For bCSDH patients who underwent unilateral evacuation, the postoperative enlargement of contralateral hematoma is not rare. However, there were a few studies researching the factors associated with CHP. In this current study, the progression rate was 13.41%, which was smaller than most previous studies[6, 7, 10, 16]. Five factors, including pre- and postoperative contralateral hematoma volume, the wide contralateral hematoma, age and opposite hematoma locating on the curved side had statistical significance between the CHP group and non-CHP group. The wide type of opposite hematoma and contralateral hematoma situated on the curved side were independent risk factors for CHP.
Age is an important factor in CSDH patients. Many studies had demonstrated that the age was associated with the incidence and recurrence rate of chronic subdural hematoma[2, 18]. Our investigation explored that among CHP group, the age was significantly higher than non-CHP group. This phenomenon might be explained by the brain atrophy and the increasing fragility of vascular. Blatter et al. described that the volume of brain parenchyma decreased with age, which meant older people would suffer more severe brain atrophy[19]. This severe atrophy was likely to increase the risk of arachnoid tearing which might cause the appearance of subdural effusion[20]. Moreover, Spallone et al. considered that the vascular would be more fragile among older CSDH patients, meaning the vessels in the dura mater was more likely to rupture. Therefore, for bCSDH patients with older age, the subdural effusion might be more common and the bleeding tendency in the subdural space was likely to be more active. The newly appearing blood from the fragile vessels and the subdural effusion might enter into the contralateral hematoma cavity, leading to the development of the non-surgical hematoma. However, none of the five previous investigations reported that the age was associated with CHP. So we consider that research with a large sample size to identify the correlation between CHP and older patients is still needed.
Radiological presentation has always been a central issue in investigations with regard to the progression of chronic subdural hematoma. According to the previous studies, the MRI intense, CT density, hematoma location, volume and maximal thickness of contralateral hematoma, the VR ratio and the midline shift were considered to be significantly correlated with CHP[6, 7, 10, 16]. In this current research, both the pre- and postoperative contralateral hematoma volume of CHP patients are statistically significant in univariate analysis. One possible hypothesis is that a bigger contralateral hematoma is more likely to result a rapid decrease of intracranial pressure in non-surgical side and make a greater postoperative parenchymal shift, which may be correlated with the development of contralateral hematoma. More investigations are needed to verify this hypothesis. As for homogeneous type of contralateral hematoma, although it had no significance in our research, some investigations demonstrated that it was associated with CHP[6, 7]. Shen et al. thought that homogeneous type meant the hematoma had been liquefied totally, making the expansion and enlargement of hematoma more easily. However, some researches revealed that compared with other hematoma density, the bleeding tendency and inflammatory reaction of homogeneous hematoma are milder, suggesting that this type is less active to progress[4, 11, 21]. This finding can be applied to explain the outcome of our study reasonably.
The wide type of contralateral hematoma is an independent risk factor about contralateral hematoma progression. A previous research reported that hematoma location was correlated with the spontaneous resolution of CSDH[22]. However, study exploring the relationship between CHP and hematoma location was rare. Only Shen et al. reported that the hematoma location might be associated with the CHP[7]. According to the description of Shen et al., the contralateral hematoma locating in three lobes was easier to progress, which might be associated with the thickness of hematoma envelope. We considered that this phenomenon was associated with the surface area and the feeding vessels of hematoma membrane. Previous researches demonstrated that the hematoma cavity was enclosed by the inner and external membrane[23]. The inner membrane barely works in the process of CSDH formation, while the external membrane plays an extremely important role in the growth of subdural hematoma[20]. Many studies found that some inflammatory mediators which promote the development of CSDH, for the instance of vascular endothelial growth factor (VEGF), angiopoietin-2 (Ang-2), and thrombomodulin, are originally from the outer membrane[24–27]. Moreover, a large quantity of macro-capillaries which is full of loose gap junction also exist in the outer membrane. The blood may exudate from the capillaries and transfer into the hematoma cavity through these junctions without any hindrance, which is considered as the main reason for the development of contralateral hematoma[20, 25]. Compared with limited type, wide type has a larger area of outer membrane, which may have a greater number of inflammatory factors and macro-capillaries. Therefore, among wide contralateral hematoma, the inflammatory response may be more severe and more blood is likely to transfer from the capillaries into hematoma cavity, causing the contralateral progression. The middle meningeal artery (MMA) was also hypothesized to play a role in the progression of contralateral hematoma. Nakaguchi et al. reported that the hematoma locating on more lobes was fed by more vessels from the MMA, which would suffer a higher bleeding risk[11]. For wide contralateral hematoma, the feeding vessels would be distributed more widely than limited type. So we think that the wide type of contralateral hematoma will have a higher bleeding tendency which may promote the development of non-operated hematoma.
The other independent risk factor is the opposite hematoma which locates on the curved side. As we know, none of the five previous studies researched this variable and our investigation is the first study to explore the relationship between skull morphology and CHP. The cranial morphology is always considered to be correlated with the laterality of hematoma[12]. Lee et al. found that bCSDH were more common among symmetrical cranial[28], which was in accordance with our investigation. In this current study, we found that when the non-operative hematoma occurred on the curved side, the patient was more likely to suffer CHP. This phenomenon was hypothesized to be associated with the subdural space between the two cranial sides. A previous research reported that patients with CSDH preferred to lie on the less curved side[29]. In such condition, the subdural space of the curved side would be enlarged due to the brain gravity, which facilitated the separation of dura-arachnoid interface and increased the risk of arachnoid tearing. This interface separation might influence the arachnoid and vascular permeability[8]. Moreover, Akhadder et al. also demonstrated that the enlarging subdural space would easily result the rupture of bridging veins[30]. Therefore, when the contralateral hematoma locating on the curved side, the appearance of effusion and veins rupture would be more frequent, contributing to the hematoma enlargement.
Our study demonstrated that the hematoma location and the cranial structure, which are ignored by most investigations, may contribute to CHP. These two independent risk factors are intuitive and can be easily obtained from the routine CT scan. The previous studies had reached a consensus that patients with higher CHP risk needed a rigorous follow-up to evaluate the change of non-operative hematoma. In our research, both the independent risk factors are correlated with the vessels existing on the outer membrane or the dura mater. Therefore, we think that the MMA embolization, a surgery method which is less invasive, can be applied for contralateral hematomas which have no surgical indication but are considered to progress easily. This treatment method can shorten the interval from initial surgery to the last follow-up[31], making more benefits for patients. However, the safety and effectiveness of this new treatment strategy is still needed to be investigated by researches with large sample size in the future.
There are some limitations in this investigation. Firstly, our research is a retrospective analysis which may introduce some biases affecting the outcomes unavoidably. Secondly, the statistical analysis about age and the classification of hematoma density are influenced by the small sample size of our research. These variables are worthy to be investigated with a larger research population.