INTRODUCTION

TOP ABSTRACT INTRODUCTION MODEL DESCRIPTION RESULTS DISCUSSION REFERENCES

THE CARDIOVASCULAR RESPONSE to arterial hypoxia displays a well-defined pattern in conscious subjects; this is characterized by a significant increase in cardiac output (CO) and heart rate (HR), a decrease in total systemic resistance, a moderate rise in arterial pressure, and redistribution of blood flow toward the organs with higher metabolic need (brain, heart, and skeletal muscle) (9, 13, 18, 19). Various mechanisms are known to participate concurrently in these cardiovascular adjustments (1, 3). However, because their interaction is quite complex and exhibits significant nonlinearities, it is arduous to assess the exact role of each mechanism in different pathophysiological conditions.

It has been known for many decades that stimulation of peripheral chemoreceptors (by hypoxia or cyanide) induces a significant pressor response; in contrast, a distinct fall in arterial pressure is observed if hypoxia is induced after acute sinoaortic denervation in the anesthetized animal (16). These results might lead to the conviction that peripheral chemoreceptors play a major role in the cardiovascular regulation. Other authors, however, observed that the hemodynamic response to hypoxia remains substantially unchanged after sinoaortic denervation in the awake animal (17, 18); it was thus hypothesized that in the conscious subject a large portion of the cardiovascular response to hypoxia arises outside the sinoaortic zone.

Large differences have also been observed when the response to hypoxia is compared in subjects with spontaneous or artificially controlled ventilation (2, 4, 8, 22). These differences led several authors to think that pulmonary stretch receptors play a primary role in the hypoxic state. Interpretation of the latter results, however, is difficult, since experiments with controlled ventilation are generally performed in anesthetized animals; hence, one cannot exclude the possibility that differences are caused by the anesthetic per se, rather than by deactivation of lung stretch receptors.

A deeper understanding of the role of each mechanism in the cardiovascular adjustments to hypoxia is not only an important physiological problem, but one that may have profound clinical implications as well. For instance, knowledge of alterations in cardiovascular control induced by anesthesia and artificial ventilation may be of value in the management of patients in intensive care units, whereas assessment of the importance of chemoreflex and baroreflex adjustments may be useful in the treatment of cardiovascular disorders associated with hypoxia, especially in patients (such as elderly, diabetic, or uremic) who may be characterized by poor sinoaortic reflex responses.

To attain a more complete theoretical understanding of the cardiovascular adjustments during hypoxia, in the companion study (24) we developed a new mathematical model of the short-term cardiovascular control. The model was able to reproduce the typical behavior of the main hemodynamic quantities during deep hypoxia reasonably well, in accordance with physiological data, with emphasis on the existence of a biphasic time pattern.

The aim of this second, related study is to analyze the role of each control mechanism included in the model [peripheral chemoreceptors, arterial baroreceptors, central nervous system (CNS) hypoxic response, and lung stretch receptors] in greater detail. To this end, a sensitivity analysis is performed by simulating the cardiovascular response to deep hypoxia after selective elimination of each mechanism individually. Furthermore, the cardiovascular adjustments to isocapnic hypoxia have been simulated in different experimental conditions characterized by various degrees of mechanism interactions: various levels of hypoxia with all mechanisms intact (awake subjects), deep hypoxia in anesthetized subjects with controlled ventilation, and deep hypoxia in awake or anesthetized subjects after sinoaortic denervation. Anesthesia is presumed to suppress or reduce the CNS hypoxic response. In the previous examples, model predictions were compared with physiological data obtained from animals or humans.

	MODEL DESCRIPTION

TOP ABSTRACT INTRODUCTION MODEL DESCRIPTION RESULTS DISCUSSION REFERENCES

A complete description of the model, including mathematical equations and assignment of parameter numerical values, has been presented in the companion study (24). Hence, only the main aspects are summarized here in a qualitative way.

The model includes a pulsating heart, systemic and pulmonary circulation, and the action of various feedback regulatory mechanisms. The heart embodies passive atria (each described by means of a linear pressure-volume curve) and pulsating ventricles described through a variable-elastance model. The systemic and pulmonary circulations include the series arrangement of large arteries, peripheral circulation, and venous circulation. Hemodynamics in each compartment are mimicked by means of a linear pressure-volume curve (specified by compliance and unstressed volume) and a hydraulic resistance. Only in the large artery compartment is the effect of blood acceleration also taken into account by means of an inertial term. Moreover, in the description of the systemic circulation, we further distinguish between the peripheral and venous circulation of five different parallel compartments: the organs that exhibit local metabolic regulation (brain, heart, and skeletal muscle), the splanchnic circulation, and the remaining extrasplanchnic systemic vascular beds. This distinction is necessary, since the regulation mechanisms are known to have a different impact on these five compartments.

The action of feedback regulation mechanisms includes a direct effect of O₂ on the peripheral resistance in the organs under metabolic control and various kinds of reflex neural responses. According to the physiological literature, we assumed that a decrease in local venous O₂ concentration is able to induce a two- to threefold increase in blood flow in the autoregulated vascular beds. Moreover, O₂ consumption rate in the heart increases with the average power of the cardiac pump.

In the description of reflex mechanisms, we distinguished between the afferent neural pathways, efferent (sympathetic and vagal) activities, and effector responses.

The afferent pathways transport information from arterial baroreceptors, peripheral chemoreceptors, and slowly adapting lung stretch receptors with myelinated A fibers. We assumed that the latter receptors are sensitive to changes in tidal volume, which, in turn, is controlled by peripheral chemoreceptor activation. The action of each afferent pathway is mimicked by means of a static relationship arranged in series with a linear dynamic term.

The efferent neural responses embrace three different kinds of fibers: sympathetic fibers to systemic vessels, sympathetic fibers to the heart, and the vagus. The activity in sympathetic and vagal fibers is a nonlinear function of the weighted sum of all afferent information described above.

Chemoreceptor activation causes a significant increase in sympathetic activity to peripheral vessels, a moderate increase in cardiac sympathetic activity [mainly mediated by the aortic body chemoreceptors (12)], and an increase in vagal activity [mainly mediated by carotid chemoreceptors (5, 19)]. Furthermore, activation of lung stretch receptors causes a decrease in sympathetic activity to vessels (thus inducing vasodilation) and a significant decrease in vagal activity (thus contributing to the increase in HR). Finally, activation of arterial baroreceptors has an inhibitory effect on heart and vessels, causing a decrease in peripheral and cardiac sympathetic activities and an increase in vagal activity.

The model also incorporates the hypoxic response of the CNS. To this end, we included an offset term in the description of the sympathetic activity to the heart and peripheral vessels. When arterial PO₂ (Pa_O2) in the brain falls below a given threshold (50-60 mmHg for cardiac fibers and 35-40 mmHg for peripheral fibers), the offset term decreases quickly, leading to an increase in the corresponding sympathetic activity.

The effector responses embrace the influence of peripheral sympathetic nerves on peripheral hydraulic resistances and venous unstressed volumes in the splanchnic, extrasplanchnic, and skeletal muscle compartments, the effect of cardiac sympathetic nerves on heart period and on end-systolic elastances, and the effect of the vagus on heart period.

According to the previous description (24), the skeletal muscle compartment is under double control (metabolic and reflex).

	RESULTS

TOP ABSTRACT INTRODUCTION MODEL DESCRIPTION RESULTS DISCUSSION REFERENCES

Hypoxia with all mechanisms intact. To attain further model validation, we simulated the cardiovascular response to different levels of hypoxia (Pa_O2 from 80 to 28 mmHg) in steady-state conditions with all mechanisms working normally. The results are then compared with experimental data by Koehler et al. (13) in isocapnic conditions (Figs. 1 and 2). As clearly shown in Figs. 1 and 2, the model is able to predict the pattern of the main hemodynamic quantities [mean systemic arterial pressure (SAP), CO, HR, total peripheral resistance, and blood flow distribution in the autoregulated and nonautoregulated vascular beds] quite well in the overall range of Pa_O2 examined. In particular, mean SAP and the extrasplanchnic resistance increase moderately but monotonically over the entire range of Pa_O2. Coronary blood flow increases moderately during mild hypoxia but exhibits a three- or fourfold rise during deep hypoxia. In contrast, splanchnic blood flow exhibits only a mild increase, despite the rise in arterial pressure. As a consequence of the previous behavior, total peripheral resistance exhibits a mild increase during moderate hypoxia (when the effect of reflex vasoconstriction prevails over the local vasodilatory effect of O₂); then it decreases sharply during deep hypoxia because of maximal vasodilation of the autoregulated vascular beds. Finally, HR and CO show a moderate increase as long as Pa_O2 remains at >40-50 mmHg; then they increase significantly.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Percent changes of mean aortic pressure (A), cardiac output (CO) (B), heart rate (HR; C), and total peripheral resistance (TPR; D) measured by Koehler et al. (13) at different levels of isocapnic hypoxia compared with model simulation results.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Percent changes of extrasplanchnic resistance (A), splanchnic blood flow (B), and coronary blood flow (C) measured by Koehler et al. (13) at different levels of isocapnic hypoxia compared with model simulation results. Experimental percent changes of extrasplanchnic resistance have been computed from data on renal resistance reported by Koehler et al. Experimental percent changes of splanchnic blood flow have been computed from data on mesenteric blood flow.

Sensitivity analysis on the role of the individual mechanisms during deep hypoxia. To elucidate the specific role of each reflex mechanism included in the model (i.e., chemoreceptors, baroreceptors, lung stretch receptors, and the CNS hypoxic response) we compared the model's cardiovascular adjustments to deep hypoxia (28 mmHg) when all mechanisms are intact with those obtained after the selective elimination of a single mechanism. To eliminate a single mechanism without affecting the basal normoxic equilibrium, the receptor input quantity (local PO₂ in the case of peripheral chemoreceptors, brain PO₂ in the case of CNS hypoxia, tidal volume in the case of lung stretch receptors, and intravascular pressure in the case of arterial baroreceptors) was maintained at its constant basal level, which is different from the level in the rest of the circulation.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Sensitivity analysis of role of different reflex mechanisms [chemoreceptors, lung stretch receptors, central nervous system (CNS) hypoxia, and baroreceptors] in response to deep hypoxia. Percent changes in main hemodynamic quantities [mean systemic arterial pressure (SAP), CO, HR, and TPR] were evaluated with the model in the steady-state condition after deep hypoxia, first when all mechanisms are intact (basal) and then after selective inactivation of a single mechanism. Inactivation of only one mechanism, without affecting the normoxic equilibrium point, was achieved by cutting the corresponding feedback loop and artificially maintaining the input quantity at the receptor at its normal value. Elimination of chemoreceptors also excludes regulation by lung stretch receptors.

Response to deep hypoxia after sinoaortic denervation. Some authors (17, 18) observed that the hemodynamic response to deep hypoxia in awake dogs is not significantly altered by chronic sinoaortic denervation, with the only exception consisting of a reduction in the rise in SAP. In contrast, sinoaortic denervation in anesthetized animals results in a significant fall in mean SAP. To look for the model's explanation of these results, we simulated the time pattern of the cardiovascular response to deep hypoxia (28 mmHg) in three different conditions. In the first condition, all of the mechanisms are intact. The second condition consisted of sinoaortic denervation in awake subjects. To simulate this condition, baroreceptor and chemoreceptor responses were eliminated by maintaining baroreceptor arterial pressure and chemoreceptor PO₂ at their basal levels throughout the simulation. Chemoreceptor inactivation abolishes the response from lung stretch receptors as well. Hence, in this condition, the only regulatory mechanisms operating in the model are the CNS hypoxic response and the direct O₂ effect on coronary, skeletal muscle, and brain circulation. For the third condition, sinoaortic denervation with suppression of the CNS hypoxic response, we assumed that chemoreceptors and baroreceptors are inactivated, and we excluded the CNS hypoxic response by maintaining constant PO₂ as input to the CNS mechanism. Hence, the only active regulatory mechanism is the direct O₂ effect. Suppression of the CNS hypoxic response might occur, for instance, as a consequence of anesthesia (6, 20).

View larger version (39K): [in this window] [in a new window]   Fig. 4.   Time pattern of SAP, blood flow in the autoregulated vascular beds (brain + skeletal muscle + heart), and blood flow in all the remaining systemic vascular beds (splanchnic + extrasplanchnic circulation) simulated with the model in response to a rapid decrease in arterial PO2 (PaO2). In all simulations, PaO2 is lowered from 95 to 28 mmHg between 50 and 60 s and then maintained at the reduced level throughout the rest of the simulation. A: basal condition (all mechanisms intact). B: after inactivation of chemoreceptors and baroreceptors (i.e., simulating chronic sinoaortic denervation in awake subjects). C: after elimination of the CNS hypoxic response as well (i.e., the only active mechanism is the direct O2 effect on coronary, skeletal muscle, and brain vessels; this condition simulates chronic sinoaortic denervation in anesthetized subjects).

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Comparison between the normalized values of the main hemodynamic quantities (mean SAP, CO, HR, and TPR) evaluated with the model in the steady-state condition after deep hypoxia (28 mmHg) in conditions simulating awake subjects after chronic sinoaortic denervation (i.e., the same case as in Fig. 4B). Model results are then compared with those measured by Krasney and Koehler (17) in analogous conditions. The 100% value represents the value of the corresponding quantity during normoxia.

Response to deep hypoxia in anesthetized subjects with controlled ventilation. Various experimental works in previous years analyzed the cardiovascular response to isocapnic hypoxia in anesthetized subjects in whom ventilation is artificially kept constant. To simulate these conditions with our model, we eliminated the contribution of lung stretch receptors by setting tidal volume to its constant basal value throughout the simulation; moreover, we assumed that anesthesia reduces the strength of the CNS hypoxic response. Figure 6 shows the normalized values of SAP, CO, and HR simulated in steady-state conditions during deep hypoxia (28 mmHg) in conditions of constant tidal volume and with the assumption that the gain of the CNS hypoxic response is reduced to one-third of its basal value. Of course, the latter assumption is only hypothetical, although it has some experimental support (6, 20). However, anesthesia might have other effects, not considered here, on the cardiovascular regulatory response (see DISCUSSION).

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Comparison between the normalized values of the main hemodynamic quantities (mean SAP, CO, and HR) evaluated with the model in the steady-state condition after deep hypoxia (28 mmHg) in an anesthetized subject (strength of the CNS hypoxic response reduced to one-third of normal) with controlled ventilation (tidal volume artificially maintained at its basal level). Model results are compared with those measured by Kontos et al. (Table 1 in Ref. 15) under analogous conditions. Note the significant decrease in HR and the consequent strong attenuation of the CO response (cf. Fig. 5).

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Temporal pattern of the changes in some hemodynamic quantities (SAP, A; splanchnic blood flow, B; coronary blood flow, C; and total systemic blood volume, D) simulated with the model in response to progressive hypoxia from 95 to 15 mmHg in a subject with controlled ventilation (i.e., constant tidal volume), first with intact chemoreceptor and baroreceptor responses and then in conditions simulating chronic sinoaortic denervation. In these simulations, CO from the left heart was artificially maintained at a constant nonpulsating value (87.1 ml/s) to simulate the experiments by Hoka et al. (10). The chemoreflex response causes a significant redistribution of blood flow from the nonautoregulated to the autoregulated vascular beds and the mobilization of a great amount of blood volume (~18.0 ml/kg). Both phenomena are significantly reduced by denervation.

View larger version (24K): [in this window] [in a new window]   Fig. 8.   Changes in CO (A) and HR (B) vs. PaO2 simulated with the model at different levels of isocapnic hypoxia (continuous line) compared with data measured by Kontos et al. (14) in 35 experiments on 26 young normal human subjects (*). In the experimental trials, arterial PCO2 decreased from 37.7 mmHg during normoxia to 29.4 mmHg at PaO2 = 39.8 mmHg (isocapnic hypoxia).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MODEL DESCRIPTION RESULTS DISCUSSION REFERENCES

The multitude of factors involved in the cardiovascular response to hypoxia and their complex mutual relationships make assessment of the exact contribution of each regulatory mechanism in different pathophysiological conditions difficult. Redundancy, nonlinearities, and individual variability might account for the large differences in the physiological and clinical literature. Hence, the present work was designed with two main purposes: to analyze critically the possible etiology of these differences and to ascertain the specific contribution of the main reflex mechanisms in the response to hypoxia. The study was conducted with an original mathematical model of cardiovascular regulation (24).

Simulation results suggest that the model is able to provide a reliable description of the cardiovascular adjustments to hypoxia in a variety of different conditions: various levels of hypoxia (ranging from moderate hypoxia to severe hypoxia <30 mmHg), the possible effect of anesthesia, chronic sinoaortic denervation, and controlled ventilation. The reliability of the results suggests that the model may be of didactic value, may help researchers in the confirmation or rejection of physiological hypotheses, and may help integrate different experimental findings into a single theoretical structure.

An important result, arising from the sensitivity analysis, is that a significant redundancy exists in the interaction among cardiovascular regulatory mechanisms. In particular, looking at Fig. 3, one can observe that inactivation of a single mechanism (e.g., chemoreceptors or CNS hypoxia) does not induce vital changes in the cardiovascular response to deep hypoxia. This might lead an experimental researcher to the false hypothesis that the mechanism under study does not exert a strong effect on cardiovascular parameters. However, this hypothesis would be incorrect. The chemoreceptors and CNS hypoxia have a significant influence on cardiovascular parameters and, hence, contribute to regulation; nonetheless, inactivation of a single mechanism alone does not lead to a substantial impairment of the entire response to hypoxia, since the remaining mechanisms are able to compensate for the lack rather well and ensure a great deal of the compensatory reserve.

Figure 3 suggests that the chemoreceptors contribute chiefly to the elevation of mean SAP during hypoxia; in the absence of this mechanism, arterial hypertension is attenuated. However, the chemoreflex exerts a smaller control on CO and HR. This observation permits us to explain data by Kontos et al. (14) on human volunteers subjected to hypocapnic hypoxia. The increase in CO and HR reported by these authors can be reproduced reasonably well with the model (Fig. 8). In contrast, our model predicts arterial hypertension during hypoxia, whereas Kontos et al. did not observe any hypertension, even when Pa_O2 was reduced to <40 mmHg. Hypocapnia attenuates the chemoreceptor afferent activity. Hence, the results by Kontos et al. can be explained by our results in Fig. 3, which were obtained after elimination of the chemoreceptor response.

The CNS hypoxic response makes a significant contribution to the increase in SAP and (through a strong impact on HR and contractility), especially in CO.

Furthermore, the sensitivity analysis suggests that the pulmonary reflex response only slightly contributes to set the hypoxic SAP and CO values, since, in the absence of this mechanism, both quantities exhibit only minor changes. However, the role of lung stretch receptors is not negligible; in the absence of this reflex, similar SAP and CO values are obtained by reducing HR, increasing sympathetic activity to peripheral vessels (which increases peripheral resistance in the reflexly regulated vascular beds and mean filling pressure), and evoking a different blood flow redistribution among vascular beds in favor of brain, heart, and skeletal muscle circulation.

Elimination of lung stretch receptor input in Fig. 3 does not cause the cardiovascular adjustments commonly observed in the diving response (i.e., bradycardia and an increase in total systemic resistance). The reason is that the diving response also involves a reflex input from nasopharyngeal receptors and progressive hypercapnic hypoxia, which augments the chemoreceptor gain.

Finally, the sensitivity analysis resolutely suggests that the role of the baroreflex is very important in avoiding excessive changes in mean SAP, CO, and HR during hypoxia. In the absence of a baroreceptor response alone (a condition that is not easily found in conscious subjects), the simultaneous action of chemoreceptors, lung stretch receptors, and CNS hypoxia might lead to unacceptable hypertension and tachycardia.

The subsequent simulations (Figs. 4-7) demonstrate that the model is able to satisfactorily elucidate results of several physiological experiments performed under different conditions. In particular, simulations performed in circumstances analogous to chronic sinoaortic denervation (i.e., after simultaneous inactivation of baroreceptors and chemoreceptors) permit us to clarify the differences reported in the literature between anesthetized and awake animals reasonably well (13, 16-18). The model ascribes these differences to the action of the CNS hypoxic response working, through sympathetic cardiac efferent fibers, on HR and contractility.

The latter result, however, crucially depends on two important hypotheses included in the model, which now deserve a critical discussion. The first is that the CNS hypoxic response is significantly reduced by anesthesia. The second is that, in awake subjects, cardiac sympathetic activity is increased by CNS hypoxia during moderate hypoxia (Pa_O2 = 25-60 mmHg), whereas sympathetic activity to peripheral vessels is increased by CNS hypoxia at lower Pa_O2 (25-40 mmHg). Various experimental results support these two hypotheses. Koehler et al. (13) analyzed the cardiovascular response to isocapnic hypoxia in conscious dogs at different levels of hypoxia before and after sinoaortic denervation. In the awake sinoaortic-denervated animal (in which the only reflex response can be ascribed to CNS hypoxia), isocapnic hypoxia increased HR and left ventricular contractility (dP/dt_max), suggesting an increased sympathetic activity to the heart, from Pa_O2 of 50-60 mmHg. The authors concluded that "the most likely explanation for these responses is a direct effect of hypoxia on the CNS." Moreover, an increase in the reflexly regulated resistance could be observed in the sinoaortic-denervated animal as Pa_O2 decreased from 40 to 25 mmHg.

The CNS response to ischemia and hypoxia in anesthetized dogs was carefully assessed by Downing et al. (6) in the early 1960s. These authors reported that CNS hypoxia induces a significant increase in peripheral vascular resistance and HR as well as in cardiac contractility, which can be ascribed to a simultaneous increase in the activity of cardiac and peripheral sympathetic nerves. However, in this study the threshold for the CNS response to hypoxia was estimated to be Pa_O2 of 30 mmHg, which is much lower than the threshold observed by Koehler et al. (13) in awake animals. We may thus hypothesize that, in awake subjects, the CNS is able to increase sympathetic activity to the heart and peripheral vessels, even during moderate hypoxia, but this response is significantly depressed (or shifted to a lower Pa_O2) by anesthesia.

With incorporation of these hypotheses, the model described in our companion study (24) is able to account for the experimental findings by Krasney et al. (17, 18). Although hypoxia in anesthetized subjects results in a significant fall in mean SAP and in only a moderate increase in CO after sinoaortic denervation (Fig. 4C), hypoxia in awake denervated animals results in a significant increase in CO (comparable to the increase observed in intact animals) and in only a minor attenuation in the rise in SAP. Hence, model predictions support the idea that the CNS hypoxic response alone is able to ensure quite a normal cardiovascular response to hypoxia, even in the absence of other reflex regulatory mechanisms.

The model also contributes to clarification of the cardiovascular behavior observed in anesthetized animals with controlled ventilation. Our results agree with those of Kontos et al. (15) and Angell James and Daly (2), who reported a significant fall in HR in controlled ventilated subjects. Angell James and Daly ascribed the significant fall in HR in this experimental condition mainly to the absence of lung stretch receptors. However, our simulations suggest that the fall in HR may only in part be ascribed to the absence of lung stretch receptors, whereas a significant responsibility should also be attributed to anesthetic depression of the CNS hypoxic response.

When simulating the experimental work by Kontos et al. (15), we assumed that the CNS hypoxic response is not completely abolished, but significantly reduced, by anesthesia. This assumption is hypothetical, although it has been supported by experimental data (6, 20). However, it is important to recognize that anesthesia may have other effects on the cardiovascular system that are not considered in the present work. For instance, anesthesia may reduce baroreflex and peripheral chemoreflex as well as the CNS hypoxic response. Moreover, some anesthetics (especially barbiturates) may almost completely suppress vagal activity. Indeed, in many experiments on animals, HR is elevated, suggesting a slight vagal tone. In this regard, Gupta and Singh (7) proposed an alternative explanation for the bradycardic response to hypoxia in anesthetized artificially ventilated animals. Under different anesthetics, these authors observed that a tachycardic response occurs in dogs with low HR (suggesting high vagal tone), but this response is converted to bradycardia in dogs with high control HR (i.e., with reduced resting vagal tone and, hence, inherently smaller reserve to increase HR). A further factor that might complicate interpretation of experimental results is surgical stress. The latter might shift sympathetic and hormonal working point, thus affecting the normal cardiovascular response.

Finally, simulations performed in conditions of constant CO (anesthetized animals with controlled ventilation) confirmed that, during hypoxia, activation of efferent sympathetic nerves by chemoreceptors is able (especially through a decrease in venous unstressed volume) to induce a marked reduction in total systemic blood volume (~18.0 ml/kg in our model), thus contributing to the observed increase in mean systemic filling pressure (21, 23). Hoka et al. (10) recently measured a decrease in total systemic blood volume as great as 22 ml/kg in analogous experimental conditions in the dog. In an earlier experiment performed under analogous conditions, Kahler et al. (11) showed that 16 ml/kg of blood can be mobilized toward an external reservoir during severe hypoxia.

In conclusion, the present study demonstrated that various experimental results on the cardiovascular response to hypoxia, reported in the physiological literature, can be assessed well by the mathematical model described in the companion study (24). These circulatory adjustments ensue from the complex interplay of several factors, which superimpose themselves with a high degree of redundancy. As a consequence, similar steady-state levels of mean SAP and CO can be attained during hypoxia by means of various combinations of regulatory actions, causing a different internal redistribution of blood flow among vascular beds and dissimilar changes in HR, stroke volume, mean systemic filling pressure, and peripheral resistance. Quantitative analysis of the nonlinear interplay among regulatory mechanisms should be considered in the interpretation of physiological studies and, in perspective, might be helpful in the management of cardiovascular disorders associated with hypoxia.