The present study evaluates a novel CDH treatment protocol that maximizes use of lung protection strategies immediately after birth. The results show a dramatic increase in survival rates with a decrease in ECMO requirement to zero.
The goal of MLP is to avoid VILI, which has been shown to be the main cause of death in CDH patients[7]. VILI may begin in the delivery room (DR), caused by the initial positive pressure ventilations [9]. Tidal Volume (TV) is hard to evaluate in the delivery room (DR) and even “standard” TV’s have the potential of causing volutrauma to heterogeneous hypoplastic lungs characteristic of CDH [10]. Moreover, the ideal tidal volume for each individual neonate with CDH is unknown and depends on severity; even low conventional ventilation pressures may actually deliver excessive regional tidal volumes (TV’s). For this reason, we believe that initiation of HFOV immediately after intubation in the delivery site, using a relatively low MAP and low HFTV (around 1.5 ml/kg) at a frequency of 15 Hz has the potential to minimize volutrauma and protect the lungs
The appropriate ventilation mode for neonates with CDH is controversial. The VICI trial, which was a multicenter prospective randomized clinical trial found HFOV for CDH to be inferior to conventional ventilation (CV) in some of the study’s secondary outcomes [11].
But the HFOV settings used in the VICI trial were different than the ones used in the current study: MAP 13–17 vs. 9–11 cm H2O, frequency 10–12 vs 15 Hz. In case of hypercarbia, the VICI trial reduced frequency or increased amplitude and in case of hypoxia, increased MAP or FiO2. Contrary to this approach, to minimize VILI, we monitor DCO2 trend and try to maintain it stable by only modifying amplitude and we prefer to increase FiO2 rather than MAP. Furthermore, there are other important differences: in the VICI trial, the median o/e l-h ratio was 47% vs. 25% in the MLP group of the current study, which is consistent with less severe disease in the former; gestational age (median (IQR)) was higher: 38 (37.3–39.0) vs 37 (36.5–38) and ECMO rates were about 50% vs. 0% in the MLP group of the present study. In fact, mortality in both groups of the VICI trial [23% for the CV group and 31% in the HFOV group (NS)] was higher than in the current study. Other investigators using HFOV reported similar results to the ones of the current study, however the percent of severe CDH ranged between 18–30%, much lower than in the present study [12–14].
Our MLP protocol uses relatively low MAP and higher frequencies. Our initial ventilator settings have the physiological rational of achieving adequate alveolar ventilation and oxygenation, at a minimal pressure cost, to minimize VILI [15]. When oxygenation is inadequate, we increase FiO2 trying to refrain from high MAP’s which, by compressing the pulmonary vessels, increase pulmonary vascular resistance aggravating pulmonary hypertension [16]. Lower frequencies (10–12 Hz) deliver higher HFTV and have the potential of increasing volutrauma. The use of frequencies above 15 Hz has been demonstrated to reduce lung injury in animal models of RDS. Gonzalez Pacheco N et al compared CV to HFOV at 10 Hz and 20 Hz [17]. Histological examination of the lungs showed the most severe signs of lung injury in the 10 Hz group; the 20 Hz group had the lowest grade of injury. The same group published outcomes of premature patients after introducing a new bundle for respiratory care using frequencies higher than 15 Hz as compared to the previous epoch [18]. They found that the second epoch was associated with an increased rate of survival free bronchopulmonary dysplasia (OR 2.28; CI 95% 1.072–4.878).
As higher frequencies generate smaller HFTV’s, amplitude must be increased to maintain the same DCO2. At first glimpse, this could be interpreted as counterproductive. However, the values of amplitude are measured at the level of the ventilator’s flow sensor, but pressure damps down as the air flows along the endotracheal tube and airways. At higher frequencies this damping is much higher; therefore, the pressures reaching the alveolar compartment are inferior and expected to be less harmful as compared to lower frequencies. Zannin et al. demonstrated that greater attenuation of oscillatory pressure at higher frequencies protects from barotrauma [15].
Timing of repair
Timing for surgery is controversial. The advocates of delayed surgery claim that pulmonary hypertension has to be controlled before surgery because of the potential for destabilizing the patient during the procedure[19]. Conversely, a recent publication showed no difference in survival between early vs late surgery, but later repair was associated with longer time to reach full oral feeds, increased post-repair ventilator days, and increased need for tube feeds and supplementary oxygen at discharge CDH patients often present a “honeymoon period” during which there is adequate ventilation and oxygenation; this means that their lungs, in spite of being hypoplastic, have enough reserve[20]. We speculate that the end of this “honeymoon period” is caused in part by the time it takes for the iatrogenic lung injury to become clinically apparent. Increasing ventilator support to improve oxygenation and ventilation requires using higher pressures, which further compress the pulmonary vessels and increase pulmonary vascular resistance and pulmonary hypertension. This vicious circle leads to additional impairment of the respiratory condition and VILI [21 − 2]. During the pre-surgery period we use minimal sedation allowing for spontaneous breathing that contributes to gas exchange at a lower ventilation cost. Early surgery under HFOV during the honeymoon period, has the potential to decrease lung injury. In the present study this was possible for the vast majority of neonates and evident in the mean age at surgery of 1.4 ± 0.9 vs. 6.6 ± 10.5 days in the MLP vs. standard therapy groups. HFOV seemed to grant respiratory stability during early surgery.
The subpopulation of newborns who underwent FETO due to severe CDH and were managed with the new protocol, had an 81.3% survival rate. When adding the neonates with GA below 33 weeks who underwent FETO, survival was 65%. This result is superior to that reported by Deprest et al (36%) in a recent multicenter randomized controlled study evaluating FETO for severe CDH (p = 0.03) [6]. Furthermore, the percentage of survival in the current study is similar to the findings of the trial for moderate CDH (63%) by the same group [23].
A particularly interesting finding was the absence of pneumothorax (before surgery) in the neonates treated with the new protocol, vs. 18% in the historical cohort. We believe that HFOV from birth with no IPPV was successful in avoiding air leaks, which have been associated with high ventilator pressures [24 − 5]. Lengths of stay, duration of ventilatory support and other chronological indicators were longer in the study group. We consider that this was caused by the decreased mortality in the MLP group. Severely sick neonates treated with the new protocol who survived, needed extended periods of support.
Study strengths and limitations
The present study was performed in a single center with a clear MLP protocol. We also received support from the surgical and anesthesiology teams to aim for early surgery under HFOV. Senior neonatologists were responsible for the treatment of all neonates and strictly adhered to the protocol. Most women with antenatal diagnosis of CDH were followed in the adjacent Obstetrical High-Risk unit, allowing us to plan elective Cesarean delivery. As a result, the neonates in the present study are all inborn and there is no selection bias. Some limitations of the study should be acknowledged. The study is a retrospective study with well-known limitations and obviously there was no randomization of neonates. The proportion of severe CDH in the MLP group 17/33 (52%) was higher than in most studies because of referrals to our center for FETO procedure since 2018; actually, all neonates with severe CDH in the MLP group underwent FETO procedure as opposed to only 3 of 39 (8%) in the control group. The decrease in mortality in the MLP group may be underestimated because this group had 52% severe CDH compared to 31% in the standard treatment group. FETO procedure may, in part, be responsible for the decrease in mortality of severe CDH. Although it is not clear to what extent FETO affected the outcome in the current study, we must note that compared to the multicenter study by Deprest et al. for severe CDH, the mortality rate of neonates who underwent FETO in the present study was much lower, suggesting that it was the MLP strategy the intervention leading to significantly decreased mortality rate rather than the FETO procedure [6, 23].