Patient data for 151 consecutive TBI who received bilateral DC from Jan. 2008 to Jan. 2022 were prospectively in a database and retrospectively reviewed. This study was approved by the ethic committee of Huashan Hospital Fudan University. Patients with TBI who were admitted to the Department of the Neurosurgery Neurotrauma Center at Huashan Hospital Fudan University were included. Informed consent was obtained from all individual participants. If a patient could not sign informed consent by himself/herself, such as patients with a Glasgow Outcome Scale (GOS) score of 1, informed consent was signed by his/her statutory guardian.

Inclusion criteria are as follows. First, participants had to be computed tomography (CT) scan confirmed patients with TBI. CT signs of TBI included epidural hematoma, subdural hematoma (SDH), intraparenchymal hemorrhage (IPH), brain contusion, and brain laceration. Second, patients had to be > 18 years of age. Third, patients received bilateral frontal-temporal-parietal DC at discharging. Finally, participants had to be admitted within 6 h after injury. Patients with TBI with traumatic injury to a body region other than the brain with an Abbreviated Injury Severity score > 3, with penetrating brain injury and those already received unilateral DC in other hospitals were excluded. However, if a patient only received ICP monitoring implanted in other hospitals at admission, the patient would be enrolled (Fig. 1).

Baseline characteristics, including age, sex, mechanism of injury, pupillary reaction to light, GCS score at admission, and type of injury, were recorded for all patients. Injury types were assessed based on initial CT scan on admission. Coagulopathy was defined as regular coagulation test results meeting one or more of the following criteria: platelet counts (PLT) < 100 × 109/L, international normalized ratio (INR) > 1.25, prothrombin time (PT) > 14 s, or activated partial thromboplastin time (APTT) > 36 s as our previously studies [14-16]. Peripheral blood analysis was performed for all patients within 6 h of injury at the Central Clinical Laboratory in Huashan Hospital.

The frontal-temporal DC was performed as previously reported to provide maximal decompression [5]. The area of bone flaps was 12*15 cm2 was performed as Guidelines for the Management of Severe Traumatic Brain Injury [7, 17]. If bilateral DC were performed, a strip of midline bone bridge covering the superior sagittal sinus would be left [5]. During the DC process, intracerebral lesions, including hematoma, contused or lacerated brain tissue and et al. were removed as totally as possible, in addition, de-stretching duroplasty was applied. ICP probe and intraventricular drainage system were implanted into lateral ventricle as previously reported. For indications of ICP implantation varied in the duration of this retrospective research, whether ICP monitoring was applied was decided by neurosurgeons who performed DC. If ICP probes were applied, it was preformed accompanied with or prior to 1st DC, and no ICP iprobes were implanted between initial DC and unplanned contralateral DC.

According to different operation plans. we divided all bilateral DC patients into four categories: (1) Pre-operative scheduled bilateral DC: bilateral lesions were detected at admission. (2) Intraoperative scheduled contralateral DC: Unilateral lesions were identified at admission and received unilateral DC. But according to intraoperative ICP values, abnormal brain swelling and et al, these patients were suspected to suffer contralateral lesions. After confirmed by intraoperative CT scans, these patients received immediately contralateral DC. (3) Post-operative immediately contralateral DC patients: Unilateral lesions were identified at admission, and after initial DC, no abnormal signs, symptoms or intraoperative CT scans findings were detected during operations. Abnormal high ICP value (> 20 mmHg) were detected before these patients were sent back to NICU. These patients received immediately CT scans and contralateral lesions were detected, and then these patients were sent back to operation rooms without delay to receive contralateral DC. (4) Unplanned contralateral DC patients: These patients were safely sent back to NICU after unilateral DC. Contralateral lesions were identified and contralateral DC were performed during hospitalization. These patients were identified as unplanned secondary surgery (USS) patients in this study.

The patients were treated in accord with the latest guidelines, but detailed therapeutic approaches were determined by neurosurgeons who performed operations. If ICP implantation were applied, the cerebral perfusion pressure (CPP) was maintained at 75–90 mm Hg at all times by keeping the mean arterial pressure at 90–100 mm Hg and the ICP at < 20 mm Hg. Systolic blood pressure was required to be maintained under 140 mmHg. Corticosteroids were not used. Body temperature, respiratory rate, heart rate, blood pressure, cardiac rhythm, and oxygen saturation were monitored continuously. Serum glucose, blood gas, and serum electrolyte values were measured regularly and kept within normal range. Intraoperative CT scans were performed before operations ending or when abnormal brain swelling, abnormal ICP values and et al. occurred. Postoperative CT scans were routinely performed 24 h, 72 h and 7 days after operations or when neurological deficits occurred. All patients were evaluated and treated by full-time neurosurgeons with specific training in critical care.

A specialist in physical medicine and rehabilitation determined the neurological outcome at 3 months after injury. The primary outcome of patients with TBI was assessed using the mortality (Grade 6) and functional outcome at 90 days after admission by using the modified Rankin scale (mRS) as our previous studies [14], through outpatient interviews or over the telephone. mRS score of 4–6 was considered a poor outcome, and mRS score of 0–3 was considered a good outcome. The rates of operation, length of ICU stay and the rate of serious adverse events, including kidney dysfunction, brain infarction and et al., were collected at discharging.

Continuous variables were expressed as mean ± standard deviation (SD) or median (interquartile range), and categorical variables were expressed as percentages. The univariate analyses of categorical data were performed using the χ2 test. Equality of variance was assessed using the Levene test. Normally distributed variables were compared using Student t test or one-way analysis of variance, whereas nonnormally distributed variables were compared using the Kruskal–Wallis or Mann–Whitney U test.

After the univariate analyses, a forward stepwise logistic regression analysis of the 3-month outcome was used to develop the prediction models and adjust for multiple predictors of 3-month outcome. All statistical tests were 2-tailed, and P < 0.05 was considered statistically significant. Statistical analysis was carried out using SPSS 23.0 (IBM, Armonk, New York, USA) and MedCalc statistical software (version 15.2.2, MedCalc Software bvba, Ostend, Belgium).