Prevalence of nocturnal hypoglycemia in first trimester of pregnancy in patients with insulin treated diabetes mellitus

Background Excellent metabolic control before conception and during diabeti c pregnancies is the aim in order to avoid malformations and perinatal morb idity. Since an inverse correlation between median blood glucose concentrat ion (BG) and hypoglycemia as well as a high prevalence of nocturnal hypogly cemia have been described, we investigated the frequency of nocturnal hypog lycemia and the predictive value of bedtime blood glucose concentration for development of this condition in insulin treated diabetic patients. Methods. During hospitalization, with no other changes in the patients' nor mal schedules, hourly blood samples were drawn from an iv-cannula from 22.0 0 h to 07.00 h for one night. BG (venous whole blood) and hemoglobin Al, we re determined the following day. Results. Fifty-three patients participated; subsequently ten were excluded due to discontinuation of blood sampling during the night, caused by either discomfort or cannula problems. Of the remaining 43 patients, 16 (37%) had at least one blood glucose <3.0 mmol/l. The duration of hypoglycemia was 2 .4 (1-7) h with the highest prevalence at 05 h. Only one patient felt hypog lycemic during the night. Hemoglobin Al, was similar in patients with (7.1 <plus/minus> 1.2%, mean +/- s.d.) and without (6.8 +/- 0.8%) nocturnal hypo glycemia. Women with nocturnal hypoglycemia had significantly lower BG befo re bedtime compared to patients without hypoglycemia, 6.4 +/- 3.6 mmol/l vs . 7.9 +/- 2.4 mmol/l, p < 0.05. The best predictive value for nocturnal hyp oglycemia was a BG below 6.0 mmol/l at 23.00, which resulted in a risk of n octurnal hypoglycemia of 71%. Conversely, if the BG was <greater than or eq ual to>6.0 mmol/l, the chance of avoiding nocturnal hypoglycemia was 83%. Conclusion. Nocturnal hypoglycemia was seen with a prevalence of 37% during a night in the first trimester of pregnancy in insulin treated patients. O nly one patient registered the hypoglycemia. Nocturnal hypoglycemia could b e predicted in the majority of patients by measurements of BG before bedtim e.