Intensive care management of arterial carbon dioxide in acute intracerebral haemorrhage: Case report of influences on cerebral haemodynamics

Case report

A 72-year-old man with a history of alcoholism and hypertension developed sudden onset headache following a short period of altered level of consciousness. The patient presented to hospital unconscious, one hour after symptoms started, with a Glasgow Coma Scale of 6 (range 3–15), with arterial hypertension (160/90 mmHg), though pupil and corneal reflexes were present and symmetric. A non-contrast head computed tomography (CT) scan demonstrated a significant right-sided caudate intracerebral haemorrhage (ICH) with ventricular extension and hydrocephalus (ICH score 4, Figure 1). ¹ Importantly, the patient’s coagulation profile was within the normal range.

OPEN IN VIEWER

Figure 1.

Axial non-contrast computed tomography images demonstrating an acute haemorrhage in the temporal lobe which has spread to the lateral, third and fourth ventricles and associated right-to-left midline shift.

At admission to the Intensive Care Unit (ICU), the patient was without analgosedation and hypertensive (160/90 mmHg) without vasoactive drugs. The patient did not have iatrogenic coagulopathy. He was intubated with mechanical ventilation, analgosedation was initiated, and arterial and venous catheters were placed. The central venous catheter was inserted in the jugular venous site under ultrasound guidance. Ultrasound of optic nerve sheath diameter (ONSD) using a 12 MHz linear transducer (SonoSite M-Turbo®, EUA, Figure 2) and TCD were performed (Table 1, Figure 4(c)) by a trained intensivist physician (JRC) with experience in neurocritical care. Cerebral blood flow velocity (CBFV) in both middle cerebral arteries was recorded with 2MHz pulsed range-gated transducers (DWL, Dopplerbox, Germany) held in place with a head frame, at insonation depths from 50 to 55mm, with slight anterior angulation (15–30°) of the transducer through the temporal window (Figure 3). At the time, dynamic cerebral autoregulation (CA) was impaired with an autoregulation index (ARI) of 3.9 by transfer function analysis (Table 1).² The ARI ranges from 0 to 9, with normal values around 6±1. The respiratory rate of mechanical ventilation was manipulated to adjust blood pressure and PaCO₂. Thereafter, the patient received an external ventricular drain (EVD) and placement of an intracranial pressure monitor.

OPEN IN VIEWER

Figure 2.

Ultrasonographic image of abnormal optic nerve sheath diameter (ONSD) measurement. The probe was adjusted to give a suitable angle for displaying the entry of the optic nerve into the globe and measurement was performed at the depth of 3 mm behind the ocular globe (A). Right and left ONSD were measured in the transversal plane, with slight rotation of the probe to obtain optimal optic nerve visualization. Distance B (between the yellow crosses) is the ONSD (0.69 mm), suggestive of intracranial hypertension.

OPEN IN VIEWER

Table 1.

Peripheral and cerebral haemodynamic parameters.

Variables	Admission ICU	24-hour ICU	High CO2PaCO2 > 40 mmHg	LOW CO2PaCO2 < 30 mmHg
Systolic CBFV, cm/s	80.9	73.4	99.7	90.9
Diastolic CBFV, cm/s	24.8	19.0	33.1	30.4
Mean CBFV, cm/s	43.8	33.5	57.3	52.7
Pulsatily index	1.3	1.5	1.2	1.1
ARI	3.9	6.8	4.8	6.0
CrCP, mmHg	21.2	34	45.12	24.14
RAP, mmHg.s/cm	2.07	1.92	1.08	1.55
HR, bpm	100.5	50.6	46.3	47.1
MAP, mmHg	111.4	98.7	106.9	105.5
ETCO2, mmHg	36	26	36	22
pH	7.348	7.406	7.359	7.424
PaO2, mmHg	148.3	125.7	138.1	161.0
PaCO2, mmHg	42.6	32.6	42.5	28.1
BE	−2.7	−4	−5	−1.7
SatO2	98.2	97.2	99.1	99.1
Lactate mg/dl	9.1	9.4	13.2	11.6

ICU: Intensive Care Unit; CBFV: cerebral blood flow velocity; ARI: autoregulation index ² ; CrCP: critical closing pressure; RAP: resistance area-product; HR: heart rate; MAP: mean arterial pressure; ETCO₂: End-Tidal Carbon Dioxide: PaO₂: arterial oxygen tension; CO₂: carbon dioxide; BE: base excess; SatO₂: arterial oxygen saturation.

OPEN IN VIEWER

Figure 3.

Screen shot of transcranial Doppler showing the ultrasound spectrogram of cerebral blood flow velocity (CBFV) for the left (top left box) and right (top right box) MCAs, with the maximum velocity envelope represented by the white tracing. CBFV in the right MCA was reduced, compared to the left side due to the intracerebral hemorrhage in the right hemisphere. The bottom box allows comparison of the CBFV tracings (blue and red) with the ECG and the arterial blood pressure (BP) waveform (yellow tracing). Both BP and CBFV are needed to obtain estimates of the autoregulation index (ARI).

Twenty-four hours after ICU admission and within the critical care setting, new TCD measurements were performed (Table 1). ARI was 6.8 (normal compared with healthy subjects), but due to low values of end-tidal CO₂ (ETCO₂), it was decided to increase ETCO₂ to achieve normocapnia. As different targets of ETCO₂ were attempted, it was noted that this had a strong influence on cerebral haemodynamics and ARI (Table 1; Figure 4(a), (b) and (d)).

OPEN IN VIEWER

Figure 4.

Longitudinal changes in (a) end-tidal carbon dioxide (ETCO₂) and its effects on (b) cerebral blood flow velocity (CBFV) as a function of time; (c) normalized CBFV responses to a step change in blood pressure at admission to the ICU (continuous line) and at 24 hours (dashed line); and (d) normalized CBFV responses to a step change in blood pressure for assessments during manipulation of CO₂, low CO₂ (PaCO₂ < 30 mmHg, dashed line) and high CO_{2 (}PaCO₂ > 40 mmHg, continuous line).

Discussion

This patient suffered a significant ICH, evidenced by the size of the haematoma, GCS and ARI at ICU admission, and intracranial complications, including intraventricular extension and hydrocephalus. The clinical course involved surgical intervention (EVD) aimed to minimise the detrimental effects of raised ICP. The presence of intraventricular haemorrhage is an independent poor prognostic marker in ICH. ¹ ICH is fairly common and has devastating consequences, accounts for 20% of cases of stroke; however, it is associated with the greatest rate of mortality; 1-year survival from ICH is approximately 40%. Furthermore, the rate of functional independence at follow- up varied from 12% to 39%. ³ Hypertension is the primary risk factor for ICH; in addition, heavy alcohol consumption has been associated with increased ICH risk. Our patient had both risk factors. ⁴

Current international guidelines recommend invasive intracranial pressure monitoring in patients with large intraventricular haemorrhage, hydrocephalus, transtentorial herniation and those with a low level of consciousness. ³ Both, non-invasive strategies (head position and pharmacological agents) ¹ and invasive strategies to manage raised intracranial pressure remain limited and associated with high morbidity and mortality. ¹ However, novel approaches to managing impaired dynamic CA in ICH using hyperventilation-induced hypocapnia are in progress. ⁵ TCD assessment of cerebral blood flow in acute stroke has permitted significant advances in our understanding of CA impairment and its relationships with stroke severity, stroke recovery ⁶ and radiological assessment of haematoma. ⁷ The ARI is a non-invasive metric to assess the effectiveness of dynamic CA and one main advantage of the ARI is that it uses all the information in the gain and phase frequency responses calculated by transfer function analysis. The value of ARI = 3.9 in this patient suggests impaired CA, considering that healthy individuals have values around 6 ± 1.^6,8 The value of CA assessment in acute ICH is found in the ability to evaluate the dynamic impact of blood pressure lowering therapies on cerebral perfusion, to detect situations where the brain is at risk of hypo- or hyperperfusion, and to establish associations with clinical outcome. This is crucial as randomised controlled trials to date have failed to clarify whether intensive versus standard blood pressure lowering is indeed a viable intervention for ICH patients with large haematoma volume requiring ITU admission. ⁹

Interestingly, this case demonstrates that beyond pharmacological and surgical interventions, simple manipulation of PaCO₂ has the propensity to improve cerebral haemodynamic parameters. In subarachnoid haemorrhage, we are aware that spontaneous hyperventilation exists and is associated with delayed cerebral ischaemia and poor neurological outcome. However, in a non-vasospasm driven pathological process like acute ICH, where mass effect and evolving oedema and acute inflammatory processes drive deterioration, the potential value of vasoconstriction could be clinically beneficial.

CO₂ is a potent vasodilator in the brain that can cause hyperaemia, as has been shown in Table 1. Patients with ICH, high blood pressure and hypercapnia can have expanded haematomas. Furthermore, previous studies have shown the negative impact of hypercapnia on CA, ¹⁰ confirmed in this patient with the worsening of CA, as expressed by the ARI, with high PaCO₂. Nowadays, normocapnia is recommended; however, different targets may be helpful for a carefully selected group of patients, such as patients with impaired CA.

The evaluation of dynamic CA in an ICU setting has been previously reported. ⁸ This case corroborates previous reports of the many benefits of having TCD available in the ICU as an easy‐to‐use, safe, and low‐cost bedside tool.^11,12 This tool allowed us not only to observe changes in CBFV but also monitor CA at the bedside. Although the CA was normal 24 hours after the admission and critical care treatment, other variables, such as PaCO₂, can modify CA and it may have a subsequent impact on the patient’s clinical outcome.

Of considerable relevance is the clinical utility of ultrasound-acquired ONSD in the ICU. This parameter has been studied in different populations and clinical presentations associated with intracranial hypertension and a recent meta-analysis about this topic found good accuracy in detecting ICP > 20 mmHg or > 25 cmH₂O. ¹³ In our case, the highly probable aetiology of intracranial hypertension was the presence of hydrocephalus. TCD, cerebral haemodynamic assessment at the bedside and sonographic ONSD have the potential advantages of being repeatable, portable, safe and low-cost tools, with no radiation or side effects.

Conclusions

In this patient with ICH, we demonstrated the relevance of TCD monitoring and its ability to reflect changes in CA linked to alterations in PaCO₂, as well as the additional benefits of ONSD in a neurocritical care setting. Our results demonstrate the negative effects of hypercapnia on CA and the benefits of having TCD available in the ICU point of care. Clearly, formal clinical trials are needed to produce guidelines that will permit extending this approach to routine clinical management of such patients.