1. Introduction

Diabetes mellitus is one of the most significant public health problems. The number of patients with diabetes is increasing due to population growth and a growing prevalence of physical inactivity leading to obesity.[1] In 1985, approximately 30 million people were reported to have diabetes worldwide.[2] A decade later, this estimate had reached 140 million.[3] In 2003, the global prevalence of diabetes was estimated to be 250 million.[3] Including undiagnosed cases of diabetes, this number is expected to reach 438 million by 2030 (approximately 7.5% of the adult population).[1,3,4] There seems to be no region of the world that will not be affected by the disease. While some of this increase will be observed in Australia, Europe and North America, the majority will be seen in countries undergoing westernization (the African continent, South America, China, India and the Middle East).[1] The majority of diabetic patients are between 45 and 65 years of age in developing countries, while they are ≥65 years of age in developed countries.[3,5] Although type 2 diabetes was previously considered to be a chronic disease of middle and late adulthood, it is increasingly becoming prevalent in adolescents.[6] The number of deaths attributable to diabetes has been estimated at around 3 million per year, which is approximately 5% of global mortality.[7]

Two types of diabetes predominate in the population. Type 1 diabetes is characterized by an inability of the body to produce insulin, whereas type 2 diabetes reflects deficient insulin secretion and/ or insulin resistance (approximately 90% of all cases of diabetes).[8] Some patients are difficult to classify as having type 1 or 2 diabetes, thus other specific types of the disease have been described (e.g. gestational diabetes, malnutrition-related diabetes, etc.). Gestational diabetes represents a separate entity as do other specific types of diabetes associated with genetic defects, diseases of the exocrine pancreas (i.e. pancreatitis, cancer or cystic fibrosis), drug- or chemical-induced, infection-related or immune-mediated diabetes. Risk factors for diabetes are anthropometric factors (e.g. obesity), metabolic factors (e.g. parameters of glucose metabolism), lifestyle factors (e.g. physical activity), metabolic syndrome, age, sex and ethnicity.[9] Most people with type 2 diabetes are obese and the degree of duration of obesity correlates with the risk of the diabetes. Obesity and inactivity are two of the most important lifestyle risk factors for type 2 diabetes.[10] It has been observed that changes in lifestyle and a moderate level of physical activity can prevent type 2 diabetes.[11,12]

One of the major concerns with the diabetes epidemic is the increase in both morbidity and mortality related to complications of the disease. Diabetes is associated with a progression of microvascular (e.g. nephropathy, neuropathy, retinopathy) and macrovascular (e.g. myocardial infarction, stroke) complications.[13] Thus, the use of drugs by patients with diabetes is expectedly higher than in an age-matched population without the disease.[14,15] Total diabetes costs were estimated at $US174 billion in the US[16] and £3.5 billion in the UK in 2007.[17] On the other hand, Aggarwal[18] reported that US sales of insulin products reached $US8.4 billion.

Although the effect of diabetes on chemically induced animal models of the disease has been studied extensively, minimal data are available on the effect of diabetes on pharmacokinetics and pharmacodynamics in humans. An understanding of diabetes-mediated changes as well as the source of the variability should lead to the improvement of medical management and clinical outcomes in patients with this widespread disease. Since the last review covering the effect of the diabetes on the pharmacokinetics and pharmacodynamics of drugs,[19] a number of related original publications have appeared. In this review, general information related to diabetes-mediated changes in clinical pharmacokinetics and pharmacodynamics is discussed. We conducted a systemic review of previously published literature (end date January 2012) by searching the PubMed and relevant bibliographies. Only papers from the English and German literature were included in the review.

2. Effect of Diabetes Mellitus on Pharmacokinetics

Not only is drug use more widespread, but the pharmacokinetics of these drugs may also be altered by the disease itself.[19–21] Diabetes affects protein, lipids and carbohydrate metabolism, and the systems that regulate these biochemical pathways are also in many cases involved in drug biotransformation. Thus, it is not surprising that diabetes affects the pharmacokinetics of drugs. Diabetes may influence the pharmacokinetics of numerous drugs by affecting (i) absorption, due to changes in subcutaneous/muscle blood flow and delayed gastric emptying; (ii) distribution, due to non-enzymatic glycation of albumin; (iii) biotransformation, due to differential regulation of enzymes involved in drug biotransformation and drug transporters; and (iv) excretion, due to nephropathy.[20–22] However, pharmacokinetic studies in diabetic patients are limited compared with pharmacokinetic studies in animals.[23] Moreover, different results were obtained between these studies and, thus, animal data should be extrapolated only very carefully to humans.[23]

2.1 Absorption

Micro- and macrovascular changes are probably the main long-term complication of diabetes. Gastric mucosal blood flow has been reported to be significantly reduced in diabetic patients (2.81 ± 0.15 V) compared with non-diabetic subjects (4.77 ± 0.29 V).[24] Because it has been demonstrated that gastric mucosal blood flow varies in the same direction as gastric acid secretion,[25–27] changed mucosal blood flow might explain gastric pH changes in diabetic patients.[28]

It is long known that longstanding type 1 and 2 diabetes can be associated with a change in the gastric emptying time and intestinal transit time,[29–50] which was first reported by Boas[51] in 1925. The prevalence of delayed gastric emptying in diabetic patients has ranged from 28% to 65%;[52,53] however, the relationship between delayed emptying and symptoms is variable[52,54,55] (it should be pointed out that the number of diabetic patients with emptying delay was underestimated in early studies that employed less sensitive diagnostic methods to quantify gastric emptying[56]). Numerous studies have observed delayed gastric emptying time in poorly controlled diabetic patients[30] and showed a strong correlation with the progression of the disease.[52,57–59] In most cases, the emptying of solid meat components and low-nutrient liquids is delayed to a varying degree in diabetic patients,[39,48,60,61] whereas the emptying of liquids has not been found to be affected by the disease.[39,48,60–62] However, there are discrepancies in the literature; for example, gastric emptying rates in patients with type 2 diabetes have been reported to be decreased,[33,36,45,60,63,64] unchanged[65] or accelerated.[66,67] Some of these discrepancies may be explained by (i) differences in the study design; (ii) uncontrolled hyperglycaemia;[68–70] or (iii) diabetes-related changes in intestinal hormones.[71] Increased knowledge of the pathophysiology of diabetes has contributed to a development of new treatments such as glucagon-like peptide-1 agonists and dipeptidyl peptidase-4 inhibitors. These new classes of antidiabetic drugs can mask the effect of the disease on gastrointestinal tract motility and make it harder to distinguish the effect of the disease and the effect of the drug.[72] It has also been reported that cystic fibrosis reduces level of free incretin hormones,[73] resulting in decreased gastric emptying time.[73,74] Although many studies have reported diabetes-mediated changes in gastric emptying time, it seems that the magnitude of the delay is modest and should not be clinically important.

Transit time in diabetic patients has been reported to be significantly lower than in non-diabetic patients (gastric transit time: 24 [4–108] vs 87 [1–478] min;[75] small bowel transit time: 302.0 ± 62.7 vs 261.2 ± 55.5 min;[75] transit time: 20.4 ± 15.6 vs 34.9 ± 29.6 hours,[76] 24.3 ± 11.9 vs 43.2 ± 22 hours,[77] 9.9 ± 6.1 vs 14.4 ± 8.3 hours[78] and 35.4 ± 4.7 vs 53.8 ± 5.5 hours[79]). Delayed transit time is probably due to vagal denervation[80] and seems to be more frequent in patients with autonomic neuropathy.[45,81,82] Recent studies have demonstrated that the blood glucose concentration itself has a major impact on gastrointestinal tract motility; marked hyperglycaemia appears to affect every region of gastrointestinal tract.[68,76,83] The mechanisms by which hyperglycaemia affects gastric emptying include suppression of antral contraction,[68] increased pyloric contractions,[68] proximal stomach relaxation[68] and induction of gastric electrical dysrhythmias.[84]

Della-Coletta and Eller[85] reported slower absorption (26%) of tolazamide in diabetic patients than in healthy volunteers. In addition, Adithan et al.[86] reported a decrease (26%) in the absorption of orally administered ampicillin in patients with type 2 diabetes. By contrast, O’Connell et al.[87] reported that basic pharmacokinetic parameters of metoclopramide after single- and multiple-dose administration were not affected by diabetes. In addition, Wahlin-Boll et al.[88] reported complete absorption of the oral sulfonylurea glipizide (administered as a solution as well as tablet form) in patients with type 2 diabetes as compared with previously reported studies with healthy volunteers.[89] Prokinetic drugs in general result in dose-related improvement in gastric emptying time; however, variations in the blood glucose concentration may account for variability in results from studies that are related to the effect of these drugs on gastric emptying in diabetic patients.[90–94]

It has been suggested that differences in the absorption rate and bioavailability between patients with diabetes and healthy subjects depend on several factors (table I). Previously published studies have reported that the rate of absorption of subcutaneously administered insulin strongly correlates with subcutaneous blood flow.[99–102] It is well known that abdominal subcutaneous fat tissue blood flow is increased in subjects with type 1 diabetes, probably due to reduced subcutaneous fat tissue (5.0 ± 2.3 vs 11.3 ± 9.1 mL/100 g/min[95]).[103] Nosadini et al.[95] reported faster absorption of insulin in subjects with type 1 diabetes; however, no difference was found in bioavailability of the drug. On the other hand, in insulin resistant and/or obese patients subcutaneous adipose tissue blood flow is reduced compared with healthy normal-weight subjects (by approximately 50%[104]).[105–109]

Table I
figure Tab1

Effect of diabetes mellitus on drug absorptiona b

Full size image

Diabetic patients exhibit compromised muscle blood flow (59.4 ± 12.9 vs 46.7 ± 14.1 mL/100 g/min[97]) at rest[97,110] and also after exercise.[111] This may be a result of blunted endothelium-dependent vasodilation,[111–113] an increased plasma level of endothelin-1,[114] a reduction in capillary density[115,116] or decreased mitochondrial volume[117] and function.[118] These changes may lead to a reduced rate of absorption for drugs with an intramuscular route of administration. This finding was supported by studies when insulin,[96] aminoglycosides[119] and benzylpenicillin[98,120] were administered to diabetic patients.

2.2 Distribution

The volume of distribution of a drug correlates with the degree of obesity, and because obesity is one of the most important factors in the development of insulin resistance and diabetes, the volume of distribution of lipophilic drugs is affected by the disease[121,122] (for more information, see the review by Hanley et al.[123]). On the other hand, obesity does not affect serum albumin and drug binding to albumin; however, data regarding α1-glycoprotein are contradictory.[123] A modest increase in concentration of free diazepam (1.9 vs 1.5 L/kg)[124] nitrazepam (19.7 vs 17.9 L/kg)[125] and oxazepam (5.1 vs 4.0 L/kg)[126] reported in obese individuals might be explained by elevation in serum free fatty acids found in obese individuals. Reduced tissue blood flow as well as alterations in cardiac structure and function has also been reported in obese patients.[107,127]

Diabetic patients have higher level of circulating glucose in the blood, leading to non-enzymatic glycation of several proteins including albumin. Glycated albumin exhibits atherogenic effects in various cells.[128] Non-enzymatic glycation of albumin produces conformational changes in the structure of albumin (affinity of the phenytoin binding site on albumin based on a modification of the lysine group[129]), which can increase the free fraction of acidic drugs in patients with type 1 and 2 diabetes (for more detail, see table II).[21,137–144] Worner et al.[138] reported a 50% decrease in binding of dansylsarcosine to albumin in diabetic patients, whereas the concentration of circulating albumin was the same in diabetic patients.[139,145,146] Glycation of blood and plasma proteins leads to reduction in protein binding capacity.[21,22,147] A linear relationship has been reported between the degree of albumin glycation and the unbound fraction of drug in the serum of diabetic patients.[137,139,140] Thus, for highly albumin bound acidic compounds the reduction in the plasma serum protein binding capacity has been shown in diabetic patients.[19,137] However, studies employing a smaller number of diabetic patients might easily fail to show this linear relationship. There are also inconsistencies in the reported studies; for example, the volume of distribution of antipirine has been increased,[132] unchanged[133–135] or decreased[136] in diabetic compared with non-diabetic patients. In addition, the volume of distribution of paracetamol (acetaminophen)[130] as well as theophylline[86,130] was increased in diabetic patients. These inconsistencies may reflect the degree of control of the diabetes, the duration of the disease and/or the presence of complications of the disease. It has been reported that diabetes can affect drug binding in several ways: (i) changing the amount and concentration of circulating free fatty acids, the level of which is increased in diabetic patients[137,148,149] (e.g. the binding of valproic acid was significantly decreased in diabetic patients;[150] a strong correlation has been previously observed between the free fraction of valproic acid and serum free fatty acid concentration[145]); (ii) increasing blood concentration of substrates possibly inhibits drug binding; and (iii) conformation changes of plasma proteins (as shown in table III).

Table II
figure Tab2

Effect of diabetes mellitus on volume of distribution (values are mean ± SE)

Full size image
Table III
figure Tab3

Effect of diabetes mellitus on drug protein binding

Full size image

2.3 Biotransformation

The effect of obesity on drug biotransformation is very variable. Obesity has been linked to decreased hepatic blood flow, non-alcohol fatty acid disease and the accumulation of fat in the liver tissue.[153,154] The effect of obesity on cytochrome P450 (CYP) appears to be enzyme specific: the enzyme activity of CYP2E1 has been increased in obese individuals.[155] Oral clearance of triazolam (CYP3A subfamily marker) has been decreased in obese compared with non-obese individuals (340 ± 44 vs 531 ± 38 mL/min).[156] On the other hand, obesity did not significantly affect clearance of midazolam (CYP3A subfamily marker) [471 ± 38 vs 530 ± 54 mL/min].[156] The effect of the disease on the activity of other CYP enzymes remains controversial or insufficiently studied.[155,156]

The clearance of drugs undergoing glucuronidation and sulfatation appears to be increased in obese individuals.[156] Abernethy et al.[126] reported increased biotransformation of oxazepam (3.1 times greater) and lorazepam (1.6 times greater) in obese individuals.

In diabetic patients, abnormal hepatic function occurs frequently, especially non-alcohol steatohepatitis, macrovesicular steatosis, liver fibrosis/cirrhosis and focal fatty liver.[157,158] Hence, it is not surprising that diabetes affects drug biotransformation. There have been numerous studies on the effect of chemically induced diabetes on drug biotransformation in both mice and rats.[159–161] Animal models of diabetes fall into two main groups: type 1 (chemically induced diabetes using, for example, alloxan and streptozotocin) and type 2 (spontaneous autosomal recessive mutations, e.g. C57BL/6J ob/ob, C57BL/KsJ db/db, yellow obese mice, Zucker diabetic rats), consistent with the aetiologies of the two main types of the disease in humans. However, data from animal studies (especially from small laboratory animals[23]) should be evaluated with care when trying to predict changes in humans.[162]

Hepatic-mediated biotransformation in diabetic humans is characterized only for a few compounds.[163] In addition, a correlation between diabetes and the activity of enzymes involved in the biotransformation is poorly known. In general, it is apparent that uncontrolled diabetes causes an overall increase in CYP enzymes although phase II enzymes are significantly decreased. Total hepatic CYP content in biopsies of diabetic patients is significantly increased in type 1 and decreased in type 2 diabetic patients.[164] The majority of studies have examined the effect of diabetes on the modulation of CYP2E1; however, varied data exist regarding whether diabetes affects expression and activity of this CYP enzyme.[165–170] Song et al.[169] reported that CYP2E1 protein levels were elevated in the lymphocytes of children and adolescents with type 1 diabetes, and a strong correlation has been observed between the enzyme level and glycosylated haemoglobin. On the other hand, enzyme activity was not changed in a group of seven type 1 and 15 type 2 patients with diabetes using chlorozoxazone as a marker of CYP2E1.[165] In addition, the protein level of CYP2E1 has not been affected by diabetes.[169] Several studies that focused on the effect of diabetes on the clearance of theophylline or caffeine (a marker of CYP1A2 enzyme activity) have suggested that there is no difference in the enzyme activity of non-diabetic and type 1 and 2 diabetic patients.[130,171–174] Sotaniemi et al.[164] found that among type 2 diabetic patients, women biotransformed antipyrin normally whereas men over 40 years of age showed reduced biotransformation. In addition, no change was found in the biotransformation of tolbutamide (a marker of CYP2C9 enzyme activity)[175] and hepatic aryl hydrocarbon hydroxylase (mediated by CYP1A1).[164] Dyer et al.[176] did not observe any change in quinidine (a non-validated marker of CYP3A subfamily activity) between non-diabetic and diabetic patients. However, other studies have described significant downregulation of liver CYP3A4 enzyme activity[177] as well as in the biotransformation of lidocaine (non-validated marker of CYP3A subfamily marker)[178,179] [for more details, see table IV].

Table IV
figure Tab4figure Tab4figure Tab4

Effect of diabetes mellitus on phase I drug biotransformation

Full size image

Minimal data exist on the effect of diabetes on the phase II biotransformation in human.[23] Evidence suggests that oxidative stress is increased in diabetes because of extensive production of reactive oxygen species and an impaired antioxidant defence mechanism.[183] Thus, the majority of studies have studied the effect of diabetes on the activity of enzymes that are part of the antioxidant defence system such as the glutathione-S-transferase (GST) superfamily. Controversy exists regarding the effect of diabetes on phase II biotransformation enzymes.[23] Some studies did not show a difference in erythrocyte GST between patients with type 1 and 2 diabetes and healthy controls.[184,185] On the other hand, McRobie et al.[180] has reported significant downregulation in placental GST activity in type 1 diabetic patients. Dostalek et al.[177] described significant downregulation in gene expression, protein levels and enzyme activity of liver uridine diphosphate glucuronosyltransferase (UGT) 2B7 in liver samples from diabetic donors (for more details, see table V). No changes have been observed for selected enzymes from the UGT 1A subfamily.[177] The effect of diabetes on drug clearance is summarized in table VI.

Table V
figure Tab5

Effect of diabetes mellitus on phase II drug biotransformation

Full size image
Table VI
figure Tab6

Effect of diabetes mellitus on hepatic drug clearance

Full size image

Clinical studies investigating the effect of diabetes on gene expression and protein levels of drug transporters are limited. It has been demonstrated that gene and protein expression of ABCA1 and ABCG1 in leukocytes from patients with type 2 diabetes was reduced and strongly correlated with the level of glycaemia.[188–190]

2.4 Excretion

The effect of obesity on kidney function is not clear.[125,191] However, glomerular filtration and tubular secretion has been shown to be increased in obese individuals, whereas tubular reabsorption has been decreased by the disease.[125] Clearance does not linearly increase with body weight in obese individuals.[123]

Diabetes-related nephropathy is the leading cause of end-stage renal disease in industrialized countries.[192,193] Nephropathy develops in approximately 40% of diabetic patients.[194,195] It has been observed that function of the renin-angiotensin system is stimulated in diabetic patients with nephropathy and angiotensin II affects the synthesis of glomerular proteins.[196] Inhibitors of angiotensin-converting enzyme are used in type 2 diabetic patients with nephropathy to slow the progression of the disease.[197–199] Diabetes causes micro- and macrovascular changes and leads to hyperfiltration and, hence, an increased glomerular filtration rate.[200,201] It has also been observed that lowering medication doses correlates with a lowering of the glomerular filtration rate.[201] Although poor glycaemic control is an important risk factor, the glucose level does not fully explain why only a subset of patients with diabetes progress to end-stage renal disease.[202]

Numerous studies have been conducted to examine the effect of diabetes on the disposition of antibacterials in children. For example, the half-life and glomerular filtration rate of carbenicillin were determined in diabetic children[203] and the glomerular filtration rate was significantly increased. However, serum concentrations were not been affected for kanamycin, bekanamicin and amikacin.[204] Although diabetic patients are more prone to developing renal failure, studies show weak correlations between renal function and the clearance of drugs.

3. Effect of Diabetes Mellitus on Pharmacodynamics

Information regarding the effect of diabetes on the pharmacodynamics of drugs is very limited; however, evidence that the effect of diabetes is not limited to the pharmacokinetics of drugs exists. Almost all previously published studies have reported the effect of diabetes on the pharmacodynamics of cardiovascular and immunosuppressive drugs. However, the question is whether these influences reflect diabetes-mediated changes in pharmacokinetics rather than pharmacodynamics?

Cardiovascular diseases are the cause of death in approximately 80% of patients with diabetes and account for 75% of all hospitalizations in diabetes patients. Hence, not surprisingly, cardiovascular drug therapy is frequently started in subjects with diabetes. It has been suggested that diabetes may alter the pharmacokinetics of several cardiovascular drugs.[19] Recently published studies show a marginal effect; discrepancies occur in lipid-lowering[205–207] and antihypertensive drugs.[151,178,179] However, very few clinically significant diabetes-mediated changes in the pharmacokinetics of cardiovascular drugs have been reported. It has been demonstrated that diabetic patients exhibited significant reduction in isoprenalin-induced heart rate as compared with non-diabetic patients.[142,208] Similar data have been observed for both atropine and propranolol (pharmacokinetic data have not been reported).[209] Reduced biological activity of catecholamines due to changes protein binding has been suggested as one possible mechanism. However, no changes were found in the protein binding of catecholamines.[141] Tissue adaptation on changes in the insulin level is also suggested as pathophysiological,[210,211] and the hypothesis of changed G-protein function has been proposed.[212] In addition, Terada et al.[213] reported an increase in the chronothropic response to administration of exogenous epinephrine in diabetic patients as compared with non-diabetic patients. Packer et al.[214] reported a greater reduction in left ventricular filling pressure and mean right atrial pressure in diabetic patients as compared with non-diabetic patients. These changes can be attributed to a renin/aldosterone synthesis and angiotensin-mediated vasoregulation that are known to exist in diabetic patients. On the other hand, no difference in mean prothrombin time has been reported between diabetic and non-diabetic patients when warfarin was administered.[215]

Because diabetes is associated with renal microvascular damage, the glomerular filtration rate finally declines below 15 mL/min, leading to end-stage renal disease, which is irreversible and fatal if untreated, but a patient’s life can be sustained through haemodialysis followed by kidney transplantation. Approximately 40% of all kidney transplant recipients in the US have diabetes at the time of transplantation surgery.[216] In addition, post-transplant diabetes occurs in 4–20% of renal transplant recipients.[217,218] However, post-transplant diabetes is not only typical for renal transplant recipients, it also develops in 2.5–25% of liver transplant recipients, 4–40% of heart transplant recipients and 30–35% of lung transplant recipients.[219] To prevent organ rejection, a transplant recipient remains dependent on lifelong therapy with a cocktail of immunosuppressive agents (a calcineurin/mammalian target of rapamycin [mTOR] inhibitor, antiproliferative drug and corticosteroid).

The literature is consistent about the effect of the disease on the pharmacokinetics of immunosuppressive drugs, reporting no change in the unbound fraction of pharmacologically active ciclosporin,[220,221] or the area under the plasma concentration-time curve of tacrolimus[222,223] and mycophenolic acid.[224–226] However, diabetes has been shown to reduce expression of B-cell surface markers and markers of T-cell activity,[227,228] and increase the risk of infection-related mortality in diabetic subjects.[229,230] Previously published studies also demonstrated diabetes-mediated reductions in (i) protein leakage;[231–233] (ii) microvascular responses to inflammatory mediators;[234,235] (iii) degranulation of mast cells;[236] (iv) leukocyte-endothelial cell interactions;[237–242] (v) lymph node retention capacity;[243] and (vi) release of tumour necrosis factor-α and interleukin-1β by leukocytes upon exposure to lipopolysaccharide.[244]

Although no effect of diabetes on the mycophenolic acid plasma concentration has previously been reported,[227,245,246] the disease highly significantly reduced gene expression, protein level and enzyme activity of both inosine 5′-monophosphate (IMP) dehydrogenase (IMPDH) type 1 and type 2.[227,246,247] The enzyme catalyses nicotinamide adenine dinucleotide-dependent oxidation of IMP to xanthosine 5′-monophosphate. This is the rate-limiting step in the de novo pathway for guanine nucleotide biosynthesis in B and T lymphocytes.[248] Although a clear relationship between increased infection-related morbidity and downregulation of IMPDH has not been reported, the present investigation provides novel observations that may partially explain the increased susceptibility to infections in diabetic patients.[249,250] Because of lower immune system activity, diabetic patients may require an increased dose of antibacterials or immunosuppressive drugs to achieve an adequate level of immunosuppressive activity. However, further clinical studies in a larger number of patients with diabetes are needed to verify whether optimization of antibacterial/immunosuppressive drug dosing is required for these patients.

Last, but not least, we would like to discuss antidiabetic drugs, which are used to keep diabetes under control. It is hard to study the effect of diabetes on the pharmacokinetics and pharmacodynamics of antidiabetic drugs because they are used mainly for the treatment of the disease. The major organs involved in elimination of these drugs are the liver and kidney (e.g. rosiglitazone and pioglitazone are eliminated mainly by CYP enzymes, whereas metformin is eliminated mainly via glomerular filtration and tubular secretion and to some extent via liver CYP enzymes). CYP2C and 3A subfamilies are the main enzymes involved in the biotransformation of antidiabetic drugs. Although diabetes affects the activity of CYP3A (discussed in section 2.3), no changes in the pharmacokinetics of antidiabetic drugs have previously been reported.[131,182,251–253]

4. Discussion and Conclusion

There are always ethical limitations in performing clinical studies. In order to overcome these limitations, chemically induced animal models of diabetes have been established. The effects of diabetes on the pharmacokinetics and pharmacodynamics of drugs have been well described in experimental models of diabetes; however, in humans only minimal data exists and the effects of the disease on such properties are not yet clear. Nevertheless, it has been shown that the pharmacokinetics and pharmacodynamics of drugs are changed in diabetes. Available data also indicate that there is a significant variability in drug response in diabetic patients. For example, gastric emptying rates in patients with type 2 diabetes have been reported to be decreased, unchanged or accelerated (as discussed in section 2.1). The reasons for the discrepancies between individual clinical studies as well as between experimental and clinical studies are probably due to (i) the restricted and focused population of subjects in clinical studies; (ii) failure to consider type, severity and duration of the disease; (iii) histopathological characteristics generally being missing; and (iv) other factors, including varying medication use, dietary protein intake, obesity, age and sex. An understanding of diabetes-mediated changes in pharmacodynamics as well as the source of the variability in patient responses to treatment should lead to better medical management of diabetic patients and improvement of clinical outcomes in this population.

In 1991, Gwilt et al.[19] concluded their excellent comprehensive review on the effect of diabetes on pharmacokinetics and pharmacodynamics by saying that further clinical studies are warranted to explain the variability in observed data and to understand the mechanisms behind diabetes-mediated changes. Although clinical research has begun to make significant progress in characterizing the pathophysiology and genetic basis of diabetes, available information regarding the effect of the disease on pharmacokinetics and pharmacodynamics, efficacy and safety of drugs is limited and inconsistent. In addition, diabetes-related changes in the pharmacodynamics of drugs have been studied less than changes in pharmacokinetics. Previously published data also suggest that these diabetes-mediated changes in the pharmacokinetics of a particular drug cannot be translated to others. Unfortunately, our conclusions are similar to those published by Gwilt and colleagues[19] 20 years ago. The effect of diabetes on pharmacokinetics and pharmacodynamics remains unclear and further clinical studies are required to understand the clinical significance of the effects.