JOSHUA KORNBLUTH, MDFEB 5 T H , 2013

NCCU Journal Club

Objective: to review a recent randomized trial of intracranial pressure monitoring in sTBI

To encourage thought and discussion of our own practices

To consider future directions

Nothing for me to disclose

Study Overview

Does intracranial pressure monitoring improve outcomes in severe TBI?

What is the current standard of care in America?

Why does this question matter?

Multicenter, 324 adult pts, severe TBI, randomized to two separate protocols

Looked at in-hospital events, survival time and 3- and 6- month outcomes

You heard it from the BTF…

Study Overview

Can we conduct this type of study in America?

Highly unlikely, so let's take advantage of alternative standards of care to investigate.

Study Design Multicenter, parallel-group trial Random assignment to ICP-monitoring group vs

imaging-clinical examination group Study started in Bolivian hospitals, and additional

hospitals were added later to increase enrollment

Study Design – Inclusion Criteria

Traumatic brain injuryGCS < 8 on admission or within first 48 hours after

injury (Motor score ≤ 5 if intubated)Admission to study hospital within 24 hours of injuryNo foreign object in the brain parenchyma.Age > 12Randomized:

within 24 hours of injury [for patients with GCS < 8 on admission] or

within 24 hours of deterioration [patients deteriorating to GCS < 8 within 48 hours of injury

Randomization stratified according to site, injury severity score, and age

Study Design – Exclusion Criteria

GCS of 3 with bilateral fixed and dilated pupilsNo consentPregnantPrisonerNo beds available in ICUNo ICP monitor availableNon-survivable injuryOther (e.g., Pre-injury life expectancy under 1

year)Pre-existing neurological disability that would

confound outcome

Study Design - protocol

Place patient on mechanical ventilation (VM)Place continuous SaPO2 and EtCO2 monitorsInsert indwelling urinary catheter to monitor urine outputInsert arterial catheter for arterial mean pressure

monitoringInsert central venous catheter for infusion of solutions

and central venous pressure monitoring.Monitor neurological clinical status each hour

Pupils GCS

Brain CT To evaluate evolution 48 hours after the admission CT To evaluate evolution 5-7 days after the admission CT p.r.n.

Study Design – Standards of Critical Care

Clearly delineate standard basic Critical Care

• Head positioning 30º• Head and neck in neutral position and aligned• Avoid hyperthermia (Defined as central temperature > 38 º C)

• Non-drug measures (cooling)• Dipirona (Metamizole sodium)

• Early enteral nutritional support• Before 48 hours• 25 Kcal/kg weight

• Pharmacologic prophylactic of post traumatic seizures (Phenytoin (IV or PO))• Load and maintenance dose as is being giving in each hospital

• Gastric bleeding prophylaxis• Ranitidine or Omeprazol

• Avoid decubitus lesions• Deep venous thrombosis prophylaxis• Frequent tracheal suctioning with sterile technique to prevent pulmonary

infections

Study Design – ICP group

Had parenchymal monitor ASAP (i.e. after randomization and resolution of coagulopathy if present)

Position was not specifiedTreat if ICP≥20mmHg x 5minIf CSF drainage indicated, EVD placedCPP goal 50-70mmHg

Study Design – Treatments (ICP Group)

Treatments based on a “Therapeutic Intensity Level”

If signs of intracranial HTN, clinical or imaging 1 – hyperosmolar therapy (mannitol)

5% NaCl only if hypotenisve, hypovolemic, hyponatremic 2 – optional mild hyperventilation (pCO2 30-35mmHg) 3 – Ventricular drainage if possible*

Study Design – Definitions (ICP Group)

Intracranial Pressure Definitions: Treatable intracranial hypertension:

ICP > 20 mmHg for > 5 minutes Treatment failure:

ICP not reduced to ≤ 20 mmHg within 20 minutes after a treatment intervention is initiated, and

Persistent elevation in ICP > 20 mmHg requires increase in therapeutic intensity level

Study Design – “Neuroworsening”

Neuroworsening = Inc’d TIL 1. Decrease in the motor GCS > 2 2. New loss of pupil reactivity 3. Interval development of pupil asymmetry of > 2mm 4. New focal motor deficit 5. Herniation syndrome

Give mannitol 0.25-1mg/kg to sOSM<320Hyperventilate to pCO2 25-30If no response thiopental x 3dCraniectomy for space-occupying lesions

Study Design – Imaging only Group

After optimized sedation and analgesia, hyperventilation and hyperosmotic therapy should be started simultaneously if there is evidence of edema on CT, as indicated as following: 1. Compressed peri-mesencephalic cisterns 2. Midline shift 3. Cortical sulcal compression / effacement

Otherwise, same metrics and goals of ICP monitored group

Corticosteriods prohibitedAED’s for prophylaxis >28d

Study Design - Outcomes

Primary outcome – 21-point composite of survival, duration and level of impaired consciousness, 3-month GOSe and GOAT, 6-month GOSe and neuropsych testing

Secondary Outcomes – ICU LOS, number of days that patients received at least 1 brain-specific treatment, days of MV, treatment with high-dose barbiturates, decompressive crani

Results

Results - Demographics

Results

MVA’s accounted for most injuries (51% of randomized pts)

45% of pts were brought in by ambulanceRemainder were transferred from other

facilitiesDid not publish pre-hospital demographics or

interventions as these we not uniformly recorded

Results - Demographics

24% of randomized patients had clinical decline to GCS within eligibility criteria

49% of patients had localizing signs on clinical exam

33% of participants required surgical treatment of mass lesions

On initial CT, 85% had cisternal compression and 36% had >5mm midline shift

Results – Clinical Outomes

HR for death at 6mos =1.10, slightly in favor of ICP group

Results – Subgroup Analysis

Hospital LOS was slightly shorter in the ICE group (iqr 12 for ICP, 9 for ICE)

No significant differences in MV days, of non-neurologic complications Except ICP-monitored pts had a higher incidence of decubitus ulcers

(12%vs 5%, P=0.03)Median time of ICP monitoring was 3.6dIncidence of Neuroworsening after randomization was 25%

in the entire study, and was similar in both groupsMedian interval for brain-specific treatments was longer in

ICE groupUse of barbiturates was significantly higher in the ICP

group (24% vs 13%)HTS and HV were used more in the ICE group (72% vs

58%, 73% vs 60%)

Results

Almost every variable, including LOS, mortality, anf functional outcomes favored ICP monitoring with a HR>1.

The study was powered to detect statistical significance of HR>1.5

Subgroup analysis of HR accounting for Marshall CT Classification

Results Summary

Composite endpoints between the two groups were similar (P=0.49) ICP group = 56 ICE group = 53

Mortality at 6 months (P=0.06) ICP group=39% ICE group=41%

ICE group had more days of brain specific treatments (hyperosmolar therapy, HV)

Discussion

So what did the trial show?Clinicians act on ICP, without clinical correlate as

evidenced by the increased use of barbituratesClinicians also act on clinical findings without

quantitative evidence of intracranial hypertension as evidenced by more brain-specific treatments overall in the ICE group.

Is this because increased ICP could herald clinical changes and early interventions abort herniation events?

Also, radiographic signs may not translate to the parenchymal monitor.

Discussion - Skepticism

South America – differences in pre-hospital, and post-hospital care

Less might survive to hospital or to hospital transferRehabilitation standards are different and may not

translate to the same cognitive recovery35%death in all groups after 14dAdjusted estimates of sTBI mortality in the US varies

from 41%-25% (J Neurotrauma. 2012 Jan 1;29(1):47-52., J Neurotrauma. 2012 Jan 1;29(1):47-52.)

The “Thereapeutic Intensity Level” is a good overall metric but others such as %responders to ICP-lowering therapy has proven predictive and should have been incorporated (J Neurosurg. 2011 May;114(5):1471-8)

Discussion - Skepticism

Technology – parenchymal monitors as standard

Triggers for treatment? – ICP>20 x 5 min vs radiographic signs with or without clinical correlates

ICP group – ICP triggers ICP is too simplistic a reflection of intracranial pathophysiology. No account for CPP

Clinical signs don’t always reflect global pressures and vice versa

No discussion of inclusion/exclusion of polytrauma and surgical interventions

Discussion - Skepticism

Variability in treatments (i.e. more mannitol and HV in the ICE group) may be because the ICE group had scheduled scans and interventions and the ICP group had more event-related treatment triggers

Conversely, that may explain why the ICP group had more barbiturates and HTS

Discussion - benefits

Very rigorous treatment algorithm for management of elevated ICP (either qualitative or quantitative)

Homogenous population across countriesBoth groups had intracranial HTN treated…

that isn’t in questionIn truth, this study did not test ICP

monitoring, only a very specific treatment algorithm to an ICP threshold compared with clinical exam

In the end, the neurologic exam might STILL be the best tool in our disposal.

Further Discussion

This was probably the only way that this type of trial could be done

Authors were careful not to compare South American patients to our own, only report their findings

The goal of therapies was to lower the average ICP within the head – this doesn’t accurately reflect mechanical compression and injury to diepnephalic which may portend a worse prognosis. The clinical signs of elevated ICP (pupil dilations, posturing, coma) are directly related to these areas.

Further Discussion

How would you alter the study?Could multimodal monitoring be the next

step?Are composite endpoints more useful that

single variable? i.e. mortality? Return to work?

Is a 6-month outcome long enough?

Don’t forget that even the most rigorous study cannot account for all possible variables and that this data might not apply to every patient.