# Decompressive craniectomy in traumatic brain injury: insights from a 15-year multicentre cohort in Sweden

## TL;DR

This study examines decompressive craniectomy practices and outcomes in Sweden over 15 years, finding variable practices and mixed real-world results compared to clinical trials.

## Contribution

The study provides real-world insights into DC practices and outcomes in Sweden, highlighting discrepancies with randomized trial findings.

## Key findings

- DC rates remained stable, with regional variations in timing and techniques.
- Radiological improvement post-DC was significant, but re-operations and complications occurred in notable proportions.
- Real-world outcomes were more favorable than in recent RCTs, though 60% had unfavorable 6-month outcomes.

## Abstract

Decompressive craniectomy (DC) is a last-resort treatment for severe traumatic brain injury (TBI) with refractory intracranial hypertension. Randomized controlled trials (RCTs) report mixed and sometimes conflicting results, leaving uncertainties regarding indications, timing, and long-term benefits. This study explored DC practices and outcomes in a contemporary Swedish setting contextualised in modern RCT evidence.

This retrospective multicentre study included 299 TBI patients who underwent DC between 2008 and 2022 across four Swedish neurosurgical centres. Clinical, radiological, surgical, and outcome data (6-months Glasgow Outcome Scale) were collected. Differences across centres and between adults/children were analysed.

Annual DC rate remained stable over 15 years, modestly declining from 3.6 to 3.2 per million inhabitants. Significant regional differences were observed in timing, indications, and techniques. Proportion of primary versus secondary DC and surgery timing remained unchanged, though bifrontal DC decreased. Patients were young (median age 37), predominantly male (76%), severely injured (GCS M < 6), and 48% had unreactive pupils. Radiological improvement in mass effect post-DC (midline shift, basal cisterns) was significant (p < 0.001). Re-operation for haemorrhage occurred in 10%, complementary decompression, surgical-site infection, and subdural hygroma each occurred in ~ 5%. At 6 months, 60% had unfavourable outcomes and 11% were deceased. Higher age, lower GCS, comorbidities, impaired pupillary reactivity and obliterated basal cisterns independently predicted unfavourable outcome.

Landmark RCTs appear to have had limited influence on Swedish DC practice, which remains variable across centres. Real-world outcomes were more favourable than in recent RCTs and other acute brain injuries.

The online version contains supplementary material available at 10.1186/s13049-026-01585-6.