1 Introduction

Electrical impedance tomography (EIT) has been increasingly utilized in various clinical conditions, including ARDS, acute hypoxemia, [1,2,3,4] general anaesthesia, [5, 6] and postoperative cardiac surgery [7, 8].

It is an efficient, non-invasive,radiation-free, real-time bedside lung imaging technology to identify atelectasis, overdistension, and recruitment. It reconstructs impedance changes through the application of imperceptible currents and measurement of changes in voltage around the thorax, thereby allowing for individualized PEEP titration. It assesses regional respiratory system compliance (Crs) changes by identifying cyclic opening & closing, regional overdistension, and early lung recruitment changes. EIT-derived regional ventilation delay inhomogeneity (SDRVD) correlates with alveolar cycling, offering a potential tool for adjusting ventilation settings. Its ability to estimate ventilation and perfusion depending upon impedance variations and real-time assessment of regional Crs changes, identifying atelectrauma and overdistension, makes it a comprehensive tool [9,10,11,12].

Mechanical ventilation is vital in acute respiratory distress syndrome (ARDS), which has morphological features such as regional atelectasis, overdistension, and lung inhomogeneities. Cyclic opening and closing of lung units pose risks like ventilator-induced lung injury (VILI) [13,14,15].

The stress–strain concept quantifies the harm caused by high lung volumes. Adjusting positive end-expiratory pressure (PEEP) based on global Crs aims to balance overdistension and cyclic opening/closing but lacks consistent benefits, possibly because global Crs inadequately predicts recruitability, especially in atelectasis dynamics [16,17,18,19].

Studies demonstrate that EIT-guided positive end-expiratory pressure (PEEP) titration is valuable for optimizing PEEP in animals and humans [20,21,22]. However, the question remains whether EIT-guided PEEP titration can enhance ventilation and optimize global oxygenation, which requires further investigation in this specific clinical context.

Thus, the current study aimed to assess the clinical utility of EIT-guided PEEP titration compared to traditional protocols in terms of better ventilation and oxygenation in adherence to the "Preferred reporting items of systematic review and meta-analysis" (PRISMA) statement [23].

2 Methods

2.1 Literature search

The comprehensive search spanned various electronic databases (PubMed, Medline, and Embase), Google Scholar (https://scholar.google.com), preprint platforms MedRxiv (https://www.medrxiv.org), and Clinical trial database (https://ClinicalTrials.gov) until November 30, 2023, using the following keywords "EIT" OR " Electrical Impedance Tomography" AND " PEEP" OR " Positive End-Expiratory Pressure" AND " Acute Respiratory Distress Syndrome" OR "ARDS" OR "General Anesthesia" OR "Surgery". Using the above MeSH terminology, all the articles were screened & reviewed by BY and SS. Disagreements were sorted out by taking PK's opinions.

We adopted the PICO format for structuring the findings, where "P" denoted the population (Adults requiring mechanical ventilation due to ARDS), "I" referred to interventions (EIT-guided PEEP titration), "C" pertained to comparisons (conventional treatment), and "O" encompassed outcomes (P/F ratio, driving pressure, PEEP optimization, successful weaning).

2.2 Inclusion and exclusion criteria

Randomized controlled trials, controlled clinical trials, prospective and retrospective comparative cohort studies, case-control studies, comparing the clinical utility of EIT-guided PEEP titration with conventional management in patients with ARDS were included.

Case reports, narrative reviews, expert opinions, studies other than those in English, without appropriate control groups, were excluded.

2.3 Study selection and data extraction

SS and BY conducted independent screenings of abstracts to remove duplication and eliminate irrelevant articles. Eligible studies underwent full-text screening for inclusion criteria. A pre-designed data extraction sheet facilitated information collection such as the first author, publication year, study nature, country, patient count, P/F ratio, driving pressure, optimized PEEP, and successful weaning. Discrepancies were resolved through discussions with PK.

2.4 Risk of bias assessment

SS & PK independently determined any potential bias in the included RCTs using the “RoB 2.0″ [24] assessment tool, and non-randomized trials using the "Risk of Bias in Non-randomized Studies—of Interventions (ROBINS-I)” [25]. The disagreements were settled by discussing with MB.

2.5 Quality of the evidence

PK and SS evaluated each outcome individually as either "High" or "Moderate" or "Low" or "Very low" with the "Grading of Recommendations Assessment, Development and Evaluation (GRADE)" [26, 27] tool, comprising five downgrading factors: "study limitations, indirectness, imprecision, consistency of effect, and publication bias" and three upgrading factors: "dose-response relation, large magnitude of the effect, and plausible confounders or biases", and disagreements were resolved by MB.

2.6 Data synthesis

The meta-analysis was carried out by SS using Review Manager Software (RevMan V.5.4.1) & Trial sequential analysis (TSA) Copenhagen Trial Unit (Version 9.5.10 Beta, Copenhagen, Denmark) [28], The mean difference was utilized as the outcome measure for continuous variables & the log odds ratio for dichotomous variables. Outcome heterogeneity (tau2) was estimated using I2 statistic. Mild heterogeneity was considered for I2 < 30%, moderate for I2 = 30% to 70%, and significant for I2 > 70%

Sensitivity analysis was conducted by removing one study at a time, noting effects on outcome heterogeneity. Funnel plot was used to assess publication bias.

3 Results

3.1 Basic characteristics

This review comprised five controlled cohort studies [29, 30, 33,34,35] and two randomized controlled trials, [31, 32] out of 341 screened publications. (Fig. 1) While three studies [29, 32, 35] included severe ARDS patients, two of them [30, 33] included mild, moderate, and severe ARDS, and one [31] had moderate and severe ARDS patients, as per Berlin definition. One of the studies included chronic obstructive pulmonary disease (COPD) [33], and the other one incorporated coronavirus disease (COVID-19) [34]-related ARDS patients.

Fig. 1
figure 1

PRISMA-2020-Flow-Diagram

Full size image

While one study described 49% of patients required prone positioning, [30] in two studies, the prevalence was 2–4% [31, 35] in the EIT group compared to 41%, 0–3% in the control group. Two studies describe the use of inhaled nitric oxide in 54% and 66% of the patients who received EIT-guided PEEP optimization and 51.1% and 96.8% in the control group. 33.3% to 38.1% of patients in the EIT group received ECMO compared to 11.1 to 16.1% of patients in the control [31, 35].

Prevalence of tracheotomized patients ranges from 6- 11% in the control group and 14–28% in patients with EIT guided PEEP optimization group. The protocol for EIT application also varied across the studies (Table 1).

Table 1 Characteristics of included studies
Full size table

None of them had any serious concerns about the risk of bias (Fig. 2).

Fig. 2
figure 2

a ROBINS-I assessment for the included non-randomized cohort studies, b ROB-2 assessment for the included randomized controlled studies

Full size image

3.2 Meta-analysis

3.2.1 Mortality

The mortality risk was lower with EIT-guided PEEP optimization in five studies with 298 patients [Risk ratio (RR) = 0.68; 95% CI: 0.49 to 0.95; I2 = 0%; p = 0.02). (Fig. 3a).

Trial sequential analysis: The required information size was estimated at 276 for a power of 95% based on the 50% relative risk reduction for mortality. The cumulative Z-score crossed the alpha spending boundary and did not dip below the inner wedge of futility, indicating a significant effect size has been achieved. (Supplementary Fig. 1).

3.2.2 Dynamic compliance

A meta-analysis of seven studies involving 515 patients found better dynamic compliance with EIT-guided PEEP optimization by an average of 3.46 ml/cm H2O (95% CI: 1.59 to 5.34; I2 = 0%; p < 0.003) compared to the conventional methods. (Fig. 3b).

Trial sequential analysis: The required information size was estimated at 510 for a power of 95% based on the 10% improved dynamic compliance. The cumulative Z-score crossed the alpha spending boundary and did not dip below the inner wedge of futility, indicating a significant effect size has been achieved. (Supplementary Fig. 2).

3.2.3 Oxygenation

There is no substantial improvement in PaO2/ FiO2 with EIT-guided PEEP optimization compared to conventional methods, [Mean Difference (MD) = 6.5; 95%CI -13.86 to 26.76; I2 = 74%; p = 0.53] found in 515 patients of seven studies (Fig. 4a).

Trial sequential analysis: The required information size was estimated at 879 for a power of 95% based on the observed change of mean PaO2/ FiO2 by at least 10. (Supplementary Fig. 3).

3.2.4 Successful weaning

There is no substantial improvement in successful weaning incidence with EIT-guided PEEP optimization than conventional methods, [Odds Ratio (OR) = 6.5; 95%CI 0.07 to 3.01 to, I2 = 51%; p = 0.32] found in 259 patients of three studies (Fig. 4b).

3.2.5 Driving Pressure

No significant difference in the requirement of driving pressure was found in the groups of 515 patients from seven studies. [MD = -0.3; 95%CI -0.99 to 0.37; I2 = 82%; p = 0.38] (Supplementary Fig. 4a).

3.2.6 Optimized PEEP

No significant difference in the requirement of PEEP was found in the groups across 461 patients in six studies. [MD = 0.95; 95%CI -0.40 to 2.29; I2 = 96%] (Supplementary Fig. 4b).

Fig. 3
figure 3

The impact of EIT-guided individualized PEEP titration on overall mortality (a), and dynamic compliance (b) in ARDS patients

Full size image
Fig. 4
figure 4

The impact of EIT-guided individualized PEEP titration on PaO2/FiO2 ratio (a), and successful weaning (b)

Full size image

3.3 Quality of evidence

The evidence regarding the utility of EIT-guided PEEP titration in ARDS patients in terms of mortality and dynamic compliance is of low quality owing to considerable indirectness, and PaO2/FiO2, successful weaning, driving pressure, optimized PEEP, is of very low quality additionally due to inconsistency, imprecision. (Table 2).

Table 2 GRADE evidence profile of EIT-guided PEEP titration
Full size table

3.4 Publication bias

An assessment of publication bias regarding dynamic compliance suggests its unlikelihood, as indicated by the absence of Funnel plot asymmetry through the Egger's Regression and Begg & Mazumdar Rank Correlation (p = 0.98 and p = 0.56, respectively). (Supplementary Fig. 5).

4 Discussion

This systematic review found reduced risk of mortality, better dynamic compliance with no significant improvement in oxygenation and successful weaning with EIT-guide PEEP titration in ARDS patients than conventional methods.

A recent systematic review of 202 participants found a higher PaO2/FiO2 ratio [standardized mean difference (SMD) = 0.636; 95% CI 0.364 to 0.908.] with no significant change in compliance compared to alternative PEEP titration strategies [SMD = -0.085; 95% CI -0.342 to 0.172.]. However, it needed to assess the generalizability of the findings and the adequate size of the population for identifying the effect with optimum power calculation [36].

An impressive 94% positive predictive value for better oxygenation during prone ventilation was reported using electrical impedance tomography to monitor collapse in dependent lung areas [37].

Another multicentric study on COVID-19 patients reported that median EIT-based PEEP varied across groups: 10, 13.5, and 15.5 cm H2O for low, medium, and high recruit ability, respectively (P < 0.05) [38].

One of the reasons behind the conflicting outcome regarding oxygenation in our study is multifactorial. Primarily, the disease severity of included patients varied from mild to severe ARDS along with COPD and COVID-19. The use of the prone position was also not uniform and standardized. This itself can influence a wide variety of primary outcomes. ARDS is an acute hypoxemic state caused by the sudden development of diffuse injury to the terminal respiratory units with exudative pulmonary oedema. The application of PEEP to recruit the collapsed alveoli is crucial. However, determining the optimal PEEP striking a balance between preventing over-distension and de-recruitment while avoiding hemodynamic compromise remains a complex task.

Thirdly, the duration of EIT intervention was also variable, ranging from a few hours to two days. Frequent EIT-guided PEEP titration was not done, unlike ARDS net tables, where PEEP was variable according to the FiO2. Moreover, application EIT-guided PEEP titration was based on finding the intercept point between the overall collapse and distension curves, and the selection of this point was subjective, which can be a source of PEEP discrepancy. At the same time, the control groups included either the ARDS net tables or pressure-volume loops (by calculating the maximal hysteresis/setting the PEEP above the lower inflexion point).

Another critical issue is the difference in the operator experience. EIT is a complex technique compared to other techniques for setting optimal PEEP. Secondly, the position of the EIT belt affects its measurement. The impedance changes are measured in a lens-shaped slice of the thorax. As a result, it cannot visualize the ventilation distribution of the whole lung, which is a major drawback. The PEEP level with the best regional compliance differs for cranial and caudal lung regions in mechanically ventilated patients. While performing a decremental PEEP trial to find the optimal PEEP in the caudal lung regions, the diaphragm can enter the measurement field, producing artefacts and causing erroneous results [39].

Various methods are available for titrating PEEP through EIT, which include the overexpansion and collapse (OD/CL) method, end-expiratory lung impedance (EELI) method, GI index method and regional ventilation delay (RVD) method [33]. However, a consensus on the best EIT-derived parameter has yet been reached.

Zhao et al. demonstrated that the global inhomogeneity (GI) index positively correlated with PEEP adjustment with global dynamic compliance and intra-tidal compliance method [20].

EIT reflects regional changes better, i.e. recruitment of dependent and non-dependent lung areas can be visualized separately, compared to the latter techniques, which do not provide any information on ventilation distribution. For the same reason, overall lung compliance is expected to increase, corresponding to the decline in driving pressures.

Heines et al. concluded that the difference in EIT-based PEEP and clinician set PEEP was clinically relevant in 28% of the patients, whereas the EIT-based PEEP disagreed with the PEEP settings according to the ARDS network [39].

PEEP plays an integral part in the acute phase of ARDS by preventing the collapse of recruited airways, thereby improving oxygenation. It minimizes ventilator-associated lung injury (VALI) in the long term by preventing repeated alveolar collapse/distension cycle (atelectotrauma). As a result of this effect, optimal PEEP can contribute to weaning success by minimizing VALI. The studies included in our review did not demonstrate weaning success by EIT-guided PEEP titration. This can be explained by these studies being underpowered to demonstrate a significant difference.

5 Strengths & limitation

The present study comprises the most considerable population reported so far with the necessary minimum population required size for identifying the effect.

Diverse populations and protocols contribute to significant heterogeneity; some of the studies have retrospective control and are prone to selection bias. Moreover, the minimum population required size was not achieved to identify a relevant change in oxygenation and successful weaning.

6 Conclusion

Electrical impedance tomography-guided individualized lung protective ventilation strategies are required to improve the overall outcome, with further requirements of prospective multicenter randomized control trials to demonstrate its utility.