Musical murmurs in human cerebral arteries after subarachnoid hemorrhage

Transcription

1 J Neurosurg 60:32-36, 1984 Musical murmurs in human cerebral arteries after subarachnoid hemorrhage RUNE AASLID, PH.D., AND HELGE NORNES, M.D. Clinic of Neurosurgery, University Hospital, Bern, Switzerland ~," A transcranial ultrasonic method for the recording of murmurs from cerebral vessels is described. Using the new approach the authors have observed musical murmurs of pure tone quality in 15 patients with increased flow velocities in the cerebral arteries after spontaneous subarachnoid hemorrhage (SAH). The frequency range of the pure tones was from 140 to 820 Hz, corresponding to flow velocities between 73 and 215 cm/sec. The musical murmurs occurred as a transitional state between silent flow and the well known phenomenon of bruit. They were observed between the 4th and the 20th day after SAH. The most likely cause of the musical murmur is a periodic shedding of vortices in the cerebral arteries, commonly referred to as "a von Kfirmfin vortex street." Clinically the presence of musical murmurs indicated that pathologically increased blood velocities were present in the artery under investigation. This probably reflected the degree of spasm. KEY WORDS ~ ultrasound recording 9 subarachnoid hemorrhage 9 blood flow velocity A TRANSCRANIAL ultrasound Doppler technique for recording blood flow velocity in basal cerebral arteries has been described previously.~ 2 In the course of a routine recording in a patient with spontaneous subarachnoid hemorrhage (SAH), we noticed tones of a musical quality from the loudspeaker of the instrument. We realized that an instrument designed to detect Doppler shifts would also act as a demodulator for phase- and amplitude-modulated ultrasonic signals, and thus could be used as a focused microphone. This type of ultrasonic detection is used routinely for the recording of fetal heart sounds. The present paper describes our findings in a series of patients with spontaneous SAH, using the transcranial ultrasonic approach. Recording Principle Clinical Material and Methods A range-gated Doppler instrument emits regularly repeated bursts of ultrasound. Vibrations of anatomical structures within the ultrasonic beam will cause phase shifts in the reflected echoes, and thus modulate the signal received. The frequency of the vibration can be obtained through demodulation in the same way as the Doppler shift signal is extracted. The phase-modulated signals from vibrations can be differentiated from Doppler signals by directional frequency analysis. The Doppler frequency shift originating from the velocity of the flow is unidirectional, having frequency components on only one side of the baseline (see Fig. 2) while the spectral components from vibrations are symmetrical around the baseline. Furthermore, the intensity of Doppler signals is relatively constant because the reflected ultrasonic energy from the blood is independent of the flow velocity. In contrast, the intensity and character of the phase-modulated signals from the vibrating structures are strongly dependent on the velocity of the flow, as shown in Fig. 2. Because the instrument is sensitive only to phenomena within a relatively small sample volume (in our system, a cylinder approximately 5 mm in diameter and 10 mm in length), noises and interference from other locations are avoided. Thus, the vibrations can be located by scanning the area for the maximum amplitude of the signal. This noninvasive procedure is somewhat similar to probing the exposed arteries with a miniature microphone during surgery: By holding the ultrasonic transducer over the vibrating membrane of a loudspeaker excited by a sinusoidal voltage, we could verify that the fundamental harmonic frequency of the vibration was correctly rendered in the output from the Doppler instrument. However, with large amplitude vibrations, overharmonics in the spectral distribution were produced. Such overharmonics were created by the modulation/demodulation process when the modulation depth increased; this phenomenon is encountered when phase modulation is used in 32 J. Neurosurg. / Volume 60 / January, 1984

2 Ultrasonic artery recording after SAH TABLE 1 Summary of hemodynamic data in patients with musical murmurs* Case Sex, Location of Occurrence of Recording Musical M ur- Frequency Corresponding Range of Musi- Range of Blood No. Age (yrs) Diagnosis Musical Period (days murs (days cal Murmurs Velocity Murmurs after SAH) after SAH) (Hz) (cm/sec) 1 F, 47 aneurysm, rt ICA rt bif M, 54 aneurysm, ACoA rt bif F, 38 aneurysm, ACoA rt bif F, 22 aneurysm, ACoA rt bif It bif F, 33 aneurysm, rt MCA rt bif M, 48 aneurysm, It MCA rt bif M, 34 aneurysm, basilar rt PCA M, 53 aneurysm, basilar l! bif l 17 9 M, 43 aneurysm, rt MCA rt MCA F, 45 aneurysm, rt MCA rt MCA F, 51 no aneurysm rt bif F, 45 no aneurysm rt bif F, 41 no aneurysm It bif t 14 M, 55 no aneurysm It bif F, 60 no aneurysm It ACA average * SAH = subarachnoid hemorrhage; 1CA = internal carotid artery; MCA = middle cerebral artery; ACA = anterior cerebral artery; ACoA = anterior communicating artery; bif = bifurcation of ICA into MCA and ACA; PCA = posterior cerebral artery. ";- No reliable recordings of flow velocity could be made in Case 13. radio Communication. Thus, the presence of higher harmonics in the demodulated signal does not necessarily mean that these are present in the original murmurs. Clinical Material The study was conducted on a consecutive series of 29 patients with recent spontaneous SAH. All patients had a full angiographic investigation. Fifteen patients demonstrated musical murmurs in one or more of the transcranial ultrasonic recordings. A total of 178 Doppler investigations were made in this group (Table 1). A saccular aneurysm was verified as the source of hemorrhage in 10 of these patients. The 14 patients without any musical murmur were subjected to a total of 128 recordings. Angiography disclosed an aneurysm in eight of these patients. The aneurysms, all verified as the source of hemorrhage, were located on the anterior communicating artery in tbur patients, the pericallosal arted in two, the internal carotid artery in one, and the carotid-ophthalmic artery in one. Investigation Procedure All patients underwent transcranial Doppler ultrasound investigations at intervals of 1 to 2 days, sometimes several times per day. The procedure for flow velocity recording in basal cerebral arteries has been reported in detail previously. 2 The best area for penetration of ultrasound was located in the temporal region, and the Doppler frequency shifts in the middle (MCA), anterior (ACA), and posterior cerebral arteries (PCA) were recorded with range-gated Doppler techniques. Musical murmurs were recorded with the same instrument. When such phenomena were heard among the Doppler signals, then the relevant section of the artery was scanned to locate the best signal. Flow velocities in these locations were also determined. The signal from the Doppler instrument was taperecorded. Real-time spectral analysis was performed by a fast-fourier transform.* The frequency resolution of this instrument was 40 Hz. The frequency range of the pure tone was determined from the spectral display, together with the harmonic of corresponding flow velocity range. The velocity readings were not corrected for the angle of incidence between the ultrasonic beam and the direction of flow. The signal was transferred to a digital computert using a sampling frequency of 4 khz and a resolution of 8 bits. A standard fast-fourier algorithm was used to obtain a frequency resolution of 16 Hz. Results Spectral analyses of the musical murmurs in Case 1, are shown in Fig. 1. This patient demonstrated a relatively high-frequency tone of 560 Hz in systole with a silent diastole. The corresponding velocity was 215 cm/ *Angioscan with forward/reverse flow adapter TRA-I, manufactured by Unigon Industries Inc., Mr. Vernon, New York. qhp-85 with HP Math-Pac manufactured by Hewlett-Packard Corporation, Corvallis, Oregon. J. Neurosurg / Vohtme 60/January

3 R. Aaslid and H. Nornes FIG. 1. Case 1. A: Spectral display of the musical murmurs (lower) and the Doppler signal (upper) in the region where the right internal carotid artery divides into the middle and anterior cerebral arteries. The musical murmur had a relatively constant frequency in systole. The diastole was silent. B: High-resolution (t6 Hz) spectral analysis of the musical murmur in mid-systole. The pure tone quality is seen as a narrow spectral line at 560 Hz. sec. The MCA exhibited extended spasms in the angiogram taken 2 days before detection of the sounds. Tones were also present in the recordings on the following day. Subsequently the sounds disappeared and the velocities decreased, indicating a decline in spasm. A highresolution spectral analysis of the signal in systole (Fig. 1B) demonstrated a very narrow spectral component. In Case 2 (Fig. 2), the musical murmur was found in early diastole. The systole was characterized by a strong broad-banded noise or bruit, while the late diastole was silent. A second harmonic of the tone was also seen in the spectral display; however, as discussed above, this was probably produced by the signal processing. Musical murmurs with pure tone qualities were heard in 37 recordings from 15 patients (Table 1). In five of the patients, no aneurysm was demonstrated in the angiograms. In two patients (Cases 5 and 6), the musical murmurs were present both before and after the surgical procedure of aneurysm clipping. In one patient (Case 9), the sounds appeared after operation. No musical murmurs were heard from any patient during the first 3 days after SAH or later than 3 weeks after the first bleed. There was no apparent correlation between the clinical grading of the patient and the presence of a musical murmur. In the individual case, the musical murmurs were present only over a comparatively narrow range of velocities. Furthermore, the musical murmur was never continuously present in both systole and diastole. All patients who demonstrated tones during diastole had broad-banded noise during systole. On the other hand, the diastole was silent in patients who demonstrated the musical murmurs during systole. The relationship between velocity and the frequency of the fundamental harmonic of the tone is shown in Fig. 3. In the individual case, there seems to be an almost direct proportionality between flow velocity and the frequency of the musical murmur. The correlation between these two variables in the present cases was relatively good (r = 0.74). In the entire series, the mean frequency of the murmur was 352 Hz, while the corresponding velocity was 137 cm/sec. In 12 of the patients, the anatomical location of the musical murmur was the area near the bifurcation of the ICA into the MCA and the ACA. Two of the patients (Cases 9 and 10) demonstrated musical murmurs in the distal MCA near the bifurcation. Case 15 presented musical murmurs from the ACA near the midline and, in Case 7, they were found to originate from the PCA. 34 J. Neurosurg. / Volume 60 / January, 1984

4 Ultrasonic artery recording after SAH FIG. 2. Spectral display of the musical murmur (lower) and the Doppler signal (upper) from Case 2. The recordings were made from the right middle cerebral artery, slightly distal of the bifurcation of the internal carotid artery. Three sound modes are recognized in the lower tracing: A systolic bruit ( 1 ), a musical murmur displayed as a narrow band in early diastole (2), and a silent phase in late diastole (3). The second harmonic of the musical murmur is also seen in the recording, this is probably an artifact from the ultrasonic modulation/demodulation process. Since an intracranial aneurysm has been suggested as the possible origin ofintracranial musical murmurs, 4 7 ~~ we made an effort to scan the location of the aneurysmal sac with the sample volume of the ultrasonic instrument. We sometimes detected an abnormal bruit or a thump from this area, but in no cases did we find murmurs with musical qualities like those reported above. Discussion Sounds generated by the cardiovascular system in general have a noisy character which is qualitatively different from the pure tone of a musical instrument. However, pure tones have been observed both in the cerebral and the peripheral circulations. 7"8""~ The hypothesis that musical murmurs are produced by aneurysms acting as Hemholtz resonators 1~ finds no support in our results: in no case was the anatomical site of the murmurs associated with the aneurysm as revealed on angiography. There is also speculation that a periodic formation of vortices, a so-called "yon Kfirmfin trail" can be responsible for the generation of pure tones. 3 6 Ft6. 3. The relationship between the velocity of the blood flow and the fundamental frequency of the musical murmur in this series of patients (excluding Case 13). Each case is represented by two points connected with a line indicating the range of frequencies and velocities observed. The dotted line, y = x, is found by regression analysis of all points: r = A theory that all cardiovascular murmurs are produced by such vortex formation and not by turbulence was presented by Bruns. 3 This is referred to as the "Aeolian theory of cardiovascular murmurs," named after an ancient musical instrument that produced tones when placed in the wind. The classical von Kfirmfin vortex street can be demonstrated at low Reynolds numbers (laminar flow) in the wake behind a cylinder. ~ Vortices with alternating rotational directions are shed with high regularity, and they create periodic fluctuations in the fluid pressure. Proportionality between the frequency of vortex shedding and the flow velocity has been found in such experiments. ~3 In our material, the proportionality between the frequency of the musical murmurs and the flow velocity was striking, and suggests that a periodic vortex street was responsible for this phenomenon. The musical murmurs occurred in a transitional phase between silent, probably laminar, flow and the well known phenomenon ofnonmusical murmurs. The pure tones and the bruits never occurred simultaneously, and in the transition between the two modes no silent gap was observed. This suggests that the two types of murmur were different modes of expression of the same basic mechanism: 1) a certain range of velocities favoring periodic vortex formation, and 2) higher velocities (and Reynolds numbers) producing irregular vortex shedding. The latter state is called "turbulence" by some authors. 4~1 J. Neurosurg. / Volume 60/January,

5 R. Aaslid and H. Nornes The theory that cardiovascular murmurs are caused by turbulence cannot properly account for the production of musical murmurs from the flow of blood. Turbulence, by definition, implies a random motion of the fluid particles, and it is difficult to understand how this should give rise to pure tones. Ferguson 4 observed murmurs of musical quality in aneurysms of the cerebral arteries during surgery. Unfortunately, no spectral analysis of these sounds was published. However, his use of the turbulence theory to explain this phenomenon leads us to doubt that his murmurs had the pure tone quality seen in our material. As early as 1954, McKusick, et al., 9 published a spectral analysis of a musical murmur recorded over the pregnant uterus. The pitch or frequency of the tone varied throughout the pulse cycle and suggested that the frequency was proportional to the velocity of the blood flow. The arteries of the pregnant uterus probably exhibit increased flow velocities due to the increased flow demand. The increased velocities in the cerebral arteries following SAH are caused by a narrowing of the arterial lumen due to vasospasm. 1 Thus, the production of musical murmurs in peripheral arteries seems to be associated with flow velocities above the normal range. Criticism of the Aeolian theory of cardiovascular murmurs has been advanced by supporters of the turbulence theory, ~1 because the complex geometrical configuration of the arteries differs from that of hydrodynamic experimental set-ups (flow past a cylinder, or jets directed at edges). However, Stehbens ~2 demonstrated in experimental models that periodic vortices occurred distal to bifurcations, although only within a limited range of flow velocities. In our series of patients, the maximum amplitude of the musical murmur was located near the bifurcation of the ICA into the MCA and the ACA in the majority of cases. Elevated velocities must be necessary to produce musical murmurs, since no such phenomenon has been observed in a series without cerebrovascular disease. 2 However, elevated velocity does not automatically give rise to musical murmurs. We observed no significant difference in the range of velocities in the 17 patients without musical murmurs compared to those listed in Table 1. It seems probable that the geometrical configuration of the arteries and possibly pathological changes in the vessel wall play some role in the production of the musical murmurs. This problem is, however, very complex and difficult to study because of the variety of the vascular geometry. The clinical significance of the pure tones described is related to the development and decay of arterial spasms. The frequency of the tones correlates with the degree of spasm, with high frequencies signifying pathologically increased velocities and a narrow arterial lumen. J The musical murmurs as well as the bruits imply an added mechanical vibrational stress on the arterial wall. The sound detection method used in the present study is highly sensitive, and it remains uncertain whether the vibrations observed are of sufficient magnitude to have a mechanical influence on vascular tissue. Some authors have speculated that flow-induced vibrations in conjunction with resonance of the arterial wall represent an additional mechanical stress. 5 6 Further experimental studies in this field would be of clinical interest. References 1. Aaslid R, Huber P, Nornes H: Evaluation of cerebrovascular spasm with transcranial Doppler ultrasound. J Neurosurg 60:37-41, Aaslid R, Markwalder TM, Nornes H: Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries. J Neurosurg 57: , Bruns DL: A general theory of the causes of murmurs in the cardiovascular system. Am J Med 27: , Ferguson GG: Turbulence in human intracranial saccular aneurysms. J Neurosurg 33: , Foreman JEK, Hutchison K J: Arterial wall vibration distal to stenoses in isolated arteries of dog and man. Circ Res 26: , Hussain AKMF: Mechanics of pulsatile flow of relevance to the cardiovascular system, in Hwang NHC, Normann NA (eds): Cardiovascular Flow, Dynamics and Measurements. Baltimore: University Park Press, 1975, pp Kosugi Y, Goto T, Ikebe J, et al: Sonic detection of intracranial aneurysm and AVM. Stroke 14:37-42, McKusick VA, Murray GE, Peeler RG, et al: Musical murmurs. Bull Johns Hopkins Hosp 97: , McKusick VA, Webb GN, Brayshaw JR, et al: Spectral phonocardiography: clinical studies. Bull Johns Hopkins Hosp 95:90-110, linger CP, Wasserman JF: Electronic stethoscope for detection of cerebral aneurysm, vasospasm and arterial disease. Surg Neuro 8: , Roach MR: Poststenotic dilatation in arteries, in Bergel DH (ed): Cardiovascular Fluid Dynamics, Vol 2. London: Academic Press, Stehbens WE: Flow in glass models of arterial bifurcations and berry aneurysms at low Reynolds numbers. Q J Exp Physioi 60: , Tritton DJ: Experiments on the flow past a circular cylinder at low Reynolds numbers. J Fluid Mech 6: , 1959 Manuscript received June 8, Address reprint requests to: Rune Aaslid, Ph.D., Department of Neurosurgery, National Hospital, Rikshospitalet, Oslo, Norway. 36 J. Neurosurg. / Volume 60~January, 1984