Organic nitrate esters, such as glyceryl trinitrate and isosorbide dinitrate, are potent vasodilators that have been used extensively in cardiovascular therapy. About one third of patients with diabetes who have coronary artery disease may be receiving treatment with organic nitrates [7]. Increasing evidence suggests that the action of organic nitrates derives from metabolic conversion to nitric oxide, which relaxes the underlying vascular smooth muscle by increasing the production of cyclic guanosine monophosphate [8]. The availability of intracellular sulfhydryl groups, probably from glutathione and cystine, is thought to be important in the biotransformation of organic nitrates to nitric oxide [9]. The amount of intracellular thiol groups is likely to be depleted in patients with diabetes because of the increased oxidative stress brought about by hyperglycemia [10]. This may reduce the vascular effects of organic nitrates in patients with diabetes. Consistent with this possibility, impaired forearm vasodilatory response to glyceryl trinitrate, but not to the direct-acting nitric-oxide-donor sodium nitroprusside, has been seen in patients with diabetes compared with persons without diabetes [11].

Blood viscosity is now considered to be a major cardiovascular risk factor [12, 13] and has been implicated in vascular complications in diabetes [14, 15]. Surprisingly, knowledge of the influence of organic nitrates on blood rheology is lacking. We investigated the hemorrheologic and hemodynamic effects of short- and long-acting glyceryl trinitrate preparations in patients with diabetes compared with controls. We also evaluated the effects of glutathione and vitamin E to find out whether antioxidant supplementation could modify responses to glyceryl trinitrate in patients with diabetes.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

We studied 80 persons who were 44 to 65 years of age. Forty patients with diabetes were recruited from the outpatient departments of our institutions. All had non-insulin-dependent diabetes mellitus, had been older than 40 years of age at diagnosis, had had diabetes for fewer than 10 years, and had no clinical or laboratory evidence of coronary or peripheral artery disease. They were normotensive and were receiving no drugs other than therapy for diabetes (sulfonylureas or biguanides). Controls were 40 healthy laboratory staff members matched to the patients with diabetes for age, sex, and smoking status. They were receiving no drugs. The clinical characteristics of both groups are shown in Table 1. The study was approved by the local ethics committee, and all participants gave written informed consent.

Procedures

All participants were studied while fasting in the morning and lying supine. Special care was taken to ensure that drugs containing aspirin or related compounds had not been used recently. Smoking was not permitted on the test day. For automatic recording of systolic, diastolic, and mean arterial pressure and heart rate, the participants were instrumented with a noninvasive device (Finapres, Ohmeda 2300, Englewood, California) that has been shown to be as accurate as intra-arterial blood pressure measurement techniques [16]. Ventricular function was assessed using echocardiography. The study started after the participants had rested for at least 30 minutes and after three consecutive measurements of blood pressure and heart rate had differed by less than 5%. Twenty patients with diabetes and 20 matched controls were each given a single sublingual dose of glyceryl trinitrate (0.3 mg), and hemodynamic values were recorded 15 and 60 minutes later. At these times, adequate blood samples without stasis were taken for rheologic measurements. The participants repeated the test twice more, after receiving vitamin E, 300 mg/d orally, for 7 days (Evitum, Lipha, Calenzano, Italy) and during glutathione infusion (Tationil, Boehringer Mannheim, Milan, Italy). Glutathione infusion (600 mg as an intravenous bolus followed by 600 mg/h) was started 30 minutes before and was continued for 60 minutes after administration of glyceryl trinitrate. The three tests were separated by at least 7 days and were done in random order.

The effects of transdermal glyceryl trinitrate patches (Adesitrin, Pharmacia, Milan, Italy; 10 mg/d) were investigated in the other 20 patients with diabetes and 20 matched controls; hemorrheologic and hemodynamic measurements were taken after 1 hour and after 12 hours. The protocol was repeated after the participants had received vitamin E, 300 mg/d, and glutathione, 600 mg/d intramuscularly, for 7 days. One half of participants put the patch on at 8:00 a.m., and the other half put it on at 8:00 p.m., in random order. The sequence of the three tests was randomized.

Blood Rheology

Platelet aggregation response induced by 0.5 and 1.25 µmol of adenosine diphosphate was determined according to the method of Born [17]. Aliquots of blood anticoagulated with 0.77 mol/L of ethylenediaminetetraacetic acid (the ratio of blood to ethylenediaminetetraacetic acid was 1:20) were used to assess blood viscosity at high shear rates (450 s^–1) using a Brookfield Digital Viscosimeter 0.8-degree cone (Brookfield Engineering Laboratories, Stoughton, Massachusetts). Blood filterability was determined according to the method of Reid and coworkers [18]. Hematocrits were determined by centrifuging blood samples in glass capillary tubes for 5 minutes at 12000 revolutions per minute. All determinations were made in duplicate by a person blinded to participants and treatments. Coefficients of variation were 2% for blood viscosity, 3% for blood filterability, and 5% for platelet aggregation.

In Vitro Studies

Studies were done in vitro on aliquots of blood taken from 10 patients with diabetes and 10 matched controls. Platelet aggregation response to adenosine diphosphate was done after incubation of platelet-rich plasma with glyceryl trinitrate (100 and 200 ng/mL, end concentration) for least 5 minutes at 37 °C. Glyceryl trinitrate was freshly prepared in distilled water, shielded from light, and incubated for at least 5 minutes before blood viscosity and filterability were determined. Glutathione was tested at the end concentration of 200 µg/mL.

Statistical Analysis

All data are presented as mean ± SE; 95% CIs are provided when appropriate. Pairs of means were assessed using the Student t-test. Hemorrheologic and hemodynamic responses to drugs in vivo were analyzed by repeated measures analysis of variance. The test primarily analyzed the differences between the matched pairs (for example, patients with diabetes compared with controls and patients with diabetes compared with themselves) as "between-subjects" effects and then tested time and group as "within-subjects" effects. In the in vitro studies, dose was considered to be a within-subjects factor. Each group of patients with diabetes was studied three times (with glyceryl trinitrate alone, with glyceryl trinitrate and vitamin E, and with glyceryl trinitrate and glutathione). To correct for the correlations between measurements (both within patients and over time), an adequate wash-out period between different tests was allowed to elapse, and a Square Latin model (program 4V of BMDP) was introduced into the linear model of the analysis. Post hoc testing was done using the Scheffe test. A further adjustment for baseline differences was made considering the differences between baseline and follow-up measures seen in the four groups (differences of differences). Relations were determined using linear regression analysis. BMDP software was used.

Results

Top Methods Results Discussion Author & Article Info References

The two groups were well matched (Table 1). At baseline, platelet aggregation and blood viscosity were significantly higher and blood filterability was significantly lower in patients with diabetes than in controls. In all participants, there was a significant inverse relation between blood viscosity and filterability (r = –0.51 [95% CI, –0.6 to –0.42]; n = 80; P < 0.001). In patients with diabetes, there was a significant positive relation between hemoglobin A_1c levels and blood viscosity (r = 0.32 [95% CI, 0.24 to 0.41]; n = 40; P = 0.029). Heart rate increased from 76 ± 1.5 beats/min to 82 ± 1.5 beats/min (15-minute value; P < 0.01), whereas ejection fraction (from 0.51 ± 0.023 to 0.42 ± 0.022; P < 0.01) and mean blood pressure (from 96 ± 1.7 to 93 ± 1.7 mm Hg; P = 0.041) decreased significantly after sublingual administration of glyceryl trinitrate in controls. Baseline values were re-established within 60 minutes of administration. Platelet aggregation responses to adenosine diphosphate (both doses), blood viscosity, and blood filterability showed significant improvements after glyceryl trinitrate; these persisted until 60 minutes after administration (Figure 1).

View larger version (18K): [in this window] [in a new window]   Figure 1. Blood filterability and viscosity and platelet aggregation at baseline and 15 and 60 minutes after administration of sublingual glyceryl trinitrate (0.3 mg) in 20 controls and 20 patients with non-insulin-dependent diabetes. Patients with diabetes were also studied after vitamin E (vit E) supplementation and during glutathione administration (GSH). * equals significance between patients with diabetes and controls (P < 0.05 to 0.001); star equals significance compared with baseline values (P < 0.05 to 0.01).

Heart rate (from 77 ± 1.8 beats/min to 78 ±1.7 beats/min), mean blood pressure (from 96 ± 2 mm Hg to 95 ± 2 mm Hg), and ejection fraction (from 0.52 ± 0.022 to 0.55 ± 0.029) did not change significantly after administration of glyceryl trinitrate in patients with diabetes. Glyceryl trinitrate paradoxically increased platelet aggregation (24% [95% CI, 15% to 33%]; P = 0.023) and blood viscosity (6% [CI, 2.9% to 8.8%]; P = 0.045) and decreased blood filterability (–7% [CI, –9.5% to 4.4%]; P = 0.04). Both glutathione and vitamin E significantly improved baseline hemorrheologic variables and normalized the response to glyceryl trinitrate (Figure 1 and Table 2). Glutathione and vitamin E did not significantly influence hemodynamic values in the basal state or after administration of glyceryl trinitrate.

Interaction analysis did not show any effect of time (morning or evening) on responses to transdermal glyceryl trinitrate patches, and the data were pooled. In controls, glyceryl trinitrate patches increased heart rate (from 74 ± 1.4 beats/min to 79 ± 1.5 beats/min; P = 0.019) and decreased ejection fraction (0.53 ± 0.02 to 0.46 ± 0.02; P = 0.042) at 1 hour but not at 12 hours. Mean blood pressure decreased slightly but significantly over time (from 95 ± 1.9 mm Hg to 93 ± 1.8 mm Hg at 1 hour to 93 ± 1.7 mm Hg at 12 hours; P = 0.039). In patients with diabetes, glyceryl trinitrate patches did not significantly change blood pressure (from 96 ± 2 mm Hg to 95 ± 1.9 mm Hg to 95 ± 2.1 mm Hg), heart rate (from 77 ± 2 beats/min to 78 ± 1.8 beats/min to 79 ± 2 beats/min), or ejection fraction [from 0.52 ± 0.022 to 0.55 ± 0.029 to 0.51 ± 0.029]. The hemorrheologic responses to glyceryl trinitrate patches in controls and in patients with diabetes were similar to those obtained after sublingual administration of glyceryl trinitrate Figure 2, left). In particular, platelet aggregation to adenosine diphosphate, blood viscosity, and blood filterability improved after administration of glyceryl trinitrate patches in controls and deteriorated in patients with diabetes. Both glutathione and vitamin E improved these variables at baseline and normalized responses to glyceryl trinitrate patches in patients with diabetes.

View larger version (79K): [in this window] [in a new window]   Figure 2. Left. Blood filterability and viscosity and platelet aggregation at baseline and 1 and 12 hours after transdermal glyceryl trinitrate patches (10 mg) in 20 controls and 20 patients with diabetes. Patients with diabetes were also studied after vitamin E (vit E) supplementation and during glutathione administration (GSH). * equals significance between patients with diabetes and controls (P < 0.05 to 0.001); star equals significance compared with baseline values (P < 0.05 to 0.01). Right. Effects of glyceryl trinitrae on platelet aggregation and blood viscosity and filterability in vitro. Blood was taken from 10 patients with diabetes and 10 matched controls. * equals significance between patients with diabetes and controls (P < 0.05 to 0.01); star equals significance compared with baseline values (P < 0.01 to 0.02). Glutathione (GSH) was tested at the end concentration of 300 mg/mL.

Figure 2(right) shows the results of in vitro studies. The effects of glyceryl trinitrate on platelet aggregation and blood viscosity were similar to those obtained in ex vivo studies. In particular, patients with diabetes presented a paradoxical deterioration of these variables after administration of glyceryl trinitrate; this deterioration was normalized by glutathione.

Discussion

Top Methods Results Discussion Author & Article Info References

We found that erythrocyte filterability was lower and blood viscosity and platelet aggregation were higher in patients with non-insulin-dependent diabetes mellitus than in carefully matched controls. Various alterations of platelet function have been found in diabetes [19]. Blood viscosity is now thought to be a major cardiovascular risk factor [12] and is elevated in patients with diabetes and vascular complications [10, 14, 15]. Our results show that patients with diabetes and no evidence of cardiovascular disease also have increased blood viscosity.

The difference we found in blood viscosity between patients with diabetes and controls is unlikely to be due to known confounding factors [20], including sex, age, obesity, blood pressure, and lipid levels, because the two groups were well matched for these variables. The significant relation between blood viscosity and hemoglobin A_1c levels suggests a role for the hyperglycemia of diabetes. Although it contributes substantially to plasma viscosity, fibrinogen contributes little to blood viscosity [20]; a difference in plasma fibrinogen levels between patients with diabetes and controls is therefore unlikely to have played an important role. Erythrocyte rigidity and aggregability are determinants of blood viscosity [20], a finding confirmed by the significant relation that we found between erythrocyte filterability and viscosity.

To our knowledge, ours is the first study to show that glyceryl trinitrate improves blood viscosity and filterability in normal persons for at least 60 minutes after sublingual administration and 12 hours after transdermal patch. Because these changes were also seen when glyceryl trinitrate was added directly to blood in vitro, they are probably direct effects not mediated by compensatory hemodynamic responses. The antiplatelet effect of organic nitrates has long been known [21]. However, nitrates are thought to cause hypotension at the doses required for platelet inhibition in humans [22]. We found that platelet inhibition persisted when the weak hemodynamic responses to glyceryl trinitrate preparations had disappeared. Nitrates are considered to be exogenous nitric oxide donors, and nitric oxide inhibits platelet aggregation by increasing the level of platelet cyclic guanosine monophosphate [8]. Nitric oxide may also inhibit interactions of erythrocytes, as has been shown for other blood elements [23]. Further studies are needed. The modest hemodynamic changes that follow administration of glyceryl trinitrate are consistent with the recognized dose-dependent dilation of veins and arterioles after administration of organic nitrates.

The effects of glyceryl trinitrate are not apparent in patients with diabetes in whom paradoxical deterioration was seen. Impaired forearm arteriolar dilatory response to glyceryl trinitrate has been found in patients with diabetes [11] and has been linked to the depletion of intracellular thiol groups; this depletion is thought to contribute to the occurrence of nitrate tolerance in the clinical setting [9]. Limited availability of reduced glutathione is likely in diabetes. The increased production of free radicals brought about by hyperglycemia coupled with impaired regeneration of protective endogenous antioxidants (glutathione and vitamin E) are the two main causes of the increased oxidative stress of the patient with diabetes [11, 24]. Consistent with this, short-term antioxidant supplementation in patients with diabetes improved basal rheologic variables and normalized their responses to glyceryl trinitrate. This argues against important structural changes in erythrocyte membranes, such as glycation of membrane proteins [25]. Because glucose transport into erythrocytes is independent of insulin [26], metabolic pathways strictly associated with hyperglycemia, including polyol pathway and glucose autoxidation, can increase the production of free radicals [11, 27]. Consistent with this interpretation, increased release of malondialdehyde has been found in the erythrocyte membrane of patients with diabetes, together with a depressed erythrocyte glutathione content [28].

The paradoxical effects of glyceryl trinitrate in patients with diabetes could also be due to drug-drug interactions, either of a chemical nature or at the tissue level. For example, metformin has been shown to decrease platelet aggregability in patients with diabetes at doses of 1.5 to 3 g per day for as long as 6 months [29]. Although platelet aggregation did not differ among patients with diabetes receiving different treatment regimens (diet, sulfonylurea, or metformin), the small number of patients receiving metformin (n = 6) precluded appropriate statistical evaluation. This should be the subject of further research involving more patients.

Aside from their effects in improving the biotransformation of nitrates to nitric oxide, glutathione and vitamin E may also decrease the production of superoxide anion, which quenches nitric oxide [30]. In addition, vitamin E binds to endothelial cell receptors releasing prostacyclin, a potent antiaggregant agent [31]. The reason antioxidants seem to have little effect on hemodynamic values might be the modest changes induced by the doses of glyceryl trinitrate used and the possible presence of cardiovascular autonomic neuropathy in diabetes [32].

Our results may have clinical implications. The 35% reduction in mortality from myocardial infarction with the use of organic nitrates [33] may also involve an antithrombotic effect. However, patients with diabetes, especially non-insulin-dependent diabetes, continue to have excess cardiovascular mortality. The hospital mortality of patients with diabetes and acute myocardial infarction is between 25% and 40%, and the incidence of total reinfarction is high [34]. It is unclear whether reduced effects of classic nitric oxide donors (such as glyceryl trinitrate) might have been implicated, at least in those patients with diabetes who were treated with them. On the other hand, recognition in clinical practice of the lack of effects of organic nitrates in patients with diabetes has presumably been hampered by exclusion of patients with diabetes from clinical trials investigating the effect of organic nitrates in the acute ischemic syndromes. Antioxidant supplementation may represent a novel therapeutic approach enabling the antithrombotic effects of organic nitrates in diabetic patients.

Dr. Acampora: Centro Antidiabetico Lepanto, Via Lepanto, 80100 Napoli, Italy.

Drs. Donzella and Quatraro: Casa di Cura S. Rita, 74100 Taranto, Italy.