CSDH is common and more prevalent in the aged population. Surgical treatment might generally be indicated in conditions of a large hematoma width (> 10 mm) or a shift of the midline (> 5 mm) on CT scan, and surgical intervention has been accepted as the treatment of choice24. However, invasive treatment for CSDH in advanced-age patients who take long-term anticoagulation/antiplatelet medications or have poor physical health carries a considerable risk of complications, such as subdural empyema, tension pneumocephalus, brain contusion, subdural or epidural hematoma, intracerebral hemorrhage, catheter penetration to the brain and even death10,25. In particular, when a brain CT scan reveals slight compression of the brain parenchyma caused by hematoma and the patients have no/mild neurological symptoms, a surgical decision cannot be justified.
From the literature, it is apparent that observation therapy for CSDH could be used in some appropriately selected patients with asymptomatic, small-volume CSDH, patients refusing surgery, or those for whom surgery carries a high risk, which might result in potential health cost savings and eliminate perioperative risks, such as but not limited to general anesthesia9,10,12,14,19. What kind of patients are prone to spontaneous resolution of CSDH is unknown, and there are few studies in the literature concerning the natural course of CSDHs. Kim, H. C. found that 13/16 CSDH patients undergoing simple wait-and-scan management experienced spontaneous total resolution. These patients had an MGS that ranged from grade I to grade II at admission. These results suggest that close watching could be chosen when the hematoma volume or thickness was less than 43 ml or 13 mm, respectively, and the MLS was less than 5 mm on brain CT scan (n = 16)19.
However, in a review, the rate of spontaneous resolution of CSDH with or without medical therapy ranged from 2.4–18.5%13,26. This variability likely stems from heterogeneous disease severity, small sample sizes, and differences in conservative measures. In addition, in most of the patients above, the evaluated conservative treatments were given as adjuvants after surgery, which may affect the evaluation of the real efficacy of each treatment alone. Therefore, we conducted this subgroup analysis on primary (not previously treated) mild CSDH. Although many factors were included in our logistic regression model, we found that only hypodense hematoma and VOH were independent predictors of the efficacy of wait-and-watch management in our study.
Hematoma stability varies depending on the stage of the bleeding cycle27, namely, bleeding, coagulation, and rebleeding28. In clinical cases, CT is the preferred radiological tool for making a CSDH diagnosis. CT-confirmed CSDH may present a variety of imaging characteristics, and it is well known that the radiological subtype of CSDH may change over time29. The radiological subtypes that can present at the time of diagnosis possibly represent different pathophysiological stages of CSDH30. Different hematoma densities relative to the brain parenchyma on CT images represent the proportion of fresh blood, with hypodense areas representing hematoma of an older age and hyperdense components of more recent or active bleeding31–34.
Not only is “wait and watch” or “wait and scan” management more appropriate for patients with no or minor symptoms, but some authors also advocate that patients who have small lesions with low density on CT have a greater chance for spontaneous resolution of their hematoma16. Nomura et al. suggested that the mixed-density type has a high tendency to rebleed, while a low-density hematoma is stable with a low tendency to rebleed or exhibit fibrinolytic abnormalities35. When the interstitial hematoma matrix changes from an isodense to a low-density signal on CT scans, on surgical inspection it changes from a dark reddish to a xanthochromic translucent liquefied hematoma, while it diminishes in volume over time36. If a low-density hematoma is considered to be the resolution stage of CSDH, this might explain the significantly higher rate of low-density hematoma in the case group compared with the other three hematoma density groups.
Preoperative VOH and MTH have been reported as prognostic factors of CSDH, consistent with our results in the univariate analysis37,38. However, our results only found a significant relationship between VOH and hematoma stability in the multivariate analysis. Additionally, there was a significant difference between the SDH nearly disappearing and the visible subgroups (P = 0.001). We hypothesized that the degree of compression from the hematoma might be a result of multiple factors, including hematoma thickness, MLS, patient age, and brain atrophy39. In addition, brain atrophy or expandability of the skull may increase the reserve capacity. When the reserve capacity is sufficient, CSDH may be asymptomatic for a longer period, even more than a year40.
In our present study, under the premise of brain atrophy in the elderly, hematoma volume may contribute more directly to intracranial pressure growth and neurological dysfunction than MTH. With the application of ROC analysis, we estimated the predictive capacity of VOH and hypodensity as a combination in our study. The AUC was under .90, which suggests that the relationship between the predicting factors and treatment outcome is more of an association than a causal relationship. There may be more complex interacting factors that were not included in our analysis, such as physical activity and alcohol use41. The mean hematoma volume of the case group in our study was 70.0 ml, which implies not only size but also manifestations of patient-specific factors. In our center, many patients are old and have significant brain atrophy, and sometimes a relatively large hematoma does not necessarily mean that the patient is in poor condition and needs immediate surgery. When a dilemmatic situation is encountered, such as subtle symptomatic patients with a relatively large hematoma or MLS, we require early hospitalization and close monitoring of the patients. If their symptoms worsen and/or their condition deteriorates, we will carry out surgical interventions. In summary, regardless of the selection of treatment (whether it is surgical or not), a close scan/follow-up of these patients is paramount.
A clinical classification might help in deciding which therapy modality might be more appropriate. However, no consensus exists about the best CSDH treatment for each grade. It also seems that there are no clear clinical or radiological signs indicating whether the CSDH will resolve spontaneously. In our study, the ADL-BI score, which measures the daily living activities of patients with neurological diseases, was more sensitive than the MGC-GCS and ASA-PS for predicting the observation treatment outcome in the univariate analysis. The ASA-PS is a method of characterizing the patient operative risk on a scale of 1–5, with 1 being normal health and 5 being moribund, which has strong, independent associations with postoperative medical complications and mortality42. The GCS was originally developed to evaluate the recovery of patients with traumatic brain injury, while the ADL-BI scores measuring the daily living ability, such as being able to defecate and urinate independently, may reveal more subtle information regarding the outcomes of our observation cohort.
Symptoms and neurological deficits from primary and recurrent CSDH arise most often due to the mass effect of the subdural collection. However, observation treatment cannot help reduce this compression and thereby alleviate symptoms and deficits. Furthermore, the prolonged presence of a hematoma on the cerebral surface, as the majority of the patients’ hematomas were still present in our study throughout the follow-up, requires longer follow-up periods, which may lead to a lengthy medical procedure and bedridden status. In addition, in some cases in our case group, although the CSDH of the patient had nearly disappeared, there remained the problem of a constant headache, which might seriously affect the patient's quality of life. These arguments suggest that wait-and-watch therapy alone may be less efficient for symptomatic CSDH, especially if the patient is in an anxious mood, while increasing the risk of hematoma enlargement and the duration of disturbing symptoms. In addition, a systematic review found that corticosteroids as either monotherapy or a surgical adjuvant improved the clinical outcomes of adult patients with CSDH43; statins are another medication considered a conservative treatment for CSDH. A clinical trial including 200 patients demonstrated that atorvastatin is safe and effective in reducing hematoma volume and improving the neurological function of patients compared to placebo44. Therefore, given the above, medical interventions, including drug treatment, should be combined with observational treatment.
Limitations
First, our study has a retrospective design, which usually produces a risk of selection bias. Second, it is unknown whether some patients in the case group may have used traditional herbal medicine to prevent the CSDH increase. Third, the Coniglobus Formula for calculating hematoma measurements can lead to an overestimation of hematoma volume45,46. Fourth, our limited follow-up is not the end-point of these patients with CSDH, and there may be some incidence of hematoma recurrence. We have encountered cases with recurrence of a hematoma after strenuous exercise and/or bathing one’s feet in hot water. Hence, the results from the patients in the present series are self-limiting, and longer follow-up is needed in the future. Despite these limitations, this study still provides useful information to predict the outcome of CSDH patients receiving observation treatment and can provide more perspectives on CSDH treatment strategies for neurosurgeons.