OBJECTIVE— To examine the contribution of nerve-axon reflex-related vasodilation to total acetylcholine-induced vasodilation in the skin of normal and diabetic subjects.

RESEARCH DESIGN AND METHODS— The skin microcirculation was evaluated at the forearm level in 69 healthy subjects and 42 nonneuropathic diabetic patients and at the foot level in 27 healthy subjects and 101 diabetic patients (33 with neuropathy, 23 with Charcot arthropathy, 32 with peripheral vascular disease and neuropathy, and 13 without complications). Two single-point laser probes were used to measure total and neurovascular vasodilation response to the iontophoresis of 1% acetylcholine, 1% sodium nitroprusside, and deionized water.

RESULTS— The neurovascular response to acetylcholine was significantly higher than the response to sodium nitroprusside and deionized water (P < 0.01). At the forearm level, the contribution of neurovascular response to the total response to acetylcholine was 35% in diabetic patients and 31% in control subjects. At the foot level, the contribution was 29% in diabetic patients without neuropathy and 36% in control subjects, while it was significantly diminished in the three neuropathic groups. A significantly lower nonspecific nerve-axon-related vasodilation was observed during the iontophoresis of sodium nitroprusside,which does not specifically stimulate the C nociceptive fibers.

CONCLUSIONS— Neurovascular vasodilation accounts for approximately one-third of the total acetylcholine-induced vasodilation at both the forearm and foot levels. The presence of diabetic neuropathy results in reduction of both the total vasodilatory response to acetylcholine and the percentage contribution of neurovascular vasodilation to the total response. Acetylcholine and sodium nitroprusside cause vasodilation in the skin microcirculation through different pathways.

Functional abnormalities of the microcirculation have gained significant attention in recent years for their potential pathogenic role in the development of diabetic complications, particularly diabetic neuropathy and diabetic foot problems(1,2,3,4,5). The microvascular tone is regulated by several humoral and neural factors. The vascular endothelium has an important role in controlling the microvascular tone by releasing several vasodilator substances such as nitric oxide,prostacyclin and endothelium-derived hyperpolarizing factor, and vasoconstrictor substances such as prostaglandins and endothelin(6). Nitric oxide is the most important vasodilator substance responsible for endothelium-dependent vasodilation. After its secretion from the endothelium, it diffuses to the adjacent smooth muscle cells and stimulates the guanylate cyclase enzyme to produce cyclic guanosine 3′,5′-monophosphate, which, in turn,leads to smooth muscle relaxation and vasodilation(7).

The normal neurovascular response conducted through the C nociceptive nerve fibers is another important mechanism for the regulation of the microcirculation. Stimulation of the C nociceptive nerve fibers leads to antidromic stimulation of the adjacent C fibers, which secrete vasodilating substances such as substance P and bradykinin, causing vasodilation at the injured or inflamed skin areas. This vasodilating response, also known as the Lewis triple-flare response, is decreased in the presence of diabetic neuropathy. Reduction of local blood flow increases the vulnerability of the neuropathic limb to severe diabetic foot problems(8,9). It has been postulated that the abnormality in the neurovascular response in the neuropathic limb further aggravates the abnormalities in the microcirculation, and a vicious cycle may ensue(10).

Several recent studies(10,11,12,13)have demonstrated reduced endothelium-dependent vasodilation in patients with either type 1 or type 2 diabetes. However, little information is available regarding the contribution of nerve-axon reflex-related vasodilatation to maximal skin vasodilation in such patients(8,9). The recent development of noninvasive techniques that can reliably quantify blood flow in the skin microcirculation has made it possible to study changes in microvascular function in patients with diabetes(14,15). In the present study, we have examined the contributing role of the nerve-axon reflex-related vasodilation response to the total skin vasodilation at both the forearm and foot levels of neuropathic and nonneuropathic diabetic patients.

Patients

We studied the skin microcirculation at the forearm level in 69 healthy subjects and 42 nonneuropathic diabetic subjects. The following exclusion criteria were applied to subjects in all groups: smoking any amount of cigarettes during the previous 6 months, subjects with diagnosed cardio-vascular disease (coronary artery disease, arrhythmia, congestive heart failure), stroke or transient ischemic attack, peripheral vascular disease(symptoms of claudication and/or absence of peripheral pulses), chronic renal disease, severe dyslipidemia (triglycerides >600 mg/dl or cholesterol>300 mg/dl), or any other serious chronic disease requiring active treatment. Subjects were also excluded if they were on any of the following medications: any type of antihypertensive drugs, lipid-lowering agents,glucocorticoids, antineo-plastic agents, psychoactive agents, or bronchodilators. In addition, diabetic patients with proliferative retinopathy, peripheral somatic neuropathy, macroalbuminuria (expressed as albumin-to-creatinine ratio >300 μg/mg), and/or on insulin or troglitazone were excluded from the study.

We also evaluated the skin microvascular reactivity at the foot level in 27 healthy subjects and 101 diabetic patients who were divided into four groups. The first group consisted of 33 diabetic neuropathic patients with a history of foot ulceration but no peripheral vascular disease, the second group of 23 diabetic patients with Charcot arthropathy, the third group of 32 diabetic patients with peripheral vascular disease and neuropathy, and the fourth group of 13 diabetic patients without any complications.

All healthy subjects were free of any illness and did not take any medications. Special emphasis was given to exclude anyone with a history of hypertension, diabetes, hypercholesterolemia, active tobacco use, history of any systemic illness, or the use of any antihypertensive, cardiac, or hormonal medication. Patients with either type 1 or type 2 diabetes were included. Patients with nephropathy (creatinine >2 mg/l), severe heart failure, or any other serious illness were excluded from the study.

Further details of the characteristics of the study population are shown in Tables 1 and 2. The study was approved by the institutional review board, and consent was obtained from all participants.

Methods

A history, physical examination, and fasting plasma glucose measurement were performed on all patients. Diabetic neuropathy was diagnosed according to the San Antonio Consensus Statement criteria(16). The symptoms were evaluated by using a neuropathy symptom score, and the clinical signs were evaluated by using a neuropathy disability score(17). Quantitative sensory testing included the assessment of vibration perception threshold using a Biothesiometer and cutaneous perception threshold using Semmes-Weinstein monofilaments(18,19). The diagnosis of Charcot neuroarthropathy was made when gross destruction of the joints of the mid-foot that resulted in significant foot deformity was present. Patients were characterized as having peripheral vascular disease based on the presence of one or more of the following clinical features:claudication, absent foot pulses, and/or abnormal invasive and abnormal noninvasive vascular tests.

Each participant was studied after a 20-min acclimatization period in a warm environment (room temperature 23-24°C). We used two single-point laser probes and a DRT4 Laser Doppler Blood Flow Monitor (Moor Instruments,Millwey, Devon, U.K.) to evaluate the skin microcirculation. Forearm microcirculatory flow was measured over the flexor surface of the forearm, and foot microcirculatory flow was measured over the dorsum of the foot. The blood flow response was measured in response to ion-tophoresis of each of three substances: 1% acetylcholine chloride solution (a substance that elicits both a neurovascular response and endothelium-dependent vasodilation), 1% sodium nitroprusside (a substance that does not elicit a neurovascular response, but induces endothelium-independent vasodilation), and deionized water (used as a control during iontophoresis to measure the vasodilation caused by the direct effect of a constant current flow)(10). Deionized water was iontophoresed in both anodal mode (the same mode in which the iontophoresis of acetylcholine is performed) and cathodal mode (the mode that the iontophoresis of sodium nitroprusside is performed). The difference in these two modes is the polarity of the iontophoresis chamber: the chamber serves as the anode for iontophoresis of acetylcholine, which has a negative electrical charge, and as the cathode for the iontophoresis of sodium nitroprusside, which has a positive electrical charge. Therefore, the constant current has an opposite direction when the polarity of the chamber is changed.

The iontophoresis instrument (MIC1 iontophoresis system; Moor Instruments)consists of an iontophoresis delivery vehicle device that sticks firmly to the skin with the help of adhesive tape. The device contains two chambers that accommodate two single-point laser probes. One probe is placed within the chamber containing the iontophoresis solution (thus measuring the direct response to acetylcholine or sodium nitroprusside iontophoresis), while the second probe is placed outside but within proximity (within 5 mm) to the iontophoresis solution chamber, thus measuring the indirect nerve-axon-related response that results from stimulation of the C nociceptive nerve fibers. A small amount (<1 ml) of test solution was applied to the iontophoresis chamber. Subsequently, a constant current of 200 μA for 60 s was applied,achieving a dose of 6 mC/cm2 between the iontophoresis chamber and a second nonactive electrode placed 10-15 cm proximal to the chamber. The two laser probes recorded changes in skin blood flow. Measurements were obtained for 40 s before the iontophoresis, during the iontophoresis, and 90 s after it(10,11). The day-to-day reproducibility of the technique was evaluated in five healthy subjects (four men and one woman, ages 23-39 years) who were repeatedly tested at their foot and forearm for 10 consecutive working days. With use of a single-point laser probe, the coefficient of variation (CV) for the baseline blood flow before iontophoresis of acetylcholine was 60.6% and for the maximal hyperemic response was 35.2% after the iontophoresis of acetylcholine.

Statistical analysis

The results were recorded and tabulated before revealing the patient category assignations. Changes in microvascular blood flow were expressed as the percentage of increase over baseline, where median, first quartile, and third quartile values are used for comparisons. Parametric data were expressed as means ± SD. Statistical analysis was performed using the Minitab computer software (State College, PA), using both parametric and nonparametric tests. All tests were two-tailed, with significance taken as P<0.05. For between-group comparisons, we used paired ttest for parametric data and Kruskal-Wallis test for nonparametric data.

Forearm level

The results of the iontophoresis are shown in Table 3. To evaluate the degree of vasodilation that is specific to the neurovascular response, we measured the capillary blood flow in a skin area in direct contact with acetylcholine and in an adjacent skin area not in direct contact with it. The latter represents the nerve-axon-related portion of the total response. The percentage contribution of the nerve-axon-related response to the total response was similar between nonneuropathic diabetic patients and the control group after the iontophoresis of acetylcholine (35 and 31%, respectively, NS). In both the nonneuropathic diabetic patients and control group, the percentage contribution of the nerve-axon-related response to the total response was significantly less after the iontophoresis of either sodium nitroprusside (13 and 10%, respectively, P < 0.01) or deionized water (16 and 17%,respectively, P < 0.01). No significant difference was seen between the percentage contribution of the nerve-axon-related reflex to the total response to sodium nitroprusside and to deionized water both in anodal and cathodal mode in both the nonneuropathic diabetic patients and control group (NS). This is consistent with the fact that acetylcholine specifically stimulates C nociceptive fibers and the nerve-axon-related reflex, whereas sodium nitroprusside and deionized water do not. The contribution of the neurovascular response to the total response to acetyl-choline is approximately one-third of the total response and is not compromised by diabetes at the forearm level.

Foot level

The results of the iontophoresis are shown in Table 4. In response to iontophoresis of acetylcholine, the percentage contribution of the nerve-axon-related response was similar to that seen at the forearm level in both the diabetic patients without complications and the healthy control subjects (29 and 36%, respectively, NS). The diabetic neuropathic patients had a significantly lower median increase of capillary blood flow over baseline in response to acetylcholine compared with the diabetic patients without complications and the control group (P < 0.01)(Fig. 1). The neurovascular response was markedly decreased in all three neuropathic groups when compared with diabetic patients without complications and the control group. The contribution of the nerve-axon-related response to the total response was 8%in diabetic patients with neuropathy (P < 0.01), 5% in diabetic patients with Charcot arthropathy (P<0.01), and 20% in diabetic patients with neuropathy and peripheral vascular disease (P <0.01). The nerve-axon-related response to sodium nitroprusside and to the anodal and cathodal iontophoresis of deionized water was similar to the response observed in the upper extremity.

Figure 1

Total and neurovascular (N) change in skin blood flow in response to acetylcholine at the foot level. The median, first quartile, and third quartile and the range are shown. The total response is significantly lower in neuropathic diabetic patients than it is in control subjects and diabetic patients without neuropathy (P < 0.01). The percentage contribution of neurovascular response to the total response is also significantly lower in neuropathic diabetic patients than in control subjects and diabetic patients without neuropathy (P <0.01).

Figure 1

Total and neurovascular (N) change in skin blood flow in response to acetylcholine at the foot level. The median, first quartile, and third quartile and the range are shown. The total response is significantly lower in neuropathic diabetic patients than it is in control subjects and diabetic patients without neuropathy (P < 0.01). The percentage contribution of neurovascular response to the total response is also significantly lower in neuropathic diabetic patients than in control subjects and diabetic patients without neuropathy (P <0.01).

Close modal

In the present study, we have shown that in healthy subjects and in nonneuropathic diabetic patients, at both the forearm and foot levels, the microvascular vasodilation that is related to the neurovascular response accounts for approximately one-third of the total vasodilation that is observed after the iontophoresis of acetylcholine. This portion is markedly decreased in the presence of diabetic neuropathy.

The total microvascular vasodilation in response to acetylcholine is currently considered to represent the sum of direct stimulation of the endothelium by acetylcholine and of the vasodilation that is related to the nerve-axon reflex (20). However, the magnitude of the contribution of each of these two factors to the total vasodilation has not been adequately studied, and the currently available data are conflicting. Thus, although some studies have suggested a considerably higher contribution of the neurovascular response, the techniques used did not allow the precise quantification of this contribution(8,21). On the other hand, another study has shown that local sensory inhibition by topical application of lignocaine and prilocaine did not have an effect on the total vasodilatory response to the iontophoresis of acetylcholine(22). The main problem in interpreting these data, though, lies in the fact that there is no evidence that topical application of lignocaine abolishes the nerve-axon reflex, since it may cause local anesthesia via mechanisms that are not affecting the antidromic stimulation of local C nociceptive fibers. This is further emphasized by the findings of a previous study that showed that deep subcutaneous injection of lignocaine does inhibit the nerve-axon-related vasodilation in response to the iontophoresis of acetylcholine(23).

In the present study, we have used a chamber that can accommodate two single-point laser probes that can measure the total and the nerve-axon reflex-related vasodilation. This technique can satisfactorily measure the two responses separately, making it possible to evaluate the relative contribution of the neurovascular response to the total response with an adequate reliability. Furthermore, we have studied subjects with and without peripheral neuropathy rather than testing with local anesthesia, which, as mentioned previously, has questionable effects on the nerve-axon reflex. Finally, it should be remembered that under conditions of stress (such as injury or inflammation), hyperemia is necessary not only in the injured area alone but in a considerably larger area that surrounds the injured site. Because this response depends mainly on a normal nerve-axon reflex, our findings make the point that this response, under normal conditions, is one-third of the maximal achievable vasodilation and that this is drastically reduced in the presence of diabetic neuropathy.

Single-point laser probe measurements are known to have a considerably high CV, whereas the use of laser scanners reduces this variability(10,16). However, with laser scanners, one cannot evaluate the nerve-axon response, a measurement that can be done only with use of the single-point laser technique. The large number of subjects in each studied group compensates for the high variability and does not affect the validity of the conclusions regarding the contribution of nerve-axon response to total vasodilation. On the other hand, the high variability does not allow the direct comparison of the vasodilatory response among the various studied groups, which makes this study prone to type 2 statistical error. Therefore, it is recommended that for reliable data regarding these questions, the reader is directed to studies that have specifically addressed this question and used the appropriate techniques, including the use of a Laser Scanner Imager(10,11,24,25,26).

In contrast to acetylcholine, sodium nitroprusside causes vasodilation by directly stimulating the vascular smooth muscle cell and does not specifically stimulate the C nociceptive fibers. This result can be seen in the present study by the small nerve-axon-related vasodilation achieved with sodium nitroprusside, similar to that achieved by deionized water, which can be attributed to a nonspecific galvanic effect of the constant current that is used for the iontophoresis(27). Therefore, we believe that the presented data also provide further evidence of different pathways through which acetylcholine and sodium nitroprusside cause vasodilation in the skin microcirculation.

The iontophoresis of deionized water in the same polarity with that of acetylcholine (i.e., with an anodal constant current) has been previously shown to lead to a small nonspecific galvanic effect(20,21). In contrast, iontophoresis with a cathodal current, as used for the iontophoresis of sodium nitroprusside, has been reported in one study to result in a significant nonspecific vasodilatory response(22). In the present study, we have not found such an exaggerated response, and both anodal and cathodal modes elicited very similar responses. The main differences between previous studies and the present study that may explain this discrepancy are the duration and amplitude of the current used for iontophoresis. Thus, in our unit, we apply 200 μA for 60 s. This produces maximal specific vasodilation with a minimal nonspecific vasodilation. This is in sharp contrast with a previous study in which three rather small pulses of iontophoresis were performed over a period of 10 min, raising the question as to whether maximal vasodilation was achieved.

In a previous study, we showed that diabetes impairs the total endothelium-dependent and endothelium-independent vasodilation at the forearm level, a skin area that is rarely affected by diabetic neuropathy(11). In addition, the present study shows that this reduction is independent of the nerve-axon-related response. A direct effect of diabetes on endothelium function or smooth muscle cells should therefore be considered as the main cause of the observed impaired vasodilation in response to acetylcholine and sodium nitroprusside. We have previously shown that differences exist between the forearm and foot microcirculation beds, with the foot vasodilatory response being approximately half that of the response at the forearm level(11). Similar results were observed in the present study.

Neuropathy has been shown to reduce the vasodilatory response at the foot level, irrespectively of the presence or absence of peripheral vascular disease(10,11). In the present study, the nerve-axon-related response in diabetic patients during specific stimulation of the C fibers with acetylcholine was markedly decreased, being similar to that observed with sodium nitroprusside. Thus,this is another indication that neuropathy renders the diabetic foot functionally ischemic, as blood flow fails to increase under conditions of stress.

In summary, we have shown in the present study that the neurovascular vasodilation response accounts for approximately one-third of the total acetylcholine-induced vasodilation response at both the forearm and foot levels of healthy subjects and non-neuropathic diabetic patients. The presence of diabetic neuropathy at the lower extremity results in a significant reduction in the total vasodilatory response to acetylcholine and to an even more pronounced reduction in the percentage contribution of the neurovascular response to the total skin vasodilatory response to acetylcholine. Acetylcholine and sodium nitroprusside cause vasodilation in the skin microcirculation through different pathways. Finally, the technique used in this study may be particularly helpful in developing new methods that can objectively evaluate the efficacy of new treatments on small-fiber function.

Abbreviations: CV, coefficient of variation.

A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

1.
Malik RA, Tesfaye S, Thompson SD, Veves A, Sharma AK, Boulton AJM,Ward JD: Endothelial localization of microvascular damage in human diabetic neuropathy.
Diabetologia
36
:
454
-459,
1993
2.
Tesfaye S, Harris N, Jakubowski JJ, Mody C, Wilson RM, Rennie IG,Ward JD: Impaired blood flow and arterio-venous shunting in human diabetic neuropathy: a novel technique of nerve photography and fluorescein angiography.
Diabetologia
36
:
1226
-1274,
1993
3.
Stevens MJ, Dananberg J, Feldman EL, Lattmir SA, Kamijo M, Thomas TP, Shindo H, Sima AA, Greene DA: The linked roles of nitric oxide, aldose reductase and (Na+, K+)-ATPase in the slowing of nerve conduction in the streptozotocin diabetic rat.
J Clin Invest
64
:
853
-919,
1994
4.
Stevens MJ, Feldman EL, Greene DA: The etiology of diabetic neuropathy: the combined roles of metabolic and vascular defects.
Diabet Med
12
:
566
-579,
1995
5.
Tesfaye S, Malik R, Ward JD: Vascular factors in diabetic neuropathy.
Diabetologia
37
:
847
-854,
1994
6.
Vane JR, Anggard EE, Botting RM: Mechanisms of disease: regulatory functions of endothelium.
N Engl J Med
323
:
27
-36,
1990
7.
Palmer RMJ, Ashton DS, Moncada S: Vascular endothelial cells synthesize nitric oxide from L-arginine.
Nature
333
:
664
-666,
1988
8.
Parkhouse N, LeQuesne PM: Impaired neurogenic vascular response in patients with diabetes and neuropathic foot lesions.
N Engl J Med
318
:
1306
-1309,
1988
9.
Walmsley D, Wiles PG: Early loss of neurogenic inflammation in the human diabetic foot.
Clin Sci
80
:
605
-610,
1991
10.
Veves A, Akbari CA, Primavera J, Donaghue VM, Zacharoulis D, Chrzan JS, DeGirolami U, LoGerfo FW, Freeman R: Endothelial dysfunction and the expression of endothelial nitric oxide synthetase in diabetic neuropathy,vascular disease, and foot ulceration.
Diabetes
47
:
457
-463,
1998
11.
Arora S, Smakowski P, Frykberg RG, Simeone LS, Freeman R, LoGerfo FW, Veves A: Differences in foot and forearm skin microcirculation in diabetic patients with and without neuropathy.
Diabetes Care
21
:
1339
-1344,
1998
12.
Johnstone MT, Creager SJ, Scales KM, Casco JA, Lee BK, Creager MA:Impaired endothelium-dependent vasodilation in patients with insulin-dependent diabetes mellitus.
Circulation
88
:
2510
-2516,
1993
13.
Williams SB, Cusco JA, Roddy M, Johnstone MT, Creager MA: Impaired nitric oxide-mediated vasodilatation in patients with non-insulin-dependent diabetes mellitus.
J Am Coll Cardiol
27
:
567
-574,
1996
14.
Celermajer DS, Sorensen KE, Gooch VM, Spiegelhalter DJ, Miller OI,Sullivan ID, Lloyd JK, Deanfield JE: Non-invasive detection of endothelial dysfunction in children and adults at risk of atherosclerosis.
Lancet
340
:
1111
-1115,
1992
15.
Tooke JE: Methodologies used in the study of the microcirculation in diabetes mellitus.
Diabetes Metab Rev
9
:
57
-70,
1993
16.
American Diabetes Association: Report and recommendations of the San Antonio Conference on Diabetic Neuropathy (Consensus Statement).
Diabetes
37
:
1000
-1004,
1988
17.
Veves A, Uccioli L, Manes C, Van Acker K, Komninou H, Philippides P, Kat-Silambros N, De Leeuw I, Menzinger G, Boulton AJM: Comparisons of risk factors for foot problems in diabetic patients attending teaching hospitals'out-patient clinics in four different European states.
Diabet Med
11
:
709
-713,
1994
18.
Wiles PG, Pearce SM, Rice PJS, Mitchell JMO: Vibration perception threshold: influence of age, height, sex, and smoking and calculation of accurate centile values.
Diabet Med
8
:
157
-161,
1991
19.
Kumar S, Fernando DJS, Veves A, Knowles EA, Young MJ, Boulton AJM:Semmes-Weinstein monofilaments: a simple, effective and inexpensive screening device for identifying diabetic patients at risk of foot ulceration.
Diabetes Res Clin Pract
13
:
63
-67,
1991
20.
Pham H, Economides PA, Veves A: The role of endothelial function on the foot microcirculation and wound healing in diabetic patients.
Clin Podiatr Med Surg
15
:
85
-94,
1998
21.
Forst T, Pfutzner A, Kunt T, Pohlmann T, Schenk U, Bauersachs R,Kustner E, Beyer J: Skin microcirculation in patients with type I diabetes with and without neuropathy after neurovascular stimulation.
Clin Sci
94
:
255
-261,
1998
22.
Morris SJ, Shore AC: Skin blood flow responses to the iontophoresis of acetylcholine and sodium nitroprusside in man: possible mechanisms.
J Physiol
496
:
531
-542,
1996
23.
Parkhouse N, Le Quesne PM: Quantitative objective assessment of peripheral nociceptive C fibre function.
J Neurol Neurosurg Psychiatry
51
:
28
-34,
1988
24.
Morris SJ, Shore AC, Tooke JE: Responses of the skin microcirculation to acetylcholine and sodium nitroprusside in patients with NIDDM.
Diabetologia
38
:
1337
-1344,
1995
25.
Caballero AE, Arora S, Saouaf R, Lim SC, Smakowski P, Park JY, King GL, LoGerfo FW, Horton ES, Veves A: Microvascular and macrovascular reactivity is reduced in subjects at risk for type 2 diabetes.
Diabetes
48
:
1856
-1862,
1999
26.
Lim SC, Caballero AE, Arora S, Smakowski P, Bashoft E, Brown F,LoGerfo FW, Horton ES, Veves A: The effect of gender and hormonal replacement therapy on the vascular reactivity of healthy individuals and individuals with type 2 diabetes.
J Clin Endocrinol Metab
84
:
4159
-4164,
1999
27.
Noon JP, Walker BR, Hand MF, Webb DJ: Studies with iontophoretic administration of drugs to human dermal vessels in vivo: cholinergic vasodilatation is mediated by dilator prostanoids rather than nitric oxide.
Br J Clin Pharmacol
45
:
545
-550,
1998