In conscious animals, the response to hemorrhage is biphasic. During phase 1, arterial pressure is maintained. Phase 2 is characterized by profound hypotension. Despite allied roles, less is known about the integrated cardiovascular and respiratory response to blood loss in conscious animals. We evaluated cardiorespiratory changes during hemorrhage to test the hypotheses that 1) respiratory rate (RR) and blood gases do not change during phase 1; 2) RR increases during phase 2; and 3) RR and blood gas changes during hemorrhage are similar in males and females. We measured mean arterial pressure, RR, and blood gases during hemorrhage in 16 conscious, chronically prepared, male and female New Zealand white rabbits. We removed venous blood until mean arterial pressure was ≤40 mmHg. Sex did not affect mean arterial pressure, heart rate, PaO2, PaCO2, or pH during hemorrhage or the blood loss required to induce phase 2. PaCO2 decreased significantly from 37 ± 1 to 33 ± 1 and 29 ± 1 mmHg (P < 0.001) during phase 1 and 2, respectively. Before hemorrhage, PaO2 was 87 ± 2 mmHg. PaO2 was unchanged in phase 1 (92 ± 2 mmHg) but increased in phase 2 (101 ± 2 mmHg; P < 0.001). Body temperature, PvCO2 (thoracic vena cava), and ventilation-perfusion mismatch (A-a gradient) were unchanged during phases 1 and 2. Neither sex increased RR during phase 1. While males doubled RR during phase 2, RR in females did not change (P < 0.001). Thus, while PaCO2 decreases in phase 1 and phase 2, the decreases are achieved in different ways across the two phases and in the two sexes.
- cardiovascular control
- respiratory control
the cardiovascular and neurohumoral response to blood loss in conscious animals is biphasic (35). In phase 1, blood pressure is maintained by increased sympathetic nervous system activity and the resultant vasoconstriction and increase in systemic vascular resistance. During phase 1, there is increased release of renin, but release of adrenal catecholamines and vasopressin does not change. When blood loss reaches a certain critical volume, phase 2 ensues with a sudden drop in blood pressure accompanied by sympathoinhibition, global vasodilation, and dramatic increases in release of renin, vasopressin, and adrenal catecholamines (35). This biphasic cardiovascular and neurohumoral response to hemorrhage is preserved in a wide range of mammalian species (35), suggesting a potential evolutionary and thus, survival, advantage for the integrated response.
While the cardiovascular and neurohumoral effects of blood loss are reasonably clear (35), there is less information available concerning the respiratory effects of progressive blood loss in conscious animals. Most previous studies have used anesthetized animal preparations (22) or have focused on the compensatory mechanisms induced by prolonged hypovolemia and hypotension after hemorrhage has been terminated (11). To our knowledge, few studies (2) have simultaneously addressed the cardiovascular and respiratory response during phase 1 and phase 2 in conscious animals. This is somewhat surprising considering the complementary roles played by the cardiovascular and respiratory systems in terms of oxygen delivery and carbon dioxide removal. Tight integration of function in these two systems is likely to be of particular importance in pathophysiological states, such as blood loss, in which normal gas exchange and/or delivery is almost certainly compromised. A complete understanding of the integrated respiratory, cardiovascular, and neurohumoral response to blood loss is central to the development of new treatment modalities.
In the present study, we measured the respiratory and cardiovascular response to hemorrhage in conscious rabbits. Our hypotheses were that 1) there would be no change in arterial blood gas values or respiratory rate during phase 1; 2) respiratory rate would increase in phase 2; and 3) the respiratory rate and resulting blood gas values during phase 1 and phase 2 would be similar in male and female rabbits. To test our hypotheses, we measured mean arterial pressure, respiratory rate, and arterial blood gases in conscious, chronically prepared, male and female New Zealand white rabbits before hemorrhage, after nonhypotensive hemorrhage (i.e., during phase 1), and after hypotensive hemorrhage (i.e., during phase 2).
MATERIALS AND METHODS
Experiments were performed on 10 male and 6 female, adult New Zealand white rabbits weighing between 2.8 and 4.0 kg (3.2 ± 0.1 kg) and 2.7 and 4.0 kg (3.1 ± 0.1 kg), respectively. Studies were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals and were approved by the University of Missouri-Columbia Animal Care and Use Committee.
Rabbits were anesthetized with halothane, and a midline laparotomy was performed. One catheter was placed in the caudal aorta for blood pressure measurements and withdrawal of arterial blood gas samples. Another catheter was placed in the vena cava for blood removal (10). Electrodes for recording the diaphragmatic electromyogram (dEMG) were placed ventrolaterally in the costal diaphragm (37). In five rabbits, a polyvinylchloride tube was also placed in the abdomen. The end of the tube was plugged to create an airtight seal. During experiments, a temperature probe (Physitemp, PT-6) was passed to the end of the tube for continuous core body temperature measurements. Catheters, dEMG wires, and the temperature tube were passed subcutaneously to exit dorsally in the interscapular region. Animals were allowed at least 2 wk to recover from surgery before beginning experimentation.
During experiments, rabbits were placed in a box (33 × 15 × 18 cm, inside dimensions). A 6-cm-diameter hole in the front of the box provided for ventilation. The rabbits were acclimated to the box before initiating experiments. The arterial catheter was connected to a pressure transducer (Statham, model P23 Db). The dEMG wires were connected to an amplifier (World Precision Instruments model DAM50). The dEMG and arterial pressure were monitored on oscilloscopes (Tektronix model 2211 and BK Precision model 2120B, respectively) during the experiment and acquired (dEMG, 4–10 kHz; arterial pressure, 50 Hz) with a computer data-sampling system (Spike2, Cambridge Electronics Design).
Rabbits were heparinized with 2,000 units of sodium heparin IV (Eli Lilly) and allowed to equilibrate (≥10 min). Following equilibration, baseline cardiovascular and respiratory measurements were collected for 5–10 min. For the experiments reported here, all rabbits were in the experimental surroundings for a minimum of 20 min before starting the hemorrhage. During hemorrhage, blood was withdrawn (2.6 ± 0.1 and 2.7 ± 0.1 ml·kg−1·min−1 for females and males, respectively) from the abdominal venous catheter into sterile 60-ml syringes by using a syringe pump (Sage Instruments, model 351). When mean arterial pressure fell to 40 mmHg, five additional milliliters of blood were drawn, and the hemorrhage was terminated. The collected blood was reinfused at the conclusion of the experiment.
Arterial blood sampling precluded measurement of arterial pressure and heart rate during sampling. Therefore, we did separate experiments for blood gas measurements (n = 12; 6 females, 6 males) and for hemodynamic and respiratory rate changes during hemorrhage (n = 13; 6 females, 7 male). Nine rabbits (6 female, 3 male) were used for both series of experiments; 4 males were used only for blood gas experiments, and 3 males were used only for the measurements of hemodynamic and respiratory rate changes. For blood gas measurements, hemorrhage was initiated as described above but was interrupted to draw the arterial blood gas samples (Fig. 1). These samples (2 ml deadspace, 0.5 ml sample) were taken at three time points: control, immediately prior to hemorrhage; during phase 1, normotensive hemorrhage when heart rate had increased by 25%; and phase 2, when MAP ≤ 40 mmHg. Arterial blood was collected anaerobically in sterile 1-cc syringes and analyzed (Osmetech Opti CCA blood gas analyzer) at the conclusion of the experiment.
Mean arterial pressure and heart rate were derived online from the pulsatile arterial pressure signal. Respiratory data were analyzed off-line with Spike2 software (Cambridge Electronics). The dEMG signal was high-pass filtered (200–270 Hz) to remove ECG artifact and was rectified and smoothed (50 ms). Respiratory rate was derived from individual bursts in the processed signal. Two characteristics of the smoothed dEMG, delta dEMG and the area under the curve (Fig. 1, inset), were evaluated offline as an index of tidal volume (1, 8, 15). Delta dEMG was measured as the peak amplitude of the smoothed dEMG burst. For analysis, delta dEMG was averaged over a 60-s period. The area under the smoothed dEMG curve was also measured over a 60-s period. This area was divided by the number of respiratory cycles over the same period to yield an average area under the curve for each diaphragmatic burst. Delta dEMG and area under the curve were measured before hemorrhage, during phase 1 of hemorrhage, and during phase 2 of hemorrhage (i.e., at the same three times as arterial blood gases; see Fig. 1). Values were normalized as a percentage change from control values measured during the control period at the beginning of the experiment.
The difference between alveolar and arterial O2 (A-a difference) was calculated for each blood gas sample to estimate the efficiency of gas exchange and the degree of ventilation perfusion mismatch (39). Arterial oxygen (PaO2) and carbon dioxide (PaCO2) tensions were measured. Alveolar oxygen tension (PaO2) was calculated using the alveolar gas equation PaO2 = FiO2 [Pb − Pwv] − PaCO2/R, where FiO2 is the fraction of inspired oxygen, Pb is the barometric pressure, Pwv is the pressure of water vapor, and R is the respiratory quotient. We assumed FiO2 = 0.21, Pb = 760 mmHg, Pwv = 47 mmHg, and R = 0.8.
Respiratory rates were normalized to a control rate measured prior to the hemorrhage. Respiratory rate for the duration of the experiment was then expressed as a percentage of control. The amount of blood removed during hemorrhage varied between rabbits. For statistical analysis, the time required for the blood loss was normalized so that 0 percent equaled the beginning of hemorrhage and 100 percent equaled the point at which mean arterial pressure reached 40 mmHg. Because 5 ml of blood were drawn after mean arterial pressure reached 40 mmHg, blood loss ranged from 0 to 105% (31).
Differences over time were assessed by ANOVA for repeated measures. During hemorrhage, the independent variables considered were blood loss (% of total, see above paragraph) and sex. If this initial two-way ANOVA failed to demonstrate a significant main effect of sex or significant interaction of sex and blood loss, we pooled the male and female rabbits and performed a one-way ANOVA with blood loss as the independent variable. Significant F values were followed by Bonferroni's t-test. Significant differences were determined if P < 0.05. Values are expressed as means + SE (calculated from the pooled sample variance).
Effects of hypotensive hemorrhage on respiratory rate.
Table 1 shows resting heart rate, mean arterial pressure, and respiratory rate for 12 rabbits used in the hemodynamic studies. These data were acquired ∼2 min before initiating hemorrhage. There were no significant differences between males and females for resting cardiorespiratory values, although females showed a tendency for a greater respiratory rate. Fig. 1 shows an individual animal record during hemorrhage. Arterial pressure is maintained and heart rate increases during phase 1. The precipitous drop in pressure and decrease in heart rate near the end of the hemorrhage marks the onset of phase 2. In this male rabbit, respiratory rate increases coincident with the fall in arterial pressure. Figure 2 shows mean arterial pressure, heart rate and respiratory rate during hemorrhage in 12 rabbits. The rate of blood removal and the total blood loss (i.e., that required to reduce mean arterial pressure to 40 mmHg) were statistically similar for males and females (Table 1). The hemodynamic response to blood loss was also not affected by sex. Arterial pressure was well maintained through ∼80% of the total blood loss. Heart rate increased progressively during hemorrhage, reaching a peak at 90% of the total blood loss, and then began to decline. In Fig. 2, males and females are plotted separately for respiratory rate as a statistically significant interaction effect of percent of total blood loss, and sex was noted. Males significantly increased respiratory rate coincident with the decrease in mean arterial pressure and relative bradycardia that mark the onset of phase 2. Females did not increase respiratory rate in either phase 1 or 2 of hemorrhage. Importantly, in both males and females, respiratory rate did not change significantly prior to 80% of the total blood loss (i.e., prior to the onset of phase 2).
Effects of hypotensive hemorrhage on blood gas values.
Arterial blood gases were measured in 13 rabbits (6 female) at three time points: control (i.e., before hemorrhage); during phase 1; and during phase 2 (Fig. 1). Mean arterial pressure was similar between control and phase 1 but had decreased significantly by phase 2 (Figs. 2A and 3A). Although there was an effect of sex on respiratory rate during hemorrhage (see above), the blood gas response did not differ between males and females (Fig. 3). However, blood loss significantly affected blood gases and pH for the pooled male and female data (n = 13). PaCO2 decreased significantly from 37 ± 1 mmHg during control to 33 ± 1 and 29 ± 1 mmHg during phase 1 and phase 2, respectively (Fig. 3C). PaO2 (Fig. 3B) tended to be higher during phase 1 (92 ± 2, P = 0.11) than during control (87 ± 2 mmHg) and increased significantly (compared with control and phase 1) during phase 2 (99 ± 2 mmHg). PvCO2 values (thoracic vena cava) were not changed by blood loss (40 ± 1, 39 ± 1, 41 ± 1 mmHg for control, phase 1, and phase 2, respectively). Arterial pH increased significantly from 7.42 ± .01 during control to 7.45 ± .01 during phase 1 and 7.47 ± .01 during phase 2 (Fig. 3D).
PaO2 was derived from the measured blood gas values (see methods). The A-a difference (i.e., ) was then calculated to assess the level of diffusion impairment. The A-a difference was 16 ± 1, 16 ± 1, and 12 ± 1 mmHg at control, phase 1 and phase 2, respectively. Despite significant changes in PaO2 and PaCO2 at each time point, the change in A-a difference over the course of hemorrhage did not reach statistical significance (P = 0.085).
Effect of hypotensive hemorrhage on indices of tidal volume.
Delta dEMG and area under the curve (see materials and methods) were measured as indirect indices of tidal volume (Fig. 4). There was a significant interaction effect of sex and blood loss on area under the curve [F(2, 20) = 7; P = 0.005]. However, the only significant changes in area under the curve occurred in phase 2 when males decreased to levels lower than females and less than prehemorrhage or phase 1. Values for females were statistically similar at all three time points. Delta dEMG (data not shown) was not affected by the main effects of sex or blood loss or interaction of sex and blood loss.
Effect of hemorrhage on body temperature.
There was no significant change in body temperature during experiments (n = 5). At an average ambient temperature of 21.6 ± 0.4°C, core body temperature was 38.8 ± 0.1, 38.7 ± 0.1, and 38.7 ± 0.1°C for control, phase 1 and phase 2, respectively.
Effect of hemorrhage on hematocrit.
Hematocrit (n = 8) was 35 ± 1% prior to hemorrhage (control) and did not change significantly during phase 1 (35 ± 1%, P = 1). However, during phase 2, hematocrit decreased significantly to 30 ± 1%.
The results of this study provide evidence that PaCO2 decreases during nonhypotensive (phase 1), as well as hypotensive (phase 2) blood loss in conscious rabbits. These findings support earlier studies in anesthetized cats (22) and conscious pigs (11) that showed decreased PaCO2 after hypotensive hemorrhage. Importantly, the current study also provides new evidence that the decrease in PaCO2 begins before the onset of phase 2 of hemorrhage. In addition, it appears that a different mechanism may account for the decrease in PaCO2 during the two phases of blood loss and that the mechanism may be different in male and female rabbits.
The two most likely explanations for the decrease in PaCO2 are increased alveolar ventilation and decreased CO2 production (39). Although we did not measure CO2 production, several features of our results are not consistent with this explanation for the decrease in PaCO2. First, body temperature did not change over the course of the experiment. We might expect a decrease in metabolic rate to be accompanied by a decrease in body temperature. Second, the time between samples in our experiments was relatively short (4.9 ± 0.4 min from control to phase 1; 6.1 ± 0.4 min from phase 1 to phase 2, and only 11 ± 0.4 min from control to phase 3) compared with the period of time before sampling began (≥20 min). Finally, we also measured PvCO2 (thoracic vena cava), and while PaCO2 went down, PvCO2 did not change over the three time points. If decreased CO2 production accounted for the decrease in PaCO2, a similar decrease in PvCO2 would be expected. Alternatively, with a constant PvCO2 and no increase in ventilation, we would have expected no change in PaCO2. Thus, it is likely that the observed decrease in PaCO2 was due to increased alveolar ventilation and not decreased metabolism and CO2 production (39).
Why might hemorrhage be associated with an increase in respiratory drive? It seems reasonable to propose that a challenge to the cardiovascular system, such as hemorrhage, poses a challenge to the respiratory system as well. The sudden decrease in blood volume, as well as the fall in hematocrit (see results), may lead to inefficient oxygen delivery or relative hypoxia and, as a result, stimulation of peripheral chemoreceptors. In anesthetized rabbits, it has been shown that carotid chemoreceptor stimulation (intracarotid NaCN) leads to marked increases in phrenic nerve activity (increased slope of rise and height of integrated signal) and respiratory rate (21). Carotid bodies have a very high blood flow under normal conditions, and decreases in blood flow, even in the absence of concurrent decreases in arterial pressure or oxygen tension, can result in stimulation of carotid chemoreceptors (18). Carotid chemoreceptor stimulation due to decreases in hematocrit may also lead to respiratory stimulation, particularly during phase 2. In anesthetized cats, small decreases in hematocrit caused carotid chemoreceptor stimulation but only after sympathetic denervation of the carotid bodies (14). This condition is similar to phase 2 of hemorrhage, which is characterized by an abrupt decrease in sympathetic tone (5, 13, 24, 25) and a relative anemia (see results) compared with the prehemorrhage state. One study in conscious dogs clearly suggested the lack of any effect of chemoreceptor stimulation on the respiratory response to blood loss (2). In that study, carotid sinus denervation did not affect the ventilatory response to hemorrhage. In addition, minute ventilation did not increase until arterial pressure had decreased significantly. However, the dogs in the earlier study were breathing 100% oxygen, making it difficult to assess what role, if any, arterial chemoreceptors played during blood loss.
There is also evidence that baroreceptor unloading leads to respiratory stimulation. Brunner et al. (4) found that stepwise decreases in carotid sinus pressure caused increases in total ventilation in anesthetized, chemodenervated dogs. Our study provides evidence consistent with an increase in ventilation in phase 1. While blood pressure does not decrease in phase 1, the simultaneous increase in sympathetic nerve activity and heart rate that occurs during phase 1 is consistent with unloading of arterial baroreceptors (35). Thus, the baroreceptor reflex may be responsible in part for the increase in ventilation observed during early hemorrhage.
One possible mediator of the apparent increase in ventilation in phase 2 is ANG II. Circulating renin levels increase dramatically during phase 2 (33), and several studies have shown that ANG II can cause respiratory stimulation. In the absence of baroreceptor feedback in anesthetized or conscious dogs, infusions of ANG II led to increases in ventilation (27, 29). In phase 2 of hemorrhage, there is activation of the renin-angiotensin system, but mean arterial pressure remains quite low. Under these circumstances, ANG II may increase ventilation, while promoting the eventual recovery of arterial blood pressure (33).
The apparent increase in alveolar ventilation during phase 1 (i.e., the decrease in PaCO2) is not due to an increase in respiratory rate but may be the result of an increase in tidal volume (39). In conscious dogs, Baue and Nara (2) reported that the initial change in ventilation during hemorrhage before blood pressure dropped was an increase in tidal volume and a decrease in respiratory rate. We did not measure tidal volume. However, we evaluated two characteristics of the rectified and integrated dEMG (i.e., the area under the curve and peak amplitude of the signal) that have been used by others as an estimate of tidal volume (1, 15, 40). Using these estimates, there was no detectable increase in tidal volume (Fig. 4). This apparent discrepancy could be due to several factors. First, only diaphragmatic electrical activity was measured. Although we did not detect increases in dEMG activity, other inspiratory muscles (e.g., external intercostals) may have been recruited to cause an increase in tidal volume (23). Second, conscious rabbits have ventilatory patterns that are influenced by factors other than ventilation (e.g., olfaction, thermoregulation). As mentioned earlier, we tried to minimize this effect by acclimating rabbits to their surroundings and maintaining ambient temperature within the recommended range for rabbits (i.e., 21.6 ± 0.4°C) (17). Finally, other studies that have used this technique to measure tidal volume have used anesthetized animal preparations (1, 15, 40), where movement and changes in respiratory rate unrelated to gas exchange (e.g., olfaction) are not a problem.
In contrast to phase 1, a significant increase in respiratory rate was noted in males (but not females) in phase 2 of hemorrhage. This sex difference was not reflected in blood gas values. Arterial CO2, O2, and pH were statistically similar between males and females in phase 1 and phase 2. This sex difference in phase 2 could be due to documented differences in cardiovascular or respiratory reflexes, differences in involved hormonal systems, or differences in central mechanisms involved in cardiovascular control. Although it is clear that sex differences exist in cardiovascular reflexes (e.g., Refs. 7 and 16), there appears to be little if any information available on differences in the respiratory effects of these reflexes. There is also evidence that respiratory control is influenced by sex (e.g., Refs. 3 and 12). In addition, although there are conflicting results, some reports suggest sex differences in chemoreflex function. For example, it has been shown that hypoxic ventilatory drive is greater in female than in male conscious cats (38) and rats (26). However, like the cardiovascular reflex studies, there do not appear to be any studies that address mechanisms accounting for sex differences in the respiratory response to blood loss. There are also sex differences in all components of the renin-angiotensin system (see Ref. 9 for a review). However, we are unaware of studies addressing sex differences for the role of ANG II in respiratory control.
We have previously described the biphasic nature of the cardiovascular (36) and neurohumoral (30, 32–34) response to hemorrhage in rabbits and other animals (35). During phase 1, arterial pressure is maintained by increased sympathetic nerve activity and the resultant regional vasoconstriction with little change in release of humoral pressor agents. With the onset of phase 2, arterial pressure and sympathetic nerve activity fall abruptly to very low levels, and release of adrenal catecholamines, vasopressin, and renin increase dramatically. This study provides evidence that the biphasic nature of the cardiovascular and neurohumoral response to hemorrhage may be accompanied by a biphasic respiratory response as well. Considering their similar goals, a close relationship between the cardiovascular and respiratory systems in responding to a stressor such as hemorrhage would be expected. For example, our study suggests that an increase in tidal volume accounts for the apparent increase in ventilation seen in phase 1 in both sexes and perhaps during phase 2 in females. An increase in tidal volume during phase 1 has also been reported in conscious dogs during hemorrhage (2). Increasing tidal volume by increasing negative pressure within the thorax may augment venous return while also increasing ventilation. This has interesting implications for development of therapies in treating blood loss. Recent studies have cited the possible detrimental effects of positive pressure ventilation on venous return when treating blood loss (28). In contrast, inspiratory impedance devices have been explored as a therapeutic measure during decreases in central blood volume (6, 19, 20). These devices force the subject to inspire through a high-resistance tube, creating greater negative intrathoracic pressure during inspiration and augmenting venous return.
In contrast to our hypothesis, we noted a decrease in PaCO2 in phase 1. This decrease appears to be due to an increase in alveolar ventilation that is not mediated by an increase in respiratory rate. The increase in respiratory drive is most likely mediated by baroreceptor unloading, although decreases in chemoreceptor blood flow during this phase may also contribute. Our finding of an increase in ventilation during phase 2 of hemorrhage was consistent with our hypothesis. In addition, our study suggests that the male and female respiratory response to hemorrhage in rabbits may be different in phase 2. While PaCO2 decreased a similar amount in both males and females, in phase 2 of hemorrhage, only males showed a significant increase in respiratory rate. Our results do not allow us to make conclusions about the cause of the apparent increase in ventilation during phase 2, but possible explanations include baroreceptor unloading, stimulation of peripheral chemoreceptors, and release of humoral pressor agents. Finally, the changes in ventilation in phase 1 and phase 2, at least in males, appear to be accomplished by different mechanisms. This is not unexpected, considering the differences in the cardiovascular and neurohumoral adjustments associated with these phases.
These experiments were supported by the Office of Naval Research Grant N00014-02-1-0162, the National Institutes of Health Grants HL63910 and RR007004, and the University of Missouri Comparative Medicine Training Program.
We thank Jan Ivey, Michael McKown, and Eric Scherff for technical assistance with this study. We also thank Drs. John Dodam and Heidi Shafford for assistance in interpreting the blood gas data.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society