Sudden unexpected death in epilepsy (SUDEP) is a leading cause of premature death in patients with epilepsy. One hypothesis proposes that sudden death is mediated by post-ictal central respiratory depression, which could relate to underlying pathology in key respiratory nuclei and/or their neuromodulators. Our aim was to investigate neuronal populations in the ventrolateral medulla (which includes the putative human pre-Bötzinger complex) and the medullary raphe. Forty brainstems were studied comprising four groups: 14 SUDEP, six epilepsy controls, seven Dravet syndrome cases and 13 non-epilepsy controls. Serial sections through the medulla (from obex 1 to 10 mm) were stained for Nissl, somatostatin, neurokinin 1 receptor (for pre-Bötzinger complex neurons) and galanin, tryptophan hydroxylase and serotonin transporter (neuromodulatory systems). Using stereology total neuronal number and densities, with respect to obex level, were measured. Whole slide scanning image analysis was used to quantify immunolabelling indices as well as co-localization between markers. Significant findings included reduction in somatostatin neurons and neurokinin 1 receptor labelling in the ventrolateral medulla in sudden death in epilepsy compared to controls (P < 0.05). Galanin and tryptophan hydroxylase labelling was also reduced in sudden death cases and more significantly in the ventrolateral medulla region than the raphe (P < 0.005 and P < 0.05). With serotonin transporter, reduction in labelling in cases of sudden death in epilepsy was noted only in the raphe (P ≤ 0.01); however, co-localization with tryptophan hydroxylase was significantly reduced in the ventrolateral medulla. Epilepsy controls and cases with Dravet syndrome showed less significant alterations with differences from non-epilepsy controls noted only for somatostatin in the ventrolateral medulla (P < 0.05). Variations in labelling with respect to obex level were noted of potential relevance to the rostro-caudal organization of respiratory nuclear groups, including tryptophan hydroxylase, where the greatest statistical difference noted between all epilepsy cases and controls was at obex 9-10 mm (P = 0.034), the putative level of the pre-Bötzinger complex. Furthermore, there was evidence for variation with duration of epilepsy for somatostatin and neurokinin 1 receptor. Our findings suggest alteration to neuronal populations in the medulla in SUDEP with evidence for greater reduction in neuromodulatory neuropeptidergic and mono-aminergic systems, including for galanin, and serotonin. Other nuclei need to be investigated to evaluate if this is part of more widespread brainstem pathology. Our findings could be a result of previous seizures and may represent a pathological risk factor for SUDEP through impaired respiratory homeostasis during a seizure.