Transcranial Doppler as an early predictor of neurological outcome in mild and moderate traumatic brain injury: An observational study : Bali Journal of Anesthesiology

Secondary Logo

Journal Logo

Original Articles

Transcranial Doppler as an early predictor of neurological outcome in mild and moderate traumatic brain injury: An observational study

Abdallah, Mai Kamal Mohamed; Afandy, Mohamad E; Abd El-Hafez, Ahmed Ali; Elhawary, Salama E; El-Gendy, Hala M

Author Information
Bali Journal of Anesthesiology 7(2):p 82-87, April-June 2023. | DOI: 10.4103/bjoa.bjoa_280_22
  • Open



Traumatic brain injury (TBI) is a major cause of disability and mortality globaly. The transcranial Doppler (TCD) method may show low diastolic blood flow velocity (FVd) and high pulsatility index (PI) measurements brought on by high vascular bed resistance. This study aimed to assess the usefulness of the TCD-PI for the early detection of secondary neurological deterioration (SND) in mild and moderate TBI.

Materials and Methods: 

This prospective study was carried out on 105 mild and moderate TBI patients, who had TCD measurements within 12 h of the initial trauma, and initial computerized tomography (CT) showed mild lesion or no detected abnormality.


Primary end point was assessed (SND) within 1st week post trauma. Of the 105 patients with mild and moderate TBI, 29 (27.6%) showed SND. We evaluated the value of our intervention (TCD) to predict SND after mild and moderate TBI in 1st week. PI could predict SND at cutoff 1.21 with good sensitivity of 96.5% and specificity of 94.7%, area under curve (AUC) value of 0.98, negative predictive value (NPV) of 98.6%, and positive predictive value (PPV) of 87.5%. To amplify our finding, we measured FVd, and at 25 cm/s it showed good sensitivity 86.2% and specificity 89.5% when AUC 0.93, NPV 94.4%, and PPV 75.8%.


TCD on admission may provide a valuable tool of early prediction of neurological outcome for mild and moderate TBI patients. Closing the gap in poor prediction of commonly used evaluation by CT especially with mild lesion.


Depending on the severity of the damage, the clinical presentation, and the neurological condition, traumatic brain injury (TBI) may be divided into mild, moderate, and severe categories.[1] Direct mechanical pressures cause the functional and anatomical components of the brain to distort, which eventually results in primary brain injury.[2]

While secondary neurological deterioration (SND) generally develops during the first week after an injury and may be brought on by a number of conditions, including cerebral edema, symptomatic hydrocephalus, seizures, subarachnoid hemorrhage, intracranial or subdural hematomas, or the cerebral effects of extracranial injuries.[3]

Cerebral perfusion pressure and intracranial pressure (ICP) and could be monitored in real time non-invasively using transcranial Doppler (TCD).[4] The middle cerebral artery (MCA) and the other main intracranial vessels’ pulsatility index (PI) values may be used by TCD ultrasonography to assess ICP and cerebral blood flow (CBF) and ICP, independent of the patient’s state of consciousness or whether they are under sedation.[5] Therefore, this study was conducted to evaluate the utility of TCD PI in mild and moderate TBI for the early prediction of SND.

Materials and Methods

After obtaining the research ethics committee approval (approval code: 33574/12/19) and Clinical Trial Registration (NCT05151978, dated December 9, 2021), each patient’s or her/his relative’s informed permission was obtained. This prospective study was performed in Tanta Emergency Hospitals on 105 adult patients of either sex, within 12 h after the trauma, an adult patient (aged 25–60) with mild to moderate traumatic brain injury with (Glasgow Coma Scale [GCS] 9–15).

A total of 179 individuals having mild to moderate TBI were admitted within 12 h post trauma and initial CT were recorded. We excluded patients underwent head surgical procedure before admission, patients with history of previous intracranial lesion, and patients with antiplatelet and/or anticoagulant therapy history. Patients with arterial pulse oximetry readings <90% and/or systolic blood pressure <90 mmHg in addition to mechanically ventilated patients on admission were excluded from this study. We also excluded patients with any craniotemporal lesion impeding satisfactory TCD examination, patients with no acoustic window, and missing TCD value.

The number of studied patients was 105; after exclusion of 74 patients, the rest received a TCD evaluation within of 12 h after the initial trauma. Blinded to the CT categorization, an observer measured the diastolic flow velocity (FVd), systolic blood flow velocity (FVs), and PI. Neurological examination was done to each patient on admission using GCS. Heart rate, mean arterial pressure, PCO2, sodium serum level, hemoglobin (Hb) concentration and Sao2 were recorded.

According to SND day-7 post trauma two groups of patients were formed: patients with no SND (N = 76, Group I) and patients with SND (N = 29, group II), while according to Glasgow outcome scale (GOSE) at 1 month, we divided patients into two groups: poor outcome (GOSE 1–6, N = 20) and good outcome (GOSE 7–8, N = 85).

We used the 2–5 MHz transducer of ultrasound across the temporal region just in front of ear tragus and slightly above the arch of zygoma (transtemporal window). We oriented the transducer slightly upward and anteriorly. The brain stem and sphenoidal bone’s clinoid process were first recognized. The circle of Willis might be identified due to color-coded sonography. The flow in the ipsilateral MCA was represented by a red color signal (towards to the probe) from 40 to 65 mm, whereas the outflow from the ipsilateral anterior cerebral arteries (ACA) was represented by a blue signal from 60 to 80 mm [Figure 1A].

Figure 1::
(A) TCD show right MCA, ACA, and PCA; (B) TCD show left MCA

MCA insonation and tracings were both recorded for at least 10 cardiac cycles in individuals with stable circumstances, that is, no cardiorespiratory distress and no pain or agitation. The built-in software primarily identified the M1 segment of the MCA and then used manual angle correction to estimate velocity of blood flow in each MCA (cm/s) [Figure 1B].

Over a 30 seconds recording period, tracings have to remain constant. Then, the FVd and the FVs, the time-averaged mean blood flow velocity (FVm), pulsitility index PI [(FVs – FVd)/FVm] were computed.

For statistical analysis, the greater PI and lower flow velocity between the left and right MCAs were taken into consideration. Clinical observation lasted 7 days for each patient to determine any SND and record it. SND, defined as a reduction of ≥2 points from the initial GCS or needing any surgical or medical therapy for neurological worsening during the first week after trauma, was the study’s main objective. After seven days of clinical observation, the neurologic outcome was examined physically or, if the patient had been released from the hospital, by telephone interview.

An expert panel that was blinded to the TCD results evaluated all cases with SND at the end of the trial. As a result, on post-trauma day-7, there were two groups of patients, group I: patients without SND and group II: patients with SND. The sample size calculation was done by G*Power (Universitat Kiel, Germany). The sample size was based on the following considerations: 95% confidence interval (CI), 95% power of the study and the sensitivity of GCS, PI and their combination to predict SND (the primary outcome) ranged from 40% to 76.7% according to a previous study.[6] Seventeen cases were added to each group to overcome dropout. Therefore, 105 cases were recruited in the study.

Data were fed to the computer and analyzed using IBM SPSS software package version 20.0 (IBM Corp., Armonk, New York). Quantitative parametric variables were presented as mean and standard deviation (SD) and compared between the two groups utilizing unpaired Student’s t test. Quantitative non-parametric data were presented as median and interquartile range (IQR) and were analyzed by Mann–Whitney test. Qualitative variables were presented as frequency and percentage (%) and were analyzed utilizing the χ2 test. A two-tailed P value <0.05 was considered statistically significant. It is generated by plotting sensitivity (true positive) on Y-axis versus 1-specificity (false positive) on X-axis at different cut off values. The area under the receiver operating characteristic (ROC) curve denotes the diagnostic performance of the test. Area more than 50% gives acceptable performance and area about 100% is the best performance for the test. The ROC curve allows also a comparison of performance between two tests.


The demographic data, GCS on admission, cause of trauma and length of stay in intensive care unit (days) were presented in Table 1. According to TCD measurements in group I were ranged between 58–111 cm/s, 20–50 cm/s and 0.66–1.31 with mean value 80.59 ± 10.89, 33.59 ± 6.66, and 0.97 ± 0.14 for FVs (cm/s), FVd (cm/s), and PI on admission, respectively. while TCD measurements in group II were ranged between 63–100 cm/s, 20–28 cm/s and 1.20–1.57 with mean value 78.86 ± 10.72, 23.31 ± 2.36, and 1.32 ± 0.09 for FVs (cm/s), FVd (cm/s), and PI on admission respectively, as shown in Table 2.

Table 1::
Patients characteristics and GCS on admission, cause of trauma and length of stay in ICU (days)
Table 2::
Relation between SND day-7 post-trauma with FVs (cm/s), FVd (cm/s) and PI on admission (n = 105)

According to FVs, there was no significant difference between two groups (P = 0.466) as shown in Table 2. There was significant decrease in FVd in group II compared to group I (P < 0.001) but there was significant increase in PI in group II in comparison to group I (P < 0.001) as shown in Table 2. The median GOSE at 1 month in group I was 8 (7.5–8), while in group II was 6 (4–7) with significant increase in group I in comparison with group II (P < 0.001).

Patients with good outcome (GOSE 7–8) (no disability) represented about 97.4% (No. 74) and 37.9% (No. 11) in group I and group II respectively, while patients with poor outcome (GOSE 1–6) (death or disability) represented about 2.6% (No. 2) and 62.1% (No. 18) in group I and group II, respectively. There was a highly significant difference between both groups (P < 0.001).

There was a negative correlation between PI on admission and GOSE at one month (P < 0.001) as, shown in Figure 2A. Also, there was a negative correlation between PI on admission and GCS on admission (P < 0.001), as shown in Figure 2B. According TCD measurement, ROC curve showed that the cut off of FVd was equal to or less than 25 (cm/s) to predict SND with positive predictive value (PPV) 75.8% and negative predictive value (NPV) 94.4% with sensitivity 86.2% and specificity 89.5%, as shown in Figure 3A.

Figure 2::
(A) Correlation between PI on admission with GOSE at 1 month (n = 105), GOSE: extended Glasgow outcome scale, PI: pulsitility index; (B) correlation between PI on admission with GCS (n = 105), PI: pulsitility index, GCS: Glasgow outcome scale
Figure 3::
(A) Receiver operating characteristic curve for FVd (cm/s) on admission to predict SND day 7 post-trauma (n = 29 vs. 76); (B) Receiver operating characteristic curve for PI on admission to predict SND day 7 post-trauma (n = 29 vs. 76); (C) Receiver operating characteristic curve for PI on admission, GCS and combined to predict SND day 7 post-trauma (n = 29 vs. 76), PI 0: pulsitility index on admission, GCS: Glasgow coma scale

Receiver operating characteristic curve showed that PI could predict SND at cutoff 1.21 with good PPV 87.5% and NPV 98.6% with sensitivity 96.5% and specificity 94.74%, as shown in Figure 3B. Receiver operating characteristic curve showed that the cut off of GCS on admission was less than 11 to predict SND with PPV 56.1% and NPV 90.6% with sensitivity 75.3% and specificity 73.3%, as shown in Figure 3C. Receiver operating characteristic curve showed that combined PI and GCS on admission to predict SND increase the sensitivity to 99% and specificity reached 96%, while PPV 90.6% and NPV 99% as shown in Figure 3C.


Identifying individuals at risk for early SND following mild to moderate TBI at the time of admission has not received much attention. The large majority of protocols of emergency management rely only on CT scans to detect a very small percentage of patients who may need early neurosurgery interventions.[7] TCD allows a bed side, noninvasive tool for monitoring the cerebral blood flow.[4] We conducted our study on those groups of patients who constitutes about 80% of all TBI to add viable, bed side and non-invasive tool in prediction of early SND within 1st week following mild and moderate TBI.

Our research studied 105 adult patients having mild to moderate TBI who were admitted to emergency room within 12 h after the trauma, initial assessment of GCS and TCD were done to each patient on admission. Also, CT was done and (TCDB) classification was made by a senior radiologist to include (TCDB I-II).

Primary end point was assessed SND within the first week post trauma. Of the 105 patients, 29 (27.6%) showed SND. Aiming to evaluate the value of our intervention (TCD) to predict SND after mild and moderate TBI in 1st week, we used receiver-operating characteristic analysis (ROCA), and we discovered the optimal threshold limit (cut off) for PI to be 1.21 which has good sensitivity 96.5% and specificity 94.7% when AUC 0.98. We found NPV 98.6% and PPV 87.5% (95% CI, 95%–100%) with this cut off.

To amplify our finding, we measured FVd which represents the magnitude of resistance of downstream in vessels. We used ROCA, we discovered the optimal threshold limits to be 25 cm/s with good specificity 89.5% and sensitivity 86.2% when AUC 0.93. We found NPV 94.4% and PPV 75.8% (95% CI, 88%–98%) with this threshold.

According to previous study, it may be possible to evaluate a proper pairing of PI and the TCDB classification at admission for predicting subsequent neurological deterioration for 1 week following TBI. Also, they recommended determining TCD thresholds in a multicenter study.[8]

Corroborative to our results, 98 individuals with mild to moderate TBI were studied, 21 cases (21%) demonstrated SND within the first week following trauma, which is in accordance with our findings. They discovered that FVd and PI were the most effective discriminators between patients with SND and those without SND based on the AUCs.[9]

They established thresholds of 25 cm/s (specificity, 76%; sensitivity, 92%; area under curve [AUC], 0.93) in FVd and 1.25 (specificity, 91%; sensitivity, 90%; AUC, 0.95) in PI in order to enhance specificity and sensitivity for these two parameters. They found that a significant probability of neurological deterioration during the 1st week following TBI was suggested by mild CT results combined with decreased FVd and/or higher PI values upon admission. They suggested the TCD measurements’ predictive value in TBI patients on admission who later suffer neurological deterioration.[8] In the previous study, they reported cut off values for PI and FVd from single center. So, they tried to verify TCD thresholds in order to better predict outcomes after mild and moderate TBI in large multicenter cohort study population. They discovered that (79%) of patients with no SND had normal values of TCD, whereas only (21%) had an abnormal pattern of TCD. consequently, thresholds of TCD (PI 1.25 and FVd 25cm/s) demonstrated 79% specificity and 80% sensitivity in predicting early neurologic deterioration. The TCD thresholds’ sensitivity and specificity for spotting early neurologic deterioration were satisfactory. They demonstrated that, after admission following mild to moderate TBI, TCD was achievable in a variety of EDs.[10]

In our study, we refined the role of TCD as a good negative test for prediction of SND in TBI patients with mild lesion in initial CT (i.e., TCDB II). Closing the gap in prediction of SND in this category of patients where CT scan has a weak prediction. Coming, the importance of TCD as non-invasive, bed side tool in initial evaluation, early prediction of subsequent SND in mild and moderate TBI patients. This observation may be due inherent characters of TCD on evaluation of cerebral circulation, mainly anterior.[11]

Additionally, TCD may be used to track the early onset of cerebral vasospasm following trauma-related subarachnoid hemorrhage as well as the early or delayed onset of increased ICP after a traumatic brain injury. This is accomplished by routine, non-invasive monitoring of the intracranial vasculature’ CBF and ICP.[12]

We utilized the extended GOSE, which has grown to be one of the most used result tools to measure overall impairment and recovery following TBI, to evaluate the long-term outcome of individuals with TBI at the end of 1 month. PI at admission and GOSE place at a single month had a significant negative relationship in our research (P value < 0.001).

The Disability Rating Scale (DRS) and TCD pattern have been used previously as a measure of long-term result at 28 days. When compared to patients with normal TCD, those with abnormal TCD (PI ≤ 1.25; FVd > 25 cm/s) at admission had markedly more significant DRS on day 28.[11] Other studies were conducted on sever TBI patients to assess long term follow up and concluded that TCD measurements were associated with long term outcome.[13-15] In extension to those previous results, our study observed equal predictivity of TCD measurements in cases of mild and moderate TBI. So, in all TBI cases, it may be included as a reliable indicator of the long-term outcome.

This study had some limitations such as it was a single center study with limited sample size in addition to the clinical utility of TCD in TBI, in which its use is limited due to operator dependency and in patients who lack an adequate acoustic temporal window for insonation.


TCD on admission may provide a valuable tool of early and after 1 month prediction of neurological outcome for mild and moderate TBI patients. Closing the gap in poor prediction of commonly used evaluation by CT especially with mild lesion.


Nothing to declare.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.

Availability of data and material

The datasets used and/or analyzed during the current study are available as MS Excel files (.xlsx) from the corresponding author upon reasonable request.

Ethical approval and protocol registration

The study was done after approval from the Ethical Committee of Faculty of Medicine, Tanta University, Egypt (approval code: 33574/12/19) and registration on (ID: NCT05151978).

Author contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.


1. Thapa K, Khan H, Singh TG, Kaur A. Traumatic brain injury: Mechanistic insight on pathophysiology and potential therapeutic targets. J Mol Neurosci 2021;71:1725-42
2. Wiles MD. Management of traumatic brain injury: A narrative review of current evidence. Anaesthesia 2022;77:102-12
3. Povlishock JT, Katz DI. Update of neuropathology and neurological recovery after traumatic brain injury. J Head Trauma Rehabil 2005;20:76-94
4. Bouzat P, Oddo M, Payen JF. Transcranial Doppler after traumatic brain injury: Is there a role?. Curr Opin Crit Care 2014;20:153-60
5. Cardim D, Robba C, Schmidt E, Schmidt B, Donnelly J, Klinck J, et al. Transcranial Doppler non-invasive assessment of intracranial pressure, autoregulation of cerebral blood flow and critical closing pressure during orthotopic liver transplant. Ultrasound Med Biol 2019;45:1435-45
6. Abdelhaleem NF, Othman HA, Abdel Razek GM, Abdel Galel EF. Is the combination of Glasgow coma scale and transcranial Doppler pulsatility index improving the prediction of outcome in traumatic brain injury patients?. ZUMJ 2019;25:529-38
7. Fabbri A, Servadei F, Marchesini G, Stein SC, Vandelli A. Early predictors of unfavourable outcome in subjects with moderate head injury in the emergency department. J Neurol Neurosurg Psychiatry 2008;79:567-73
8. Jaffres P, Brun J, Declety P, Bosson JL, Fauvage B, Schleiermacher A, et al. Transcranial Doppler to detect on admission patients at risk for neurological deterioration following mild and moderate brain trauma. Intensive Care Med 2005;31:785-90
9. Bouzat P, Francony G, Declety P, Genty C, Kaddour A, Bessou P, et al. Transcranial Doppler to screen on admission patients with mild to moderate traumatic brain injury. Neurosurgery 2011;68:1603-9; discussion 1609
10. Robba C, Goffi A, Geeraerts T, Cardim D, Via G, Czosnyka M, et al. Brain ultrasonography: Methodology, basic and advanced principles and clinical applications. A narrative review. Intensive Care Med 2019;45:913-27
11. Bouzat P, Almeras L, Manhes P, Sanders L, Levrat A, David JS, et al. Transcranial Doppler to predict neurologic outcome after mild to moderate traumatic brain injury. Anesthesiology 2016;125:346-54
12. Wolf ME. Functional TCD: Regulation of cerebral hemodynamics—Cerebral autoregulation, vasomotor reactivity, and neurovascular coupling. Front Neurol Neurosci 2015;36:40-56
13. Ract C, Le Moigno S, Bruder N, Vigué B. Transcranial Doppler ultrasound goal-directed therapy for the early management of severe traumatic brain injury. Intensive Care Med 2007;33:645-51
14. Zaytoun T, Fayed A, Elbeheiry A, Elsefi T. Role of transcranial Doppler ultrasound as a predictor of outcome in severe traumatic brain injury and its correlation with Glascow coma scale and full outline of unresponsiveness score. J Med Sci Clin Res 2017;5: 44-53
15. Moreno JA, Mesalles E, Gener J, Tomasa A, Ley A, Roca J, et al. Evaluating the outcome of severe head injury with transcranial Doppler ultrasonography. Neurosurg Focus 2000;8:18-7

Secondary neurological deterioration; transcranial doppler; traumatic brain injury

© 2023 Bali Journal of Anesthesiology