Factors determining the appearance of glucose in upper and lower respiratory tract secretions.

(a). To describe the glucose content of normal human airways secretions; (b). to observe the effects of hyperglycemia and airways inflammation on airways glucose. Observational studies. (a). St George's Hospital Medical School; (b). diabetes mellitus outpatient clinics; (c). adult general intensive care unit. Nineteen healthy volunteers, 24 volunteers with acute rhinitis, 20 patients with diabetes mellitus, and 60 patients admitted to a general adult intensive care unit. (a). Non-ventilated patients: simultaneous measurement of blood and nasal glucose concentrations; (b). ICU patients: simultaneous blood, nasal, and endotracheal (ET) glucose concentrations. Nasal glucose was undetectable in all healthy volunteers. Glucose was detected in 12/24 volunteers with acute viral rhinitis [1 (1-2) mmol l(-1)] and 18/20 people with diabetes [4 (2-7) mmol l(-1)]. Glucose was detected in the ET secretions of 31/60 ventilated patients on ICU. Patients with ET glucose had significantly higher blood glucose (9.8+/-0.4 mmol l(-1)) than patients without ET glucose (7.2+/-0.3 mmol l(-1), P<0.001), and all patients with blood glucose >10.1 mmol l(-1) had glucose in ET secretions. Enteral nutrition did not affect the presence or concentration of glucose in ET secretions. Glucose is not normally present in airways secretions, but appears where hyperglycaemia or epithelial inflammation occur. The detection of glucose cannot reliably be used to detect enteral feed aspiration.