This study showed that it is feasible to use EIT to visualise regional changes in OD, CL and homogeneity of ventilation distribution during different ventilator settings in patients with ARDS and control patients in a clinical setting. These indices may help the physician decide on ventilator settings in an individual patient. Optimal PEEP based on GI showed a significant difference compared to optimal PEEP based on ODCL and best Cdyn for both ARDS and CTS patients.
No different response of ODCL and GI between patients with injured lungs (ARDS) and patients with healthy lungs during PEEP changes was observed. However, the difference in the increase of Cdyn in control patients was more pronounced. This might be explained by the fact that healthy lungs, in this case, CTS patients, are more susceptible to alveolar recruitment due to the prevalence of postoperative atelectasis . Moreover, ARDS patients had a trend towards a lower Cdyn. This may explain the larger increase of OD compared to the decrease in CL in these patients. The decrease in Cdyn with increasing PEEP may indicate evident OD with the higher PEEP levels in these patients.
End expiratory impedance distribution varies widely among patients, especially in ARDS. Franchineau et al. found a broad variability in optimal PEEP in patients with severe ARDS under extracorporeal membrane oxygenation. In these patients, they found no PEEP level combining no CL and no OD, meaning that there is always a combination of a certain level of CL and OD . No single optimal PEEP level exist for the whole lung . This reinforces the need for personalised titration, not only of PEEP level but also tidal volume because the latter also contributes to tidal recruitment and OD [22, 23]. Therefore, EIT might be a helpful tool because it can estimate regional OD and CL percentages in individual patients.
It has been suggested that the optimal PEEP level is determined when air is most homogenously distributed in the lung . Homogenisation of ventilation distribution somewhat became synonymous with a lung-protective ventilation strategy, assuming that recruited alveoli improve ventilation distribution as part of the tidal volume, thereby minimising OD as well [24, 25] We found that increasing airway pressure enhances homogenous ventilation distribution in ARDS and control patients. Nevertheless, the most homogenous ventilation resulted in the largest amount of OD in both groups. In ARDS, the increase in homogeneity with increasing airway pressure was less pronounced in severe ARDS compared to mild and moderate ARDS, suggesting that higher pressures are probably required for more pronounced homogeneous distribution of ventilation, but at the price of a higher risk of OD . Using the GI solely to adjust PEEP settings in ARDS can lead to severe OD and barotrauma . Solely trying to minimise inhomogeneity without limiting the upper level of PEEP may lead to severe OD and can be harmful . In fact, some heterogeneity in ventilation distribution is physiologic and thus occurs in healthy subjects as well . Hochhausen et al. used the combination of GI and ODCL for PEEP setting. Optimal PEEP was defined when GI had the lowest value, provided ODCL was ≤ 10% . Combining these EIT indices may lead to a feasible and safe PEEP setting. Furthermore, we found that mainly in ARDS, the more homogeneously ventilation was distributed, the lower Cdyn was. This may explain the increase in OD. Maybe the GI helps identify responders and non-responders to alveolar recruitment. We hypothesise that in patients who do not respond to alveolar recruitment, GI does not change, while in responders, the GI improves due to alveolar recruitment, inducing changes in ventilation distribution [31–33]. The GI is highly correlated with lung recruitability. The percentage of recruitable lung regions decreases when the GI decreases [34, 35]. The lack of improvement in GI following a PEEP increase indicates a negative response to PEEP and warns against high PEEP levels.
To this point, there is no agreement on the gold standard method for PEEP titration. A commonly used bedside tool to determine optimal PEEP is the PEEP level with the best Cdyn during a decremental PEEP trial . We found that optimal PEEP according to the best Cdyn was not in agreement with the lowest value in GI (Fig. 4). Therefore, we consider the GI solely is not suitable for optimal PEEP determination. Simultaneously, the single value Cdyn does not reflect the PEEP dependent changes in respiratory system mechanics (i.e., CL, OD and atelectrauma) . Dynamic respiratory system compliance is an average over the whole tidal volume and does not give any information on the regional ventilation distribution . Our results show that the application of Cdyn may lead to an underestimation of the optimal PEEP level in patients with ARDS. The optimal PEEP based on the best Cdyn in ARDS in this paper may be erroneous lower in cases no plateau for Cdyn was observed in the decremental PEEP phase. Dynamic respiratory system compliance could increase by further decreasing PEEP, which could not be assessed using our protocol. In this retrospective analysis, PEEP was set according to EIT information and not based upon best Cdyn. The advantage of EIT over Cdyn is that it can identify the level of PEEP where derecruitment begins, even if global Cdyn still increases due to some relief of OD, which can also be visualised with EIT . Also, Franchineau et al. showed that PEEP set at the best Cdyn did not necessarily correspond to the PEEP level with the lowest level of CL and OD in patients with ARDS . Global respiratory mechanics parameters like Cdyn for PEEP titration underestimate measures of regional ventilation distribution. Hyperinflation exists when the best respiratory system compliance is used for PEEP titration [38, 39]. A low ODCL does not exclude the presence of OD or CL, most notably if ventilation is heterogeneously distributed.
There is a strong tendency to personalise ventilator settings, particularly in conditions like ARDS where lung damage is heterogeneously, and large individual differences exist across patients . However, from a physiological point of view, an individual approach to select the level of PEEP and tidal volume to the patient's specific lung mechanics seems reasonable . Furthermore, individual responses to PEEP and tidal volume should be expected in a patient with ARDS and in patients with healthy lungs .