The Role of Hyperventilation in ICP Control

Management of arterial partial pressure of carbon dioxide in the first week after traumatic brain injury: results from the CENTER-TBI study

Written by – Dr Jose Chacko & Dr Swapnil Pawar

Context – 

Recent BTF Guidelines suggested the use of Hyperventilation (HV) to target PaCO₂ 32–35 mmHg for controlling ICP only as a second-tier treatment. These guidelines did not support lower PaCO₂ levels and recommended against routine hyperventilation to PaCO₂ below 30 mmHg.

However, the practices vary across the globe.

Objectives – 

1. to assess, in a real-world context, PaCO₂ management and the lowest target of PaCO₂ in a large cohort of mechanically ventilated TBI patients and
2. To assess practice variability between centres to evaluate the association between the use of profound HV and 6-month clinical outcomes.

Setting and design

The CENTER-TBI core study is a prospective observational study across centers in Europe that addressed the epidemiology, management and outcomes after TBI, conducted over a 3-y period between 2014–2017. The current study was performed to evaluate the management of PCO2 in patients with severe TBI, and optimal PCO2 targets in relation to clinical outcomes

Subjects

1176 patients with TBI from 51 centers who were intubated, ventilated and had at least 2 arterial PCO2 values in the first 7 days. From this cohort, 1100 patients were included from centers that enrolled at least 10 patients. The median age was 48 (29–64) years, 64.7% of patients had severe TBI (GCS 8 or less). The baseline PCO2 was 39.1 mm Hg. ICP monitoring was carried out in 68% of patients.

To adjust for confounders, the propensity of centers to use profound hypocapnia was evaluated measured as the relative proportion of PCO2 <30 mm Hg. This was included in a logistical regression model.

Results

The PCO2 levels in the first week varied widely between centers and ranged between a mean level of 32.3 – 38.7. Mean lowest PCO2 in the first week: 35.2 ± 5.4.

Relatively few patients [144, (13%)] patients had all PCO2 between 35–45 (relatively small number of patients who were maintained between the recommended range)

Wide variation was observed in the use of HV

There was a 1.72 fold difference between centers for odds of PCO2 between 35–45.

After exclusion of patients with intracranial hypertension, the difference in the median odds ratio decreased (MOR) to 1.4

Patients who underwent profound HV (PCO2 <30) in the first 7 days ranged between 1–30% between centers

At least one episode of profound hyperventilation (HV) was observed in 36% of patients during the first week. There was considerable variation between centres in the use of profound HV (MOR of 2.0). The use of HV was greater in patients with intracranial hypertension (odds of HV 3 times higher). The use of HV decreased from days 1–7.

Other monitoring modalities including jugular bulb oxygen saturation and brain tissue oxygenation were used infrequently.

Profound HV is associated with more aggressive treatment – as evidenced by a higher therapeutic intensity level (TIL). Patients who underwent profound HV also more likely to undergo decompressive surgery (8.6 vs. 4.8%), and low or high-dose osmotherapy

Clinical outcomes

165/1100 (15%) died in ICU

970 patients had 6-month outcomes available

775/970 (79.9%) had poor outcomes: death or GOSE 4 or less

246/970 (25.4%) died at 6 months

529/970 (54.5%) GOSE 4 or less

In patients with at least one episode of profound HV:

Mortality at 6 months: 29% with profound HV vs. 23% who did not undergo profound HV (p=0.045)

GOSE 4 or less: 64% vs. 49% (p <0.001)

On instrument variable analysis the propensity to use profound HV did not impact mortality or unfavorable outcome, after adjusting for the level of intracranial pressure. After adjustment for baseline variables, no difference in mortality or 6-month outcomes with the use of HV

Centres that used profound HV more often, (by 10%) had numerically higher odds for mortality, though not significant OR: 1.06; 95% CI: 0.77–1.45, p value=0.7166

For GOSE 4 or less: OR: 1.12; 95% CI=0.90– 1.38, p value=0.3138

Main findings:

• The wide variation between centres regarding PCO2 management
• ICP monitored patients had lower PCO2 levels
• PCO2 levels were lower in patients with intracranial hypertension
• No difference in mortality or 6-month neurological outcomes after propensity analysis 

Comments

• Seattle international severe traumatic brain injury consensus conference (SIBICC) recommends the use of HV as second-tier therapy
• BTF guidelines recommend it only as a temporizing measure for raised ICP
• The study represents an overview of the current management of PCO2 levels in Europe
• Limitations of an observational study, with no adequately matched controls
• Wide heterogeneity in the type of injury that could impact outcomes
• The PCO2 level represents a snapshot in time; the actual levels over a longer duration may vary widely. Capnography may be a better tool to assess trends over time
• Assessment of longer-term outcomes (1 y) may be meaningful in TBI.

Our Recommendations – 

There is not enough evidence to completely abandon the current practice of HV to control ICP.

However, the dogmatic practice of using HV in every single patient with severe TBI should be avoided.

Its use can be considered as a rescue measure in the setting of raised ICP, which is refractory to first-line treatment modalities.

A large RCT is needed to address this controversy further.