INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Polycythemia occurs in the course of hypoxemic lung disease to maintain oxygen delivery in the face of impaired gas exchange. Although widespread use of supplemental oxygen therapy has decreased the severity of polycythemia in patients with chronic hypoxemic lung disease (CHLD), an increase in hematocrit (Hct) level remains a common finding in this setting, especially when sleep-disordered breathing is also present. Convincing evidence exists that a high hematocrit level produces adverse effects, including an increase in pulmonary vascular resistance (PVR) and decreases in blood flow and oxygen supply to the brain and to other organs. Several studies have shown that in polycythemic patients, moderate reduction in Hct decreases systemic and pulmonary arterial pressures, enhances cardiac output and oxygen delivery, and improves exercise tolerance (1). Moreover, cerebral blood flow, limb blood flow, and limb oxygen transport can be improved dramatically by reduction of Hct to normal (2, 3).

The most often cited mechanism by which polycythemia contributes to increased resistance to blood flow involves an increase in blood viscosity (4). However, vasoconstriction may also play a role. Basal release of endothelium-derived nitric oxide (NO) is determined by vessel-wall shear stress (5, 6). Because blood viscosity is an important component of shear stress exerted on the vascular wall, the increased blood viscosity induced by polycythemia may enhance basal NO release through activation of endothelial NO synthase (NOS). On the other hand, hemoglobin (Hb) is a potent inhibitor of endothelium-derived NO, and several observations suggest that Hb in intact red blood cells is as effective as free Hb in blocking endothelium-dependent relaxation (7, 8). In patients with polycythemia, a reasonable hypothesis is that a high circulating Hb level may increase vascular resistance through inactivation of endothelium-derived NO. Although some experimental and human studies point to a key role for Hb in modulating NO-mediated vasodilation (9), this role has not been examined in patients with polycythemia. In previous studies using acetylcholine (ACh) as a screening agent for pulmonary vasodilation in patients with CHLD, we observed that the extent of ACh-induced vasodilation varied with the Hct level (12). We therefore designed the present study to assess the specific effects of polycythemia on endothelium-dependent relaxation of hemodynamics in patients with CHLD. In the first part of this study, we investigated the vasodilator response to intravenous ACh in a large series of patients, classified as normocythemic or polycythemic according to whether their Hb level was less or more than 15.5 g/dl. In a subgroup of polycythemic patients defined by use of isovolemic hemodilution, we examined the response to ACh infusion before and after hemodilution. Because inhaled NO is a direct-acting pulmonary vasodilator (13), we also examined the hemodynamic response to NO inhalation in the normocythemic and polycythemic patients.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

Twenty-one patients (mean age: 59 yr; range: 38 to 72 yr) were studied. Most of the patients had chronic obstructive lung disease (COLD) as established from a previous history of chronic bronchitis and evidence of chronic airflow limitation on standard pulmonary function tests (Table 1). Several patients had predominantly restrictive lung disease, and one patient had morbid obesity. All patients experienced dyspnea on exertion but were in a stable phase of their disease as defined by a lack of hospitalization during the 2 mo preceding the study. The absence of electrocardiographic abnormalities suggesting ischemic left heart disease, and the absence of echocardiographic left ventricular dysfunction were verified in each patient prior to the study. Several patients had a previous history of right heart failure, but none had peripheral edema at the time of the study. Patients were treated with oral and inhaled bronchodilators but took no medications during the period immediately before the measurements made in the study. Four patients were studied while receiving supplemental oxygen. None of these patients received steroids or nonsteroidal antiinflammatory drugs. The study was approved by the ethics committee of our institution, and informed consent was obtained from each patient before the study.

Patients were then classified into two groups based on whether their Hb levels were less or more than 15.5 g/dl. Physiologic parameters recorded in the normocythemic (n = 10) and polycythemic (n = 11) patients are summarized in Table 1. Results obtained in eight of the normocythemic and three of the polycythemic patients have been published elsewhere (12).

Hemodynamic Studies during Right-heart Catheterization

All patients were studied while in the supine position during right-heart catheterization. A size 7 French Swan-Ganz thermodilution catheter was inserted into the femoral vein and placed in the pulmonary artery. An arterial line was inserted into a homolateral femoral artery to record systemic arterial pressure (Psa) and to sample arterial blood (Edwards Laboratories, Inc., Santa Ana, CA). Systemic and pulmonary artery pressures were measured with Gould P50 pressure transducers, linked to pressure modules and to Gould TA 2000 multichannel recorder (Gould Electronics, Ballainvilliers, France). All transducers were referenced to the mid-chest level with patients in the supine, semirecumbent position. Mean Psa and Ppa were obtained by electronic integration. Pulmonary arterial wedge (Ppaw) and right arterial pressure (RAP) were measured at end expiration over at least three respiratory cycles. Cardiac output was determined by the thermal dilution method, using a bedside cardiac-output computer (Edwards 9520-A), and each cardiac-output value was taken as the mean of at least three determinations. Heart rate (HR) was measured from a continuously recorded electrocardiogram (ECG) lead. Derived hemodynamic variables were calculated according to standard formulas: cardiac index (CI) = cardiac output/body surface area L/min/m²); systemic vascular resistance index (SVR) = Psa  RAP/CI (mm Hg/L/ min/m²); pulmonary vascular resistance index (PVR) = Ppa  Ppaw/ CI (mm Hg/L/min/m²). Steady-state conditions during catheterization were defined by the absence of more than 10% changes in Ppa and HR within at least 30 min before the infusion of ACh.

Measurements of Gas Exchange

Simultaneous arterial and mixed venous blood samples were collected in heparinized syringes for determination of blood-gas tensions and pH (ABL 30; Radiometer, Copenhagen, Denmark). Total Hb, methemoglobin and Hb oxygen saturation were measured spectrophotometrically, using an OSM 3 hemoximeter (Radiometer). Venous admixture (Qs/Qt, percentage of total blood flow) was calculated with the equation of Berggreen as capillary O₂ content minus arterial O₂ content/capillary O₂ content minus mixed venous O₂ content. Oxygen consumption (ml/min/m²) was calculated as the product of CI and the difference between arterial and mixed venous O₂ contents. Systemic oxygen transport (SOT) was obtained as the product of CI and arterial oxygen content (ml/min/m²).

Protocol

Infusion of ACh. After baseline measurements, ACh (Assistance Publique, Paris) was infused during three sequential 10-min periods at incremental rates of 5, 10, and 15 mg/min (infusion pump; Vial-Medical, Grenoble, France). Hemodynamic parameters were continuously monitored. Blood samples for gasometric measurements were drawn prior to ACh infusion and at the end of each 10-min infusion period.

Inhalation of NO

The procedure for NO administration used in the present study was similar to that previously described in patients with COLD (12). Patients breathed through a face mask connected to a nonrebreathing circuit consisting of a 5-L reservoir bag and a one-way valve to separate inspired from expired gas. Expired gas was scavenged and discarded. The residence half-time of NO in the gas reservoir was less than 20 s, with a fresh gas flow of 18 L/min. The NO/N₂ mixture was delivered into the mixing reservoir bag at increasing flows to obtain inspired NO concentrations of 40 ppm in the inspiratory circuit. Precise flows of the NO/N₂ mixture were achieved by using a flowmeter that was volumetrically calibrated prior to the study. Fractional concentration of O₂ in the inspiratory circuit was kept constant by using an O₂ blender and by adjusting the FI_O2 in the gas mixture delivered to the reservoir bag. Continuous measurement of the fractional concentration of O₂ in inspired gas was obtained with an oxygen analyzer (Servomex, Crowborough, UK).

NO (stored as 300 ppm in pure N₂) was obtained from the Compagnie Française des Produits Oxygénés (Meudon La Forêt, France). Chemiluminescence analysis of this tank revealed less than 5 ppm nitrogen dioxide (NO₂). During the studies, NO and NO₂ concentrations of the breathing mixtures were assessed through chemiluminescence analysis of NO/NO₂ (Topaze 2020; Cosma SA, Igny, France).

Hemodilution

In six polycythemic patients (Patients 6 to 11), the response to ACh infusion was studied prior to and after isovolemic hemodilution. In this subgroup of patients, stepwise isovolemic hemodilution was performed by substituting saline for blood drawn by phlebotomy in steps of 300 ml. Hemodilution was stopped when the Hb concentration fell below 15.5 g/100 ml from 18.0 ± 2 to 15.1 ± 2 g/100 ml. Thirty minutes after the last hemodilution step, hemodynamic measurements were performed and a further infusion of ACh was given.

Statistical Analysis

All data are presented as means ± SEM. Between-group comparisons of baseline parameters were made with the unpaired t test. Two-way analysis of variance (ANOVA) was used to evaluate the effects of the ACh infusion and NO inhalation in the normocythemic and polycythemic groups. Because ANOVA indicated significant differences between groups, these were compared through a modified t test in which the t value was computed from the residual value. Two-way ANOVA for repeated measures was used to evaluate the effects of hemodilution and ACh infusion, and the interaction of both. When the interaction was significant, pairwise comparison was done using Wilcoxon's paired test; values of p < 0.05 were considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hemodynamic Responses to Intravenous ACh and NO Inhalation in Normocythemic and Polycythemic Patients

Hemodynamic and gas-exchange parameters recorded during the control phase prior to infusion of ACh and shown in Table 1 and Figure 1. Polycythemic and normocythemic patients did not differ with respect to Ppa, Psa, PVR, SVR, and CI. Arterial PO₂ and pulmonary venous admixture did not differ between the two groups.

View larger version (20K): [in this window] [in a new window]   Figure 1.   Effects of intravenous infusions of 5, 10, and 15 mg/min ACh on Ppa, PVR, SVR, CI, PaO2, and venous admixture in 10 normocythemic and 11 polycythemic patients. Results are expressed as means ± SEM. *p < 0.05; **p < 0.01 versus control values. There were no differences between groups at baseline.

In the normocythemic group (Hb = 13.6 ± 0.3 g/100 ml), the overall effects of ACh consisted of pulmonary and systemic vasodilation and dose-dependent impairment of gas exchange (Table 2). ACh at the highest infusion rate decreased Ppa, from 30 ± 2 mm Hg to 26 ± 2 mm Hg (p < 0.01); PVR, from 5.1 ± 0.4 U/m² to 3.4 ± 0.3 U/m² (p < 0.001); Psa, from 111 ± 4 mm Hg to 108 ± 4 mm Hg (p < 0.05); and SVR, from 27 ± 2 U/m² to 22 ± 2 U/m², and also increased CI, from 3.8 ± 0.2 L/min/m² to 4.7 ± 0.3 L/min/m² (p < 0.001). Pa_O2 fell from 59 ± 3 mm Hg to 48 ± 3 mm Hg (p < 0.001) and Qs/Qt rose from 32 ± 4% to 44 ± 4% (p < 0.01). VO₂ remained unchanged (180 ± 8 ml/min/m² versus 165 ± 8 ml/min/m², p = NS), whereas SOT increased slightly, from 62 ± 4 ml/min/m² to 69 ± 4 ml/min/m² (p < 0.05). In contrast, patients with polycythemia (Hb = 17.7 ± 0.5 g/100 ml) did not show a change in their Ppa (34 ± 1 mm Hg versus 33 ± 1 mm Hg, p = NS), PVR (6.7 ± 0.9 U/m² versus 6.9 ± 0.8 U/m², p = NS), or Psa, but their SVR decreased, from 28 ± 2 U/m² to 26 ± 2 U/m² (p < 0.01) and their CI increased slightly, from 3.6 ± 0.3 L/min/m² to 3.9 ± 0.3 L/min/m² (p < 0.01). Although Qs/Qt increased slightly from 40 ± 4% to 45 ± 3% (p < 0.05), Pa_O2 remained unchanged (50 ± 2 versus 52 ± 2 mm Hg, p = NS), as did VO₂ (164 ± 6 ml/min/m² versus 167 ± 6 ml/min/m², p = NS) and SOT (78 ± 6 ml/min/m² versus 75 ± 6 ml/min/m², p = NS). Thus, differences were seen between normocythemic and polycythemic patients with respect to ACh-induced changes in Ppa, PVR, SVR, CI, and venous admixture. In the overall study population, the Hb concentration correlated negatively with Ppa (r = 0.46, p < 0.05), PVR (r = 0.57, p < 0.01), SVR (r = 0.47, p < 0.05), CI (r = 0.50, p < 0.05), Pa_O2 (r = 0.79, p < 0.01), and Qs/Qt (r = 0.79, p < 0.01) (Figure 2). In response to NO inhalation, normocythemic and polycythemic patients showed decreases in Ppa, from 31 ± 2 mm Hg to 25 ± 2 mm Hg (p < 0.01) and from 35 ± 1 mm Hg to 30 ± 1 mm Hg (p < 0.01), respectively, and in PVR, from 4.9 ± 0.4 U/m² to 3.3 ± 0.3 U/m² (p < 0.01) and from 6.7 ± 1.2 U/m² to 5.1 ± 0.9 U/m² (p < 0.01), respectively, with no significant changes in CI. There was no difference between normocythemic and polycythemic patients with respect to changes in Ppa (20 ± 5% versus 14 ± 7%, p = NS) and PVR (31 ± 12% versus 24 ± 7%, p = NS) in response to NO inhalation (Figure 3). Two-way ANOVA was not significant when testing the effects of NO inhalation on gas-exchange parameters. In the normocythemic group, inhaled NO did not alter Pa_O2 (62 ± 3 mm Hg versus 59 ± 4 mm Hg, p = NS), Qs/Qt (31 ± 3% versus 35 ± 4%, p = NS), VO₂ (183 ± 9 ml/min/m² versus 167 ± 7 ml/ min/m², p = NS), or SOT (71 ± 4 ml/min/m² versus 67 ± 4 ml/ min/m², p = NS). Similarly, in the polycythemic group, there was no change in Pa_O2 (55 ± 3 mm Hg versus 55 ± 4 mm Hg, p = NS), Qs/Qt (37 ± 3% versus 37 ± 3%, p = NS), VO₂ (167 ± 9 ml/min/m² versus 161 ± 8 ml/min/m², p = NS), or SOT (82 ± 7 ml/min/m² versus 78 ± 7 ml/min/m², p = NS) in response to NO inhalation.

View larger version (24K): [in this window] [in a new window]   Figure 2.   The relationship between Hb concentration and changes in PaO2, venous admixture, CI, SVR, Ppa, and PVR induced by ACh given intravenously at 15 mg/min in 21 patients.

View larger version (10K): [in this window] [in a new window]   Figure 3.   Effects of inhaled NO (40 ppm) on Ppa and PVR in normocythemic and polycythemic patients. Results are expressed as means ± SEM. *p < 0.01 versus control values.

Effect of Hemodilution on the Hemodynamic Response to Intravenous ACh in Polycythemic Patients

In six polycythemic patients (Patients 16 to 21, Tables 2 and 3) treated with isovolemic hemodilution, responses to ACh infusions were studied twice, at mean Hb levels of 18.6 ± 0.9 g/dl and 15.3 ± 0.3 g/dl, respectively. A significant interaction between the effect of hemodilution and ACh infusion was found when testing the effects of these treatments on Ppa (p < 0.001), PVR (p < 0.01), SVR (p < 0.05), CI (p < 0.01), and Pa_O2 (p < 0.05), but not on venous admixture. However, there was a significant effect of hemodilution (p < 0.05) and of ACh infusion (p < 0.05) on venous admixture. Hemodilution induced decreases in Ppa, from 33 mm Hg to 27 mm Hg (p < 0.01); PVR, from 7.7 ± 1.5 U/m² to 5.7 ± 0.8 U/m² (p < 0.05); and SOT, from 73 ± 10 ml/min/m² to 63 ± 6 ml/min/m² (p < 0.05), as well as an increase in Pa_O2, from 50 ± 2 mm Hg to 59 ± 4 mm Hg (p < 0.05) (Tables 2 and 3, Figure 4). Hemodilution did not significantly alter Psa, CI, SVR, or total VO₂. Infusion of ACh prior to hemodilution did not induce pulmonary or systemic vasodilation in this group of patients, nor did it alter gas-exchange parameters (Table 2, Figure 4). However, when infused after hemodilution, ACh produced decreases in Ppa, from 28 ± 1 mm Hg to 23 ± 1 mm Hg (p < 0.001); PVR, from 5.7 ± 0.8 U/m² to 3.6 ± 0.5 U/m² (p < 0.02); Psa, from 101 ± 5 mm Hg to 97 ± 5 mm Hg (p < 0.05); and SVR, from 29 ± 3 U/m² to 23 ± 2 U/m² (p < 0.01), and also increased CI, from 3.4 ± 0.4 L/min/m² to 4.1 ± 0.4 L/min/m² (p < 0.05) (Table 3, Figure 4). ACh infusion also decreased Pa_O2, from 60 ± 4 mm Hg to 49 ± 3 mm Hg (p < 0.05), and increased Qs/Qt, from 28 ± 5% to 39 ± 5% (p < 0.05), and SOT, from 63 ± 6 ml/ min/m² to 71 ± 5 ml/min/m² (p < 0.05).

View larger version (17K): [in this window] [in a new window]   Figure 4.   Effects of intravenous infusions of ACh (15 mg/min) on CI, Ppa, SVR, PVR, PaO2, and venous admixture in six polycythemic patients studied before and after hemodilution. A significant interaction between hemodilution and ACh infusion was found when testing the effects of these treatments on Ppa (p < 0.001), PVR (p < 0.01), SVR (p < 0.05), CI (p < 0.01), and PaO2 (p < 0.05), but not on venous admixture. Results are expressed as means ± SEM. *p < 0.01; **p < 0.001 versus control values.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our results show that the vasodilator activity of ACh in patients with CHLD varies with the level of circulating Hb. Although ACh infusion decreased pulmonary and systemic vascular resistances and impaired arterial oxygenation in normocythemic patients, its effects were considerably less marked in polycythemic patients. Isovolemic hemodilution, used to treat polycythemia, caused reductions in pulmonary artery pressure and resistance, and also restored the vasodilator response to ACh. We conclude that circulating Hb contributes significantly to vascular reactivity in patients with CHLD, and that polycythemia is associated with a loss of NO-mediated endothelium-dependent vasodilator responses to ACh. Our data also suggest that reduced endothelium-dependent vasodilation contributes to the increases in pulmonary and systemic vascular resistances seen in patients with polycythemia.

A vast literature indicates that a functionally intact endothelium is required for ACh to induce vasodilation in humans, and that most of the vasodilator activity of exogenously administered ACh is mediated by endothelial formation of NO (14). When infused into the brachial artery, ACh causes an increase in forearm flood flow. This effect is substantially blunted by coinfusion of L-arginine analogues to inhibit local synthesis of NO. In human vascular diseases such as atherosclerosis or systemic hypertension, ACh has been widely used to investigate endothelium-dependent vasodilation. Impairment of ACh- induced vasodilation is usually ascribed to impaired endothelial synthesis and/or release of NO (15). In addition to its systemic vasodilator activity, ACh is a potent pulmonary vasodilator and one of the first vasodilators examined in patients with pulmonary hypertension (16). Recently, ACh was proposed as a substitute for prostacyclin when screening for pulmonary vasodilator responsiveness in patients with primary pulmonary hypertension (17). In an earlier study, we compared hemodynamic responses to ACh infusion and NO inhalation in patients with COLD and moderate pulmonary hypertension (12). We found that the overall effect of ACh on pulmonary hemodynamics was a decrease in PVR of similar magnitude to that seen with inhaled NO. However, a few patients did not respond to ACh in this study (12). Because these patients had polycythemia, we undertook the present study to examine whether resistance to intravenous ACh might be related to increased Hb concentration.

In the normocythemic patients in the study, we found that ACh decreased Ppa, PVR, and SVR, and increased cardiac output in a dose-dependent fashion. These effects were associated with a marked alteration in arterial oxygenation, as reflected by a decrease in arterial PO₂ and an increase in venous admixture. In contrast, infusion of ACh to polycythemic patients only slightly increased cardiac output and decreased SVR, with no significant changes in Ppa, PVR, or arterial PO₂. In the whole study population Hb concentration was negatively correlated with the extent of systemic vasodilation, degree of pulmonary vasodilation, and alteration in gas exchange. The effects of ACh on gas exchange therefore parallelled its pulmonary vasodilator activity, with an impairment of arterial oxygenation being most marked in the normocythemic patients. Reciprocally, the absence of ACh-induced gas-exchange alterations in the polycythemic patients was related to the absence of a vasodilator response. The finding that polycythemic patients exhibited pulmonary vasodilation in response to exogenous inhaled NO but not to ACh infusion is consistent with impaired release or activity of endogenous NO by the pulmonary endothelium in response to ACh. Since only four patients received supplemental O₂ in our study, it is unlikely that combining inhalation of NO and O₂ in these patients would have interfered with the results (18).

Our results therefore strongly suggest that circulating Hb contributes significantly to ACh-induced vasodilation in both the systemic and pulmonary circulations. To investigate whether resistance to ACh could be reversed by normalization of Hb concentration, we examined ACh responses in a subgroup of polycythemic patients treated with isovolemic hemodilution. ACh infused after normalization of the Hb level increased cardiac output, decreased SVR and PVR, and impaired arterial oxygenation, whereas it was without effect prior to hemodilution. Moreover, hemodilution was associated with improvements in pulmonary hemodynamics and arterial oxygenation. That a reduction in Hct improves cardiopulmonary function in patients with cor pulmonale and erythrocytosis has been amply documented in earlier studies, supporting the concept of overcompensating erythrocytosis in cor pulmonale (1, 19). Experimental studies have also established that polycythemia contributes to the development of hypoxic pulmonary hypertension. Naeye and Jenssens observed that repeated phlebotomy to reduce the Hct to normocythemic values during hypoxia blunted the severity of pulmonary hypertension and right ventricular hypertrophy without affecting vessel structure (20, 21). In addition, Fried and colleagues found that acute phlebotomy to achieve normocythemia in chronically hypoxic rats reduced Ppa by 30% as compared with unphlebotomized hypoxic control rats (22). These effects of polycythemia have usually been ascribed to the increase in blood viscosity associated with polycythemia. However, in CHLD patients, they may also be caused by an increase in vascular tone related to impaired ACh-induced vasodilation, which may involve the endothelial L-arginine/NO pathway.

The mechanism by which polycythemia may effect the vasodilator response to ACh in patients with CHLD can only be suspected from our data. ACh is known to stimulate the endothelial L-arginine/NO pathway via muscarinic receptors, and alterations of this receptor-operated mechanism in polycythemic patients may cause resistance to ACh. Our finding that hemodilution restored vasodilation in response to ACh militates against this hypothesis. Another well established fact is that the endothelial L-arginine/NO pathway can be activated by shear forces exerted by circulating blood, which leads to flow-dependent vasodilation (5, 6). Since the red blood cell is a major determinant of blood viscosity, polycythemia, which increases shear stress via an increase in viscosity, may also increase the release of NO. Several observations support this hypothesis. In isolated rabbit lungs, inhibition of NO formation as associated with an increase in PVR only when the lung is perfused with blood (23). Experimental polycythemia induced in rats by chronic administration of erythropoietin (EPO) is associated with an increase in Psa and enhancement of the pressor response to the L-arginine analogue N-monomethyl-L-arginine (L-MMA), in accord with an increase in NO release (24). Conceivably, the failure to exhibit a vasodilator response to ACh observed in our polycythemic patients may have been related to previous activation of the endothelial L-arginine/NO pathway by shear forces exerted by circulating blood. In that case, however, inhaled NO would not have decreased Ppa or PVR in the polycythemic group. Moreover, such a mechanism is not in keeping with earlier reports that cerebral blood flow, limb blood flow, and limb oxygen transport improve dramatically after reduction in Hct (2, 3).

Alternatively, the failure of ACh to induce vasodilation in our polycythemic patients may have been caused by depletion of NO as a result of their increased Hb concentration. This hypothesis is based on indirect data. Free Hb has been shown to increase vascular resistance in vivo, as a result of its ability to inactivate NO, and to inhibit endothelium-dependent vasodilation (25, 26). Infusion of stroma-free Hb causes an immediate and pronounced increase in SVR in hemorrhagic shock (27). That hemoglobin in red blood cells is accessible to NO released in the vascular lumen is also widely accepted. Using in vitro vascular preparations, Gillepsie and coworkers have shown that Hb in intact red blood cells is as effective as free Hb in blocking endothelium-dependent vasodilation (7). Moreover, the ability of blood to inhibit NO-induced vasodilation has been shown to be inversely correlated with Hb oxygen saturation, a fact that may have a substantial impact in chronically hypoxemic patients, especially in the pulmonary circulation (28). In portal-hypertensive rats treated with EPO, development of polycythemia has also been shown to blunt the blood-pressure response to sodium nitroprusside and to attenuate the gastric and mesenteric blood-flow response to L-nitro arginine methyl ester (L-NAME), an inhibitor of NO formation (11). In isolated rat lungs, the presence of red blood cells in the perfusate has been shown to impair NO-mediated vasodilation (9). That EPO-induced changes in vascular tone may be due to a depletion of NO by increased levels of Hb in the blood has also been suggested by other studies (29). Moreover, the possibility that circulating Hb could influence the magnitude of NO-mediated vasodilation has been investigated in previous clinical studies. In patients with chronic severe anemia, inhibition of local NO synthesis by L-MMA caused a threefold greater decrease in blood flow than that seen in controls, suggesting that basal endothelium-derived relaxing factor (EDRF) activity was enhanced in the anemic patients (10). Moreover, red-blood-cell transfusion, which increased vascular resistance in anemic patients, reduced the vasococonstrictor response to L-MMA to the level observed in controls. It was therefore concluded that enhanced basal endothelium-derived NO contributes substantially to the low systemic vascular resistance seen in chronic severe anemia (11). The observation that NO-mediated vasodilation is enhanced in anemic patients is consistent with our finding of an impaired vasodilator response to ACh in polycythemic patients. Our data showing that hemodilution in polycythemic patients improved ACh-induced vasodilation and decreased vascular resistance are evidence in favor of an inhibition of vascular NO by circulating Hb, a mechanism that may contribute to the high vascular resistance seen in polycythemic patients.

In conclusion, our study suggests that polycythemia, which contributes substantially to the increases in Ppa and central blood volume seen in patients with chronic hypoxic lung disease, may act at least in part through an alteration in NO-mediated vasocodilation. Because ACh-induced vasodilation is inversely correlated with the level of Hb, a modest increase in the Hct may affect vascular reactivity in both the systemic and the pulmonary circulations.