Abstract
Background
Diabetes mellitus (DM) is a global health issue associated with increased cardiovascular disease risk. Endothelial dysfunction is a known precursor to atherosclerosis and cardiovascular disease, and its role in the pathogenesis of DM complications is well-documented. There is limited information in the evaluation of endothelial function in prediabetic patients using flow-mediated dilation (FMD), and studies have not excluded patients with known atherosclerosis or coronary artery disease. Thus, in this study, we aimed to evaluate the endothelial functions using FMD from the brachial artery of DM and prediabetes patients who had normal coronary arteries.
Methods
This study included 73 participants: 25 with DM, 25 with prediabetes, and 23 normoglycemic controls, all with normal coronary arteries on angiography. FMD measurements were conducted following established protocols, and statistical analysis was performed using standard methods to compare FMD levels among groups.
Results
The groups were comparable in clinical and demographic characteristics, except for fasting plasma glucose levels. Significant differences in FMD levels were observed: 10.1% in the DM group, 16.5% in the prediabetes group, and 14.8% in the control group (P = 0.004). Diabetic patients had significantly lower FMD levels than both prediabetic and normoglycemic individuals. No significant difference in FMD was found between prediabetic and control groups.
Conclusions
Diabetic patients exhibited significant endothelial dysfunction compared to normoglycemic individuals, while prediabetic patients did not show similar dysfunction. These findings suggest a window of opportunity in the prediabetic stage for early intervention to prevent advanced endothelial dysfunction.
Trial registration number
ISRCTN Registry ISRCTN15351014. Registry date: 23/09/2024. Retrospectively registered.
Keywords: Flow mediated dilation, Endothelial dysfunction, Atherosclerosis, Diabetes mellitus, Prediabetes, Coronary artery disease, Coronary angiography
Introduction
Diabetes mellitus (DM) is an important global pandemic disease and is associated with an increased incidence and mortality of cardiovascular disease [1]. Endothelial dysfunction is considered a precursor to atherosclerosis and cardiovascular disease (CVD) and the role of endothelial dysfunction in the pathogenesis of both microvascular and macrovascular complications is well-documented in patients with DM [2]. One of the most widely used non-invasive methods for assessing endothelial function in arteries is flow-mediated dilation (FMD) of the brachial artery [2]. The assessment of peripheral endothelial function via FMD has been shown to correlate with coronary artery endothelial function [3].
Prediabetes is an important diagnosis, enabling physicians to implement measures to prevent the progression of DM and its related complications or to slow the disease process. Evaluating endothelial function in prediabetic patients may allow for the early detection and treatment of issues before further complications arise [2]. However, there is limited information in the literature regarding the evaluation of endothelial function in prediabetic patients using FMD, and studies have not excluded patients with known atherosclerosis or coronary artery disease [4]. The relationship between the presence and severity of coronary artery disease and the degree of endothelial dysfunction is well established [5]. Therefore, assessing endothelial function in a prediabetic population without excluding the presence of coronary artery disease presents limitations in demonstrating a causal relationship between endothelial dysfunction and prediabetes. Thus, in this study, we aimed to evaluate the endothelial functions using FMD from the brachial artery of DM and prediabetes patients who had normal coronary arteries.
Materials and methods
Patients evaluated in the Cardiology outpatient clinic who were found to have normal coronary arteries on coronary angiography were included in the study. The study was conducted in compliance with the Declaration of Helsinki. The study protocol was approved by the Institutional Committee on Human Research and Ethics. All patients provided written informed consent.
A total of 73 patients were included in the study, comprising 25 individuals with DM, 25 with prediabetes, and 23 normoglycemic participants. All patients with any coronary artery lesions and/or coronary artery wall irregularities that may be associated with atherosclerosis in the coronary arteries on coronary angiography were excluded from the study. Patients with a history of previous percutaneous coronary intervention, myocardial infarction, atrial fibrillation, heart failure, more than mild valvular regurgitation, valvular stenosis, uncontrolled hypertension (systolic blood pressure ≥ 140 mmHg and diastolic blood pressure ≥ 90 mmHg in the last 6 weeks), stroke, transient ischemic attack or cardiomyopathy were excluded from the study.
Diagnosis of prediabetes and diabetes mellitus
Blood samples were collected from the patients in the morning after an 8-hour fasting period. Prediabetes and DM were diagnosed according to American Diabetes Association criteria [1]. Prediabetes was defined as fasting plasma glucose (FPG) level between 100 and 125 mg/dL and a 2-hour plasma glucose after a 75 g oral glucose challenge ranging from 140 to 199 mg/dl or HbA1c level was 5.7–6.4%. Diabetes mellitus was diagnosed if FPG level ≥ 126 mg/dL or the 2-hour plasma glucose level after a 75 g oral glucose loading was ≥ 200 mg/dl or HbA1c ≥ 6.5% or the patient was on antidiabetic therapy. Patients with FPG levels < 100 mg/dL and HbA1c levels < 5.7% were included in the normal control group.
Measurement of flow mediated dilation
Study subjects refrained from smoking and consuming caffeinated beverages for 24 h before FMD measurement. FMD procedures were conducted in the morning hours following an 8-hour fasting period.
All echocardiographic and FMD measurements were conducted and assessed by a single observer who was unaware of the patient’s glycemic status. The FMD measurements were performed following the guidelines outlined in previously published literature [2, 6–8]. The study took place after an 8-hour fasting period in a serene, temperature-controlled room. The measurements were carried out by an investigator proficient in vascular ultrasound techniques, utilizing an Acuson CV70 ultrasound system from Mountain View, CA, USA equipped with a 5–10 MHz linear array transducer. To facilitate measurements, a sphygmomanometric cuff was positioned proximally to the antecubital fossa. A baseline image was captured, following which the cuff was inflated to 250 mmHg or at least 50 mmHg above the systolic pressure to induce arterial occlusion. After 5 min, the cuff was deflated. The longitudinal image of the artery was continuously recorded from 30 s before to 2 min after cuff deflation. In order to measure the hyperemic blood flow velocity, the mid-artery pulsed Doppler signal was immediately assessed after cuff deflation. Ten minutes after reactive hyperemia, when baseline conditions were reestablished, resting arterial flow was measured again. Subsequently, a 5 mg sublingual tablet of isosorbide dinitrate was administered to determine the maximum vasodilator response. Images were continuously recorded for 4 min after nitrate administration. The maximal brachial artery diameter was measured from an image taken at the peak of T wave on ECG. FMD was calculated using the formula below:
FMD (%): (Peak diameter-baseline diameter)/Baseline diameter x100.
Statistical analysis
Statistical analysis was conducted using a commercial software program (Statistical Package for the Social Sciences, Version 17.0, SPSS Inc., Chicago, IL, USA). All continuous variables checked with Kolmogorov-Smirnov normality test to show their distributions. Continuous variables with normal distributions were compared using the paired and unpaired t-tests. Continuous variables with abnormal distributions were compared using the Mann–Whitney U and Wilcoxon tests. Continuous variables are expressed as means ± standard deviation (median). For categorical variables, the chi-square test was used. Values for P less than 0.05 were considered statistically significant.
Results
Clinical and laboratory variables
The clinical, demographic characteristics, and laboratory findings of the three groups were comparable, except for gender and fasting plasma glucose (FPG) levels, as illustrated in Tables 1 and 2. The FPG levels of the DM group were significantly higher than those of the prediabetic and control groups (DM vs. prediabetic: P = 0.001; DM vs. control: P < 0.001). Although the FPG levels of the prediabetic group were higher than those of the control group, this difference was not statistically significant (P = 0.44). The medications, including antihypertensive and lipid-lowering agents, prescribed to the patients were similar across all groups, ensuring consistency in the study parameters.
Table 1.
Demographic and clinical characteristics
| Control group N: 23 |
Prediabetes group N: 25 |
Diabetes group N: 25 |
P value | |
|---|---|---|---|---|
| Age (years) | 54,1 ± 6,0 (51) | 55,9 ± 6,6 (54) | 56,7 ± 5,6 (57) | 0,454 |
| Body mass index (kg/m2) | 30,8 ± 7,1 (29.3) | 29.9 ± 4.9 (28.8) | 30,7 ± 3,7 (30.1) | 0,816 |
| Waist circumference (cm) | 104,2 ± 14,2 (103) | 99.9 ± 17,8 (99) | 102,7 ± 10,0 (100) | 0,589 |
| Hip circumference (cm) | 116,1 ± 15,3 (115) | 118,0 ± 15,9 (116) | 118,7 ± 12,9 (118) | 0,825 |
| Gender, Female, n (%) | 12 (52,2) | 20 (80) | 18 (72) | 0,105 |
| Hypertension, n (%) | 17 (73,9) | 16 (64) | 18 (72) | 0,726 |
| Systolic blood pressure* | 130,3 ± 12,8 (125) | 131,1 ± 17,7 (125) | 132,9 ± 11,7 (130) | 0,859 |
| Diastolic blood pressure* | 80,3 ± 4,5 (80 | 81,6 ± 7,9 (80) | 81,6 ± 6,5 (80) | 0,796 |
| Smoking, n (%) | 6 (26,1) | 7 (28) | 3 (12) | 0,331 |
| Dyslipidemia, n (%) | 12 (52,2) | 11 (44) | 18 (72) | 0,123 |
| Ejection Fractions (%) | 64,3 ± 3,8 (63) | 63,4 ± 3,1 (62) | 63,6 ± 6,3 (62) | 0,803 |
| Drugs | ||||
| Aspirin, n (%) | 1 (4,3) | 3 (12) | 3 (12) | 0,555 |
| Beta-blockers, n (%) | 2 (8,7) | 5 (20) | 6 (24) | 0,342 |
| CCB, n (%) | 3 (13) | 1 (4) | 3 (12) | 0,555 |
| ACEI, n (%) | 2 (8,7) | 1 (4) | 3 (12) | 0,635 |
| ARB, n (%) | 2 (8,7) | 7 (28) | 10 (40) | 0,045 |
| Statin, n (%) | 3 (13) | 2 (8) | 5 (20) | 0,523 |
CCC: calcium channel blocker, ACEI: angiotensin converting enzyme inhibitor, ARB: angiotensin receptor blocker. * The blood pressure values taken before the FMD measurement
Table 2.
Laboratory results
| Control group N: 23 |
Prediabetes group N: 25 |
Diabetes group N: 25 |
P value | |
|---|---|---|---|---|
| BUN (mg/dL) | 14,9 ± 5,6 (13.1) | 15,4 ± 3,6 (14.9) | 16,1 ± 4,3 (15.9) | 0,646 |
| Creatinine (mg/dL) | 0,8 ± 0,2 (0.8) | 0,7 ± 0,2 (0.7) | 0,8 ± 0,2 (0.7) | 0,119 |
| Uric acid (mg/dL) | 5,7 ± 1,5 (5.5) | 5,3 ± 1,6 (5.1) | 5,0 ± 1,5 (4.5) | 0,429 |
| ALT (U/L) | 26,5 ± 16,3 (20) | 24,2 ± 11,3 (20) | 22,7 ± 15,8 (20) | 0,725 |
| CRP (mg/L) | 4,8 ± 2,1 (4.1) | 4,8 ± 3,0 (3.73) | 4,8 ± 4,3 (4.0) | 0,997 |
| HbA1c | - | 5,5 ± 0,1 (5.5) | 8,7 ± 1,8 (8.2) | 0,036 |
| Hemoglobin (g/dL) | 14,2 ± 1,5 (13.5) | 14,2 ± 1,1 (13.8) | 13,6 ± 1,7 (13.2) | 0,359 |
| TSH (uU/mL) | 1,5 ± 0,4 (1.2) | 2,2 ± 1,5 (2.0) | 2,0 ± 1,7 (1.5) | 0,436 |
| Total cholesterol (mg/dL) | 197,1 ± 36,8 (189) | 190,5 ± 39,5 (179) | 196,6 ± 37,0 (183) | 0,798 |
| LDL (mg/dL) | 122,0 ± 36,6 (114) | 117,5 ± 39,1 (125) | 115,0 ± 32,2 (109) | 0,790 |
| HDL (mg/dL) | 43,0 ± 9,8 (40.1) | 44,5 ± 15,8 (44.8) | 42,5 ± 9,4 (44.2) | 0,828 |
| Triglyceride (mg/dL) | 156,0 ± 67,1 (138) | 136,0 ± 56,2 (130) | 188,1 ± 93,5 (149) | 0,053 |
| Fasting blood glucose (mg/dL) | 91,6 ± 5,7 (92) | 105,2 ± 8,6 (103) | 145,2 ± 63,3 (126) | < 0,001 |
Subgroup analysis for fasting blood glucose (Tukey): Diabetes vs. prediabetes: p = 0,001; Diabetes vs. Control: p < 0,001; Prediabetes vs. Control: p = 0,440
Analysis of FMD measurements
There were no significant differences observed between the basal brachial artery diameters across the three groups (Fig. 1). Basal brachial artery diameters were as follows: DM group 4.14 ± 0.47 mm (4.1); prediabetes group 4.13 ± 0.52 (4.1) mm; control group 4.18 ± 0.59 (3.9) mm (P = 0.942). Throughout the FMD study and following the administration of isosorbide dinitrate, the patients remained stable, and no adverse effects were observed. Hyperemic brachial artery diameters were less in the diabetic group than in the prediabetic and control groups, but this difference did not reach statistical significance. Hyperemic brachial artery diameters were noted as follows: DM group 4.56 ± 0.46 (44) mm; prediabetes group 4.81 ± 0.66 (47) mm; control group 4.8 ± 0.72 (46) mm (P = 0.272).
Fig. 1.
Basal and hyperemic brachial artery diameters of the three groups. The values given in this figure are “mean” values. Detailed data are given in the “Results” section, and the applied statistical analysis methods are given in the “statistical analysis” section
Significant differences were observed in the FMD levels among the three groups, as depicted in Fig. 2. FMD levels were as follows: DM group 10.1%; prediabetes group 16.5%; control group 14.8% (P = 0.004). It was found that FMD levels in the DM group were markedly lower than those in the prediabetic and control groups (DM vs. prediabetes: P = 0.006; DM vs. control: P = 0.030). However, the FMD levels in the prediabetic group did not show a notable difference when compared to the control group (P = 0.878).
Fig. 2.
Flow mediated dilation results of the three groups. The values given in this figure are “mean” values. Detailed data are given in the “Results” section, and the applied statistical analysis methods are given in the “statistical analysis” section
Discussion
In this study, we evaluated the endothelial functions of prediabetic and DM patients with normal coronary arteries and no known atherosclerotic vascular disease using FMD. We found that diabetic patients exhibited significant endothelial dysfunction compared to normoglycemic individuals, whereas prediabetic patients had endothelial function similar to that of normoglycemic individuals.
The distinction of our study lies in its focus on patients with DM and prediabetes who had normal coronary arteries and no known atherosclerotic vascular disease. This specific subgroup has received limited attention in previous research. Prior studies investigating endothelial function in DM and prediabetes patients often lacked comprehensive data regarding coronary anatomy or included patients with obstructive coronary artery disease, making direct comparisons challenging. According to our available literature review, there are no other studies evaluating the relationship between DM, prediabetes, and endothelial function using FMD in a patient group without known atherosclerotic vascular disease. The correlation between impaired FMD and coronary artery disease (CAD) and heart failure has been demonstrated in patients with CAD [9–12]. Manganaro et al. [5] showed that the degree of reduction in FMD correlates with the severity of coronary artery disease, suggesting that FMD at the brachial artery is a reliable indicator of CAD burden. Considering these findings, we believe that excluding patients with CAD from our study helped to more accurately evaluate the relationship between prediabetes and endothelial function.
A study similar to ours, conducted by Su et al. [4], evaluated endothelial function with FMD in patients with DM and prediabetes (impaired glucose tolerance - IGT, and impaired fasting glucose- IFG). Unlike our study, this research showed that compared with subjects with normal glucose tolerance, those with IFG and IGT had impaired FMD, and this impairment was more pronounced in subjects with type 2 diabetes mellitus than those with IFG and IGT. We believe that the differences between our study and that of Su et al. can be attributed to several factors. First, our study excluded patients with coronary artery disease and known atherosclerotic vascular disease, while Su et al. [4] study did not consider this factor. The relationship between the presence of coronary artery disease and atherosclerotic vascular disease and endothelial dysfunction is well-known [2, 5]. The inclusion of such patients could have led to worse endothelial function results and complicated the findings in prediabetic patients. Second, Su et al. study [4] excluded patients with a diagnosis of hypertension, whereas our study included a similar distribution of hypertensive patients across the groups to represent the normoglycemic, diabetic, and prediabetic populations in the community. However, all patients in our study had controlled blood pressure. The relationship between hypertension and endothelial function is well-established [8]. These factors likely contribute to the differences in findings between our study and that of Su et al.
Endothelial dysfunction is a key player in the development of vascular processes and occurs early in the development of atherosclerosis [2, 8] and endothelial dysfunction is a critical part of the pathogenesis of microvascular and macrovascular complications in DM [2]. Endothelium-dependent vasodilation in peripheral and coronary arteries of patients with DM is blunted, principally due to the loss or reduction of nitric oxide. Proposed mechanisms of endothelial dysfunction in the setting of DM include increased oxidative stress, uncoupling of endothelial nitric oxide synthase, pro-inflammatory activation of endothelial cells, mitochondrial dysfunction, impaired endothelial repair potential, and increased permeability [2]. Endothelial functions can be effectively assessed by FMD measurement from the brachial artery, which is the most commonly used method for evaluating endothelial function. While FMD has predicted cardiovascular risk in some large clinical trials, this correlation was not observed in all studies [2]. Nevertheless, FMD-derived assessments of brachial artery function are strongly correlated with coronary endothelial function [5–7].
Study limitations
In our study, to diagnose normal coronary arteries, a quantitative test such as IVUS or OCT in addition to coronary angiography could have made our data more objective and valuable. However, due to costs, we could not perform IVUS or OCT on our patients. We did not perform routine Hba1c testing in the control group because their fasting blood sugar levels were within normal limits. Performing routine Hba1c testing in addition to fasting blood sugar measurement in the control group could have made our data more valuable.
In this study, we used FMD, the most commonly employed non-invasive method for assessing endothelial function. There is no universally optimal method for evaluating coronary function, as both invasive and non-invasive techniques have distinct advantages and limitations. Invasive procedures, which assess coronary epicardial and microvascular function, offer high accuracy but are associated with considerable costs, risks, and time demands. These methods require sophisticated equipment and specialized expertise. Conversely, Flow-Mediated Dilation (FMD) of the brachial arteries presents a non-invasive alternative that is generally safer and less resource-intensive. Nevertheless, FMD is technically complex and necessitates extensive training and rigorous standardization to ensure reliable results. While FMD provides valuable information regarding endothelial function and can offer indirect insights into coronary health, it may not be as precise in evaluating coronary microvascular function compared to invasive techniques.
Conclusions
Our study demonstrated that early prediabetic patients with normal coronary arteries do not yet exhibit endothelial dysfunction. This observation underscores the potential window of opportunity during the prediabetic stage. Implementing early interventions, such as lifestyle modifications and close monitoring, could potentially mitigate the risk of developing advanced endothelial dysfunction and related complications in these patients. Additionally, our study supports the guidelines recommending aggressive management of risk factors in patients with DM, even in the absence of atherosclerotic lesions in the coronary arteries.
Author contributions
A.I.A, T.K.A, H.B, I.A and M.E.K designed the research study. A.I.A, T.K.A and H.B performed the research. A.I.A and I.A analyzed the data. All authors contributed to writing of the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.
Funding
None.
Data availability
Data that support the findings of this study are available from the authors upon request.
Declarations
Ethics approval and consent to participate
This study was approved by the Institutional Review Board of Ankara Guven Hospital and the protocols used in the study were approved by this board. The approval number is 190.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.ElSayed NA, Aleppo G, Aroda VR, Bannuru RR, Brown FM, Bruemmer D, Collins BS, Hilliard ME, Isaacs D, Johnson EL, Kahan S, Khunti K, Leon J, Lyons SK, Perry ML, Prahalad P, Pratley RE, Seley JJ, Stanton RC, Gabbay RA. On behalf of the American Diabetes Association. 2. Classification and diagnosis of diabetes: standards of Care in Diabetes-2023. Diabetes Care. 2023;46(Suppl 1):S19–40. 10.2337/dc23-S002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Alexander Y, Osto E, Schmidt-Trucksäss A, Shechter M, Trifunovic D, Duncker DJ, Aboyans V, Bäck M, Badimon L, Cosentino F, De Carlo M, Dorobantu M, Harrison DG, Guzik TJ, Hoefer I, Morris PD, Norata GD, Suades R, Taddei S, Vilahur G, Waltenberger J, Weber C, Wilkinson F, Bochaton-Piallat ML, Evans PC. Endothelial function in cardiovascular medicine: a consensus paper of the European Society of Cardiology Working Groups on Atherosclerosis and Vascular Biology, Aorta and Peripheral Vascular diseases, coronary Pathophysiology and Microcirculation, and thrombosis. Cardiovasc Res. 2021;117(1):29–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Anderson TJ, Uehata A, Gerhard MD, Meredith IT, Knab S, Delagrange D, Lieberman EH, Ganz P, Creager MA, Yeung AC, Selwyn AP. Close relation of endothelial function in the human coronary and peripheral circulations. J Am Coll Cardiol. 1995;26:1235–41. [DOI] [PubMed] [Google Scholar]
- 4.Su Y, Liu XM, Sun YM, Wang YY, Luan Y, Wu Y. Endothelial dysfunction in impaired fasting glycemia, impaired glucose tolerance, and type 2 diabetes mellitus. Am J Cardiol. 2008;102(4):497–8. 10.1016/j.amjcard.2008.03.087. [DOI] [PubMed] [Google Scholar]
- 5.Manganaro A, Ciracì L, Andrè L, Trio O, Manganaro R, Saporito F, Oreto G, Andò G. Endothelial dysfunction in patients with coronary artery disease: insights from a flow-mediated dilation study. Clin Appl Thromb Hemost. 2014;20(6):583–8. 10.1177/1076029614524620. [DOI] [PubMed] [Google Scholar]
- 6.Corretti MC, Anderson TJ, Benjamin EJ, Celermajer D, Charbonneau F, Creager MA, et al. Guidelines for the ultrasound assessment of endothelial-dependent flow-mediated vasodilation of the brachial artery: a report of the International Brachial artery reactivity Task Force. J Am Coll Cardiol. 2002;39(2):257–65. [DOI] [PubMed] [Google Scholar]
- 7.Thijssen DH, Black MA, Pyke KE, Padilla J, Atkinson G, Harris RA, Parker B, Widlansky ME, Tschakovsky ME, Green DJ. Assessment of flow-mediated dilation in humans: a methodological and physiological guideline. Am J Physiol Heart Circ Physiol. 2011;300(1):H2–12. 10.1152/ajpheart.00471.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ambrosino P, Bachetti T, Ennio S, D’Anna SE, Galloway B, Bianco A, D’Agnano V, Papa A, Motta A, Perrotta F. Mechanisms and clinical implications of endothelial dysfunction in arterial hypertension. J. Cardiovasc. Dev. Dis. 2022;9:136. [DOI] [PMC free article] [PubMed]
- 9.Kaku B, Mizuno S, Ohsato K. The correlation between coronary stenosis index and flow-mediated dilation of the brachial artery. Jpn Circ J.,1998;62(6):425-30. [DOI] [PubMed]
- 10.Wu TC, Chen YH, Chen JW, et al. Impaired forearm reactive hyperemia is related to late Restenosis after coronary stenting. Am J Cardiol. 2000;85(9):1071-6. [DOI] [PubMed]
- 11.Kuvin JT, Patel AR, Sliney KA, et al. Peripheral vascular endothelial function testing as a noninvasive indicator of coronary artery disease. J Am Coll Cardiol. 2001;38(7):1843-9. [DOI] [PubMed]
- 12.Cassano V, Armentaro G, Magurno M, et al. Short-term effect of sacubitril/valsartan on endothelial dysfunction and arterial stiffness in patients with chronic heart failure. Front Pharmacol. 2022;13:106982810.3389/fphar.2022.1069828. eCollection 2022. [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data that support the findings of this study are available from the authors upon request.