Published ahead of print on June 8, 2006, doi:10.1164/rccm.200603-351OC
© 2006 American Thoracic Society doi: 10.1164/rccm.200603-351OC
Lung Recruitment Using Oxygenation during Open Lung High-Frequency Ventilation in Preterm InfantsDepartment of Neonatology, Emma Children's Hospital, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Correspondence and requests for reprints should be addressed to A. H. van Kaam, M.D., Ph.D., Department of Neonatology (Room H3-150), Emma Children's Hospital AMC, University of Amsterdam, P.O. Box 22700, 1100 DD Amsterdam, The Netherlands. E-mail: a.h.vankaam{at}amc.uva.nl
Rationale: Changes in oxygenation are often used to guide the recruitment procedure during open lung high-frequency ventilation in preterm infants. However, data on the feasibility and safety of this approach in daily clinical practice are limited. Objective: To prospectively collect data on ventilator settings, gas exchange, and circulatory parameters before and after surfactant therapy during open lung high-frequency ventilation. Methods: In 103 preterm infants with respiratory distress syndrome, the opening, closing, and optimal pressures were determined during high-frequency ventilation by increasing and decreasing stepwise the continuous distending pressure, defining optimal recruitment as adequate oxygenation using a fraction of inspired oxygen not exceeding 0.25. This procedure was repeated after each surfactant treatment. Measurements and Main Results: The mean presurfactant opening and optimal continuous distending pressures were, respectively, 20.5 ± 4.3 and 14.0 ± 4.0 cm H2O, with a fraction of inspired oxygen of 0.24 ± 0.04. Surfactant treatment enabled a reduction in the mean optimal pressure of almost 6 cm H2O without compromising oxygenation. Blood pressure and heart rate remained stable and no air leaks were observed during the recruitment procedures. The mortality rate and the incidence of severe intracranial hemorrhage or periventricular leukomalacia and chronic lung disease at 36 wk were comparable to previously reported data. Conclusion: Open lung high-frequency ventilation using oxygenation to guide the recruitment process is feasible and safe in preterm infants and enables a reduction of the fraction of inspired oxygen below 0.25 in the majority of preterm infants with respiratory distress syndrome.
Key Words: air leaks • chronic lung disease • respiratory distress syndrome • surfactant Animal studies have shown that high-frequency ventilation (HFV) optimizes gas exchange and attenuates ventilator-induced lung injury compared with conventional positive-pressure ventilation (1, 2). On the basis of the assumption that this reduction in ventilator-induced lung injury might also reduce the incidence of chronic lung disease, HFV has been extensively investigated in preterm infants with respiratory failure. However, metaanalyses of the randomized controlled trials (RCTs) comparing HFV with conventional positive-pressure ventilation have shown at most a small but inconsistent reduction in chronic lung disease in favor of HFV (3, 4). It has been suggested that this disappointing efficacy of HFV to reduce the incidence of chronic lung disease is, in part, caused by the fact that some RCTs did not apply an "optimal lung volume" or "open lung" strategy during HFV (5). Animal studies have clearly shown that recruiting and stabilizing collapsed alveoli are essential in the process of attenuating ventilator-induced lung injury during HFV (1, 2). One of the major "handicaps" in optimizing lung volume during HFV in preterm infants is the inability to accurately measure direct changes in lung volume at bedside (6). This is why most studies exploring HFV in preterm infants have used changes in oxygenation as an indirect parameter for changes in lung volume, assuming that recruitment of collapsed alveoli will reduce the intrapulmonary shunt fraction and thus improve oxygenation (7). In theory, achieving an optimal lung volume (i.e., optimal recruitment) would enable ventilation with little or no supplemental oxygen while still maintaining adequate oxygenation. However, data on the feasibility of using oxygenation as an indirect tool for lung recruitment during HFV in daily clinical practice are limited. Most RCTs on HFV in preterm infants report data only on mean airway pressure and oxygenation several hours after lung recruitment and surfactant therapy. In addition, there is no information on changes in circulatory parameters during the actual recruitment procedure, information that seems to be important in light of the ongoing concern that HFV might increase the risk of intraventricular hemorrhages (8). The present study therefore prospectively collected data on ventilatory and circulatory parameters and gas exchange in preterm infants with neonatal respiratory distress syndrome (RDS) ventilated with primary HFV. HFV was combined with a standardized open lung strategy using oxygenation as the indirect parameter to guide changes in airway pressure during the recruitment procedure. In addition, we determined the influence of exogenous surfactant on ventilatory pressures and oxygenation. Some of the results of this study have been reported previously in abstract form (9, 10).
See the online supplement for a detailed description of methods.
Patients
Ventilation Strategy
Starting at 6–8 cm H2O, the continuous distending pressure (CDP) was increased stepwise as long as transcutaneous oxygen saturation (StcO2) and transcutaneous oxygen partial pressure (PtcO2) improved. The FIO2 was reduced stepwise, keeping StcO2 within the target range (86–94%). The recruitment procedure was stopped if oxygenation no longer improved or if the FIO2 did not exceed 0.25. The corresponding CDP was called the opening pressure (CDPO). Next, the CDP was reduced stepwise until the StcO2 deteriorated. The corresponding CDP was called the closing pressure (CDPC). After a second recruitment maneuver, the optimal CDP (CDPOPT) was set 2 cm H2O above the CDPC. The pressure amplitude was adjusted to maintain the transcutaneous partial carbon dioxide pressure (PtcCO2) between 5.3 and 8.0 kPa.
Surfactant Treatment
Data Collection and Analysis In addition, the following outcome parameters were recorded: surfactant use, duration of HFV, air leaks, dexamethasone use, chronic lung disease at 36 wk postmenstrual age, intracranial abnormalities, and mortality. Statistical analysis was performed with SPSS version 12 (SPSS, Chicago, IL). Data are presented as means ± SD, unless stated differently. Serial data were analyzed with a paired t test or linear mixed model analysis, followed by a Bonferroni post hoc test. Possible associations between the presurfactant CDPOPT and patient characteristics were analyzed by multiple regression analysis. A p value less than 0.05 was considered statistically significant.
Characteristics at Baseline Five hundred and seven preterm infants were admitted to neonatal intensive care during the inclusion period. Of these, 142 required mechanical ventilation in the first 72 h after birth and were started on HFV. Twenty-four of these infants were excluded because they did not have major lung disease as indicated by a normal chest radiograph, low airway pressure (CDP  8 cm H2O), and no supplemental oxygen to maintain adequate oxygenation. In addition, 11 infants were excluded for the following reasons: lung hypoplasia (5), pleural effusion (1), severe circulatory shock (2), pneumonia (1), and severe persistent pulmonary hypertension confirmed by echocardiography (2). In four patients data on the recruitment procedure before and after surfactant treatment were not recorded by the attending physician. The remaining 103 patient were included in the present cohort and Table 1 shows the basic characteristics.
Presurfactant Recruitment As shown in Figure 2A, the mean CDPO was 20.5 ± 4.3 (range, 11–30) cm H2O, and applying this pressure resulted in a significant reduction of the FIO2, from 0.70 ± 0.27 (range, 0.27–1.0) at the start of ventilation to 0.24 ± 0.04 (range, 0.21–0.50) after recruitment. The CDPC was 12.0 ± 4.0 (range, 7–21) cm H2O, resulting in a mean CDPOPT of 14.0 ± 4.0 (range, 9–24) cm H2O. However, despite this reduction of 6 cm H2O, the FIO2 remained stable at a mean value of 0.24 ± 0.04. In 77 (75%) infants lung recruitment resulted in an FIO2 reduction to no more than 0.25 and in 98 (96%) infants to no more than 0.30.
The StcO2 improved after the first recruitment procedure (CDPO) and showed, as expected, a significant drop when determining the CDPC (Figure 2B). However, after a second recruitment with the known CDPO and setting the optimal CDP 2 cm H2O above the closing pressure, the mean StcO2 increased to a level comparable to the CDPO time point. The PtcO2 showed a similar pattern, although data on this variable were not available at all time points in some patients, because of the equilibration process and/or detachment of the sensor (Figure 2B). Despite the fact that the pressure amplitude and frequency did not change significantly over time, the mean PtcO2 showed a significant decrease from 7.7 ± 2.3 kPa, at the start of ventilation, to 6.5 ± 1.4 kPa after optimal recruitment (p < 0.001, n = 95). Despite the relatively large differences between the CDP at the start of ventilation and the CDPO, mean arterial blood pressure and heart rate did not differ significantly during the recruitment process (Figure 2C). In some patients data on mean arterial blood pressure and heart rate were not available at all time points. None of the patients required volume expansion or vasoactive medication during or in the first hour after lung recruitment. Multiple regression analysis showed a significant association between the CDPOPT and the modified oxygenation index at the start of HFV (b = 0.44, p < 0.001), but not with gestational age or birth weight. This is also illustrated in Figure 3, showing that the CDPOPT and the corresponding FIO2 did not differ between the different gestational age strata.
Postsurfactant Recruitment Ninety-nine infants (96%) received a first dose of exogenous surfactant after the presurfactant recruitment procedure. In one infant, data on the recruitment procedure performed after the first surfactant administration were not recorded. In response to this first dose of surfactant, the CDP could be significantly reduced by 5.8 ± 3.2 (range, 0–12) cm H2O, resulting in a mean postsurfactant CDPOPT of 9.3 ± 2.6 (range, 6.0–17.0) cm H2O (Figure 4A). In 26 infants (26%) the closing procedure was stopped, as oxygenation remained stable at a CDP of 6 cm H2O. Surfactant treatment also resulted in a significant reduction in the CDPO. After the first surfactant treatment and lung recruitment, the percentage of infants with an FIO2 not exceeding 0.25 increased to 87%, resulting in a mean FIO2 of 0.23 ± 0.03 (range, 0.21–0.35) after optimal recruitment.
The changes in oxygenation after the first surfactant dose were comparable to the presurfactant recruitment procedure, showing a significant drop in StcO2 and PtcO2 when determining the CDPC (Figure 4B). The decrease in PtcO2 postrecruitment after surfactant administration was less evident than during the presurfactant period (7.5 ± 1.8 vs. 7.0 ± 2.2 kPa, p < 0.01, n = 93). However, surfactant treatment did enable a reduction in the pressure amplitude (24.4 ± 0.6 vs. 22.2 ± 0.6 cm H2O, p < 0.01, n = 98). Comparable to the presurfactant period, mean arterial blood pressure and heart rate did not change significantly during the recruitment procedure after the first surfactant dose (Figure 4C). Second and third doses of surfactant were administered in, respectively, 35 (34%) and 7 (7%) infants. The pattern of changes in CDP, FIO2, oxygenation, blood pressure, and heart rate was similar to the first surfactant dose. However, the reduction in CDP after the second (3.2 ± 2.9 cm H2O) and third (2.6 ± 3.7 cm H2O) surfactant doses was less impressive than during the first dose.
Other Outcome Parameters
The median duration of the initial HFV period was 2 d and only four patients received systemic dexamethasone to facilitate extubation. Twenty-five infants developed chronic lung disease. A total of six patients died at a median postnatal age of 8 d, for the following reasons: necrotizing enterocolitis (2), cardiac tamponade (1), pulmonary hemorrhage (2), and septic shock (1). Ultrasonography of the brain revealed a grade 3 or 4 intraventricular hemorrhage in 14 infants, but no cases of periventricular leukomalacia.
The present study shows that changes in oxygenation can be used to guide the recruitment procedure during open lung HFV and that optimal recruitment allows ventilation with little or no supplemental oxygen (FIO2 0.25) in the majority of preterm infants with neonatal RDS. Furthermore, this study provides, for the first time, detailed information about the pressures needed to recruit and stabilize the lung before and after surfactant treatment. Animal studies have clearly shown that avoiding both volutrauma and atelectrauma during mechanical ventilation attenuates ventilator-induced lung injury (11). This is why optimization of lung volume has been considered an essential part of HFV when aiming to reduce the incidence of chronic lung disease in preterm infants (12). Traditionally, lung volumes in (ventilated) newborn infants have been measured by washout techniques, but this method is difficult to apply bedside and washing out the tracer gas requires conventional (tidal) ventilation (6). Most clinicians have therefore adopted oxygenation as the (indirect) bedside tool with which to assess changes in lung volume and to guide ventilator settings during HFV (12). The basic idea is that increasing lung volume in the (atelectatic) neonatal RDS lung is accompanied by alveolar/saccular recruitment, which reduces intrapulmonary shunt and improves oxygenation. In theory, optimal recruitment would enable ventilation with little or no supplemental oxygen. The open lung HFV strategy used in the present study therefore adopted a target FIO2 of 0.25 or less to define optimal lung recruitment. Assuming that each patient had a different severity of lung disease, we started the recruitment procedure at a relatively low CDP of 6–8 cm H2O, trying to minimize the risk of overdistending lungs with mild RDS. As expected, most infants needed a high FIO2 to achieve adequate oxygenation, indicating a high degree of alveolar/saccular collapse at this stage of the recruitment procedure. However, increasing the CDP stepwise, as long as oxygenation improved, resulted in a reduction of the FIO2 to 0.25 or less in the majority of the infants. To our knowledge, the mean presurfactant CDPO of almost 21 cm H2O needed to reach this low FIO2 has not been previously reported. We believe it is important to mention that this high CDPO was applied only for several minutes, because mathematical, experimental, and human studies have suggested that airway pressures can be safely reduced after lung recruitment without resulting in significant loss of lung volume (i.e., lung hysteresis) or, more importantly, derecruitment (13–16). The present study seems to indicate that hysteresis is also present in the early phase of neonatal RDS, as we could lower the CDP almost 8 cm H2O before oxygenation started to deteriorate, indicating loss of lung volume due to alveolar/saccular collapse. By once more recruiting the lung with the known CDPO and setting the optimal CDP 2 cm H2O above the CDPC, we were able to regain and maintain the optimal lung volume at the lowest possible pressure for each patient. The positive correlation between the severity of lung disease and the CDPOPT seems to support this conclusion. However, this finding should be interpreted cautiously as we did not use arterial PO2, which was not available at the start of ventilation, but StcO2 to calculate a modified oxygenation index. After lung recruitment the StcO2 stabilized at about 94% with a corresponding PtcO2 of approximately 8.0 kPa. On the basis of this PtcO2 the recruitment procedure probably did not completely reverse intrapulmonary shunt and thus atelectasis. However, we believed that a further increase in the recruitment pressure targeted on PtcO2 would increase the risk of alveolar/saccular overdistention and hyperoxia. Studies in conventionally ventilated preterm infants have shown that exogenous surfactant increases functional residual capacity (17). This is why we determined the opening, closing, and optimal CDP after each surfactant administration, anticipating improved alveolar stability. This resulted in an almost 6–cm H2O reduction in the CDPOPT after the first surfactant administration, while still maintaining adequate oxygenation with an FIO2 of 0.25 or less. This reduction in stabilization pressure after surfactant treatment is consistent with previous experimental data (18). Only a minority of the infants required a second (35%) or third (7%) dose of surfactant, which is similar to previous reports (19). The reduction in CDPOPT after the second and third dose was, however, not as impressive as after the first dose of surfactant. In the present study, we administered exogenous surfactant after lung recruitment for the following reasons. First, in our department the position of the endotracheal tube is checked by chest radiograph before administering exogenous surfactant, thereby minimizing the risk of unilateral deposition. While awaiting the results of this investigation, patients must be ventilated with conventional ventilation or low-pressure HFV, which may be injurious to the lung (1, 20). Second, there is some experimental evidence that the efficacy of (rescue) surfactant treatment is improved after lung recruitment, although there are no human data to substantiate this (21). Applying relatively high recruitment pressures, as reported in the present study, might compromise hemodynamics or increase the risk for air leaks. The results from the present study seem reassuring as mean arterial blood pressure and heart rate did not change significantly during the different steps of lung recruitment, which is consistent with previous findings (22). In addition, chest radiographs obtained after the pre- and postsurfactant recruitment procedures did not reveal air leaks. The observed incidences of other important outcome parameters such as mortality, intraventricular hemorrhages, periventricular leukomalacia, and chronic lung disease are difficult to interpret as this was an observational and not an intervention study. Comparison with previous RCTs on HFV in preterm infants should be done cautiously, as many variables (gestational age, birth weight, and dexamethasone use) affecting the described outcome parameters differed to some extent with our study. Comparing the present study with RCTs on HFV that included a similar population showed a lower or comparable incidence of the above-mentioned outcome parameters, indicating that the open lung strategy used in the present study was probably not associated with an increased incidence of more long-term adverse effects (23, 24). The present study has several limitations that need to be addressed. First, we did not measure actual changes in lung volume during the recruitment procedure, so we cannot confirm that lung volumes were truly optimal at CDPOPT. Newer, promising techniques such as electrical impedance tomography and respiratory inductance plethysmography should be able to address this problem in the near future (16, 25). Second, only a small group of infants included in the present study had a gestational age below 26 wk. Although we did not find a clear correlation between gestational age and the CDPOPT or corresponding FIO2, we cannot conclude that the open lung approach used in this study is also feasible and safe in this group of infants. Finally, the results of this study are applicable only to preterm infants with RDS. We excluded infants with severe and complex lung pathology, which is often accompanied by extrapulmonary right-to-left shunting and makes oxygenation a less reliable tool with which to monitor lung volume. The severity of lung disease in this group of excluded infants was also illustrated by a high mortality rate of 73%. Taken these limitations into account, we do believe the present study has important implications for clinical practice and future research on HFV. Most of the previously conducted RCTs, comparing HFV with conventional positive-pressure ventilation in preterm infants, provided only limited information about the ventilation strategy during HFV and reported only data on CDPOPT and FIO2 several hours after lung recruitment and surfactant therapy (24, 26, 27). This study provides a detailed description of an individualized open lung HFV strategy guided by changes in oxygenation and, more importantly, provides reference data on opening, closing, and optimal CDPs and corresponding changes in oxygenation and FIO2. The present study also shows that implementation of such a protocolized approach is feasible in a large neonatal intensive care unit, where patient care is guided by a large group of physicians and nurses. The fact that the changes in CDP, FIO2, StcO2, and PtcO2 during the pre- and postsurfactant recruitment procedure were done according to protocol seems to substantiate this conclusion. Factors that may have contributed to successful implementation are adequate training of both physicians and nurses in the theoretical and practical issues of open lung HFV, and the use of bedside (online) parameters for oxygenation (StcO2 and PtcO2) to guide recruitment. Although these results are especially of interest to neonatologists, they might also prove to be important to pediatric and adult intensivists, as HFV is increasingly used in the treatment of acute respiratory distress syndrome (28, 29). We would like to point out that the open lung HFV strategy as described in the present study is more time consuming and demanding than the respiratory care procedures used by clinicians generally. In conclusion, this study shows that an open lung ventilation strategy using oxygenation to guide the recruitment process is feasible and safe during HFV in preterm infants with RDS. Applying such a ventilation strategy enables a reduction of the FIO2 below 0.25 in the majority of infants after optimal recruitment.
The authors thank M. Merkus, epidemiologist from the Center for Pediatric Clinical Epidemiology, Emma Children's Hospital, Amsterdam, for statistical support.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200603-351OC on June 8, 2006 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form March 10, 2006; accepted in final form June 7, 2006
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
p < 0.001 versus start time point; 
p < 0.001 versus open and optimal time points.
