Continuous Thermal Diffusion-Based Cerebral Blood Flow Monitoring in Adult Traumatic Brain Injury: A Scoping Systematic Review

Abstract

Thermal diffusion flowmetry (TDF) is an appealing candidate for monitoring of cerebral blood flow (CBF) in neurocritical-care patients as it provides absolute measurements with a high temporal resolution, potentially allowing for bedside intervention that could mitigate secondary injury. We performed a systematic review of TDF-regional(r)CBF measurements and their association with (1) patient functional outcome, (2) other neurophysiological parameters, and (3) imaging-based tissue outcomes. We searched MEDLINE, EMBASE, SCOPUS, BIOSIS, GlobalHealth, and the Cochrane Databases from inception to October 2018 and relevant conference proceedings published over the last 5 years. Nine articles that explored the relationship between TDF-rCBF, mortality, and Glasgow Outcome Scale (GOS) or GOS-Extended (GOS-E) at various intervals were included. Despite being based on an overall weak body of evidence, our analysis suggests a link between sustained low or high CBF and poor functional outcome. Twenty-five studies reporting associations with neurophysiological parameters were included. The available data also point to an association between low or high TDF-rCBF and intracranial hypertension. TDF-rCBF appears to correlate well with regional brain tissue oxygenation measurements. We found no studies reporting on imaging-based tissue outcome in relation to TDF. In conclusion, despite being based on a relatively weak body of evidence, the available literature suggests a link between consistently abnormal TDF-rCBF values, intracranial hypertension, and poor functional outcome. TDF-rCBF also appears to correlate well with regional measurements of brain tissue oxygenation. Currently, such monitoring should be considered experimental, requiring much further evaluation prior to widespread adoption.

Introduction

Thermal diffusion flowmetry (TDF) allows for real-time continuous measurement of absolute cerebral blood flow (CBF). There is growing interest in employing this monitoring technique in patients with severe acute neurological insults such as traumatic brain injury (TBI) as it theoretically facilitates early bedside detection of ischemic brain tissue at risk for secondary injury.^1

–4 In contrast, the traditional positron emission tomography (PET) and computed tomography (CT)-based radiographic methods for CBF measurement only allow for snapshot assessments and may not be suitable for unstable patients who are not able to tolerate transport for image acquisition,⁵ and it is uncertain if invasive brain tissue oxygen (PbtO₂) monitors measure a surrogate of true CBF versus extracellular oxygen diffusion capacity.

Commercially available TDF probes (Bowman Perfusion Monitor and Saber) work by calculating the power required to maintain a temperature difference between a proximal and a distal thermistor, which is directly proportional to cerebral tissue perfusion. The probes can either be placed on a cortical surface of interest (Saber) or directly in the brain parenchyma (Bowman Perfusion Monitor). Despite long-standing use of TDF in experimental models and recent interest in incorporating it into a multi-modal monitoring regimen, there is currently no consensus on thresholds for intervention based on TDF-CBF measurements, or a comprehensive understanding of how to employ such a device for standard monitoring of the critically ill TBI patient. The consensus summary by the International Multidisciplinary Consensus Conference on Multimodality Monitoring provides little additional guidance, merely stating that a “TDF probe may be used to identify patients with focal ischemic risk within the vascular territory of the probe,” a “weak” recommendation based on “very low quality evidence.”⁶ However, a comprehensive breakdown and review of the literature was not provided, particularly the association between TDF-measured CBF and patient-oriented outcomes such as global functional outcome, other physiological measures, and imaging-based tissue outcomes.

The purpose of this scoping systematic review was to synthesize the available evidence on TDF in adult TBI in a robust and comprehensive manner to understand the association between this monitored variable and: global patient outcome, other measured aspects of cerebral physiology, and documented imaging-based tissue fate. Our review was designed to address the following three key questions:

Is there an association between TDF-based CBF measurements and functional patient outcomes?

Is there an association between TDF-based CBF measurements and commonly monitored neurophysiological parameters?

Is there an association between CBF parameters measured by thermal diffusion and imaging-based tissue outcomes?

Methods

A scoping systematic review of the literature was conducted using the methodology described in the Cochrane Handbook for Systematic Reviewers.⁷ Data items were reported in accordance with the preferred reporting system for systematic reviews and meta-analyses (PRISMA).⁸ The search questions and search strategy were developed by consensus between the primary author and the senior author (FM and FAZ).

Search questions

The primary and secondary research questions for this review were:

Primary

Is there an association between CBF parameters measured by thermal diffusion and functional patient outcome in TBI patients in need of neurocritical care?

Secondary

(A) Is there an association between CBF parameters measured by thermal diffusion and other commonly measured neurophysiological parameters in TBI patients in need of neurocritical care? (B) Is there an association between CBF parameters measured by thermal diffusion and imaging-based tissue outcomes?

Functional outcome

For our primary research question, functional outcome measures of interest included any description of morbidity, mortality, functional outcome scale score (e.g., Glasgow Outcome Scale [GOS], or GOS-Extended [GOS-E]) or cognitive function assessments. Complications associated with TD monitoring were also recorded as a secondary outcome measure when available.

Neurophysiological outcome and imaging-based outcome

For our secondary research questions, measures of interest included any monitoring correlates of cerebral physiology such as intracranial pressure and compliance, autoregulation indices, brain tissue oxygenation, microdialysis analytes, entropy, and non-TD measures of cerebral CBF. Tissue outcome of interest was radiological evidence of ischemia, infarction, or atrophy of brain parenchyma during the acute hospital stay or at long-term follow-up, utilizing CT, computed tomographic angiography (CTA), computed tomographic perfusion (CTP), xenon CT (Xe-CT), magnetic resonance imaging (MRI), MR angiography (MRA), MR perfusion (MRP), MR spectroscopy (MRS), and PET.

Inclusion and exclusion criteria

Inclusion criteria were all studies enrolling adult (18 years or older) moderate and severe TBI (GCS score 3–12) patients in which TDF was used and any of the primary or secondary outcomes of interest were reported. Studies that enrolled pediatric or mild TBI patients exclusively, non-English studies, and case reports or studies that used non-TD-based CBF probes were excluded.

Search strategy

MEDLINE, EMBASE, SCOPUS, BIOSIS, GlobalHealth, and the Cochrane Databases were searched from inception to October 2018 using the strategy detailed in Supplementary Appendix A. Relevant meeting proceedings for the last 5 years were also screened to identify unpublished work on TDF use in the TBI population. A list of society meetings searched can be found in Supplementary Appendix B. Lastly, all the review articles discussing TDF use in TBI that we identified through our database search were set aside and their reference lists were searched for additional primary research articles meeting our inclusion criteria.

Study selection

Two of the authors (FM, FAZ) reviewed all of the articles returned by our search strategy in a two-step process. A first filter was conducted based on manuscript titles and abstracts to identify articles potentially meeting our inclusion criteria. A second filter was then performed on the identified articles by reviewing their full text to confirm that they met the inclusion criteria and that the primary or secondary outcome measures of interest were reported. Any discrepancies between the two reviewers were discussed and resolved by involving a third reviewer if required.

Data collection

Data of interest pertaining to demographics, study characteristics, and outcome measures were extracted from the final list of relevant articles and abstracts and compiled in an electronic database. Specific data fields included article location, study type, sample size, mean age, functional and physiological outcome measures, and complications.

Bias assessment

For all the final articles and abstracts to be included in our review, risk of bias was assessed using the RTI item bank.⁹ Based on scores for each of the 29 bank items, studies were classified as either high risk (H), low risk (L), or unclear risk (U) for each of the bias type sub-domains. This assessment can be found in Supplementary Appendix C.

Statistical analyses

The heterogeneity in outcome reporting and lack of robust quantitative data for most of the studies included in this review did not allow for a meta-analysis.

Results

Search strategy results

A PRISMA flow diagram outlining our search strategy results at each step can be found in Figure 1. A total of 1711 articles were identified through our database search. Of those, 141 duplicates were removed. Five additional relevant articles were found through a reference list check of review articles, and one abstract was identified from the gray literature search. The first filter through titles and abstracts applying our inclusion and exclusion criteria yielded 48 eligible articles. Our subsequent full text review provided 25 final articles that met our criteria for the functional and/or physiological outcome measures. We identified no studies that reported imaging-based tissue outcome in relation to TDF measurements.

FIG. 1.

PRISMA flow diagram.

Study description and demographics

Of the 25 articles included in this review, 17 are formal publications^{10

–27} and 8 are meeting abstracts.^{18,28

–33} Of these articles, 23 have a prospective ^{10,11,13

–33} and 2 have a retrospective observational design.^12,34 In total, 18 studies included only severe TBI (defined as admission Glasgow Coma Scale [GCS] score 3–8) patients,^{11
–13,15–18,28
–30,33} 3 studies also included patients with moderate TBI (GCS score 9–13),^14,21,27 and 2 studies included TBI patients of unspecified severity.^31,34 Of the studies enrolling severe TBI patients only, 3 studies focused on patients undergoing a neurosurgical procedure.^10,16,34 Lastly, 2 studies also enrolled patients with a non-TBI acute neurological insult such as a spontaneous intracerebral hemorrhage²³ or subarachnoid hemorrhage.²⁶ The number of subjects enrolled ranged from 3 to 56, but one abstract did not specify the total sample size.³² Across all studies, there were a total of 443 patients with TBI who underwent CBF monitoring by a thermal diffusion technique. The mean age ranged from 21 to 52 years, but 11 studies did not report any information related to patient age. Details of concurrent neurocritical care interventions were not described in most article. Tables 1 through 3 detail the study populations and design for each included article.

Table 1.

Study Characteristics and Patient Demographics

Reference	Article location	Study type	Number of patients	Patient characteristics	Mean age (years)	Device type	Primary and secondary goals of study
Abdullah et al.	Manuscript	Prospective observational	17	Severe TBI undergoing decompressive craniectomy	21.35 (18–27)	Cortical	Primary: determine the effect of decompressive craniectomy on CBF, cerebral compliance, cerebral autoregulation, pressure-volume index, jugular oximetry, and PbtO₂
Brennan et al.	Meeting abstract	Prospective observational	6	Severe TBI	Unknown	Cortical	Primary: determine the safety and feasibility of continuous CBF monitoring by thermal diffusion monitor, and efficacy of a CBF-targeted approach Secondary: determine the efficacy of a CBF-based approach in terms of preventing malignant cerebral edema
Carter et al.	Meeting abstract	Prospective observational	12	Severe TBI	Unknown	Cortical	Primary: determine the safety and feasibility of monitoring CBF continuously in patients with severe head trauma
Dias et al. (2014)	Manuscript	Prospective observational	18	Severe TBI	40 (21–64)	Parenchymal	Primary: determine the effects of hypertonic saline boluses on CBF, ICP, PbtO₂, and cerebrovascular reactivity (PRx, CBFx) in patients with severe TBI Secondary: determine the individual variations in ETCO₂ and sodium serum levels for each hypertonic saline bolus administered
Dias et al. (2015)	Manuscript	Retrospective	18	Severe TBI	42 (SD 16)	Parenchymal	Primary: assess staff compliance and outcome impact of autoregulation-guided treatment in patients with severe TBI, based on continuous evaluation of PRx Secondary: compare CPPopt values obtained with PRx with CPPopt estimates obtained with other cerebrovascular reactivity indices (CBFx-CPPopt, COx-CPPopt, ORxs-CPPopt)
Dias et al. (2016)	Manuscript	Prospective observational	18	Severe TBI	42 (SD 15)	Parenchymal	Primary: describe the relationship between plateau waves of ICP, and CBF, CPP, PbtO₂, CO, and cerebrovascular reactivity indices
Dickman et al.	Manuscript	Prospective observational	12	Moderate and severe TBI	31 (7–65)	Cortical	Primary: determine the relationship between rCBF measured continuously, ICP, and clinical examination, and determine the link with patient outcome
Garcia-Perez et al.	Meeting abstract	Prospective observational	5	Severe TBI	Unknown	Parenchymal	Primary: determine the correlation between continuously monitored CBF with ICP, PbtO₂, CPP, MAP, central temperature, and BIS in severe TBI, and determine the link with patient outcome
Gopinath et al.	Manuscript	Prospective observational	35	Severe TBI	33 (SD 2.3)	Cortical	Primary: determine the clinical usefulness of the TD-rCBF method and the relationship between rCBF and global CBF measured via Kety-Schmidt technique
Hinzman et al.	Manuscript	Prospective observational	24	Severe TBI undergoing craniotomy	48 (23–79)	Parenchymal	Primary: evaluate neurovascular coupling to CSD after craniotomy for severe TBI using rCBF monitoring and assess the link with patient outcome
Hirsch et al.	Meeting abstract	Prospective observational	17	TBI (unspecified)	Unknown	Parenchymal	Primary: determine the relationship between cerebral venous-to-arterial CO₂ gradient, CBF, brain tissue oxygenation, and patient outcome
Jaeger et al.	Manuscript	Prospective observational	3	Severe TBI	Unknown	Parenchymal	Primary: examine the relationship between continuously monitored rCBF and PbtO₂
Krasberg et al.	Meeting abstract	Prospective observational	>50	Severe TBI	Unknown	Parenchymal	Primary: assess the feasibility and clinical usefulness of continuous rCBF monitoring and its correlation with CPP and ICP
Lee et al.	Manuscript	Prospective observational	20	Moderate and Severe TBI	35.7 (17–61)	Cortical	Primary: examine the correlation between rCBF, ICP, and GCS on admission, and outcome in response to various therapeutic interventions
Oshorov et al.	Meeting abstract	Prospective observational	5	Severe TBI	Unknown	Parenchymal	Primary: determine the effects of hyperthermia on MAP, CPP, ICP, PRx, CBF, and CBFx
Phan et al.	Meeting abstract	Prospective observational	>20	Severe TBI	Unknown	Parenchymal	Primary: determine the clinical usefulness of TD-CBF within a multi-modal monitoring regimen Secondary: evaluate the effects of arterial and cerebral venous pressure changes on CBF
Rosenthal et al. (2011)	Manuscript	Prospective observational	20	Severe TBI	37 (15–65)	Parenchymal	Primary: determine the safety and feasibility of using intra-parenchymal rCBF probe as part of a monitoring regimen Secondary: assess the rCBF response to hyperventilation and MAP challenges to gain information on CVR and autoregulatory status
Rosenthal et al. (2008)	Manuscript	Prospective observational	14	Severe TBI	36.8 (15–65)	Parenchymal	Primary: examine the relationship between PbtO₂, local CBF, and measures of cerebral oxygen diffusion and delivery
Schröder et al.	Manuscript	Prospective observational	13	Moderate and severe TBI	27 (median; 15-82)	Cortical	Primary: determine the correlation between continuous monitoring of CBF with a TD probe and sXe-CT CBF studies
Sena et al.	Meeting abstract	Retrospective	52	TBI (unspecified)	Unknown	Parenchymal	Primary: investigate relationship between ICP and focal CBF
Sioutos et al.	Manuscript	Prospective observational	56	Severe TBI	Unknown	Cortical	Primary: examine the relationship between rCBF, initial GCS, and ICP, and the link with patient outcome
Soukup et al.	Manuscript	Prospective observational	10	Severe TBI orICH	52 (SD 15.7)	Parenchymal	Primary: investigate the clinical feasibility and reliability of continuous monitoring of rCBF during mild hyperventilation using a thermodiffusion probe
Tackla et al.	Manuscript	Prospective observational	7	Severe TBI undergoing neurosurgical procedure	41 (27–61)	Parenchymal	Primary: assess cerebral autoregulatory status using rCBF in surgically managed severe TBI patients
Tenjin et al.	Manuscript	Prospective observational	15	Severe TBI treated surgically	46 (6–87)	Cortical	Primary: assess CO₂ reactivity in severe head-injury patients using thermal diffusion CBF measurements
Vajkoczy et al.	Manuscript	Prospective observational	16	Severe TBI and SAH	Unknown	Parenchymal	Primary: clinically validate a new TD-CBF microprobe via comparison with sXE-CT technique for rCBF measurement

BIS, bispectral index; CBF, cerebral blood flow; CBFx, cerebral blood flow index; CBFx-CPPopt, cerebral-blood-flow-index-based optimal cerebral perfusion pressure; CO, cerebral oximetry; CO₂, carbon dioxide; COx-CPPopt, cerebral-oximetry-based optimal cerebral perfusion pressure; CPP, cerebral perfusion pressure; CPPopt, optimal cerebral perfusion pressure; CSD, cortical spreading depression; ETCO₂, end-tidal carbon dioxide; GCS, Glasgow Coma Scale; ICH, intracerebral hemorrhage; ICP, intracranial pressure; MAP, mean arterial pressure; ORx-CPP, brain-tissue-oxygen-based optimal cerebral perfusion pressure; PbtO₂, brain tissue oxygen tension; PRx, pressure reactivity index; rCBF, regional cerebral blood flow; SAH, subarachnoid hemorrhage; sXE-CT, stable xenon-enhanced computed tomography; sXE-CT CBF, cerebral blood flow measured by sXE-CT; TBI, traumatic brain injury; TD, thermal diffusion; TD-CBF, cerebral blood flow measured by thermal diffusion; TD-rCBF, regional cerebral blood flow measured by thermal diffusion.

Thermal diffusion flowmetry technique

The type of thermal diffusion device used was parenchymal in 16 studies^{11
–13,16–20,23,24,26,30

–34} and cortical for the remaining 9 studies.^{10,14,15,21,22,25,27
–29} CBF measurements are reported in terms of mL/100 g/min. The measurement period varies between studies and patients, but was usually initiated within hours of admission to an intensive care unit and lasted for up to 10 days after admission. Placement was achieved via a bedside cranial access kit burr hole, via a small dedicated craniotomy, or at time of craniotomy or craniectomy for a TBI-related indication. Most studies targeted relatively normal appearing cortex or white matter in the frontal region, but a few studies specifically targeted perilesional parenchyma to monitor tissue at high risk for secondary injury.^13,16,25 In 2 studies, the authors specifically targeted the least-injured hemisphere,^19,20 but the rationale is not discussed. Most studies described confirming probe placement by routine post-insertion CT scan of the head. Although the commercially available TDF probes are said to monitor blood flow “continuously,” they necessitate frequent recalibration periods (every 30 min to 2 h, a calibration that usually takes 2–5 min) during which recordings are interrupted and this must be kept in mind when interpreting the data. Currently available methods also suffer from lapses in conduction in high pulsatility environments or when the intracranial temperature exceeds 39.5°C, further interrupting the continuity of measurements. Ranges of CBF considered normal also varied between studies, but generally hovered around a lower threshold of 20–40 mL/100 g/min and upper threshold of 70–100 mL/100 g/min depending on whether the technique was cortical- or parenchymal-based. Methods used to increase CBF included expanding plasma volume, inotropic support, and osmotic agents. Interventions used to lower CBF included hyperventilation and barbiturate coma.

Functional outcome studies

Functional outcome was most commonly reported using the GOS or GOS-E at 3 months, 6 months, or at an unspecified interval.^{11,15,16,19,22,24,25,27,30} That said, 2 studies only report mortality at an unspecified interval^10,14 and 15 studies do not report any functional outcome measure in relation to CBF parameters.^{11,13,15,17

–21,23,26,28,33,34} Table 2 summarizes the studies on TDF and association with global patient outcome.

Table 2.

TDF-rCBF and Patient Functional Outcomes

Reference	Functional outcome measures	Patient outcome	Complications	Limitations	Conclusions
Abdullah et al.	Mortality (unspecified interval)	No significant difference in the proportion of patients with normal CBF in those who survived post-DC and those who had a fatal outcome	Unspecified	Small sample size	The proportion of patients with normal CBF was not significantly greater in patients who survived post-DC
Carter et al.	Dichotomized into good and poor outcomes (unspecified)	Patients with hyperemia or low flow (level unspecified) did “poorly”	Two patients developed a CSF leak around probe, no infections or major complications	Small sample size	Hyperemia and low CBF were associated with a poor outcome
Dias et al. (2015)	Dichotomized GOS at 6 months (Good: ≥3; Poor: <3)	Patients with overall preserved autoregulation (PRx <0.25) had better outcome Larger discrepancy between real CPP and CPPopt was associated with unfavorable outcome	Unspecified	Small sample size	Preserved autoregulation was associated with better outcome
Dickman et al.	Mortality (unspecified interval) and unspecified outcome scale at unspecified time interval	A sustained reduction in rCBF preceded the development of brain death in 6 patients. Five patients with sustained hyperemia progressed to brain death. One patient with sustained reduced CBF associated with ACA and MCA vasospasm on angiography developed a large MCA infarct. This patient had an outcome of vegetative state.	No complications	Small sample size. Statistical analyses not performed or not reported.	Sustained reduced CBF or hyperemia was associated with poor outcome
Garcia-Perez et al.	GOS at 3 months	Three patients had a baseline CBF >20 mL/100 g/min, which was associated with GOS of 4-5 at 3 months, whereas two patients with initial CBF <20 mL/100 g/min had a GOS of 1-3 at 3 months. One patient with initial CBF <5mL/100 g/min progressed to brain death 36 h after admission.	Unspecified	Small sample size	Baseline CBF >20 mL/100 g/min was generally associated with a favorable outcome, whereas a CBF <20 mL/100 g/min was associated with a poor outcome
Hinzman et al.	Dichotomized GOS-E at 6 months (Good = 5 to 8; Poor = 1 to 4)	Average rCBF was significantly higher in patients with a good outcome (46.8 ± 6.5 mL/100 g/min) than those with a poor outcome (32.2 ± 3.7 mL/100 g/min; t(22) = 2.08, p = 0.0487). Patients with a poor outcome suffered longer total times of ischemia (<18 mL/100 g/min) compared with those with a good outcome (23.2 ± 6.5 h vs. 5.3 ± 2.4 h; t(22) = 1.90, p = 0.071).	No hemorrhagic or infectious complications	Substantial amount of data was lost by placing flow probe near thermally significant vessels or tissues with unstable thermal conductivity	Patients with a good outcome had significantly higher regional CBF than those with a poor outcome. Patients with a poor outcome trended toward longer total ischemia time.
Hirsch et al.	Dichotomized into good (discharge to home or rehabilitation facility) and poor (death or discharge to a long-term care facility) outcomes	CBF was significantly lower in patients with a good outcome (17 ± 13 cc/100 g/min) vs. poor outcome (29 ± 14 cc/100 g/min, p < 0.05)	Unspecified	Small sample size	Patients with a good outcome had significantly lower CBF
Lee et al.	Dichotomized GOS at 6 months (Good = 4 or 5; Poor = 1 to 3)	Patients with good outcome had a mean initial rCBF of 55 mL/100 g/min (range 44–64) during the first hour post-operatively, whereas patients with poor outcome had mean rCBF of 48 mL/100 g/min (range 23–110). Patients with persistently low rCBF (<20 mL/100 g/min) (n = 3) suffered brainstem failure within 2 weeks. Three patients with a hyperemic response (peak CBF >100 mL/100 g/min) had a poor outcome (GOS 1-2).	No complications	Small sample size. Statistical comparisons not performed between groups.	Patients with persistently low rCBF or hyperemic rCBF patterns had a poor outcome
Sioutos et al.	Dichotomized GOS at 6 months (Good = 4 or 5; Poor = 1 or 2)	Baseline mean rCBF was similar in patients with good outcome and those with poor outcome (38.6 and 37.8 mL/100 g/min respectively, p > 0.05). The change in mean rCBF (final minus initial) was significantly greater in patients with good outcome (increase, +19.8 mL/100 g/min) than in patients with poor outcome (decrease, −7.18 mL/100 g/min, p < 0.05).	Three patients developed an infection at the scalp entry site, cured with antibiotic therapy; no other complications	Small sample size	Patients with good outcome had a significant increase in rCBF from baseline when compared with patients with poor outcome, in whom mean rCBF decreased from baseline
Tenjin et al.	Dichotomized GOS at 6 months (Good = 4 or 5; Poor = 1 to 3)	The good outcome group showed a CO₂ reactivity index (%ΔCBF/ΔP_aCO₂) of 3.3 ± 1.7, whereas the poor outcome one showed a CO₂reactivity index of 1.3 ± 1.7 (p < 0.05)	Unspecified	Small sample size	Disturbance of CO₂ reactivity measured using TDF correlated with poor outcome

ACA, anterior cerebral artery; CBF, cerebral blood flow; CO₂, carbon dioxide; CPP, cerebral perfusion pressure; CPPopt, optimal cerebral perfusion pressure; CSF, cerebrospinal fluid; DC, decompressive craniectomy; GOS, Glasgow Outcome Scale; GOS-E, extended Glasgow Outcome Scale; MCA, middle cerebral artery; P_aCO₂, arterial partial pressure of carbon dioxide; PRx, pressure reactivity index; rCBF, regional cerebral blood flow; TDF, thermal diffusion flowmetry.

An association between a sustained reduction in CBF below 20 mL/100 g/min and increased mortality was seen across a number of studies.^14,22,29,30 Conversely, hyperemia—defined as CBF greater than 80–100 mL/100 g/min—was generally associated with a poor outcome.^14,27,29 For patients with regional (r)CBF values predominantly falling within the normal range, 3 studies saw higher average CBF in the favorable outcome group (46.8 vs. 32.2 mL/100 g/min¹⁶; 55 vs. 48 mL/100 g/min^22,27). One study reported significantly lower CBF in patients with a good outcome compared with poor outcome (17 vs. 29 mL/100 g/min),³¹ but the authors do not describe the functional outcome evaluation that was used. One study did not observe a difference between the proportion of patients with normal CBF in those who survived after a decompressive craniectomy for refractory intracranial hypertension and those who had a fatal outcome.¹⁰ In one patient, a consistent reduction in rCBF correlated with severe ipsilateral anterior cerebral artery (ACA) and middle cerebral artery (MCA) vasospasm leading to a hemispheric infarct and a vegetative outcome.¹⁴

Neurophysiological outcome studies

Various measured neurophysiological variables were correlated with continuously measured CBF via TDF, including: intracranial pressure (ICP), cerebral perfusion pressure (CPP), PbtO₂, and cerebrovascular reactivity. Table 3 details the studies assessing the association between TDF-based CBF and neurophysiological measures.

Table 3.

Association between TDF-rCBF and Other Neurophysiological Measures

Reference	Other physiologic parameters measured	Physiologic outcomes	Conclusions
Abdullah et al.	Autoregulation (unspecified index) CI PbtO₂ PVI SjVO₂	Value for each parameter dichotomized into normal and abnormal. Proportion of patients with normal CBF, SjVO₂, PbtO₂, and autoregulation not significantly different pre- and post-DC. Proportion of patients with normal CI and PVI significantly greater post-DC (p < 0.05)	Compliance and PVI improved significantly post-DC in surviving patients
Carter et al.	ICP	Two patients had hyperemic flow preceding development of elevated ICP.	Hyperemia preceded intracranial hypertension in two patients.
Dias et al. (2014)	ICP CBFx CPP CVR ETCO₂ PbtO₂ PRx	ICP and CPP improved over time after hypertonic saline bolus (p < 0.001). This was preceded by an increase in mean CBF of 7.8 mL/min/100 g (p < 0.001) and a decrease in mean CVR of 0.4 mm Hg^min^100 g/mL (p = 0.01). No significant change in PbtO₂ and in ETCO₂ over time after hypertonic saline bolus. Autoregulation as measured by PRx and CBFx showed a slight improvement on average after hypertonic saline bolus and followed a linear decreasing model (p-values for slopes 0.01 and 0.04, respectively).	Recovery of CBF appeared before a rise in CPP and a decrease in ICP after hypertonic saline boluses for intracranial hypertension.
Dias et al. (2015)	CBFx-CPPopt COx-CPPopt ORxs-CPPopt PRx-CPPopt	Comparison between PRx-CPPopt and CPPopt values obtained from other reactivity indices, including CBFx-CPPopt, revealed similar limits of precision	Limits of precision of CPPopt derived from PRx, CBFx, COx, and ORxs were similar
Dias et al. (2016)	CBFx CO CPP CVR ICP ORxs PbtO₂ PRx PAx	At the peak of an ICP plateau wave, there was a significant increase in mean ICP pulse amplitude (p = 0); a significant decrease in mean CPP (p < 0.0001), brain oxygenation (p = 0.0004), and CO (p = 0.021); and a non-significant trend toward reduced CBF and CVR. Plateau waves were associated with a disruption in autoregulatory status as measured by PRx (p = 0.02) and ORxs (p = 0.04). After plateau waves, a majority of patients showed increased CBF (p = 0.03) and brain oxygenation (p = 0.009) above baseline	Plateau waves are associated with a decrease in CPP and brain oxygenation, and a disruption in autoregulatory status. A significant proportion of patients experienced hyperemia and increased oxygenation after the end of an ICP plateau wave.
Dickman et al.	ICP	Reduced CBF was often inversely proportional to increased ICP. Several patients with diffuse brain injury had elevated ICPs, and hyperemic CBF patterns.	High ICPs were associated with both reduced CBF and hyperemia.
Garcia-Perez et al.	BIS CPP ICP MAP PbtO₂	CBF showed a statistically significant positive correlation with ICP (r = 0.370, p < 0.01), but a poor correlation with PbtO₂, MAP, BIS. and CPP	ICP positively correlated with CBF measured by TDF
Gopinath et al.	Global CBF (Kety-Schmidt technique)	The average global CBF value was 50.5 ± 0.9 mL/100 g/min vs. an average of 60.5 ± 1.4mL/100g/min for TD-CBF. The overall slope of the regression between the global CBF and TD-CBF measurements was 0.618.	Changes in regional CBF measured by TDF generally correlated with similar changes in global CBF measured via Kety-Schmidt technique
Hinzman et al.	ECoG PbtO₂ rCBFx	Inverse neurovascular coupling (transient hypoperfusion) was observed for 74 CSDs in 5 patients, whereas physiological coupling (hyperemia) was observed for 34 CSDs in 3 patients. Physiological hemodynamic response to CSD only occurred in the presence of intact autoregulation (measured by rCBFx).	Loss of cerebral autoregulation was associated with an inverse hemodynamic response during CSD
Hirsch et al.	PbtO₂ v-a(CO₂)	Patients with good outcome had significantly higher mean v-a(CO₂) when compared with patients with poor outcome (9.00 ± 6.57 mm Hg vs. 5.75 ± 4.00 mm Hg, p < 0.03). This was associated with lower CBF in patients with good outcome (17 ± 13 cc/100 g/min vs. 29 ± 14 cc/100 g/min, p < 0.05).	Patients with good outcome had significantly higher mean v-a(CO₂) and lower CBF when compared to patients with poor outcome
Jaeger et al.	CPP ICP PbtO₂	There was a significant overall correlation between rCBF and PbtO₂ (r = 0.36). Mean overall rCBF was 42.6 ± 19.2 cc/100 g/min. Mean ICP was 13.9 ± 5.1 mm Hg and mean CPP was 78.8 ± 8.9 mm Hg.	Changes in PbtO₂ were positively correlated with simultaneous changes in rCBF
Krasberg et al.	Autoregulatory status (unspecified measure) CPP	In patients with less severe TBI, CBF did not fluctuate significantly with changes in perfusion pressure More severely injured patients demonstrated changes in CBF that correlate closely with instantaneous changes in perfusion pressure suggesting loss of autoregulation	CBF did not fluctuate significantly with changes in perfusion pressure in patients with less severe injury, suggesting intact autoregulation More severely injured patients demonstrated changes in CBF that correlate closely with instantaneous changes of perfusion pressure suggesting loss of autoregulation
Lee et al.	ICP	Patients with extremely high ICP refractory to treatment had a rCBF persistently below 20 mL/100 g/min. Two of three patients with hyperemic responses showed persistent high ICPs. Patients with moderately low to normal CBF usually had ICP within the normal range.	Persistently very low rCBF and hyperemia were associated with high ICPs, whereas low to normal rCBFs were associated with normal ICPs.
Oshorov et al.	CBFx CPP ICP MAP PRx	There was an increase in CBF in 55.5% of 18 hyperthermia episodes, whereas it remained constant in 5.5% and decreased in 39% of cases. Hyperthermia episodes were associated with an increase in ICP in 89%. CPP increased in 27.5% of hyperthermia episodes, remained the same in 17%, and decreased in 55.5%. MAP increased in 61% of hyperthermia episodes. Of cases, 55.5% showed associated autoregulatory failure as measured by PRx; CBFx impairment was observed in 28% of cases.	Hyperthermia episodes are associated with increased ICP and possibly autoregulatory failure
Phan et al.	CPP CVR	In a first group of patients, the average change in CVR during a 10 mm Hg MAP challenge using phenylephrine was –0.17 ± 3.9 mm Hg/cc/100 g/min. In a second group of patients, mean increase in CPP during a MAP challenge was 19.6 ± 9.4 mm Hg and this was associated with a 112 ± 30% average increase in CVR. During venous challenges by CSF drainage via an external ventricular drain in the same second patient group, CPP increased by 7.2 ± 4.5 mm Hg and this resulted in a significant decrease in CVR to 63 ± 19% of baseline (p < 0.01).	There was a significant decrease in CVR (increase in CBF) when lowering cerebral venous pressure by CSF drainage compared with increasing MAP.
Rosenthal et al. (2011)	Cerebral CO₂ reactivity CPP CVR ICP	Mean change in ICP was +1 ± 2 mm Hg (not significant) during a 10 mm Hg MAP challenge by phenylephrine drip and –3 ± 4 mm Hg (p < 0.0003) during an hyperventilation challenge. Mean CBF decreased substantially with hyperventilation (−7 ± 8 mL/100 g/min, p = 0.0002) and non-significantly increased during MAP challenges (+1 ± 7 mL/100 g/min, p = 0.17). CVR increased in 83% of patients during a hyperventilation challenge (mean change of +3.5 ± 3.8 mm Hg/mL/100 g/min, p = 0.0002), whereas CVR response to MAP challenge was variable with no significant trend.	Hyperventilation was associated with a mean reduction in CBF and ICP. There were no significant changes in CBF, CVR, or ICP during MAP challenges. CVR increased in a majority of patients during a hyperventilation challenge suggesting preserved cerebral CO₂ reactivity.
Rosenthal et al. (2008)	AVDO₂ PbtO₂	Multi-variate analysis showed a strong association between brain tissue oxygen tension and the product of CBF and AVDO₂ (r = 0.89 from Michaelis-Menten curve, p < 0.001)	A strong association between PbtO₂ and the product of CBF and AVDO₂ was observed suggesting a relationship between PbtO₂ and diffusion of dissolved plasma oxygen across the blood–brain barrier
Schröder et al.	sXe-CT CBF	The sXe-CT CBF data from the 2-cm and 3- to 5-mm regions of interest adjacent to the TDF probe showed no significant correlation with TD-CBF probe measurements (r = 0.57, p = 0.17, and r = 0.53, p = 0.44, respectively)	Continuous CBF measured by TDF did not correlate significantly with regional or global sXe-CT CBF values
Sena et al.	ICP	Slopes of best-fit lines for ΔICP against Δfocal-CBF across the study population ranged from 0 to 0.62 (mean = 0.13, SD = 0.13). Individual correlation coefficients describing ΔICP against Δfocal-CBF ranged from 0.01 to 0.27 (mean = 0.12, SD = 0.07) with each respective p-value <0.01.	ΔICP positively correlated to Δfocal-CBF in some of the TBI patients in the study but the ΔICP-to-Δfocal-CBF relationship is not well characterized by a linear model
Sioutos et al.	ICP	Patients with increased ICP had either hyperemic or ischemic blood flow patterns.	Patients with increased ICP demonstrated either a hyperemic or an ischemic rCBF pattern
Soukup et al.	PbtO₂	Mild controlled hyperventilation resulted in a rCBF reduction from 30 ± 3 mL/100 g/min to 25 ± 2.4 mL/100 g/min (−17%, p < 0.05) on the first day of admission and from 31 ± 3.6 mL/100 g/min to 22 ± 4.9 mL/100 g/min (−29%, p < 0.05) on the second day of admission. Of note, the baseline parameters P_aCO₂ and rCBF did not differ significantly on the two days. These decreases in rCBF under controlled hyperventilation were associated with a reduction of regional PbtO₂ from 20 ± 2.9 mm Hg to 15 ± 4 mm Hg (−25%, p < 0.05) on the first day of admission and from 20 ± 3.1 mm Hg to 14 ± 1.5 mm Hg (−30%, p < 0.05) on the second day of admission.	Both rCBF and PbtO₂ were significantly reduced by mild controlled hyperventilation
Tackla et al.	Autoregulation (rCBFx) CPP	All rCBF values fell below the ischemic threshold (18 mL/100 g/min) when CPPs were <50 mm Hg compared with 11% ischemia when CPPs were >50 mm Hg (p < 0.05). Percent time when autoregulation was intact and rCBF exceeded the ischemic threshold was maximized in the 75-80 mm Hg CPP range, termed CPPideal, for the overall patient population, but individual CPPopt ranged from 60 to 100 mm Hg.	CPP lower than 50 mm Hg was associated with ischemia. Autoregulatory status and CBF were optimized at CPPs between 75 and 80 mm Hg for the overall patient population but there was considerable variability between individual patients.
Tenjin et al.	CO₂ reactivity (%ΔCBF/ΔP_aCO₂)	Patients with preserved CO₂ reactivity responded to hyperventilation with a reduction in ICP	Patients with preserved CO₂ reactivity responded to hyperventilation with a reduction in ICP
Vajkoczy et al.	sXe-CT rCBF	A comparison of TD-rCBF and sXe-CT rCBF revealed a good correlation (r = 0.89, p < 0.0001) and a mean difference of 1.1 ± 5.2 mL/100 g/min between the two techniques	rCBF measured by TD correlated strongly with sXe-CT rCBF

AVDO₂, arterio-venous difference in oxygen; BIS, bispectral index; CBF, cerebral blood flow; CBFx, cerebral blood flow index; CBFx-CPPopt, cerebral-blood-flow-index-based optimal cerebral perfusion pressure; CI, cerebral compliance; CO, cerebral oximetry; CO₂, carbon dioxide; COx-CPPopt, cerebral-oximetry-based optimal cerebral perfusion pressure; CPP, cerebral perfusion pressure; CPPopt, optimal cerebral perfusion pressure; CSD, cortical spreading depression; CVR, cerebrovascular resistance; DC, decompressive craneictomy; ECoG, electrocorticography; ETCO₂, end-tidal carbon dioxide; ICP, intracranial pressure; MAP, mean arterial pressure; ORxs, short oxygen reactivity index; ORxs-CPP, short-oxygen-reactivity-index-based optimal cerebral perfusion pressure; P_aCO₂, arterial partial pressure of carbon dioxide; PAx, pulse amplitude index; PbtO₂, brain tissue oxygen tension; PRx, pressure reactivity index; PRx-CPPopt, pressure-reactivity-index-based optimal cerebral perfusion pressure; PVI, pressure-volume index; rCBF, regional cerebral blood flow; SjVO₂, jugular venous oxygen saturation; sXE-CT, stable xenon-enhanced computed tomography; sXE-CT CBF, cerebral blood flow measured by sXE-CT method; sXE-CT rCBF, regional cerebral blood flow measured by sXE-CT method; TD, thermal diffusion; TD-CBF, cerebral blood flow measured by thermal diffusion; v-a(CO₂), venous-arterial difference in carbon dioxide.

TDF-CBF association with ICP

A number of studies found an association between high ICPs and either significantly reduced CBF or hyperemia.^14,22,24,29 Five studies describe a relationship between high ICPs and hyperemic CBF patterns.^{11,14,22,27,29} In those patients, hyperemia tended to precede the development of increased ICP by a few hours.^12,14 One study reported a positive correlation between CBF and ICP (r = 0.370, p < 0.01) even within the normal range.³⁰ A negative association between CBF and ICP was seen in 3 studies.^13,14,22 One study reported a negative association across the entire range of CBF values, whereas in the other 3 studies the inverse correlation was seen at CBF values below the ischemic threshold.^14,22,27 In one study, improvement of ICP after hypertonic saline (HTS) bolus was preceded by a mean increase in CBF.¹³

One study reported a positive correlation between delta-focal-ICP and delta fCBF within individual patients, but no clear correlation between the two parameters when looking at the overall study population.³⁴

TDF-CBF association CPP

One study describes optimal CBF patterns when CPP was in the 75–80 mm Hg range and CBF was consistently below the ischemic threshold when CPP was less than 50 mm Hg.²⁴ Another study saw no significant fluctuation in CBF with changes in CPP in less severely injured patients, although more severely injured patients demonstrated changes in CBF that correlated closely with instantaneous changes in perfusion pressure, suggesting disturbances in autoregulatory mechanisms.¹⁸

TDF-CBF association with PbtO₂

Of the studies that explored the relationship between CBF and PbtO₂, most found a positive correlation between the two variables.^17,19,23 A reduction in CBF induced by hyperventilation in one study was associated with a concomitant reduction in regional PbtO₂.²³ Multi-variate analysis in another study showed a strong association between brain tissue oxygen tension and the product of CBF and arteriojugular venous difference of oxygen (AVDO₂; p < 0.001).¹⁹ No studies reported a negative association between CBF and PbtO₂. A poor overall correlation between CBF and PbtO₂ is reported by Garcia-Perez and colleagues,³⁰ but their study lacked statistical power with only 5 subjects.

TDF-CBF association with cerebrovascular reactivity

In one study looking at a CBF-based vascular reactivity index (CBFx), the percentage of time when vascular reactivity was preserved and CBF exceeded the ischemic threshold was maximized in the 75–80 mm Hg CPP range for the overall patient population, with individual optimal CPP (CPPopt) ranging from 60 to 100 mm Hg.²⁴ As described above, Krasberg and associates¹⁸ observed a close relationship between changes in CBF and instantaneous changes in CPP only in the most severely injured members of their cohort, presumably because of a failure of autoregulation. Lastly, vascular reactivity as measured by pressure reactivity index (PRx) and CBFx showed a slight improvement on average after a hypertonic saline bolus and this was associated with a mean increase in CBF.¹³

TDF-CBF association with other CBF measures

Changes in regional CBF measured by TDF correlated with changes in global CBF determined via the Kety-Schmidt technique in one study (regression slope = 0.636).¹⁵ One study comparing CBF values obtained via TDF and stable Xe-rCBF (sXe-rCBF) found a good correlation between the two techniques (r = 0.89, p < 0.0001),²⁶ whereas a similar study found no significant correlation between TDF measurements and sXe-CT CBF data from 2-cm and 3- to 5-mm regions of interests adjacent to the TD probe.²¹

Complications of TDF

Eleven studies totalling 214 patients commented on complications observed during TDF monitoring.^{14
–16,19,21,22,24,26

–29} Of those, 7 reported no probe-related complications.^{14
–16,19,24,26
–28} The overall complication rate—excluding studies that failed to disclose adverse events—was 3.7% (8/214), most of which were of infectious nature. Four patients developed a cerebrospinal fluid (CSF) leak around the probe^21,29; which led to meningitis in one case²¹; one patient developed an intracerebral abscess requiring surgical management²¹; and three patients developed a superficial soft-tissue infection at the scalp exit site.²² There were no clinically significant intracranial hemorrhages related to probe insertion.

Risk of bias

Risk of bias in individual studies was assessed via the RTI item bank. Overall, all studies were deemed at high risk for bias given their observational nature, relatively small size, and lack of reporting regarding concurrent interventions. A table summarizing individual risks of bias for different bias domains derived from the RTI questionnaire can be found in Supplementary Appendix C. Meeting abstracts in particular were considered to have very high risk of bias for most of these domains.

Discussion

This scoping systematic review of cerebral TDF summarized all the major available studies documenting the relationship between TDF-rCBF measurements, patient functional outcome (11 studies), and neurophysiological parameters (25 studies). Although the relatively small scale and heterogeneity of the included studies make it difficult to draw robust conclusions, a few consistent findings emerged through our synthesis of the evidence. With respect to TDF-rCBF and functional outcome, a sustained reduction in rCBF and hyperemia were both consistently associated with a poor outcome. Similarly, very low initial rCBF was predictive of catastrophic neurological outcome or progression to brain death. On the other hand, the relationship between TDF-rCBF and functional outcome remains unclear for patients who tended to stay within the normal perfusion range. Importantly, more than half of the studies included did not comment on the link between rCBF and functional outcome. Of those who did, intervals for outcome measurement and type of outcome scale used were very inconsistent between studies, highlighting the need for more concerted research efforts to elucidate the predictive value of this technique. As for relationships between TDF-rCBF and other neurophysiological parameters, very low or very high rCBF dependably correlated with intracranial hypertension. A positive correlation between changes in rCBF and PbtO₂ was also observed across a number of studies. This is in line with findings from an Xe-CT CBF study showing a significant correlation between regional blood flow and brain tissue oxygenation.³⁵ Unfortunately, there are too few studies reporting on associations with other neurophysiological variables such as autoregulation indices, cerebral microdialysis analytes, and global CBF measures to draw meaningful conclusions in these domains. The technique does appear to be well tolerated with an overall complication rate of 3.7%, which is comparable to other invasive neuromonitoring techniques.^36,37

An important factor identified from the available literature is that TDF-based continuous CBF monitoring has really only been employed in the experimental setting and has yet to receive widespread adoption. Prior issues with the technique have been identified, including the need for repeated zeroing of the device when in situ.³⁸ Further, the invasive nature of both parenchymal and subdural probes may dissuade clinicians from placement in patients already requiring multiple invasive monitoring devices. In general, this scoping review confirms that much more prospective work is needed to determine the exact role—if any—that TDF devices should play in the multi-modal monitoring of critically ill patients.

First steps for future work should focus on better characterizing the association between continuously measured CBF from TDF methods, continuous brain tissue oxygenation, and validated measures of cerebrovascular reactivity. Such analysis should focus on high-frequency time-series relationships to determine the degree of correlation between these three variables under high temporal resolution windows. In addition, such relationships should be interrogated through dynamic physiological challenges (through CO₂, mean arterial pressure [MAP], or ICP manipulation). These type of analyses would help determine if any complementary information is added by directly monitoring CBF, as it is currently unclear how much explained variance in prediction models is added when introducing additional CBF measures on top of CPP, which is the surrogate used by most centers today.

From here, if such analysis confirms added information from CBF monitoring, larger multi-center work should be conducted to assess the added value of such monitoring in terms of both coarse global outcome prediction and more detailed neuropsychiatric/functional outcome measures at various time-points post-injury, including longer-term assessments up to and beyond 12 months. Such outcome analysis should also include acute and long-term high-resolution MRI to assess the association between changes in TDF-based CBF and lesion progression as well as long-term cortical atrophy and connectivity via diffusion tractography. Surrogate markers of tissue outcome, such as protein markers of brain tissue fate, could also be used to see if the extent of affected tissue is associated with abnormal rCBF as measured with TDF.³⁹ It is through such an integrated approach that we should be able to begin to understand the role of such monitoring in adult TBI care.

Limitations

This review suffers from a number of limitations that primarily arise from heterogeneity in outcome reporting and high risk of bias in the included studies. All of the studies are relatively small, of observational nature, and often had a primary focus that was not well aligned to our primary and secondary research questions. A quantitative synthesis of the evidence via a meta-analysis was also not possible given the lack of common outcome measures.

Further, the majority of studies do not adequately report on possible confounding effects from concurrent interventions or on the influence of medical and surgical therapies on CBF measurements. Without this information, we could not comment on the impact of CBF-directed therapy on outcome. Of the 9 articles that reported functional outcome measures, only 2 superficially discuss the relationship between CBF response to intervention and patient outcome. Lee and colleagues²⁷ describe how 4 patients with high ICP and ischemia who responded well to osmotherapy were noted to have a good outcome. They also reported that for patients with an initial CBF lower than 20mL/100 g/min, intravascular volume expansion did not improve CBF and these patients universally had a poor outcome. However, the small number of patients and lack of information with respect to what was considered a positive “response” in terms of CBF values makes it difficult to draw meaningful conclusions.

In a pilot study investigating the value of CPPopt-guided therapy, Dias and colleagues¹² reported that patients with larger discrepancy between real CPP and CPPopt were more likely to have an adverse outcome. However, these conclusions are drawn from a PRx-derived CPPopt. Although they describe a reasonably good agreement between PRx-CPPopt and optimal CPP calculated by other cerebrovascular reactivity indices including CBFx-CPPopt, they do not discuss the impact of their interventions on CBF values and the direct relationship between CBFx-CPPopt and GOS.

Another critical limitation is the variability in probe placement between studies, which confounds comparisons of rCBF measurements. rCBF values are likely greatly influenced by positioning of the probe tip in healthy versus lesional or perilesional tissue. Whereas most studies targeted non-lesioned cerebral tissue, 3 studies targeted perilesional parenchyma, which has been shown to have aberrant flow patterns on Xe-CT studies.⁴⁰ There was also an uneven distribution of studies using cortical versus parenchymal-based technique, further limiting cross-study comparisons of absolute rCBF values. Establishing reliable thresholds of abnormal rCBF requiring intervention will benefit from a more systematic framework for probe placement commensurate with the clinical scenario.

Conclusion

Despite being based on a relatively weak body of evidence, the available literature suggests a link between consistently abnormal TDF-rCBF values, intracranial hypertension, and poor functional outcome. TDF-rCBF also appears to correlate well with regional measurements of brain tissue oxygenation. Currently, such monitoring should be considered experimental, requiring much further evaluation prior to widespread adoption.

Footnotes

Acknowledgments

FAZ has received salary support for dedicated research time, during which this project was completed. Such salary support came from: the Cambridge Commonwealth Trust Scholarship and the University of Manitoba Clinician Investigator Program. FAZ's research program is supported through the University of Manitoba Thorlakson Chair in Surgical Research Establishment Fund. EPT received post-doctoral grants from the Swedish Society for Medical Research. FM has received salary support for dedicated research time from the Canada Cambridge Scholarship sponsored by the Cambridge Commonwealth Trust.

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

Supplementary Appendices A–C

References

Bouzat

, Sala

, Payen

J.F.

, and Oddo

(2013). Beyond intracranial pressure: 0ptimization of cerebral blood flow, oxygen, and substrate delivery after traumatic brain injury. Ann. Int. Care, 3, 1–9.

Figaji

, Padayachy

, Fieggen

, and Rohlwink

(2012). Continuous local cerebral blood flow monitoring in children with TBI. Child's Nerv. Syst. 28, 1590.

Frontera

, Ziai

, O'Phelan

, Leroux

P.D.

, Kirkpatrick

P.J.

, Diringer

M.N.

, and Suarez

J.I.

(2015). Regional brain monitoring in the neurocritical care unit. Neurocrit. Care, 22, 348–359.

Miller

, Armonda

, Le Roux

, Menon

D.K.

, Vespa

, Citerio

, Bader

M.K.

, Brophy

G.M.

, Diringer

M.N.

, Stocchetti

, Videtta

, Badjatia

, Boesel

, Chesnut

, Chou

, Claassen

, Czosnyka

, De Georgia

, Figaji

, Fugate

, Helbok

, Horowitz

, Hutchinson

, Kumar

, McNett

, Naidech

, Oddo

, Olson

, O'Phelan

, Provencio

, Puppo

, Riker

, Robertson

, Schmidt

J.M.

, and Taccone

(2014). Monitoring of cerebral blood flow and ischemia in the critically ill. Neurocrit. Care, 21, 121–128.

Steiner

L.A.

, and Czosnyka

(2002). Should we measure cerebral blood flow in head-injured patients?. Br. J. Neurosurg. 16, 429–439.

Le Roux

, Menon

D.K.

, Citerio

, Vespa

, Bader

M.K.

, Brophy

G.M.

, Diringer

M.N.

, Stocchetti

, Videtta

, Armonda

, Badjatia

, Boesel

, Chesnut

, Chou

, Claassen

, Czosnyka

, De Georgia

, Figaji

, Fugate

, Helbok

, Horowitz

, Hutchinson

, Kumar

, McNett

, Miller

, Naidech

, Oddo

, Olson

, O'Phelan

, Provencio

J.J.

, Puppo

, Riker

, Robertson

, Schmidt

, Taccone

, Neurocritical Care Society, and European Society of Intensive Care Medicine. (2014). Consensus summary statement of the International Multidisciplinary Consensus Conference on Multimodality Monitoring in Neurocritical Care: a statement for healthcare professionals from the Neurocritical Care Society and the European Society of Intensive Care Medicine. Intensive Care Med. 40, 1189–1209.

Higgins

J.P.T

.G.S. Cochrane Handbook for Systematic Reviews of

Interventions

, version 5.1.0. http:handbook.cochrane.org (Last accessed October 28, 2018 ).

Moher

, Liberati

, Tetzlaff

, and Altman

D.G.

(2009). Preferred reporting items for systematic reviews and meta-analysis: the PRISMA statement. Ann. Int. Med. 151, 264–269.

Viswanathan

, and Berkman

N.D.

(2011). Development of the RTI Item Bank on Risk of Bias and Precision of Observational Studies. Report no. 11-EHC028-EF. Agency for Healthcare Research and Quality: Rockville, MD.

10.

Abdullah

, Zamzuri

, Awang

, Sayuthi

, Ghani

, Tahir

, and Naing

N.N.

(2005). Preliminary report on spiegelberg pre and post-operative monitoring of severe head-injured patients who received decompressive craniectomy. In: Acta Neurochir. Suppl. 95, 311–314.

11.

Dias

, Maia

, Cerejo

, Smielewski

, Paiva

J.A.

, and Czosnyka

(2016). Plateau waves of intracranial pressure and multimodal brain monitoring. Acta Neurochir. Suppl. 122, 143–146.

12.

Dias

, Silva

M.J.

, Pereira

, Monteiro

, Maia

, Barbosa

, Silva

, Honrado

, Cerejo

, Aries

M.J.

, Smielewski

, Paiva

J.A.

, and Czosnyka

(2015). Optimal cerebral perfusion pressure management at bedside: a single-center pilot study. Neurocrit. Care, 23, 92–102.

13.

Dias

, Silva

M.J.

, Pereira

, Silva

, Cerejo

, Smielewski

, Rocha

A.P.

, Rita Gaio

, Paiva

J.-A.

, and Czosnyka

(2014). Post-traumatic multimodal brain monitoring: response to hypertonic saline. J. Neurotrauma, 31, 1872–1880.

14.

Dickman

C.A.

, Carter

L.P.

, Baldwin

H.Z.

, Harrington

, and Tallman

(1991). Continuous regional cerebral blood flow monitoring in acute craniocerebral trauma. Neurosurgery, 28, 467–472.

15.

Gopinath

S.P.

, Valadka

, Contant

C.F.

, and Robertson

C.S.

(1999). Relationship between global and cortical cerebral blood flow in patients with head injuries. Neurosurgery, 44, 1273–1279.

16.

Hinzman

J.M.

, Andaluz

, Shutter

L.A.

, Okonkwo

D.O.

, Pahl

, Strong

A.J.

, Dreier

J.P.

, and Hartings

J.A.

(2014). Inverse neurovascular coupling to cortical spreading depolarizations in severe brain trauma. Brain, 137, 2960–2972.

17.

Jaeger

, Soehle

, Schuhmann

M.U.

, Winkler

, and Meixensberger

(2005). Correlation of continuously monitored regional cerebral blood flow and brain tissue oxygen. Acta Neurochir. (Wein) 147, 51–56.

18.

Krasberg

, Sena

, Mead

, Pappu

, Nemoto

E.M.

, Olin

, Malkoff

, Stippler

, and Yonas

(2013). Autoregulation of CBF following TBI: an important variable that is now clinically accessible. J. Neurotrauma, 30, A128.

19.

Rosenthal

, Hemphill

J.C.

3 rd

Sorani

, M., Martin

, Morabito

, Obrist

W.D.

, and Manley

G.T.

(2008). Brain tissue oxygen tension is more indicative of oxygen diffusion than oxygen delivery and metabolism in patients with traumatic brain injury. Crit. Care Med. 36, 1917–1924.

20.

Rosenthal

, Sanchez-Mejia

R.O.

, Phan

, Hemphill

J.C.

3 rd

Martin

, C., and Manley

G.T.

(2011). Incorporating a parenchymal thermal diffusion cerebral blood flow probe in bedside assessment of cerebral autoregulation and vasoreactivity in patients with severe traumatic brain injury. J. Neurosurg. 114, 62–70.

21.

Schröder

M.L.

, Muizelaar

J.P.

(1993). Monitoring of regional cerebral blood flow in acute head injury by thermal diffusion. Acta Neurochir. Suppl., 47–49.

22.

Sioutos

P.J.

, Orozco

J.A.

, Carter

L.P.

, Weinand

M.E.

, Hamilton

A.J.

, and Williams

F.C.

(1995). Continuous regional cerebral cortical blood flow monitoring in head-injured patients. Neurosurgery (Baltimore) 36, 943–950.

23.

Soukup

, Bramsiepe

, Brucke

, Sanchin

, and Menzel

(2008). Evaluation of a bedside monitor of regional CBF as a measure of CO2 reactivity in neurosurgical intensive care patients. J. Neurosurg. Anesthesiol. 20, 249–255.

24.

Tackla

, Hinzman

J.M.

, Foreman

, Magner

, Andaluz

, and Hartings

J.A.

(2015). Assessment of cerebrovascular autoregulation using regional cerebral blood flow in surgically managed brain trauma patients. Neurocrit. Care, 23, 339–346.

25.

Tenjin

, Yamaki

, Nakagawa

, Kuboyama

, Ebisu

, Kobori

, Ueda

, and Mizukawa

(1990). Impairment of CO2 reactivity in severe head injury patients: an investigation using thermal diffusion method. Acta Neurochir. (Wien) 104, 121–125.

26.

Vajkoczy

, Roth

, Horn

, Lucke

, Thomé

, Hubner

, Martin

G.T.

, Zappletal

, Klar

, Schilling

, and Schmiedek

(2000). Continuous monitoring of regional cerebral blood flow: experimental and clinical validation of a novel thermal diffusion microprobe. J. Neurosurg. 93, 265–274.

27.

Lee

S.C.

, Chen

J.F.

, and Lee

S.T.

(2005). Continuous regional cerebral blood flow monitoring in the neurosurgical intensive care unit. J. Clin. Neurosci. 12, 520–523.

28.

Brennan

, Tompkins

, Stevens

F.A.

, and Carter

L.P.

(1998). Treatment of severe brain injury based upon cerebral blood flow. In: Acta Neurochirurgica Supplement; Intracranial pressure and neuromonitoring in brain injury. Marmarou

, Bullock

, Avezaat

, et al. (eds). Springer-Verlag Wien: Vienna, pp. 384.

29.

Carter

L.P.

, Dickman

, and Williams

F.C.

(1990). Feasibility of monitoring cerebral blood flow continuously in patients with severe head trauma. J. Trauma, 30, 916.

30.

Garcia-Perez

M.L.

, Montero

M.J.

, Badenes Quiles

, De Fez

, Maruenda

, and Belda

F.J.

(2011). Cerebral blood flow by thermal diffusion in severe traumatic brain injury. Correlation with neuromonitoring and prognostic implications. European Society of Intensive Care Medicine 24th Annual Congress, September, Berlin, Germany. Int. Care Med. S46.

31.

Hirsch

K.G.

, Yue

, Choukalas

, and Singh

(2013). Cerebral venous-to-arterial carbon dioxide gradient is a novel biomarker for brain injury. Neurocritical Care Society 11th Annual Meeting, October. 1–4, Philadelphia. Neurocrit. Care, S39.

32.

Phan

, Rosenthal

, and Manley

(2009). Bedside assessment of cerebral pressure autoregulation and vasoreactivity in severe traumatic brain injury: insight in arterial vs. venous control. J. Neurotrauma, 26, A66.

33.

Oshorov

A.S.

, I;

Popugaev

, K.; Potapov

(2014). Influence of hyperthermia on the parameters of MAP, CPP and Prx in patients with severe TBI. Brain Injury. 28.

34.

Sena

, Krasberg

, Pappu

, Olin

, Mead

, Nemoto

E.M.

, and Yonas

(2013). Thermal dilution CBF in white matter is a useful clinical measurement that does provide new insight into ICP alterations. J. Neurotrauma, 30, A129.

35.

Valadka

A.B.

, Hlatky

, Furuya

, and Robertson

C.S.

(2002). Brain tissue PO2: correlation with cerebral blood flow. Acta Neurochir. Suppl. 81, 299–301.

36.

Volovici

, Huijben

J.A.

, Ercole

, Stocchetti

, Dirven

C.M.F.

, van der Jagt

, Steyerberg

E.W.

, Lingsma

H.F.

, Menon

D.K.

, Maas

A.I.R.

, and Haitsma

I.K.

(2018). Ventricular Drainage Catheters versus Intracranial Parenchymal Catheters for Intracranial Pressure Monitoring-Based Management of Traumatic Brain Injury: A Systematic Review and Meta-Analysis. J. Neurotrauma. doi: 10.1089/neu.2018.6086 [Epub ahead of print].

37.

Zeiler

F.A.

, Thelin

E.P.

, Helmy

, Czosnyka

, Hutchinson

P.J.A.

, and Menon

D.K.

(2017). A systematic review of cerebral microdialysis and outcomes in TBI: relationships to patient functional outcome, neurophysiologic measures, and tissue outcome. Acta Neurochir. (Wien) 159, 2245–2273.

38.

Lang

E.W.

, Kasprowicz

, Smielewski

, Santos

, Pickard

, and Czosnyka

(2015). Short pressure reactivity index versus long pressure reactivity index in the management of traumatic brain injury. J. Neurosurg. 122, 588–594.

39.

Thelin

E.P.

, Zeiler

F.A.

, Ercole

, Mondello

, Buki

, Bellander

B.M.

, Helmy

, Menon

D.K.

, and Nelson

D.W.

(2017). Serial sampling of serum protein biomarkers for monitoring human traumatic brain injury dynamics: a systematic review. Front. Neurol. 8, 300.

40.

Furuya

, Hlatky

, Valadka

A.B.

, Diaz

, and Robertson

C.S.

(2003). Comparison of cerebral blood flow in computed tomographic hypodense areas of the brain in head-injured patients. Neurosurgery, 52, 340–345; discussion 345–346.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB

Continuous Thermal Diffusion-Based Cerebral Blood Flow Monitoring in Adult Traumatic Brain Injury: A Scoping Systematic Review

Abstract

Introduction

Methods

Search questions

Primary

Secondary

Functional outcome

Neurophysiological outcome and imaging-based outcome

Inclusion and exclusion criteria

Search strategy

Study selection

Data collection

Bias assessment

Statistical analyses

Results

Search strategy results

Study description and demographics

Thermal diffusion flowmetry technique

Functional outcome studies

Neurophysiological outcome studies

TDF-CBF association with ICP

TDF-CBF association CPP

TDF-CBF association with PbtO2

TDF-CBF association with cerebrovascular reactivity

TDF-CBF association with other CBF measures

Complications of TDF

Risk of bias

Discussion

Limitations

Conclusion

Footnotes

Acknowledgments

Author Disclosure Statement

Supplementary Material

References

Supplementary Material

TDF-CBF association with PbtO₂