Introduction

With over 50 million people affected annually, moderate to severe traumatic brain injury (TBI) is a substantial cause of morbidity and mortality on a global scale [1]. Approximately 60% of patients with moderate or severe TBI suffer severe long-term disability or death [2, 3]. The mechanisms of TBI are mediated by two principal components: primary insult related to the impact of trauma itself, which consists of structural damage such as axonal shearing injury or intracranial hemorrhage, and secondary brain injuries, which are a result of a complex interplay of metabolic, cytotoxic and vasogenic factors that lead to an increase in the intracranial pressure (ICP) and are responsible for excess neurological morbidity following TBI [4].

Since its introduction by Theodor Kocher in the early 20th century, decompressive hemicraniectomy (DHC) has been the standard surgical procedure for the treatment of refractory post-traumatic intracranial hypertension (in patients with diffuse injury and swelling with or without mass lesion, with midline shift or impending herniation, usually Marshall ≥ 3) [5, 6]. Based on the Monroe-Kellie doctrine, the underlying rationale of DHC consists in expanding the volume of the cranial vault to combat rising intracranial pressure (ICP) [7].

The RESCUEicp randomized controlled trials showed that DHC resulted in a reduction of the overall mortality rate but also in higher rates of vegetative state and neurological disability at 6 months follow-up [8]. Various other treatment options have been investigated but did not show a clinical benefit when evaluated in RCTs [9,10,11,12,13,14]. With a track record of over 200 failed clinical trials, new treatment options for severe TBI are desperately needed [15, 16].

Recent laboratory studies have provided novel mechanistic insights into CSF pathophysiology [17,18,19,20]. One model assumes that CSF is absorbed via the pericapillary Virchow-Robin spaces (VRS) at the basal cisterns, which are part of the glymphatic system [17,18,19,20,21,22]. Increased cisternal pressure following the obstruction of the VRS – such as due to traumatic subarachnoid haemorrhage (tSAH) – may lead to a reversal of CSF flow from the subarachnoid space back into the brain parenchyma (known as the “CSF-shift edema” hypothesis), resulting in increased ICP, disruption of the permeability of the glymphatic system and stagnation of toxic metabolic factors [23,24,25].

This emerging concept provided the rationale for Cherian et al. to introduce basal cisternostomy (BC) with basal cisternal drainage in 2013 as a novel treatment modality in moderate to severe traumatic brain injury [26]. Following the promising results of their study, this technique has been increasingly adopted by several neurosurgeons worldwide [27,28,29,30,31]. The procedure usually consists of a craniotomy and fenestration of all basal cisterns including the prepontine cistern, passive CSF drainage, and leaving in place an external cisternal drain within the prepontine cistern [32]. However, with few published studies and without any systematic evidence summary, the effect of BC on clinical outcome following TBI remains unclear [30, 33,34,35,36,37,38]. The present review aims to evaluate the effect of BC as an adjunct to decompressive hemicraniectomy (termed “adjuvant BC” = BC + DHC) in patients with moderate to severe traumatic brain injury (TBI).

Materials and methods

Overview

A systematic review was carried out to identify any studies in patients with moderate to severe TBI and reporting intra-/perioperative, or clinical outcomes of adjuvant BC (BC + DHC), including (1) mean Glasgow Outcome Scale (GOS) at follow-up, (2) GOS ≥ 5 at follow-up, (3) mortality at follow-up, (4) complications, (5) operative time, (6) mean ICP (preoperative / after the first burr hole / after hemicraniectomy and/or cisternostomy / at closure / in the ICU), (7) mean brain outward herniation, (8) osmotherapy, (9) complications, and (10) length of stay in the ICU. Title and abstract screening, full-text review, and data extraction were handled independently by three reviewers (VES, VP and OC) using Covidence (Covidence systematic review software, Melbourne, Australia) [39], and disagreements at any stage were resolved by discussion and consensus. This systematic review followed the methodological framework described by Arksey and O’ Malley for Systematic Reviews [40] and the PRISMA Statement [41].

Search strategy

The PubMed/MEDLINE and EMBASE database was searched to identify eligible articles. The search strategy was as follows for both databases: TBI OR “traumatic brain injury” OR “head injury” OR “brain trauma”) AND (cisternostomy OR cisternal OR cisterns). Word variations and exploded medical subject headings were searched for whenever feasible. References were screened to identify additional relevant articles. The search included articles published between 1995 and 2023 as we aimed to include patients managed under contemporary TBI guidelines and the technique was first described in 2007. Last comprehensive search was conducted on December 26th, 2023.

Study selection

Only in vivo studies in English enrolling humans aged ≥ 18 years were considered. Systematic reviews, case reports, small case series with less than 5 patients and studies dealing reporting data exclusively on treatment of mild TBI as classified by GCS were excluded. Studies looking at patients with moderate to severe TBI receiving adjuvant BC (BC + DHC) were included. To be eligible for inclusion, studies had to assess at least one of the abovementioned variables of interest at preoperative, intraoperative and postoperative timepoints. The primary comparison of interest was adjuvant BC (BC + DHC) versus standalone DHC. However, we also included studies looking at (1) adjuvant BC (BC + DHC) as a standalone treatment or (2) comparing adjuvant BC (BC + DHC) versus standalone BC.

Data extraction and quality assessment

The following data were extracted from all included publications: Study design and year of publication, number of patients, mean patient age and sex distribution, adjuvant BC (BC + DHC) and standalone DHC, technical nuances, as well as at least one of the outcomes of interest. The methodological quality of the included studies was graded using the Newcastle-Ottawa Quality Assessment Scale for Cohort Studies [42] and the Cochrane Risk-of-Bias (RoB-2) tool for RCTs [43].

Statistical meta-analysis

Based on the anticipated heterogeneity and the lack of larger RCTs for certain outcomes, a random-effects analysis model was applied: Mantel-Haenszel models evaluating odds ratios as effect measure for dichotomous outcomes, and inverse variance models evaluating mean differences as effect measure for continuous outcomes [44]. Cochran’s Q and I2 were used to evaluate the strength of evidence for heterogeneity and the level of heterogeneity, respectively. A p ≤ 0.1 was considered as relevant heterogeneity. We performed subgroup analyses for RCTs. Reporting bias was assessed by visual inspection of funnel plots. The meta-analyses were carried out in Review Manager 5.4 (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark) [45]. P ≤ 0.05 on 2-tailed tests were considered statistically significant for the assessment of overall effect.

Results

Literature search

A PRISMA flowchart is shown in Fig. 1. Of ten publications [27,28,29, 46,47,48,49,50,51,52] which originally met the inclusion criteria, Vemula et al. [51] was based on the same cohort as the publication of Chandra et al. [29], and only the latter publication was regarded. This was also excluded since it deals with standalone BC and not to adjuvant BC (BC + DHC) and did not meet inclusion criteria. A total of 8 publications was included for qualitative analysis [27, 28, 46,47,48,49,50, 52]. A total of five articles had sufficient outcome data in appropriate format to be eligible for quantitative meta-analysis [27, 46, 48, 50, 52].

Fig. 1
figure 1

PRISMA Flowchart

Full size image

Included study characteristics and quality

Tables 1 and 2 summarizes the characteristics of the included studies and their quality assessment while clinical outcomes are detailed in Tables 3 and 4. Details of the included study characteristics are provided in Supplementary Content 1 – Supplementary Table 1.

Table 1 Overview of the study-related characteristics of the included studies
Full size table
Table 2 Overview of the baseline clinical and surgical characteristics of the included studies
Full size table
Table 3 Overview of outcomes related to neurological outcome, mortality, complications, brain outward herniation and patients requiring osmotherapy among the included studies
Full size table
Table 4 Overview of outcomes related to intracranial pressure, length of stay, and cranioplasty among the included studies
Full size table

The primary aim of this study was to compare the outcomes of adjuvant BC (BC + DHC) versus standalone DHC. However, for the sake of completion in summarizing the clinical evidence on adjuvant BC in TBI, we also included other secondary comparisons:

  • Primary comparison:

    • Adjuvant BC (BC + DHC) versus standalone DHC:

      • Kumar et al., 2022 (40 patients) [48]: Prospective, quasi-experimental study.

      • Giammattei et al., 2020 [27] (40 patients): retrospective single-centre study.

      • Youssef et al., 2020 [52] (40 patients) RCT.

      • Singh et al. 2021 [50] (54 patients): RCT.

  • Secondary comparisons:

    • Adjuvant BC (BC + DHC) versus standalone BC:

      • Parthiban et al., 2021 [28] (40 patients): retrospective single-centre study.

      • Encarnación Ramirez et al., [49] (30 patients): prospective, three-centre study.

    • Adjuvant BC (BC + DHC) versus standalone BC versus standalone DHC:

      • Cherian et al., 2019 [46] (1032 patients): Retrospective single-centre study.

    • Adjuvant BC (BC + DHC) standalone:

      • Goyal and Kumar, 2021 [47] (9 patients): Prospective observational study reporting the results o f nine patients treated with adjuvant BC (BC + DHC) standalone.

With regard to quality, most cohort studies analyzed an at least somewhat representative cohort according to the definition of the Newcastle-Ottawa Scale but most suffered from poor comparability and provided too short follow-up data to draw valid conclusions on the final outcome of the patients.

According to the Cochrane risk-of-bias tool, the remaining two RCTs were assessed to be at high risk of bias as the randomization process was not detailed in the study by Singh et al. [50] and was generated by alternation in the study by Youssef et al. [52].

Results of the primary comparison (adjuvant BC = BC + DHC) are reported in the following paragraph. The results of the secondary comparisons are reported separately in Supplementary Content 2 – Supplementary Results 1, which includes the following comparisons: Adjuvant BC (BC + DHC) vs. standalone BC, Adjuvant BC (BC + DHC) versus standalone BC versus standalone DHC and Adjuvant BC (BC + DHC) standalone.

Primary comparison: adjuvant BC (BC + DHC) versus standalone DHC

Intraoperative parameters

Operative time

Among two RCTs [50, 52] and a retrospective study [27] with a total of 132 patients reporting duration of surgery (Fig. 2), there was a statistically significant increase in the duration of surgery in patients who underwent adjuvant BC compared to DHC (Mean difference [MD]: 33.80 min, 95% confidence interval [95% CI]: 26.0 to 41.59, p < 0.001). Statistical heterogeneity was low with an I2 of 0% (p = 0.73). Among two RCTs [50, 52] (94 patients), the duration of surgery was longer in patients with adjuvant BC (MD: 34.75, 95% CI: 26.49 to 43.01, p < 0.001). Statistical heterogeneity was low with an I2 of 0% (p = 0.69).

Fig. 2
figure 2

Forest plots detailing the meta-analyses of intraoperative parameters

BC + DHC, adjuvant basal cisternostomy + decompressive hemicraniectomy; DHC, standalone decompressive hemicraniectomy; RCT, randomized controlled trial; RCT, randomized controlled trial;

Full size image
Closing ICP

Among a retrospective [27] and prospective study [48] (80 patients) reporting the closing ICP, there was no statistically significant difference in the closing ICP (MD: 0.79 mmHg, 95% CI: -9.00 to 10.58 mmHg, p = 0.87). Statistical heterogeneity was considerable with an I2 of 96% (p < 0.001).

Perioperative parameters

Length of stay in the ICU

There was a statistically significant reduction in the length of ICU stay for adjuvant BC (MD: -3.25 days, 95% CI: -5.41 to -1.09 days, p = 0.003, Fig. 3), among two prospective RCTs [50, 52], a prospective [48] and a retrospective [27] study (174 patients) reporting appropriate data. Statistical heterogeneity was low with 0% (p = 0.85). Among two RCTs [50, 52] (94 patients), there was no statistically significant intergroup difference (MD: 2.55 days, 95% CI: -5.33 to 0.22 days, p = 0.07). Statistical heterogeneity was low with an I2 of 0% (p = 0.97).

Fig. 3
figure 3

Forest plots detailing the meta-analyses of perioperative parameters

BC + DHC, adjuvant basal cisternostomy + decompressive hemicraniectomy; DHC, standalone decompressive hemicraniectomy; RCT, randomized controlled trial; RCT, randomized controlled trial;

Full size image
ICP in the ICU

Among a prospective and a retrospective study (80 patients) reporting ICP in the ICU, there was no statistically significant intergroup difference (MD: -2.44 mmHg, 95% CI: -7.53 to 2.66 mmHg, p = 0.35). Statistical heterogeneity was considerable with an I2 of 97% (p < 0.001).

Complications

Within a prospective [48] and a retrospective study [27] (80 patients) reporting appropriate data, there was no statistical difference in postoperative complications (OR 3.19, 95% CI: 0.13–77.54, p = 0.48). Statistical heterogeneity was substantial with an I2 of 68% (p = 0.08).

Osmotherapy

A RCT [50] and a retrospective study [27] reported data on patients requiring osmotherapy (94 patients). Among these, there was a statistically significant reduction in the odds of requiring osmotherapy in patients with adjuvant BC (Odds ratio [OR] 0.09, 95% CI: 0.02 to 0.41, p = 0.002). Statistical heterogeneity was moderate with an I2 of 50% (p = 0.16).

Brain outward herniation

Among one RCT [50] and a retrospective study [27] reporting brain outward herniation (94 patients), there was a significant reduction in brain outward herniation (measured as described by Bruno et al. [53]) in patients with adjuvant BC (MD: -0.68 cm, 95% CI: -0.90 to -0.46 cm, p < 0.001). Statistical heterogeneity was low with an I2 of 0% (p = 0.38).

Clinical outcomes

Mean glasgow outcome scale (GOS) at follow-up

Forest plots for meta-analyses of clinical outcomes are summarized in Fig. 4. Among a RCT [52] and a retrospective [27] study (80 patients) reporting mean GOS at final follow-up a median follow-up of four weeks, there was no significant change in patients with adjuvant BC (MD: 0.77, 95% CI: -0.08 to 1.63, p = 0.07). Statistical heterogeneity was low with an I2 of 0% (p = 0.48).

Fig. 4
figure 4

Forest plots detailing the meta-analyses of clinical outcomes

BC + DHC, adjuvant basal cisternostomy + decompressive hemicraniectomy; DHC, standalone decompressive hemicraniectomy; RCT, randomized controlled trial; RCT, randomized controlled trial;

Full size image
GOS ≥ 5 at final follow-up

Among a RCT [50], a prospective [48] and a retrospective [27] study (134 patients) reporting GOS ≥ 5 at a median follow-up of six months, there was no statistically significant intergroup difference (OR 2.50, 95% CI: 0.95 to 6.55, p = 0.06).

Statistical heterogeneity was low with an I2 of 0% (p = 0.74).

Mortality at follow-up

Among two RCTs [50, 52] and a prospective study [48] (134 patients), there was no statistical intergroup difference (OR 0.80, 95% CI: 0.17 to 3.74, p = 0.77). Statistical heterogeneity was high with an I2 of 72% (p = 0.03). Among two RCTs [50, 52] (94 patients), patients with adjuvant BC had borderline significantly reduced odds for mortality at a median follow-up of four weeks (OR 0.39, 95% CI: 0.15 to 1.00, p = 0.05). Statistical heterogeneity was low with an I2 of 0% (p = 0.43).

Publication bias

Funnel plots are provided for each analysis in Supplementary Content 3 – Supplementary Figs. 1–12. Due to the low number of studies, a proper qualitative or quantitative analysis was not possible, and no assumptions can be made regarding publication bias.

Discussion

The present study is the first systematic review and meta-analysis to quantify the effect of adjuvant BC (BC + DHC) in moderate to severe TBI. We evaluated eight relevant studies including a total of 1305 patients, of which three were RCTs analysing a total of 144 patients. Despite technical variations, methodological differences and limited cohort sizes of the respective studies, the results of our meta-analysis indicate that adjuvant BC (BC + DHC) compared to standalone DHC may potentially be associated with a reduction in the length of stay in the ICU and a lower mean brain outward herniation. In terms of clinical outcome, adjuvant BC (BC + DHC) appeared to reduce in-hospital mortality and long-term mortality at least in randomized studies, while there was no significant effect on neurological outcome.

The principal paradigm in TBI treatment has been to decrease ICP to minimize subsequent secondary brain injury [54,55,56]. DHC constitutes the standard last-resort surgical procedure for refractory intracranial hypertension that aims at reducing the ICP by providing more space for the brain to expand [57]. Optimization of ICP does not seem to correlate directly with clinical outcome, and the underlying reasons may be manifold: Although DHC creates an outlet for the herniating brain and reduces ICP, it does not prevent the formation of cytotoxic and vasogenic oedema [56]. The expansion of the brain comes at the cost of additional brain injury and progressive neurological deterioration caused by axonal shear stress, laceration of vessels and ischemia [35]. In addition, highly invasive approaches such as bifrontal craniectomy are frequently associated with a considerable rate of treatment-related complications and often necessitate a second surgery for the replacement of the bone flap [8, 56].

BC theoretically represents a physiological approach to the treatment of TBI-related prolonged ICP elevation by opening the cisterns to atmospheric pressure, relieving cisternal pressure, restoring glymphatic clearance, reversing the CSF-shift-oedema and reducing brain herniation [25, 58, 59]. Several previous case reports in the literature reported a beneficial effect of adjuvant BC on clinical outcomes following moderate to severe TBI [30, 34,35,36, 60, 61].

In our systematic review, we mainly identified studies comparing adjuvant BC (BC + DHC) versus standalone DHC, which allows to draw conclusions only about the additive effect of adjuvant BC (BC + DHC) in standalone DHC. Physiologically, standalone BC would constitute an adequate treatment of CSF-shift-oedema, however, proper comparative studies are lacking. As standalone DHC is an established last-resort treatment modality for refractory ICP, evaluating standalone BC in clinical studies would probably not be feasible without including DHC as the established treatment.

Concerning studies on adjuvant BC (BC + DHC) versus standalone DHC, there is evidence at least from randomized studies suggesting that in-hospital and long-term mortality are reduced with adjuvant BC (BC + DHC), however, neurological outcome was only non-significantly trending towards improvement for adjuvant BC (BC + DHC). While the mean GOS and the percentage of patients with GOS ≥ 5 at follow-up were consistently higher across all evaluated studies, likely due to the small sample size. Except for Kumar et al. [48], the majority of studies reported that patients with adjuvant BC (BC + DHC) had also lower mortality rates at final follow-up compared to patients with standalone DHC. Importantly, the study of Kumar et al. [48] suffers from several considerable limitations: the distribution of patients in the intervention groups was highly inequal, with only 9 patients in the adjuvant BC (BC + DHC) group and 31 patients in the standalone DHC group. Therefore, their study may be prone to random variation and overestimate the mortality rate in the adjuvant BC (BC + DHC) group, causing substantial heterogeneity in the meta-analysis. In contrast, the odds of mortality at follow-up were reduced by 60% when evaluating only the two relevant RCTs [50, 52], however, there is still insufficient evidence to draw valid conclusions.

Several additional factors may contribute to improved clinical outcome in patients with adjuvant BC (BC + DHC): Our meta-analysis demonstrated a notable reduction in the length of ICU stays in patients who underwent adjuvant BC (BC + DHC) compared to standalone DHC. In addition, there was a marked reduction in the odds of requiring osmotherapy in the adjuvant BC group (BC + DHC) – whose administration may itself potentially cause electrolyte abnormalities and renal failure [62].

It has been previously demonstrated that it is possible to drain a larger amount of CSF through the cisterns and thus achieve a more sustainable reduction of the ICP [27]. The rationale for an in-situ drain following BC is based on recent evidence suggesting that interstitial fluid communicates with cisternal CSF through the Robin-Virchow-spaces [58] – which may also explain why the placement of an external ventricular drainage has not shown many beneficial effects on clinical outcomes despite lowering the ICP [63, 64]. In our meta-analysis, the mean difference in the closing ICP and ICP in the ICU did not reach statistical significance as the two evaluated studies reported conflicting results.

From a technical point of view, basal durotomy followed by the release of CSF from the basal cisterns reduces cerebral edema and relaxes the brain relatively early during the procedure [65]. This allows a more gentle durotomy preventing kinking of cortical veins, lacerations of cerebral cortex on the bone edge and additional brain injury caused by cerebral swelling [66]. Importantly, the relaxation of the brain prevents herniation and enables the immediate reimplantation of the bone flap in some patients where it would otherwise have been impossible, thus avoiding the need for a second surgery for cranioplasty [29]. Our meta-analysis confirmed that patients undergoing adjuvant BC (BC + DHC) compared to standalone DHC had a significantly reduced mean brain outward herniation on postoperative CT-scans. In cases where a second procedure for cranioplasty was required, Giammattei et al. noted that the time to cranioplasty was shorter in patients who underwent adjuvant BC compared to DHC only [27].

While the reportedly positive effects of adjuvant BC (BC + DHC) on clinical outcome have encouraged neurosurgeons worldwide to advocate for its widespread adoption, there has been a resistance to implement BC as a complementary procedure in the treatment of TBI [26]. BC is a microneurosurgical procedure requiring training in vascular and skull base neurosurgery, appropriate instrumentation and the availability of a microscope [61]. Of note, TBI has the highest prevalence in developing countries where limited access to microneurosurgical equipment in trauma care centres may remain an obstacle for the widespread implementation of BC as a standard complementary procedure for TBI treatment [67]. In clinical practice, standalone DHC for TBI is frequently performed by neurosurgeons in training who may not always be able to safely perform adjuvant BC (BC + DHC) [68].

Due to considerable qualitative and clinical heterogeneity (including differences in inclusion criteria and clinical protocols) of the currently available evidence, it remains difficult to assess the effect of adjuvant BC (BC + DHC) on clinical outcome after TBI. As the majority of these studies originate from developing countries, it is also questionable to which extent those data would reflect the results of studies in Western healthcare systems. To determine the efficacy and safety of adjuvant BC (BC + DHC) in treatment of treatment of refractory hypertension following moderate to severe TBI, larger international RCTs are warranted.

Limitations

Moderate to severe TBI summarizes a variety of pathologies with different prognoses which may itself cause a variance in outcomes. As a result, the heterogenous populations / inclusion criteria and varying indications may limit the comparability and generalizability of the studies. Due to the heterogeneity in patient cohorts and the low number of events reported outcomes, the evaluated studies are prone to random effects. The heterogeneity in study design limits the comparability of studies and decreased the number of eligible studies for meta-analysis to an average two to three, leading to a high statistical heterogeneity. Meta-analyses of just two or three studies therefore should be very cautiously interpreted. Including further databases and opting also for “gray literature” could expand the number of identified studies, potentially at the cost of study quality. Due to a considerable variance in the length of follow-up, our data may not allow for an accurate evaluation of final outcome of the patients. Similarly, different surgical protocols, e.g. adjuvant BC (BC + DHC) vs. standalone DHC, parcellate the available evidence and preclude drawing generalizable conclusions about the effects of these treatments with the currently limited published data. Overall, the abovementioned methodological limitations limit the ability to draw valid conclusions on the effect size of adjuvant BC (BC + DHC) on clinical outcome.

Conclusion

Our systematic review and meta-analysis indicates that there is insufficient moderate- to high-quality data to demonstrate a potential beneficial effect of adjuvant BC (BC + DHC) on the clinical outcome following moderate to severe TBI. Despite some evidence for reduced mortality and length of stay, there is no effect on neurological outcome. However, these results need to be interpreted with caution as they carry a high risk of bias due to overall scarcity of published clinical data, technical variations, methodological differences, limited cohort sizes, and a considerable heterogeneity in study design and reported outcomes. Consequently, it remains difficult to draw valid and generalizable conclusions about the effect of adjuvant BC (BC + DHC) on clinical outcome.