Inhaled nitric oxide (NO), a selective vasodilator, improves oxygenation in many patients with adult respiratory distress syndrome (ARDS). Vasoconstrictors may also improve oxygenation, possibly by enhancing hypoxic pulmonary vasoconstriction. This study compared the effects of phenylephrine, NO, and their combination in patients with ARDS.
Twelve patients with ARDS (PaO2/FIO2 <le> 180; Murray score <me> 2) were studied. Each patient received three treatments in random order: intravenous phenylephrine, 50-200 micrograms/min, titrated to a 20% increase in mean arterial blood pressure; inhaled NO, 40 ppm; and the combination (phenylephrine+NO). Hemodynamics and blood gas measurements were made during each treatment and at pre- and posttreatment baselines.
All three treatments improved PaO2 overall. Six patients were "phenylephrine-responders" (delta PaO2 > 10 mmHg), and six were "phenylephrine-nonresponders." In phenylephrine-responders, the effect of phenylephrine was comparable with that of NO (PaO2 from 105 +/- 14 to 132 +/- 14 mmHg with phenylephrine, and from 110 +/- 14 to 143 +/- 19 mmHg with NO), and the effect of phenylephrine+NO was greater than that of either treatment alone (PaO2 from 123 +/- 13 to 178 +/- 23 mmHg). In phenylephrine-nonresponders, phenylephrine did not affect PaO2, and the effect of phenylephrine+NO was not statistically different from that of NO alone (PaO2 from 82 +/- 12 to 138 +/- 28 mmHg with NO; from 84 +/- 12 to 127 +/- 23 mmHg with phenylephrine+NO). Data are mean +/- SEM.
Phenylephrine alone can improve PaO2 in patients with ARDS. In phenylephrine-responsive patients, phenylephrine augments the improvement in PaO2 seen with inhaled NO. These results may reflect selective enhancement of hypoxic pulmonary vasoconstriction by phenylephrine, which complements selective vasodilation by NO.
Adult respiratory distress syndrome (ARDS) is a syndrome with diverse causes, characterized by profound hypoxemia, pulmonary hypertension, and poor lung compliance.  Pathologic changes in the lung include alveolar hemorrhage, fluid accumulation, and interstitial-alveolar fibrosis. Pulmonary vascular changes include progressive thickening of arterial wall muscle, thrombosis, narrowing, compression, and occlusion. [2,3] The lung parenchymal and vascular involvement is heterogenous: there are completely consolidated regions and normally ventilated ones, with similar variation in perfusion. Not surprisingly, intrapulmonary shunt and dead space fractions are high. Hypoxemia in these patients is primarily a result of shunting through consolidated lung. 
Inhaled nitric oxide (NO), a locally acting vasodilator, selectively dilates pulmonary vessels in ventilated lung regions. This reduces shunt (QS/QT) and improves arterial oxygenation (PaO2) in patients with ARDS. [5–9] Conversely, nonselective vasodilators tend to increase QS/QTand decrease PaO2, presumably by opposing hypoxic pulmonary vasoconstriction (HPV) in the unventilated lung regions. [5,10]
Almitrine bismesylate, a selective pulmonary arterial vasoconstrictor, decreases QS/QTand increases PaO2in patients with ARDS, with a tendency to increase pulmonary artery pressure (PAP). These data suggest that almitrine increases HPV, diverting flow away from the most hypoxic lung regions. [11,12] Studies of almitrine and inhaled NO together suggest that the two drugs have additive effects on PaO2, perhaps because they have complementary mechanisms of redirecting blood flow away from hypoxic lung regions. [13–16]
Theory  and preliminary animal studies  suggest that even a nonselective vasoconstrictor, given alone, may improve PaO2in the setting of significant shunt. Such an effect would reflect increase of HPV. [17,18] Further, inhaled NO would reverse local vasoconstrictor, such that the combination may reduce shunt, and increase PaO2more than either drug alone.
Phenylephrine, an alpha-receptor agonist, has pulmonary and systemic vasoconstrictor effects.  Phenylephrine is nonselective (compared with almitrine) and is commonly used in intensive care. We investigated the response to intravenous phenylephrine and to inhaled NO, separately and combined, in a group of patients with ARDS, to determine whether phenylephrine alone would improve PaO2and whether the combination of phenylephrine and inhaled NO would have an additive effect on PaO2.
Based on previous studies with other drugs, we expected substantial variability in response to phenylephrine. It is already known that not all patients with ARDS respond to inhaled NO with increased Pa sub O2, decreased QS/QT, or increased PAP. [9,20–25] Similarly, not all patients with ARDS respond to almitrine. [11,12,14,16] We sought to identify patients who responded to phenylephrine and to examine responses relative to patient variables such as baseline physiologic values and severity and duration of disease.
Materials and Methods
The study was carried out with approval from the Food and Drug Administration (FDA) for experimental drug use (NO) and with approval from the Institutional Human Studies Review Board. Written informed consent was obtained for studies in 12 patients in medical and surgical intensive care units (ICUs). Inclusion criteria were 1) ARDS, defined as PaO2/FI sub O2 < 180, infiltrates on chest radiograph, and Murray Score <me> 2.0,  and 2) presence of arterial and pulmonary artery catheters. Exclusion criteria were 1) age < 18 yr, or 2) hemodynamic instability requiring treatment with any vasopressor or vasodilator medication. Dopamine infusion in renal vasodilator doses (< 3 micro gram [center dot] kg sup -1 [center dot] min sup -1) was accepted, as long as the infusion continued unchanged throughout the study period.
Each patient received all three treatments in random order: intravenous phenylephrine, 50–200 micro gram/min, infused to achieve a 20% increase in mean arterial blood pressure (mABP); inhaled NO, 40 ppm; or both together (phenylephrine + NO). Twenty minutes of equilibration during each treatment condition preceded each set of measurements. Baseline periods of equal duration preceded and followed each treatment period, with 20-min equilibration periods before measurements. FIO2was 0.90 for the entire study. Only one baseline period occurred between treatments; for example, if phenylephrine followed NO treatments, the “postphenylephrine” baseline was also the “pre-NO” baseline, and so on.
Phenylephrine was administered in saline (200 micro gram/ml), via a standard intravenous infusion pump. NO was delivered from a tank containing 800 ppm NO (BOC Gases, Port Allen, Louisiana). Stock NO was diluted with medical grade nitrogen (N2, Puritan-Bennett, Boston, MA) using an air-oxygen blender (Ohmeda, Austell, GA); the gas mixture was delivered to the high-pressure air inlet of an ICU ventilator (Puritan-Bennett 7200, Puritan Bennett, Boston, MA), as described elsewhere.  Waste gas was directed to a scavenging container (Boehringer, Laboratories, Wynnewood, PA) and from there evacuated to the hospital's waste gas system. NO and NO2concentrations were continuously monitored from the inspiratory limb of the ventilator, using an aspirating dual-channel chemiluminescence analyzer (Eco Physics CLD 700AL, Eco Physics, Ann Arbor, MI). NO2concentrations never reached 5 ppm (OSHA limit ).
Measurements were made using standard ICU monitors: electrocardiogram, arterial cannula, pulmonary artery catheter with thermodilution cardiac output capability, and central venous catheter. Pressure transducers were connected to a Hewlett Packard ICU monitor (Hewlett Packard, San Diego, CA). Values for heart rate (HR), blood pressures (BP), pulmonary artery pressures (PAP), and central venous pressures (CVP) were simultaneously recorded every minute automatically; five consecutive measurements were averaged to obtain values for each period. At the same time, respiratory rate, peak pressures, and PEEP were recorded from the ventilator's spirometer. Four consecutive measurements were averaged to obtain values of each of these quantities. Minute ventilation (VE) was counted over 1 min. Expired gas passed through a baffled 5-l chamber, which ensured complete mixing (PCO2independent of respiratory cycle) with equilibration well within each 20-min stabilization period.  Mixed expired carbon dioxide (PECO2) was measured by an aspirating capnograph (Ohmeda 5200) at the outlet of the chamber.
After hemodynamic and ventilatory measurements were complete (5 min), one arterial and one mixed venous blood sample were drawn to obtain PO2, PCO2, pH, O2saturation (SO2), hemoglobin (Hb), and methemoglobin (metHb). The two Hb determinations were averaged to obtain Hb. Four thermodilution cardiac output (CO) determinations were then performed at random throughout the respiratory cycle. These values were averaged. Finally, pulmonary artery occlusion pressure (PAOP) was determined once at end-expiration, concluding the measurements for each period.
Systemic and pulmonary vascular resistances and oxygen consumption (VO2) were calculated using standard equations. Carbon dioxide production (VCO2) was determined as VEx FECO2, and the resulting value was used to determine respiratory quotient, R (R [equivalence sign] VCO2/VO2). R was used to calculate PaO2and hence QS/QT, using standard equations. Of note, the contribution of R to QS/QTis small at high FIO2, so that substituting R = 0.8 for “measured” R did not change calculated QS/Q sub T by more than 1%.
Physiologic data for the various baselines and treatments were compared using analysis of variance (ANOVA) for repeated measures, with pairwise comparisons by the Newman-Keuls method. Other between-subgroup comparisons were performed using unpaired t tests (for parametric data, such as initial baseline values), Wilcoxon rank-sum test (for nonparametric data, namely, Murray scores), and Fisher's exact test (for binomial data, namely, survival).  In all analyses, a P value less than 0.05 was accepted as statistically significant.
Twelve patients were enrolled. As shown in Table 1, causes and severity of ARDS were variable, with Murray scores ranging from 2.00 to 3.75 (median, 3.00). Duration of ARDS at time of study ranged from 1 to 13 days. All patients had arterial hypoxemia, pulmonary hypertension, and elevated shunt fractions (QS/QT).
Effects of Treatments
Baseline data for PaO2, mPAP, and QS/QTwere grouped by period (first, second, third, or fourth baseline) and then by treatment (prephenylephrine, pre-NO, or prephenylephrine + NO baseline). Grouping baseline values by treatment required dropping the fourth baseline because it did not precede a treatment. Baseline values varied significantly by patient (two-factor ANOVA for repeated measures; P <0.0001). However, baseline values (shown in Table 2) did not vary significantly by period or by treatment (two-factor ANOVA for repeated measures).
Nitric oxide alone increased PaO2by > 10 mmHg in 11 of the 12 patients, we called these patients “NO-responders,” using a definition proposed by other investigators. [9,14] Phenylephrine alone increased PaO2by > 10 mmHg in six patients, which we called “phenylephrine-responders.” With six in each group, we were able to compare phenylephrine-responders with phenylephrine-nonresponders, as will be described.
All three treatments-phenylephrine, NO, and phenylephrine + NO-improved PaO2when data from all patients were analyzed together. Figure 1shows each patient's pretreatment, treatment, and posttreatment values of PaO2. These were analyzed by two-factor ANOVA for repeated measures, with treatment values of PaO2compared with each other and to pre- and posttreatment baselines, using the Newman-Keuls method of pairwise comparisons.  Average PaO2values with all treatments were significantly (P < 0.05) higher than baseline values. PaO2values with phenylephrine + NO were not statistically different from those with NO alone in the total patient group; both were significantly higher than PaO2values in the group receiving phenylephrine alone. Table 2shows average baseline and treatment values of PaO2.
When the patients were subdivided into phenylephrine-responders and phenylephrine-nonresponders, additional trends emerged. Figure 2shows average pretreatment, treatment, and posttreatment values of PaO2for the two subgroups; numerical data are given in Table 3. Data and calculated values were analyzed by ANOVA as described in the previous paragraph.
Within the phenylephrine-responders subgroup, PaO2increased significantly (P < 0.05) with every treatment. PaO2values with phenylephrine alone were not statistically different from those with NO alone. PaO2values with phenylephrine + NO were significantly higher than PaO2values with either phenylephrine or NO alone. Within the phenylephrine-nonresponders subgroup, PaO2did not increase significantly with phenylephrine (as per the definition of the subgroup), although PaO2did increase significantly (P < 0.05) with NO and with phenylephrine + NO. The PaO2values with phenylephrine + NO were not statistically different from those receiving NO alone.
Mean systemic arterial BP increased significantly (P < 0.05) with phenylephrine in every patient by design, as phenylephrine was titrated to a 20% increase in mABP. Within the phenylephrine-responder subgroup (as defined previously), mean pulmonary artery pressure (mPAP) increased significantly with phenylephrine. Within the phenylephrine-nonresponder subgroup, mPAP decreased significantly with NO. CO, HR, and PAOP did not change significantly with any treatment in the total patient group or within either subgroup. However, the power of this study to detect clinically significant changes in these hemodynamic quantities was low (< 50%) because of the small size of the sample. Table 2and Table 3give values of mPAP, mABP, HR, CO, and PAOP.
QS/QTdecreased with NO and with phenylephrine + NO in the total patient group and within the phenylephrine-nonresponder subgroup. Within the phenylephrine-responder subgroup, QS/QTdid not change significantly with any treatment. PVR and VD/VTdid not change significantly with any treatment in the total patient group or within either subgroup. However, as with the hemodynamic quantities discussed previously, the power of this study to detect clinically significant changes in QS/QT, PVR, or VD/VTwas very low (about 70% for QS/QT; < 30% for the others). Table 2and Table 3give values of QS/QT, PVR, and VD/VT.
Phenylephrine-responders versus Phenylephrine-nonresponders
Phenylephrine-responders required less phenylephrine (P < 0.05 by unpaired t test) to achieve the 20% systemic blood pressure increase than did the phenylephrine-nonresponders. The dose (infusion rate) of phenylephrine averaged 75 +/- 11 micro gram/min in phenylephrine-responders and 169 +/- 29 micro gram/min in phenylephrine-nonresponders. Data are mean +/- SE.
Initial (first) baseline values of measured and calculated quantities did not differ significantly (by unpaired t test) between the two subgroups; however, the study had limited power to detect differences, as noted previously. The subgroups did respond differently to treatment as described.
Phenylephrine-responders did not differ significantly from phenylephrine-nonresponders in average age, duration of ARDS at time of study, Murray scores, or survival (discharged vs. passed away in hospital). Ages and ARDS duration were compared by t test; Murray scores were compared by Wilcoxon rank-sum test, and survival data were compared using Fisher's exact test. Mean age was the same in the two subgroups. Median values of ARDS duration, Murray score, and mortality were higher in phenylephrine-nonresponders, but the differences between groups were not statistically significant. As noted with regard to other variables, the small size of the groups limited the power of the study to detect differences in these characteristics.
In a group of patients with ARDS treated with phenylephrine infusion, some responded with increases in PaO2of 10 mmHg or more comparable with the increases in PaO2achieved by inhalation of NO. These patients were identified as “phenylephrine-responders” after a definition in the NO literature. [9,14](Other investigators have defined response using a ratio, requiring an increase in PaO2to 110% of baseline; in our patient group, this yields the same subgroups as the definition used herein.) The PaO2response to phenylephrine suggests that, in certain patients, phenylephrine acts to increase HPV.
In the phenylephrine-responsive subgroup, the combination of phenylephrine + NO increased PaO2significantly more than did either treatment alone. The changes in PaO2were large, such that the study had 80–90% statistical power to detect differences of such magnitudes despite the small size of the subgroups. The result supports the hypothesis that the two agents act by complementary mechanisms to improve PaO2. According to the hypothesis, phenylephrine would increase HPV, thereby primarily affecting poorly ventilated and shunt regions, whereas NO causes pulmonary vasodilation, primarily affecting better-ventilated regions. Phenylephrine may increase HPV by a nonspecific effect of pulmonary vasoconstriction, as similar improvements in PaO2have been observed with other pulmonary vasoconstrictors. [11–18]
Phenylephrine increased mPAP significantly in phenylephrine-responders but not in phenylephrine-nonresponders, suggesting that regional pulmonary vasoconstriction may be necessary and sufficient to improve oxygenation. Calculated PVR values did not change significantly with treatments in any groups, but the statistical power to detect changes was low. Another possibility is that systemic vasoconstriction by phenylephrine leads to reduced CO and thus to reduced absolute and relative shunt flow. Lower CO may lead to decreased SvO2, which increases the stimulus for HPV. CO did not change significantly with treatments in any groups, but the power to detect changes was low, so we cannot rule out CO as a determinant of PaO2changes.
Our results provide a potential explanation of the difference between phenylephrine-responsive and phenylephrine-nonresponsive patients with ARDS. As noted previously, phenylephrine-responders manifested an increase in mPAP with phenylephrine, suggesting that the pulmonary vasculature is more able to constrict than in phenylephrine-nonresponders. Two pathologic features of ARDS could limit phenylephrine-induced pulmonary vasoconstriction: 1) extensive consolidation, i.e., preexisting maximal constriction (caused by HPV), or 2) vascular fibrosis, which limits the vessels' intrinsic ability to constrict. Average mPAP declined with NO in phenylephrine-nonresponders, suggesting that they had more reversible pulmonary vasoconstriction than phenylephrine-responders, which is consistent with maximal constriction resulting from hypoxia. We can only speculate regarding the presence of vascular fibrosis. A third phenomenon also may play a role: phenylephrine-nonresponders required significantly more phenylephrine to reach mABP endpoints, suggesting that intrinsic systemic responsiveness to phenylephrine also affects whether phenylephrine will increase PaO2in patients with ARDS.
In conclusion, we found that phenylephrine and NO can increase PaO2in patients with ARDS and that phenylephrine alone increases Pa sub O2 and mPAP in some of these patients (phenylephrine-responders). This suggests that phenylephrine-induced vasoconstriction effectively increases HPV. In phenylephrine-responders, the increase in PaO2with phenylephrine + NO was greater than that with either treatment alone, as would be expected from separate, complementary mechanisms of action.