Main

Alkalosis has long been used in treating neonatal and pediatric pulmonary hypertension (1–4). This practice is based on animal (5–8) and human (9, 10) studies showing that alkalosis caused acute pulmonary vasodilation. However, in these and other studies, responses were measured within 5–30 min of initiating alkalosis. In contrast, alkalosis therapy is generally maintained for hours or days in clinical practice, and despite sustained alkalosis, many children develop significant pulmonary hypertension after the initial vasodilator response. Although this may reflect the natural course of their disease, several recent studies of isolated pulmonary arteries and lungs have found that vascular reactivity was enhanced after as little as 15–60 min of alkalosis (11–16). Whether in vivo pulmonary vascular reactivity is enhanced after relatively brief periods of sustained alkalosis is unknown.

Mechanical ventilation designed to achieve hypocapnic alkalosis (or even eucapnia in the presence of severe lung disease) can result in significant barotrauma. This has led several groups to propose the use of lower peak inflating pressures and/or lower tidal volumes, resulting in “permissive” hypercapnia (17–19). However, this may aggravate pulmonary hypertension, for acidosis has been found to cause acute pulmonary vasoconstriction in several studies of animals (5, 20) and humans (9, 10). On the other hand, hypercapnic acidosis failed to increase PVR in intact dogs (21), and acidosis appeared to decrease smooth muscle contractility in vitro (22, 23). The relative effects of acute and sustained hypercapnic acidosis on pulmonary vascular reactivity in vivo have not been described.

In view of the widespread use of alkalosis therapy and growing use of permissive hypercapnia in clinical practice, we sought to elucidate the consequences of acute and sustained changes in pH on the pulmonary vasculature of an intact animal. Two-week-old piglets were studied because, although their baseline PVR is low, hypoxia induces significant vasoconstriction at that age (24–26), making the model useful for assessing the effects of altering pH during acute, hypoxia-induced pulmonary hypertension. Based on previous in vitro studies cited above, we hypothesized that the acute vasodilator or vasoconstrictor responses to alkalosis or acidosis, respectively, would not be sustained. To test this hypothesis, the effects of acute (20 min) and sustained (60–80 min) respiratory alkalosis (PCO2 approximately 25 torr, pH approximately 7.60) or acidosis (PCO2 approximately 65 torr, pH approximately 7.23) on baseline (O2 = 35%) PVR and the hypoxia-induced (O2 = 12%) increase in PVR were compared with responses measured during eucapnia (PCO2 approximately 44 torr, pH approximately 7.40). In addition, because recent in vitro studies suggest that pH-induced changes in pulmonary vascular reactivity may be mediated by the nitric oxide–cyclic GMP pathway (27, 28), eNO production was measured under the different study conditions.

METHODS

This study was approved by the Institutional Animal Care and Use Committees of the Medical College of Wisconsin and the Zablocki Veterans Administration Medical Center in Milwaukee. Two-week-old piglets (n= 21) were prepared as previously described (25). Briefly, after anesthesia with ketamine (40 mg/kg i.m.), acepromazine (1.5 mg/kg i.m.), and pentobarbital (25 mg/kg, i.p.), a midline incision was made in the neck and a tracheostomy tube was placed. Catheters were then inserted in one carotid artery for constant systemic blood pressure measurement and intermittent blood gas sampling and in one external jugular vein for CVP measurement and cardiac output injections. A 5F balloon-tipped thermodilution cardiac output pulmonary artery catheter (model 132F5, Baxter Healthcare Corp, Irvine, CA) was introduced in the other external jugular vein and advanced into the pulmonary artery while continuously measuring and recording pressure.

Protocol.

Ventilation (model 613, Harvard Apparatus, South Natick, MA) was initiated and pentobarbital (25–50 mg i.v.) and pancuronium bromide (0.1 mg/kg i.v.) were administered. Anesthesia was maintained throughout the experiment with pentobarbital (25–50 mg i.v.) administered every 45–60 min as needed. All piglets were ventilated at a rate of approximately 22 breaths/min, tidal volume of approximately 18 mL/kg, and end-expiratory pressure of 2–3 mm Hg (3–4 cm H2O). These settings were held constant throughout each experiment to preclude any effect of changes in tidal volume or minute ventilation on the hemodynamic responses to acidosis or alkalosis. The desired PaCO2 and pH were achieved by varying the inspired CO2. Metabolic acidosis that was frequently noted immediately after instrumentation of the animals was corrected with 1–2 mEq/kg of 1 N NaHCO3 as needed. No further NaHCO3 was required during the study. Figure 1 illustrates the protocol used in this study. All piglets were initially ventilated with 35% O2 and approximately 3% CO2 to achieve a PaO2 of approximately 200 torr and PaCO2 of approximately 40 torr (target pH approximately 7.40). After 60 min of stabilization, baseline hemodynamics and eNO concentration were determined (Fig. 1, 60 min). The inspired O2 was then reduced to 12%, and after 20 min, stable hemodynamic responses and eNO concentration during eucapnic hypoxia were measured (Fig. 1, 80 min).

Figure 1
figure 1

Experimental protocol (see “Methods” for details).

Full size image

The piglets were then assigned to one of three treatment groups: CON (n= 8), ALK (n= 7), or ACID (n= 6). The order of experiments using piglets from each group was mixed during the course of the study to preclude any effect of improved expertise or season on the results. One failure in the ACID and one failure in the ALK group were not included. The study was terminated before completing eight experiments in each group because differences among the groups achieved significance with the lower numbers of animals in the ACID and ALK groups. CON piglets were maintained eucapnic with a pH of approximately 7.40 throughout the experiment. ALK piglets were made hypocapnic by reducing the inspired CO2 to less than 0.5%, resulting in a PaCO2 of approximately 25 torr (target pH approximately 7.60). ACID piglets were made hypercapnic by increasing the inspired CO2 to approximately 6.5%, resulting in a PaCO2 of approximately 70.0 torr (target pH approximately 7.20). After 20 min of hypoxia under eucapnic, hypocapnic, or hypercapnic conditions, hemodynamic and eNO measurements were repeated to determine the responses to acute acidosis or alkalosis (Fig. 1, 100 min). CON, ALK, and ACID piglets were then returned to 35% O2 while maintaining eucapnia, hypocapnia, or hypercapnia, respectively, and the effects of sustained changes in pH on baseline hemodynamics and eNO concentrations were determined after 40 min (Fig. 1, 140 min). A second challenge with 12% O2 was then applied to determine the response to hypoxia during sustained alkalosis or acidosis, and hemodynamic and eNO measurements were repeated after 20 min (Fig. 1, 160 min). All piglets were then returned to eucapnia with 35% O2, and baseline measurements were repeated after 20 min. Responses to a final eucapnic hypoxic challenge were then again determined after 20 min (Fig. 1, 200 min).

Measurements.

Airway pressure, PAP, SAP, and CVP (all in millimeters of mercury) were constantly measured (Statham Gould model P23id pressure transducers, Spectramed, Oxnard, CA) and recorded (model 7D polygraph, Grass Instruments, Quincy MA). CO (in milliliters per minute; model 9520, Edwards Laboratories, Santa Anna, CA) and PCWP (in millimeters of mercury) were measured at end expiration at the end of each experimental condition. CO measurements were repeated three to four times under each study condition, and the mean CO for that condition was calculated from the three measurements falling within 5% of each other. The mean CO was then indexed to body weight (CI = CO/wt). PVRI and SVRI were calculated as (PAP − PCWP)/CI and (SAP − CVP)/CI, respectively, and are expressed in resistance units (mm Hg·min·kg·mL−1). Exhaled O2 and CO2 concentrations were constantly monitored and blood gases were analyzed (model 278 blood gas analyzer, CIBA-Corning, Medfield, MA) during each study condition.

In three CON, five ALK, and three ACID piglets, eNO was continuously measured by passing a portion of the exhaled gas through a chemiluminescence nitric oxide analyzer (model 270B, Sievers, Boulder, CO) at a sampling rate of 200 mL/min. The analyzer was calibrated with authentic nitric oxide (Matheson, Chicago, IL) diluted in N2 to achieve a concentration range of 0 to 150 parts per billion (ppb) before each experiment, and data were recorded on the polygraph throughout the experiment. The eNO production in nanomoles per minute was then calculated (eNO concentration × minute ventilation divided by 22.4 L/mol).

Data analysis.

All data are expressed as mean ± SEM. Hemodynamic data (PAP, SAP, CVP, PCWP, CI, PVRI, and SVRI) are shown for each group. Baseline eNO production during eucapnic 35% O2 ventilation for each group is shown followed by percent change relative to the initial baseline eNO production. Data were analyzed using one-way or two-way repeated measures ANOVA and the least significant difference test for post hoc comparisons if the ANOVA was significant at p< 0.05.

RESULTS

There were no differences among the CON, ALK, or ACID treatment groups in terms of animal weight (4.6 ± 0.1 kg), initial hematocrit (29.5 ± 1.0%), peak airway pressures (14.3 ± 0.2 mm Hg [19.1 ± 0.3 cm H2O]), end-expiratory pressure (2.4 ± 0.1 mm Hg [3.2 ± 0.1 cm H2O]), and ventilator rate (22 ± 0.8 breaths/min) during the study. Hy- poxia (12% O2) caused a significant decrease in PaO2 in all groups, but PaO2 during normoxia (181 ± 5.4 torr) and during hypoxia (50.8 ± 0.7 torr) did not differ among treatment groups. PaCO2 (44.2 ± 2.9 torr) and pH (7.39 ± 0.01) did not differ among treatment groups during eucapnic experimental conditions. There was a significant decrease in PaCO2 (26.7 ± 2.1 torr) and increase in pH (7.58 ± 0.02) when the inspired CO2 was reduced in the ALK group. There was a significant increase in PaCO2 (67.2 ± 1.9 torr) and decrease in pH (7.23 ± 0.01) when the inspired CO2 was raised in the ACID group.

CVP ranged from 1.9 to 4.1 mm Hg and PCWP ranged from 2.0 to 4.3 mm Hg under the different experimental conditions in all treatment groups. There were no significant effects of treatment group or experimental condition on CVP or PCWP. CI data from the different experimental conditions for all treatment groups are shown in Table 1. Although the effects of experimental condition on CI did not differ among the treatment groups, mean CI for all piglets did increase significantly during hypoxia.

Table 1 Cardiac index (mL•min•kg−1) measurements under different experimental conditions* * Treatment group did not alter the effects of experimental condition on CI, but hypoxia caused a significant increase in mean CI in all piglets.
Full size table

PAP and PVRI increased significantly in response to the first eucapnic hypoxic challenge in all groups (Fig. 2 80 min), and the hypoxic response was unchanged at 100 min in the CON group. However, 20 min of alkalosis caused a significant decrease in hypoxic PAP and PVRI in the ALK group and 20 min of acidosis caused a significant increase in hypoxic PAP in the ACID group (Fig. 2, 100 min). After returning to 35% O2, PAP and PVRI decreased in all groups. Baseline PAP and PVRI were similar in piglets during eucapnia (CON) or sustained hypocapnia (ALK), but were significantly higher during sustained hypercapnia (Fig. 2, 140 min). The increase in PAP in response to the hypoxic challenge administered during sustained alkalosis (ALK) was significantly attenuated compared with CON (Fig. 2A, 160 min). However, the increase in CI during the second hypoxic challenge was also less in the ALK (16.7 ± 8.4 mL·kg−1·min−1) compared with the CON (41 ± 11.1 mL·kg−1·min−1) groups. Thus, the hypoxia-induced increase in PVRI was similar during sustained alkalosis (ALK) and eucapnia (CON) (Fig. 2B, 160 min). The hypoxic challenge administered during sustained acidosis (ACID) resulted in a significantly greater increase in both hypoxic PAP and hypoxic PVRI than occurred during eucapnia (CON) or sustained alkalosis (ALK) (Fig. 2, 160 min).

Figure 2
figure 2

Effects of pH on normoxic (35% O2) and hypoxic (12% O2) PAP and PVRI (shaded areas indicate hypoxia). Treatment group significantly altered the effects of experimental condition on PAP and PVRI. * indicates significantly higher than control, † indicates significantly less than control, ** indicates both ACID and ALK are higher than CON.

Full size image

Baseline PAP and PVRI did not differ among groups after returning to eucapnic ventilation with 35% O2, (Fig. 2, 180 min). However, the final eucapnic hypoxic challenge resulted in a significantly greater increase in PAP and PVRI in piglets previously exposed to alkalosis compared with those ventilated with a eucapnic gas mixture throughout the experiment (ALK versus CON in Fig. 2, 200 min). The final eucapnic hypoxic PAP and PVRI in the ACID group were lower than those measured during the preceding hypercapnic period, but remained higher than in the CON group (Fig. 2, 200 min).

SAP did not change significantly under the different experimental conditions in either CON or ACID, but decreased significantly (approximately 7–10 mm Hg) during hypoxic alkalosis in ALK (Fig. 3A). Treatment group did not alter the effects of experimental condition on SVRI (Fig. 3B), but the simple effect of experimental condition on mean SVRI for all piglets was significant owing to the hypoxia-induced decrease in SVRI.

Figure 3
figure 3

Effects of pH on normoxic (35% O2) and hypoxic (12% O2) SAP and SVRI (shaded areas indicate hypoxia). Treatment group did not change the effects of experimental condition on SAP or SVRI. One-way ANOVA showed that SAP was significantly lower during alkalotic hypoxia compared with eucapnic hypoxia in the ALK group (†). Hypoxia caused a significant decrease in mean SVRI for all piglets.

Full size image

Baseline eNO production did not differ significantly among groups (CON, 1.31 ± 0.18 nmol/min; ALK, 1.03 ± 0.65 nmol/min; and ACID, 0.74 ± 0.28 nmol/min), and there was considerable variability in baseline eNO production within groups. The effects of experimental condition on the percent change in baseline eNO production are shown in Figure 4. The percent decrease in eNO production was significant after 40 min of hypoxia in CON and ACID piglets, then remained unchanged for the rest of the experiment. The percent eNO production did not change significantly from baseline during alkalosis in the ALK group and remained significantly higher than the CON or ACID piglets until eucapnic ventilation was restored.

Figure 4
figure 4

Effects of pH on the percent change in eNO concentration during normoxia (35% O2) and hypoxia (12% O2 in shaded areas). The initial decrease in eNO during hypoxia seen in CON and ACID piglets was blunted during alkalosis. * indicates significantly higher in ALK group.

Full size image

DISCUSSION

In this study of normal intact piglets, alkalosis caused acute pulmonary vasodilation when PVR was elevated by hypoxia, but pretreatment with hypocapnic alkalosis failed to attenuate the subsequent pressor response to hypoxia. Furthermore, eucapnic HPV appeared to be accentuated in piglets previously exposed to 80 min of alkalosis. Acute hypercapnic acidosis caused only a small and statistically insignificant increase in hypoxic PAP, but 60–80 min of sustained acidosis resulted in a marked increase in both baseline and hypoxic PVR.

Alkalosis failed to alter baseline PVR in this study of normal intact piglets (Fig. 2B, 140 min), likely because vascular tone was already low. However, after raising PVR with hypoxia, alkalosis caused acute pulmonary vasodilation (Fig. 2B, 100 min), as has previously been reported in intact animals, isolated lungs, and isolated pulmonary arteries (6–8, 11, 16, 28). Although the mechanism underlying this response is not fully defined, alkalosis has been shown to increase endothelial cytosolic Ca2+ concentration (29), a potent stimulus for synthesis of endothelium-derived modulators such as nitric oxide and prostacyclin. In this study, hypoxia caused a significant decrease in eNO production in CON and ACID piglets, as previously reported in isolated piglet lungs (30), but eNO production did not decrease from baseline during hypoxic alkalosis in ALK piglets (Fig. 4). Although the numbers of animals were small, the variability within groups was considerable, and nonendothelial sources of eNO cannot be excluded, these data are consistent with our recent study of isolated piglet pulmonary artery rings suggesting that alkalosis-induced vasodilation is mediated by the nitric oxide–cyclic GMP pathway (28). In contrast, prostacyclin appeared to mediate alkalosis-induced vasodilation in isolated rat and lamb lungs (16, 31), suggesting that the endothelium-derived modulator responsible may differ among species. Nonendothelium-derived modulators may also contribute to the response, for neither nitric oxide nor prostacyclin appeared to play a role in alkalosis-induced vasodilation in intact lambs and rabbit lungs (7, 27, 32). Furthermore, recent patch-clamp studies found that pulmonary artery smooth muscle K+ channels were activated by alkalosis (33, 34), suggesting that alkalosis may act by inducing pulmonary artery smooth muscle hyperpolarization. Further investigation of the mechanism of acute alkalosis-induced pulmonary vasodilation is required.

Previous in vitro studies have shown that preexistent alkalosis either failed to attenuate or actually enhanced hypoxia- and thromboxane-induced pulmonary vasoconstriction (6, 11–14, 16). To our knowledge, this is the first in vivo study showing that preexistent alkalosis failed to attenuate pulmonary vasoconstriction (Fig. 2B). The factors responsible for the apparently discordant effects of acute and sustained alkalosis are unknown. There does not appear to be a decrease in vasodilator modulator activity, for the eNO concentration remained unchanged throughout alkalosis in this study (Fig. 4). An alternative explanation is suggested by studies of both pulmonary and systemic vessels showing that alkalosis enhanced smooth muscle contractility through several mechanisms, including increased vascular smooth muscle cytosolic Ca2+ and enhanced Ca2+-calmodulin binding (12, 13, 15, 22, 23, 35). We speculate that alkalosis induces an acute increase in modulator activity sufficient to attenuate preexistent pulmonary hypertension, but the concurrent enhancement of vascular reactivity during sustained alkalosis results in subsequent pressor responses that exceed the vasodilator effects of alkalosis.

Clinical experience and reports dating back to the initial use of alkalosis therapy reveal that marked pulmonary vasoconstriction may occur when alkalosis therapy is weaned (1, 2). This has been generally believed to be caused by the underlying pulmonary hypertension. However, in this study, HPV was heightened after only 80 min of alkalosis in previously normal newborn piglets (Fig. 2, 200 min). Inasmuch as eNO concentrations decreased markedly after alkalosis (Fig. 4), we speculate that the enhanced HPV may have been caused by persistence of the alkalosis-induced increase in vascular smooth muscle contractility in the absence of continued enhancement of modulator synthesis. Further study of the time course and mechanisms involved in this phenomenon are required.

Several studies have shown that acidosis caused an acute increase in PVR (5, 9, 10, 20). In the current study, the increase in hypoxic PVR after 20 min of acidosis approached, but did not reach, significance (Fig. 2B). However, sustained acidosis markedly enhanced both baseline and subsequent HPV (Fig. 2). These findings differ from observations in cardiac and vascular smooth muscle cells, in which acidosis caused a transient increase in contractility, associated with an increase in cytosolic Ca2+, followed by a decrease in contractility related to several factors, including diminished Ca2+-calmodulin binding (22, 23). Although our data suggest that sustained acidosis may enhance pulmonary vascular reactivity, thus aggravating pulmonary hypertension, the design of the study precluded evaluation of the effects of changes in lung volume or airway pressure on PVR, systemic hemodynamic measurements, or lung mechanics. Therefore, we could not assess whether ventilator strategies leading to permissive hypercapnia offer therapeutically important benefits such as reduced barotrauma and increased venous return that outweigh the potential risks of increased PVR.

The systemic effects of alkalosis and acidosis were minimal in this study. Alkalosis led to a small decrease in hypoxic SAP (Fig. 3A), and hypoxia caused a decrease in SVRI in all groups (Fig. 3B). However, neither acidosis nor alkalosis significantly altered the effects of hypoxia on SVRI (Fig. 3B). The relatively minor effects of acidosis and alkalosis on systemic hemodynamics in this study may reflect the narrow range of pH used and the fact that ventilator settings were not changed under any experimental condition.

In summary, this study showed that hypocapnic alkalosis acutely reduced HPV, but preexistent alkalosis failed to prevent subsequent pressor responses. Moreover, hypoxic reactivity appeared to be enhanced after cessation of alkalosis. In contrast, acidosis led to a progressive increase in both baseline and hypoxic PVR. Although extrapolation of data from normal 2-wk-old piglets with low baseline PVR and acute hypoxia-induced pulmonary hypertension to infants or children with preexistent pulmonary hypertension of various etiologies must be done with caution, our results suggest that the efficacy of sustained alkalosis therapy may be limited and the risks of pulmonary hypertension must be considered when using ventilator strategies that result in permissive hypercapnia. Further studies of the effects of alkalosis and acidosis in an animal model of preexistent pulmonary hypertension and clinical trials evaluating the safety and efficacy of alkalosis therapy and permissive hypercapnic acidosis in the treatment of infants and children are needed.