Diabetologia (1983) 24:412-417 Diabetologia Springer-Verlag 1983 Glibenclamide-Associated Hypoglycaemia: A Report on 57 Cases K; Asplund 1, B.-E. Wiholm 2, 3 and F. Lithner I 1Department of Medicine, Ume~ University Hospital, Umefi, 2Department of Drugs, National Board of Health and Welfare, Uppsala and 3Department of Clinical Pharmacology, Karolinska Institute, Huddinge Hospital, Huddinge, Sweden Summary. In the largest series of patients with glibenclamide- patients responded immediately to treatment, 24 had protract- associated hypoglycaemia reported so far, 51 cases reported to ed hypoglycaemia of 12-72 h duration and 10 died. Fatal out- the Swedish Adverse Drug Reactions Advisory Committee come was observed even with small doses of glibenclamide and six additional cases are reviewed and related to sales and (2.5-5 mg/day). Previous strokes and cardiac disorders were prescription data of glibenclamide. Median age of the patients isolated as two independent determinants of a serious course with hypoglycaemia was 75 years and 21% were 85 years or of the hypoglycaemia. Other contributing factors included im- above. For comparison, the median age of a random sample paired renal function, low food intake, diarrhoea, alcohol in- (1 in 288 of all patients prescribed glibenclamide) was 70 years take and interaction with other drugs. Thus, it is not uncom- and only 5% were 85 years or older. In eight out of 40 cases mon for glibenclamide, like the first-generation sulphonyl- where duration of glibenclamide treatment was recorded, the ureas, to cause serious, protracted and even fatal hypoglycae- hypoglycaemic event occurred during the first month of treat- mic events. ment. The median daily dose of glibenclamide prescribed was 10 mg both in the hypoglycaemic cases and in the prescription Key words: Sulphonylurea, glibenclamide, drug-induced hy- Sample. Coma or disturbed consciousness was the most com- poglycaemia, hypoglycaemic coma, drug-related death, drug mon clinical presentation in this series and the minimum interactions blood glucose value was 1.3 retool/1 (median). Twenty-two The value of sulphonylureas in the long-term treatment In Sweden, glibenclamide was introduced in 1971 of Type 2 (non-insulin-dependent) diabetes has been and it is now the most frequently prescribed oral anti- questioned [1]. The potential benefits of an oral anti- diabetic drug [5]. Up to mid-1981, a total of 51 reports diabetic drug must be carefully weighed against the risk on hypoglycaemia in glibenclamide-treated diabetic of hazardous adverse reactions, hypoglycaemia in par- patients, including 10cases with fatal outcome, had ticular. In 1972, Seltzer reviewed more than 200 cases of been submitted to the Swedish Adverse Drug Reactions severe hypoglycaemia induced by sulphonylureas [2]. Advisory Committee (SADRAC). Of these, the majority were attributed to chlorpropa- This report outlines some clinical characteristics of mide. During the last decade, newer, more potent sul- these 51 patients and also of six additional cases, not phonylurea preparations have been introduced. There reported to SADRAC. Sales and prescription data on have been only few reports on severe hypoglycaemic glibenclamide are also analysed as background infor- events caused by these agents and, therefore, little is mation. known about the salient features of hypoglycaemia in- duced by the so-called second generation of sulphonyl- ureas. The biological half-life of single doses of gliben- Patients and Methods clamide is relatively short [3, 4]. Therefore, when the drug was first marketed, it was inferred that hypogly- In Sweden, adverse drug reactions have been reported to SADRAC of the National Board of Health and Welfare since 1965. From 1975, caemia induced by glibenclamide might be less com- reporting is compulsory for suspected reactions that are fatal or other- mon and easier to control than that induced by chlor- wise serious, new, unexpected or remarkable. In the present study, we propamide. examined all 51 reports on glibenclamide-associated hypoglycaemia K. Asplund et al.: Glibenclamide-Associated Hypoglycaemia 413 method and concomitant clinical symptoms in 54 events occurring in 50 patients. In the remaining seven cases, all had symptoms suggestive 5~ 16 of hypoglycaemia and responded promptly to the administration of intravenous glucose. Blood glucose was determined quantitatively in five of them, showing moderately low levels (2.8-3.0 mmol/l) and uJ 12 semi-quantitatively by Dextrostix in one instance. Five patients were ~ 10 concomitantly treated with biguanides. Total annual sales of glibenclamide in Sweden [5] were converted to and expressed as numbers of so-called defined daily doses [6]. For glibenclamide one defined daily dose is set to 10 mg. Since 1974, age and sex of the patient and name and amount of the drug have been recorded from a random 1-in-288 sample of all pre- scriptions delivered in Sweden [71. Since 1976 the prescribed daily I I / I I I I I I dose has also been recorded. Data from this prescription survey have 1972 1974 1976 1978 1980 been extracted and compared with age, sex and daily dose of the YEAR patients with hypoglycaemia. Between 1975 and 1980, a total of Fig.l. Annual numbers of glibenclamide-associated hypoglycaemic 3,557 prescriptions of glibenclamide was recorded and the annual episodes reported and sales of glibenclamide in Sweden [5]. The sample size averaged 593 with a standard deviation of 71. defined daily dose of glibenclamide is 10 mg [6]. Only first 6 months of 1981 Results Glibenclamide ~'//2A ~////A Reporting to the Swedish Adverse Drug Reactions 40 N Advisory Committee Figure 1 gives the sales of glibenclamide and the an- 30 P'///A nual number of hypoglycaemic episodes reported to SADRAC. N There is a great variation in the reporting from year 20 to year and a marked geographical clustering in the reporting which does not correspond to differences in lO sales (data not shown). Patients 0-29 30-39 40-49 50-59 60-69 70-79 80-89 90-99 AGE (years) There was a female preponderance both among the pat- Fig, 2. Age distribution of 57 patients with glibenclamide-associated ients with hypoglycaemia (61%) and in the prescription hypoglycaemia [~: distribution of patients from the prescription sur- material (55%). Median age was 70years among the vey 1975-1980 [] (annual sample size, mean = 593, SD = 71) 57 hypoglycaemic patients and 75years among the 3,557 patients from the prescription survey (Fig. 2), The proportions of patients aged 85 years or more were 21% Table 1. Time since diagnosis of diabetes mellitus and duration of glibenclamide treatment in 57 patients with glibenclamide-associated and 4.8% respectively (p< 0.01). Only one of the hypo- hypoglycaemia glycaemic patients was < 60 years of age. As shown in Table 1, diabetes was diagnosed during Duration Number of patients the 12 months preceding the index event in more than Diabetes Glibenclamide one-third of the patients and during the last month in treatment seven. However, one-third had diabetes of more than < 1 month 7 8 5 years' duration. 1-3 months 4 8 4-12 months 9 14 1-5 years 13 10 Gtibenclamide Treatment > 5 years 19 0 Data not available 5 17 Data on the duration of glibenclamide treatment were available in 40patients. Of these, 30individuals had Total 57 57 been on gtibenclamide for < 1 year. In eight cases hy- poglycaemia developed within the first month and in four cases within the first week of glibenclamide treat- that were classified as 'probable' or 'possible'. To clarify details in the ment. reports, complete medical records were obtained. During the course The prescribed daily doses of glibenclamide in the of our work, we became aware of six additional cases of glibencla- patients with hypoglycaemia and in the prescription mide-associated hypoglycaemia not reported to SADRAC. These cases are also included in this report. survey are displayed in Figure 3. The distributions of The diagnosis of hypoglycaemia (blood glucose < 2.5 mmol/1) daily doses were similar and median daily dose was was based on blood glucose determinations by a glucose oxidase 10 mg in both groups. 414 K. Asplund et al.: Glibenclamide-Associated Hypoglycaemia PRESCRIPTION SURVEY 1979 GLIBENCLAMIDE PATIENTS WITH GLIBENCLAMIDE - n=591 INDUCED HYPOGLYCAEMIA (mg/day) / n=57 E 30 27,5 25 22,5 [ 20 ~ 17,5 --'-1 Fig. 3. Distribution of prescribed daily doses [ 15 I .~ I of glibenclamide as related to outcome of the r 12,5 ~, hypoglycaemic event (fight-hand part of the figure). [] = short duration ( < 12 h), [] = pro- tracted bypoglycaemia (12-72 h), = death, ] 7,5 I = [] = surviving, but duration unknown. For ,5 i comparison, prescribed daily doses of gfiben- t i 2,5 I = clamide in the prescription survey 1979 are shown on the left. n denotes number of pat- E 1,2s I ients I 215 20 15 10 5 0 5 10 15 20 25 % % ponded immediately to therapy and did not relapse; (2) 3.0 24 patients had protracted hypoglycaemia of at least 12 h duration, two of these remained hypoglycaemic for 3 days despite continuous IV glucose therapy; (3) ten patients had deep coma leading to death. The duration of hypoglycaemia was not stated in one non-fatal case. 000 The clinical notes were seldom sufficient to deter- 0 2.0 mine whether the patients were left with sequelae to the hypoglycaemic event. The youngest patient (aged E 00 E 32years), however, was left with permanent brain 000 dysfunction after a protracted hypoglycaemic coma. 0 00 _J 0 Among the ten cases with fatal outcome, one died E~ within a few hours of the onset of hypoglycaemia. Eight 0 0 had hypoglycaemia lasting for at least 12 h despite cor- 1,0 0000 rect diagnosis and treatment with parenteral glucose. When the hypoglycaemia was alleviated they all re- mained comatose until death which occurred within 1-20 days.