Spontaneous intracerebral hemorrhage is a major cause of death and disability all over the world [ 1 ]. Emergency hematoma removal with decompressive craniectomy (DC) surgery has been widely used in these patients as a primary intervention, or as an auxiliary intervention when first-line treatment fails [ 1 , 2 , 3 ]. An optimal intracranial pressure (ICP) management is critical in perioperative care. Several studies have indicated that DC can relieve hyperintracranial pressure and enhance cerebral perfusion, which can improve long-term neurofunctional outcomes [ 4 , 5 , 6 ]. However, two previous randomized clinical trials have failed to demonstrate that DC can improve the prognosis at 6 months after traumatic brain injury [ 7 , 8 , 9 ]. In addition, several studies have put forward opposite conclusions on the effectiveness of ICP monitoring [ 10 , 11 , 12 ]. Therefore, whether ICP real-time monitoring can improve the prognosis of patients undergoing DC is still controversial, and importantly, it might be the target-specific ICP management that makes the difference beyond ICP monitoring. In this study, we mainly focused on the value of different parameters in perioperative ICP monitoring for predicting mortality of patients with spontaneous intracerebral hematoma undergoing emergent hematoma removal and DC, to identify potential ICP target for improving prognosis.

The primary purpose of cerebral hemorrhage surgery is craniotomy decompression, followed by the removal of hematoma to relieve local mass effect [ 16 ]. However, the effect of DC on the prognosis of severe cerebral hemorrhage remains controversial [ 4 , 5 , 6 , 7 , 8 , 9 ]. It should be noted that ICP monitoring is one of the core indicators of perioperative treatment of patients with cerebral hemorrhage. In this study, we summarized 44 patients with ICP monitoring and 132 patients without ICP monitoring in our institution. Without an identified ICP management target, we found no statistically significant differences in perioperative mortality between the ICP group and non-ICP group, as was the disability rate at discharge. A higher peak ICP value, rather than a continuous ICP of more than 25 mmHg, was an independent predictor for perioperative death, and the cut-off value of peak ICP value for predicting mortality was found as 31 mmHg.

Whether DC can improve the prognosis of patients with severe cerebral hemorrhage has been controversial. The guideline for TBI proposed by AANS in 2016 indicated that for patients with diffuse severe TBI (without local occupying lesion) with ICP > 20 mmHg and cumulative time > 15 min/h and failing first-line treatment, dual frontal DC cannot improve prognosis, but can reduce ICP and shorten the length of intensive care unit (ICU) stay [ 16 ]. One prospective, multicenter, randomized controlled trial (RCT) confirmed that for adult patients with diffuse severe TBI and refractory intracranial hypertension, early DC can reduce ICP and shorten the length of ICU stay, but it cannot reduce the mortality of 6 months after injury, and the proportion of patients with poor prognosis (GOS score 1-4) was increased at 6 months after injury [ 7 ]. However, the study was widely questioned because of its design flaws, such as the broader inclusion criteria (ICP > 20 mmHg and cumulative time > 15 min/h), and significant baseline differences between groups, etc. Another subsequent RCT study RESCUEicp adopt different inclusion criteria based on ICP value (ICP > 25 mmHg and cumulative time > 1-12 h) and indicated that DC can reduce the mortality at 6 months and 12 months after injury, especially significantly reducing the early mortality after injury. However, the proportion of lower severe disability and upper severe disability increased, and the proportion of moderate disability and good recovery was similar with conservative management [ 17 ]. In this study of spontaneous intracerebral hemorrhage, the perioperative mortality was 20.5%. 28.4% were mild coma, 21.0% were moderate coma, and 30.1% were severe coma. 29.0% have deteriorated GCS scores at 2 weeks after the operation compared with admission.

DC is an emergency response to severe intracranial hypertension, and the control of ICP and restoration of cerebral perfusion pressure is an important content in the perioperative treatment. However, the ICP management target is still poorly understood, and even the ICP threshold in the case of cerebral hernia is still unknown [ 18 ]. In clinical practice, doctors usually need to make a comprehensive evaluation on the clinical manifestations (such as cognitive state, pupil), imaging manifestations (such as midline shift, brain edema, or swelling), ICP monitoring (if any), to decide whether DC is necessary [ 19 ]. On the other hand, postoperative intracranial pressure monitoring is helpful for clinical workers to adjust the dosage of dehydrating drugs such as mannitol in real time based on the ICP value. Moreover, real-time monitoring of ICP after operation is helpful to timely find the changes of the condition and effectively reduce the incidence of serious complications such as electrolyte disturbance and renal injury after operation [ 20 ]. However, one previous study reported no effects on ICP monitoring patients whatever mortality or functional outcome [ 21 ]. In this study, we found no difference in the perioperative prognosis between the ICP group and non-ICP group, which indicated that the management protocol beyond monitoring alone might improve the prognosis. Therefore, a ICP management target should be explored based on the association between ICP monitoring parameters and patient outcomes. One recent study proposed that neurocritical care management of patients using multimodal monitoring can improve outcomes [ 22 ]. In this study, the peak ICP value was found statistically significantly associated with perioperative death, which means that strict control of the peak ICP value may reduce perioperative mortality. In previous studies, the reported cut-off ICP for DC in refractory high ICP is different [ 23 ]. In this study, the cut-off ICP was preset to 25 mmHg. However, an ICP value > 25 mmHg was found no significant association with perioperative death. The further ROC consistency test showed the cut-off value of peak ICP value was 31 mmHg, which means keeping the peak value of ICP below 31 mmHg at all times may be more meaningful than other measures.

According to the position of the ICP monitoring probe in the cranial cavity, it can be divided into ventricular monitoring method, brain tissue monitoring method, and subdural or epidural monitoring method [ 24 ]. Intraventricular monitoring is the most accurate method. However, in clinical practice, the ventricle on the bleeding side is often compressed to become smaller or displaced, resulting in the difficulty of placing the ICP probe. Repeated puncture may lead to new intracranial hemorrhage and increase the risk of postoperative complications. In this study, the ICP monitoring probes were placed intracerebral to minimize shift and monitoring inaccuracy. However, changes in postures (sitting position or supine position) tend to lead to large variations in ICP values. One recent study confirmed that the postural ICP difference remained constant at around 8 mmHg (sitting position and supine position) [ 25 ]. In this study, we collected ICP values measured in supine position. Besides, most of the severe intracranial hemorrhage patients showed irritability during the perioperative period, which can lead to a higher ICP. Therefore, sedation and analgesia may be necessary for these patients to control ICP [ 22 , 26 , 27 , 28 ].

A high ICP peak value might provide a better prediction for the mortality of patients with spontaneous intracerebral hemorrhage evacuation and DC, suggesting a tailored ICP management protocol to decrease ICP peak value.