One of the final steps in several neurosurgical operations is dural defect closure, which can be achieved by re-approximating the existing dura or using additional dural substitutes. When primary closure of the dura is not feasible, surgeons often need to use grafts to achieve either watertight or onlay dural closure. This procedure is necessary to prevent cerebrospinal fluid (CSF) leakage, reduce the risk of infection after surgery, and prevent scar formation. CSF leakage is a potential complication of cranial duraplasty, particularly after intradural surgery. The incidence rate of CSF leakage following cranial duraplasty varies depending on the type of surgery but is generally reported between 4.6% and 19.6% [15,31]. CSF leaks can increase the risk of infection, headaches, or the need for reoperation [8,18].

The incidence rate of aseptic meningitis after cranial duraplasty is not well established, but case reports suggest that it may occur in 0.5% to 1.5% of patients [18]. Aseptic meningitis following cranial duraplasty can result from an immune reaction to the dural graft or from contamination during surgery [18]. The incidence rate of pseudomeningocele after cranial duraplasty varies depending on the type of surgery, graft material, and other factors. Some studies report incidence rates ranging from 6.7% to 27% [1,2,25]. The risk of pseudomeningocele can be reduced by achieving watertight closure of the dura mater, avoiding excessive CSF drainage, and promptly treating any infection [2,25].

Duraplasty is commonly used to ensure dural closure in various cranial procedures, including tumor resection, decompressive craniectomy (e.g., during Chiari decompression), traumatic subdural hematoma (SDH) evacuation, and transsphenoidal approach for parasellar tumor resection [4]. A wide variety of dural grafts are available for duraplasty, broadly categorized into autologous and non-autologous materials. Autologous dural materials include pericranial tissue, temporal muscle fascia, nuchal ligament, cervical fascia, and fascia lata. Non-autologous materials can be selected from allografts, xenografts, and synthetics [22].

Theoretically, and supported by several studies, using foreign material such as non-autologous dural grafts may increase the risk of infection and is less effective in preventing CSF leakage. However, numerous studies on duraplasty techniques and materials have yielded varied results and opinions. In the late 1980s, ideal criteria for dural repair were defined as 1) materials that do not induce or minimize, scarring formation, 2) prevention of CSF leakage and pseudomeningocele, 3) reduction of infection risk, 4) ease of manipulation, and 5) availability and affordability. These criteria were established through experimental studies on canine spinal dura in the lumbar region, as well as clinical experience [16].

Previous studies on duraplasty techniques and materials have often yielded statistically insignificant results, indicating a need for further research and analysis to achieve comprehensible conclusions. Due to these inconclusive findings, many surgeons rely on personal judgment when selecting materials for duraplasty, taking into account some factors such as the pathology, graft characteristics, duration of surgery, length of hospital stay, cost-effectiveness, and convenience. Recent reviews of duraplasty have highlighted the importance of choosing appropriate dural graft materials. Bolly et al. [22] suggested seven criteria for the ideal dural graft, including non-toxicity, immunological inertness, sterility, availability, non-adhesiveness to bone and brain tissue, affordability, and the ability to promote native dural healing. A qualitative review by Azzam et al. [4] identified the most suitable material for various clinical settings. For example, xenografts were preferred for rapid closure in traumatic SDH evacuation during decompressive craniectomy, allografts for watertight closure in Chiari decompression, and synthetics for their non-adhesive properties to neural tissue during tumor resection procedures, which help prevent complications in recurrent tumor operations. However, this study identified moderate risks in both groups and recommended further research to analyze the effectiveness and characteristics of dural graft materials [4].

In this systematic review, we aim to compare the risk of complications between autologous and non-autologous dural graft materials.

The search strategy yielded a total of 1,646 articles, with an additional 4 potential studies identified through a manual search. After the screening process, 13 studies were selected for data extraction. However, one clinical trial by Turchan et al. [36] was excluded due to non-estimable data for analysis, leaving 12 studies included in the final analysis (Fig. 1).

The characteristics and qualitative analyses of these included studies are summarized in Table 1 and 2. Four studies favored the use of autologous grafts, while two studies favored non-autologous grafts. The remaining six studies showed similar efficacy between autologous and non-autologous graft groups.

A study by Hoffman et al. [12] demonstrated that the use of allografts (AlloDerm) in Chiari decompression increased the odds of CSF leakage compared to autologous materials (pericranium). Another study of duraplasty in Chiari malformation, conducted by Vanaclocha et al. [37], also showed more favorable outcomes with pericranium compared to allografts (freeze-dried cadaveric dura). Zhao et al. [41] concluded that the meningitis rate was lower in the cervical fascia group compared to the artificial dura group. Wang et al. [39] conducted a study using the nuchal ligament, demonstrating that it was safer, more cost-effective, less time-consuming, and significantly reduced postoperative complications compared to non-autologous dural grafts.

On the other hand, Totten et al. [35] showed that duraplasty using xenograft (porcine small intestinal submucosa) resulted in a lower CSF leak rate compared to the autologous graft group. A similar reduction in CSF leak rate using synthetic grafts was observed in a study by Walcott et al. [38] A summation of the qualitative analysis indicated that the autologous graft group was generally preferable, with better outcomes compared to the non-autologous group.

The risk of bias in each included study is presented in Table 3. Fig. 2 and Fig. 3 showed forest plot and funnel plots based on the results of a study comparing complication risks between autologous and non-autologous dural graft materials. The funnel plots are symmetric, indicating no publication bias.

Twelve studies, comprising a total of 1,877 patients, were included in the analysis. Of these, 965 (51.4%) procedures used autologous dural graft materials, while 912 (48.6%) procedures used non-autologous graft materials. The cases included in the analysis involved Chiari Malformation type I, neoplasms, vascular lesions, and trauma.

Autologous grafts were used in 629 of the 965 patients (65.2%). Among these, 415 (66%) received a pericranial graft, which was the most frequently used in 7 studies, 144 (22.9%) received cervical fascia (3 studies), 67 (10.7%) received nuchal ligaments (3 studies), and 3 (0.5%) received fascia lata (2 studies). Additionally, 336 unspecified autologous grafts were reported in 2 studies. For the non-autologous group, 407 patients (44.6%) received xenografts (3 studies), 365 (40%) received synthetic grafts (7 studies), and 110 (15.4%) received allografts (5 studies).

Autologous dural graft materials generally elicit a better recipient response than artificial grafts due to their inherent properties. Autologous grafts are immunologically inert and do not trigger tissue scarring. However, when autologous materials are not feasible, artificial dura serves as an alternative. This study compares the outcomes between these two types of grafts and discusses which materials are preferable.

In terms of hydrodynamic profile, we found no overall difference between autologous and non-autologous dural graft materials in preventing CSF leakage. However, pairwise comparisons showed that pericranium and nuchal ligament had superior outcomes compared to some other graft subtypes. Xenografts were more effective at preventing CSF leakage than fascia lata (p=0.0310) and allografts (p=0.0108). This finding aligns with a study by Laun et al. [20], which compared allografts (human cadaveric dura) with xenografts (bovine pericardium). Their results demonstrated similarly low complication rates (wound infection) for both, but the characteristics of bovine pericardium were superior to cadaveric dura mater [20].

Other non-autologous graft subtypes also exhibited excellent outcomes in terms of hydrodynamic performance. An experimental study by Turchan et al. [36] analyzed the watertightness of an allograft by injecting saline into the subdural cavity during cranioplasty several months after decompressive craniectomy. The study revealed that the use of an allograft (human amniotic membrane) was as effective as an autologous graft (temporal fascia) in preventing CSF leaks.

Amniotic membrane patches formed a uniform layer with the dura, well-demarcated from the cerebral cortex in almost all patients, serving as an excellent scaffold for subsequent reconstructive surgery during cranioplasty. Outcomes in the amniotic membrane group were consistent with those of biological dural substitutes. Microscopic analysis of specimens showed perfect integration of the amniotic membrane with the autologous dura, along with the disappearance of the epithelium and the formation of new connective tissue, thin plates of fibrous tissue, and no inflammatory reaction or necrosis. These results are likely attributed to the anti-adhesive and pluripotential differentiation properties of amniotic membranes [21]. Human amniotic membrane is also a non-autologous dural substitute known for its antifibrotic effects, which suppress scarring [30].

In a study involving reoperative craniotomy for recurrent glioma, an amniotic membrane previously placed during the original craniotomy was found intact and fully integrated into the surrounding dura, with minimal fibrosis or scar tissue [9]. Kshettry et al. [17] observed that dural reconstruction with a non-autologous collagen onlay graft resulted in a higher rate of incisional CSF leaks in surgical approaches near the fourth ventricular outflow tract (i.e., midline suboccipital and far lateral approaches).

In general, autologous materials are superior to non-autologous materials when it comes to the prevention of infection and tissue rejection. This was demonstrated by the analysis, which showed lower rates of meningitis (OR=0.31; 95% CI 0.17-0.54) and wound infection (OR=0.34; 95% CI 0.14-0.80) in the autologous graft group compared to the non-autologous group. While synthetic dural grafts made from polypropylene have been used safely for both single- and double-layered duraplasty, they can occasionally trigger foreign body reactions and, in rare cases, present as an extradural abscess [29]. In another study, excessive granulation tissue formation as part of a foreign body reaction was observed in four patients shortly after receiving ReDura as a dural substitute, leading to diagnostic challenges and increased healthcare costs [23]. Pairwise comparisons showed that pericranium, nuchal ligament, allograft, and xenograft were preferred over synthetic materials (p=0.0001; p=0.0145; p=0.0005; p=0.0005). This finding is supported by a recent meta-analysis discussing the risk of meningitis in Chiari decompression, which highlights lower rates of meningitis and reoperation associated with autologous grafts compared to various non-autologous grafts [14].

The pseudomeningocele rate was also reduced with autologous grafts compared to non-autologous grafts. The pooled effect showed a significant reduction with minimal heterogeneity (OR=0.50; 95% CI 0.32-0.79; P=0.003; I²=10%). Pairwise comparisons indicated that pericranium and nuchal ligament had better outcomes compared to other graft materials in preventing pseudomeningocele formation. This finding is supported by studies from Yahanda et al. [40] and Balasa et al. [5], which reported higher pseudomeningocele rates with non-autologous graft materials.

The hydrocephalus rate (OR=0.79; 95% CI 0.39-1.60) and revision surgery rate (OR=1.80; 95% CI 0.89-3.62) were similar between the two groups. However, the studies included in the analysis were limited, and further research is needed to explore these complications in duraplasty procedures.

Prior studies have suggested that complication rates with various types of dural graft materials are similar. Our findings, however, indicate that autologous grafts are preferable compared to other graft types. Nevertheless, the choice of material also depends on local protocols and individual surgeon preferences. Other factors to consider when selecting dural graft materials include operation time and cost-effectiveness. A study by Sabatino et al. [28] concluded that autologous grafts, particularly galea-pericranium, were preferable due to their ease of harvesting and short procedure duration (1.45 ± 0.38 minutes on average), with no additional costs. Other autologous graft materials, such as the nuchal ligament, were also tested by Cools et al. [7], who found that the nuchal ligament was cost-efficient, easily harvested, and durable, with favorable outcomes. On the other hand, a clinical trial by Filippi et al. [10] demonstrated that solvent-preserved bovine pericardium is a safe, suitable, and affordable material for duraplasty, with favorable implantation characteristics and clinical outcomes. These findings suggest that cost, harvest time, and individual graft characteristics play a significant role in determining the overall outcome and length of hospital stay for patients.

Our findings indicate that autologous dural graft materials may reduce the risk of complications such as meningitis, pseudomeningocele, and wound infections compared to non-autologous materials. However, the choice of graft material remains a subject of debate due to multiple influencing factors, including surgical approach, patient-specific considerations, and institutional protocols.

One critical limitation of this meta-analysis is the heterogeneity of the included studies. The sample sizes varied widely, ranging from 12 to 399 patients per study. Additionally, nearly half of the cases analyzed in this meta-analysis were related to Chiari malformation, a condition that may not be representative of other neurosurgical indications requiring duraplasty. As a result, the applicability of our findings to broader populations, including patients undergoing duraplasty for trauma, tumor resection, or vascular surgery, may be limited.

The heterogeneity index for CSF leakage rates (I²= 49%) suggests moderate variability across the included studies. This variation likely stems from differences in study design (retrospective vs. prospective), patient demographics, surgical techniques, and follow-up durations. Although we used a random-effects model for the analyses to account for this variability, residual heterogeneity remains an important factor to consider when interpreting the results.

Furthermore, while autologous grafts demonstrated favorable outcomes in reducing infection rates and pseudomeningocele formation, the benefits were not uniformly observed across all complications. For instance, our analysis found no significant difference in CSF leakage, hydrocephalus, or revision surgery rates between autologous and non-autologous graft materials. These findings suggest that while autologous grafts may be preferable for reducing certain postoperative complications, other factors such as surgical technique, patient comorbidities, and intraoperative handling of graft materials also play crucial roles in determining outcomes.

Given these considerations, we recommend that future research focus on randomized controlled trials with larger sample sizes and standardized protocols to better delineate the benefits and limitations of different dural graft materials in diverse neurosurgical populations.

Our meta-analysis results indicate that the use of autologous dural graft materials may reduce the risk of complications, including meningitis, pseudomeningocele, and wound infection, compared to non-autologous materials. Therefore, when the dural defect can be sufficiently covered using available autologous grafts, we strongly recommend their use for dural closure. However, when autologous materials are not an option, we recommend selecting a non-autologous dural substitute that meets the criteria and considerations previously discussed. We also highlight the need for further research to determine the optimal choice among non-autologous dural substitutes.