INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bronchiolitis is the most frequent cause of hospitalization for respiratory infection in early childhood (1). Approximately 8% of hospitalized children will require intensive care (2), increasing to more than 30% in "high risk" infants such as those with bronchopulmonary dysplasia (BPD) or congenital heart disease (3).

The treatment for infants with bronchiolitis is largely supportive. There are no therapies demonstrated to be effective in shortening either hospitalization or length of intensive care stay in infants with established respiratory syncytial virus (RSV) infection. Ribavirin may actually increase the duration of hospitalization and mechanical ventilation in this population (4), and immunomodulators such as Palivizumab and RespiGam are not recommended for use and have not been shown to be beneficial in patients with established RSV disease (5, 6). Albuterol has been subjected to numerous studies, including some in mechanically ventilated infants with little evidence of benefit in either clinical scores or respiratory mechanics (7). Steroids are frequently used in the acute phase of the illness without convincing evidence of efficacy (13, 14). Anticholinergics are similarly unimpressive (8, 15).

Airway obstruction in bronchiolitis has a number of causes. Constriction of bronchial smooth muscle represents a relatively minor contribution; hence, the limited effect of albuterol. More important factors are thought to be mucosal edema mucous plugging and airway obstruction by epithelial cellular debris (16).

In 1978, Wohl and Chernick (16) hypothesized that alpha agonists might provide effective relief of airway obstruction caused by mucosal edema. Eight studies investigating the use of epinephrine in bronchiolitis have since been published (9, 17). Some investigators have used racemic epinephrine (9, 18, 23), in doses ranging from 2.25 mg (9) to 0.5 mg/kg (23), whereas the majority have administered l-epinephrine (17, 19- 22), in doses ranging from 0.5 mg/kg (19) to 8 mg (17). Most investigators have administered the drug by nebulization; however, in two studies parenteral administration was used either alone (22) or in combination with nebulization (21). All studies have reported the clinical response of patients to epinephrine administration, four have additionally recorded pulse oximetry values (19, 23), and in three studies an assessment of respiratory mechanics was made (9, 17, 18). All investigators documented responses in acutely unwell patients, except for one study, which tested patients in the convalescent phase of their illness (17). The majority of reports indicate an improvement in one or more of the outcome parameters, although two groups do mention some deterioration in clinical score in one patient (9) and pulse oximetry in three patients (19) after administration of epinephrine.

Patients requiring mechanical ventilation for bronchiolitis represent only a minority of cases but have the potential to benefit most from relief of airway obstruction. These patients are typically ventilated for relatively long periods, are more likely to have underlying immunologic, cardiac, or respiratory disease, and occasionally have a fatal outcome (2, 3). To date no study has reported the use of epinephrine in intubated infants or reported the results of arterial blood gas analysis. Furthermore, no studies have limited enrollments to patients with proven RSV disease, choosing instead to enroll on clinical grounds alone; the inclusion of other disorders may obfuscate responses that are limited to patients with RSV infections.

The aim of this study was to investigate changes in respiratory mechanics and arterial blood gases in response to a single nebulized dose of epinephrine in infants requiring mechanical ventilation for RSV-positive bronchiolitis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study was approved by the ethics committee of the local Area Health Service.

Patients were eligible for enrollment if they met the following criteria: (1) age < 2 yr; (2) intubated and mechanically ventilated for proven RSV infection (defined as positive immunoflourescence for RSV from nasopharyngeal secretions (Bartels Respiratory Kit; Bartels Inc., WA, Australia); (3) heart rate < 200 beats/min; (4) not treated with bronchodilators for at least 12 h; (5) informed consent obtained from one or both parents.

Infants with previous episodes of wheeze were eligible for enrollment provided this was not thought to have been due to asthma. Infants with a history of prematurity and/or chronic lung disease were also eligible for inclusion in the study. Prior to testing, a neuromuscular blocking agent was administered (vecuronium or pancuronium 0.1 mg/kg given intravenously, repeated as necessary throughout the test period). In patients who were not receiving these drugs as part of their routine therapy, neuromuscular blockers were administered at least 15 min prior to the collection of the first arterial blood gas. All patients were sedated with morphine (20 to 40 µg/kg/h) and/or midazolam (1 to 4 µg/kg/min) infusions while under neuromuscular blockade. The endotracheal tube was suctioned to clear secretions prior to the commencement of testing (at least 10 min before collection of the first arterial blood gas) and as necessary throughout the testing procedure. Arterial blood gases were collected and analyzed immediately (ABL 725; Radiometer, Copenhagen, Denmark). Arterial blood pressures and heart rates were recorded by indwelling radial or femoral arterial catheters in all patients (Spacelabs Medical, Richmond WA). Throughout the study period patients were ventilated in pressure- control mode using Siemens SV 300 ventilators (Siemens-Elema AB, Solna, Sweden). Ventilator settings were not altered during the study period, but inspired oxygen concentration (FI_O2) was adjusted if necessary to maintain Sa_O2 > 90%.

Immediately after the collection of baseline arterial blood gas and mechanics data, l-epinephrine (adrenaline acid tartrate, Astra Zeneca, North Ryde, NSW, Australia) was administered by ultrasonic nebulizer (Siemens SUN 345 with 10 ml holding chamber; Siemens- Elema AB, Solna, Sweden). All patients received 0.5 mg/kg epinephrine (undiluted, 1 mg/ml) with a maximum dose of 5 mg.

At the completion of the epinephrine nebulization, a further set of arterial blood gases were immediately collected and analyzed as above, immediately after which a final set of flow-volume curves was collected. Although the study was neither blinded nor placebo-controlled, the objective nature of the primary end points (arterial blood gases and respiratory mechanics) meant that each patient could adequately serve as his or her own control.

Oxygenation index (OI) and ventilation index (VI) were calculated using the following equations: OI = (mean airway pressure × FI_O2)/Pa_O2 and VI = (peak inspiratory pressure × rate × Pa_CO2)/1,000.

Measurement of Respiratory Mechanics

Respiratory mechanics were assessed using a single-breath occlusion passive deflation flow-volume technique (24). An automated system (Sensormedics 2600 lung function cart; Sensormedics, Yorba Linda, CA) was used for all measurements. A shutter valve occluded the airway for 200 to 300 ms at end-inspiration and the end-inspiratory pressure was measured by a transducer (Validyne MP45; Validyne Engineering Corp., Northridge, CA), after which the patient exhaled passively through a pneumotach (Model 4500; Hans-Rudolph, Kansas City, MO). The passive expiratory flow signal was integrated to obtain a flow-volume relationship, which was displayed graphically and stored numerically. A minimum of five curves was collected for each patient.

Flow-volume data points for each curve were obtained from computer memory and entered into a statistical package (SPSS Version 6.1.4; SPSS Inc, Chicago, IL) for further analysis. The flow-volume data (at incremental volume points of 0.95 to 1.00 ml for each curve) were integrated to yield a volume-time curve, which was analyzed using a single or double compartment model as described below.

Flow-Volume Curve Analysis

For flow-volume relationships that were linear over at least 50% of the exhaled volume, the volume-time data points were analyzed using nonlinear regression and a single-compartment lung model (25):

where V_(t) is volume at time t, V₍₀₎ is volume at the beginning of the passive exhalation, R is resistance of the respiratory system, and C is compliance of the respiratory system. The nonlinear regression generates values for V₍₀₎ and the time constant RC. The plateau pressure (P) at the beginning of exhalation is used to obtain compliance (C): C = V/P. Resistance (R) can then be calculated from: R = RC/C.

For many patients, flow-volume relationships were nonlinear. This phenomenon has been observed previously in patients with obstructive airway disease (26, 27). For such patients the volume-time curves were analyzed using a two-compartment model as described by Jarriel and colleagues (26). This model assumes two independent compartments of different volumes and time constants emptying in parallel. The two compartments were labeled "slow" and "fast":

In this case the nonlinear regression yields values for V_(0)fast, V_(0)slow, RC_fast, and RC_slow. As the plateau pressure at the beginning of exhalation is applied to both compartments equally, values for C and R can be obtained as for the single compartment model:

Resistance is then calculated for each compartment as indicated above using the values for C and RC.

Statistical Analysis

The majority of patients (n = 8) had nonlinear (two-compartment) flow-volume relationships both before and after administration of epinephrine. For these patients, direct paired comparisons of resistance for each compartment were performed (R_fast pre-epinephrine versus R_fast post-epinephrine, R_slow pre-epinephrine versus R_slow post-epinephrine). A further five patients had linear (single-compartment) flow-volume relationships both before and after administration of epinephrine. The single values for R in these patients were arbitrarily labeled "slow" and grouped with the slow compartment values of the eight patients with two-compartment mechanics for the purposes of statistical analysis. Two patients had nonlinear (two-compartment) curves before epinephrine administration, which improved to linear (single-compartment) curves after epinephrine. For these patients the single value of resistance obtained post-epinephrine was used for comparison, with both values of resistance (R_fast and R_slow) obtained pre-epinephrine.

For the comparison of compliance, a value for total compliance was calculated for each patient: C_total = (V_fast + V_slow) / P.

This value was used in all patients to compare pre- and post-epinephrine changes in compliance.

Each set of flow-volume curves was averaged to obtain a mean value for resistance (for each of two compartments if appropriate) and compliance and a coefficient of variation for each curve set was then calculated. Response was defined as a change from baseline (pre-epinephrine) values by more than 2 coefficients of variation (7).

All paired data were compared using Wilcoxon's Rank Sign test. Relationships between changes in parameters were analyzed using Pearson's correlation coefficients.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A total of 17 patients admitted to the intensive care unit (ICU) during the study period (May 1, 1998 to October 31, 1999) fulfilled enrollment criteria. Consent to participate in the study was not obtained from one patient, and one patient was excluded because of technical problems with test equipment; thus, 15 patients (10 male, five female) were tested over two RSV seasons (southern hemisphere winter of 1998 and 1999). The study population therefore represents 88% of all patients ventilated for RSV respiratory disease in the ICU during the period of the study. The majority of patients (nine of 15) were tested within 24 h of initiation of mechanical ventilation, with a further three tested on Day 2, two on Day 3, and one on Day 4.

In the pre-epinephrine data collection, 5 to 12 (median, 10) curves were collected and analyzed for each subject, after epinephrine, a further five to 15 (median, nine) curves were collected and analyzed for each subject. The median coefficients of variation for curves for each parameter (R_fast, R_slow, C_total) were 7.1, 9.0, and 5.8, respectively, before epinephrine and 9.1, 7.8, and 5.6, respectively, after epinephrine.

Correlation coefficients (r) for the nonlinear regressions ranged from 0.9833 to 0.9999, indicating an excellent fit of the model to the observed data. A typical nonlinear curve is illustrated in Figure 1, with observed and modeled flow-volume curves superimposed.

View larger version (11K): [in this window] [in a new window]   Figure 1.   Two compartment modeling of nonlinear flow-volume relationship. Flow (y) versus Volume (x). Solid line = observed data. Dashed line = flow-volume curve constructed from values of resistance (R), compliance (C), and volume (V ) for each of two compartments described by the mathematical model detailed in the text.

Patient demographics are described in Table 1. As expected, these patients were young infants with moderately severe respiratory disease. The acute illness represented the first episode of wheezing in 11 of 15 patients (73%); three infants had previously been labeled as having acute bronchiolitis (one influenza A-positive, two not tested for viruses), and a single infant had a previous episode of suspected aspiration. Only one patient had chronic lung disease at the time of enrollment; this infant was born at 28 wk gestation, had been ventilated for 1 wk immediately after delivery, and required supplemental O₂ for 10 wk, but was discharged home breathing room air at 12 wk. At 19 wk he contracted RSV and was enrolled in the study. All patients had evidence of airway obstruction and only four patients had baseline (pre-epinephrine) values for C_total within the normal range as defined by Hammer and colleagues (7); the remainder had low values for C_total and therefore had mixed restrictive and obstructive disease.

All patients tolerated epinephrine administration with minimal alteration in hemodynamics (Table 2). Although heart rate and systolic blood pressure increased to a statistically significant degree, these changes were not clinically significant. There was no significant change in diastolic blood pressure. Mean arterial pressure increased and just failed to reach statistical significance (p = 0.0552).

Respiratory mechanics changed substantially in almost all patients; results are summarized in Table 3. Resistance of the slow compartment changed significantly in response to epinephrine in 13 of 15 patients; in all 13 patients the response was an improvement (i.e., decrease) in resistance. Although a response was defined as a change of at least 2 coefficients of variation in magnitude, 10 of the 13 responders had a change in resistance of > 4 coefficients of variation. The magnitude of the changes was both statistically and clinically significant, with median resistance of the slow compartment falling to less than half the baseline (pre-epinephrine). Resistance of the fast compartment changed significantly in seven of 10 patients who had comparisons performed (i.e., excluding the five patients who had single-compartment mechanics before and after epinephrine) and in all seven "responders" a statistically significant improvement was noted. All patients who had a response to epinephrine in the fast compartment also responded in the slow compartment.

A significant change in C_total was observed in six patients; four patients demonstrating an improvement (ranging from 38.5 to 114.3% change over baseline) after epinephrine, whereas in two patients epinephrine led to a decrease in C_total. In only one of these patients was the change potentially clinically significant (a 29.1% fall in compliance); the other patient had only a 6.9% decrease. Taken as a group, these changes were not statistically significant, although the number of patients analyzed was small (Table 3). All patients who demonstrated a significant change (either improvement or deterioration) in total compliance also showed improved resistance in at least one compartment after administration of epinephrine.

Although individual responses varied widely, no significant changes were noted in either OI or VI after administration of epinephrine when data were analyzed for the group (Table 4). Individual responses for OI ranged from a decrease (improvement) of 59.9% to an increase (deterioration) of 114.0% and for VI from a decrease (improvement) of 14.7% to an increase (deterioration) of 37.9% of baseline. For OI, nine of 15 patients improved, whereas six of 15 deteriorated, and for VI, eight of 15 improved, six of 15 deteriorated, and one of 15 had no change. Individual changes in OI and VI after epinephrine are detailed in Figure 2.

View larger version (19K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Figure 2.   Individual patient responses in oxygenation index (Top panel ) and ventilation index (Bottom panel ) before and after administration of epinephrine with changes in compliance and resistance also indicated as follows: open circles = no change in compliance after epinephrine; crosses = compliance decreased after epinephrine; open triangles = compliance increased after epinephrine; thin line = resistance improved (decreased) in one or both compartments after epinephrine; bold line = no change in resistance after epinephrine.

Patients who responded positively to epinephrine with an improvement in either R_slow (n = 13), R_fast (n = 7), or C_total (n = 4) after treatment did not show statistically significant improvements in either OI or VI when these three subgroups were analyzed separately. Similarly there was no significant correlation observed between changes in R_fast, R_slow, or C_total and changes in OI or VI in those subgroups of patients who responded positively or negatively to epinephrine in each of these parameters.

A substantial number of study patients (five of 15) had received either nebulized (2), oral (1), or parenteral glucocorticoids (2) during their acute respiratory illness before being admitted to the ICU. Administration of steroids did not effect response to epinephrine. Both R_slow nonresponders, two of three R_fast nonresponders, and three of nine C_total nonresponders had received steroids.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has demonstrated that the majority of patients requiring mechanical ventilation for RSV-positive bronchiolitis experience a substantial improvement in respiratory system resistance after administration of nebulized epinephrine. Using a two-compartment model, slow compartment resistance improved in 13 of 15 patients (87%), whereas fast-compartment resistance improved in seven of 10 patients (70%). Total compliance was unaffected in the majority of patients (nine of 15) but improved substantially in four of 15 patients, whereas two patients demonstrated decreases in compliance after epinephrine, which was potentially clinically significant in one (this patient also had a marked improvement in R_slow after epinephrine).

The patients enrolled in this study are typical of those encountered by most pediatric intensive care units. They were young (all except one being less than 12 mo of age) and, with the exception of one infant with mild BPD, did not have underlying lung disease or airway hyperreactivity. Only four of 15 patients had experienced previous episodes of wheeze, and these episodes were clearly attributable by aspiration in one case, influenza A infection in another, and in the remaining two cases by illnesses highly suggestive of viral bronchiolitis (although viral testing had not been performed).

Respiratory mechanics in patients with bronchiolitis treated with epinephrine have been assessed in only three previously published studies (9, 17, 18). Two studies reported some improvements in mechanicsin tidal flow-volume loops (18) and inspiratory and expiratory dynamic lung resistance (9)whereas one (17) demonstrated no change in respiratory system resistance measured by forced oscillation. There may have been technical problems with the measurement of mechanics in this last study, however, as the investigators did not consider potential effects of phase shifts between pressure and flow (16).

As our patients were intubated, mechanically ventilated, and under neuromuscular blockade, it was not possible to assess respiratory effort; however, an improved resistance and/or compliance would be expected to be reflected in an improved "clinical score" in spontaneously breathing patients. In six of the eight studies previously reported an improved clinical score after treatment with epinephrine has been noted (9, 18, 19, 21), whereas two studies reported no change in clinical assessment (17, 20). Sanchez and colleagues (9) reported one of 24 patients who had a clinical deterioration after inhalation of epinephrine.

Despite the substantial improvement in airway resistance in the majority of patients, there was no significant change in either oxygenation or ventilation indices in the group when analyzed as a whole (although certain patients demonstrated marked improvement or deterioration in both indices). Although arterial blood gas determinations have not previously been reported in studies of epinephrine in bronchiolitis, several investigators have reported changes in oxygen saturation with mixed results. Reijonen and colleagues (21) found no change in oxygen saturation using a dose of 0.9 mg/kg racemic epinephrine administered via a jet nebulizer. Kristjánsson and colleagues (23) demonstrated a small improvement in Sa_O2 measured immediately after inhalation of approximately 0.5 mg/kg racemic epinephrine. The improvement in saturation was statistically significant but amounted to a clinically insignificant change of around 2% (interpreted from graphical data; numeric data not published). In this study, transcutaneous PO₂ also improved immediately and at 30, 45, and 60 min after inhalation. Menon and colleagues (20) demonstrated statistically (but probably not clinically96% versus 94%) significantly higher oxygen saturations in patients treated with epinephrine than in those treated with salbutamol 1 h after inhalation but did not comment on the values compared with those taken before treatment with either agent. Rusconi and Sideri (19) reported a deterioration in Sa_O2 (of 1.2 to 4.6%) in three of 9 patients studied during administration of 0.5 mg/kg of nebulized l-epinephrine. Values for the other six patients were not detailed; however, all nine patients, including the three who desaturated after epinephrine, had an improvement in clinical score, indicating that epinephrine may have conflicting effects on respiratory mechanics and oxygenation.

There are several possible explanations for the unimpressive response in oxygenation and ventilation indices. Epinephrine is well known to increase oxygen consumption and CO₂ production. Newth and colleagues (28) demonstrated a 13.2% increase in O₂ consumption in monkeys after a 10-min nebulization of 1 mg epinephrine via a jet nebulizer. It is possible, indeed likely, that our patients experienced substantial increases in metabolic rate during therapy. Although there are no dose- response data in the literature regarding epinephrine and oxygen consumption, our patients received substantially higher doses of epinephrine than did Newth's monkeys (which had a mean weight of 9.1 kg). Our patients received 0.5 mg/kg (maximum dose of 5 mg) epinephrine administered by ultrasonic nebulization, which is likely to result in substantially greater delivery of the drug when compared with jet nebulization (29). We would therefore expect increases in metabolic rate of at least the same magnitude (and possibly much greater) as those observed in the study of Newth and colleagues (28). Thus any improvement in respiratory mechanics and ventilation may well have been negated by the increased metabolic demands of the patient and therefore not reflected in the arterial blood gas determinations.

It may be that the dose-response relationships for metabolic rate and for effects on respiratory mechanics are different; thus, it may be possible to administer a dose of epinephrine that improves respiratory mechanics while having a minimal effect on metabolic rate. It is also not clear why there was such a large individual variation in oxygenation and ventilation indices, with some patients demonstrating a significant improvement, whereas others experienced substantial deterioration.

Another possible explanation for the lack of improvement in arterial blood gases is that epinephrine simultaneously relieved airway obstruction but at the same time caused vasoconstriction in the pulmonary vasculature, increasing ventilation/perfusion mismatch. The majority of our patients had mixed restrictive and obstructive disease, and at enrollment all required increased FI_O2 to maintain acceptable oxygenation. Desaturation caused by ventilation perfusion mismatch has been described with the use of salbutamol in young infants with RSV-positive bronchiolitis (12). As epinephrine is both an alpha and a beta adrenergic agonist it may well be expected to have a similar or even greater effect.

It is noteworthy that our patients did not have purely obstructive airway disease; the median C_total was < 50% of normal, and only four of 15 patients had normal values for C_total at baseline. Although this might represent loss of compliance because of hyperinflation, we believe it is more likely to be a manifestation of restrictive lung disease coexisting with airway obstruction. Although we did not measure lung volumes in this cohort of patients, several studies have demonstrated that in ventilated infants with RSV-positive bronchiolitis and significant airway obstruction FRC is normal, whereas TLC and C_rs are substantially reduced, indicating a significant restrictive component to the disease (7, 30, 31). Furthermore, compliance improved significantly in only a minority of patients (four of 15) after epinephrine therapy, suggesting that the low compliance was not a result of hyperinflation in the majority of patients. If the low compliance was due to hyperinflation we would expect to see a more widespread improvement after administration of a drug that substantially improved airway resistance in 13 of 15 patients. We can reasonably assume therefore that our patients did have significant restrictive lung disease, which substantially contributed to their need for mechanical ventilation. If this is the case, the beneficial effect of epinephrine on airway resistance is unlikely to impact substantially on the disease course in the ICU. Epinephrine may potentially have a more beneficial effect on gas exchange in patients with purely obstructive disease; such patients are more likely to be managed out of the ICU.

Although some investigators cite the theoretical hemodynamic advantages of racemic epinephrine over the l-isomer (reduced cardiovascular effects secondary to the d-isomer blocking the beta₁ stimulation) (32), most reports indicate that l-epinephrine is well tolerated, and this was the case in our patients. Changes in heart rate observed in this study were of less magnitude than those observed in a similar population of children given salbutamol by MDI (7). Menon and colleagues (20) also noted that infants receiving 3 mg nebulized epinephrine had a significantly lower heart rate 90 min after inhalation than did a similar group receiving 1.5 mg nebulized salbutamol (165 ± 13 versus 151 ± 15, p = 0.003).

Despite the dramatic effect on respiratory mechanics, we are unable to recommend nebulized epinephrine as a routine therapy in patients with severe RSV-positive bronchiolitis at this time. Although some patients respond with an improvement in mechanics and oxygenation and ventilation indices, a significant number show deterioration in OI and VI after treatment despite improvement in their mechanics. If nebulized epinephrine is to be used it must be on a case-by-case basis, with close observation of the individual response. Different doses of epinephrine may have relatively different effects on mechanics and OI or VI, and it may be that a lower or higher dose gives a more consistent and beneficial response. This area is worthy of further investigation as the drug clearly has the potential to dramatically improve some patients. It should also be noted that the responses to epinephrine in patients with less severe illness (who are more likely to have a purely obstructive airways problem rather than the mixed restrictive/obstructive disease seen in the majority of our patients) may be quite different from those observed here. In 1963, well before the advent of modern intensive care units, Reynolds and Cook (33) noted "oxygen is vitally important in bronchiolitis and there is little convincing evidence that any other therapy is consistently or even occasionally useful." This observation appears to be almost as valid currently as it undoubtedly was in 1963.