between plasma metformin and phenformin concentrations and metabolic effects. Drug levels were measured in 37 type II diabetic patients by HPLC.
Plasma biguanide levels are correlated with metabolic effects in diabetic patients Metabolic abnormalities occur in biguanide-treated diabetic patients. We investigated the relationship between plasma metformin and phenformin concentrations and metabolic effects. Drug levels were measured in 37 type II diabetic patients by HPLC. The method was sensitive, specific, and linear over a wide range of drug concentrations. Metformin and phenformin values ranged from 236 to 718 ng/ml and from 28 to 114 ng/ml, respectively. The plasma metformin level was correlated with triglycerides (r = 0.55; P < 0.05) but not with drug dosage, plasma glucose, HbA creatinine, creatinine clearance, lactate, pyruvate, lipid, and clinical parameters. Plasma phenformin concentrations correlated with lactate (r = 0.49; P < 0.05) and HbA, (r = 0.50; P < 0.05) but not with drug dosage, parameters of diabetes control, creatinine, creatinine clearance, pyruvate, and clinical parameters. The clinical usefulness of this HPLC method, the evidence that the increase of lactate is related to the circulating phenformin levels, and the demonstration that the metformin effect on triglyceride metabolism is correlated to plasma drug levels are the positive findings of this work. (CLIN PHARMACOL THER 1987;41:450-4.)
P. Marchetti, M.D., L. Benzi, M.D., P. Cecchetti, R. Giannarelli, M.D., C. Boni, D. Ciociaro, M.D., A. M. Ciccarone, M.D., G. Di Cianni, M.D., A. Zappella, M.D., R. Navalesi, M.D. Pisa, Italy
Metabolic abnormalities occur during treatment with the antidiabetic biguanides phenformin and metformin.' The way by which the two drugs exert their hypoglycemic action may explain the changes in circulating metabolites. If enhanced glycolysis in extrahepatic tissues' and inhibition of hepatic gluconeogenesis' are important mechanisms to reduce plasma glucose levels in type II diabetic patients, changes in circulating gluconeogenic substrates are to be expected. Moreover, an action of biguanides, in particular metformin, on lipid metabolism has been reported.' It remains to be elucidated whether and in which way the plasma concentration of phenformin and metformin are correlated with the metabolic changes produced by the drug, a major difficulty being the lack of simple, quick, sensitive, and specific methods to assay both drugs in human plasma. Recently we described a new technique to accurately measure plasma concentrations of phenformin and met-
From the Cattedra di Malattie del Ricambio, Istituto di Clinica Medica II and Istituto di Fisiologia Clinica del C.N.R. , Universita di Pisa. Supported by the Italian National Research Council, Special Project "Diabetes," grant No. 82.02254.56, and by Ministero Pubblica Istruzione. Received for publication June 10, 1986; accepted Sept. 10, 1986. Reprint requests: Dr. P. Marchetti, Cattedra di Malattie del Ricambio, Istituto di Clinica Medica II, Via Roma, 67, 56100 Pisa, Italia.
450
formin by HPLC.5 In this article we investigate the relations among circulating biguanide levels and plasma levels of endogenous substances in a group of type II diabetic patients, and evidence is reported that the plasma levels of some compounds are related with the concentration of biguanides in blood, especially when phenformin is considered.
MATERIAL AND METHODS Patients. The study was carried out in 37 type II diabetic patients (16 men and 21 women) from an outpatient clinic. Their C-peptide levels were 1.6 ± 0.6 ng/ml. The therapy with biguanides had started 1 year before the study. Thirteen diabetic patients were taking metformin and 24 were taking phenformin in combination with sulphonylureas. It has been shown that phenformin produces similar metabolic abnormalities whether used alone or in combination with a sulphonylurea.'No patient had clinical or biochemical evidence of hepatic or renal disease. Of the patients taking metformin, one was taking clonidine and phenobarbital, and another one glyceryl trinitrate; three diabetic patients taking phenformin were taking other drugs: verapamil and digoxin (one patient), digoxin and diuretics (one patient), and clonidine (one patient). After informed consent had been obtained, a blood sample was taken between 8:00 and 9:00 AM after an overnight fast.
VOLUME 41 NUMBER 4
Biguanide levels and metabolic effects
451
Table I. Clinical details of diabetic patients subdivided according to the type of antidiabetic therapy Length of diabetes (yr)
MIF
Age (yr)
(kg/m2)
13
5/8 11/13
67 ± 9 63 ± 11
28 ± 3
24
No.
Patients taking metformin Patients taking combined phenformin and sulphonylurea
BMI
27 ±
3
16 16
± 11 ± 9
Length of therapy (yr)
Biguanide dose (mg/day)
Biguanide dose (mg/kg/day)
±
1240 ± 560 66 ± 20
15.4 ± 7.0 0.9 ± 0.3
5 5
±
3 3
BMI, body mass index.
Table II. Biochemical parameters of diabetic patients subdivided according to the type of antidiabetic therapy
Patients taking metformin Patients taking combined phenformin and sulphonylurea *P
< 0.05
Creatinine (mg/dl)
Creatinine clearance (ml/mm)
0.9 ± 0.1 1.0 ± 0.2
107 106
± ±
15 18
Fasting plasma glucose (mg/dl) 120 ± 30 124 ± 34
Total HbA, (%)
8.3 ± 0.8* ± 1.4
cholesterol (mg/dl) 228 ± 36
Triglycerides (mg/dl) 148
± 62
HDLcholesterol (mg/dl) 53 ± 15
10.1
vs. patients taking combined phenformin and sulphonylurea.
All patients had rested for at least 15 minutes and had taken the last biguanide tablet the evening before. In 16 normal subjects (age 63 ± 8 years; body mass index 27 ± 4 kg/m2), blood lactate and pyruvate levels (see below) were tested in the fasting state. Analytic procedures. The HPLC determination of metformin and phenformin in human plasma has been described.' Briefly, isocratic separations were performed at room temperature on a prepacked 300 by 3.9 mm inside diameter C18 Bonde-Pek (10 p.m) reversedphase liquid chromatographic column (Waters Associates, Milford, Mass.). Plasma samples were prepared as follows: in 0.3 aliquots acetonitril was added to obtain deproteinization (final ratio 1/5, v/v), and centrifugation was performed at 2000 x g for 15 minutes. After centrifugation the solvent was evaporated from the supernatant. Samples containing metformin (range 75 to 10,000 ng/ml) and phenformin (25 to 5,000 ng/ml) were used as reference to assess the identity of the peaks obtained chromatographing human plasma, the linearity of the method, and its precision, sensitivity, and specificity.' The within-assay coefficient of variation was 4.5% for metformin and 3% for phenformin at low and high concentration; the betweenassay coefficient of variation was 10% and 4.3% for metformin and phenformin, respectively. Blood lactate and pyruvate were determined as de-
scribed.' The within-assay and between-assay coefficient of variation was 4.6% and 5.5%, respectively. Plasma glucose was measured by the glucose-oxidase method on a Beckman glucose analyzer (Beckman Instruments, Galway, Ireland); HbA, concentrations were dosed after the separation of the labile HbA, fraction by column chromatography (Bio-Rad Laboratories, Richmond, Va.). The intra- and between-assay coefficients of variation for the two methods were lower than 3%. Total serum cholesterol, triglycerides, and HDLcholesterol were tested in 13 patients taking metformin by fully enzymatic procedures (Boheringer Biochemia, Mannheim, West Germany). The coefficients of variation for lipid determinations were
