Craniocerebral War Missile Injuries: Clinical and Radiological Study S. Jankovic  1, N. Bes Ïenski2,Z Ï . Bus Ïic  3,Z Ï . Dujic  4, A. Buc Ïa1, J. Mas Ïkovic  1, I. Lus Ïic  5, and D. Primorac6 1 Department of Radiology, University Hospital Split, Split, Croatia 2 Department of Radiology, University Hospital Rebro, Zagreb, Croatia 3 Department of Neurosurgery, University Hospital Split, Split, Croatia 4 Department of Physiology, Split University School of Medicine, Split, Croatia 5 Department of Neurology, University Hospital Split, Split, Croatia 6 Laboratory for Clinical and Forensic Genetics, University Hospital Split, Croatia Summary In this study we reviewed the initial clinical and radiological management and early outcomes of 176 consecutive patients from the war in Croatia. Keywords: War injury; missile injury; intracranial foreign bodies; computed tomography. Introduction Despite the fact that numerous articles have been published on this topic it is clear that di¨erent authors suggest slightly distinct approaches in diagnosis and treatment of patients with craniocerebral missile injuries [2±4]. Here, we present our experiences in the management of 176 patients with craniocerebral mis- sile injuries sustained during the war in Croatia from 1991 to 1995. Patients and Methods During the war in Croatia, 176 patients (86,4% were soldiers, and 13,6% civilians) with craniocerebral missile injuries were hospitalized at the Department of Neurosurgery, University Hospital Split. The mean GCS score for 156 patients on admission was 10 (range 3±15), while twenty patients were comatose. After initial evaluation and stabilization all patients underwent plain X-ray skull radiography and computed tomography. Patients with extensive entry wounds were operated on and after craniec- tomy, debridement of devitalized tissue and debris, removal of the visible foreign bodies and some bone or metal fragments visualized on the pre-operative NCT was done, followed by dural closure with fascia lata, periosteum or rarely with lyophilized dura. For statistical analysis chi-square test with Yates correction was used. Probability of less than 0,05 was considered statistically signi®cant. Results By using clinical and radiological procedures we observed the following neurological complications: CSF ®stula in 19, meningo-encephalitis in 13 patients (7 of them also had CSF ®stula), seizures in 8, hydro- cephalus in 2, while brain abscess occurred in 9 patients. Nineteen patients (10,7%) died during the hospitalisation. Discussion Complications which occurred in 38,6% of all cases, were mostly postoperative cerebrospinal ¯uid ®stulas (39%) and infections (16,5%). In our work, we did not ®nd any statistically important connection between postoperative cerebrospinal ¯uid ®stulas and the kind of the substitute used for the dural defect closure. Furthermore, we found that the basal cerebrospinal ¯uid ®stulas signi®cantly a¨ected the occurrence of complications, especially infections (chi square  4,57; p  0;032). We also observed a statistically im- portant connection between remaining bone fragments and the occurrence of infection (chi square  9; 42; p  0;002) a di¨erent result from Aarbi [1]. It is important to emphasise that all patients with CNS infection and retained foreign bodies had at least two other risk factors such as GCS U 5, extensive brain injuries, CSF ®stula, and craniobasal penetrat- ing injuries [5]. Of the total number of all patients 10 (5,6%) died immediately after hospitalisation, and 9 (5,1%) died during the treatment as a result of compli- Acta Neurochirurgica > Springer-Verlag 2000 Printed in Austria Acta Neurochir (Wien) (2000) 142: 101±102