Article, Neurology

Cerebral oximetry with blood volume index and capnography in intubated and hyperventilated patients

a b s t r a c t

Objective: Hyperventilation-induced hypocapnia leads to cerebral vasoconstriction and hypoperfusion. Intubated patients are often inadvertently hyperventilated during resuscitations, causing theoretical risk for Ischemic brain injury. Current emergency department monitoring systems do not detect these changes. The purpose of this study was to determine if cerebral oximetry (rcSO2) with Blood volume index (CBVI) would detect hypocapnia-induced cerebral tissue hypoxia and hypoperfusion.

Methods: Patients requiring mechanical ventilation underwent end-tidal CO2 (ETCO2), rcSO2, and CBVI monitoring. Baseline data was analyzed and then the effect of varying ETCO2 on rcSO2 and CBVI readings was analyzed. Median rcSO2 and CBVI values were compared when above and below the ETCO2 30 mmHg threshold. Subgroup analysis and descriptive statistics were also calculated.

Results: Thirty-two patients with neurologic emergencies and potential increased intracranial pressure were in- cluded. Age ranged from 6 days to 15 years (mean age, 3.1 years; SD, 3.9 years; median age, 1.5 years: 0.46-4.94 years). Diagnoses included Bacterial meningitis, viral meningitis, and seizures. ETCO2 crossed 30 mm Hg 80 times. Median left and right rcSO2 when ETCO2 was below 30 mmhg was 40.98 (35.3, 45.04) and 39.84 (34.64, 41) re- spectively. Median left and right CBVI when ETCO2 was below 30 mmhg was -24.86 (-29.92, -19.71) and -22.74 (-27.23, – 13.55) respectively. Median left and right CBVI when ETCO2 was below 30 mmHg was -24.86 (-29.92, -19.71) and -22.74 (-27.23, -13.55) respectively. Median left and right rcSO2 when ETCO2 was

above 30 mmHg was 63.53 (61.41, 66.92) and 63.95 (60.23, 67.58) respectively. Median left and right CBVI

when ETCO2 was above 30 mmHg was 12.26 (0.97, 20.16) and 8.11 (-0.2, 21.09) respectively. Median duration ETCO2 was below 30 mmHg was 17.9 minutes (11.4, 26.59). Each time ETCO2 fell below the threshold, there was a sig- nificant decrease in rcSO2 and CBVI consistent with decreased cerebral blood flow. While left and right rcSO2 and CBVI decreased quickly once ETCO2 was below 30 mmHg, increase once ETCO2 was above 30 mmHg was much slower. Conclusion: This preliminary study has demonstrated the ability of rcSO2 with CBVI to noninvasively detect the real- time effects of excessive hyperventilation producing ETCO2 b 30 mmHg on cerebral physiology in an emergency de- partment. We have demonstrated in patients with suspected increased intracranial pressure that ETCO2 b 30 mmHg causes a significant decrease in cerebral blood flow and regional Tissue oxygenation.

(C) 2016

? Work previously presented at the 22nd Pediatric Critical Care Colloquium; poster ses-

sion; September 11-13, 2015; Little Rock, AR.

?? Grants/funding: none obtained for this study.

? Conflicts of interest: The authors declare that they have no conflict of interest.

?? Author contributions: TAB and TJA conceived the study, designed the trial, ensured data collec-

tion, and drafted figures/tables and manuscript. TJA, GWA, JWO, EAS, NWHP, TMT, and EO assisted with conduction of the study and data collection. ZH and TN provided statistical planning/design and data analysis. All authors contributed significantly to manuscript revision and editing.

* Corresponding author at: University of Arkansas for Medical Sciences, 1 Children’s Way, Slot 512-16, Little Rock, AR 72202. Tel.: +1 501 364 1089.

E-mail address: [email protected] (T.A. Bagwell).

1 Previously worked at: Department of Pediatrics, Division of Pediatric Emergency Med- icine, Vanderbilt School of Medicine, Nashville TN.

+ The author has passed away last year after his contribution to this work.

Introduction

There is ample evidence to support the physiologic premise that hy- perventilation leads to decreased arterial blood carbon dioxide partial pressure (PaCO2) with subsequent decreased cerebral blood flow [1-3]. Hypocapnia-induced vasoconstriction and associated decreased cerebral blood flow lead quickly to subtle clinical signs and symptoms of compromised cerebral perfusion. If severe or prolonged, this hypo- perfusion may increase a patient’s risk for ischemic brain injury [2-4]. Variations in PaCO2 and corresponding capnography or end-tidal CO2

http://dx.doi.org/10.1016/j.ajem.2016.03.005

0735-6757/(C) 2016

(ETCO2) are frequently encountered during trauma and medical resusci- tations and postresuscitative care; the etiologies are numerous [2]. For example, intubated pediatric patients are often inadvertently hyperventilated while being manually ventilated or bagged during re- suscitations [5]. In addition, therapeutic hyperventilation has been used in the acute management of known or suspected increased intra- cranial pressure (ICP) for more than 40 years [4]. Standard prehospital and emergency department (ED) monitoring systems include cardiac telemetry, pulse oximetry, and capnography. Capnography or ETCO2 trends can be influenced by changes in ventilation, pulmonary and sys- temic perfusion, and/or pulmonary gas diffusion. Standard resuscitation goals aim to keep PaCO2 at normal values between 35 and 46 mm Hg or between 30 and 35 mm Hg if instituting therapeutic hyperventilation. Unfortunately, ETCO2 does not always correlate with PaCO2, and without confirmation via blood gas, PaCO2 may not be well controlled during resuscitation. Current monitoring systems are unable to detect hypocapnia-induced changes in regional cerebral perfusion and/or physiology, and there is no validated Noninvasive tool for monitoring changes in cerebral perfusion during resuscitation and postresuscitative care. However, the noninvasive use of cerebral oximetry or regional ce- rebral tissue oxygen saturation (rcSO2) with cerebral blood volume index (CBVI) is capable of detecting real-time physiologic changes and providing feedback to clinicians regarding local cerebral tissue perfu- sion, oxygenation, and metabolism/Oxygen extraction [6-7,9-11]. Cere- bral oximetry readings are obtained using near infrared spectroscopy probes placed over the forehead [7,8]. There are standardized normal values, and acquired readings/trends reflect dynamic changes in local tissue perfusion, oxygenation, and metabolism/oxygen extraction re- ferred to as Regional cerebral tissue oxygen saturation [6-7,9-11]. Values are expressed as a ratio of venous oxyhemoglobin to deoxyhemoglobin (rcSO2), and pediatric normal values range from 60% to 80% [6,9,10-11]. Abnormally low values (rcSO2 b 50%), abnormally high values (rcSO2 N 80%), and interhemispheric discordance (difference in left and right of N 10%) reflect pathology [11]. Studies have shown that rcSO2 values less than 50% or a change (in either direction) of 20% or greater from a patient’s baseline may be indicative of increased risk for hypoxic brain injury [11]. From these data, information on regional blood flow can also be extrapolated, and this is referred to as cerebral blood volume index with readings ranging from -50 to +50 [9]. Negative CBVI read- ings are interpreted as low-flow states, and readings are not pulse de- pendent, which is especially useful in resuscitations [9].

We present information regarding a convenience sampling of 32 intubated and manually ventilated patients with viral meningitis, bacte- rial meningitis, and/or seizures at risk for increased ICP. Data will be pre-

at risk for increased ICP, unstable multisystem trauma, respiratory fail- ure, or cardiac arrest [6-10]. When in use, patients have left and right cerebral oximetry probes (INVOS 5100C Somanetics) placed on the forehead, and data (left and right rcSO2 and left and right CBVI) are sam- pled and recorded at 5-second intervals [6-10]. Intubated PED pateints also have inline capnography with ETCO2 recordings every 5 seconds.

PED patients with neurologic emergencies–viral meningitis, bacterial meningitis with positive cerebrospinal fluid cultures, and/or sei- zures requiring active resuscitation and intubation with periods of cap- tured unintentional hyperventilation (ETCO2 b 30 mmHg) –were initially selected from the pediatric ED EMR, PED intubation, and cerebral oximetry REDCap (Research Electronic Capture Database) databases. Patients were selected by one of the primary investigators (after questioning the patient’s respiratory therapist [RT] or trainee). Episodes of hypocapnia (ETCO2 b 30 mmHg greater than 2 minutes) were selected as inclusion criteria because our review was designed to detect the effects of deviation from the standard clinical practice for increased ICP manage- ment targeting ETCO2 30-35 mmHg. At both PEDs, there is a dedicated PED RT who supervised accredited RT trainees. Upon patient resuscitation with intubation, the PED RT performs continuous bagging until the pa- tient is transferred to the pediatric intensive care unit. The PED ETCO2 pro- tocol was a target range between 35 and 45 mmHg unless the PED attending or fellow suspects increased ICP; then the ETCO2 target range is 30-35 mmHg. Only patients with inclusion and exclusion criteria and PED left and right rcSO2, CBVI, and ETCO2 monitoring with no gaps in data recording were selected by the primary investigator for analysis. Table 1 gives a stepwise process for patient selection and exclusion. Per PED cerebral Near-infrared spectroscopy standard protocol, left and right probes were placed, and ETCO2 monitoring was used during resusci- tation, RSI, intubation, ventilation, and cerebral oximetry and ETCO2 data were internally recorded every 5 seconds. The left and right rcSO2, CBVI se- quence, date, and times were matched to the corresponding ETCO2. Corre- lating continuous bagging ventilation rate by the PED RT (or trainee or non-PED RT) to the ETCO2 recording was not possible. monitoring devices remained in place for the duration of ED care. Health Insurance Portability and Accountability Act compliance for all aspects of patient data was maintained. The institutional review board at each facility granted ap- proval including a waiver of informed consent.

Table 1

Data collection and patient inclusion and exclusion methods

sented detailing ETCO2, rcSO2, and CBVI during resuscitation including Rapid Sequence Intubation and postresuscitative care. All patients experienced nontherapeutic hyperventilation at some point during their care. Data analysis will be reviewed detailing a distinct correlation between falling EtCO2 values and falling rcSO2 and CBVI values. These findings illustrate the role cerebral oximetry can play during resuscita- tion as a noninvasive and real-time cerebral physiology assessment tool. Findings also suggest that the utilization of this additional informa- tion may allow clinicians to more accurately tailor neurocardiovascular resuscitative and postResuscitative efforts in the pediatric ED (PED).

Methods

We present an observational retrospective case series detailing the rcSO2, CBVI, and ETCO2 readings of 32 intubated and manually ventilated patients during PED resuscitation and stay. Patients were selected and data was collected at 2 level 1 trauma centers (pediatric EDs) at pediat- ric tertiary care facilities from 2011 to 2015. Data was pulled from both facilities’ PED RSI continuous quality improvement programs database, PED electronic medical record(EMR) and vital signs, and PED cerebral oximetry database files. At both facilities, cerebral oximetry is a stan-

Study inclusion criteria (in stepwise

Selection process):

PED patients with meningitis: viral or bacterial with positive CSF cultures, with or without seizures.
  • Medical(nontrauma history) resuscitation with RSI intubation
  • Potential for medically induced increased ICP (nontrauma) as indicated by PED records or by CT scan report indicating possible increased ICP
  • Bag ventilation by RT or trainee (never placed on ventilator) with simultaneous ETCO2 monitoring with simultaneous cerebral oximetry
  • Unintentional ETCO2 b 30-mm Hg periods (N 2 min) due to excessive bagging upon questioning the patient’s RT or RT student and not PED attending directed
  • Had continuous (5-s interval) left and right rcSO2 with left and right CBVI correlating with ETCO2 monitoring with no missing data
  • Study exclusion criteria:

    Episodes of bradycardia after intubation
  • Episodes of hypotension by BP measurements for age during PED stay
  • Requiring pressor support during PED stay especially 10 min before and 20 min after low ETCO2
  • Trauma history
  • CT scan indicative of Physical abuse pattern (old or new)
  • Cerebrovascular accidents
  • Hydrocephalus with cerebral spinal
  • fluid shunts

    brain tumors new or previously diagnosed
  • Prior neurosurgical interventions
  • Known cerebrovascular anomalies
  • Lapse or lack of corresponding left and right rcSO2, left and right CBVI with ETCO2
  • dard monitoring tool in pediatric patients with neurologic emergencies

    CT, computed tomography; BP, blood pressure.

    2.1. Statistical analysis

    Collected measurements (left and right rcSO2, CBVI, and ETCO2) in the first 5 minutes were analyzed. Our prior studies have demonstrated a correlation between rcSO2 b 60 and abnormal cerebral physiology/ pathology, so based on the initial value of rcSO2, subjects were split into 2 groups. Group 1 consisted of patients with b 60 rcSO2 (abnormal cerebral physiology/pathology) and group 2 consisted of patients with 60-80 rcSO2 (normal cerebral physiology/pathology). Patients were subgrouped and analyzed to assess whether those with initial abnormal rcSO2 had greater changes than patients with initial normal rcSO2. Mea- surements were also analyzed to detect changes in rcSO2 and CBVI in re- lationship to changes in ETCO2 above and below a threshold of 30 mmHg. Every time ETCO2 crossed the threshold of 30 mmHg, it was considered 1 time period of change. Relative changes in rcSO2 and CBVI were calculated as percentages from baseline and were determined sep- arately for each patient using the rcSO2 and CBVI values at the ETCO2 threshold of 30 mmHg. This variability in rcSO2 and CBVI was expressed as percent change from patient baseline (10%, 15%, 20%, 50%) by both number and percentage of occurrences. Descriptive statistics such as median and interquartile ranges of time to relative change on measure- ments were also reported. Measurements were compared via utilization of the Wilcox on sum rank test. All P values less than .05 were consid- ered statistically significant. The analyses and figures were produced using R v.3.2.3 (R Foundation for Statistical Computing, Vienna, Austria).

    Results

    A total of 32 patients met the study criteria. Patients ranged in age from 6 days to 15 years (mean age, 3.1 years; SD, 3.9 years; median age, 1.5: 0.46-4.94 years). All patients were intubated with episodes of captured hypocapnia (ETCO2 b 30 mmHg) which upon review with the patient’s RT, were deemed secondary to unintentional excessive bag- ging by non-PED RT or RT trainees. None of these events were secondary to the initiation of the therapeutic hypocapnia for suspected increased ICP. Diagnoses included bacterial meningitis (with and without sei- zure), viral meningitis (with and without seizure), and encephalitis with seizures. Viral and encephalitis CSF cultures revealed adenovirus, influenza A, enterovirus, arbovirus (West Nile), and human herpesvirus

    6. The bacterial meninigits CSF cultures revealed group B Streptococcus,

    Streptococcus pneumoniae, Haemophilus influenzae, and Escherichia coli. Data was analyzed to compare rcSO2 and CBVI readings in patients when ETCO2 was both above and below a threshold of 30 mmHg. Adjusting for age and sex due to small sample size was not possible. Line charts were made for each patient as a visual representation of the changes in rcSO2 and CBVI in relationship to ETCO2; a sampling of this data is displayed in Fig. 1. When hypocapnia was present (ETCO2 b 30 mmHg), there was an associated precipitous drop in rcSO2 and CBVI. ETCO2 crossed the threshold of 30 a total of 80 times in 32 patients. Upon further review of the EMR and PED RT interviews, changes in ETCO2 (ETCO2 b 30 mmHg) were driven by changes in ventilation alone. Median left and right rcSO2, median left and right CBVI, and median ETCO2 were calculated for the first 5 minutes of resuscitation including both normal cerebral oximetry readings and abnormal cerebral oxime- try readings. This data is reflected in Table 2. Table 2 also details the changes in left and right rcSO2 and left and right CBVI when ETCO2 was above and below the threshold of 30 mmHg. Findings were expressed as Percentage change in left and right rcSO2 and left and right CBVI from patient baseline by 10%, 15%, 20%, and 50%. In addition, Table 2 de- tails median left and right rcSO2, and median left and right CBVI when ETCO2 was above and below the threshold of 30 mmHg. Median left and right rcSO2 when ETCO2 was below 30 mmHg was 40.98 (35.3, 45.04) and 39.84 (34.64, 41) respectively. Median left and right CBVI

    when ETCO2 was below 30 mmHg was -24.86 (-29.92, -19.71) and

    -22.74 (-27.23, – 13.55) respectively. Median left and right rcSO2 when ETCO2 was above 30 mmHg was 63.53 (61.41, 66.92) and

    63.95 (60.23, 67.58) respectively. Median left and right CVBI when ETCO2 was above 30 mmHg was 12.26 (0.97, 20.16) and 8.11 (-0.2, 21.09) respectively. Median duration ETCO2 was below 30 mmHg was 17.9 minutes (11.4, 26.59). When ETCO2 fell below a threshold of 30 mmHg, there was a resultant change in rcSO2 and CBVI consistent with decreased cerebral blood flow. Each time ETCO2 fell below a thresh- old of 30 mmHg, there was a corresponding consistent decrease in left and right rcSO2 and left and right CBVI. In each instance, criteria for low flow state were met as illustrated by a change in rcSO2 baseline of greater than 20%, rcSO2 value less than 50, and/or negative CBVI value. While left and right rcSO2 and CBVI decreased quickly once ETCO2 was 30 mmHg, increase once ETCO2 was below above 30 mmHg was much slower, as demonstrated in Table 2.

    Discussion

    This case series has demonstrated the ability of rcSO2 with CBVI to noninvasively detect the effects of excessive hyperventilation during re- suscitation on cerebral physiology in patients with a neurologic emer- gency with potential increased ICP in the ED setting. We have demonstrated that hypocapnia (ETCO2 b 30 mmHg) secondary to exces- sive hyperventilation leads to a significant decrease in left and right rcSO2 and CBVI, which is reflective of regional cerebral blood flow, perfu- sion, and oxygenation [13-19]. These patients consistently demonstrat- ed changes reflective of decreased cerebral blood flow during times of excessive hyperventilation. In addition, our patients consistently dem- onstrated changes in cerebral perfusion significant enough to be consid- ered low-flow states indicating a potential increased risk for ischemic brain injury [3-19].

    The PaCO2 has a profound effect on Cerebral autoregulation. Elevation of PaCO2 (hypercapnia) produces cerebral vasodilation, increased cere- bral blood flow, and subsequently increased ICP. Inversely, decreased PaCO2 (hypocapnia) produces vasoconstriction, decreased cerebral blood flow, and subsequently decreased ICP. This phenomenon is re- ferred to as cerebrovascular CO2 reactivity [1-3]. Normal PaCO2 values are generally accepted to range from 35 to 45 mmHg. The relationship between PaCO2 and cerebral vascular tone is not linear, and studies have reported that effects are most prominent when PaCO2 levels range from 30 to 50 mmHg [2-4,7,11,13]. Understanding the clinical implications of cerebrovascular CO2 reactivity is essential, as variations in PaCO2 are often encountered during medical or trauma resuscitations [2,11-18]. However, it is also important to keep in mind that cerebral autoregulation is multifactorial and dynamic (depending on cerebral perfusion pressure and nonperfusion pressure processes) [2,4,11-14]. As such, studies have shown that the effects of PaCO2 on cerebral auto- regulation may be decreased during periods of hypotension. However, studies have also shown that a combination of hypotension (decreased cerebral perfusion pressure) and hypocapnia may have deleterious ef- fects on cerebral blood flow [2-4,11-18].

    The effects of PaCO2 on cerebral autoregulation can have important clinical implications. Studies have shown a correlation between cerebral vasoconstriction, decreased cerebral blood flow, poor local tissue perfu- sion, and risk for ischemia [2-4,11-18]. Studies on vasovagal syncope (in healthy hyperventilating subjects) have demonstrated that vasocon- striction and decreased cerebral blood flow can lead to signs of compro- mised cerebral perfusion such as poor hemoglobin oxygen saturation, aura, generalized discomfort, electroencephalogram changes, and ulti- mately syncope [2,4,11-15]. Signs of compromised perfusion developed rapidly in many, typically within 20-40 seconds from onset of hyper- ventilation [3,13,11-18]. Studies have also demonstrated a relationship between poor functional outcomes and hypocapnia in head trauma and stroke patients [11-16].

    Hyperventilation for treatment of increased ICP has been used for more than 40 years [4]. General consensus has been to avoid PaCO2 levels less than 30 mmHg [4]. Much of the literature concerning cere- brovascular CO2 reactivity has been completed in adult patients with

    Fig. 1. Selected patient ETCO2 graphs demonstrating changes in left and right rcSO2 and CBVI correlating with inadvertent hyperventilation-induced hypocapnia (ETCO2 b 30 mm Hg).

    Traumatic brain injury . Stocchetti et al [4] published a review of data concerning TBI and the role of hyperventilation in 2005, which in- cluded analysis of studies dating back to 1966. Numerous studies were reviewed illustrating the profound impact of PaCO2 on cerebral blood flow and blood volume. It is estimated that there is a 3% change in cere- bral blood flow for every 1 mmHg change in PaCO2 [4]. Studies cited in this review seemed to support this premise. One study (Fortune et al) found that decreasing PaCO2 to 26 mm Hg in healthy individuals resulted in a reduction of cerebral blood volume by 7.2% and a reduction of cere- bral blood flow by 30.7% [4]. Another study (Raichle et al) found that a baseline decrease in PaCO2 of 15-22 mmHg in healthy volunteers caused a subsequent 40% reduction in cerebral blood flow [4]. In addition, Stocchetti et al found a mean blood volume change of 0.72 mL (+-0.42 mL) for each mmHg change in PaCO2, and (Yoshihara et al) found that a change in cerebral blood volume of merely 0.5 mL was enough to cause a subsequent change in ICP of 1 mmHg [4]. Although significant, these effects seem to be transient. Muizelaar et al and Raichle et al both demonstrated a relatively short-lived effect (4-24 hours) and even highlighted risk for a rebound increase in cerebral blood flow (and ICP) after departure from hyperventilation [4].

    Certainly, the concern with utilization of therapeutic hyperventila- tion in treatment of increased ICP (in traumatic brain injury) seems to be a risk for induction or exacerbation of preexisting brain ischemia [4]. Therefore, the safety of therapeutic hyperventilation is debatable. One study cited in the 2005 Stocchetti et al review concluded that while clearly decreasing cerebral blood flow, short-lived “moderate” hyperventilation failed to alter cerebral metabolism or oxygen extrac- tion [4]. In contrast, a prospective randomized trial by Muizelaar et al demonstrated poorer outcomes in TBI patients who were profoundly hyperventilated (PaCO2 of 25 mmHg for 5 days postinjury) compared

    with those who were maintained at normoventilation (PaCO2 of 35 mmHg for 5 days postinjury) [4]. After extensive review, the 2005 Stocchetti et al article seemed to conclude that while inherent risk ex- ists, judicious use of brief hyperventilation in the treatment of increased ICP is appropriate. However, the authors also suggested that therapeutic hyperventilation should be reserved for those patients with suspected or confirmed increased ICP and that physiologic monitoring should be employed during its use [4]. In light of these findings, health care pro- viders should be judicious with the use of hyperventilation/hypocapnia as therapy and should remain cognizant of the risks of inadvertent hy- perventilation with the use of manual ventilation or bagging during re- suscitation [4].

    Unfortunately, inadvertent hyperventilation during medical and trauma resuscitations seems to be commonplace, and pediatric patients are particularly vulnerable. According to a recent retrospective observa- tional study of ventilation rates during mock codes at a large tertiary care pediatric facility, hyperventilation took place at some point in all evaluated resuscitations (number 68) [5]. Rates of hyperventilation did not vary with provider type. In fact, no significant difference was found regardless of level of training [5]. Additional studies have shown that inadvertent hyperventilation commonly occurs in the prehospital setting. Thomas et al found that 70% of helicopter transport patients were hyperventilated en route to trauma centers (patents with TBI) [4]. Emergency medicine providers should be aware of this tendency and of the possibility that real-time physiologic monitoring with adjunctive tools such as cerebral oximetry and capnography may be of benefit.

    The use of noninvasive monitoring devices including capnography (ETCO2) and cerebral oximetry (cerebral near-infrared spectroscopy) is becoming more common in PEDs. These devices allow real-time

    Table 2

    The effect of varying ETCO2 on left and right rcSO2 and BVI values comparison of median (Q1, Q3) between the groups: left and right rcSO2 and CBVI and ETCO2 were defined by whether ETCO2 b 30 mm Hg N. Time for rcSO2 and BVI to have a 15% change when ETCO2 dropped to b 30 mm Hg (minutes). Time for rcSO2 and BVI to have a 15% change after ETCO2 increased beyond 30 mm Hg (minutes). Comparison of measurements based on their initial 5-minute left and right rcSO2 and CBVI and ETCO2 values in patients with significant cerebral pathology (rcSO2 b 60) vs normal cerebral physiology(rcSO2 = 60-80)

    10% change (yes) n (%)

    15% change (yes) n (%)

    20% change (yes) n (%)

    50% change (yes) n (%)

    Left cerebral rcSO2 (normal range, 60-80)

    78 (98)

    78 (98)

    78 (98)

    60 (75)

    Right cerebral rcSO2 (normal range, 60-80)

    79 (99)

    79 (99)

    79 (99)

    53 (66)

    Left BVI (normal range, -50 to +50)

    80 (100)

    79 (99)

    79 (99)

    76 (95)

    Right BVI (normal range, -50 to +50)

    80 (100)

    79 (99)

    79 (99)

    78 (98)

    (Above) Frequency of change in left and right rcSO2 and left and right BVI during the time ETCO2 is below 30 mm Hg (Below) Frequency of change in left and right rcSO2 and left and right BVI during the time ETCO2 is above 30 mmHg

    10% change (yes) n (%)

    15% change (yes) n (%)

    20% change (yes) n (%)

    50% change (yes) n (%)

    Left cerebral rcSO2 (normal range, 60-80)

    32 (40)

    31 (39)

    31 (39)

    16 (20)

    Right cerebral rcSO2 (normal range, 60-80)

    32 (40)

    32 (40)

    31 (39)

    16 (20)

    Left BVI (normal range, -50 to +50)

    32 (40)

    32 (40)

    32 (40)

    32 (40)

    Right BVI (normal range, -50 to +50)

    32 (40)

    32 (40)

    32 (40)

    32 (40)

    Comparison of median left and right rcSO2 and left and right CBVI and between ETCO2 groups both above and below 30mmHg

    ETCO2 b 30 ETCO >= 30 Difference

    P value

    Left cerebral rcSO2 (normal range, 60-80) 40.98 (35.3, 45.04) 63.53 (61.41, 66.92) 25.54 (18.1, 27.7)

    Right cerebral rcSO2 (normal range, 60-80) 39.84 (34.64, 41) 63.95 (60.23, 67.58) 24.85 (19.88, 29.85)

    Left BVI (normal range, -50 to +50) -24.86 (-29.92, -19.71) 12.26 (0.97, 20.16) 36.68 (21.6, 47.58)

    b.001 b.001 b.001

    Right BVI (normal range, -50 to +50) -22.74 (-27.23, -13.55) 8.11 (-0.2, 21.09) 30.66 (16.74, 45.27)

    b.001

    Comparison of measurements based on their initial value Initial rcSO2 60-80 n = 23 Initial rcSO2 b 60 n = 9

    P value

    Left cerebral rcSO2 (normal range, 60-80) 64.92 (63.61, 69.28) 56.12 (52.17, 57.28)

    Right cerebral rcSO2 (normal range, 60-80) 65.2 (61.98, 68.05) 54.8 (54.17, 55.05)

    Left BVI (normal range, -50 to +50) 6.03 (-5.91, 14.53) 11.3 (-14.68, 16.87)

    b.0001 b.0001

    .8668

    Right BVI (normal range, -50 to +50) 10.75 (-14.2, 17.97) 15.72 (12.23, 23.23)

    .2945

    ETCO2 42.12 (41.44, 46.03) 42.52 (23.92, 46.03)

    .6443

    Time for rcSO2 and BVI 15% change after ETCO2 b 30 (min) n(%) = 79 Mean (SD +) Time (min) for rcSO2 and BVI 15% change after ETCO2 N 30 n(%) = 32

    Mean (SD +)

    Left cerebral rcSO2 (normal range, 60-80) 0.83 (+1.68) Left cerebral rcSO2 (normal range, 60-80)

    16.48 (+ 14.71)

    Right cerebral rcSO2 (normal range, 60-80) 0.76 (+1.61) Right cerebral rcSO2 (normal range, 60-80)

    13.68 (+ 14.13)

    Left BVI (normal range, -50 to +50) 0.56 (+1.17) Left BVI (normal range,-50 to +50)

    2.28 (+4.92)

    Right BVI (normal range, -50 to +50) 0.49 (+0.99) Right BVI (normal range, -50 to +50)

    1.13 (+1.74)

    physiologic feedback, which can be especially useful during resuscita- tions [6-11,13-19]. ETCO2 monitoring is available via nasal cannula, bag mask, or endotracheal tube and allows trending of changes in con- centration of CO2 in a patient’s expired gas [7]. Value trends should cor- relate with expected PaCO2 values and reflect changes in patient ventilation and/or metabolism. There are standardized normal values, and acquired readings/trends reflect dynamic changes in local tissue perfusion, oxygenation, and metabolism/oxygen extraction referred to as regional cerebral tissue oxygen saturation [6,7,9-11,13-19]. Values are expressed as a ratio of venous oxyhemoglobin to deoxyhemoglobin (rcSO2), and pediatric normal values range from 60% to 80% [6,9-11]. Ab- normally low values (rcSO2 b 50%), abnormally high values (rcSO2 N 80%), and interhemispheric discordance (difference in left and right of N 10%) reflect pathology [7-11,13-19]. Studies have shown that rcSO2 values less than 50% or a change (in either direction) of 20% or greater from a patient’s baseline may be indicative of increased risk for hypoxic brain injury [11,13-19]. Negative CBVI readings are interpreted as low- flow states, and readings are not pulse dependent, which is especially useful in resuscitations. [7-11,13-19] The physiologic information available via use of these adjunctive tools may aid clinicians in tailoring care during resuscitations.

    As detailed above, CO2 has a profound effect of cerebral autoregula-

    tion. Variations in PaCO2 are commonly encountered during both medi- cal and trauma resuscitations. In addition, pediatric patients are often inadvertently hyperventilated during resuscitations. This physiologic relationship has been the basis for use of therapeutic hyperventilation in the acute management of increased ICP. The therapeutic use of hyper- ventilation is not without risk, and clinicians should be cognizant of the

    advantages and disadvantages of this practice, as detailed above. Chang- es in cerebral perfusion secondary to variations in PaCO2 occur rapidly. Hypocapnia-induced cerebral vasoconstriction puts patients at risk for ischemic brain injury and poorer functional outcomes [13-19]. Current ED Standard monitoring systems such as telemetry, pulse oximetry, and capnography are unable to detect changes in cerebral perfusion and physiology. However, rcSO2 with CBVI has the capability to detect abnor- mal cerebral hemispheric physiology, tissue hypoxia, and hypoperfu- sion, providing-real time physiologic feedback to clinicians [13-19].

    This study has illustrated the physiologic premise that hyperventila- tion leads to decreased cerebral blood flow and demonstrated the abil- ity of cerebral oximetry monitoring to noninvasively detect the effects of excessive hyperventilation during resuscitation on cerebral physiolo- gy. It has preliminarily highlighted the risks associated with excessive hypocapnia by showing that inadvertent hyperventilation causing ETCO2 less than 30 mmHg leads to a significant decrease in left and right rcSO2 and CBVI in patients with a neUrologic emergency and potentially increased ICP in the ED setting [13-19]. These changes are reflective of decreased regional cerebral blood flow, perfusion, and oxy- genation [13-19]. These changes would not have been detected by rou- tine monitoring in the ED.

    While ETCO2 monitoring has become a standard tool during resusci- tation and management of the critically ill and injured, cerebral oxime- try may not be available to all providers. For these providers, this study has supported the tenant that hyperventilation producing ETCO2 b 30 mmHg has a significant effect on cerebral physiology. Therefore, main- taining a patient’s ETCO2 N 30 mm Hg should prevent hypocapnia in- duced vasoconstriction and decreased cerebral blood flow.

    Limitations

    This is a preliminary study limited by its observational and conve- nience sample perspective. Analyzing for sex, age, and disease state (viral vs bacterial meninigits) was not obtainable because of the small patient sample size.

    We were unable to correlate manual ventilation rate with ETCO2 reading. In addition, our study suggests that rcSO2 and CBVI decreased quickly when ETCO2 was below 30 mmHg but that they increased slow- ly when ETCO2 was above 30 mmHg. This could be related to the cere- bral physiologic and perfusion changes associated with viral or bacterial meningitis. It could also be related to a higher precapillary sen- sitivity to decreased regional pCO2 vs increased pCO2 [11-18]. These findings in patients with abnormal cerebral physiology warrant further investigation.

    Conclusion

    This preliminary study in pediatric patients with neurologic emer- gencies and the potential for underlying increased ICP has demonstrat- ed the potentially deleterious effects of excessive hyperventilation (whether intentional or inadvertent, defined as ETCO2 b 30 mmHg) on interhemispheric regional cerebral physiology as detected by cerebral oximetry. Current standard ED monitoring methods would not have de- tected these hyperventilation induced effects on interhemispheric cere- bral physiology. In light of these findings, clinicians should be cognizant of the risk associated with therapeutic or inadvertent hyperventilation during resuscitation. Clinicians should consider the use of cerebral ox- imetry with CBVI and ETCO2 during resuscitation and postresuscitation for all patients, especially those with known or suspected neurologic emergencies with potential for increased ICP. Capnography or ETCO2 is an expected monitoring tool for emergency intubation and resuscitative endeavors; for clinicians who lack cerebral oximetry monitoring, this study has shown that ETCO2 b 30 mmHg does have an effect on cerebral physiology and needs to be avoided.

    Conflicts of interest

    The authors declare that they have no conflict of interest.

    References

    1. Kuznetsova DV, Kulikov VP. Cerebrovascular and systemic hemodynamic response to carbon dioxide in humans. Blood Press Monit 2014;19(2):81-9.
    2. Meng L, Gelb AW. Regulation of cerebral autoregulation by carbon dioxide. Anesthe- siology 2015;122(1):00-00 [E-publication ahead of print].
    3. Lagi A, Cencetti S, Corsoni V, Georgiadis D, Bacalli S. Cerebral vasoconstriction in va- sovagal syncope: any link with symptoms? A transcranial Doppler study. Circulation 2001;104(22):2694-8.
    4. Stocchetti N, Maas AI, Chieregato A, Vander Plas AA. Hyperventilation in head injury: a review. Chest 2005;127(5):1812-27.
    5. Niebauer JM, White ML, Zinkan JL, Youngblood AQ, Tofil NM. Hyperventilation in pe- diatric resuscitation: performance in simulated pediatric Medical emergencies. Pedi- atrics 2011;128(5):e1195-200.
    6. Kane I, Abramo TJ, Meredith M, Williams A, Crossman K, Wang W, et al. Cerebral ox- ygen saturation monitoring in pediatric altered mental status patients. Am J Emerg Med 2014;32(4):356-62.
    7. Abramo TJ, Harris ZL, Meredith M, Crossman K, Seupaul R, Williams A, et al. Cerebral oximetry with cerebral blood volume index in detecting pediatric stroke in a pediat- ric ED. Amer J Emerg Med 2015;33:1622-9.
    8. Gross T, Marcin J, Auerbach M, Middlebrooks L, Putzier P, Jaeger J, et al. New technol- ogies in emergency medical services for children. Clin Pediatr Emerg Med 2014; 15(1):67-78.
    9. Abramo TJ, Zhou C, Estrada C, Merideth M, Miller R, Pearson M, et al. Innovative ap- plication of cerebral rSO2 monitoring during Shunt tap in pediatric ventricular malfunctioning shunts. Pediatr Emerg Care 2014;00(00):00-00 [E-publication ahead of print.
    10. Abramo TJ, Meredith M, Jaeger M, Schneider B, Bagwell H, Ocal E, et al. Cerebral ox- imetry with blood volume index in asystolic pediatric cerebrospinal fluid malfunctioning shunt patients. Am J Emerg Med 2014;32(11):1439 [e1-1439.e7].
    11. Abramo TJ, Zhou C, Estrada C, Drayna PC, Locklair MR, Miller R, et al. Cerebral region- al oxygen saturation monitoring in Pediatric malfunctioning shunt patients. Am J Emerg Med 2013;31(2):365-74.
    12. Drayna PC, Abramo TJ, Estrada C. Near-infrared spectroscopy in the critical setting. Pediatr Emerg Care 2011;27(5):432-42.
    13. Ghanayem NS, Wernovsky G, Hoffman GM. Near infrared spectroscopy as a hemo- dynamic monitor in critical illness. Pediatr Crit Care Med 2011;12:S27-32.
    14. Murkin JM, Arango A. Near-infrared spectroscopy as an index of brain and tissue ox- ygenation. Br J Anaesth 2009;103:i3-i13.
    15. Figaji AA, Kent SJ. Brain tissue oxygenation in children diagnosed with brain death. Neurocrit Care 2010;12:56-61.
    16. Andropoulos D, Stayer S, McKenzie D, Fraser CD. Regional low-flow perfusion pro- vides comparable blood flow and oxygenation to both cerebral hemispheres during neonatal aortic arch reconstruction. J Thorac Cardiovasc Surg 2003;126(6):1712-7.
    17. Sheeren TW, Schober P, Schwarte LA. Monitoring tissue oxygenation by near infra- red spectroscopy (NIRS): background and current applications. J Clin Monit Comput 2012;26:279-87.
    18. Al-Rawi P, Kirkpatrick P. Tissue oxygen index: threshold for cerebral ischemia using near infrared spectroscopy. Stroke 2006;37:2720-5.
    19. Schneider B, Abramo TJ, Albert G. An update on cerebral oxygenation monitoring, an innovative application in cardiac arrest and neurological emergencies, annual up- date in intensive care and emergency medicine. Springer Publication; 2015 273-88.