Can Transcranial Color Doppler Spectral Signatures be a Novel Biomarker for Monitoring Cerebrovascular Autoregulation and Intracranial Pressure? A Speculative Synthesis.

BACKGROUND: Trans Cranial Colour Doppler (TCCD) has been extensively used in various neurological and neurosurgical conditions causing severe raise in the intracranial pressure (ICP). MATERIAL AND
METHOD: Our study explores the sequential evolution of TCCD flow pattern by correlating with pupillary reactivity, Glasgow coma scale (GCS), and imaging. Our cohort consisted of thirty patients with ten patients in each subgroup admitted to the neuro-Intensive Care Unit (NICU) for various neurological and neurosurgical causes. Middle cerebral artery was insonated through the transtemporal window at the time of admission to NICU. Doppler waveform and parameters such as peak systolic velocity, end-diastolic velocity, systolic by diastolic ratio, pulsatility index, and resistivity index were recorded. The clinical variables for evaluating the degree of raised ICP were the GCS and pupil size. Other systemic parameters such as mean arterial pressure, heart rate, and respiratory rate were also considered and these results were further correlated with TCCD findings. The groups were divided into three groups based on GCS, pupillary reactivity, and imaging. Imaging was done to indicate the etiology for ICP changes and also to look for signs of raised ICP.
RESULTS: Ten distinct types of waveform patterns were noted, and these waveforms correlated with various physiological parameters suggestive of raised ICP.
CONCLUSION: The sequential evolution of distinct patterns of Doppler waveform with increasing degree of raise in ICP has been described and can act as a quick screening tool in NICU and helps stratify patients for treatment and prognostication.

INTRODUCTION

The initial application of ultrasound in neurology was using A-mode sonography to determine midline shift (MLS) in suspected intracranial mass lesions. However, with the introduction of computed tomography (CT) and magnetic resonance tomography, it became a redundant technique. The use of transcranial Doppler (TCD) sonography by Aaslid et al.[1] gave way to newer applications in neurosciences.

Trans Cranial Colour Doppler (TCCD) is a technical development that combines noninvasive imaging of intracranial vessels and parenchymal structures at a high spatial resolution.[2345] It can rapidly and noninvasively image blood flow in the major basal intracranial arteries. Hence, the intracranial vascular structures can be displayed in the correct anatomical relationships which make angle corrected velocity measurements possible. Flow velocities and direction can be color coded based on the Doppler shift, resulting from moving erythrocytes (frequency-based TCCD).[6789] The main advantage of TCCD is the low cost of imaging, ease of repeatability, and excellent safety and tolerability.[10] The main drawback is its limited field of view and also is not technically feasible in 10% of cases.

TCCD provides useful information on cerebral circulation. Lindegaard ratio considers the peak systolic velocity (PSV) of middle cerebral artery (MCA) versus internal carotid artery (ICA) and gives a rough estimate between intracranial and extracranial cerebral circulation and degree of vasospasm and compensatory hyperemia occurring.[11] A systematic study using TCCD to noninvasively and objectively substantiate and prognosticate the evolution and progression of rise in intracranial pressure (ICP) and detect it early on is lacking in the present literature. There are several papers which use indices such as PSV, pulsatility index (PI), and resistivity index (RI)[12131415161718] as a marker for severely raised ICP. However, clinical studies with the correlation of the spectral waveform and also Doppler parameters such as end-diastolic velocity (EDV) and systolic by diastolic ratio (S/D) with clinical parameters are currently unavailable. A study on spectral pattern has been published using rat model with sequential raise of ICP, but human studies are lacking.[19] Another study by Chen et al.[20] has used pupil reactivity as a reliable biomarker for predicting the evolution of raised ICP. They divided their cohort into three groups based on pupil reactivity and correlated it with ICP values. Group A had normal pupils and it correlated to ICP of 19 mmHg, Group B had abnormal pupillary reactivity and it correlated with ICP of 30 mmHg, and Group C had nonreactive pupils and it correlated with ICP of 33 mmHg.

When ICP increases from normal, compensated, to decompensated phase, associated clinical parameters such as Glasgow coma scale (GCS) and pupils along with mean arterial pressure (MAP) and heart rate (HR) also alter in these three phases in accordance with Monro–Kellie model. GCS is a reliable marker for the grading level of consciousness, and acutely raised ICP can cause rapid deterioration of level of consciousness.[21] In view of the nonreliability of clinical parameters as indicators of intracranial hypertension, in focal and diffuse models, CT imaging is also considered for assessing the degree of acutely raised ICP. A study by Asil et al.[22] using TCD noted that unilateral lesions such as malignant MCA infarct with a PI of >1.5 and an MLS of >9 mm correlated with raised ICP. Although ICP monitoring is the accurate method of ICP calculation, various studies[23] have concluded that the outcome of ICP monitored group was similar to cases based on only clinical and imaging and was in no way superior. Invasive ICP monitoring was not a routine protocol in our institute in the neuroICU.

In this study, we propose a method of building an ICP model and the physiological basis for interpreting TCCD spectral pattern by correlating with clinical and imaging parameters. For this purpose, we have divided the patients into three subgroups based on clinical parameters and imaging findings which suggest acute raise in ICP. The basic aim of our study was to identify unique spectral patterns on Doppler in different phases of raised ICP.

MATERIALS AND METHODS

This was a prospective study and patients admitted to neuro-Intensive Care Unit (NICU) with brain dysfunction for various etiologies were examined. TCCD was performed at the point of admission to Intensive Care Unit (ICU) and physiological parameters at the point of admission were considered for correlation with TCCD. The study was approved by the Ethics Committee, and informed consent was obtained from patient's relatives to perform the study. The data collection and interpretation were by licensed neuroradiologist and interventionist (SM) with 10 years of experience in sonology and dedicated 7-year experience in neuroradiology. The stratification of patients into different groups was based on GCS and pupillary status and imaging. Other systemic parameters such as HR, MAP, and respiratory rate were also considered. All patients were on medical management to maintain the HR and blood pressure in the physiological range. Ten patients in each group were considered with a total of thirty patients in the cohort. Group A comprised patients who had GCS between 8 and 14 and normal pupil reactivity. Imaging had no features of raised ICP, Group B comprised patients in whom GCS was <8 had abnormal pupillary reactivity with sluggishly reactive pupils. Group C comprised patients with GCS <8 and deeply comatose with pupils dilated and fixed at the time of admission. Groups B and C patients had imaging features of raised ICP.

Imaging criteria on nonenhanced CT suggestive of increased ICP were decreased ventricle size, decreased basilar cistern size, effacement of sulci, transfalcine herniation (MLS), transtentorial herniation, and loss of gray-white matter differentiation.

Inclusion criteria

The patient cohort consisted of both neurological and neurosurgical causes, brain dysfunction, and admission into NICU for the same [Supplementary Table 1]. Written consent was obtained before TCCD. Acute neurological conditions suggestive of causing sudden changes in ICP were considered.

Exclusion criteria

Patients with technical problems wherein TCCD could not be performed optimally such as thick skull leading to a poor window for insonation of the vessel were excluded. Postoperative cases were also excluded from the study. Patients with a long-standing history suggestive of chronic intracranial hypertension were also excluded.

Trans Cranial Colour Doppler (TCCD) was performed with an ultrasound machine (LOGIQ e, GE Healthcare, Little Chalfont, Buckinghamshire, UK) with a low-frequency phased array (1.75 ± 3.0 MHz) probe. Although MCA, posterior cerebral artery (PCA), and anterior cerebral artery (ACA) bilaterally were insonated through the temporal window, MCA vessel was preferred to be the region of interest in our study as it is very accessible through the temporal window and shows less anatomical variability. Symmetric changes, when noted in both the MCA, reflect global changes in brain perfusion. MCA vessel, unlike ACA and PCA, is less affected by herniation or MLS. After identifying the vessel of interest (MCA) through the temporal window, we measured Doppler parameters such as PSV, EDV, and S/D. Dimensionless variables such as PI and RI are not dependent on the insonation angle and hence were also calculated. PI and RI were calculated as and , respectively.[2425] These parameters were considered to indicate the peripheral vascular resistance. The spectral signature was also noted. Although bilateral MCA was insonated, the waveform and parameters considered were from MCA showing most deviation from the normal range of PSV and EDV.

The normal flow pattern is pulsatile forward flow with EDV being half of PSV. The spectral waveform will be a low-resistance waveform with a continuous diastolic flow. Complete filling of the spectral window is noted in these vessels because of the small diameters of these intracranial vessels. A large quantity of forward flow continues throughout diastole phase in intracranial vessels as opposed to extracranial vessels.[26] The normal PSV, EDV, PI, RI, and S/D of MCA in normal healthy individuals are in the range of 55–75 cm/s, 30–40 cm/s, 0.6–1, 0.3–0.5, and 1.7–2.1, respectively. The normal spectral signature in MCA vessel is shown in Figure 1.

Figure 1

Left middle cerebral artery transcranial Doppler waveform (bottom) with color Doppler in a normal healthy control through transtemporal window. Low resistance type of waveform with continuous diastolic flow typical of cerebral circulation noted

Clinical correlation was performed in all patients with respect to the etiology, response to surgical decompression, and prognosis at the time of discharge from ICU. Surgical decompression was indicated if deterioration was noted clinically with evidence of herniation on CT.

The prognosis was considered “good” if recovery from altered consciousness was noted with stabilization of systemic parameters and the patient was discharged from ICU. Prognosis was considered “bad” if deterioration of patient's condition was noted even after aggressive management and life support measures.

RESULTS

Thirty cases of brain dysfunction were examined and based on clinical parameters such as GCS, pupil reaction, HR, and MAP were subdivided into three groups [Supplementary Table 1]. Sequential changes in PSV, EDV, S/D, RI, and PI at every stage of the Doppler flow pattern are shown in Supplementary Table 2 and Table 1a and b. Ten different types of waveforms were recognized and these waveforms have been correlated clinically.

Table 1a

The Doppler flow patterns have been described and range of different Doppler parameters such as peak systolic velocity, end-diastolic velocity, and Doppler ratios such as systolic by diastolic, pulsatility index, and resistivity index in each of this pattern has been given

Table 1b

The Doppler flow patterns have been described as compared to normal peak systolic velocity, end-diastolic velocity, and Doppler ratios such as systolic by diastolic, pulsatility index, and resistivity index in each of this pattern and behavior of middle cerebral artery vessel has been described

Group A

Group A consisted of cohort of patients placed in ICU for altered sensorium who had GCS between 8 and 15 and with normal pupil reaction and had spontaneous breathing and with an HR either normal or increased. MAP was either normal or increased (65–120 mmHg). Imaging had no features to suggest raised ICP.

On TCCD, two types of waveforms were noted as given below.

Pattern I (Blunted-B)

The waveform showed a delayed systolic acceleration as compared to normal. The PSV was decreased to as low as 30–37 cm/s and EDV was decreased to as low as 16 cm/s. The Doppler ratios were maintained in the normal range as both PSV and EDV were decreased proportionately (S/D = 1.8–2.3, PI = 0.63–0.78, RI = 0.47–0.55) [Figure 2a].

Figure 2

TCCD demonstrates a decreased end-diastolic velocity when intracranial pressure is mildly raised. Two waveforms are noted. (a) Both peak systolic velocity and end-diastolic velocity is in normal range - Pattern I – Blunted. (b) Peak systolic velocity is normal and end-diastolic velocity is reduced - Pattern II – Dampened. This waveform is consistent with low resistance flow as a part of cerebral autoregulation to maintain cerebral perfusion

Pattern II (Dampened-D)

The waveform showed a low resistance flow. PSV was in normal range (53–88 cm/s), but EDV reduced in the range of 18–27 cm/s. The Doppler ratios were increased as compared to normal as EDV which is the denominator was reduced but not the PSV (S/D = 2.6–4, PI = 1–1.68, RI = 0.61–0.75) [Figure 2b].

Overall, in Group A, EDV was less than half of PSV and the velocity was decreased up to 16 cm/s for EDV and up to 30 cm/s for PSV. The maximum increase in Doppler parameters was observed (S/D = 4, PI = 1.7, RI = 0.75).

Both waveforms were indicative of low resistance forward diastolic flow. At physiological level, this pattern is indicative of setting in of cerebral autoregulatory vasodilatory response to maintain the cerebral blood flow (CBF).

Decompressive surgery was not done in any of these cases. Only elective surgery for space-occupying lesions such as subdural hematoma (SDH) and tumor was done. Prognosis was good in all cases and they were discharged from NICU.

Group B

Group B consisted of patients placed in ICU for altered sensorium who had a GCS <8 and were placed on the ventilator. Pupils were sluggishly reactive. Tachycardia with or without intermittent bradycardia was noted in this group. MAP in this group was increased up to 120 mmHg. Imaging had features of raised ICP.

On TCCD, two waveforms were noted. Since aliasing was noted at regular Doppler setting of 0–50 cm/s in this group, baseline velocity range was increased to 0–150 cm/s by increasing the pulse repetition frequency.

Pattern I (prominent systolic peak without Doppler window)

The PSV was increased with aliasing noted (95–130 cm/s), but EDV velocity was maintained in the normal range (25–40 cm/s). The Doppler parameters were increased with upper limits (S/D = 4, PI = 1.6, and RI = 0.75) as the PSV which is the numerator was increased but EDV was in normal range [Figure 3a].

Figure 3

TCCD demonstrates an increased peak systolic velocity when intracranial pressure is moderately raised. Two waveforms are noted. (a) Peak systolic velocity is increased and end-diastolic velocity is in normal range - Pattern I - prominent systolic peak without Doppler window. (b) Peak systolic velocity and end-diastolic velocity is increased - Pattern II - prominent systolic peak with Doppler window. This waveform is consistent with setting in vasoconstriction as maximum vasodilatory capacity is reached

Intracerebral vessels including MCA show complete filling of the spectral window because of the small diameters of these vessels and this was indicative of a normal laminar flow pattern.

Pattern II (prominent systolic peak with Doppler window)

Prominent PSV with aliasing and high-velocity EDV with Doppler window in both the phases was noted. Both PSV and EDV were increased in the range of 122–189 and 48–74 cm/s, respectively. The Doppler parameters such as S/D, PI, and RI were normal as the increase in PSV was accompanied by a proportionate increase in EDV [Figure 3b].

The Doppler window is the clear black zone between the spectral line and the baseline and is accompanied by a narrow spectral line. This type of waveform is normally seen in medium-sized vessels such as ICA but not in MCA and indicates a plug flow due to high velocity as seen in this waveform.

Overall, in Group B, the PSV was increased up to 189 cm/s and EDV was increased up to 74 cm/s. The maximum increase in Doppler parameters was noted (S/D = 4, PI = 1.6, RI = 0.75). MAP was increased in this group.

The EDV was one-third of the PSV in both patterns. The EDV was either normal or increased; physiologically, these flow patterns were indicative of a failure of vasodilatation to maintain CBF and hence were accompanied by systemic hemodynamic compensation with an increase in MAP and HR clinically. This pattern is also indicative of an extreme end of maintaining cerebral autoregulation.

Surgical decompression was done in three cases when CT had features of herniation. TCCD in two cases showed Pattern I and one case showed Pattern II waveform. Surgical outcome was good in both these waveforms. Overall, in Group B, prognosis was good irrespective of medical management or surgical decompression.

Group C

Patients in Group C were in the ICU with GCS <8, on a ventilator with pupils fixed and dilated and with bradycardia. MAP was either normal or decreased (40-50 mmHg). Imaging had features of raised ICP.

TCCD showed six different patterns of waveforms.

Pattern I (sharp wave with or without flow reversal)

There was prominent a systolic spike (SS) with very low diastolic of <12 cm/s accompanied by intermittent reverse flow and indicated that some forward flow is still present and diastolic arterial pressure is still greater than ICP. MAP was maintained [Figure 4a].

Figure 4

(a-f) TCCD demonstrates sequential changes in peak systolic velocity and end-diastolic velocity when intracranial pressure is severely raised six types of waveforms are noted. All these waveform are consistent with loss of cerebral autoregulation in maintaining cerebral perfusion. (a) Peak systolic velocity is increased and end-diastolic velocity reduced to <12 cm/s (topmost row) with intermittent flow reversal (middle row) - Pattern I - sharp wave with or without flow reversal. (b) Peak systolic velocity is in normal range with flow reversal in diastolic phase - Pattern II - systolic spike with flow reversal (bottom row). (c) Peak systolic velocity is reduced and diastolic wave shows reverse flow and continuity between waveforms is lost (decoupling) - Pattern III - systolic and diastolic spike. (d) Prolonged systolic acceleration with diminished amplitude with total absence of diastolic flow Pattern IV - tardus parvus waveform. (e) Peak systolic velocity is decreased with discontinuity of waveform Pattern V - systolic spike with absent diastolic. (f) Total absence of systolic and diastolic flow Pattern VI - no flow waveform

Pattern II (systolic spike with flow reversal)

PSV is in normal range or decreased and flow reversal (FR) is noted in diastolic phase and this oscillating pattern occurs when ICP becomes higher than diastolic arterial pressure. MAP was maintained [Figure 4b].

Pattern III (systolic and diastolic spike)

PSV was reduced and the diastolic wave showed a reverse flow and continuity between waveforms was lost. MAP was below normal [Figure 4c].

Pattern IV (tardus parvus waveform)

Tardus refers to prolonged or delayed early systolic acceleration. Parvus refers to diminished amplitude and rounding of the systolic waveform with a total absence of diastolic flow. MAP was below normal [Figure 4d].

Pattern V (systolic spike)

PSV is decreased with decoupling of the waveform. A total absence of diastolic waveform is noted. At this stage, MAP was decreased to <40 mmHg even after ionotropic support [Figure 4e].

Pattern VI (no flow waveform)

On color Doppler, no flow (NF) was detected in the MCA location and power Doppler also could not pick up signals at the lowest settings [Figure 4f].

Decompressive surgery was performed in two patients showing Pattern II and Pattern III waveform. However, prognosis was bad in all cases including the surgical decompression group and all were declared brainstem dead. It was also observed that the survival period of the Pattern I and Pattern II after aggressive management was up to 6–9 days and abruptly dropped to 1 day in Patterns III, IV, V, and VI.

Between-group comparison

On comparing Group A and Group B, there was a difference in waveform and also the PSV and EDV velocities. Group A had decreased EDV and Group B had increased PSV and the same was reflected in the waveform. The Doppler ratios were either normal or increased in both these groups and hence did not have a predictive value to differentiate between these two groups. In short, PSV was more than twice the EDV in Group A and PSV was three times more than EDV in Group B. The Doppler ratios S/D, PI, RI were twice the normal values in both these groups.

Irrespective of type of waveform, the prognosis was good in both the groups after either medical or surgical management and hence is a marker for reversibility of brain dysfunction.

On comparing Groups A and B with Group C, a waveform with EDV <12 cm/s or showing FR and with S/D of 12 or 0, PI of 0 or 2.4, and RI of 0 or 0.98 showing erroneous extreme values was a marker of irreversibility and autoregulation failure. Irrespective of the type of waveform, all indicated a cascading effect of both cerebral and systemic autoregulatory failure and both medical and surgical management failed in this group.

When the waveforms from the bilateral MCA were compared in all three groups, the waveforms fitted into similar groups, irrespective of the type of etiology and side of the injury. This similarity between hemispheres may be due to a good collateral circulation in the brain and its role in cerebral autoregulation.

Prognosis-wise Group A and Group B showed a good recovery and were discharged from NICU, whereas the Group C had a bad prognosis and all cases did not recover even after aggressive medical and surgical management. Group B responded to surgical management and the cascade of cerebral ICP autoregulatory failure was halted, whereas Group C wherein the cascade effect of autoregulation failure had started showed poor response to both medical and surgical treatment.

On within-group pattern analysis in Group C, the survival time showed a sudden decrease from 1 week to 1 day after Pattern III onward. The outcome in Group A and Group B did not vary with the pattern of the waveform.

On imaging, Group A had no MLS. Group B and Group C had features of transtentorial herniation, with effaced basal cisterns and diffuse cerebral edema.

DISCUSSION

TCCD has been extensively used in neurosciences. Studies have correlated the importance of PSV, PI, and RI as marker for severely raised ICP and in brain dead patients.[1213141516] Physiologically, under conditions of progressive intracranial hypertension, disturbances of CBF begin at the microcirculatory level and then extend to the larger vessels.[27] Precapillary arteriolar vasodilation occurs. However, no systematic study has been done to study the sequential raise in ICP using Doppler and with clinical correlation.

In our study, we have tried to explain the sequential changes in the hemodynamic waveform with increasing degree of brain dysfunction and raise in ICP using TCCD. The cerebral circulation is different from the hemodynamic model of other organs because of the unique collateral circulation in the brain (circle of Willis). Furthermore, unlike most other organs which are end organs and the changes in Doppler have a linear relation, cerebral system also exhibits another unique phenomenon based on Monro–Kellie hypothesis wherein the volume of the brain is fixed and hence ICP, cerebrospinal fluid, and CBF are in equilibrium. There is a dynamic coupling occurring in these three compartments.[28] Hence, irrespective of etiology in any of these compartments, the ICP can increase. Furthermore, at the end point, brain function and survival are not dependent on ICP but on CBF to maintain the cerebral metabolic rate of oxygen demands.[28] Many studies existed in the literature try to correlate these parameters. However, these studies are both cumbersome and are not cost-effective.

In pathological conditions when intracranial compliance is reduced as a result of an intracranial mass lesion or brain edema, autoregulation sets in the initial phase and vasodilation occurs resulting in an increase in cerebral blood volume and that may further raise the ICP. Further increased ICP leads to decrease MAP and cerebral perfusion pressure (CPP). Subsequently, failure of autoregulation sets in at the extremes of CPP and passive vessel collapse occurs.

The degree of raise in ICP in the ICU is assessed based on the combined findings of clinical parameters such GCS, pupils reactivity along with HR, and MAP values. Although GCS is a reliable marker for the grading level of consciousness, an acutely raised ICP can cause rapid deterioration of level of consciousness[21] and a preserved GCS does not rule out a raised ICP. Hence, GCS cannot be used in isolation to predict the degree of raise in ICP. A recent study by Chen et al.[20] has used pupil reactivity as a marker for raised ICP and correlated with ICP values and concluded that it can be used as an indirect marker for the degree of raised ICP. As the ICP increases, HR is maintained in the initial phase and changes to tachycardia pattern as an autoregulatory response, and when autoregulation fails in, severely raised ICP bradycardia sets in. MAP is also initially maintained within a normal range in the initial phase, and as an autoregulatory response, MAP will increase; however, with the failure of autoregulatory response, MAP will decrease. A combination of these clinical parameters is used in the NICU setup as indirect markers to measure the degree of raise in ICP and brain dysfunction.

In our study, the evolution of different Doppler waveforms was studied with various degrees of brain dysfunction as a marker for severity of raised ICP. Our study was divided into subgroups in a manner very similar to Chen et al. study;[20] however, since invasive pressure recording for raised ICP was not a part of our protocol, we used both pupil reaction and GCS status as clinical markers for brain dysfunction associated with raised ICP along with imaging findings. Changes in HR and MAP are also noted with raised ICP and hence were also considered to confirm the degree of raised ICP clinically.

Our study was designed to explore whether TCCD parameters and spectral signatures correlated with the degree of a raise in ICP and whether TCCD spectral signatures could be a biomarker for the different degree of increases in the ICP.

Physiologically, autoregulation is triggered by changes in CPP, which in turn may be the result of changes in ICP with or without changes in MAP. Cerebral arteriolar vasodilatation or vasoconstriction occurs as a result of this response.

When CPP drops as a result of ICP elevation, vasodilatory response ensues as long as the CPP does not fall below the lower limit of autoregulation (50–70 mmHg) and autoregulation is preserved. At CPP values close to the lower limit of autoregulation, vasodilatation and hence cerebral blood volume are maximal, potentially leading to ICP elevation.

When the CPP drops below the lower limit of autoregulation under conditions of progressive intracranial hypertension, there is a passive collapse of vessels, leading to cerebral hypoperfusion. These changes at microcirculatory level due to CPP and ICP changes are reflected in the Doppler pattern of larger vessels such as MCA.

Many systemic compensatory mechanisms take place such as an increase in HR and MAP to maintain perfusion as the ICP increases. As already described, HR and MAP increase in the initial phase followed by a decrease in both HR and MAP as autoregulation fails.

Doppler parameters and spectral pattern can be used to objectively reflect the degree vasoconstriction or vasodilation. Under conditions of progressive intracranial hypertension, disturbances of CBF begin at the microcirculatory level and then extend to the larger vessels.[27] Hence, with progressive increase in ICP, if the vessel of interest as in our study is the MCA, the initial pattern noted will reflect the changes occurring at microcirculatory levels distal to MCA, that is, MCA spectral pattern behavior is like a prestenotic segment. However, with the further increase in ICP, the MCA also undergoes vasoconstriction, and hence, spectral pattern will be like a stenotic segment and with further deterioration acts like a poststenotic segment.

Doppler studies in intracranial arterial stenosis have shown that flow velocity depends on the diameter, length of involvement of the vessel of interest, and also the degree of distal stenoses. In a study by Sharma et al.[29] on correlating PI with flow velocity, in diffuse intracranial disease, it has been shown that high PI with increased flow velocity indicates stenotic flow. A high PI and a gradual increase in velocity parameters have been noted in Groups A and B in our study and reflect the gradual increase in degree of vasoconstriction. Cardiac status can also be assessed indirectly using Doppler. A high PI and a normal velocity indicate an increased cardiac output as seen in Groups A and B cases and indicate a systemic hemodynamic compensation. Similarly, a low PI with low velocity indicates reduced cardiac output as noted in Group C and reflects a loss of systemic compensation.

Group A patients had a normal pupillary reaction and based on other clinical and imaging parameters were termed as patients with mildly raised ICP. These findings, when extrapolated with Chen et al.[20] findings, can be used to indicate a maximum ICP pressure of 19 mmHg. On correlating with the waveform noted in this group in our study, it may reflect Stage I of cerebral autoregulation which consists of initial vasodilatation to maintain the CBF. The MAP also correlated and was either normal or increased in this group.

Two types of waveforms were noted in Group A which we termed as Pattern I and Pattern II type of flow. The Doppler waveform was suggestive of low resistance type of flow with reduction in both PSV and EDV (Pattern I - Blunted) due to vasodilatation, and as ICP increased, the PSV also increased (Pattern II - Dampened) and may reflect transition from zone of maximum vasodilatation to a zone of vasoconstriction as an autoregulatory phenomenon.

Group B patients had sluggishly reactive pupils and clinically required aggressive medical management and were on a ventilator. Some of them also required surgical decompression if CT showed features of herniation. These findings, when extrapolated with Chen et al.[20] findings, can be used to indicate a maximum ICP pressure of maximum of 30 mmHg. The range of waveforms noted in this group may reflect maximum limit of autoregulatory capacity in response to raised ICP and the CBF is decreased. The MAP was increased in this group.

The Doppler waveform in this group was indicative of increase in distal resistance wherein there was increase in PSV (Pattern I - prominent systolic peak [PSP] without Doppler window [DW]) initially followed by increase in both PSV and EDV (Pattern II - PSP-DW). This pattern marks end of vasodilatory capacity and setting in of vasoconstriction and the beginning of autoregulation failure. Systemic hemodynamic compensation occurs at this point to maintain the perfusion as reflected by the increase in MAP and HR in this group of patients.[28] This pattern may also be used to indicate an upper limit of a salvageable situation by treatment.

Group C patients had fixed dilated pupils and based on clinical and imaging parameters represented patients with severely raised ICP. From this stage onward, the cascade of cerebral autoregulatory failure occurs wherein severe increases in ICP lead to reduced flow velocity in the diastolic phase more rapidly than in the systolic phase implying a total loss of intracranial compliance. Furthermore, the systemic hemodynamic compensation which plays a major role in maintaining cerebral perfusion also fails from this stage onward as noted by a decrease in HR and MAP. The patients did not have a good prognosis even with aggressive management. According to the study by Chen et al.,[20] mid-dilated pupils reaction represents an ICP >33 mmHg.

The waveform noted in this group reflects autoregulatory failure. In the initial phase of this failure, there were increase in PSV and decrease in diastolic velocity, which indicates increased peripheral vasoconstriction at microcirculatory/capillary level. The presence of EDV in this waveform in conditions of severe raised ICP indicates that diastolic arterial pressure is still trying to maintain above ICP, but there is a decrease in brain compliance. This pattern may be accompanied by intermittent FR in diastolic phase to maintain perfusion (Pattern I - sharp wave with or without FR). Bradycardia was noted in all patients in this group (as a part of Cushing reflex). Bradycardia indicates a decompensatory phenomenon though MAP was still maintained in the normal range using inotropes.

The waveform as ICP goes on progressing further shows an oscillating pattern with PSV decreasing as vasoconstriction starts extending from microcirculation to the more proximal vessels. Reversal of flow in diastolic phase marks the beginning of ICP rise being greater than diastolic pressure (Pattern II - SS with FR).

Beyond this waveform pattern, four more patterns were observed; however, at physiology level, the hemodynamic model was unlike cerebral circulation and behaved like a high peripheral resistance system. In a hemodynamic model with high peripheral resistance such as the lower limb, the presence of a systolic and diastolic waveform indicates the presence of an antegrade and retrograde flow, respectively, and absence of a severe flow limiting lesion in the proximal larger vessels.[30] However, when the vasoconstriction extends proximally into larger vessels, the Doppler pattern changes. A decoupling of the diastolic waveform may indicate a time lag in retrograde flow during diastole due to high resistance at the proximal segment.

In Pattern III (SS and diastolic spike [DS]), a decrease in PSV along with diastolic FR and diastolic waveform decoupling was noted. Since there was a failure of systemic autoregulation, along with bradycardia, the MAP was also reduced from this waveform onward even on treatment (Pattern III - SS and DS) with decoupling of waveform. This decoupling was responsible to give a DS waveform. A decrease in PSV may reflect decreased hemodynamic compensation and FR indicates ICP > diastolic pressure. A decoupled diastolic flow as in Pattern III may indicate a delay in passive retrograde flow due to high resistance at the proximal segment.

A tardus parvus (TP) waveform was noted in Pattern V in which the PSV was prolonged and of low amplitude waveform type in systolic phase with total absence of diastolic flow (Pattern IV - TP). A TP waveform indicates a poststentotic segment.[31] In our case, this type of waveform in the MCA segment indicates that MCA acts as a poststenotic segment and indicates a migration of vasospasm proximal to MCA that is to involve the ICA part also. The next waveform noted was a systolic waveform with decreased amplitude and showed a decoupling between waveforms (Pattern V - SS with absent diastolic).

The low amplitude waveform may indicate an ICP which is even greater than the systolic pressure in the vessels. A decoupled systolic flow may indicate severe stenoses in ICA leading to antegrade flow into MCA after a time lag. This again indicates further raise in ICP causing vasospasm to extend up to ICA segment.

NF pattern was the last pattern wherein a flat line was seen, and NF can be detected at the least in the given Doppler and window setting in both systolic and diastolic phase and indicates a total lack of perfusion to the brain.[28] This waveform indicates a severe degree of vasospasm wherein the raised ICP prevents perfusion to the brain completely.

In our study, progressive changes in the TCCD waveform with increasing brain dysfunction and raise in ICP were studied. These changes in the flow pattern paralleled the physiological hemodynamic changes in intracranial tension that ultimately lead to an arrest. The different types of waveforms have been described physiologically and this may help in the stratification of the patient in the NICU with respect to prognosis and treatment. Objective evidence for stratification of patients using TCD has been highlighted in our paper for the first time. Our study has highlighted the role of PSV and EDV and waveform to differentiate between Groups A and B patients which indicate mild to moderate increases in ICP, respectively. The Doppler ratios such as PI and RI are almost similar between these two groups and have no predictive value to differentiate these two groups. We have highlighted the correlation between the characteristic flow patterns on TCCD and its relation to ICP, raising the possibility of monitoring increased ICP noninvasively using TCCD.

On correlating the findings clinically with respect to prognosis and surgical management, it was noted that when the Doppler waveform as noted in Groups A and B indicating a maximum ICP of 19 mm and 30 mm of Hg; there was still chance of recovery of brain perfusion and surgery or medical management can help contain the cascade of events as cerebral autoregulatory failure has not yet set in.

After this stage, when diastolic velocity is very low <12 cm/s, it marked the beginning of a severe form of raised ICP in Group C wherein perfusion to the brain could not be maintained even after surgical decompression and aggressive medical management. This may correlate with ICP of 33 mmHg wherein the critical threshold of autoregulatory failure has reached and clinically there was a total downhill course with recovery failing. An interesting finding that was also noted in our study was that even in the severely raised ICP (Group C), there was an abrupt change in survival pattern in. Patients showing Patterns III–VI had a survival period in the ICU of 24 h whereas Patterns I and II showed a survival of 1 week. In Patterns I and II, the waveform indicated that vasospasm was at the capillary level. Beyond this stage, the vasospasm extended proximally to involve the MCA and then up to the ICA in the final stage. This Doppler pattern is accompanied by an abrupt change in survival pattern and may help in the stratifying patients for brain organ transplantation. Most studies focusing on brain dead do not correlate the length of survival of these brain dead patients. This too has been highlighted in our study for the first time. The ICP and cerebral autoregulation model which has been built based on TCD spectral parameters with clinical correlation have been well depicted in Figure 5 and summary of TCD values in Table 1a and b.

Figure 5

The intracranial pressure and cerebral autoregulation model built based on transcranial Doppler spectral waveform and clinical correlation

When the MCA spectral pattern showed a prestenotic behavior, the situation was salvageable and showed a good response to treatment and cerebral autoregulation was still preserved. When the MCA vessel had a stenotic segment behavior based on spectral pattern, the prognosis and treatment/surgical response could be predicted. Further, when the MCA had a poststenotic segment behavior, the patient had a survival rate of less than a week even on treatment indicating a progression of vasospasm.

Although it can be speculated that focal and diffuse ICP changes can change the dynamics of ICP in a different manner, a study by Asil et al.[22] stated that PI >1.5 and MLS of >9 mm were always associated with raised ICP irrespective of focal or diffuse injury. In Group A, there was no sign of MLS or diffuse herniation, so the question of focal/diffuse ICP rise models does not arise. In Group B, all patients had PI >1.5 except Patient 8 and nine patients who had dengue encephalopathy and traumatic SDH. However, on CT, both had diffuse cerebral edema and transtentorial herniation both of which indicates a diffuse increase in ICP. In Group C, all had PI >3 and hence indicates a diffusely raised ICP. In another study by Poca et al.,[32] they quoted that invasive ICP monitoring was normal (<20 mm) in a subset (12 out of 25 cases) of malignant infarction. Hence, this study advised that both radiological and clinical follow-up was advocated to decide on invasive ICP monitoring. Our protocol was to follow clinical and imaging assessment rather than invasive monitoring of ICP. Our TCCD study has added another parameter to monitor such cases in a noninvasive manner.

CONCLUSION

Different Doppler signature can predict, differentiate, and help prognosticate clinical outcomes. Surgery responders and nonresponders can be identified by waveform alone. Differentiating clinically severe raised ICP group into subtypes for the purpose of prognostication is also possible to predict the duration of survival. Understanding of the range of types of waveform pattern that can be noted in the MCA has been highlighted and has an important role especially in TCD, wherein the measurement are made in a blinded fashion and wherein MCA is identified based on depth and direction of waveform.

Although studies have shown that etiology does not have a confounding effect in severely ICP raised patients,[33] further studies using homogenous patient population are indicated to ascertain whether these waveforms can also predict duration of stay in ICU till discharge in Groups A and B and whether rate of recovery can be predicted. In addition, correlation of these waveforms with direct measures such as invasive ICP monitoring and electrocochleography, CPP also needs to be explored. Longitudinal change in the TCCD patterns following surgical intervention and whether TCCD pattern correlates with clinical improvement needs to be explored. Further validation of this study is definitely required. More elaborate work is warranted using direct quantification parameters of cerebral autoregulation such as pressure reactivity index or mean flow index, partial pressure of carbon dioxide levels for a more accurate interpretation of the observed cerebral vasoresponses.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.