Discussion

This systematic review synthesized the current evidence regarding early versus late mobilization on outcomes in adult patients undergoing surgical evacuation for cSDH. Our findings suggest that EM is safe and may reduce postoperative complications and improve functional recovery, without any increased risk of hematoma recurrence. All four studies reported no significant differences in cSDH recurrence rates between groups. This was consistent with our supplemental pooled analysis, which likewise found no statistical differences between groups.

Two studies demonstrated a significant reduction in medical complications with EM. This reduction appeared to be attributed to lower rates of pneumonia, urinary tract infections and gastrointestinal complications such as constipation and ileus. These findings are consistent with the broader surgical literature, which shows how early ambulation reduces complications associated with prolonged immobility, particularly respiratory and urinary issues. ^{Reference Chen, Wan and Wang23,Reference Rupich, Missimer and O’Brien24,Reference Turan, Khanna and Brooker30,Reference Rosowicz, Brody and Lazar31} However, this was not replicated in all of the included studies. Notably, the paper by Staartjes et al. did not find a significant difference in overall adverse events despite a comprehensive enhanced recovery after surgery protocol. The authors acknowledged that protocol adherence likely varied between patients, potentially diminishing the observed impact. Additionally, their protocol included multiple elements, making it difficult to interpret the impact of mobilization timing on their reported complication rates. Adeolu et al. only identified one complication in each group, suggesting the study was likely underpowered to detect any significant change in complication rates.

The assessment of functional recovery across included studies revealed mixed findings. The GET-UP trial by Sousa et al. found a significant improvement in functional outcomes, with a higher proportion of patients in the EM group achieving favorable scores on the GOSE (GOSE ≥ 7) at one month postoperatively. Notably, this study also demonstrated a significant reduction in postoperative complications, which is known to impact functional outcomes after neurosurgical procedures. ^{Reference Reponen, Tuominen, Hernesniemi and Korja32,Reference Sarnthein, Stieglitz, Clavien and Regli33} This supports the idea that fewer complications may have contributed to better outcomes, though it is unclear if these occurred in the same patients. In contrast, Staartjes et al. found no difference in functional outcomes between groups. This was attributed to protocol complexity and variations in adherence, which may have diluted the benefits of EM. Adeolu et al. also reported no significant difference in functional outcomes (GOS) at discharge. However, the majority of patients in both groups achieved a maximum GOS score, suggesting a possible ceiling effect, limiting the ability to detect further improvement. Overall, these differences may reflect variations in study design, follow-up duration, mobilization timing and how outcomes were measured.

The impact on the length of hospital stay also remains mixed. While Sousa et al. reported a significant reduction in hospital stay with EM, Staartjes et al. did not observe this effect. This variability might reflect differences in healthcare systems, discharge criteria or the specific components of the “early mobilization” protocols. As the Enhanced Recovery After Surgery protocol adopted by Staartjes et al. implemented multiple changes, these differences in the timing of other interventions, such as the initiation of thrombosis prophylaxis and earlier restarting of oral anticoagulants or antiplatelets, may have obscured the effect of EM.

This review identifies several limitations in the literature on mobilization after cSDH surgery. First, the number of included studies was small, limiting the ability to perform a meaningful quantitative analysis. Second, there were considerable differences in study design, with only one RCT and three observational studies. Observational designs carry a greater risk of bias due to confounding factors, even when methods such as propensity matching, as used in Staartjes et al., are applied. The single-center nature of all studies further limits the generalizability of their findings. There were also major inconsistencies in the definitions of key variables, such as what constituted “early” and “late” mobilization, definitions of recurrence and differences in functional outcome assessments. The timing of EM ranged from 6 to 24 hours postoperatively, while DM ranged from after 2 days to beyond 7 days. Definitions of recurrence and functional outcome measures also varied across studies. In addition, differences in postoperative care, including positioning protocols, use of thromboprophylaxis and drain management, may have influenced outcomes. Finally, the follow-up periods varied and were not consistently reported beyond two months across all studies. These factors limited our ability to perform further quantitative analysis of the available literature.

Additionally, there were limitations with our study that need to be addressed. This study was not preregistered (e.g., in PROSPERO), which limits independent verification of pre-specified methods. Nonetheless, the analyses were conducted according to a predefined plan established prior to data collection. Another limitation is that a formal risk-of-bias assessment (e.g., ROB-2, ROBINS-I or Newcastle-Ottawa Scale) was not performed. This decision was made because the included studies were heterogeneous in design and largely retrospective in nature, with only one study being an RCT. ^{Reference Sousa, Pinto and da Silva27} However, the lack of a structured bias evaluation limits the ability to formally quantify the methodological quality of included studies. Some degree of bias cannot be excluded, and several bias domains should be acknowledged. Selection bias may exist due to the inclusion of mostly single-center observational studies, which could reflect differences in patient selection and perioperative management. Measurement bias may have been introduced by heterogeneity in how early and late mobilization were defined and by the use of different functional outcome scales (GOS, GOSE, mRS), which may have resulted in inconsistent classification of recovery status across studies. Reporting bias is also possible given the inconsistent documentation of complications and follow-up intervals. These factors collectively limit both the internal validity and external generalizability of our results and should be carefully considered when interpreting these results and the overall strength of evidence in the literature.

Another limitation is that we did not perform formal sensitivity analyses or quantitative exploration of heterogeneity, such as subgroup analyses or meta-regression. Because only four studies were eligible and differed substantially in design, definitions of mobilization and reported outcomes, these analyses were not methodologically feasible. However, the absence of such analyses limits the ability to assess how factors like study design and mobilization timing might have influenced the pooled results.

We propose that future research should prioritize several key areas. Future, ideally larger, multicenter RCTs that are sufficiently powered would help strengthen the evidence base to help provide more robust guidelines for clinicians. Second, the development and adoption of standardized definitions for the timing of mobilization, the definition of recurrence and which outcome measures are reported are crucial to maintain consistency in this field. As shown by the fact that all the included studies had different definitions of early and late mobilization, these definitions are poorly standardized. We propose that future studies aim to maintain similar definitions when describing different mobilization timings. Although recurrence is variably defined in the literature, we propose that clinically significant recurrences, including patients with neurological sequelae or those that require surgical intervention, be emphasized in future research. ^{Reference Desai, Scranton and Britz34} Reporting outcomes using a clinically validated scale should be strongly considered to facilitate comparison of results between trials. We propose that future studies adopt the GOSE, with a score of ≥5 and ≥7 being defined as a moderate and good outcome, respectively, as a consistent functional endpoint to facilitate comparison between studies. This should be evaluated at both discharge and 30 days after surgery. This is consistent with the grading used by the GET-UP Trial. ^{Reference Sousa, Pinto and da Silva27} Finally, we recommend that studies with follow-up periods beyond two months are needed to assess the long-term impact on recurrence and functional outcomes fully. This is in keeping with other cSDH studies that have followed patients over six months postoperatively. ^{Reference Jeon, Rim and Lee35,Reference Martinez-Perez, Tsimpas, Rayo, Cepeda and Lagares36}

The findings of this review suggest that EM compared to DM after surgical evacuation for cSDH is a safe and potentially beneficial strategy. Clinicians may consider incorporating EM into their postoperative protocols, particularly given the observed reduction in medical complications, which can significantly impact patient recovery and resource utilization. The absence of a clear increase in recurrence risk with EM is a key finding that can help overcome traditional hesitations.