Challenges in identifying and managing cerebral oedema at first presentation of type 1 diabetes

Abstract

An 11-year-old girl with known asthma presented to the emergency department with a 24-hour history of dyspnoea and chest pain. Examination revealed tachycardia, tachypnoea with normal oxygen saturation in air, intermittent drowsiness, poor perfusion peripherally and clinical dehydration with a central capillary refill time of 3 seconds. Initial investigations (electrocardiogram and chest X-ray) ruled out arrhythmia and/or pneumothorax. Subsequently, a venous blood gas at 2 hours (Table 1) revealed marked metabolic acidosis and hyperglycaemia and severe diabetic ketoacidosis (DKA) was confirmed in the presence of ketonuria. As per DKA guidance1 fluid resuscitation was given, followed by correction for 8% dehydration over 48 hours alongside maintenance fluids, followed by insulin after an hour. The patient continued to have severe acidosis and tachycardia, Glasgow Coma Scale dropped to 8 (E2M4V2) and cerebral oedema (CE) secondary to DKA was suspected, and thus hypertonic saline (5 ml/kg of 3%) was administered. As the patient remained cardiovascularly unstable, a further fluid bolus was administered followed by intubation and ventilation; CT-head was in keeping with CE. A third fluid bolus was administered for circulatory shock and inotropes commenced. Maintenance fluids were halved and correction time for dehydration extended to 72 hours. Intravenous calcium was administered for myocardial protection and antibiotics for presumed sepsis. Sodium bicarbonate was considered. The parents were given a grave prognosis, but the patient eventually stabilised. Subsequent treatment was tailored to allow for controlled rise of PaCO2, fall in blood glucose and correction of metabolic acidosis (Table 1). A full clinical recovery with no acute neurological deficits was observed. Retrospectively, a history of weight loss despite an increased appetite over a month was established but osmotic symptoms (polyuria, polydipsia) and lethargy were denied. HbA1c at presentation was 70 mmol/mol (8.6%) and HbA1c six months from diagnosis is 52 mmol/mol (6.9%). Cerebral oedema is an uncommon but potentially devastating consequence of DKA. In the UK, the overall incidence of CE is reported as 6.8 per 1000 episodes of DKA, with the incidence being higher in new (11.9 per 1000 episodes) cases of type 1 diabetes compared to those with an established diagnosis (3.8 per 1000).1 Cerebral oedema has a high mortality (20–25%) and morbidity with 15–35% of survivors having permanent neurologic sequelae.2 CE typically develops 4–12 hours following initiation of treatment, but may occur any time within the first 24 hours.2 The pathophysiology of onset and progression of CE remains poorly understood, although various theories have been proposed (ischaemic, vasogenic and osmotic)3 as well as factors attributed to disease state and treatment initiation. Factors associated with disease state are greater hypocapnia at presentation after adjusting for degree of acidosis, increased serum urea nitrogen, and more severe acidosis.4 Treatment-associated factors are use of bicarbonate, attenuated rise in sodium during therapy, greater fluid volumes in the first 4 hours, and administration of insulin in the first hour of fluid treatment.4 Demographic factors such as younger age at onset, new diagnosis of diabetes and long-standing symptoms have also been implicated. Risk factors associated with the disease state in our patient were the severity of acidosis at presentation, elevated urea nitrogen (8.5 mmol/L), severe dehydration (serum osmolality 332 mOsm/kg) and hypocapnia. Although osmotic symptoms were denied, the combination of weight loss despite hyperphagia and severe dehydration at diagnosis indicates longer-standing disease. The fluctuating Glasgow Coma Scale on initial assessment suggests that CE had already developed which was then aggravated by excessive use of fluids in the first 4 hours. The new DKA guidelines recommend a single fluid bolus for patients in shock only5 and advise timely discussion with a diabetes consultant to further reduce the risk of CE in these patients. However, the role of fluids in the causation of CE remains inconclusive as reduced fluid regimens have yet to demonstrate lower rates of CE. A similar incidence in the United States and the United Kingdom despite use of fluids with different sodium content highlights gaps in the understanding of the development of CE.6 The International Society for Pediatric and Adolescent Diabetes’ recommended treatment of CE includes reduction in the rate of fluid administration by one-third, use of mannitol and/or hypertonic saline, elevation of the head end of the bed, and ventilation while avoiding hyperventilation.4 Maintenance fluid administration in the new DKA guidelines is considerably lower when compared to the previous guidelines, with a fluid reduction of up to 35% in severe DKA and 60% in mild DKA in very young children. Outcome data are awaited and will determine whether the change in fluid regimen reduces the incidence of CE in children.