Magnerrc Resonance Imaging, Vol. 9, pp. 229-234, Printed in the USA. All rights reserved.
0730-725X/91 $3.00 + .oO Copyright 0 1991 Pergamon Press plc
1991
l Original Contribution
CONTROLLED VENTILATION DURING NMR SPECTROSCOPIC STUDIES: HEMODYNAMIC AND BIOCHEMICAL CONSEQUENCES M. WHALEN
AND
J.I. SHAPIRO
The Giles F. Filley Laboratory of the Webb Waring Lung Institute and Department of Medicine, University of Colorado School of Medicine, Denver, Colorado 80262, USA
The effects of different ventilation methods on cardiac output measured by the indicator-dilution method, liver blood flow measured by a deuterium washout technique using zH nuclear magnetic resonance (NMR) and liver concentrations of ATP and intracellular pH determined with 31P NMR were compared in anesthetized rats. No differences in mean arterial blood pressure were demonstrable with the different modes of ventilation. However, significant drops in cardiac output were observed between freely breathing and animals ventilated with positive pressure but not the high frequency oscillatory method (407 f 46 and 520 + 88 vs. 633 f 86 ml/min/kg, p < 0.05 and p = NS, respectively). Moreover, liver blood flow was significantly reduced during positive pressure but not high frequency oscillatory ventilation compared with free breathing rats (32 f 4 and 43 + 10 vs. 46 f 8 ml/100 g, p < 0.05 and p = NS, respectively). 31P NMR spectroscopy revealed no effects of either ventilation method on tissue ATP or intracellular pH as estimated by the chemical shift of inorganic phosphate. These data suggest that controlled ventilation in normal rats accomplished with standard positive pressure methods is associated with major decreases in cardiac output and liver blood flow despite maintenance of normal blood pressure. High frequency oscillatory ventilation appears to effect less compromise of cardiac output and hepatic perfusion than positive pressure ventilation and may, therefore, be preferable for some biological studies. Keywords: 31P NMR spectroscopy; 31P MRS.
The purpose of this study was to carefully quantitate the hemodynamic responses of the whole animal as well as the hepatic circulatory and metabolic responses to positive pressure ventilation and high frequency oscillatory ventilation under the experimental conditions of in vivo NMR spectroscopic study.
INTRODUCTION In vivo NMR spectroscopy is an important tool in the
investigation of metabolic processes in a variety of biomedical areas. For many of these applications, in particular the study of acid-base disturbances, controlled ventilation is an essential part of the experimental methodology. l-8 However, the manner in which controlled ventilation is accomplished may have significant impact on the underlying physiology and metabolism. Specifically, positive pressure ventilation may induce decreases in right ventricular filling as well as changes in the geometry of the heart, possibly resulting in decreases in cardiac output, mean arterial pressure as well as organ perfusion.9-13 In contrast, high frequency jet ventilation may avoid some of these hemodynamic consequences. I4
METHODS Animals
Sprague Dawley rats (approx. 400-500 g body weight) were used in these experiments. Rats were initially anesthetized with ketamine (0.4 ml, 100 mg/dl) and xylazine (0.4 ml, 20 mg/ml) given intramuscularly. Additional doses of ketamine (0.1 ml) were given as needed to suppress a tail pinch response. The right jugular vein and carotid artery cannulated with
RECEIVED 6/ 15/90; ACCEPTED 9/ 13/90. Acknowledgments-The authors would like to thank Drs. Laurence Chan and Richard Kucera for their helpful suggestions. We would also like to thank Ms. Marsha Underwood and the Infrasonics company for the generous loan of their Infant-Star ventilator. This work was supported by a grant-in-aid from the American Heart Association. Dr.
Shapiro is supported by a Squibb Corporation-American Heart Association Clinician Scientist award. Address all correspondence to Joseph I. Shapiro, M.D., Associate Director, NMR Laboratory, Assistant Professor of Medicine, C-281 University of Colorado Health Sciences Center, 4200 E 9th Ave., Denver, CO 80262, USA.
229
230
Magnetic Resonance Imaging 0 Volume 9, Number 2, 1991
PE-50 tubing for all magnetic resonance studies. For the cardiac output portion of the study both right and left jugular, a carotid artery, femoral artery and vein catheters were placed. All animals were placed on a warming pad to maintain a core temperature of 35”-37°C. Ventilation Animals had a tracheostomy performed (14 ga quick cath) and were ventilated using an Infant Star Ventilator (Infrasonics, Palo Alta, CA) capable of both positive pressure ventilation and high frequency oscillatory ventilation (HFOV) modes of operation. The tubing used for ventilation was made from Tygon tubing having 3/8 and 5/8 inch diameter inside and outside, respectively. The inspiratory and expiratory branches of the circuit were 18 ft in length with a oneway valve system at the animal to ensure no rebreathing of gas occurred. A standard three-way stopcock was used to connect the rat’s tracheostomy tube to the valve arrangement. During pilot studies, PPV and HFOV settings were chosen to place the arterial PC02 about 40 torr with the least deleterious effects on mean arterial pressure. The PPV mode parameters chosen were a tidal volume of 8.0 ml and 42 breaths per minute, a peak inspiratory pressure of 8-10 cmH*O, 2 cmHzO of positive and expiratory pressure and an inspiratory time of 0.14 sec. In HFOV mode 7-10 Hz frequencies were used with an amplitude of 2 cmH*O and 2-3 cmHzO positive end expiratory pressure. Measurements Arterial blood pressure was monitored continuously using a pressure transducer and a chart recorder (Gould Instruments, Cleveland, OH) during each experiment. Arterial blood obtained from either the carotid artery or the femoral artery was analyzed for PO*, PC02 and pH using a blood gas instrument (Radiometer Instruments, Copenhagen, Denmark). Cardiac output measurements were performed using the indicator dilution meth0d.r’ To accomplish this, a densitometer (Waters Inc., Baltimore, MD) interfaced via A/D converter (Metrabyte Corp. Taunton, MA) to a personal computer (IBM PC, International Business Systems, Boca Raton, FL) which recorded the green dye indicator-dilution curves (Cardio-Green@, Hynson, Westcott and Dunning, Inc., Baltimore, MD) that were analyzed to determine cardiac outputs (vida infra). The densitometer was placed ahead of a pulsatile pump which created an arterio-venous shunt between the carotid and right jugular vessels. Bolus injections of green dye (10 ~1) were made through the left jugular catheter using a 500 ~1 syringe equipped
with a 10 ~1 repeating dispenser which generated an evolution and decay curve of arterial green dye concentration. The concentration of green dye vs time was analyzed as described by Trautman and Newboner” so as to integrate only the mono-exponential portion of the concentration decay (A) and ignore the contribution from recirculation. This allowed us to solve the equation CO = Q/A in ml/min where Q was a densitometer reading on a standard concentration of indocyanine green using software written in BASIC by the authors. *’ Magnetic Resonance Spectroscopy Studies Determination of liver phosphorous metabolites and intracellular pH or DzO washout time curves were performed using a 1.89 Tesla, 30-cm horizontalbore cryomagnet with Biospec spectrometer (Oxford Research Systems, Oxford, England and Bruker Instruments, Billerica, MA). ‘rP NMR spectroscopy studies utilized a 2.0-cm diameter three-turn surface coil made from 2-mm thick copper wire, home built by the authors, which was placed on the anterior surface of the liver through a midline abdominal incision. Signal from other sources, particularly skeletal muscle was minimized using cotton gauze packing. Using variable capacitors, the probe was then turned to the resonance frequency of 31P (32.6 MHz) at the field strength used. After positioning the rat in the homogenous volume of the magnet, the Be field was shimmed on a water proton signal from the rat liver until the line width was less than 60 Hz. 3’P MRS spectra were then continuously acquired every 20 min using a sweep width of 3000 Hz, IK data arrays zero-filled to 4K and 120 transients employing a pulse width of 15 ps (corresponding to go-degree-tip at the center of the coil) and repetition time of 4 sec. Based on the spin-lattice relaxation time (Tr) of beta ATP being less than 1 sec,i6 these acquisition parameters can be considered “fully relaxed” for beta ATP. The free induction decay was then Fourier transformed following exponential multiplication with 10 Hz line broadening. Based on the relatively low intensity of the creatine phosphate resonance observed on these spectra, we estimate that >95% of our observed adenosine triphosphate (ATP) and inorganic phosphate (Pi) was derived from liver tissue and not skeletal muscle. Quantitation of liver tissue ATP was then compared between each mode of ventilation using the absolute intensity of the beta ATP peak integrated using a multiple Lorenzian linefit routine (Linesim@ version 880101 .O, Spectrospin) compatible with the Aspect 3OtIOcomputer system (Bruker Instruments, Billerica, MA) (Fig. 1). Comparisons of ATP concentrations observed with the different modes of
Controlled
ventilation during NMR spectroscopic studies 0 W.
EXPERIMENTAL
I
I
10.0
5.0
I
0.0
I
I
I
I
-5.0
-10.0
-15.0
-20.0
PPM
Fig. 1. Representative 31P NMR liver spectrum and simulation linefit. Experimental spectrum obtained with 150 transients employing a pulse width of 15 psec, 4-set relaxation delays, a sweep width of 3000 Hz, and 1 K data arrays. The free induction decay (FID) was deconvoluted using the convolution difference method prior to Fourier transformation. Specifically, the FID was multiplied by 10 Hz line broadening for signal to noise enhancement after subtracting an FID broadened with 200 Hz line broadening for resolution enhancement. Peak assignments 1-phosphomonoesters, 2inorganic phosphate, 3-phosphodiesters, 4-gamma ATP, 5-alpha ATP, 6-beta ATP, and 7-creatine phosphate. Sim-
ulation spectrum generated from Lorenzian lines fit to experimental spectrum by the simplex method of iteration.
ventilation with that observed under control conditions (free breathing period 1, vida infra) assume that the coil position, Q of the coil and other factors important in determining signal intensity did not change during the course of the experiment. Estimates of ATP concentration were reproducible within 10% error. The intracellular pH was estimated from the chemical shift of inorganic phosphate. Because the observed creatine phosphate peak intensities were low, the gamma resonance from ATP was used as an internal pH standard set at -2.5 ppm as suggested by Gadian. ” Simulation peaks obtained with our linefitting routine were used for these chemical shift assignments. Intracellular pH was then calculated based on the formula pHi = 6.9 + log,, ((csPi-3.4)/(5.7-csPi)) where csPi is the chemical shift of inorganic phosphate in ppm, 3.4 and 5.7 are the experimentally determined intercepts
WHALEN AND
J.1.
SHAPIRO
231
and 6.9 is the experimentally determined pKa for the chemical shift of inorganic phosphate observed in a pH titration curve performed on solutions with pH ranging from 4 to 9 (unpublished data). Because of signal to noise limitations, the reproducibility of these intracellular pH determinations was only within 0.08 pH units. This method of assigning the intracellular pH agreed quite well with that described by Ackerman and coworkers using the proton resonance frequency as an internal standard. “-*O *H magnetic reso nance spectroscopy studies were performed to measure hepatic blood flow using the D20 washout technique described by Ackerman and Briefly, a mesenteric vein was cannucolleagues. 21~22 lated with PE-10 tubing, a *H 1.5 cm diameter three turn surface coil made from 2-mm thick copper wire also homebuilt by the authors was placed on the liver in an identical fashion to that described for “P NMR spectroscopy above. The probe was tuned to the frequency of *H (12.3 MHz) and the B. field was shimmed, again using the ‘H signal from tissue water until the linewidth was less than 60 Hz. ‘H files were obtained with a pulse width of 20 psec, a sweep width of 3000 Hz and a 256-word data array. This small array size was used to expedite postacquisition processing and did not result in any major truncation of the *H free induction decay. An automatic program which collected 25 files of one *H scan every five seconds was initiated. Following collection of two baseline deuterium scans, 0.3 ml of D20 was infused rapidly into the mesenteric vein. These files were converted to spectra using exponential multiplication with 10 Hz linebroadening followed by Fourier transformation. The signal to noise of these *H spectra following injection of D20 exceeded 3O:l. Plotting the logarithm of the intensity of *H signal detected with the liver coil versus time, a calculation of the time constant describing D20 washout was made (Simfit@ version 871201.2, Spectrospin) also available on the Aspect 3000 computer system (Bruker Instruments, Billerica, MA). As the T, for D20 mixed with Hz0 (1: 1 mixture) was 0.43 + .02 set-’ and all of our spectra showed little truncation artifact, we assume that T, or T2 changes which could have resulted from our ventilation maneuvers did not effect the interpretation of our data. A representative stacked plot of *H spectra in a liver blood flow experiment is shown in Fig. 2 with the intensity of the *H signal versus time plotted in Fig. 3. From the time constant (T) derived from this decay curve and the ratio of water weight to total tissue weight of liver tissue (0.83),*’ liver blood flow could be estimated from the formula: Liver blood flow (ml/100 g) = 83/T.*‘***
232
Magnetic Resonance Imaging 0 Volume 9, Number 2, 1991
TIME
(seconds)
Fig. 2. Representative stacked plot of deuterium washout curve used to calculate liver blood flow from a free breathing rat.
Experimental Design For the cardiac output studies animals were allowed to stabilize for 10 minutes after placement on the arterio-venous shunt and incrementing flow to 2.9 ml/min. Three cardiac output measurements were then obtained at three-min intervals and an arterial blood gas determination was performed on the freely breathing animals. Rats were then ventilated using the positive pressure method and the high frequency oscil-
-
D20
Statistics Data were analyzed using two-way analysis of variance to demonstrate differences between methods of ventilation, Individual group means were compared using the Student’s t-test for paired data employing Bonferroni’s correction for multiple comparisons.
EOLUS
I
~5 lE-l+0
3
30
latory ventilation method (with order randomized) with three more output determinations and an arterial blood gas measurement made following 10 min for equilibration. Following these experimental ventilations, measurements were repeated in the freely breathing state to serve as a time control. MRS liver studies were performed in a similar manner. After 10 min stabilization in the magnet, either “P NMR liver spectra or 2H NMR spectra studying the decay of 2H signal versus time were acquired while the animal was breathing freely. After this, a switch was made to positive pressure and high frequency oscillatory ventilation with determinations made after lo-15 min equilibration with each ventilation mode. Following experimental ventilation, either 3’P or 2H NMR spectroscopic studies were repeated with the rat breathing freely.
60
90
120
TIME (seconds)
Fig. 3. Representative deuterium intensity vs time curve during a deuterium washout study. Time constant of decay is used to calculate the liver blood flow (see Methods).
RESULTS
Rats appeared to be stable throughout the entire study based on the comparable results obtained for all parameters measured between free breathing animals
Controlled
ventilation
during
studies 0 W. WHALEN AND
NMR spectroscopic
Table 1. Physiologic and biochemical effects of ventilation
Method FBI PPV HFOV FB2
PaCOz (mmHg)
Art pH 7.40 7.41 7.41 7.39
f t t -t
.02 .03 .03 .03
36 32 37 35
-t ? i f
3 3 5 6
PaOz (mmHg) 67 + 4 71+4 68 f 4 71 + 5
MAP (mm%) 92 87 90 96
+ + t +
(ml/rYZ/kg)
5 7 5 6
633 407 520 588
+ f t f
86 46+ 88 76
J.I.
SHAPIRO
233
method
&c) 109+20 159 + 18* 117 + 28 102 * 22
% [ATP] 100 96 f 5 119 + 12 105 + 7
pHi 7.19 7.11 7.19 7.19
f f + *
.07 .I0 .07 .07
Results expressed as mean + sem. FB(1) and FB(2) refer to freely breathing PPV = positive pressure
ventilation.
HFOV
= high frequency
oscillatory
rats prior to and following controlled ventilation, respectively. ventilation. *p < 0.05 compared with FB(1). Art pH = arterial pH,
n = 15 determinations for each ventilation method. PaC02 = arterial CO* tension, n = 15. PaOz = arterial 02 tension. MAP = mean arterial pressure, n = 15. CO = cardiac output, n = 5 sets of three triplicate determinations. T = time constant of D20 washout curve, n = 5. % [ATP] = percentage of liver tissue concentration of ATP during FB(l), n = 5. pHi = intracellular liver pH, n = 5.
at the beginning of the study (FBl) and at the end of the study (FB2). The different modes of ventilation employed did not differ in their effects on mean arterial pressure or in the resultant arterial pH or pCOZ. In contrast, mechanical ventilation using positive pressure resulted in a 35% reduction in cardiac output. High frequency oscillatory ventilation did not significantly decrease cardiac output. Examining the 2H MRS studies, in freely breathing rats, the liver blood flow (LBF) was estimated to be 46 f 8 ml/100 g (liver tissue weight). The time constant (7’) determined on the D20 washout curve was 33% longer with positive pressure ventilation than with freely breathing rats, implying a reduction in LBF to 32 f 4 ml/100 g (p < 0.05 compared with control). However, high frequency oscillatory ventilation did not significantly effect Tor LBF (43 k 10 ml/100 g, p = NS compared with control). Using 3’P NMR spectroscopy, no differences among the ventilation methods on liver ATP concentration or intracellular pH were noted. All data are summarized in Table 1. The similarity of the ATP concentrations between groups FBI and FB2 support the validity of our assumptions concerning coil placement and coil sensitivity discussed in the Methods. DISCUSSION We observed that despite efforts to minimize the hemodynamic effects of positive pressure ventilation, considerable decreases in cardiac output and liver blood flow were observed compared with free breathing animals. These changes, however, were not sufficient to affect liver ATP concentrations or intracellular pH. This implies that a 30% decrease in liver blood flow does not, in and of itself, cause a derangement in the steady state concentration of ATP and intracellular pH. In contrast to the results observed with positive pressure ventilation, high frequency oscillatory
ventilation had no major effect on any hemodynamic or metabolic parameter studied. The effects of the manner of ventilation on hepatic blood flow could be separated from the effects of laparotomy which is known to decrease splanchnic blood flo~.‘~ Controlled ventilation is essential for many NMR studies, especially those specifically addressing acidbase metabolism. For the most part, positive pressure ventilation has been employed in these investigations performed in other laboratories or in the authors’ labis known oratory. I-* As positive pressure ventilation to have important clinical consequences which are attributed primarily to decreases in left ventricular preload, it is likely that the ventilation maneuvers used in these studies do have hemodynamic consequences, independent of the physiologic maneuvers studied.13*i4 In the normal animals studied, these hemodynamic alterations were not sufficient to effect the hepatic concentrations of phosphorus metabolites or intracellular pH. However, it is quite likely that during physiologic maneuvers which could make the animals more sensitive to decreases in left ventricular preload, more pronounced effects of positive pressure ventilation could be observed. The high-frequency oscillatory ventilation method had less hemodynamic effects which is probably due to the much lower mean airway pressure obtained. Although the expense of this type of ventilation is considerably greater than positive pressure ventilation, it may offer important advantages under some physiologic circumstances. REFERENCES Bates, T.E.; Williams, S.R.; Gadian, D.G. Phosphodiesters in the liver: The effect of field strength on the P-31 signal. Mugn. Reson. Med. 12:145-150; 1989.
Ross, B.D.; Freeman, D.; Chan, L. Contributions of nuclear magnetic resonance to renal biochemistry. Kidney Int. 29:131-141; 1986. Shapiro, J.I.; Whalen, M.; Kucera, R.; Kindig, N.; Fil-
234
4.
5.
6.
7.
Magnetic Resonance Imaging 0 Volume 9, Number 2, 1991
ley, G.; Chan, L. Brain pH responses to sodium bicarbonate and Carbicarb during systemic acidosis. Am. J. Physiof. 256:H1316-H1321; 1989. Freeman, D.; Lowry, M.; Radda, G.K.; Ross, B.D. P31 NMR analysis of the renal response to respiratory acidosis. Biochem. Sot. Trans. 10:399-402; 1982. Kucera, R.R.; Shapiro, J.I.; Whalen, M.A.; Kindig, N.B.; Filley, G.F.; Chan, L. Brain pH effects of NaHC03 and Carbicarb in lactic acidosis. Crit. Care Med. 17: 1320-1321; 1989. Corbett, J.T.; Laptook, A.R.; Hassan, A.; Nunnally, R.L. Quantitation of acidosis in neonatal brain tissue using the P-31 NMR resonance peak of phosphoethanolamine. Magn. Reson. Med. 6:99-106; 1988. Ross, B.D. Acid-base regulation: has 31P any answers?
Contrib. Nephrol. 63:53-59; 1988. 8. Shapiro, J.I.; Whalen, M.; Chan, L. Hemodynamic and
hepatic intracellular pH responses to bicarbonate and Carbicarb during acidosis. Magn. Reson. Med. (in press). 9. Bedran De Castro, M.T.B.; Downey, H.F.; Crystal, G.J.; Bashour, EA. Effect of controlled ventilation on renal and splanchnic blood flows during nicotine. Am. J. Physiol. 248:H360-H365; 1985. 10. Halden, E.; Jakobson, S.; Janeras, L.; Norlen, K. Effects of positive end-expiratory pressure on cardiac output distribution in the pig. Acta Anaesth. Stand. 26:403-408;
1982.
11. Manny, J.; Justice, R.; Hechtman, H.B. Abnormalities in organ blood flow and its distribution during positive end-expiratory pressure. Surgery 85:425-432; 1979. 12. Johnston, E.E. Splanchnic hemodynamic response to passive hyperventilation. J. Appl. Physiol. 38: 156- 162; 1975. 13. Pinsky, M.R. The influence of positive pressure venti-
lation on cardiovascular function in the critically ill. Crit. Care C/in. 1:699-717; 1985. 14. Truog, W.E.; Standaert, T.A. Effect of high-frequency ventilation on gas exchange and pulmonary vacular resistance in lambs. J. Appl. Physiof. 59:1104-l 109; 1985. 15. Trautman, E.D.; Newbower, R.S. The development of indicator-dilution techniques. Trans. Biomed. Eng. 3 1: 800-807; 1984. 16. Shapiro, J.I.; Chan, L. P-31 nuclear magnetic resonance study of urinary obstruction in the rat. J. Clin. Invest. 80:1422-1427; 1987. 17. Gadian, D.G. Nuclear Magnetic Resonance and Its Ap-
plications to Living Systems. Clarendon Press; Oxford;. 1982. 18. Ackerman, J.J.H.; Gadian, D.G.; Radda, G.K.; Wong, G.G. Observation of H-l NMR signals with receiver coils tuned for other nuclides. The optimization of BO homogeneity and a multinuclear chemical shift reference. J. Mag. Reson. 42:498-500; 1981. 19. Moon, R.B.; Richards, J.H. Determination of intracellular pH by P-31 magnetic resonance. J. Biol. Chem. 248:7276-7278; 1973. 20. Roos, A.; Boron, W.P. Intracellular pH. Physiol. Rev. 61:296-434; 1981. 21. Ackerman, J. J.; Ewy, C.S.; Becker, N.N..; Shalwitz,
R.A. Deuterium nuclear magnetic resonance measurements of blood flow and tissue perfusion employing 2H20 as a freely diffusible tracer. Proc. Natl. Acad. Sci. (USA) 84:4099-4102; 1987. 22. Ackerman, J.J.; Ewy, C.S.; Kim, S.G.; Shalwitz, R.A.
Deuterium magnetic resonance in vivo: The measurement of blood flow and tissue perfusion. Ann. N. Y. Acad. Sci. 508:89-98; 1987. 23. Gelman, S. Carbon dioxide and hepatic circulation. Anesthesia anafg. 69:149-151; 1989.
0730-725X/91 $3.00 + .oO Copyright 0 1991 Pergamon Press plc
1991
l Original Contribution
CONTROLLED VENTILATION DURING NMR SPECTROSCOPIC STUDIES: HEMODYNAMIC AND BIOCHEMICAL CONSEQUENCES M. WHALEN
AND
J.I. SHAPIRO
The Giles F. Filley Laboratory of the Webb Waring Lung Institute and Department of Medicine, University of Colorado School of Medicine, Denver, Colorado 80262, USA
The effects of different ventilation methods on cardiac output measured by the indicator-dilution method, liver blood flow measured by a deuterium washout technique using zH nuclear magnetic resonance (NMR) and liver concentrations of ATP and intracellular pH determined with 31P NMR were compared in anesthetized rats. No differences in mean arterial blood pressure were demonstrable with the different modes of ventilation. However, significant drops in cardiac output were observed between freely breathing and animals ventilated with positive pressure but not the high frequency oscillatory method (407 f 46 and 520 + 88 vs. 633 f 86 ml/min/kg, p < 0.05 and p = NS, respectively). Moreover, liver blood flow was significantly reduced during positive pressure but not high frequency oscillatory ventilation compared with free breathing rats (32 f 4 and 43 + 10 vs. 46 f 8 ml/100 g, p < 0.05 and p = NS, respectively). 31P NMR spectroscopy revealed no effects of either ventilation method on tissue ATP or intracellular pH as estimated by the chemical shift of inorganic phosphate. These data suggest that controlled ventilation in normal rats accomplished with standard positive pressure methods is associated with major decreases in cardiac output and liver blood flow despite maintenance of normal blood pressure. High frequency oscillatory ventilation appears to effect less compromise of cardiac output and hepatic perfusion than positive pressure ventilation and may, therefore, be preferable for some biological studies. Keywords: 31P NMR spectroscopy; 31P MRS.
The purpose of this study was to carefully quantitate the hemodynamic responses of the whole animal as well as the hepatic circulatory and metabolic responses to positive pressure ventilation and high frequency oscillatory ventilation under the experimental conditions of in vivo NMR spectroscopic study.
INTRODUCTION In vivo NMR spectroscopy is an important tool in the
investigation of metabolic processes in a variety of biomedical areas. For many of these applications, in particular the study of acid-base disturbances, controlled ventilation is an essential part of the experimental methodology. l-8 However, the manner in which controlled ventilation is accomplished may have significant impact on the underlying physiology and metabolism. Specifically, positive pressure ventilation may induce decreases in right ventricular filling as well as changes in the geometry of the heart, possibly resulting in decreases in cardiac output, mean arterial pressure as well as organ perfusion.9-13 In contrast, high frequency jet ventilation may avoid some of these hemodynamic consequences. I4
METHODS Animals
Sprague Dawley rats (approx. 400-500 g body weight) were used in these experiments. Rats were initially anesthetized with ketamine (0.4 ml, 100 mg/dl) and xylazine (0.4 ml, 20 mg/ml) given intramuscularly. Additional doses of ketamine (0.1 ml) were given as needed to suppress a tail pinch response. The right jugular vein and carotid artery cannulated with
RECEIVED 6/ 15/90; ACCEPTED 9/ 13/90. Acknowledgments-The authors would like to thank Drs. Laurence Chan and Richard Kucera for their helpful suggestions. We would also like to thank Ms. Marsha Underwood and the Infrasonics company for the generous loan of their Infant-Star ventilator. This work was supported by a grant-in-aid from the American Heart Association. Dr.
Shapiro is supported by a Squibb Corporation-American Heart Association Clinician Scientist award. Address all correspondence to Joseph I. Shapiro, M.D., Associate Director, NMR Laboratory, Assistant Professor of Medicine, C-281 University of Colorado Health Sciences Center, 4200 E 9th Ave., Denver, CO 80262, USA.
229
230
Magnetic Resonance Imaging 0 Volume 9, Number 2, 1991
PE-50 tubing for all magnetic resonance studies. For the cardiac output portion of the study both right and left jugular, a carotid artery, femoral artery and vein catheters were placed. All animals were placed on a warming pad to maintain a core temperature of 35”-37°C. Ventilation Animals had a tracheostomy performed (14 ga quick cath) and were ventilated using an Infant Star Ventilator (Infrasonics, Palo Alta, CA) capable of both positive pressure ventilation and high frequency oscillatory ventilation (HFOV) modes of operation. The tubing used for ventilation was made from Tygon tubing having 3/8 and 5/8 inch diameter inside and outside, respectively. The inspiratory and expiratory branches of the circuit were 18 ft in length with a oneway valve system at the animal to ensure no rebreathing of gas occurred. A standard three-way stopcock was used to connect the rat’s tracheostomy tube to the valve arrangement. During pilot studies, PPV and HFOV settings were chosen to place the arterial PC02 about 40 torr with the least deleterious effects on mean arterial pressure. The PPV mode parameters chosen were a tidal volume of 8.0 ml and 42 breaths per minute, a peak inspiratory pressure of 8-10 cmH*O, 2 cmHzO of positive and expiratory pressure and an inspiratory time of 0.14 sec. In HFOV mode 7-10 Hz frequencies were used with an amplitude of 2 cmH*O and 2-3 cmHzO positive end expiratory pressure. Measurements Arterial blood pressure was monitored continuously using a pressure transducer and a chart recorder (Gould Instruments, Cleveland, OH) during each experiment. Arterial blood obtained from either the carotid artery or the femoral artery was analyzed for PO*, PC02 and pH using a blood gas instrument (Radiometer Instruments, Copenhagen, Denmark). Cardiac output measurements were performed using the indicator dilution meth0d.r’ To accomplish this, a densitometer (Waters Inc., Baltimore, MD) interfaced via A/D converter (Metrabyte Corp. Taunton, MA) to a personal computer (IBM PC, International Business Systems, Boca Raton, FL) which recorded the green dye indicator-dilution curves (Cardio-Green@, Hynson, Westcott and Dunning, Inc., Baltimore, MD) that were analyzed to determine cardiac outputs (vida infra). The densitometer was placed ahead of a pulsatile pump which created an arterio-venous shunt between the carotid and right jugular vessels. Bolus injections of green dye (10 ~1) were made through the left jugular catheter using a 500 ~1 syringe equipped
with a 10 ~1 repeating dispenser which generated an evolution and decay curve of arterial green dye concentration. The concentration of green dye vs time was analyzed as described by Trautman and Newboner” so as to integrate only the mono-exponential portion of the concentration decay (A) and ignore the contribution from recirculation. This allowed us to solve the equation CO = Q/A in ml/min where Q was a densitometer reading on a standard concentration of indocyanine green using software written in BASIC by the authors. *’ Magnetic Resonance Spectroscopy Studies Determination of liver phosphorous metabolites and intracellular pH or DzO washout time curves were performed using a 1.89 Tesla, 30-cm horizontalbore cryomagnet with Biospec spectrometer (Oxford Research Systems, Oxford, England and Bruker Instruments, Billerica, MA). ‘rP NMR spectroscopy studies utilized a 2.0-cm diameter three-turn surface coil made from 2-mm thick copper wire, home built by the authors, which was placed on the anterior surface of the liver through a midline abdominal incision. Signal from other sources, particularly skeletal muscle was minimized using cotton gauze packing. Using variable capacitors, the probe was then turned to the resonance frequency of 31P (32.6 MHz) at the field strength used. After positioning the rat in the homogenous volume of the magnet, the Be field was shimmed on a water proton signal from the rat liver until the line width was less than 60 Hz. 3’P MRS spectra were then continuously acquired every 20 min using a sweep width of 3000 Hz, IK data arrays zero-filled to 4K and 120 transients employing a pulse width of 15 ps (corresponding to go-degree-tip at the center of the coil) and repetition time of 4 sec. Based on the spin-lattice relaxation time (Tr) of beta ATP being less than 1 sec,i6 these acquisition parameters can be considered “fully relaxed” for beta ATP. The free induction decay was then Fourier transformed following exponential multiplication with 10 Hz line broadening. Based on the relatively low intensity of the creatine phosphate resonance observed on these spectra, we estimate that >95% of our observed adenosine triphosphate (ATP) and inorganic phosphate (Pi) was derived from liver tissue and not skeletal muscle. Quantitation of liver tissue ATP was then compared between each mode of ventilation using the absolute intensity of the beta ATP peak integrated using a multiple Lorenzian linefit routine (Linesim@ version 880101 .O, Spectrospin) compatible with the Aspect 3OtIOcomputer system (Bruker Instruments, Billerica, MA) (Fig. 1). Comparisons of ATP concentrations observed with the different modes of
Controlled
ventilation during NMR spectroscopic studies 0 W.
EXPERIMENTAL
I
I
10.0
5.0
I
0.0
I
I
I
I
-5.0
-10.0
-15.0
-20.0
PPM
Fig. 1. Representative 31P NMR liver spectrum and simulation linefit. Experimental spectrum obtained with 150 transients employing a pulse width of 15 psec, 4-set relaxation delays, a sweep width of 3000 Hz, and 1 K data arrays. The free induction decay (FID) was deconvoluted using the convolution difference method prior to Fourier transformation. Specifically, the FID was multiplied by 10 Hz line broadening for signal to noise enhancement after subtracting an FID broadened with 200 Hz line broadening for resolution enhancement. Peak assignments 1-phosphomonoesters, 2inorganic phosphate, 3-phosphodiesters, 4-gamma ATP, 5-alpha ATP, 6-beta ATP, and 7-creatine phosphate. Sim-
ulation spectrum generated from Lorenzian lines fit to experimental spectrum by the simplex method of iteration.
ventilation with that observed under control conditions (free breathing period 1, vida infra) assume that the coil position, Q of the coil and other factors important in determining signal intensity did not change during the course of the experiment. Estimates of ATP concentration were reproducible within 10% error. The intracellular pH was estimated from the chemical shift of inorganic phosphate. Because the observed creatine phosphate peak intensities were low, the gamma resonance from ATP was used as an internal pH standard set at -2.5 ppm as suggested by Gadian. ” Simulation peaks obtained with our linefitting routine were used for these chemical shift assignments. Intracellular pH was then calculated based on the formula pHi = 6.9 + log,, ((csPi-3.4)/(5.7-csPi)) where csPi is the chemical shift of inorganic phosphate in ppm, 3.4 and 5.7 are the experimentally determined intercepts
WHALEN AND
J.1.
SHAPIRO
231
and 6.9 is the experimentally determined pKa for the chemical shift of inorganic phosphate observed in a pH titration curve performed on solutions with pH ranging from 4 to 9 (unpublished data). Because of signal to noise limitations, the reproducibility of these intracellular pH determinations was only within 0.08 pH units. This method of assigning the intracellular pH agreed quite well with that described by Ackerman and coworkers using the proton resonance frequency as an internal standard. “-*O *H magnetic reso nance spectroscopy studies were performed to measure hepatic blood flow using the D20 washout technique described by Ackerman and Briefly, a mesenteric vein was cannucolleagues. 21~22 lated with PE-10 tubing, a *H 1.5 cm diameter three turn surface coil made from 2-mm thick copper wire also homebuilt by the authors was placed on the liver in an identical fashion to that described for “P NMR spectroscopy above. The probe was tuned to the frequency of *H (12.3 MHz) and the B. field was shimmed, again using the ‘H signal from tissue water until the linewidth was less than 60 Hz. ‘H files were obtained with a pulse width of 20 psec, a sweep width of 3000 Hz and a 256-word data array. This small array size was used to expedite postacquisition processing and did not result in any major truncation of the *H free induction decay. An automatic program which collected 25 files of one *H scan every five seconds was initiated. Following collection of two baseline deuterium scans, 0.3 ml of D20 was infused rapidly into the mesenteric vein. These files were converted to spectra using exponential multiplication with 10 Hz linebroadening followed by Fourier transformation. The signal to noise of these *H spectra following injection of D20 exceeded 3O:l. Plotting the logarithm of the intensity of *H signal detected with the liver coil versus time, a calculation of the time constant describing D20 washout was made (Simfit@ version 871201.2, Spectrospin) also available on the Aspect 3000 computer system (Bruker Instruments, Billerica, MA). As the T, for D20 mixed with Hz0 (1: 1 mixture) was 0.43 + .02 set-’ and all of our spectra showed little truncation artifact, we assume that T, or T2 changes which could have resulted from our ventilation maneuvers did not effect the interpretation of our data. A representative stacked plot of *H spectra in a liver blood flow experiment is shown in Fig. 2 with the intensity of the *H signal versus time plotted in Fig. 3. From the time constant (T) derived from this decay curve and the ratio of water weight to total tissue weight of liver tissue (0.83),*’ liver blood flow could be estimated from the formula: Liver blood flow (ml/100 g) = 83/T.*‘***
232
Magnetic Resonance Imaging 0 Volume 9, Number 2, 1991
TIME
(seconds)
Fig. 2. Representative stacked plot of deuterium washout curve used to calculate liver blood flow from a free breathing rat.
Experimental Design For the cardiac output studies animals were allowed to stabilize for 10 minutes after placement on the arterio-venous shunt and incrementing flow to 2.9 ml/min. Three cardiac output measurements were then obtained at three-min intervals and an arterial blood gas determination was performed on the freely breathing animals. Rats were then ventilated using the positive pressure method and the high frequency oscil-
-
D20
Statistics Data were analyzed using two-way analysis of variance to demonstrate differences between methods of ventilation, Individual group means were compared using the Student’s t-test for paired data employing Bonferroni’s correction for multiple comparisons.
EOLUS
I
~5 lE-l+0
3
30
latory ventilation method (with order randomized) with three more output determinations and an arterial blood gas measurement made following 10 min for equilibration. Following these experimental ventilations, measurements were repeated in the freely breathing state to serve as a time control. MRS liver studies were performed in a similar manner. After 10 min stabilization in the magnet, either “P NMR liver spectra or 2H NMR spectra studying the decay of 2H signal versus time were acquired while the animal was breathing freely. After this, a switch was made to positive pressure and high frequency oscillatory ventilation with determinations made after lo-15 min equilibration with each ventilation mode. Following experimental ventilation, either 3’P or 2H NMR spectroscopic studies were repeated with the rat breathing freely.
60
90
120
TIME (seconds)
Fig. 3. Representative deuterium intensity vs time curve during a deuterium washout study. Time constant of decay is used to calculate the liver blood flow (see Methods).
RESULTS
Rats appeared to be stable throughout the entire study based on the comparable results obtained for all parameters measured between free breathing animals
Controlled
ventilation
during
studies 0 W. WHALEN AND
NMR spectroscopic
Table 1. Physiologic and biochemical effects of ventilation
Method FBI PPV HFOV FB2
PaCOz (mmHg)
Art pH 7.40 7.41 7.41 7.39
f t t -t
.02 .03 .03 .03
36 32 37 35
-t ? i f
3 3 5 6
PaOz (mmHg) 67 + 4 71+4 68 f 4 71 + 5
MAP (mm%) 92 87 90 96
+ + t +
(ml/rYZ/kg)
5 7 5 6
633 407 520 588
+ f t f
86 46+ 88 76
J.I.
SHAPIRO
233
method
&c) 109+20 159 + 18* 117 + 28 102 * 22
% [ATP] 100 96 f 5 119 + 12 105 + 7
pHi 7.19 7.11 7.19 7.19
f f + *
.07 .I0 .07 .07
Results expressed as mean + sem. FB(1) and FB(2) refer to freely breathing PPV = positive pressure
ventilation.
HFOV
= high frequency
oscillatory
rats prior to and following controlled ventilation, respectively. ventilation. *p < 0.05 compared with FB(1). Art pH = arterial pH,
n = 15 determinations for each ventilation method. PaC02 = arterial CO* tension, n = 15. PaOz = arterial 02 tension. MAP = mean arterial pressure, n = 15. CO = cardiac output, n = 5 sets of three triplicate determinations. T = time constant of D20 washout curve, n = 5. % [ATP] = percentage of liver tissue concentration of ATP during FB(l), n = 5. pHi = intracellular liver pH, n = 5.
at the beginning of the study (FBl) and at the end of the study (FB2). The different modes of ventilation employed did not differ in their effects on mean arterial pressure or in the resultant arterial pH or pCOZ. In contrast, mechanical ventilation using positive pressure resulted in a 35% reduction in cardiac output. High frequency oscillatory ventilation did not significantly decrease cardiac output. Examining the 2H MRS studies, in freely breathing rats, the liver blood flow (LBF) was estimated to be 46 f 8 ml/100 g (liver tissue weight). The time constant (7’) determined on the D20 washout curve was 33% longer with positive pressure ventilation than with freely breathing rats, implying a reduction in LBF to 32 f 4 ml/100 g (p < 0.05 compared with control). However, high frequency oscillatory ventilation did not significantly effect Tor LBF (43 k 10 ml/100 g, p = NS compared with control). Using 3’P NMR spectroscopy, no differences among the ventilation methods on liver ATP concentration or intracellular pH were noted. All data are summarized in Table 1. The similarity of the ATP concentrations between groups FBI and FB2 support the validity of our assumptions concerning coil placement and coil sensitivity discussed in the Methods. DISCUSSION We observed that despite efforts to minimize the hemodynamic effects of positive pressure ventilation, considerable decreases in cardiac output and liver blood flow were observed compared with free breathing animals. These changes, however, were not sufficient to affect liver ATP concentrations or intracellular pH. This implies that a 30% decrease in liver blood flow does not, in and of itself, cause a derangement in the steady state concentration of ATP and intracellular pH. In contrast to the results observed with positive pressure ventilation, high frequency oscillatory
ventilation had no major effect on any hemodynamic or metabolic parameter studied. The effects of the manner of ventilation on hepatic blood flow could be separated from the effects of laparotomy which is known to decrease splanchnic blood flo~.‘~ Controlled ventilation is essential for many NMR studies, especially those specifically addressing acidbase metabolism. For the most part, positive pressure ventilation has been employed in these investigations performed in other laboratories or in the authors’ labis known oratory. I-* As positive pressure ventilation to have important clinical consequences which are attributed primarily to decreases in left ventricular preload, it is likely that the ventilation maneuvers used in these studies do have hemodynamic consequences, independent of the physiologic maneuvers studied.13*i4 In the normal animals studied, these hemodynamic alterations were not sufficient to effect the hepatic concentrations of phosphorus metabolites or intracellular pH. However, it is quite likely that during physiologic maneuvers which could make the animals more sensitive to decreases in left ventricular preload, more pronounced effects of positive pressure ventilation could be observed. The high-frequency oscillatory ventilation method had less hemodynamic effects which is probably due to the much lower mean airway pressure obtained. Although the expense of this type of ventilation is considerably greater than positive pressure ventilation, it may offer important advantages under some physiologic circumstances. REFERENCES Bates, T.E.; Williams, S.R.; Gadian, D.G. Phosphodiesters in the liver: The effect of field strength on the P-31 signal. Mugn. Reson. Med. 12:145-150; 1989.
Ross, B.D.; Freeman, D.; Chan, L. Contributions of nuclear magnetic resonance to renal biochemistry. Kidney Int. 29:131-141; 1986. Shapiro, J.I.; Whalen, M.; Kucera, R.; Kindig, N.; Fil-
234
4.
5.
6.
7.
Magnetic Resonance Imaging 0 Volume 9, Number 2, 1991
ley, G.; Chan, L. Brain pH responses to sodium bicarbonate and Carbicarb during systemic acidosis. Am. J. Physiof. 256:H1316-H1321; 1989. Freeman, D.; Lowry, M.; Radda, G.K.; Ross, B.D. P31 NMR analysis of the renal response to respiratory acidosis. Biochem. Sot. Trans. 10:399-402; 1982. Kucera, R.R.; Shapiro, J.I.; Whalen, M.A.; Kindig, N.B.; Filley, G.F.; Chan, L. Brain pH effects of NaHC03 and Carbicarb in lactic acidosis. Crit. Care Med. 17: 1320-1321; 1989. Corbett, J.T.; Laptook, A.R.; Hassan, A.; Nunnally, R.L. Quantitation of acidosis in neonatal brain tissue using the P-31 NMR resonance peak of phosphoethanolamine. Magn. Reson. Med. 6:99-106; 1988. Ross, B.D. Acid-base regulation: has 31P any answers?
Contrib. Nephrol. 63:53-59; 1988. 8. Shapiro, J.I.; Whalen, M.; Chan, L. Hemodynamic and
hepatic intracellular pH responses to bicarbonate and Carbicarb during acidosis. Magn. Reson. Med. (in press). 9. Bedran De Castro, M.T.B.; Downey, H.F.; Crystal, G.J.; Bashour, EA. Effect of controlled ventilation on renal and splanchnic blood flows during nicotine. Am. J. Physiol. 248:H360-H365; 1985. 10. Halden, E.; Jakobson, S.; Janeras, L.; Norlen, K. Effects of positive end-expiratory pressure on cardiac output distribution in the pig. Acta Anaesth. Stand. 26:403-408;
1982.
11. Manny, J.; Justice, R.; Hechtman, H.B. Abnormalities in organ blood flow and its distribution during positive end-expiratory pressure. Surgery 85:425-432; 1979. 12. Johnston, E.E. Splanchnic hemodynamic response to passive hyperventilation. J. Appl. Physiol. 38: 156- 162; 1975. 13. Pinsky, M.R. The influence of positive pressure venti-
lation on cardiovascular function in the critically ill. Crit. Care C/in. 1:699-717; 1985. 14. Truog, W.E.; Standaert, T.A. Effect of high-frequency ventilation on gas exchange and pulmonary vacular resistance in lambs. J. Appl. Physiof. 59:1104-l 109; 1985. 15. Trautman, E.D.; Newbower, R.S. The development of indicator-dilution techniques. Trans. Biomed. Eng. 3 1: 800-807; 1984. 16. Shapiro, J.I.; Chan, L. P-31 nuclear magnetic resonance study of urinary obstruction in the rat. J. Clin. Invest. 80:1422-1427; 1987. 17. Gadian, D.G. Nuclear Magnetic Resonance and Its Ap-
plications to Living Systems. Clarendon Press; Oxford;. 1982. 18. Ackerman, J.J.H.; Gadian, D.G.; Radda, G.K.; Wong, G.G. Observation of H-l NMR signals with receiver coils tuned for other nuclides. The optimization of BO homogeneity and a multinuclear chemical shift reference. J. Mag. Reson. 42:498-500; 1981. 19. Moon, R.B.; Richards, J.H. Determination of intracellular pH by P-31 magnetic resonance. J. Biol. Chem. 248:7276-7278; 1973. 20. Roos, A.; Boron, W.P. Intracellular pH. Physiol. Rev. 61:296-434; 1981. 21. Ackerman, J. J.; Ewy, C.S.; Becker, N.N..; Shalwitz,
R.A. Deuterium nuclear magnetic resonance measurements of blood flow and tissue perfusion employing 2H20 as a freely diffusible tracer. Proc. Natl. Acad. Sci. (USA) 84:4099-4102; 1987. 22. Ackerman, J.J.; Ewy, C.S.; Kim, S.G.; Shalwitz, R.A.
Deuterium magnetic resonance in vivo: The measurement of blood flow and tissue perfusion. Ann. N. Y. Acad. Sci. 508:89-98; 1987. 23. Gelman, S. Carbon dioxide and hepatic circulation. Anesthesia anafg. 69:149-151; 1989.
