Blood glucose and prognosis in children with presumed severe malaria: is there a threshold for 'hypoglycaemia'?

摘要: Summary objectives Hypoglycaemia (glucose 8.3 mmol ⁄l. Associations between glycaemia and case fatality were analysed for 418 children using logistic regression models a receiver operator curve (ROC). results There was significant difference blood glucose levels in who died (median 4.6 ⁄l) survivors 7.6 ⁄l, P < 0.001). Case declined from 61.5% of the hypoglycaemic to 46.2% those with low glycaemia, 13.4% normal 7.6% hyperglycaemia (P Logistic showed an adjusted odds ratio (AOR) 0.75 (0.64‐0.88) per 1 ⁄l increase baseline glucose. Compared glucose, hypoglycaemia both significantly increased death (AOR 11.87, 2.10‐67.00; 5.21, 1.86‐14.63, respectively), whereas reduced 0.34, 0.13‐0.91). The ROC [area under at 0.753 (95% CI 0.684‐0.820)] indicated that had moderate predictive value identified optimal threshold <6.1 (sensitivity 64.5% specificity 75.1%). conclusions If there is which defines worse prognosis, it higher level than current definition 2.2