Randomized trial compares autologous and synthetic grafts for cranial duraplasty safety

When neurosurgeons cannot close the dura mater directly after cranial surgery, they must patch the tough membrane that envelops the brain with a substitute graft. The trial, conducted between January and December 2021 and reported in the Chinese Neurosurgical Journal, found that early postoperative complications were numerically more frequent with the synthetic material, but the difference did not reach statistical significance, and the authors explicitly caution that the result should not be read as proof that the two materials are equivalent.

The study enrolled 121 consecutive patients who required dural repair or substitution because their dural defects were unsuitable for primary closure. All provided informed consent, and a computer-generated allocation sequence assigned them to one of two groups using sealed opaque envelopes opened only at the time of surgery. Group A received a galea-pericranial graft harvested from the reflected scalp flap, tailored to the defect size and kept moist in normal saline until implantation. Group B received a high-density polypropylene patch trimmed to size and prepared according to the manufacturer’s recommendations, Surgiwear Ltd. of India. In both arms, the graft was sutured to the native dura with interrupted or continuous non-absorbable sutures to achieve watertight closure, with additional sutures placed to eliminate gaps at the graft–dura interface.

Not every randomized patient contributed to the headline numbers. Twenty-one patients died within 30 days of surgery, nine in the pericranium group and twelve in the G-Patch group, and were excluded from the primary complication analysis because their outcomes could not be reliably assessed. The investigators report that these early deaths occurred predominantly in patients with severe traumatic brain injury and malignant intracranial lesions and were attributed to progression of the underlying neurological disease rather than to graft-related problems. The final analyzed cohort therefore comprised 100 patients, 50 per arm, all of whom completed the 30-day follow-up with no losses. The authors acknowledge that excluding early deaths may introduce survivorship bias, and they addressed this concern with a sensitivity analysis that folded the deaths into a composite adverse outcome of any complication or death.

The cohort reflected the trauma-heavy case mix of a busy Indian neurosurgical service. Traumatic brain injury was the most common indication at 58 percent of patients, with brain tumors accounting for 33 percent. Sixty-one percent of the analyzed patients were male, and the largest age band was 30 to 44 years at 27 percent. Most patients, 82 percent, had no comorbidities, with hypertension at 7 percent and diabetes mellitus at 6 percent the most frequent among those who did. Two-thirds reported no substance use; among users, alcohol at 18 percent led smoking at 9 percent and tobacco chewing at 6 percent. Surgical location was broadly balanced: 47 supratentorial and 3 infratentorial procedures in the pericranium group versus 46 supratentorial and 4 infratentorial in the synthetic group, though these figures were presented descriptively only.

The primary outcome was the occurrence of any postoperative complication within 30 days, spanning cerebrospinal fluid leak, wound infection, subdural empyema, subcutaneous CSF collection, meningitis, wound dehiscence, and bone flap osteitis, each with a pre-specified clinical definition. Daily wound examinations continued through hospitalization and at the 30-day visit. Complications arose in 26 percent of the synthetic graft group, 13 of 50 patients, compared with 12 percent, 6 of 50, in the autologous group, meaning 74 percent versus 88 percent of patients enjoyed an uneventful recovery. Wound infection was the most common complication in both arms, affecting 16 percent of polypropylene patients and 8 percent of pericranium patients, while the remaining complication types appeared at low frequencies across both groups.

Statistically, the unadjusted odds ratio for any complication with the synthetic graft was 2.58, with a 95 percent confidence interval of 0.88 to 7.54 and a p-value of 0.08. Because the confidence interval crosses 1, the data cannot rule out either no difference or a substantially larger difference. In a limited multivariable logistic regression adjusting for primary diagnosis, graft type remained a non-significant predictor, with an adjusted odds ratio of 1.84, a 95 percent confidence interval of 0.54 to 6.27, and a p-value of 0.33, while traumatic brain injury itself showed a stronger though still non-significant association with complication risk. The composite sensitivity analysis including the 21 early deaths yielded an odds ratio of 1.98, with a 95 percent confidence interval of 0.91 to 4.30 and p equal to 0.08, and did not materially change the overall interpretation.

Recognizing that randomization had not fully balanced the underlying surgical indications, the team ran a post-hoc exploratory analysis confined to the traumatic brain injury subgroup of 58 patients. There, complication rates were nearly identical, 15 percent with high-density polypropylene versus 14 percent with autologous pericranium, but the authors stress that this comparison is limited by its small sample size and deserves caution. Exploratory descriptive analyses of comorbidities and substance use found no clear trends; wound infections occurred predominantly in patients without recorded comorbidities, and most complications clustered among patients without substance abuse histories, but sparse event counts and overlapping patient categories meant the investigators drew no inferential statistical conclusions from these tables.

The study’s design choices shape how much weight its findings can bear. It was reported according to CONSORT guidelines, performed by a single experienced neurosurgeon using a standardized operative protocol to minimize technical variability, and involved no external funding or industry support, with synthetic grafts procured through routine institutional supply. Blinding was impossible given the nature of the intervention, and no a priori sample size or power calculation was performed; the final sample simply reflected the eligible cases treated during the twelve-month window. Notably, this investigator-initiated pragmatic randomized study was never prospectively registered in a clinical trial registry, a transparency gap the authors disclose openly.

The limitations section is unusually candid. Beyond the modest sample size and lack of power for stratified or multivariable analyses, the imbalance in underlying diagnoses, traumatic brain injury predominating in the synthetic group and tumor procedures in the pericranium group, introduces potential confounding by indication, since etiology, defect characteristics, contamination risk, and adjuvant therapies all influence postoperative complication rates. The cohort also predominantly involved cranial procedures with relatively low risk of high-flow CSF leakage, so the results may not generalize to skull base surgery, transsphenoidal approaches, cranialization procedures, or spinal dural repair, where CSF dynamics differ significantly. The 30-day window captures early outcomes only, saying nothing about long-term graft integration or delayed complications, and no intraoperative variables, neurological recovery, functional status, or quality-of-life measures were assessed.

Against this backdrop, the authors’ conclusion is measured: both autologous pericranium and high-density polypropylene demonstrated acceptable short-term safety profiles within this heterogeneous cohort, with no statistically significant difference in early complications even though complications were more numerous in the synthetic arm. They emphasize that the absence of significance should not be interpreted as evidence of equivalence, given the limited sample size, short follow-up, diagnostic imbalance, and absence of stratified analyses. Practical considerations may still guide graft selection in resource-constrained settings: pericranium harvesting added minimal operative time and avoided material costs, while the synthetic patch offered immediate availability, with institutional supply ensuring no additional financial burden fell on patients. Larger studies with longer follow-up, indication-specific analyses, and functional outcome measures, the authors write, are needed to define the comparative performance of dural substitutes more definitively.

Subject of Research: Early postoperative complications of autologous pericranium versus high-density polypropylene dural substitutes in cranial duraplasty

Article Title: A comparative analysis of early postoperative outcomes of autologous pericranium versus high-density polypropylene in cranial duraplasty in a tertiary care center in Central India

Article References: Gupta, S., Gupta, R., Sheikh, Z., Jain, K., & Kumar, S. (2026). A comparative analysis of early postoperative outcomes of autologous pericranium versus high-density polypropylene in cranial duraplasty in a tertiary care center in Central India. Chinese Neurosurgical Journal, 12(1), Article 21. https://doi.org/10.1186/s41016-026-00440-y

Image Credits: AI Generated

DOI: 10.1186/s41016-026-00440-y

Keywords: cranial duraplasty, dural substitute, autologous pericranium, high-density polypropylene, G-Patch, randomized trial, postoperative complications, CSF leak, wound infection, neurosurgery, traumatic brain injury, 30-day outcomes