Serum fructosamine as a marker of 5-year risk of developing diabetes mellitus in patients exhibiting stress hyperglycaemia.

AIMS: We examined whether the level of random serum glucose (RSG) in subjects exhibiting stress hyperglycaemia is a useful marker of the future risk of developing diabetes mellitus (DM), and whether serum fructosamine is of any additional value.
METHODS: All non-diabetic adults attending Accident and Emergency in 1994-1995, who had venesection, were studied. Serum fructosamine and RSG were routinely measured in all such patients. Using the laboratory biochemistry database the number of subjects with stress hyperglycaemia (RSG > 11.1 mmol/l) was determined, and their corresponding fructosamine values were recorded. The number of subjects who developed DM over the following 5 years was determined.
RESULTS: Three hundred and seventeen patients had stress hyperglycaemia, and follow-up data were available on 224 patients. Of these patients, 63 (28%) had developed DM over the 5 years follow-up period. RSG and fructosamine levels at baseline of patients subsequently developing DM were (mean +/- sd (range)) 16.7 +/- 7.0 (11.2-55.0) mmol/l and 3.3 +/- 0.6 (1.3-4.5) mmol/l, respectively. The patients who did not develop DM had a similar baseline RSG, 15.9 +/- 3.3 (11.2-30.6) mmol/l; P = 0.170, but lower baseline fructosamine, 2.4 +/- 0.4 (1.6-3.8) mmol/l; P < 0.001. Receiver-operating characteristics showed that a serum fructosamine > or = 2.8 mmol/l was a useful marker of the future risk of DM (75% sensitivity, 74% specificity, 53% positive and 88% negative predictive power).
CONCLUSIONS: The level of RSG in stress hyperglycaemia does not predict the future development of DM. Raised serum fructosamine is a more useful marker of future DM risk than RSG alone. Further prospective studies are needed.