Collins, Patricia O'Connor,. Michael 1. Cullen and John Feely. Departments of Pharmacology and Therapeutics and Endocrinology, Trinity College Medical ...
ORIGINAL R SEARCH ARTICLE
C1in. Pharmacokinel. 24 (2): 183·186, 1993 0312·5963/93/ 0002·0183/$02.00/0 © Adis International Limited. All rights reserved. CPKt 270
Plasma Protein Binding of Lidocaine and Warfarin in Insulin-Dependent and N on-Insulin-Dependent Diabetes Mellitus Sharon a 'Byrne, Michael G. Barry, William CJ. Collins, Patricia O'Connor, Michael 1. Cullen and John Feely Departments of Pharmacology and Therapeutics and Endocrinology, Trinity College Medical School, St James's Hospital, Dublin, Ireland
Summary
We examined the plasma protein binding of an acidic drug (warfarin bound to albumin) and a basic drug [lidocaine (lignocaine) bound to ai-acid glycoprotein] in 15 patients with insulindependent diabetes mellitus (IDDM) and 15 matched controls. We also examined protein binding of warfarin and lidocaine in 30 patients with non-insulin-dependent diabetes (NIDDM) and 25 controls. Compared with control, the binding of both warfarin (98.81 ± 0.02 vs 98.57 ± 0.03%, mean ± SEM) and of lidocaine (69 ± 2 vs 58 ± 2%) was significantly reduced in IDDM. This group had lower concentrations of both albumin and ai-acid glycoprotein (AAG), achieving statistical significance vs control for albumin only. In the patients with NIDDM, who had a similar level of glycosylated haemoglobin, while there was no significant difference in the binding of lidocaine there was a significant increase in warfarin binding compared with the control population (99.01 ± 0.03 vs 98.82 ± 0.04%), This study suggests that binding of both acidic and basic drugs is altered in both IDDM and NIDDM.
The extent of plasma protein binding may significantly affect both pharmacokinetics and pharmacodynamics of drugs. In addition to genetic, physiological and environmental factors most of the variation seen in the population is explicable largely on the basis of disease, particularly renal and hepatic, There has been recent interest in the effect of endocrine disorders on drug handling (O'Connor & Feely 1987). An earlier study by RuizCabello and Erill (1984) reported an increased free concentration of sulfisoxazole and diazepam in patients with diabetes, However, no details were given as to whether the patients had insulin-
dependent (100M) or non-insulin-dependent (NIOOM) diabetes, Furthermore, whether this observation extended to acidic drugs with a low therapeutic ratio such as warfarin, or to basic drugs such as lidocaine (lignocaine), which are bound to cq-acid glycoprotein (AAG), was not known. In addition, the site of binding of warfarin to albumin (site 1) is distinct from that for diazepam (site 2) and displacement interactions are site-specific, Our objective was therefore to compare the binding of an acidic drug (warfarin, bound to albumin) and a basic drug (lidocaine, bound to AAG) in a group of patients with 100M and NIOOM with matched controls,
184
Clin. Pharmacokinet. 24 (2) 1993
Materials and Methods Patients 15 well controlled patients with IDDM (mean 41 years, range 18 to 63 years, 13 male) were age and sex matched to 15 healthy controls. Initially 15 patients with NIDDM (mean 59 years, range 32 to 72 years, 9 female, 8 on oral hypoglycaemic drugs) were compared with 10 control patients; subsequently, a further 15 patients with NIDDM (mean 61 years, range 32 to 78 years, II female, 5 on oral hypoglycaemics) were age and sex matched to 15 healthy controls. Blood samples were drawn by venepuncture into a glass system and centrifuged at 490g, plasma was separated and stored at -20°e. No patient received any medication or suffered from a concurrent illness known to influence drug binding.
monofluor (National Diagnostics, New Jersey). Radioactivity was measured using a liquid scintillation counter. The percentage of unbound or free drug was expressed as a ratio of disintegrations per minute (dpm) in buffer and plasma. We also examined the effect of in vitro addition of insulin (Human Monotard, Novo) to produce 10, 100 and 200 mUlL concentrations in control plasma samples (n = 8). Protein binding was determined as previously. Albumin concentration was measured using an automated analysis (SMAC, Technicon) and AAG concentration was assayed in duplicate using radial immunodiffusion (lCL Scientific). Statistical analysis was performed using Student's t-test, Wilcoxon rank sum test and correlation by least square regression. Results are expressed as the mean ± standard error of the mean (SEM).
Drug Binding
Results Plasma protein binding was determined by equilibrium dialysis using semi-macrocells (Dianorm Diachema AG, Zurich, Switzerland) and semi-permeable membranes (Medical International Ltd, UK) with a molecular weight cut-ofT of 10 OOOD. 14C-radiolabelled drug solutions of warfarin (Radiochemical, Amersham) and lidocaine (New England Nuclear), both with radiochemical purity of 98.5%, were added, in duplicate, to plasma to achieve concentrations of 2.0 mg/L and 1.0 mg/L, respectively, and were dialysed against Na2HP04/KH2P04 buffer solution at pH 7.45 for 4h at 3Te. After equilibration, 0.5ml aliquots of both sample and buffer were added to 4.5ml of
Concentrations of binding proteins and glycosylated haemoglobin are given in table I. Glycosylated haemoglobin (HbAI) was as expected significantly elevated in the diabetic population compared with nondiabetic controls (table I). Albumin concentration was significantly lower in IDDM as opposed to control patients. The extent of reduction in AAG in the patients with IDDM did not, however, achieve statistical significance (p = 0.15). There was no difference in either AAG or albumin compared with controls in the patients with NIDDM (table I). The results of plasma protein binding studies
Table I. Mean (± SEM) plasma protein concentration and glycosylated haemoglobin (HbA1) in patients with insulin-dependent (100M) and non-insulin-dependent (NIOOM) diabetes mellitus and control healthy volunteers Parameter
Control (n
AAG (mg/dl) Albumin (gIL) HbA1 (%)
= 15)
77 ± 7.2 46.8 ± 0.6 6.8 ± 0.2
Abbreviations and symbols; AAG from control.
100M (n
= 15)
63 ± 4 44.8 ± of' 8.6 ± Of'
= £>1-acid glycoprotein; • = p < 0.001
Control (n
= 25)
96.4 ± 3.5 43.8 ± 0.9 6.6 ± 0.2 different from 100M; ..
NIOOM (n
= 30)
100.9 ± 3.6' 44.2 ± 0.5 9.3 ± 0.5'"
= P < 0.01; ... = p < 0.001
different
185
Protein Binding in Diabetes
Table II. Protein binding (mean % ± SEM) of lidocaine and warfarin in patients with insulin-dependent (100M) and non-insulindependent (NIOOM) diabetes mellitus
Drug
Control (n = 15)
100M (n = 15)
Control (n = 25)
NIOOM (n = 30)
Lidocaine Warfarin
69 ± 2 98.81 ± 0.02
58 ± 2' 98.57 ± 0.03'
66.6 ± 1 98.82 ± 0.04
67 ± 1.3 99.01 ± 0.03"
Symbols: •
= p < 0.05; •• = P < 0.001
from control subjects.
are summarised in table II. The extent of both warfarin and lidocaine binding was significantly reduced in IODM (table II). Lidocaine binding remained unchanged in NIDDM; however, the results in the initial 15 patients with NIDDM suggested a trend towards increased binding of warfarin (99.06 ± 0.03% vs 98.93 ± 0.05%) and we thus studied a further 15 NIDDM with age and sex matched controls. In this group alone (98.96 ± 0.05% vs 98.75 ± 0.05%, p < 0.01) and in the group as a whole (n = 30) there was a significant increase in warfarin binding in the patients with NIDDM (table II). There was no difference in binding between patients with NIDDM treated with oral hypoglycaemics and those managed on diet alone. The addition of insulin to control samples did not alter warfarin binding (table III). There was no relationship between HbAI concentration and the binding of either warfarin or lidocaine in either group of patients with diabetes.
Discussion The results of this study in relation to binding of acidic drugs in IDDM are in agreement with the earlier study of Ruiz-Cabello and Erill (1984) and suggest that the differences are seen both with drugs bound to site I [warfarin and sulfafurazole (sulfisoxazole)] and site 2 (diazepam) on albumin. It would appear, however, that the mechanisms involved in altered binding of drugs to albumin in diabetes mellitus differ depending on location of the binding site. Various ways by which interference in drug
binding may occur have been demonstrated. Posttranslational modification of albumin due to glycosylation causes interference with binding to site I without affecting site 2 (Ruiz-Cabello & Erill 1984) and will vary according to the level of diabetic control. There is in vitro information from these investigators that increased glycosylation of albumin increased the free fraction of sulfafurazole (site I); there was no relationship with glycosylation of haemoglobin. Similarly, in our study there was no correlation between HbAI levels and protein binding of warfarin (site I). Similar changes can occur in uraemia (Erill & Calvo 1980) and may be induced experimentally after carbamylation of albumin with cyanate which causes a chemical modification of lysine groups; glucose appears to produce its effect at the same site (Ruiz-Cabello & Erill 1984). Salicylate binding is reduced in IODM and Mereish et al. (1982) showed that the binding of salicylate decreased with glycosylation of albumin. Displacement of drug binding by free fatty acids (FFA) brings about a decrease in binding of valproic acid (site I) in IODM (Gatti et al. 1987) and diazepam (site 2) [Ruiz-Cabello & Erill 1984] - in the latter the type of diabetes is not specified. The generality of these changes in relation to drug binding to albumin in IODM is supported by the finding of reduced binding of phenytoin (site 1) [Kemp et al. 1987]. Contrary to other studies, albumin concentration was found to be significantly reduced in the patients with IDDM in our study and may act as a small contributory factor to the reduced binding of warfarin. It is unlikely that insulin has a direct role as the addition of increasing concentrations of insulin did not produce any ef-
Clin. Pharmacokinet. 24 (2) 1993
186
Table III. The effect of insulin 10, 100 and 200 mUlL on protein binding of warfarin (mean % ± SEM) in control healthy volunteers, n = 8
Warfarin binding
Baseline
10
100
200
98.13 ± 0.07
98.21 ± 0.04
98.21 ± 0.06
98.22 ± 0.05
fect. In patients with uraemia competitive inhibition by endogenous ligands phenylacetic, hippuric, vanillic and particularly indoxyl acids contribute to the decrease in drug binding (Oengler et al. 1983; Gulyassy et al. 1983; Zini et al. ) 990). Allosteric binding of FFA to albumin on the other hand can favour binding to site 1 with warfarin and sulfafurazole (Ruiz-Cabello & ErillI984). This would be in keeping with the increased binding of warfarin observed in the patients with NIOOM in this study and would imply an increased concentration of FFA, which is well recognised in NIOOM, and may playa more important role in NIOOM than in 100M. With respect to binding of drugs to albumin there may be opposing mechanisms at play in diabetes mellitus; glycosylation of albumin (site 1) and competitive inhibition by FFA (site 2) in 100M vs allosteric binding of FFA (site 1) in NIODM. The location and size of the binding site on albumin and its relationship to other defined binding sites may also be important. Our results with the basic drug lidocaine suggest reduced binding in patients with 100M but not those with NIOOM. This appears to be due in part to reduced concentrations of the main binding protein AAG although the role of glycosylation has not been explored. AAG concentrations were significantly higher in NIOOM in comparison with 100M, as AAG increases with age (Kremer et al. 1988) this may reflect an age rather than disease related phenomenon. Our study and others suggest that altered drug binding is a feature of diabetes mellitus and is characterised by a reduced binding in 100M of particularly acidic drugs, including warfarin, phenytoin and valproic acid, which have a narrow therapeutic range. In contrast we found a small but significant increase in warfarin binding in a group
of patients with NIOOM when compared with matched controls. The extent of these changes and the range of drugs involved suggest that the clinical impact of altered protein binding on drug handling and response in 100M be studied. A recent review by Gwilt et al. (1991) has outlined the need for carrying out pharmacodynamic studies on a wider range of drugs in diabetes mellitus.
References Dengler TJ. Robertz-Vaupel GM, Dengler HJ. Endogenous ligands and structural changes inhibit binding to human serum albumin in chronic renal failure. European Journal of Clinical Pharmacology 36 (Supp!. I): A65, 1989 Erill S, Calvo R. Plasma protein carbamylation and decreased acidic drug protein binding in uraemia. Clinical Pharmacology and Therapeutics 27: 612-618, 1980 Gatti G, Crema F, Attardo-Partinello G, Fratino P, Aguzzi P, et a!. Serum protein binding of phenytoin and valproic acid in insulin-dependent diabetes mellitus. Therapeutic Drug Monitoring 9: 389-391, 1987 Gulyassy PF, Bottini AT, Jarrard EA, Stanfel LA. Isolation of inhibitors of ligand: albumin-binding from uraemic body fluids and normal urine. Kidney International 16 (Supp!.): S238-S242, 1983 Gwilt PR, Nahhas RR, Tracewell WG. The effects of diabetes mellitus on pharmacokinetics and pharmacodynamics in humans. Clinical Pharmacokinetics 20: 477-490, 1991 Kemp SF, Kearns GL, Turley CPo Altered phenytoin binding in children with epilepsy. Clinical Pharmacology and Therapeutics 41: 170, 1987 Kremer JMH, Wilting J, Janssen LHM. Drug binding to human alpha-I-acid glycoprotein in health and disease. Pharmacological Reviews 40: 1-47, 1988 Mereish KA, Rosenberg H, Cobby I. Glucosylated albumin and its influence on salicylate binding. Journal of Pharmaceutical Science 71: 235-238, 1982 O'Connor P, Feely J. Clinical pharmacokinetics and endocrine disorders: therapeutic implications. Clinical Pharmacokinetics 13: 345-364, 1987 Ruiz-Cabello F, Erill S. Abnormal serum protein binding of acidic drugs in diabetes mellitus. Clinical Pharmacology and Therapeutics 36: 691-695, 1984 Zini R, Riant P, Barre J, Tillement JP. Disease-induced variations in plasma protein levels: implications for drug dosage regimens (Part I). Clinical Pharmacokinetics 19: 147-159, 1990 Correspondence and reprints: Professor John Feely, Department of Pharmacology and Therapeutics, Trinity College Dublin, St James's Hospital, Dublin 8, Ireland.
