Predicting short-term outcomes in brain-Injured patients: A comprehensive approach with transcranial Doppler and intracranial compliance assessment

doi:10.21203/rs.3.rs-3406169/v1

Download PDF

Research Article

Predicting short-term outcomes in brain-Injured patients: A comprehensive approach with transcranial Doppler and intracranial compliance assessment

https://doi.org/10.21203/rs.3.rs-3406169/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Neurocritical patients frequently exhibit abnormalities in cerebral hemodynamic (CH) impairment and/or intracranial compliance (ICC), all of which significantly impact their clinical outcomes. Transcranial Doppler (TCD) and the cranial micro-deformation sensor (B4C) are valuable techniques for assessing CH and ICC, respectively. However, there is a scarcity of data regarding the predictive value of these techniques in determining patient outcomes.

Methods

We prospectively included neurocritical patients undergoing intracranial pressure (ICP) monitoring within the first 5 days of hospital admission for TCD and B4C assessments. Comprehensive clinical data were collected alongside parameters obtained from TCD (including blood velocities, pulsatility index, estimated ICP - eICP and estimated cerebral perfusion pressure - eCPP]) and B4C (measured as the P2/P1 ratio). These parameters were evaluated individually as well as in combination. These parameters were evaluated alone as well as in combinations among them. The short-term outcomes (STO) of interest were the therapy intensity levels (til) for icp management recommended by the SIBICC, as TIL 0 (STO 1), TIL 1–3 (STO 2) and death (STO 3), at the seventh day after last data collection.

Results

A total of 98 patients were included, with 67% having experienced traumatic brain injury; ICP, P2/P1, and eCPP demonstrated the highest ability to predict early mortality (p = 0.02, p = 0.02, and p = 0.006, respectively). P2/P1 was the only parameter significant for the prediction of STO 1 (p = 0.03). Combining B4C and TCD parameters, the highest areas under the curve (AUC) were 0.85 to predict death (STO 3), using P2/P1 + eCPP, whereas AUC was 0.72 to identify ICP > 20 mmHg using P2/P1 + eICP.

Conclusion

The combined noninvasive neuromonitoring approach using TCD and B4C demonstrated improved performance in predicting outcomes during the early phase after acute brain injury. Correlation between these parameters and intracranial hypertension was moderate, supporting the role of multimodal neuromonitoring as ancillary information in the ICU.

Trial registration

NCT03144219, Registered 01 May 2017 Retrospectively registered, https://www.clinicaltrials.gov/ct2/show/NCT03144219

acute brain injury

cerebral hemodynamics

intracranial compliance

intracranial pressure

transcranial Doppler

traumatic brain injury

Noninvasive neuromonitoring demonstrated to be valuable as ancillary information complementing invasive intracranial pressure measure in neurocritical patients.
The combination of noninvasive methods as transcranial Doppler and a cranial pulsation sensor (B4C) enhanced predicting poorer outcomes in the present study.
Both techniques demonstrated high negative predictive values for ruling out either intracranial hypertension or poorer outcomes.
The combination of these techniques was important to improve their predictive value, rather than using them separately.
The strong correlation of these techniques with outcomes but only moderate correlation with intracranial pressure values seems to reinforce the inaccuracy of determining intracranial hypertension according to a predetermined cut-offs all human beings.

Neurocritical patients who require invasive mechanical ventilation are at high risk of death and poor functional outcomes^1,2. Besides primary injuries following kinetic forces in head incidents or spontaneous intracranial bleedings, further damage may occur because of the inflammatory and immune responses in the subsequent early period after an acute brain injury (ABI)^3,4. All this combination of phenomena contribute to impairment in cerebrovascular properties, raising risks for the development of intracranial hypertension (IH), a life-threatening condition for the neurocritical patient^5–7.

Furthermore, these patients require multiple systemic and brain catheterization to their monitoring and treatment, turning medical practice in a carrier of additional risks. Therefore, noninvasive methods which may diminish medical practice burdens are of great value. In this setting, transcranial Doppler (TCD) is an ultrasound-based technique able to perform multiple cerebrovascular diagnostics at the bedside⁸. Moreover, this technique has been extensively applied in neurocritical care for its potential to estimate intracranial pressure (ICP) and cerebral perfusion pressure (CPP)⁹, with remarkable performance on discarding IH. A distinct noninvasive approach tailored for evaluating loss of compliance on the intracranial compartment was recently released for application in medical standards (B4C system)¹⁰. This technique has ability to register skull pulsations and present surrogate ICP waveforms (ICPW) in real-time, allowing physicians to observe the impact of their interventions at the bedside¹¹.

The morphological changes in ICPW are markers of intracranial compliance (ICC) impairment^12–14. Such changes are the targets of the B4C system¹⁵, by means of the quotient from the division of second and first ICPW peaks (P2/P1). Since IH indicates a decrease in ICC, it results in CPP reduction¹⁶, indicating a potential role for noninvasive monitoring combination by means of TCD and B4C, although paucity of data in this regard. Hence, the present study aimed to assess the correlation between TCD features and P2/P1 either separately or in combination- with short-term outcomes in neurocritical patients (primary endpoint). Additionally, the correlation of these parameters with ICP was also tested (secondary endpoint). Hypothetically, the power of noninvasive neuromonitoring techniques correlations is enhanced when different techniques are combined^17–19.

Study Design

We conducted a retrospective analysis of a prospective study database to investigate the prediction ability of multiple parameters for the detection of IH and patient outcome differentiation. The study was conducted in the Hospital das Clínicas intensive care units (São Paulo University, Brazil). The study protocol was approved by the local Ethics Committee and retrospectively registered at clinicaltrial.gov (NCT03144219). Informed consent was obtained from legally authorized representatives/next of kin of patients before inclusion. This study was performed according to the Standards for Reporting of Diagnostic Accuracy Studies (STARD) (Supplemental Table 1).

Population

ABI patients were included exclusively when required mechanical ventilatory support and after an invasive ICP probe monitor inserted, according to the guidelines of the Brain Trauma Foundation after neurosurgical and neurointensivist teams judgement²⁰. Exclusion criteria was the evidence of fixed mydriatic/middle-sized pupils for more than 2 h after ventilatory and hemodynamic stabilization. Patients were assessed within the first 5 days after admission. Simultaneous recording of invasive arterial blood pressure, ICP, electrocardiogram, temperature and oxygen saturation were also obtained.

The process of weaning patients from mechanical ventilation was carried out by the local ICU team independently of this study, and they followed established criteria for this purpose. These criteria included assessing the patient's ability to safeguard their airway, ensuring the maintenance of airway patency, confirming the presence of a robust cough reflex, evaluating the Glasgow coma scale (GCS) score to be above 8, monitoring a PaO2/FiO2 ratio exceeding 150, and considering the necessity for FiO2 levels greater than 0.40 or a PEEP (positive end-expiratory pressure) requirement below 10. These stringent criteria were applied to ensure the safe and appropriate weaning of patients from mechanical ventilation.

Three short-term outcome (STO) groups were separated based on to SIBICC therapy intensity level (TIL) recommendation (Table 1)²¹. Group 1 (STO 1) was composed by no specific ICP directed therapy patients, whilst group 2 (STO 2) by ICP directed TIL 1–3. Finally, group 3 (STO 3) was composed by in-ICU deaths. These outcomes were registered at day 7 after last monitoring session; therefore, the outcomes of interest were registered in the second week of hospitalization, in order to allow potential associations of changes in the variables provided by noninvasive neuromonitoring (TCD and B4C) and clinical results in a short-term period.

Table 1

Expert consensus for traumatic brain injury therapy intensity level (TIL) escalation.

(Adapted from Hawryluk et al: A management algorithm for patients with intracranial pressure monitoring: the Seattle International Severe Traumatic Brain Injury Consensus Conference).
TIL 0	No specific ICP-directed therapy. Basic ICU care: sedation, vasopressors, overall variables control and head positioning not based on ICP management.
TIL 1	Mild ICP-directed therapy: Deeper sedation, vasopressors, osmotics, hypocapnia and CSF drainage according to ICP/CPP.
TIL 2	Moderate ICP-directed therapy: More rigorous TIL 2 plus mild hypothermia ˜35⁰C.
TIL 3	Extreme ICP-directed therapy: Deeper hypocapnia, hypothermia, barbiturates, neurosurgery.

Data sources

TCD (Doppler Box, DWL, Singen, Germany) recorded CBv form middle cerebral arteries bilaterally through the temporal acoustic windows. In case of more than 20% CBv values difference between sides, the side with higher injury was chosen for the statistical analysis. P2/P1 was collected using the brain4care (B4C) system (Brain4care, São Carlos, Brazil). The P2/P1 is a parameter derived from the surrogate B4C waveforms, which are waveforms correlated with the ICP changes¹⁵, obtained after the discovery of the principle of skull microexpansion^22,23. ICP was measured using the Neurovent monitoring system via optic-fiber transducer (Raumedic, Munchberg, Germany); mean ICP values were registered for every monitoring session. All sessions last about 10 minutes and were performed independently of therapeutic interventions, which were not influenced by the variables of interest for this study (Fig. 1).

Data Preprocessing

Prior to the analysis, the dataset underwent preprocessing steps to ensure data quality and compatibility. The parameters of interest were, from TCD: mean cerebral blood velocity (CBv), pulsatility index (PI), estimated ICP (eICP) and estimated CPP (eCPP)¹⁹. From B4C, the P2/P1, which represents the marker of buffering reserve exhaustion and the move towards IH²⁴. Both techniques (B4C and TCD) parameters have demonstrated significant correlations with ICP in previous studies^15,19,25. These variables were all recorded concomitantly with ICP and their mean values were calculated automatically for statistical analysis. IH was defined in this study as ICP > 20 mmHg.

Data Analysis

Data analysis was performed in two ways. First, each proposed method as evaluated separately as to their abilities to predict the presence or absence of IH and differentiating favorable outcomes and death (STO 1 and 3). Secondly, P2/P1 and each respective parameter derived from TCD were combined. The following steps were performed for this secondary analysis:

Splitting the Data: the dataset was randomly divided into a training set and a testing/validation set using a 50:50 ratio. The training set (N = 50 measurements) was used for model development, while the testing/validation set (N = 50 measurements) was used for model evaluation.

Logistic Regression Model: a logistic regression model was constructed using the training set. The response variables were “intracranial hypertension” and “outcomes” (STO levels 1 and 3) which were encoded as binary variables (0 for absence, 1 for presence) in separate columns in the dataset. The predictor variables included the selected parameters. The logistic regression model was fitted using the `glm()` function from the R statistical software, with the family set to "binomial".

Model Evaluation: the fitted logistic regression model was then used to generate predictions on the testing/validation set. Predicted probabilities were obtained using the `predict()` function. To assess the performance of the model, a receiver operating characteristic (ROC) curve analysis was conducted. The ROC curve was plotted using the `pROC` package in R²⁶, and the area under the ROC curve (AUC-ROC) was calculated as a measure of the model's predictive ability.

Statistical Analysis

Quantitative and qualitative variables were summarized differently: absolute (n) and relative frequencies (percentage) were used for qualitative variables, while mean ± standard deviation and medians with minimum and maximum values were employed for quantitative variables. The correlation between the noninvasive methods and the absolute value of intracranial pressure (ICP) was evaluated using Spearman’s correlation coefficient. To assess the effectiveness on distinguishing between favorable outcome and non-survivors (STO 1 and 3, respectively), the area under the curve (AUC) of the receiver operating characteristic (ROC) curve was calculated for each method. Additionally, an analysis was conducted to determine optimal cutoff values for predicting ICP > 20 mmHg using the Youden index. These values represent the optimal balance between sensitivity and specificity for a given threshold. An AUC exceeding 0.7 indicates reasonable predictive ability, while an AUC surpassing 0.8 signifies strong predictive ability²⁷. Statistical comparisons between ROC curves were performed using DeLong's test for two correlated ROC curves (R package pROC²⁶). To jointly analyze the relationships of these methods with outcomes, binary logistic regression models were employed, utilizing generalized estimated equations. The findings were expressed as odds ratios (ORs) along with their respective 95% confidence intervals and p-values. The dependent variables were the outcomes, whereas the independent variables were the noninvasive methods. Additionally, an analysis of variance (ANOVA) was executed to explore potential associations between the assessed variables and the different STO levels within the patient cohort. All statistical analyses were carried out using RStudio (R version 4.3.1), with statistical significance determined by a p-value threshold of 0.05.

152 patients were consecutively included but only 98 of these had both TCD and B4C satisfactorily recorded for analysis. Patients were middle-aged (± 41 y.o.) predominantly male (65%) and hospitalized because of TBI (67%) as show in Table 2. Among these, 19 instances of IH were detected for the analysis contemplating individual methods, while 11 cases were identified in the subset of 50 measurements from the testing dataset for the analysis contemplating combined methods (TCD + B4C). The analysis was also conducted at different STO levels. At levels 1 and 3, 37 measurements were considered from the original dataset, with 17 measurements showing STO 3 for the individual methods analysis. Additionally, from the testing dataset (N = 50) 14 measurements were identified at STO levels 1 and 3, and out of these, 8 measurements exhibited level 3 for the combined methods analysis. No significant differences regarding admission characteristics in the population assessed were found.

Table 2

Baseline characteristics of the patient cohort and mean (SD) values of the studied parameters.
	All (n = 98)	STO 1 (n = 20)	STO 2 (n = 62)	STO 3 (n = 16)
Age, years	41.90 (21.11)	38.55 (18.82)	40.59 (21.23)	50.65 (22.08)
Sex, n males (%)	64 (65)	14 (70)	41 (66)	9 (56)
Parameters
mICP, mmHg	14.70 (7.60)	14.10 (5.60)	12.97 (5.24)	21.59 (12.37)^c
mICP ≥ 20 mmHg, n (%)	19 (19)	4 (20)	6 (10)	9 (56)
P2/P1	1.10 (0.27)	1.05 (0.24)	1.08 (0.27)	1.27 (0.23)^b,c
P2/P1 ≥ 1.2, n (%)	40 (41)	8 (40)	21 (34)	11 (69)
MAP, mmHg	89.60 (11.40)	88.55 (10.48)	90.17 (11.77)	88.65 (11.55)
CBv, cm/s	79.10 (28.20)	78.38 (26.63)	79.83 (28.90)	77.32 (29.23)
PI	0.90 (0.25)	0.89 (0.21)	0.87 (0.20)	1.04 (0.37)
eICP, mmHg	18.80 (11)	19.56 (9.65)	16.69 (7.90)	26.02 (18.11)
eCPP, mmHg	70.50 (13.70)	71.76 (12.64)	72.47 (12.74)	61.54 (15.67)^c
SO₂ (%)	98 (3)	98.30 (2.43)	98.03 (3.32)	97.41 (2.37)
PaCO₂, mmHg	37 (5.49)	37.30 (7.03)	36.6 (5.29)	36.15 (3.67)
Hemoglobin, mg/dL	9.56 (1.65)	9.60 (1.73)	9.50 (1.53)	9.72 (2.07)
Temperature, ºC	36.12 (0.93)	36.78 (0.81)	36.06 (0.90)^a	35.56 (0.74)
Pathology
TBI	66	13	49	10
SAH	15	3	7	5
Ischemic Stroke	8	2	6	-
Hemorrhagic Stroke	8	2	6	-
Brain Neoplasm	1	-	-	1
Neurosurgery
No	27	5	14	7
Craniotomy	42	9	20	4
Craniectomy	31	6	28	5
Comorbidities
None	55	10	36	8
Metabolic Syndrome	21	4	14	3
Others	17	4	9	4
SAPS-3 Score	58.84 (12.22)	54 (10.40)	59.63 (12.23)	61.65 (13.29)
Admission GCS	5.73 (4.40)	5.75 (4.35)	5.92 (4.58)	5 (3.95)
a: indicates significant statistical difference between STO groups 1 and 2
b: indicates significant statistical difference between STO groups 1 and 3
c: indicates significant statistical difference between STO groups 2 and 3
CBv: cerebral blood velocities, eCPP: estimated cerebral perfusion pressure, mICP: mean intracranial pressure, eICP: estimated intracranial pressure, GCS: Glasgow coma score, MAP: mean arterial pressure, PaCO₂: carbon dioxide pressure, PI: pulsatility index, SAH: subarachnoid hemorrhage, SAPS3: simplified acute physiologic scale, SO₂: oxygen saturation, STO: short-term outcome, TBI: traumatic brain injury.

TCD and B4C results

Detailed results of the receiver operating characteristic (ROC) analysis, encompassing metrics such as AUC, sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), are outlined in Table 3. P2/P1 and eCPP were the parameters significantly different between both survivor groups (STO 1 and 2) and the STO 3 group (Fig. 2). For mortality prediction, the P2/P1 exhibited an AUC of 0.70 (95% CI 0.53–0.87), alongside a strong NPV of 100%. Among TCD-derived parameters, for mortality prediction, eCPP displayed the highest AUC of 0.70 (95% CI 0.52–0.88), also coupled with a strong NPV of 100%. Detailed results of the ROC analysis and the IH cutoff value, encompassing metrics such as AUC, sensitivity, specificity, accuracy, PPV and NPV are outlined in Table 4. Best linear correlations with ICP were seen for eICP and eCPP (Fig. 3).

Table 3

Results of receiver operating characteristic analysis for mortality (N = 37 measurement points of the original dataset for single methods; N = 14 measurement points of the testing dataset for combined methods). CBv: cerebral blood velocities, eCPP: estimated cerebral perfusion pressure, eICP: estimated intracranial pressure, ICP: intracranial pressure, PI: pulsatility index, P2/P1: quotient from second and first brain4care waves peaks.
Parameters	AUC (95% CI)	Sensitivity (%)	Specificity (%)	Accuracy (%)	PPV (%)	NPV (%)
ICP	0.69 (0.52–0.87)	100	5	49	47	100
P2/P1	0.70 (0.53–0.87)	100	5	49	47	100
CBv	0.47 (0.28–0.67)	100	5	49	47	100
PI	0.62 (0.42–0.81)	94	0	43	44	0
eICP	0.59 (0.40–0.78)	100	5	49	47	100
eCPP	0.70 (0.52–0.88)	100	5	49	47	100
P2/P1 + PI	0.83	100	17	64	61	100
P2/P1 + CBv	0.81	100	17	64	61	100
P2/P1 + eICP	0.71	100	17	64	61	100
P2/P1 + eCPP	0.85	100	17	64	61	100

Table 4

Results of receiver operating characteristic analysis for ICP > 20 mm Hg (N = 100 measurement points of the original dataset for single methods; N = 50 measurement points of the testing dataset for combined methods). CBv: cerebral blood velocities, eCPP: estimated cerebral perfusion pressure, eICP: estimated intracranial pressure, PI: pulsatility index, P2/P1: quotient from second and first brain4care waves peaks.
Parameters	AUC (95% CI)	IH cutoff	Sensitivity (%)	Specificity (%)	Accuracy (%)	PPV (%)	NPV (%)
P2/P1	0.66 (0.53–0.78)	1.13	100	1	20	19	100
PI	0.66 (0.51–0.82)	0.96	95	1	19	18	19
CBv	0.67 (0.41–0.69)	92.65	100	1	20	19	100
eICP	0.67 (0.50–0.83)	25.25	100	1	20	19	100
eCPP	0.67 (0.51–0.82)	64.09	100	1	20	19	100
P2/P1r + PI	0.70	-	100	3	24	22	100
P2/P1 + CBv	0.67	-	100	3	24	22	100
P2/P1 + eICP	0.72	-	100	3	24	22	100
P2/P1 + eCPP	0.67	-	100	3	24	22	100

Combined methods

In the prediction analysis of STO 3, the top-performing methods included "P2/P1 + eCPP," "P2/P1 + PI," and "P2/P1 + CBv," achieving AUCs of 0.85, 0.83, and 0.81, respectively. All these methods showcased robust NPVs of 100% and moderate PPVs of 61% (Table 3). Table 5 presents a comprehensive summary of the outcomes derived from logistic regression models based on combined noninvasive methods. To distinguish IH, the methods "P2/P1 + PI" and "P2/P1 + eICP" emerged as the best performers, showing AUCs of 0.70 and 0.72, respectively. Moreover, they demonstrated NPVs of 100% (Fig. 2) (Table 4). Despite marginally enhanced AUCs for IH assessment, these combined methods did not exhibit statistical significance when compared to the individual methods (P2/P1, PI, and eICP).

Table 5

Summary of the logistic regression results for the combined noninvasive methods.
Model	Predictor	Odds Ratio	95% CI Lower	95% CI Upper	p-value
P2/P1 + PI	Intercept	0.0006	0.00003	0.12	0.006
	P2/P1	8.81	0.38	203.76	0.17
	PI	26	1.34	503.58	0.03
P2/P1 + CBv	Intercept	0.01	0.0002	0.61	0.03
	P2/P1	6.19	0.33	116	0.22
	CBv	1.01	0.99	1.03	0.43
P2/P1 + eICP	Intercept	0.01	0.0002	0.50	0.02
	P2/P1	4.77	0.23	99.15	0.31
	eICP	1.05	0.97	1.15	0.24
P2/P1 + eCPP	Intercept	1.90	0.01	320.55	0.81
	P2/P1	10.08	0.35	286.81	0.18
	eCPP	0.92	0.86	0.99	0.03
CBv: cerebral blood velocities, eCPP: estimated cerebral perfusion pressure, eICP: estimated intracranial pressure, PI: pulsatility index, P2/P1: quotient from second and first brain4care waves peaks.

Relationship with short-term outcomes

16 patients died (STO 3) and 82 survived (STO levels 1 and 2). Twenty patients had most favorable outcomes (STO 1). Figure 3 depicts the noninvasive methods that presented at least one association with patient outcome.

The methods that presented significant difference between favorable outcome and non-survivors were eCPP and P2/P1 (Fig. 2). Among these, P2/P1 demonstrated the highest significance level (p = 0.03) to differentiate these outcomes. ICP was significantly higher in patients who died compared to those of group STO 2. STO groups 1 and 2 median ICP levels were under 20 mmHg, whereas the opposite succeeded for the mortality group.

In the present cohort of 98 acute brain injured patients, two noninvasive techniques were used to register biometrical parameters for the assessment of ICC and CH. CH and ICC impairments were indicators of poorer short-term outcomes. Furthermore, the severity of such disturbances observed in the early days after injury was a reliable prognostic factor, being ICP, P2/P1 and eCPP factors significantly associated with early death in the ICU. On the contrary, patients who exhibited early progression toward successful spontaneous breathing showed either mild or no CH and ICC impairment. These significant correlations with outcomes but moderate associations between CH and ICC with IH might reinforce the inaccuracy of such thresholds (as > 20 mmHg) to differentiate patients at stake of inadvertent secondary injuries.

The prognostic power by means of the P2/P1 was modest but with excellent NPV, indicating this parameter between 0.8–1.1 valuable to rule out ICC impairment. In the present study, a mean P2/P1 over 1.2 among was found as indicator of unfavorable outcome, which is in concordance with previous studies^24,28. Regarding the several TCD parameters, eCPP was significantly lower in the group of patients who died, but overall, TCD parameters also disclosed strong NPVs and moderate PPVs. However, when taken in combination, TCD and B4C prognostication power was enhanced. These findings remark the prognostic role of noninvasive neuromonitoring, pointing to the possibility of targeting therapies on such information in future studies.

It has been reported that neurocritical patients exhibit an extended duration of mechanical ventilation and a heightened prevalence of extubation failure in comparison to the broader critically ill²⁹. Notably, in studies specifically investigating patients with ABI, a noteworthy proportion of approximately 35% needed the implementation of tracheostomy procedures²⁹. Furthermore, a systematic review including 7929 patients indicated that several parameters are considered for the decision on mechanical ventilation weaning, being age, level of consciousness, the inspiratory maximum pressure, the rapid shallow breathing index and overall disease severity scales the most used³⁰. However, these parameters are assessed concomitantly with sedation weaning. It is important to highlight that none of these studies assessed brain dynamic conditions as CH and ICC as markers of brain health to support mechanical ventilation withdrawal, which according to the findings of the present study may be advocated as ancillary information to assess in such process.

Robba et al. previously observed a prediction power improvement of combining TCD, pupillometry and optic nerve sheath ultrasound, rather than using such techniques separately^17,31. Godoy et al. proposed a model for the combination of these same techniques but including also ICP pulse morphology with the purpose of understanding further the intracranial compartment syndrome¹⁸. Brasil et al. assessing 72 ABI patients observed the potential refinement of ICP invasive monitoring coupled with B4C waveforms on outcome prediction, advocating not only for the potential benefit on adding noninvasive techniques even when an ICP catheter is implanted²⁴.

TCD was concentrated in the past on the PI as maker of raised ICP³². Notwithstanding, many studies brought controversial results for the PI in this regard, concluding this parameter to be more an indicator of reduced CPP (when PI is raised > 1.4) than elevated ICP³³. Our study is in accordance with this latter point of view. The sample of this study was composed predominantly by TBI patients in early days after injury present with posttraumatic hyperemia³. We observed elevated mean CBv (> 70 cm/s) and low PI (˜0.9) in all groups, what reinforce this point. Therefore, calculating eICP and eCPP is advised when using TCD to evaluate neurocritical patients in the early days after injury. Additionally, as demonstrated, being B4C available, its use altogether with TCD will probably enhance diagnostic power.

The main limitation to acknowledge in this study is the absence of a longer period cohort, which would allow more robust associations between parameters and outcomes. Nevertheless, it is worth noting that there is a dearth of literature focusing on short-term outcomes within this specific patient group, and it's possible that long-term outcomes may not necessarily be correlated with cerebral hemodynamics during the initial five days of admission. Second, some IH surges may have occurred without being registered in this study. Moreover, due to the few datapoints recorded, the associations assessed do not prove causality between parameters and outcomes. It is well known that several other variables not measured here play fundamental roles in outcome determination, nevertheless, the absence of significant clinical differences between groups at least raise the possibility of the present analysis making true sense.

In the present study performed in early days after acute brain injuries, noninvasive cerebral hemodynamics and intracranial compliance assessments were correlated with early clinical outcomes. Such potential was enhanced when considering a combination of techniques results, advocating on behalf using multimodal monitoring in neurocritical care.

The manuscript complies with all instructions to authors.

All authorship requirements have been met and the final manuscript was approved by all authors.

This manuscript has not been published elsewhere and is not under consideration by another journal.

STARD reporting checklist was added as supplemental material.

Authors’ contributions: SB conceived the study, collected data and wrote manuscript. MHY collected electronic data. WSP provided supervision and training. DC performed data analysis and statistical analysis. SB, FST, CR, JC and MLO revised manuscript. All authors agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship and all those who qualify for authorship are listed. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Competing interests: SB and DC are scientific advisors for brain4care. All other authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The authors confirm compliance with ethical approval and informed consent for human studies.

Funding: The present study received no funding.

Sérgio Brasil ORCID https://orcid.org/0000-0003-2397-9947

Lahiri S, Mayer SA, Fink ME, et al. Mechanical Ventilation for Acute Stroke: A Multi-state Population-Based Study. Neurocrit Care. 2015;23(1):28–32. 10.1007/s12028-014-0082-9.
Chesnut R, Aguilera S, Buki A et al. A management algorithm for adult patients with both brain oxygen and intracranial pressure monitoring: the Seattle International Severe Traumatic Brain Injury Consensus Conference (SIBICC). Intensive Care Med 2020;46(5):919–929. 10.1007/s00134-019-05900-x.
Martin NA, Patwardhan RV, Alexander MJ, et al. Characterization of cerebral hemodynamic phases following severe head trauma: hypoperfusion, hyperemia, and vasospasm. J Neurosurg. 1997;87(1):9–19. 10.3171/jns.1997.87.1.0009.
Lazaridis C. Brain Shock-Toward Pathophysiologic Phenotyping in Traumatic Brain Injury. Crit Care Explor. 2022;4(7):e0724. 10.1097/CCE.0000000000000724.
Mataczynski C, Kazimierska A, Uryga A, Kasprowicz M. Intracranial Pressure Pulse Morphology-based Definition of Life-threatening Intracranial Hypertension Episodes. Annu Int Conf IEEE Eng Med Biol Soc. 2022;2022:1742–6. 10.1109/embc48229.2022.9871403. (In eng).
Zeiler FA, Ercole A, Cabeleira M, et al. Patient-specific ICP Epidemiologic Thresholds in Adult Traumatic Brain Injury: A CENTER-TBI Validation Study. J Neurosurg Anesthesiol. 2019. 10.1097/ANA.0000000000000616.
Brasil S, Godoy DA, Paiva WS. Doing more with less on intracranial pressure monitoring. World Neurosurg. 2023. 10.1016/j.wneu.2023.07.055.
Bittencourt Rynkowski C, Caldas J. Ten Good Reasons to Practice Neuroultrasound in Critical Care Setting. Front Neurol. 2021;12:799421. 10.3389/fneur.2021.799421.
Rasulo FA, Calza S, Robba C, et al. Transcranial Doppler as a screening test to exclude intracranial hypertension in brain-injured patients: the IMPRESSIT-2 prospective multicenter international study. Crit Care. 2022;26(1):110. 10.1186/s13054-022-03978-2.
Brasil S, Godoy DA, Hawryluk GWJ. A Point-of-Care Noninvasive Technique for Surrogate ICP Waveforms Application in Neurocritical Care. Neurocrit Care. 2023. 10.1007/s12028-023-01786-2.
de Moraes FM, Rocha E, Barros FCD, et al. Waveform Morphology as a Surrogate for ICP Monitoring: A Comparison Between an Invasive and a Noninvasive Method. Neurocrit Care. 2022;37(1):219–27. 10.1007/s12028-022-01477-4.
Ocamoto GN, Russo TL, Mendes Zambetta R, et al. Intracranial Compliance Concepts and Assessment: A Scoping Review. Front Neurol. 2021;12:756112. 10.3389/fneur.2021.756112.
Hu X, Glenn T, Scalzo F, et al. Intracranial pressure pulse morphological features improved detection of decreased cerebral blood flow. Physiol Meas. 2010;31(5):679–95. 10.1088/0967-3334/31/5/006.
Rubiano AM, Figaji A, Hawryluk GW. Intracranial pressure management: moving beyond guidelines. Curr Opin Crit Care. 2022;28(2):101–10. 10.1097/MCC.0000000000000920.
Brasil S, Solla DJF, Nogueira RC, Teixeira MJ, Malbouisson LMS, Paiva WDS. A Novel Noninvasive Technique for Intracranial Pressure Waveform Monitoring in Critical Care. J Pers Med. 2021;11(12). 10.3390/jpm11121302.
Brasil S, Nogueira RC, Salinet ASM, et al. Contribution of intracranial pressure to human dynamic cerebral autoregulation after acute brain injury. Am J Physiol Regul Integr Comp Physiol. 2023;324(2):R216–26. 10.1152/ajpregu.00252.2022.
Robba C, Frigieri G, Brasil S, Taccone FS. Early prognostic value of non-invasive intracranial pressure methods in brain-injured patients. Intensive Care Med. 2022;48(12):1812–4. 10.1007/s00134-022-06899-4.
Godoy DA, Brasil S, Iaccarino C, Paiva W, Rubiano AM. The intracranial compartmental syndrome: a proposed model for acute brain injury monitoring and management. Crit Care. 2023;27(1):137. 10.1186/s13054-023-04427-4.
Robba C, Pozzebon S, Moro B, Vincent JL, Creteur J, Taccone FS. Multimodal non-invasive assessment of intracranial hypertension: an observational study. Crit Care. 2020;24(1):379. 10.1186/s13054-020-03105-z.
Rubiano AM, Puyana JC. Intracranial-pressure monitoring in traumatic brain injury. N Engl J Med. 2013;368(18):1748. 10.1056/NEJMc1301076.
Hawryluk GWJ, Aguilera S, Buki A et al. A management algorithm for patients with intracranial pressure monitoring: the Seattle International Severe Traumatic Brain Injury Consensus Conference (SIBICC). Intensive Care Med 2019. 10.1007/s00134-019-05805-9.
Andrade RAP, Oshiro HE, Miyazaki CK, et al. A Nanometer Resolution Wearable Wireless Medical Device for Non Invasive Intracranial Pressure Monitoring. IEEE Sens J. 2021;21(20):22270–84. 10.1109/JSEN.2021.3090648.
Mascarenhas S, Vilela GH, Carlotti C, et al. The new ICP minimally invasive method shows that the Monro-Kellie doctrine is not valid. Acta Neurochir Suppl. 2012;114:117–20. 10.1007/978-3-7091-0956-4_21.
Brasil S, Frigieri G, Taccone FS, et al. Noninvasive intracranial pressure waveforms for estimation of intracranial hypertension and outcome prediction in acute brain-injured patients. J Clin Monit Comput. 2023;37(3):753–60. 10.1007/s10877-022-00941-y.
Cardim D, Robba C, Czosnyka M, et al. Noninvasive Intracranial Pressure Estimation With Transcranial Doppler: A Prospective Observational Study. J Neurosurg Anesthesiol. 2019. 10.1097/ANA.0000000000000622.
Robin X, Turck N, Hainard A, et al. pROC: an open-source package for R and S + to analyze and compare ROC curves. BMC Bioinformatics. 2011;12:77. 10.1186/1471-2105-12-77.
David W, Hosmer SL Jr., Rodney X. Sturdivant. Applied Logistic Regression. Third ed: John Wiley & Sons, Inc.; 2013.
Brasil S, Taccone FS, Wayhs SY, et al. Cerebral Hemodynamics and Intracranial Compliance Impairment in Critically Ill COVID-19 Patients: A Pilot Study. Brain Sci. 2021;11(7). 10.3390/brainsci11070874.
Cinotti R, Mijangos JC, Pelosi P, et al. Extubation in neurocritical care patients: the ENIO international prospective study. Intensive Care Med. 2022;48(11):1539–50. 10.1007/s00134-022-06825-8.
Baptistella AR, Sarmento FJ, da Silva KR, et al. Predictive factors of weaning from mechanical ventilation and extubation outcome: A systematic review. J Crit Care. 2018;48:56–62. https://doi.org/10.1016/j.jcrc.2018.08.023.
Sheehan JR, Liu X, Donnelly J, Cardim D, Czosnyka M, Robba C. Clinical application of non-invasive intracranial pressure measurements. Br J Anaesth. 2018;121(2):500–1. 10.1016/j.bja.2018.04.017.
Bellner J, Romner B, Reinstrup P, Kristiansson KA, Ryding E, Brandt L. Transcranial Doppler sonography pulsatility index (PI) reflects intracranial pressure (ICP). Surg Neurol. 2004;62(1):45–51. 10.1016/j.surneu.2003.12.007. discussion 51. (In eng).
Cardim D, Robba C, Donnelly J, et al. Prospective Study on Noninvasive Assessment of Intracranial Pressure in Traumatic Brain-Injured Patients: Comparison of Four Methods. J Neurotrauma. 2016;33(8):792–802. 10.1089/neu.2015.4134.

STARD.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Predicting short-term outcomes in brain-Injured patients: A comprehensive approach with transcranial Doppler and intracranial compliance assessment

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Trial registration

Figures

TAKE HOME MESSAGES

INTRODUCTION

METHODS

Study Design

Population

Data sources

Data Preprocessing

Data Analysis

Statistical Analysis

RESULTS

TCD and B4C results

Combined methods

Relationship with short-term outcomes

DISCUSSION

CONCLUSIONS

Declarations

References

Supplementary Files

Status:

Version 1