Transforming High Risk to High Yield

Acute Respiratory Distress Syndrome (ARDS) remains a major contributor towards perioperative morbidity and mortality. Although the crude mortality rates in earlier clinical trials from the ARDS Clinical Trials Network were comparatively higher than in more recent studies of this group (35 and 26% respectively),1 others have found that there has been little if any success in improving survival from ARDS over time.2 Hence, the lack of effective approaches for prevention and therapy of ARDS has driven multiple efforts with the goals to better understand the molecular processes in the development and endogenous recovery of lung injury,3,4 to design novel therapeutic approaches targeting these processes,5,6 and to conduct clinical trials to evaluate meaningful outcomes following interventions.7 A major challenge for the practicing clinician is to determine, which patient would be most likely to benefit from such novel, but potentially expensive and side-effect laden therapy. The clinical scientist is confronted with a similar predicament: in order to demonstrate effectiveness of any preventative intervention, studying it in a high-risk population would be desirable. This permits limiting sample size and thus makes a trial more feasible. In this issue of ANESTHESIOLOGY, Dr. Daryl J. Kor from the Department of Anesthesiology at the Mayo Clinic in Rochester, Minnesota and his colleagues from the United States Critical Illness and Injury Trials Group provide us with important new insight on the prediction of postoperative lung injury.8 Using primary data from the previously conducted prospective multicenter Lung Injury Prevention Study (LIPS)9, Dr. Kor and colleagues studied the performance of their formerly developed surgical lung injury prediction model (SLIP)10 in a large multi-center derived data-set of diverse high-risk surgical patients. While the original SLIP score did not perform well in identifying patients, who progressed to ARDS, the authors derived a modified scoring system (SLIP-2) that did. The anesthesiologist Dr. Bjorn Ibsen at the Hospital for Communicable Diseases in Copenhagen (Professor of Anaesthesiology, University of Copenhagen, Copenhagen, Denmark) (1915-2007) revolutionized the management of acute respiratory failure during the 1952 polio outbreak in Denmark.11 The innovative concept of using cuffed endotracheal tubes and manual artificial ventilation outside of the operating room marks a founding innovation in the field of critical care medicine. In the ensuing years, artificial ventilation for acute respiratory failure became a prevalent characteristic of many patients admitted to an intensive care unit. Lung injury leading to ARDS can occur as consequence from direct tissue injury, but it can also be triggered through indirect insults stemming from systemic illness such as sepsis and shock. The realization that mechanical ventilation itself can be not only therapeutic, but also the culprit and perpetrator for the development and progression of ARDS, led to the concept of ventilator induced lung injury. Hence, most approaches for therapy of ARDS revolve around strategies that limit further injury through preventative strategies. Current therapeutic concepts include low tidal volume ventilation with approriate positive end-expiratory pressure and inspired oxygen concentration, restrictive fluid management, consideration of early pharmacologic paralysis, as well as prone positioning. Multiple promising interventions, such as administration of antioxidant nutritional supplements, steroids, and high-frequency oscillation ventilation have failed to provide tangible benefits in most randomized clinical trials. Some examples of innovative pharmacologic and non-pharmacologic treatment strategies currently under investigation include utilization of bone marrow-derived multipotent mesenchymal stem cells,5 activation of regulatory T-cells,3 stabilization of hypoxia-inducible factor 1A,4 modulation of adenosine metabolism,6 as well as utilization of extracorporeal membrane oxygenation.12 Reflecting the multifactorial etiology of ARDS and consistent with previous findings, Dr. Kor and colleagues identfied the following conditions to be associated with a high risk for the development of ARDS and thus essential components of their SLIP-2 scoring algorythm: sepsis, high-risk cardiac and aortic surgery, emergency surgery, cirrhosis, admission other than from home, respiratory rate, arterial oxygen saturation, as well as a patient’s oxygen requirement. The orginal SLIP score was derived from secondary analysis of a prospective cohort investigation of 4,366 surgical patients at the Mayo Clinic. Of the patients included in the original SLIP study, 2.6% developed early postoperative acute lung injury/ARDS.10 Whereas 7.5% of the 1,562 patients in their current study carried this diagnosis.8 Clearly, the two populations studied were quite heterogeneous and the newly derived SLIP-2 algorithm is likely to perform better in more acutely ill patients. The poor performance of the previously reported SLIP algorithm in the dataset used for their current study emphasizes the need for accurate, validated prediction models. Here, the work of Kor and colleagues can eventually help bridge a critical gap between innovative therapies developed from mechanistic experimental models and the design of clinical studies. Given that the SLIP-2 score was derived from a high-risk patient cohort from multiple centers, it would be a valuable tool for designing clincial trials testing novel approaches in similar populations (Figure 1). Figure 1 Accurate predicative algorithms to identify surgical populations at high risk for acute respiratory distress syndrome (ARDS) can inform the design of high yield clinical trials Although the application of an algorithm that has not been validated in an independent cohort will preclude widespread adoption at this time, three key conclusions should be highlighted: First, the proposed SLIP-2 scoring system would likely facilitate studying preventative strategies for ARDS in high-risk surgical patients. The ability to limit sample sizes without compromising statistical power would increase feasibility and possibility for success of future clinical trials. Second, the prospect of enabling clinicians to predict the likelyhood of a surgical patient to develop postoperative ARDS, could permit deployment of perioperative interventions that are more likely to impact meaningful clinical outcomes. Third, the present study serves as yet another example of the fruitful collaboration amongst the United States Critical Illness and Injury Trials Group. In conclusion, the authors should be congratulated on their work and we are hopeful that their efforts will continue to sustainably advance the field of critical care medicine.