Lower tidal volume at initiation of mechanical ventilation may reduce progression to acute respiratory distress syndrome: a systematic review

Close to 200,000 cases of acute lung injury (ALI) or acute respiratory distress syndrome (ARDS) occur annually in the United States [1, 2]. It remains a leading cause of death in critically ill patients, with a mortality rate of approximately 30%, but as high as 60% in the elderly [1, 3, 4]. ARDS also represents a large societal burden, accounting for 3.6 million hospital days annually and significant long-term sequelae in survivors [1, 3]. These include various neuropsychological impairments (e.g. depression, cognitive decline), persistent weakness and pulmonary dysfunction, and decreased quality of life [3, 5].Unfortunately, the number of ARDS trials demonstrating improved clinical outcomes are far outnumbered by those that have not [6–21]. In patients with ARDS, the current body of evidence supports a lower tidal volume strategy as best practice [6].

In contrast, the optimal tidal volume in patients without ARDS is uncertain [22, 23]. Historically, it was thought that high tidal volume not only was safe, but also was preferred to avoid the hypoxia, atelectasis, and acidosis associated with mechanical ventilation with lower tidal volumes [24–26]. This led to the common use of tidal volumes in the 12- to 15-ml/kg range [27]. However, experimental data suggest that tidal volume is a major contributor to the development of lung injury [28]. Clinical data in intensive care unit (ICU) patients also suggests that larger tidal volume may contribute to the development of ARDS [29–31], suggesting that this syndrome may in part be preventable. However, this evidence has been inconclusive, because of observational trial design, small sample sizes, and lack of reproducibility in different patient populations [32, 33]. In patients without lung injury at the time of endotracheal intubation, progression to ARDS can occur over the course of hours and may increase mortality by >30% [30, 34, 35]. This has led to increased interest in ARDS prevention and the examination of a lower tidal volume strategy to prevent its development in patients without lung injury.

We therefore aimed to systematically review the evidence regarding the role of tidal volume as a contributor to ARDS progression in patients without ARDS. The objectives of this study were to perform a systematic review of the literature comprising a comprehensive search strategy and standardized analysis techniques to determine whether lower tidal volumes, compared with higher tidal volumes, are associated with a decreased incidence of ARDS in mechanically ventilated patients. We hypothesized that lower tidal volume ventilation is associated with a decrease in progression to ARDS in mechanically ventilated patients, but the existing literature lacks the methodologic quality to provide definitive recommendations for the best tidal volume in patients without ARDS.

This systematic review was designed, conducted, and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) and Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines [36, 37]. It did not require ethical approval from Washington University Human Research Protection Office. It is recognized that ALI and ARDS have undergone recent definitional change [38]. Although the definitional criteria in place during the systematic search for clinical trials, as well as the conduct of each study, was the American-European Consensus Conference (AECC) criteria, for consistency in reporting according to the new Berlin definition, we use the term ARDS to refer to both ARDS and ALI (as reported in previous literature) [2, 38].

In patients who do not have ARDS at the time of endotracheal intubation, the preponderance of evidence implies an association between lower tidal volume and subsequent progression to ARDS. However, this systematic review could yield no definitive recommendations regarding the most appropriate tidal volume strategy in patients without ARDS. Several important caveats exist. An overall lack of data appears in the existing body of literature, as well as a lack of methodologic quality in the data that do exist. Of the 13 trials included in this review, only one was a high quality RCT, whereas the remainder were of observational trial design. None of the observational trials referenced the STROBE guidelines, which are endorsed by 110 journals worldwide [50]. Publication guidelines like STROBE were designed to standardize the reporting of medical research to enhance transparency and minimize variability [51]. An increasing body of evidence indicates that the guidelines do improve the overall quality of research reporting [52].

There was also substantial heterogeneity appeared in the patient populations examined, each carrying perhaps vastly different risk for ARDS progression. For example, relatively healthy patients with a finite exposure to tidal volume risk (operating room cases) are inherently different in risk for progression compared with critically ill ICU patients [53]. Furthermore, even among the studies restricted to patients with longer exposure to tidal volume (ICU patients), the diagnoses leading to mechanical ventilation (Table 1) also carry very different risk for ARDS (e.g. septic shock versus postoperative). Funnel-plot inspection for the assessment of publication bias revealed asymmetry suggestive of publication bias. However, asymmetric funnel plots are not sufficient proof of publication bias. Other potential explanations for asymmetry include study-to-study heterogeneity with the intervention fidelity or outcome assessment. Also, improvement in the usual care in the "control" treatment groups as routine management evolves over time may reduce the observed effect size. It is also possible that an asymmetric funnel plot is the result of chance alone [41]. Finally, differences were noted between studies in how a patient was classified as having developed ARDS, even though each study required fulfillment of accepted consensus criteria [2].

The results of this systematic review do reveal some important trends in the data. First, each study that reported ARDS onset showed that it occurs early in mechanically ventilated at-risk patients. This may reflect the difficulty in defining the precise onset of ARDS retrospectively, but also that many mechanically ventilated patients may have early ARDS that has yet to manifest clinically. This suggests that prevention trials should target the most immediate time after initiation of mechanical ventilation (e.g. in the ED, interhospital transport, and/or the first 24 hours of ICU admission). Endotracheal intubation and initiation of mechanical ventilation is a common occurrence in the ED, yet the effect that ED-based mechanical ventilation may have on outcome has not been studied [54]. However, preenrollment data from ARDS clinical trials showed that early tidal volume did not influence mortality, further highlighting the equipoise on this topic [55].

Second, the incidence of ARDS is very low when tidal volume is studied in low risk patients in the operating room (<2%). This is in contrast to the ICU, where the progression to ARDS appears to be a significant problem (as high as 44%). However, in a cohort of higher-risk surgical patients, higher tidal volumes have been associated with other clinical outcomes, including greater nonpulmonary organ failure and an increase in duration of mechanical ventilation and ICU length of stay [56].

Third, although no definitive conclusions between tidal volume and progression to ARDS can be made, based on this systematic review, several findings in this body of evidence suggest a link between higher tidal volumes and progression to ARDS. Although almost all of the data are observational trials, consistent findings across trial design suggest a cause-effect relation [57]. In the only RCT examining this topic to date, tidal volume was predictive of progression to ARDS. Additionally, the majority of the observational data demonstrate an increased ARDS incidence with larger tidal volumes, which also suggests a causal link [57]. A dose-response relation (seen in five studies), with the proportion of ARDS patients increasing with higher tidal volumes, suggests cause-effect, as does the fact that no findings appeared in the opposite direction (no study showed lower tidal volumes to be harmful).

Fourth, the development of ARDS was associated with many other factors across these trials. These include patient-related (e.g. restrictive lung disease [29]), ventilator-related (e.g. airway pressure [47]), and non ventilator-related (e.g. fluid balance and transfusion [35, 45, 47, 49]) factors. This further highlights that the development of ARDS reflects an interaction between patient risk and treatment variables, which was demonstrated before, and suggests that the most effective ARDS-prevention strategy may involve a multimodal approach (e.g. early resuscitation, fluid management and type, lower tidal volume, high transfusion threshold, and so on) [48, 53]. In addition, future ARDS-prevention trials will need to control for each of these confounding variables while assessing patient populations with similar baseline risk of progressing to ARDS.

Finally, the progression to ARDS is associated with significant increase in morbidity and mortality. In all trials directly comparing ARDS versus no-ARDS groups, mortality (up to fivefold increase), length of stay, and organ failure were all increased with the development of the syndrome. Although the best strategy to prevent the progression to ARDS has not been fully elucidated, it is clear that its prevention should be a high priority in clinical care and further research [58].

Important limitations exist in this systematic review. The only RCT was stopped early, after an unplanned interim analysis, because of safety concerns of ARDS development in the 10-ml/kg PBW group [44]. Clinical trials stopped early for benefit are increasing in prevalence, often lack information on why the trial was stopped, and routinely show larger than expected treatment effects [59]. The benefit of clinical trials stopped early, including the trial in this review, should therefore be viewed with caution. However, with 75% of the planned enrollment achieved, and a very significant between-group difference for the primary outcome (p = 0.01), confidence is increased that these results are indeed valid.

Second, relatively few studies have examined ARDS prevention and the role of tidal volume in the development of ARDS. Existing studies are limited by their observational trial design, and therefore any attempt to analyze these studies as a group is limited by their individual weaknesses.

Third, a meta-analysis of these data could not be conducted as planned a priori, which differs from another systematic review on this topic [60]. The patients were not qualitatively similar and represent a spectrum of patients from those with very little risk for ARDS (e.g. elective surgery patients) to those with very high risk (e.g. septic shock). When contemplating whether a systematic review should become a meta-analysis, heterogeneity is typically first addressed on a macroscopic level of the individual study's design, population, intervention(s), and outcome measures. If sufficient differences exist between individual studies such that lumping them together is comparing "apples and oranges," then a meta-analysis is not appropriate. Systematic review authors should proceed to meta-analysis only if this first level of heterogeneity is assessed and deemed unlikely, an approach that has been previously advocated [61]. We were therefore able to conduct only a qualitative analysis of the data, as conducting a meta-analysis would have served only to homogenize patients artificially. We recognize that these limitations make drawing recommendations difficult based on this systematic review, but with our inclusion of an additional 10 studies (and >7,000 ICU patients) and strict adherence to reporting guidelines, we believe this current review to be the most complete and methodologically sound assessment of the literature to date.

The focus on early mechanical ventilation and ARDS prevention is likely to increase. Recently the National Heart, Lung and Blood Institute (NHLBI) recommended the development of ARDS prevention trials, as well as observational trials of patients without ARDS undergoing prolonged mechanical ventilation [58]. Given the limited treatment options for ARDS, the period of mechanical ventilation immediately after endotracheal intubation is emerging as a critical time window of clinical care and an opportunity for clinical trials to improve outcome [58]. Our results are important in further highlighting knowledge gaps, which may serve in the future design of ARDS prevention trials, and not only call attention to the limitation of the literature, but also highlight potential solutions for clinical trials going forward. This includes RCTs targeting homogeneous patient groups with higher event rates for ARDS progression and early timing of trial enrollment (e.g. immediately after endotracheal intubation in the ED). Factorial trial design aimed at limiting more than one risk modifier for ARDS progression, such as tidal volume and fluid balance, should be considered to maximize efficiency in this research arena [62]. Future research should aim to minimize bias by incorporating publishing guidelines such as the Consolidated Standards of Reporting Trials (CONSORT) for RCTs and the STROBE for observational trials and should have clinically relevant end points, such as progression to ARDS or development of nonpulmonary organ failure as a surrogate for ventilator-associated lung injury [40, 63, 64]. Finally, to reduce efficiently the preventable heterogeneity and variability in research and clinical care in a sustainable fashion, interventions should be simple, and reproducibility should be enhanced with the aid of implementation science and electronic aids [65–69]. Knowledge of factors influencing a lack of adherence to beneficial therapy in ARDS may aid clinicians in translating research to routine care for prevention of the syndrome as well [65, 70]. These trials will ultimately influence clinicians and change practice. Therefore, analyzing critically what has and has not been done is of vital importance.

Currently, no definitive recommendations can be made on the most appropriate tidal volume strategy in mechanically ventilated patients without ARDS. Data suggest an increased incidence of ARDS with higher tidal volumes, but previous studies are limited by their heterogeneity and high variability in baseline ARDS risk among the patients included. Given the increased mortality and morbidity associated with progression to ARDS, a focus on prevention should remain a high priority for future research and clinical practice.