The impact of patient-ventilator asynchrony in adult mechanically ventilated patients on outcomes：A systematic review and meta-analysis v1

1. Background Patients-ventilator asynchrony (PVA) is defined as a mismatch between the patient breathing efforts with a ventilator’s breath delivery. (1) PVA is a common problem in mechanically ventilated patients, (2, 3) which could potentially induce the exhaustion of diaphragm and respiratory muscles and deliver high tidal volume to the lung leading to ventilator-induced lung injury (VILI). (4, 5) The impact of PVA on clinical outcomes varied depending on the timing of evaluating PVA. (6, 7) Utility of closed-loop-ventilation such as Neurally Adjusted Ventilatory Assist (NAVA) and Proportional Assist Ventilation (PAV) which usually reduces PVA was shown to be associated with better outcomes in the weaning phase of mechanical ventilation using partial ventilation. (8, 9) In contrast, the impact of PVA in acute phase of mechanical ventilation on clinical outcomes were not similar across studies. Thille AW et al. reported that higher incidence of PVA was associated with longer duration of mechanical ventilation, while this was not associated with mortality. (2) Conversely, Blanch L et al. showed that ICU mortality of the patients with higher incidence of PVA was significantly higher than that with lower incidence of PVA. (10) Furthermore, to date, there is no systematic review regarding PVA in the acute phase of mechanical ventilation. Therefore, we will conduct this systematic review and meta-analysis to clarify the impact of PVA on specific clinical outcomes and the effect of interventions in order to improve PVA for mechanically ventilated patients in the acute phase of mechanical ventilation. 2. Review question Part A. The impact of PVA on clinical outcomes To clarify the impact of patient-ventilator asynchrony in adult mechanically ventilated patients in acute phase of mechanical ventilation on outcomes. Part B. The impact of interventions for mechanically ventilated patients on PVA To clarify the impact of interventions for adult mechanically ventilated patients in acute phase of mechanical ventilation on PVA. 3. Method 3.1.1 Types of study to be included Part A. The impact of PVA on clinical outcomes We will include published and unpublished observational studies and secondary analysis of randomized controlled trials (RCTs) including crossover trials, cluster-randomized trials, quasi-randomized trials. We will include studies presented only in abstract or letter form, in any language, from any country and with any length of follow-up. Part B. The impact of interventions for mechanically ventilated patients on PVA We will include published and unpublished observational studies and randomized controlled trials (RCTs) including crossover trials, cluster-randomized trials, quasi-randomized trials. We will include studies presented only in abstract or letter form, in any language, from any country and with any length of follow-up. 3.1.2 Participants Inclusion criteria Adults aged 18 years and older We define “acute phase” as within 72 hrs of initiation of mechanical ventilation, as soon as patients were not able to trigger all ventilator breaths or using assist-control ventilation mode. Patients who have received mechanical ventilation in acute phase of mechanical ventilation. Exclusion criteria We will exclude patients with weaning phase of mechanical ventilation, post-operation, do-not-resuscitate orders, suspected bronchopleural fistula or air leaks, those admitted for organ donation, less than 18 years old, pregnant patients. 3.1.3 Exposure(s)/Intervention(s) Part A. The impact of PVA on clinical outcomes We will include studies that have evaluated whether the incidence of PVA defined as asynchrony index (AI) > 10% or ineffective trigger index (ITI) > 10% or higher index defined by the trialists is associated with clinical outcomes in the acute phase of mechanical ventilation. AI is defined as the number of asynchronous breaths, according to each study, divided by the total number of breaths (both requested and delivered) multiplied by 100. (2) ITI is defined as the number of ineffective triggered breaths divided by the total number of triggered and ineffectively triggered breaths multiplied by 100. (3) Part B. The impact of interventions for mechanically ventilated patients on PVA We will include studies that have evaluated the impact of interventions that possibly affect PVA in acute phase of mechanical ventilation. 3.1.4 Control(s) Part A. The impact of PVA on clinical outcomes asynchrony index (AI) < 10%, ineffective trigger index (ITI) < 10% or lower index defined by the study authors Part B. The impact of interventions for mechanically ventilated patients on PVA Patients who have been defined as a control by the study authors 3.2 Outcomes Part A. The impact of PVA on clinical outcomes 3.2.1 Primary outcomes We will assess the association between PVA as AI, ITI or index defined by the study authors, respectively, with the following outcomes. 1. Duration of mechanical ventilation 2. ICU mortality 3. Hospital mortality 3.2.2 Secondary outcomes 4. Rate of reintubation 5. Rate of tracheostomy 6. All adverse events (as defined by the study authors). 3.2.3 summary of PVA defined by the study authors We will summarize the results as the same type of PVA such as breathe stacking, auto-triggering or double triggering, etc. Part B. Interventions for improving PVA 3.2.1 Primary outcomes 1. Incidence of PVA 2. Duration of mechanical ventilation 3.2.2 Secondary outcomes 3. ICU mortality 4. Hospital mortality 5. Rate of reintubation 6. Rate of tracheostomy 7. All adverse events (as defined by the study authors). 3.3 Searches 3.3.1 Electronic searches the Cochrane Central Register of Controlled Trials(CENTRAL) ; MEDLINE via Ovid; EMBASE; See Appendix 1, 2, and 3 for the search strategies. No restriction in languages. 3.3.2 Other resources the World Health Organization International Clinical Trials Platform Search Portal ( ICTRP) See Appendix 4 for the search strategies. ClinicalTrials.gov See Appendix 5 for the search strategies. We will also inspect the references of extracted studies and the international guideline “mechanical ventilation in adult patients with acute respiratory distress syndrome” from American Thoracic Society, European Society of Intensive Care Medicine, and Society of Critical Care Medicine. (11) 3.4 Strategy for data synthesis 3.4.1 Search strategy The search strategy described above will be used to obtain the titles and abstracts of studies that may be of relevance to the review. These titles and abstracts will be screened independently by two authors who will discard any studies which are not applicable, although any studies and reviews which might include relevant data or information on trials will be retained initially. We will contact the authors of these studies in order to assist in the screening of these studies if necessary. Two authors will then independently assess the remaining abstracts and, if necessary, their full texts, to determine whether they satisfy the inclusion criteria or not. The two authors will then compare their lists, and any differences in opinion between them will be resolved by discussion and, if this fails, through arbitration by a third author. 3.4.2 Data extraction Data extraction will then be carried out independently on the studies selected for inclusion by two authors using standard data extraction forms. Again, we will contact authors of these studies if necessary during this process, and any differences in opinion regarding data collection between the authors will be resolved by discussion, or through arbitration by a third author if required. If more than one publication of a given study exists, the reports will be grouped together and the publication with the most complete data used in the analyses. If relevant outcomes have only been published in earlier versions of studies, these data will be used, and any discrepancies between the published versions will be highlighted. 3.5 Risk of bias assessment Part A. The impact of PVA on clinical outcomes Two researchers independently will assess the risk of bias for each study using QUIPS tool (Hyden 2013) (12). We will assess the following domains: 1. Study Participation 2. Study Attrition 3. Prognostic Factor Measurement 4. Outcome Measurement 5. Study Confounding (Age, Severity score, Duration of mechanical ventilation) 6. Statistical Analysis and Reporting Agreement between the two review researchers with regard to the primary outcomes. Disagreement will be resolved by discussion and if necessary a third review researcher. Part B. Interventions for improving PVA Two researchers independently will assess the risk of bias for each randomized control study using risk of bias 2 tool (13). Two researchers independently will assess the risk of bias for each observational study using ROBINS-I tool (14). 3.6 Assessment of effects of prognostic factor and interventions For the dichotomous variables of mortality and prevalence of reintubation and tracheostomy, pooled odds ratios (ORs) with 95% CIs are provided. For the continuous outcomes including duration of mechanical ventilation (expressed in days) and incidence of PVA (expressed in percentage), the standardised mean differences or the mean differences with 95% CIs were calculated, as recommended by the Cochrane Handbook. (15) Adverse events were narratively summarised because their definition often varies across studies. 3. 8 Missing values 3.8.1 Discrete variables For discrete variables, we will analyze all the data following the concept of intention-to-treat (ITT). For continuous variables, we will not perform imputation of missing values, in accordance with the recommendations set out in the Cochrane handbook (15). We will perform meta-analyses using original data. 3. 8. 2 Missing values We will contact the study authors about missing values. 3. 8. 3Statistical measurement for missing values If only standard errors or p-values are reported, the Altman method(16) is used to calculate the standard deviation. If the author is not known, the standard deviation is calculated from the confidence intervals and t-values using the method described in the Cochrane handbook(17) or supplemented by a validated method(18). Alternatively, the standard deviations are complemented by the validated method(18). The validity of these methods will be verified by a sensitivity analysis. 3.9 Assessment of heterogeneity We will calculate I2 as a measure of variation across studies that is due to heterogeneity rather than chance, and interpreted the values as follows: 0%–40%, negligible heterogeneity; 30%–60%, mild-to-moderate heterogeneity; 50%–90%, moderate-to-substantial heterogeneity; 75%–100%, considerable heterogeneity. If heterogeneity is identified for an outcome (I2 >50%), we will investigate the underlying reasons and conduct the χ2 test, with a p value of