Introduction

Primary brainstem hemorrhage (PBSH) is a rare but fatal subtype of intracerebral hemorrhage (ICH), representing approximately 3.4–5.8% of spontaneous ICH cases1,2. PBSH is characterized by acute onset, rapid progression, and early coma, often necessitating mechanical ventilation. It is associated with poor outcomes, including high mortality and disability rates2,3. The severity of PBSH leads to prolonged hospital stays and increased medical resource use, imposing a substantial burden on healthcare systems.

While conservative management may be appropriate for small brainstem hematomas with clear consciousness and a favorable prognosis, larger brainstem hematomas can have mortality rates as high as 30 to 90%4,5,6 and typically require surgical intervention. Active treatment approaches have been shown to improve outcomes in cases with larger hematomas7,8,9. Although such instances are currently isolated, their frequency may increase with advancements in minimally invasive surgical techniques.

Hydrocephalus is a common and treatable complication following spontaneous ICH10,11,12. PBSH can easily extend into the ventricle system and obstruct cerebrospinal fluid circulation, resulting in hydrocephalus. Many patients with larger hematomas often present with hydrocephalus-related complications. According to a systematic review, hydrocephalus develops in up to 30.3% of patients with PBSH13 and has been identified as an independent prognostic factor for PBSH14,15.

PBSH treatment is inactive and passive. External ventricular drainage (EVD) is a formal indication in cases presenting with hydrocephalus. However, current treatment is still conservative intensive care for the little dataset supporting neurosurgical interventions16,17,18. Moreover, EVD alone cannot solve the mass effect and secondary injury of hematoma. Early aggressive surgical treatment, including hematoma removal surgery, may stabilize and shorten the pathological process of brain stem hematoma as soon as possible, providing a solid foundation for patients to withdraw from life system support and enter rehabilitation treatment as early as possible. This study aimed to identify clinical and radiological features associated with the development of hydrocephalus and to evaluate the effectiveness of hydrocephalus treatment in patients with PBSH.

Methods

Clinical data collection and study design

This retrospective study was conducted at a single center between January 2021 and December 2023. This study was approved by the Clinical Research Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine (2024/ETH1161). Due to the retrospective nature of the study, the need for informed consent was waived by First Affiliated Hospital, Zhejiang University School of Medicine. The study was performed in accordance with the relevant guidelines and regulations. Medical information was collected in tabular form, and serial numbers were used instead of patient`s private information. Diagnosis of PBSH was confirmed through computed tomography (CT) or CT angiography (CTA). Patients with non-spontaneous causes, such as trauma, tumors, cavernous hemangiomas, arteriovenous malformations, aneurysms, hemorrhagic cerebral infarction, or coagulopathy, were excluded from this study. Medical records and brain CT imaging from a computerized database were independently reviewed by two clinical research coordinators using structured forms.

Study group and interventions

Patients were divided into hydrocephalus group and non-hydrocephalus group. Hydrocephalus presence was determined from brain CT scans by observing dilation in the temporal horn of the ventricle, enlargement of the third or lateral ventricles, and either shallow or disappeared sulci. Intraventricular hematomas were also referenced for confirmation. Information on risk factors associated with the incidence of hydrocephalus was collected from imaging characteristics: hematoma sites, hematoma volume, ambiens cistern disappearance, and combined hematoma elsewhere, such as basal ganglia, cerebellum and ventriculohematoma. The volume of the hematoma was calculated using Tada’s formula (ABC/2)19.

Individualized surgical interventions were selected based on the following factors: (1) hematoma volume, (2) resolution of cisterna surrounding the brainstem, (3) extension of the hematoma into the ventricle, (4) presence of hydrocephalus, (5) severity of the patient’s condition, and (6) family preference. Stereotactic puncture drainage of hematoma (SPDH) and/or EVD were employed as surgical methods. The Medtronic S8 opt-magnetic integrated navigation system guided puncturing brainstem hematoma and the lateral ventricle. For EVD, a ventricular catheter (Phoenix external drainage set) was placed on the frontal horn of the right lateral ventricle. Microbiological and biochemical analysis of cerebrospinal fluid (CSF) samples is carried out daily. The velocity of CSF is regulated by adjusting the height of the drainage device. For SPDH, the navigation system was used to place the ventricular catheter into the hematoma stereotactically. The design principle of the puncture path is that the entry point is selected on the convex surface of the cerebellum, and the target is located in the center of the long axis of the hematoma; the puncture path should be as short as possible and close to the longitudinal axis of the hematoma. After postoperative CT scanning, the hematoma was aspirated with a syringe and then dissolved by injecting urokinase into the hematoma cavity once a day. The drainage catheter was removed according to CT scanning results. Patients with PBSH were categorized into different groups based on their respective surgical approach, with non-operative patients serving as the control group: (1) control group, (2) SPDH group, (3) EVD group, and (4) SPDH + EVD group. The effectiveness of the operative was evaluated by hydrocephalus remission and prognosis assessment.

Outcome assessment

Indicators of disease severity included the disturbance of consciousness at admission, cardiac arrest and cardiopulmonary resuscitation (CPR), endotracheal intubation, and ventilator-assisted breathing. Consciousness was assessed using the Glasgow Coma Scale (GCS), with unconsciousness defined as the absence of voluntary movements, unresponsiveness to verbal and external stimuli, and a GCS score below nine. For patients who underwent SPDH or EVD, the duration of drainage was used to evaluate the operation’s effectiveness. Patient prognosis was assessed based on 30-day mortality (mortality within 30 days of symptom onset), successful ventilator withdrawal (de-ventilator), and length of hospital stay (in-hospital days). Upon discharge, patients were either monitored at home or referred to a rehabilitation hospital for further treatment.

Statistical analysis

Statistical analysis was performed using R language (version 4.2.2). Univariate analyses were performed using Chi-square tests or non-parametric rank sum tests to determine: (1) the relationship between hydrocephalus and indicators of disease severity, (2) the relationship between hydrocephalus and prognosis indicators of PBSH, (3) risk factors of hydrocephalus, and (4) the impact of different surgical treatments on hydrocephalus remission and prognosis. Multivariate logistic regression analysis identified disease severity indicators and prognostic factors related to hydrocephalus, as well as the best surgical treatment for hydrocephalus in patients with PBSH. Fisher’s exact test was used to examine differences in prognosis between surgical treatment groups, considering factors such as age and sex. Statistical significance was set at p < 0.05.

Results

Prognosis analysis of patients with PBSH

Of the 169 patients with PBSH enrolled in this study, 140 were men and 29 were women, with an average age of 52.7 years (range 33–91 years) (Table 1). Severe PBSH often presents with a high incidence of coma. At admission, 62 patients were conscious, while 107 (63.3%) were unconscious. Eight patients experienced cardiac arrest and were successfully resuscitated with CPR. Ninety-three patients required endotracheal intubation and ventilator-assisted breathing upon admission. Other common symptoms included sudden headaches, dizziness, nausea, vomiting, convulsions, and limb weakness or numbness.

Table 1 Clinical characteristics of patients without hydrocephalus versus patients with hydrocephalus.
Full size table

Twenty patients (11.8%) died within 30 days of symptom onset. The remaining patients were discharged home or transferred to local or rehabilitation hospitals. The average length of hospital stay was 12.5 ± 7.2 days. By the 30th day of admission, 66 of the 107 patients (61.7%) who received endotracheal intubation and ventilator support still required ventilator assistance, while 41 (38.3%) patients were successfully weaned from ventilator support, with a mean de-ventilator time of 9.2 ± 4.8 days.

Neuroradiological findings among patients with PBSH

Radiological characteristics are detailed in Table 1. Pons hemorrhage was the most common, occurring in 97 (57.4%) cases, including basal hemorrhage (22 cases), tegmental hemorrhage (38 cases), and central hemorrhage (37 cases). Isolated midbrain and medulla hemorrhages were less frequent, with two and four cases, respectively. Additionally, 39.1% of cases involved multiple parts of the brainstem: midbrain-pons (66 cases) and midbrain-pons-medulla (one case). Massive bleeding extended to the basal ganglia in 11 cases and the cerebellum in nine cases.

The occupying effect of PBSH often led to the disappearance of the cisterna surrounding the brainstem, with the disappearance of the annular cisterna being the most common, occurring in 70 cases. Among the 169 patients, 61 had hematomas that penetrated the ventricular system, including the third ventricle (four cases), fourth ventricle (21 cases), both third and fourth ventricles (11 cases), third and lateral ventricles (10 cases), and the entire ventricular system (15 cases). The average hematoma was 9.1 mL (range 0.3–36.7 mL). Of these, 83 cases had a volume ≤ 6.1 mL, and 86 cases had a volume > 6.1 mL.

Hydrocephalus prevalence in patients with PBSH and its risk factors

In this cohort of 169 patients with PBSH, 77 (45.6%) patients developed hydrocephalus. Hydrocephalus aggravated the condition, with 44% of patients being unconscious, 39.6% requiring endotracheal intubation, and 43.8% needing ventilator support. Univariate analysis showed that patients with hydrocephalus had significantly more severe disease compared with those without hydrocephalus, as indicated by higher rates of unconsciousness, endotracheal intubation, and ventilator support (p < 0.001). There was no significant difference in the incidence of cardiac arrest between patients with and without hydrocephalus (p = 0.33).

Hydrocephalus also affects the prognosis. The 30-day mortality rate for patients with hydrocephalus was significantly higher compared with those without hydrocephalus (23.4% vs. 2.2%, p < 0.001). Conversely, the de-ventilator rate was significantly lower than in those without hydrocephalus (29.7% vs. 57.6%, p < 0.01). Multivariate logistic regression analysis revealed that hydrocephalus was associated with a low de-ventilator rate (odds ratio (OR) = 0.27, 05%CI: 0.084–0.78, p < 0.05), indicating that patients with hydrocephalus were less likely to be weaned off ventilator support.

As a direct complication of PBSH, risk factors for hydrocephalus were examined mainly through imaging findings of PBSH. Univariate logistic regression, adjusted for age and sex, identified midbrain hemorrhage, tegmental pons hemorrhage, annular cisterna disappearance, and hematoma extension into the cerebellum as increasing the risk of hydrocephalus. Additionally, hematoma penetration into the ventricles also heightened the risk. The RCS curve demonstrated a linear relationship between hematoma volume and hydrocephalus (Fig. 1). Compared with the reference group with hematoma volume ≤ 6.1 mL, a volume > 6.1 mL was associated with a significantly increased risk of hydrocephalus (Table 2).

Fig. 1
figure 1

RCS curve of hematoma volume and hydrocephalus in patients with PBSH.

Full size image
Table 2 Comparison of hydrocephalus risk factors in patients with PBSH on admission.
Full size table

Variables with p < 0.01 in univariate logistic regression were included in the multivariate logistic model for further analysis. The results are shown in Table 3. Midbrain hemorrhage, tegmental pons hemorrhage, annular cisterna disappearance, hematoma involving the cerebellum, hematoma penetrating the ventricle, and hematoma volume were highly reliable predictors of hydrocephalus development. The receiver operating characteristic (ROC) curve, based on this model (area under the curve (AUC) = 0.988, 95%CI: 0.978–0.998), indicated an optimal cutoff value of 0.350, with specificity = 0.924, sensitivity = 0.961, and Youden index = 0.885 (Fig. 2).

Table 3 Multivariate logistic regression analysis of risk factors of hydrocephalus in patients with PBSH.
Full size table
Fig. 2
figure 2

ROC curve of risk factor regression model of hydrocephalus in patients with PBSH.

Full size image

Radiological features of hematoma described above were classified into three categories: (1) multiple hematoma sites, (2) presence of intraventricular hematoma, and (3) hematoma volume > 6.1 mL. These risk factors were included in a multifactorial risk study for hydrocephalus. Logistic regression analysis showed that patients with PBSH having ≥ 2 risk factors had a significantly increased risk of developing hydrocephalus compared with ≤ 1 risk factor (OR = 1618.5, 95%CI: 288.1-17515.7, p < 0.001).

Hydrocephalus surgery effectiveness and outcomes in patients with PBSH

The surgical methods used in this study included SPDH and EVD. Of 77 patients with PBSH complicated by hydrocephalus, 47 (61.0%) received surgical treatment. This included six patients in the SPDH group, 20 in the EVD group, and 21 in the SPDH + EVD group. Thirty (39.0%) patients did not undergo surgery (Control group). The average duration of hematoma puncture drainage was 5.0 days (range 4–14 days), while the average duration of external ventricular drainage was 10.3 days (range 4–17 days). Among the 21 patients who underwent combined SPDH and EVD (SPDH + EVD group), two patients ultimately required ventriculoperitoneal (VP) shunt surgery because of the unrelieved hydrocephalus.

Among the 47 patients with PBSH and hydrocephalus who underwent surgical intervention, 30 experienced relief from hydrocephalus, yielding a remission rate of 63.8%. In contrast, only one out of the 30 (3.3%) patients who did not undergo surgery experienced relief from hydrocephalus. There were notable differences in the remission rates of hydrocephalus based on the surgical method employed (Table 4). Specifically, the remission rates were 85.1% in patients who underwent SPDH surgery and 58.5% in those who underwent EVD surgery. Correlation analysis revealed that both SPDH and EVD surgeries were effective in treating hydrocephalus in patients with PBSH (Chi-square test, both p < 0.001). Logistic regression analysis further demonstrated that EVD (OR = 15.6, 95%CI: 2.39–310), and particularly the combination of SPDH and EVD (OR = 106.5, 95%CI: 15.5-2240.4), were the most effective surgical treatments for alleviating hydrocephalus in these patients (p = 0.015 and p < 0.001, respectively).

Table 4 Hydrocephalus remission rate and prognosis in different treatment groups.
Full size table

Surgical intervention significantly improved the prognosis of patients with hydrocephalus. The 30-day mortality rate in the non-surgical (control) group was 46.7%, significantly higher than the 23.3% observed in the surgical group (p < 0.001). Additionally, the de-ventilator rate was only 3.3% in the control group, significantly lower than in the surgical group (p < 0.001). Further analysis revealed significant differences in 30-day mortality and de-ventilator rate among patients treated with different surgical methods (Table 4). The SPDH group had the lowest 30-day mortality (0%) and the highest de-ventilator (66.7%) rates, suggesting that patients with hydrocephalus due to PBSH greatly benefited from hematoma removal surgery.

Discussion

Many studies have sought to identify independent predictors of mortality or functional outcomes in patients with ICH or PBSH20,21,22,23,24,25. The development of hydrocephalus has been recognized as a strong predictor of short- and long-term mortality, as well as poor 90-day outcomes in these patients16,20,26. Hydrocephalus following PBSH is a common and treatable complication, making its management an important aspect of PBSH treatment. However, factors predicting hydrocephalus have been understudied. This study focused on identifying these factors and confirmed that PBSH had a high incidence of hydrocephalus, consistent with previous studies13. In this study, 45.6% of patients with PBSH developed hydrocephalus, which was associated with a more serious condition and poorer outcomes. Unconsciousness, endotracheal intubation, and ventilator support were statistically significant in patients with hydrocephalus. These patients also had a worse prognosis, evidenced by a higher 30-day mortality rate and a lower probability of being weaned off the ventilator. Although not statistically significant, patients with hydrocephalus also tended to have longer de-ventilator times and hospital stays than patients without hydrocephalus. Therefore, it is reasonable to recommend the inclusion of hydrocephalus in PBSH prognostication scoring systems15.

To the best of our knowledge, this is the first study to evaluate risk factors influencing the incidence of hydrocephalus in PBSH. Our study identified unconsciousness, midbrain hemorrhage, pons tegmental hemorrhage, combined cerebellum, combined intraventricular hematoma, and greater hematoma volume as the most significant predictors for the development of hydrocephalus following PBSH. Previous studies identified similar risk factors for hydrocephalus in patients with ICH, including infratentorial hemorrhage, hematoma extension to the ventricles, greater ICH volume, and hematoma expansion27,28,29,30. Our results align with these, indicating that factors such as hematoma size, location, and expansion into the ventricle are critical in both PBSH and ICH for predicting the risk of hydrocephalus.

Understandably, hematoma expansion into the ventricle is a significant risk factor for hydrocephalus in PBSH. The midbrain and tegmental pons are anatomically close to the ventricles and cerebrospinal fluid circulation pathways, making bleeding in these areas more likely to either extend into the ventricular system or press on the pathways, thus increasing the risk of hydrocephalus. The amount of hematoma is another critical factor, differing in PBSH compared with other conditions like ICH. In this study, a hematoma volume threshold of 6.1 mL was identified, indicating that patients with hematoma volumes exceeding this volume have a very high probability of developing hydrocephalus.

Bleeding in the brainstem often extends longitudinally, affecting multiple sites. In our cohort, 36.1% of patients with PBSH exhibited hematoma involving multiple parts of the brainstem, such as pons-midbrain, pons-medulla, midbrain-pons-medulla, and even extending beyond the brainstem to areas like the basal ganglia and cerebellum. This study found significant differences in the location of multiple hematomas between patients with and without hydrocephalus. A prediction model was established based on three risk indicators: hematoma involvement of more than one site, hematoma volume greater than 6.1 mL, and hematoma extension into the ventricles. The study found that hydrocephalus could be predicted in patients with PBSH who had more than one of these risk factors.

In the past, the treatment approach for PBSH was often negative and pessimistic, largely owing to the severe nature of brainstem dysfunction and the high risk of persistent coma16,17. Conservative treatment was associated with poor outcomes, particularly in cases of primary pons hemorrhage, where it was a poor predictor of 30-day mortality4,21. With advancements in minimally invasive neurosurgery, the outlook for PBSH treatment has improved. Techniques like SPDH surgery have gained popularity because of their advantages, including simple and rapid operation, precise positioning, and minimal surgical trauma9,21,22. The adoption of technologies, such as optomagnetic integrated navigation and neurosurgical robots, has further enhanced the precision and efficacy of these procedures, leading to better patient outcomes. This shift towards more proactive surgical interventions has contributed to improved prognosis and reduced reliance on long-term life support systems. Several studies have shown that early, aggressive surgical intervention can significantly reduce the 30-day mortality rate and improve prognosis8,14,31. Our findings corroborate these results, showing that surgical treatment lowers the 30-day mortality rate and enhances the likelihood of ventilator weaning. Patients with hydrocephalus resulting from PBSH benefited from hematoma removal surgery.

Surgical treatment for hydrocephalus in patients with PBSH must address both the hydrocephalus and the hematoma. While EVD is the most common and effective surgical method for managing hydrocephalus7,32, our study found that SPDH can also be effective. EVD may initially relieve hydrocephalus, but without hematoma removal, secondary damage can persist, and cerebrospinal fluid circulation may not be restored promptly. Combining SPDH with EVD addresses these issues. This combination clears the hematoma, reduces its occupying effect, and mitigates the cytotoxic impact of the hematoma decomposition products while helping to manage hydrocephalus in patients15,33. This is especially important for the brainstem, which is especially sensitive to these cytotoxic effects33,34. This study showed that SPDH combined with EVD was the best surgical method for PBSH complicated with hydrocephalus, which is very beneficial to the relief of hydrocephalus in patients with PBSH complicated with hydrocephalus.

Currently, the management of PBSH primarily involves passive and conservative intensive care. Despite the benefit of surgery for those presenting with PBSH has not been proven yet due to the limited dataset, exploratory surgical interventions are under consideration7,14,35,36. This retrospective study specifically examines hydrocephalus during the early phase of PBSH, and surgical intervention may at least benefit patients in the early stage of treatment. However, this is evidently insufficient in addressing the entire treatment process and ultimate treatment objectives for PBSH. While various factors such as the financial capacity of patients’ families, their willingness to continue treatment, and the rehabilitation and nursing capabilities of rehabilitation facilities may greatly influence later-stage treatments, it is imperative to conduct prospective studies that encompass later prognostic indicators for hydrocephalus in patients with PBSH to determine specific surgical intervention methods associated with favorable or unfavorable outcomes.

Conclusions

PBSH complicated by hydrocephalus is a severe condition and was associated with a poorer outcome. Our research identified the most important factors contributing to the development of hydrocephalus in patients with PBSH: midbrain hemorrhage, tegmental pons hemorrhage, combined cerebellar or intraventricular hematoma, and hematoma volume > 6.1 mL. Furthermore, the risk of developing hydrocephalus significantly increased when patients exhibited multiple risk factors, that is, multiple hematoma sites, combined intraventricular hematoma, or hematoma volume > 6.1 mL. Advances in minimally invasive neurosurgical techniques have enhanced early prognosis in patients with PBSH. Both EVD and SPDH are effective in alleviating hydrocephalus, with the combination of SPDH and EVD being the most effective surgical method for relieving hydrocephalus, increasing the probability of successful ventilator weaning and reducing the 30-day mortality rate, thereby laying a solid foundation for neurological rehabilitation.