Arterial Peaks in Regional Cerebral Blood Flow 133 Xenon Clearance Curves

Transcription

1 Arterial Peaks in Regional Cerebral Blood Flow 133 Xenon Clearance Curves BY STEPHEN G. ROSENBAUM, B.A., LINNETTE D. ILIFF, B.SC, M. PHIL., PH.D., J. W. D. BULL, M.D., F.R.C.P., F.F.R., G. H. DU BOULAY, M.B., M.R.C.P., F.F.R., JOHN MARSHALL, M.D., F.R.C.P. (ED.), F.R.C.P. (LOND.), D.P.M., R. W. ROSS RUSSELL, M.A., M.D., D.M., F.R.C.P., LINDSAY SYMON, F.R.C.S. (ED.), F.R.C.S. (ENG.) Abstract: Arterial Peaks in Regional Cerebral Blood Flow "'Xenon Clearance Curves In patients with normal rcbf, arterial spikes are found in all regions of the cerebral hemisphere. The normal range of spike height is 17% to 31 % of H max, except over the carotid siphon (region 15), when it is 36% to 38%. Under general anesthesia proportionate spike height increases significantly when CBF is low. Increased blood flow is not, however, significantly associated with decrease in spike height. The possible explanations for this are discussed. Additional Key Words scintillation counters isotope cerebral hemisphere arteriovenous malformations carotid siphon In examining the two-minute clearance curves during regional cerebral blood flow studies (rcbf), at least two authors have made mention of peaks occurring in the temporal region the "carotid peak." 1 ' 2 Paulson et al. 1 described these as "arterial peaks" and distinguished them from "tissue peaks" on the basis of their duration and the shape of the subsequent clearance curve. The "arterial peaks" last only three to five seconds, are about 10% of the maximum recorded counts (H max), and are followed by the usual bi-exponential clearance curve. In contrast, the "tissue peaks" are longer, though not usually exceeding one minute in duration, and are distinguished by a multi-exponential clearance curve which has a definite inflection. Both Paulson et al. 1 and Palvolgyi 2 seem to restrict the normal occurrence of the "arterial peak" to the region of the carotid siphon, believing it to represent the immediate passage of the bolus of radioactive isotope through the siphon into the superficial vessels of the hemisphere. Similar but much larger arterial peaks have been noted by Paulson et al. 1 to be associated with arteriovenous malformations or highly vascular tumors. The purpose of this study was to investigate "arterial peaks" (referred to hereafter as "spikes") more closely. In particular, we were concerned with From the Institute of Neurology, National Hospital for Nervous Diseases, Queen Square, London, WC1N 3BG, England. Stroke, Vol. 4, January-February 1973 defining those regions in which arterial spikes are commonly present, what constitutes a "normal" arterial spike, and the relation of the spike height to any abnormalities in the cerebral blood flow parameters. /Methods The data for this study were obtained from rcbf studies conducted under both local anesthesia (LA) and general anesthesia (GA) as described by Wilkinson et al. 3 Following the intracarotid injection of 8 to 10 me of radioactive Xenon ( 133 Xe) dissolved in 5 ml saline, the subsequent clearance of the isotope from cerebral tissue was recorded by 16 scintillation counters placed on the side of the head in the pattern shown in figure 1. The falling counts were recorded for 15 minutes on magnetic tape. The first two minutes of the clearance curves provided the data for observation of arterial spikes. Traces of these were obtained by playing back from the magnetic tape via a pen-recorder, two channels at a time, at a time-constant of 0.2 second. A total of 115 cases was examined, 56 under local anesthesia and 59 under general anesthesia. Since it is known that general anesthesia affects overall cerebral blood flow, 4 data from the two groups of patients have been kept separate. Tables 1 and 2 provide details of the patients studied in respect to diagnosis, age, sex, mean arterial carbon dioxide tension in mm Hg (Pi^02 ), and mean blood pressure (BP) during the study. Patients under local anesthesia were studied at their resting Paco. The data included patients with both "normal" and "abnormal" cerebral blood flow. An abnormal region was defined as one which differed from the mean 73

2 ROSENBAUM, ILIFF, BULL, DU BOULAY, MARSHALL, RUSSELL, SYMON and (4) F init, representing the "initial log slope" flow index estimated from the gradient of the first two minutes of the clearance curve. FIGURE l Position of the scintillation counters in relation to the skull. hemisphere flow by a significantly greater amount (2.6 or more standard deviations) than did the corresponding region in a group of normals. 3-4 A spike height (b in fig 2) was calculated as the portion of H max (a in fig. 2) above the zero intercept of a straight line (c to d in fig. 2) drawn through the subsequent clearance curve. This was expressed as a percentage of the total H max ( X 100). The two-minute clearance curves for b each of the 16 areas of the 115 cases were examined for spikes. Normal and abnormal regions of flow were included in the study. Four flow parameters were utilized: (1) Fg (ml/100 gm/min), which represents the flow through fast-clearing tissue (mostly gray matter); (2) Fw (ml/100 gm/min), representing the flow through slow-clearing tissue (mostly white matter); (3) F (ml/100 gm/min), representing the mean weighted flow which was derived from Fg, Fw and relative weights of perfused gray and white matter, TABLE 1 Details of 56 Patients Studied Under Local Anesthesia Age in years Diagnosis Completed stroke Transient ischemic attack (TIA) Epilepsy Others* Number of cases (moan) 55 (42-69) 51 (43-62) 39 (18-57) 50 (28H52) "Others" include: dementia, optic or pituitary lesions, and migraine. 74 Results Tables 3 and 4 show the total number of normal and abnormal rcbf observations, as measured by each of the parameters, under local and general anesthesia, respectively. There were more normal regions in awake patients, patients under general anesthesia having a significantly larger number of abnormal flow observations (indicated in tables 3 and 4 by an arrow up or down). Tables 5 and 6 show the proportionate spike heights in all regions including those with normal and abnormal flow. Notice that the data show spikes to be present in all 15 regions during normal flow, the height of the spikes varying between 17% to 38% ofh max. The spikes were more frequent over regions 11 and 15 (which correspond to the area of the carotid siphon) and moreover were significantly higher in those areas amounting on average to 38% of H max in area 15. Normal proportionate spike height was not significantly different during local or general anesthesia. The number of spikes in areas of abnormal flow under local anesthesia (table 5) is insufficient to show meaningful differences from the normal. During general anesthesia (table 6), however, the mean proportionate spike height (when based on five or more observations) showed a substantial increase in areas 11, 15 and 16; it is noteworthy that this increase in proportionate spike height was associated with abnormally low flow. When the proportionate spike heights in the regions of abnormality were pooled and compared with the spikes in the pooled normal regions, this inverse relationship of proportionate spike height to flow was shown to be highly significant (p < 0.001) (table 7). Proportionate spike height was by contrast not significantly different in areas of high flow. Sex 15 M 7F 12 M 1 F 4M 1 F 12 M 4F Moan Paco 2 (mm Hfl) 42 (33-49) 41 (33-45) 44 (42-49) 42 (31-50) Mean blood pressure (mm Ho,) 101 (75-150) 99 (75-130) 83 (74-90) 81 (57-152) Stroke, Vol. 4, January-February 7973

3 ARTERIAL PEAKS IN rcbf TABLE 2 Details of 59 Patients Studied Under General Anesthesia Diagnosis Completed stroke Transient ischemic attack (TIA) Subarachnoid hemorrhage (SAH) Epilepsy Others* Number of cate» Age in years (mean) 50 (23-68) 52 (34-63) 49 (28-67) 24 (22-26) 47 (32-68) Sex 12 M 4 F 6M 3 F 9M 9F 1 M 4F 8M 4 F Mean Paco* (mm Hg) 40 (20-51) 43 (38-50) 39 (26-45) 37 (22-51) 39 (27-52) Mean blood pressure (mm Hg) 88 (65-140) 82 (60-115) 92 (65-110) 74 (56-100) 82 (63-130) "Others" include: dementia, optic or pituitary lesions, and migraine. Discussion The first finding of importance in this investigation is that arterial spikes are commonly present in all regions of the hemisphere. This is so even when the flow in these regions is normal. The proportionate spike heights were 17% to 31% of the H max except in region 15, where the spike was 36% and 38% in the local and general anesthetic groups, respectively. The spikes undoubtedly indicate passage of the isotope through superficial pial arteries and are not representative of tissue perfusion. They do not indicate pathology; the spikes seen over arteriovenous malformations (none of which were included in this study) are much higher, sometimes amounting to 100% of H max. The fact that the spikes in region 15 are higher than elsewhere, though not FIGURE 2 Calculation of spike height. Regional count rate on the y axis is plotted against time for the first two minutes of the measurement on the x axis, f = Time of injection of lss Xe, a = total height (H max), b =z spike height to be expressed as % of "a," c to d line approximating subsequent curve. Stroke, Vol. 4, January-February 1973 reaching the dimensions seen over arteriovenous malformations, is because this region includes the carotid siphon. The second finding of importance concerns the relationship of the proportionate spike height to abnormalities in cerebral blood flow. Under general anesthesia there is a significant increase in proportionate spike height in association with decreased blood flow in focal areas where there were sufficient data to permit statistical analysis. There was no statistically significant alteration of spike height when blood flow was increased. The reason for the increased proportionate spike height with low tissue perfusion may be that, provided the volume of blood passing beneath the detector remains constant, low perfusion will be associated with slow transit, hence prolongation of the superficial (or pial) phase. This will increase the number of counts per unit time recorded by the detector. Furthermore, there will be more time before Xenon, which has entered the tissue, starts to be cleared; the takeoff of the delayed clearance curve will therefore be at a lower level, causing the proportion of spike height to H max to be increased. In considering this explanation it is important to bear in mind that, since detection of counts falls off rapidly with distance from the counter, superficial vessels will be the major contributory source of counts. An alternative hypothesis is that ischemic changes in the vessel wall associated with the low flow may interfere with the passage of the isotope from the vascular to the tissue compartment. A greater proportion of isotope may be retained in the vascular compartment than in the normal situation. This again may be reflected in an increase in the proportionate spike height. A third possibility causing retention of a greater proportion of isotope within the vascular compartment hence higher proportionate spike height is 75

4 TABLE 3 Number of Regional Blood Flow Observations in 56 Patients Under Local Anesthesia Reglorin Normal flow Fg F Fw T Fw F t F init T F init Area 14 was discarded entirely because of technically unsatisfactory data. indicates significant increase, &significant decrease in flow. TABLE 4 Number of Regional Blood Flow Observafions in 59 Patients Under General Anesthesia Pegtons b 16 Normal flow '" Fg FwT i F T F F init C g Area 14 was discarded entirely because of technically unsatisfactory data. f indicates significant increase, & significant decrease in flow. 4 " S W

5 L m TABLE 5 b a X, 3 Mean Regional Spike Height (%), Local Anesthesia - 56 Cases - 2 z ' t e Replons 1 2 I A w Normal Row '" Standard error No. of observations Fg T Standard error No. of observations Fg J " Standard errors were calculated only for five or more observations. f indicates significant increase, & significant decrease in flow

6 ROSENBAUM, ILIFF, BULL, DU BOULAY, MARSHALL, RUSSELL, SYMON Stroke, Vol. 4, January-February 1973

7 ARTERIAL PEAKS IN rcbf TABLE 7 Comparison of Accumulated Spike Heights in Areas of Normal and Abnormal Flow Parameters Under General Anesthesia Mean splks ht (%) S.E. of mean N.. of fpiket p Normal flow Fg 1 Fg. Fw Fw. F T F i F init T F init l.i indicates significant increase, i significant decrease in flow >0.9 < >0.1 < >0.1 < >0.1 <0.01 arteriovenous shunting. This is believed not to be present in the normal cerebral circulation. Extremely high spikes have been demonstrated by several centers in patients with known angiomata or arteriovenous malformations which may be considered as exaggerated forms of the phenomenon we are considering here. It may be that under pathological conditions associated with low flow, arteriovenous shunting, of lesser degree than that seen in arteriovenous malformations, appears. Whichever hypothesis is correct, it is clear that spikes may be seen in any area of the brain and do not necessarily indicate the presence of a pathological condition. The significance of spikes can be interpreted only in conjunction with other parameters of flow. Acknowledgment This work was supported by the National Fund for Research into Crippling Diseases and the Medical Research Council. References 1. Paulson OB, Cronqvist S, Risberg J, et al: Regional cerebral blood flow: A comparison of 8-detector and 16-detector instrumentation. J Nucl Med 10: , Palvolgyi R: Regional cerebral blood flow in patients with intracranial tumors. J Neurosurg 31: , Wilkinson IMS, Bull JWD, Du Boulay GH, et al: Regional blood flow in the normal cerebral hemisphere. J Neurol Neurosurg Psychiat 32: , Wilkinson IMS, Browne DRG: The influence of anaesthesia and of arterial hypocapnia on regional blood flow in the normal human cerebral hemisphere. Brit J Anaesth 42: , 1970 Stroke, Vol. 4, January-February