Hyperglycemic Crises in Adult Patients With Diabetes: A Consensus Statement From the American Diabetes Association

Response to Rosival

In response to the comments by Dr. Rosival (1), we point out that impaired mental status is present in 52% of adult patients with diabetic ketoacidosis (DKA) and 74% of patients with hyperglycemic hyperosmolar syndrome (HHS) (2). Although most patients are lethargic or stuporous on arrival, 13% of admissions for DKA and 23% for HHS present with loss of consciousness. In previous reports (3–6), we evaluated the relationship between the initial biochemical data and level of consciousness in adult patients with DKA (4). In 1980, we reported that age, admission plasma glucose, bicarbonate, blood urea nitrogen, and osmolality correlated with the level of consciousness; however, arterial pH and serum ketone bodies did not (4). Coma was uncommon in patients younger than 40 years old, and in both groups it was most frequent in patients aged >65 years. Furthermore, in a subsequent study of 132 patients with DKA, we found a significant correlation between serum osmolality and level of consciousness. A serum osmolality >320 mOsm/kg was observed in stuporous or comatose patients, while an osmolality <320 mOsm/kg was observed in alert patients with DKA (5). More recently, in 144 consecutive patients with DKA and 23 subjects with HHS, we reported that loss of consciousness correlated better with serum osmolality than with severity of metabolic acidosis. In patients with DKA, the mean ± SEM serum osmolality and arterial pH was 311 ± 3 mmol/kg and 7.18 in noncomatose patients and 345 ± 4 mmol/kg and 7.03 in comatose patients, respectively (P < 0.01 for both). Furthermore, in the absence of significant metabolic acidosis in patients with HHS, we observed a positive correlation between level of consciousness and serum osmolality (2).

Dr. Rosival reports a contradiction between our findings, which are summarized in the recent ADA consensus statement in adult subjects (6), and a recent retrospective study in 225 children with DKA (7). In this cohort, either pH or bicarbonate measurements were available on admission in 218 episodes: pH in 174, plasma bicarbonate in 212, and both measurements in 168. Only four of these patients presented with loss of consciousness. In this report, the admission consciousness level was related to the degree of acidosis rather than to blood glucose, sodium, or osmolality level. Level of consciousness also appeared to be related to age, with younger children more likely to have a reduced consciousness level.

It is not clear what explains the differences between admission levels of consciousness and admission biochemical data in adults and children with DKA. We believe that duration of DKA may be an important factor. Thus, in children, the duration of DKA may be shorter than in adults, limiting their exposure to acute hyperglycemia. Therefore, the degree of acidosis may play a more prominent role in the level of consciousness in children with DKA.