm3 (100 g"1) min"1. The beginning of the washout process appeared to start 1-2 minutes after Freon was discontinued because there was a large dead volume between the respirator and the animal, resulting from the distance (approximately 3 m) between the magnet and the respirator. The effect can be minimized by rapidly flushing the hoses with Freon-free gas as quickly as possible after the Freon is discontinued (data not shown). Repetitive measurements were made in three control cats, identified in Table 1 as 116, 117, and 118, and plotted in Figure 2. Clearance constant measurements on any single animal in the control group were reproducible over a period of 5 hours to within 12%. The average CBF-FR value obtained from independent measurements in the control animals was 37 ±7 cm3 (100 g"1) min"1 with Paco2 of 31.2±1.4 mm Hg. Table 1 summarizes CBF-FR and Paco2 measurements in the control animals and in the ischemic group before ischemia. No correlation between the measured CBF-FR and Paco2 could be identified in this limited range of Paco2. We investigated the sensitivity of the Freon-MRS technique for detecting residual flow during the ischemic injury by making CBF-FR measurements during ischemia in seven animals. The results fell into two categories represented by two cases, shown in Figure 3. In five animals, the washout could not be detected during the occlusion, indicating that complete ischemia had been achieved. In two animals, there was a small but measurable washout during the occlusion, illustrating that occlusion was not complete. In both sets of animals, washout began without delay on releasing the occlusion because the Freon had emptied from the lungs, body tissues, and respirator hoses during the occlusion. We made postischemic flow measurements in the ischemia group during the 5 hours after injury (Figure 2). The five animals that experienced complete ischemia, as indicated by the lack of Freon washout during ischemia, had Freon clearance that was statistically different from the control animals. The postischemic clearance values were 62% of control in the period from 30 minutes to 5 hours after ischemia. The two animals that experienced partial flow during the ischemic injury did not show subsequent slowed Freon clearance. Discussion The assumptions made to obtain Equation 2 only approximately describe the actual conditions. They were made to evaluate the sensitivity of the method, in its simplest form, to long-term CBF deficit. The
Freon-23 CBF Studies of Cerebral Ischemia
1441
TABLE 1. Paco 2 and Cerebral Blood Flow Determinations in Control Group and Before Occlusion in Ischemia Group
Cat no. Ischemia group 103 106 109 111 115 104 112 Control group 116
117
118
Measurement Preisch Preisch Preisch Preisch Preisch Preisch Preisch Preisch Preisch Preisch Preisch Preisch Preisch Preisch
Paco 2
Clearance constant
CBF-FR
26.3 30.5 26.0 39.6 28.1 28.1 34.1 36.0 32.4 33.8 36.1 36.1 32.1 27.2
0.24 0.21 0.28 0.20 0.28 0.24 0.23 0.27 0.24 0.24 0.31 0.28 0.39 0.35
21.5 19.3 25.3 17.8 25.3 21.8 21.0 24.3 21.5 21.3 28.3 25.2 35.4 31.1
31.1 31.1 47.0 23.6 23.7 29.6 33.0 40.9 41.3 44.8 33.0 33.0 29.4 30.2 32.0 35.4 33.3
0.35 0.33 0.24 0.51 0.58 0.52 0.54 0.55 0.49 0.57 0.36 0.32 0.42 0.37 0.37 0.37 0.34
31.8 29.9 21.2 45.7 52.0 46.6 48.7 49.0 43.7 51.1 32.5 28.8 37.8 33.0 33.3 33.3 30.8
#1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2
Control Control 300 min Control Control Control 30 min 90 min 180 min 300 min Control Control Control 30 min 90 min 180 min 300 min
CBF-FR, cerebral blood flow determined by the Freonmagnetic resonance spectroscopic technique; Preisch, measurements in Ischemia group made before occlusion; Control, measurements in Control group made before sham occlusion. Times given refer to the time between sham occlusion and measurement.
rapid arterial clearance assumptions are supported by Barranco et al,10 who found that the arterial washout function for Freon-23 in ventilated cats was biexponential, with a fast component (78% of the total amplitude) having a rate constant about 10-fold larger than brain washout, and a slow component having a rate constant similar to the brain clearance. Thus, the arterial concentration drops to about 16% of its initial concentration in the first minute and falls to 8% during the next 5 minutes. The more recent measurements of Ewing et al13 showed the arterial concentration fell to zero in 1-2 minutes. The single compartment assumption was made because our Freon MRS signal-to-noise ratios were not sufficiently high to permit fitting the washout to a multi-
Downloaded from http://stroke.ahajournals.org/ by guest on July 10, 2011
1442
Stroke
Vol 21, No 10, October 1990
1.2
RELATIVE C.B.F. 1.0
OB
0.6
0.4— • — O— —•tr—
02-
0.0 -2
2
CONTROLS (N=3) COMPL. ISCHEMIA (N=5) INCOMPL. ISCHEMIA
F I G U R E 2. Relative clearance constants (mean±SD) obtained from control animals (n=3) and animals subjected to 10 minutes of global ischemia (n=7). Separate curves are plotted for five animals in the ischemic group that experienced complete ischemia (o) and two animals that experienced incomplete ischemia ( A ) . C.B.F., cerebral blood flow.
4
TIME (hrs)
with Freon-23, Ewing et al13 also were unable to fit reliably two compartments of blood flow when using single uptake and clearance data, and they had to resort to a pulsed inhalation technique to improve their signal-to-noise ratio. Our single compartment analysis provides clearance rate constants that approximate the average washout in the entire volume of detected tissue, including bone, cortex, and subcortical matter. Therefore, the apparently good fits of our data to a single exponential should not be taken as evidence that the tissue being sampled is homogeneous. Barranco et al10 supported the method's validity by performing a comparative study with microspheres.
exponential function. Ewing et al6 observed multiexponential Freon-22 washouts, which they associated with the contributions of grey matter, white matter, and extracerebral tissues. The amplitudes of their slower exponentials were less than 5% of the amplitude of the more rapid component. The lower precision of this work is partially related to the propensity of Freon-22 for producing a stronger MRS signal at tissue saturation than Freon-23 produces. Moreover, Ewing et al6 obtained Freon signal intensities by applying heavy signal filtering and integrating over a wide frequency range. This approach was impractical here because a halothane signal was present. In a more recent study
A. Ln FREON INTENSITY
FREONOFF
m
__J
OCCUSON RELEASE
Ln FREON INTENSITY
ccausKjN
MAAMAA^ Ra£XSE
A A A A A AA
•
• \
A
•• •• » •
V • CONTROL A COMPLETE ISCKEMA
A — .
A*
Time
A
P • • A
10
A
**A
COHTROL INCOMPLETE ISCHEMIA
S
A A
A*1*
• 10
15
TIME (mln)
19
FIGURE 3. Semilogarithmic F magnetic resonance spectroscopy washout data from individuals experiencing complete (panel A) and incomplete (panel B) ischemia, together with preischemic measurements. Clearance rate constants obtained from linear regression analyses are panel A: preischemic washout: 0.237±0.003 (R2=0.99); occlusion: 0.0119±0.0013 (R2=0.81); postischemic washout: 0.23±0.01 (R2=0.98);panelB:preischemic washout: 0.27±0.01 (R2=0.99); occlusion: 0.041±0.002 (R2=0.96); 2 postischemic washout: 0.51±0.05 (R =0.96).
Downloaded from http://stroke.ahajournals.org/ by guest on July 10, 2011
Brunetti et al
Their Freon-derived CBF values showed the expected CO2 reactivity but were about 30% lower than the values determined simultaneously by microsphere methods. The microsphere measurements demonstrated that inhalation of 70% Freon-23 produced a small CBF depression, whereas the recent study of Ewing et al13 provided no such evidence. Our CBF-FR measurements (Table 1) are somewhat lower than those of Barranco et al.10 However, they are on the low end of the range of global CBF-FR measurements made in cats by several different traditional techniques reported in a review by Lacombe et al.14 Our low values are supported by a statement in that review indicating that inhalation of halothane as a supplement to N2O, as was used here, depresses CBF by 15-20%. The same review indicates that halothane uncouples the normal physiological relationship between blood flow and metabolic rate. This may explain our inability to identify a correlation between Paco2 and CBF-FR. Despite the apparent need for upward correction of CBF-FR values, our method seems to give reliable relative measures of global CBF. Thus, we have reported postischemic bloodflowmeasurements relative to preischemic measurements (Figure 2). Our CBF-FR measurements also are potentially influenced by the change in nitrous oxide concentration (20-70%) that occurred when Freon was discontinued. Presumably, this would have little influence on the relative measures of CBF reported in Figure 2 because both control and postischemic measurement should have been influenced identically. Our primary objective was to test the method as a means of making multiple CBF measurements that could be incorporated easily into metabolic MRS studies of brain ischemia. Horikawa et al3 made simultaneous MRS and CBF measurements, but the CBF measurements necessitated killing the subject. Gadian et al2 and Crockard et al1 measured CBF by H2 clearance while monitoring brain MRS signals. The Freon-MRS method is advantageous relative to this approach in that electrodes need not be placed, thereby obviating tissue damage and eliminating radiofrequency interference from the electrode leads. The repetitive Freon measurements are made quickly, without killing the subject, and the data analysis is relatively simple. Moreover, the fast arterial clearance assumptions are more valid in ischemic and postischemic measurements, because the CBF tends to be low. A significant advantage of the method is that clearance measurements can be repeated as soon as enough washout data have been acquired to estimate k. Figure 1 indicates that this occurs within 5 minutes of the time that Freon administration ceases. Resaturation as judged by the 19F-MRS method is complete with 12 minutes of gas readministration. Thus, measurements of CBF-FR can be repeated every 17 minutes. This permitted determinations before, during, and after ischemia to be made on single animals. The reflow period in vascular occlusion models is characterized by an initial period of reactive hyper-
Freon-23 CBF Studies of Cerebral Ischemia
1443
emia, which is followed by a more prolonged period of cerebral hypoperfusion.15"19 Another objective was to establish whether the Freon method is sensitive to the delayed postischemic hypoperfusion. Our observations (Figure 2) indicate that the Freon-MRS method produces results consistent with postischemic hypoperfusion, even when approximate analytical procedures are used. The technique also appears to be sensitive to small residual flows during a period of experimental cerebral ischemia (Figure 3), and thus can be used to discriminate between complete and incomplete ischemia. We found that five animals that suffered from complete cerebral ischemia all developed subsequent hypoperfusion; however, two animals in which CBFFR was not completely stopped by the occlusion did not develop hypoperfusion. One possible explanation for this finding is that complete energy failure is necessary for the subsequent development of hypoperfusion and that the residual flow during ischemia found in these two animals was sufficient to protect the brain from metabolic energy failure. While the finding and explanation remain preliminary, they point toward further MRS studies relating flow and energy metabolism using techniques that produce graded ischemia. Acknowledgments This work was performed in the National Institutes of Health In Vivo NMR Research Center. The studies benefited from discussions with Drs. G. Di Chiro, I. Klatzo, and A.C. McLaughlin and from a previous collaboration with Prof. Dr. K.-A. Hossmann. References 1. Crockard HA, Gadian DG, Frackowiak RSJ, Proctor E, Allen K, Williams SR, Ross Russell RW: Acute cerebral ischemia: Concurrent changes in cerebral blood flow, energy metabolism, pH and lactate measured with H2 clearance and P-31 and H-l nuclear magnetic resonance spectroscopy. II. Changes during ischemia. / Cereb Blood Flow Metab 1987;7:394-402 2. Gadian DG, Frackowiak RSJ, Crockard HA, Proctor E, Allen K, Williams SR, Ross Russell RW: Acute cerebral ischemia: Concurrent changes in cerebral blood flow, energy metabolism, pH and lactate measured with H2 clearance and P-31 and H-l nuclear magnetic resonance spectroscopy. I. Methodology. / Cereb Blood Flow Metab 1987;7:199-206 3. Horikawa Y, Naruse S, Hirakawa K, Tanaka C, Nishikawa H, Watari H: In vivo studies of energy metabolism in experimental cerebral ischemia using topical magnetic resonance. J Cereb Blood Flow Metab 1985;5:235-240 4. Bolas NM, Petros AJ, Bergel D, Radda GK: Use of 19F magnetic resonance spectroscopy for measurement of cerebral blood flow (abstract). Proc 4th Ann Mtg Soc Magn Res Med 1985;1:315-316 5. Eleff SM, Schnall MD, Ligetti L, Osbakken M, Subramanian VH, Chance B, Leigh JS Jr: Concurrent measurements of cerebral blood flow, sodium, lactate and high-energy phosphate metabolism using 19F, ^Na, 'H and 31P nuclear magnetic resonance spectroscopy. Magn Reson Med 1988;7:412-424 6. Ewing JR, Branch CA, Helpern JA, Smith MB, Butt SM, Welch KMA: Cerebral blood flow measured by NMR indicator dilution in cats. Stroke 1989;20:259-267 7. Kety SS, Schmidt CF: The determination of cerebral blood flow in man by the use of nitrous oxide in low concentrations. Am J Physiol 1945;143:53-66
Downloaded from http://stroke.ahajournals.org/ by guest on July 10, 2011
1444
Stroke Vol 21, No 10, October 1990
8. Kety SS, Schmidt CF: Nitrous oxide method for the quantitative determination of cerebral blood flow in man: Theory, procedure, and normal values. / Clin Invest 1948;27:475-483 9. Kety SS: The theory and applications of the exchange of inert gas at the lungs and tissues. Pharmacol Rev 1951;3:1-41 10. Barranco D, Sutton LN, Florin S, Greenberg J, Sinnwell T, Ligeti L, McLaughlin AC: Use of "F NMR spectroscopy for measurement of cerebral blood flow: A comparative study using microspheres. / Cereb Blood Flow Metab 1989;9:886-891 11. Zierler KL: Equations for measuring blood flow by external monitoring of radioisotopes. Circ Res 1965;16:309-321 12. Alger JR, Brunetti A, Nagashima G, Hossmann K-A: Assessment of postischemic cerebral energy metabolism in cat by 3IP NMR: The cumulative effects of secondary hypoxia and ischemia. / Cereb Blood Flow Metab 1989;9:506-514 13. Ewing JR, Branch CA, Fagan SC, Helpern JA, Simkins RT, Butt SM, Welch KMA: Fluorocarbon-23 measure of cat cerebral blood flow by nuclear magnetic resonance. Stroke 1990;21:100-106
14. Lacombe P, Meric P, Seylez J: Validity of cerebral blood flow measurements obtained with quantitative tracer techniques. Brain Res 1980;2:105-169 15. Hossmann K-A, Lechtape-Grueter H, Hossmann V: The role of cerebral blood flow for the recovery of the brain after prolonged ischemia. Z Neurol 1973;204:281-299 16. Kagstrom E, Smith M-L, Siesjo BK: Local cerebral blood flow in recovery period following complete cerebral ischemia in the rat. / Cereb Blood Flow Metab 1983;3:170-182 17. Levy DE, Van Uitert RL, Pike CL: Delayed post-ischemic hypoperfusion: A potentially damaging consequence of stroke. Neurology 1979;29:1245-1252 18. Pulsinelli WA, Levy DE, Duffy TE: Regional cerebral blood flow and glucose metabolism following transient forebrain ischemia. Ann Neurol 1982;ll:499-509 19. Snyder JV, Nemoto EM, Carroll RG, Safar P: Global ischemia in dogs: Intracranial pressure, brain blood flow and metabolism. Stroke 1975;6:21-27 KEY WORDS • cerebral blood flow • cerebral ischemia nuclear magnetic resonance • cats
Downloaded from http://stroke.ahajournals.org/ by guest on July 10, 2011
•
