INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

ARTERIAL COMPLIANCE IS BECOMING an increasingly important clinical parameter. When reduced, it may be a potentially useful indicator of the presence of arterial disease (30). Functional changes in the arterial wall leading to reduced compliance may precede the onset of clinically apparent disease (clinical manifestations and/or arterial wall structural changes) and may identify individuals at risk before disease onset. A variety of techniques evaluating different arterial wall functions (arterial diameter, blood flow, and cross-sectional area) have been developed to measure this parameter (3, 10, 13, 14, 31). Some of them allow the assessment of large vessel compliance only, whereas others allow the assessment of large and small vessel compliance, i.e., resulting from changes at the arteriolar level. At present, there is no consensus as to the "best" method for measuring peripheral compliance, but it is important to note that a number of available techniques do not take into account that compliance is a pressure-dependent parameter.

Independent of the technique, most of the studies published so far have generally shown that differences among cases are not detectable in unstimulated conditions but become clearly evident when measured after the administration of organic nitrates (endothelium-independent vasodilators) or cholinergic stimuli (endothelium-dependent vasodilators). Because these vasoactive substances must be administered via a catheter inserted into the brachial artery, this may become an invasive, costly, and time-consuming procedure, associated with potential risk and discomfort, thus severely limiting clinical applicability.

A possible alternative endothelium-dependent stimulus useful for identifying endothelial dysfunction in humans in a noninvasive way is reactive hyperemia (20, 27, 34). This stimulus, obtained with a more or less prolonged arterial occlusion, induces an increased flow (27, 34) causing endothelium-mediated vasodilatation (20, 36) and increased compliance (17, 19).

In studies evaluating arterial compliance changes during reactive hyperemia, the postischemic modifications have been generally measured at arbitrary time intervals of 20-30 s, and the time course curve of postischemic variations was generally ignored.

We propose a new plethysmographic method useful for measuring forearm arterial compliance (FAC) [the area under the FAC-blood pressure (BP) curve (FAC_AUC)]. This method, which takes into account nonlinear compliance/BP dependence and allows continuous measurements, was previously validated by evaluating its capacity to detect acetylcholine-induced changes as well as differences between hypercholesterolemic patients and age- and sex-matched normal controls (2).

The present study was designed to 1) evaluate the ability of this method to appreciate FAC_AUC changes occurring during reactive hyperemia; 2) assess the validity of the aforementioned noncontinuous postischemic compliance measurement procedures; 3) further validate the plethysmographic method vis-à-vis measurements made with an echotracking device; 4) assess whether apart from the maximal compliance changes, known to be impaired in the presence of several pathological conditions (e.g., hypercholesterolemia) (17), other time-dependent postischemic parameters may be related to plasma lipids and other laboratory and clinical variables.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Subjects. The experimental group consisted of 95 volunteers (53 men and 42 women, age range: 18-80 yr) recruited from the medical staff of our institutions as well as from patients attending the E. Grossi Paoletti Center for the Study of Metabolic Diseases. In view of the methodological nature of the study, subjects were enrolled without any specific selection criteria. Oral informed consent was obtained from all subjects, and the study was approved by the Internal Review Board. The characteristics of the studied population are shown in Table 1.

Lipids. At the time of blood sampling for lipid analysis, the subjects were on a free diet. Venous blood was collected from the antecubital vein after an overnight fast and anticoagulated with EDTA (1 mg/ml). Total cholesterol and triglycerides were determined by enzyme methods (7, 29). High-density lipoprotein cholesterol (HDL-C) was separated by selective precipitation of ApoB-containing lipoproteins with dextran-sulfate-MgCl₂ (38). Plasma LDL-C levels were calculated according to the Friedewald formula (15).

BP measurements. Heart rate, DBP, mean BP, and SBP were continuously recorded from the middle finger of the dominant arm using a Finapress instrument (2300 Finapress blood pressure monitor, Ohmeda). This method provides accurate continuous BP measurements comparable with intra-arterial recordings (19, 26).

FAC measurements. The FAC of the nondominant arm was measured by a previously described plethysmographic method, which allows direct assessment of the nonlinear "FAC-BP curve" relative to each single cardiac cycle (2). Briefly, measurements of forearm pulse volume (FPV) were made by means of the extensimetric hemofluximeter Angiomed (Microlab Electronic; Padua, Italy). A double-stranded loop of fine-gauge silastic rubber tubing (internal diameter 0.5 mm, external diameter 2.1 mm) containing gallium, indium, and tin was placed around the upper forearm. Simultaneous recordings of FPV (Fig. 1, top) and BP (Fig. 1, bottom) were used to establish the relationship between FPV and BP (Fig. 2, top). The mathematical model that fits best with these experimental points is a logarithmic function mathematically expressed as

(1)

with correlation coefficients ranging from r = 0.72 to 0.99. In 72.7% of the cases, the correlation coefficient was >0.9 (see Fig. 2, top). a and b are the coefficients that fit best with the experimental points. x and y are BP and FPV, respectively.

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Simultaneous recordings of the pulsatile change in the forearm pulse volume (top) and hand finger arterial blood pressure (BP; bottom) in the preischemic condition and at the peak value during reactive hyperemia.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Example of the pulse volume-BP relationship (top) and compliance-BP curve (bottom) in the preischemic condition and at the peak value during reactive hyperemia.

(2)

(3)

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Typical example of a time course of compliance changes during reactive hyperemia showing the "true" area under the forearm arterial compliance (FAC)-BP curve (FACAUC) maximal change not otherwise detectable with noncontinuous measurements. Each FACAUC value is the average of 4-5 beats. a and b are coefficients of the mathematical function that best fit with the descent time.

Statistical analyses. Mean ± SD values were used as descriptive measures of normally distributed variables; in other cases, the median and range were used. Pre- and postischemic FAC_AUC comparisons were done by a Wilcoxon signed-rank test. Correlation analyses were performed using parametric methods (Pearson's moment correlation) after log transformation of triglycerides and plethysmographic variables. Multiple stepwise regression analysis was used to determine the relative importance of each variable (total cholesterol, HDL-C, LDL-C, triglycerides, age, and body mass index) in predicting pre- and postischemic FAC_AUC parameters. Statistical significance was accepted at a value of P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

FAC and reactive hyperemia. During reactive hyperemia, arterial BP did not change, whereas FPV increased (Fig. 1). The rise of FPV produced a higher slope of the FPV-BP curve and, as a consequence, an increase in the FAC_AUC (Fig. 2).

2

2

2

1

1

Time course of FAC_AUC changes during reactive hyperemia. To assess the FAC_AUC closest to the real peak value ("true peak value"), FAC_AUC measurements were carried out continuously, and the time course of FAC_AUC changes was defined. This protocol allowed analyzing the shape of the time course curve, thus providing the opportunity to understand whether the peak FAC_AUC value was maintained for a time period long enough to be detectable with noncontinuous measurements. The time-dependent plethysmographic characteristics of the 93 subjects studied are shown in Table 2. FAC_AUC reached its maximal value ~17 s after cuff deflation (range, 2-52 s) and, in 77% of subjects, was maintained for not more than 2.5-7 s, a time interval too short for legitimate noncontinuous measurements. Thus, in most subjects, descent starts almost immediately, and FAC_AUC returns to its respective preischemic values in ~63 s (descent time range, 8-196 s).

Repeatability of pre- and postischemic plethysmographic variables. Table 3 shows the pre- and postischemic plethysmographic variables obtained at the initial visit and after 1 mo in a subset of nine subjects; the CVs are also reported. Whereas the CV was <10% for preischemic FAC_AUC, peak FAC_AUC and percent change, it ranged between 16.2 and 62% for the time-related variables.

Agreement between the plethysmographic method and echotracking device in compliance determination. The 12 subjects recruited for the agreement study had a mean age (±SD) of 37 ± 11 yr (range, 27-55 yr). Eleven (92%) were males, and, with the exception of two smokers and two hypertensives, none was exposed to significant cardiovascular risk factors (e.g., blood pressure, lipids, or diabetes).

2

2

1

1

2

2

1

1

2

2

1

3

2

2

1

3

2

2

2

2

2

2

2

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Top: agreement between the plethysmographic and ultrasound technologies. The differences between measurements obtained with the plethysmographic and with the ultrasound methods are plotted versus their means. The mean difference and the limits of agreement are also indicated. Bottom: relationship between measurements obtained with the plethysmographic system and with ultrasound technology. RACAUC, area under the radial artery compliance (RAC)-BP curve.

Agreement between plethysmography and echotracking in assessment of compliance changes after GTN administration. The FAC_AUC and RAC_AUC responses to the two doses of sublingual GTN in the five subjects selected for the GTN study are illustrated in Fig. 5. In the three subjects treated with low doses of GTN (25 and 50 µg) and in the single subject treated with 50 and 100 µg of GTN, a dose-dependent increase of RAC_AUC was observed. In contrast, in the single subject treated with the higher doses of GTN (from 200 to 400 µg), a dose-dependent reduction of RAC_AUC was detected. All the compliance changes detected with the echotracking system were also appreciated using the plethysmographic device. Interestingly, although quite different in terms of absolute values, the trends of the compliance changes are substantially identical with the plethysmographic method and with the echotracking device (Fig. 5).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Changes in area under the compliance-BP curve induced by two doses of sublingual glyceryl trinitrate (GTN) as measured using the plethysmographic method proposed here () as well as echotracking technology () in 5 control subjects.

Agreement between plethysmography and echotracking in assessment of compliance changes during reactive hyperemia. The FAC_AUC and RAC_AUC responses to reactive hyperemia in three subjects are shown in Fig. 6. Although not identical in absolute terms, the kinetics of the compliance changes during reactive hyperemia are clearly appreciable using both the plethysmographic method and the echotracking device (Fig. 6).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Time-course curve of compliance changes during reactive hyperemia in 3 control subjects as measured using plethysmography () and echotracking technology ().

Duration of ischemic test and shape of the time course curve. In the 12 subjects enrolled to investigate whether duration of arterial occlusion may influence the shape of the time course curve, two complete plethysmographic examinations were performed at a 1-h interval. The FAC_AUC changes obtained after 3 and 12 min of arterial occlusion were then compared. As shown in Table 4, a prolonged arterial occlusion clearly produced higher values for each parameter considered. However, the time required to reach the peak was again highly variable (range, 2-27 s), and the time of peak maintenance was again very brief, not exceeding, in 83% of the cases, 14 s (range, 5-45 s).

Correlation between plethysmographic parameters and laboratory and clinical variables. To investigate whether the different postischemic plethysmographic parameters were related to laboratory and clinical variables, a series of correlation analyses was performed (Table 5). The peak FAC_AUC correlated inversely with SBP, whereas percent change correlated inversely with LDL-C and the LDL-C-to-HDL-C ratio and directly with HDL-C. Peak time correlated inversely with age and HDL-C, whereas descent time and the reserve area correlated inversely with age and log-transformed triglyceride content.

Multiple regression analyses. Table 6 shows a series of five forward stepwise multiple regression analyses performed using log-transformed plethysmographic parameters as dependent variables and laboratory/clinical variables as independent variables. In this series of analyses, sex was the only parameter independently associated with preischemic FAC_AUC (R² = 0.19, P = 0.001). Sex (directly) and SBP (inversely) significantly predicted the peak FAC_AUC value (R² = 0.26, P = 0.0001 and P = 0.01, respectively), whereas HDL-C (directly) and total cholesterol (inversely) correlated independently with the percent change (R² = 0.10, P = 0.01 and P = 0.04, respectively). Finally, age (inversely) and body mass index (directly) correlated with the peak time (R² = 0.13, P = 0.003 and P = 0.02, respectively), whereas log-transformed triglyceride content was the only independent predictor of descent time (R² = 0.09, P = 0.005).

Plethysmographic parameters and lipid tertiles. To further investigate the impact of plasma lipid/lipoprotein levels in predicting pre- and postischemic plethysmographic variables, the data were stratified into tertiles according to HDL-C, LDL-C, triglycerides, and the LDL-C-to-HDL-C ratio (Table 7). According to what was observed in the correlation analysis, the percent changes increased and peak time decreased with rising HDL-C levels. In contrast, the percent changes decreased with rising LDL-C levels and LDL-C-to-HDL-C ratios. As far as time-related variables are concerned, descent time and the reserve area significantly decreased with rising triglyceride values.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

A number of studies performed using high-resolution ultrasound methods to assess arterial compliance (measurements of finger BP and radial artery diameter, allowing the analysis of the diameter-pressure curve) have shown that compliance increases markedly after prolonged ischemia, thus allowing the study of reserve above baseline (18). In the present study, it is also clearly substantiated that with the proposed plethysmographic method, based on continuous measurements of finger BP and FPV (FPV is the result of changes at the level of large and small vessels level), postischemic compliance modifications occurring during reactive hyperemia can be detected even after only 3 min of arterial occlusion. Despite the high interindividual variability in the preischemic FAC_AUC value, all patients show, after cuff deflation, a significant increase in FAC_AUC, with percent changes ranging from 16.3 to 265.5%.

To the best of our knowledge, this is also the first study attempting to explore the kinetics of compliance modification using a continuous measurement approach, thus allowing the evaluation of the validity of noncontinuous measurement procedures often used, by others, to study compliance changes in the course of reactive hyperemia. The present findings clearly show that FAC_AUC changes during reactive hyperemia are strongly related to the time factor. Thus, according to the present results, a noncontinuous measurement procedure may be questionable. These findings, in fact, not only show that the time required to reach the peak value (peak time) is quite variable, ranging from subject to subject from 2 to 52 s, but also, and more important, that the peak value is maintained for a time period too short (<7 s in ~77% of cases) to be assessed by measurements carried out at arbitrary time intervals. After a defined interval from cuff deflation, i.e., 30 s, measured values may not be representative of the true peak but, more likely, provide a simple estimate of the ascent or descent phases of the time course curve (Fig. 7). The concept that a noncontinuous measurement procedure may be questionable is also underlined by the fact that methodological repeatability is clearly lower for time-related variables versus FAC_AUC and percent change values (Table 3).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Time course curves of 3 subjects. After a definite time interval (i.e., 30 s), the measured compliance value may correspond to the true maximal compliance change (curve A) but also to a simple value of the ascent (curve B) or descent phase (curve C).

The reasons for the elevated between-patient variability in the time course compliance parameters reported here are not clear. Variability may reflect prominent influences not controlled for, such as neural effects, humoral stimuli, or regional metabolic changes. In addition, biological variability may, of course, be partially contributory. The presence of vascular risk factors, as well as a possible influence of flow-mediated vasodilatation on the time course compliance parameters, may further contribute to the described variability. Independent of the mechanism(s) of a such striking variability, the present results strongly advocate that, to determine true maximal changes of arterial compliance during reactive hyperemia, a continuous measurement is more appropriate. A possible criticism to this conclusion might be that only 3 min of arterial occlusion are too brief a stimulus to induce maximal compliance changes. Ischemia was limited to 3 min because the subject's pain, frequently resulting from a more prolonged ischemia, was a clear limiting factor. However, in the 12 subjects in whom 3 and 12 min of arterial occlusion were compared, we observed that, although 12 min of arterial occlusion effectively produced higher values for each parameter considered, the peak time was again too variable and the time of peak maintenance too brief to allow acceptance of noncontinuous measurements. Thus, in performing 12 min of arterial occlusion, a continuous measurement procedure also seems to be essential (at least for the first 1 or 2 min after cuff deflation) to identify the true peak FAC_AUC value.

Another important reason to determine the time course curve derives from the fact that this procedure may provide, in addition to maximal FAC_AUC, other time-dependent compliance parameters potentially useful to better characterize possible endothelial dysfunction. This hypothesis is supported by previous studies (20, 34) in human peripheral conduit arteries suggesting that, whereas nitric oxide is minimally involved in regulating forearm blood flow at peak reactive hyperemia, it plays a significant role in maintaining vasodilatation during the post-peak phase. On the basis of these findings, one may assume that the marked increases in forearm blood flow during the early phases of reactive hyperemia can induce increased shear stress, which may release nitric oxide from the endothelium, thus contributing to vasodilatation and possibly compliance modulation during the mid to late phase of reactive hyperemia. The mechanisms by which the endothelium, or perhaps endothelium-derived nitric oxide, can modulate the postischemic time-dependent compliance parameters were not investigated in this study. One may, however, speculate that local factors contributing to flow-mediated vasodilatation of microvessels, such as changes in interstitial potassium and hydrogen ions, osmolality, carbon dioxide, catecholamines, prostaglandins, and adenosine (1, 5, 8, 12, 20, 23, 25), can also be involved in the regulation of the parameters considered here. Although we cannot be sure about the endothelium-dependent nature of these parameters, it is interesting to note that some of them are markedly correlated with well-known traditional risk factors for endothelial dysfunction, such as age (9, 11, 16), high cholesterol (17, 33, 37), and triglycerides (28) (Table 5); such findings were also confirmed after stratification of the studied group into tertiles for lipid variables (Table 7).

Most of the studies published so far evaluating the effects of vascular risk factors on vasoactivity after arterial occlusive stimuli have been carried out completely ignoring time factors. To the best of our knowledge, the only study (32) that evaluated the effects of a major vascular risk factor (cigarette smoking) on vasoactivity using continuous measurements showed improved measurement accuracy and an improved possibility of uncovering arterial functional abnormalities in smokers not otherwise detectable when the "time" factor is ignored. Thus vascular risk factors may modulate postischemic compliance changes, affecting not only the degree of reactive response, but also overall duration, or even both of these aspects. The present results provide further support to this hypothesis. In fact, when analyzing the cardiovascular risk factors of the examined patient series, it becomes apparent that, whereas reduced HDL-C levels may affect vasoactivity by reducing percent changes and delaying the time necessary to reach peak values, age seems to act only by affecting time-dependent parameters. Thus, according to Stadler et al. (32, 37), considering time-dependent parameters, one can appreciate the effects of vascular risk factors not otherwise detectable by considering percent changes only. The direct correlation between percent changes and HDL-C and the inverse correlation between percent changes and LDL-C or LDL-C-to-HDL-C ratios confirm, also in this nonselected group of subjects, a positive role for HDL-C (21, 22, 28) and a negative role for hypercholesterolemia on endothelial function (2, 17, 33). After adjustment of data for the other variables considered, in multiple stepwise regression analysis (Table 6), sex was the only variable independently related to the preischemic FAC_AUC thus explaining ~19% (R² = 0.19) of variability. Sex and SBP independently correlated with peak FAC_AUC and together explained ~26% (R² = 0.26) of variability. Peak time was mainly affected by age and body mass index (R² = 0.13), whereas lipid variables affected mainly the percent change, significantly and independently correlated with plasma levels of HDL-C (directly) and LDL-C (inversely), and descent time, this last mainly affected by plasma levels of triglycerides. As far as the reserve area is concerned, the correlations observed with the linear regression analysis did not reach statistical significance after data adjustment for the other variables considered. Another interesting finding that can be deduced from our results is that descent time seems to be related to the presence of pathological conditions when it is reduced, whereas the same is true for the prolongation of peak time. This suggestion is well supported not only by the inverse correlation between peak time and HDL-C, a well-known protective lipoprotein parameter (21, 22, 28), but also by the same negative correlation with percent changes of arterial compliance (0.38, P < 0.001), known to be reduced when vascular risk factors or pathological conditions are present.

All of these results strongly advocate the use of the plethysmographic method proposed here as a useful tool to measure arterial compliance in vivo. Because more extensive background information was needed to validate its ability to provide measurements applicable to clinical trials, a study was performed comparing the method proposed here with the echotracking system, a more widely accepted ultrasonic approach often applied to clinical trials (17, 19).

A high between-methods correlation coefficient was observed (r = 0.85, Fig. 4); however, by following the analytical approach proposed by Bland and Altman (4), the most appropriate method for evaluating the consistency of a new method of measurement versus an established one, it became apparent that the two methods can result in a discrepancy in arterial compliance of about 0.87 × 10². Thus, whereas the plethysmographic method may not be considered as interchangeable with the echotracking one, both provide essentially the same kind of information. Indeed, as documented in Figs. 5 and 6, although different in terms of absolute values, the trends of compliance changes during reactive hyperemia as well as those induced by sublingual GTN are essentially identical for the two methods, thus indicating that the plethysmographic method described here provides reliable technology for arterial compliance measurements applicable to clinical trials.

In conclusion, the proposed technique appears to be highly sensitive and may be considered, as an alternative to ultrasound, as a potentially useful tool to detect and monitor in vivo compliance changes induced by arterial occlusion. In addition, the study strongly suggests that the time factor should not be underestimated when postischemic variables are considered. It may, in fact, provide further time-dependent markers enabling the objective assessment of the effects of risk factors, and perhaps of risk factor modifications, on compliance changes during reactive hyperemia, a stimulus often used in pharmacological and clinical trials to investigate, in a completely noninvasive way, arterial endothelial dysfunction in humans.