To characterise the spectrum of central nervous system (CNS) conditions associated with paediatric sudden death a retrospective autopsy study was conducted. The most common causes of SUD were epilepsy, infection or haemorrhage. Patients presented either with rapid onset of fulminant sepsis, with minor non-specific symptoms including fever or headache, or with no apparent symptoms or signs. The latter group of patients were most often found dead in bed, which led to clinical confusion with SIDS in those patients <1 yr. of age. Gross autopsy findings were occasionally unremarkable in some cases of sepsis which meant that diagnosis depended on adequate histological and microbiological sampling. Terminal presentations and gross autopsy findings in patients with haemorrhage were similar in spite of different underlying conditions; i.e. patients presented with sudden collapse and death and were found to have haemorrhage due to unsuspected arteriovenous malformations or tumours, a berry aneurysm, and a connective tissue disorder (Ehlers-Danlos syndrome). Diagnosis occasionally required extensive sampling for microscopy, along with specialised techniques such as collagen analysis. Several patients with bleeding diatheses due to leukaemia died suddenly from intracerebral bleeding. Other unusual causes of sudden death included a colloid cyst, an intracardiac echinococcal cyst embolus and a paradoxical Wilms’ tumour embolus. These cases demonstrate that 1) infants with unsuspected CNS pathology may present clinically as SIDS, 2) intracranial conditions more commonly found in adults may cause SUD in childhood, and 3) that quite diverse conditions with different clinical and genetic implications may present identically. Careful examination of the brain, including cerebral vessels, with adequate histological sampling, is therefore essential in all cases of paediatric SUD. Routine CSF and blood cultures are necessary, and fresh and frozen skin for fibroblast culture and collagen analysis may need to be taken when haemorrhage is detected, to exclude the possibility of an occult connective tissue disorder. To characterise the spectrum of central nervous system (CNS) conditions associated with paediatric sudden death a retrospective autopsy study was conducted. The most common causes of SUD were epilepsy, infection or haemorrhage. Patients presented either with rapid onset of fulminant sepsis, with minor non-specific symptoms including fever or headache, or with no apparent symptoms or signs. The latter group of patients were most often found dead in bed, which led to clinical confusion with SIDS in those patients <1 yr. of age. Gross autopsy findings were occasionally unremarkable in some cases of sepsis which meant that diagnosis depended on adequate histological and microbiological sampling. Terminal presentations and gross autopsy findings in patients with haemorrhage were similar in spite of different underlying conditions; i.e. patients presented with sudden collapse and death and were found to have haemorrhage due to unsuspected arteriovenous malformations or tumours, a berry aneurysm, and a connective tissue disorder (Ehlers-Danlos syndrome). Diagnosis occasionally required extensive sampling for microscopy, along with specialised techniques such as collagen analysis. Several patients with bleeding diatheses due to leukaemia died suddenly from intracerebral bleeding. Other unusual causes of sudden death included a colloid cyst, an intracardiac echinococcal cyst embolus and a paradoxical Wilms’ tumour embolus. These cases demonstrate that 1) infants with unsuspected CNS pathology may present clinically as SIDS, 2) intracranial conditions more commonly found in adults may cause SUD in childhood, and 3) that quite diverse conditions with different clinical and genetic implications may present identically. Careful examination of the brain, including cerebral vessels, with adequate histological sampling, is therefore essential in all cases of paediatric SUD. Routine CSF and blood cultures are necessary, and fresh and frozen skin for fibroblast culture and collagen analysis may need to be taken when haemorrhage is detected, to exclude the possibility of an occult connective tissue disorder.