Acute respiratory distress syndrome (ARDS) is associated with significant morbidity, mortality, and costs in critically ill patients.1,2 The leading causes of ARDS include sepsis, aspiration, and pneumonia, although recent studies have shown that iatrogenic risk factors such as positive fluid balance, high tidal volumes (TVs), high airway pressures, and transfusion of blood products are also associated with the development of ARDS.3–5 Conversely, conservative fluid management strategies have been shown to decrease the duration of mechanical ventilation and intensive care unit (ICU) stay, while also improving lung function, when compared with liberal fluid strategies in patients with lung injury.6,7 Aggressive fluid resuscitation in surgical patients has also been associated with a heightened systemic inflammatory response capable of inducing lung injury.8 Several studies have further shown that the use of lower TV with limited peak inspiratory pressure (PIP) ventilation and restrictive transfusion policies are associated with improved outcomes.4
Ventilator-induced lung injury is a known trigger for ARDS. Inadequate TV leads to atelectasis, hypoxemia, hypercarbia, and inflammation. Excessive TV or inadequate positive end-expiratory pressures (PEEPs), however, lead to ventilation/perfusion mismatch, alveolar capillary injury, inflammation, pulmonary hypertension, and barotrauma. Thus, volutrauma, barotrauma, or biotrauma to the lungs can be caused by mechanical ventilation, leading to acute lung injury and ARDS.9–13
Few studies have examined the role of mechanical ventilation in the operating room on the systemic inflammatory response,14–16 and very little, if any, research has examined the effect of intraoperative fluid resuscitation, mechanical ventilation strategies, and blood administration on the development of ARDS. Several models of ARDS describe a “first hit” leading to tissue injury and inflammation and a “second hit” leading to the development of ARDS.17,18 We hypothesized that intraoperative factors could predispose patients to the development of ARDS during their postoperative period. In this cohort study, we specifically hypothesized that the intraoperative use of aggressive fluid resuscitation (>10 mL/kg/h), high TV per ideal body weight (TV/IBW) ventilation, and the exposure to significant blood products (>5 U packed red blood cells or fresh frozen plasma [FFP]) would be independent risk factors for the development of ARDS.
The IRB at Vanderbilt University Medical Center approved this study as exempt. Our surgical ICU database was queried for patients admitted to the surgical or trauma ICU who had undergone surgery under general anesthesia immediately before or within 24 hours of their ICU admission and had acute postoperative hypoxemic respiratory failure requiring mechanical ventilation. Medical records were examined for the diagnostic features of ARDS during the first 7 postoperative days based on the ARDSNet consensus definition.1 This included respiratory failure with noncardiogenic etiology, radiographic studies with bilateral diffuse alveolar opacities, and laboratory results of an arterial PO2 to the fraction of inspired oxygen (PaO2/FIO2) ratio ≤200. The diagnosis of ARDS in these patients was adjudicated by an investigator (CGH) blinded to the patients' intraoperative course.
Intraoperative data were obtained from the patients' electronic anesthetic care records by study staff blinded to the ARDS status. This included age, sex, length of surgery, FIO2, TV, PIP, PEEP, blood products transfused (packed red blood cells and FFP), and crystalloid and colloid volume administered. For PEEP, blood products transfused, and fluid volume administered, a value of 0 was assigned if no values were documented in the record. All blood products transfused were neutrophil depleted. Intraoperative data were collected only for the initial surgery for each patient. Postoperative information was obtained via medical record review in the electronic patient medical record and included Acute Physiological and Chronic Health Evaluation (APACHE) II score,19 height, weight, arterial lactate, and the postoperative development of pneumonia or sepsis. Patients were excluded from the analysis if they were admitted to the ICU >24 hours before operation, had traumatic injuries to the chest, had thoracic surgery, had evidence of prior pneumonia, acute lung injury, or respiratory failure, or were admitted after liver transplantation with significant fluid and blood product administration because of massive intraoperative blood loss.
Simple descriptives were calculated for demographic and intraoperative variables for all subjects. Continuous variables are summarized by medians and interquartile ranges and categorical variables by percentages. Univariate comparisons by ARDS status were performed via the Wilcoxon rank sum or Pearson χ2 tests for continuous and categorical variables, respectively.
Logistic regression was used to examine the independent association between fluid resuscitation rate (mL/kg/h), TV/IBW (mL/kg), and number of blood units transfused (U) and the development of ARDS.20 TV/IBW was calculated by dividing the TV measurement by a sex-specific adjustment factor defined as 50 + 2.3 (height [cm]/2.54 − 60) for men and 45.5 + 2.3 (height [cm]/2.54 − 60) for women and was modeled as a continuous variable. Fluid resuscitation rate was categorized as <10, 10 to 20, or >20 mL/kg/h, and blood products were categorized as 0, 1 to 2, or >2 U per case. Two sets of models were considered: first an unadjusted model and a second model adjusted for possible confounders using a propensity score adjustment.21 For the TV/IBW regression, the propensity score was defined as the predicted value from a linear regression model where the outcome was TV/IBW and the exposure variables included PIP (cm H2O), PEEP (cm H2O), APACHE II score, infection (presence of either postoperative pneumonia or sepsis), and the categorical versions of blood units and resuscitation rate. Propensity score variables for blood product and resuscitation rate regression models were fitted values from proportional odds models. The proportional odds model predicting blood units included all of the same predictors as those for TV/IBW except substituting TV/IBW for blood units. Similarly for the resuscitation rate model, the same predictors were used except TV/IBW was substituted for resuscitation rate. Odds ratios and 95% confidence intervals (CIs) were calculated to summarize each model. All analyses were performed in R version 2.9.22
Ninety-two subjects met our predetermined inclusion criteria. Of these, 3 were excluded because of missing height and weight information, resulting in the inability to calculate TV/IBW and resuscitation rates. The excluded subjects were generally younger (median age = 25 years), had lower TV and PIP measurements, and required fewer blood transfusions compared with those with complete data. One of these 3 patients also developed ARDS. The surgery types for the remaining subjects (n = 89) included abdominal (48%), orthopedic (16%), vascular (16%), neurological (10%), otolaryngological (6%), and spine (4%) procedures. Of these 89 patients, 25 were characterized as having ARDS. Demographic and intraoperative summaries by ARDS status are presented in Tables 1 and 2. Subjects with ARDS were slightly older and taller than those without ARDS, but these differences were not statistically significant (P = 0.13, P = 0.16). Arterial lactate values were similar between the 2 groups in the immediate perioperative period. Intraoperative variables that were associated with ARDS status included mean FIO2 (P = 0.03) and trends with PIP (P = 0.11) and resuscitation rate (P = 0.12). The difference in FIO2 did not seem to be clinically significant. Patients with ARDS were more likely to have had pneumonia or sepsis in the postoperative period (68% vs 47%, P = 0.07). Table 3 summarizes the results from the regression models (unadjusted and adjusted). Regardless if the propensity adjustment was included, the only significant association noted involved fluid resuscitation. For patients receiving fluid resuscitation amounts >20 mL/kg/h, the unadjusted odds of developing ARDS were approximately 3 times that if <10 mL/kg/h was used (odds ratio = 3.1, 95% CI = 1.0–9.9, P = 0.05). After adjusting for the propensity score, these odds increased to 3.8 (95% CI = 1.1–13.3, P = 0.04). When the resuscitation amount was between 10 and 20 mL/kg/h, the odds of developing ARDS were 2.4 times that of those receiving <10 mL/kg/h, with and without the propensity score, but this was not statistically significant (P = 0.14). TV/IBW and the number of blood units transfused were not found to be associated with ARDS development in this study.
This cohort study suggests that aggressive intraoperative fluid resuscitation strategies increase the odds of developing postoperative ARDS in the ICU. This indicates that the operative course and intraoperative anesthetic management of patients contribute to their susceptibility to the development of ARDS. Because of our inclusion criteria, all patients in this study had evidence of acute respiratory failure on arrival to the ICU after their operation. However, only 25 of 89 patients developed ARDS per the ARDSNet criteria,1 and the mean time to development of ARDS in the postoperative period was 2.6 days, suggesting that the patients did not have ARDS before their anesthetic course, providing evidence to support the intraoperative “first hit” hypothesis. Positive fluid balance is considered to be a risk factor for the development of ARDS in ICU patients.5,6 This study shows an association between aggressive intraoperative fluid resuscitation and ARDS, which could represent the initial pulmonary insult to surgical patients.
The potential role of intraoperative fluid administration in the development of ARDS will likely shift consideration of the harm to benefit ratio of fluid administration in the operating room. The statistically significant association between greater fluid resuscitation and the development of ARDS found in this study was based on a mL/kg/h measurement, a clinically significant and relevant value for determination of intraoperative fluid administration. Whereas only a small variation existed in the TV/IBW ventilation of patients, there was a large variation in intraoperative fluid administration at our institution. Although it is possible that patients receiving more fluid were sicker, we did adjust for both APACHE II score and infection as covariates in the propensity score, thus strengthening the association between fluid administration and ARDS.
Any significant effect of the intraoperative ventilation strategy on the development of ARDS would primarily be the result of variation in the TV/IBW ratio because other significant ventilatory variables, including PIP and PEEP, were accounted for in the propensity score. The results of this study, however, did not support our initial hypothesis that high TV/IBW intraoperative ventilation increases the odds of developing ARDS. In examining the effect of TV/IBW, we observed that the majority of patients' lungs were ventilated between 8 and 10 mL/kg and very few patients at a TV/IBW >10 mL/kg. The lack of high TV/IBW ventilation intraoperatively at our institution likely represents wider acceptance of the ARDSNet study that supported the use of lower TVs in mechanically ventilated patients with ARDS.1 The median intraoperative PEEP, however, was 0 in both groups. Whether PEEP was not frequently used or was used but not routinely documented could not be accounted for in this study.
The results of this study also did not support our hypothesis that increased blood product administration increases the odds of postoperative ARDS. The statistical methods used in this study separated the contributable effects of blood product versus crystalloid/colloid fluid administration. However, there was little variation in the amount of blood product administered in our patient population, with relatively few products given overall, making it difficult to ascertain significance without a much larger patient population and larger number of products given.
This cohort study has several limitations based on the study design, number of patients, and data retrieved. The rigorous inclusion criteria and definition of ARDS diagnosis produced a small number of outcomes, limiting the ability to detect subtle changes in outcome and necessitating the use of propensity scores to adjust for important covariates. Our surgical database focused on patients with respiratory failure requiring mechanical ventilation; consequently, not all postoperative patients were included. Whereas this might limit the identification of additional risk factors for ARDS, it increased the likelihood of ARDS development in our cohort. We did not subdivide crystalloid and colloid, and blood and FFP were included together despite potential immunologic differences.23 We were unable to reliably obtain data on intraoperative hypotension, central venous pressure, and vasopressor therapy to study the potential effects of hemodynamic instability on the development of ARDS. However, arterial lactate values were similar between the 2 groups, suggesting comparable end-organ perfusion.24 We could not control for the type of surgery, and we were unable to assess for differences in the postoperative ICU course (fluid therapy, blood product administration, etc.) that patients might have experienced because of our limited sample size. However, we did adjust for APACHE II scores and the presence of sepsis or pneumonia in our propensity scores. Sepsis, in particular, warrants a more aggressive fluid administration strategy, and by controlling for it in the propensity score, we were able, in part, to account for some differences in fluid administration that may have occurred in these patients. Per ICU protocol, the postoperative ventilation strategies were set at 6 mL/kg IBW for all patients in accordance with the ARDSNet study,1 but fluid, blood, drug, and other therapy administration were not controlled in this study.
This cohort study provides evidence to suggest a relationship between intraoperative fluid resuscitation and the development of ARDS. Larger prospective trials are required to confirm these findings and possibly identify additional risk factors for ARDS by studying all postoperative patients, including those with and without postoperative respiratory failure.
CGH helped to design and conduct the study, analyze the data, and write the manuscript. LW, AB, and PPP helped to design the study, analyze the data, and write the manuscript. NDM and JSS helped to analyze the data and write the manuscript.
1. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. N Engl J Med 2000;342:1301–8
2. Amato MB, Barbas CS, Medeiros DM, Magaldi RB, Schettino GP, Lorenzi-Filho G, Kairalla RA, Deheinzelin D, Munoz C, Oliveira R, Takagaki TY, Carvalho CR. Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome. N Engl J Med 1998;338:347–54
3. Jia X, Malhotra A, Saeed M, Mark RG, Talmor D. Risk factors for ARDS in patients receiving mechanical ventilation for >48 h. Chest 2008;133:853–61
4. Plurad D, Martin M, Green D, Salim A, Inaba K, Belzberg H, Demetriades D, Rhee P. The decreasing incidence of late posttraumatic acute respiratory distress syndrome: the potential role of lung protective ventilation and conservative transfusion practice. J Trauma 2007;63:1–7
5. Sakr Y, Vincent JL, Reinhart K, Groeneveld J, Michalopoulos A, Sprung CL, Artigas A, Ranieri VM. High tidal volume and positive fluid balance are associated with worse outcome in acute lung injury. Chest 2005;128:3098–108
6. Wiedemann HP, Wheeler AP, Bernard GR, Thompson BT, Hayden D, deBoisblanc B, Connors AF Jr, Hite RD, Harabin AL. Comparison of two fluid-management strategies in acute lung injury. N Engl J Med 2006;354:2564–75
7. Calfee CS, Matthay MA. Nonventilatory treatments for acute lung injury and ARDS. Chest 2007;131:913–20
8. Cotton BA, Guy JS, Morris JA Jr, Abumrad NN. The cellular, metabolic, and systemic consequences of aggressive fluid resuscitation strategies. Shock 2006;26:115–21
9. Carney D, DiRocco J, Nieman G. Dynamic alveolar mechanics and ventilator-induced lung injury. Crit Care Med 2005;33:S122–8
10. Dreyfuss D, Soler P, Basset G, Saumon G. High inflation pressure pulmonary edema. Respective effects of high airway pressure, high tidal volume, and positive end-expiratory pressure. Am Rev Respir Dis 1988;137:1159–64
11. Gajic O, Dara SI, Mendez JL, Adesanya AO, Festic E, Caples SM, Rana R, St Sauver JL, Lymp JF, Afessa B, Hubmayr RD. Ventilator-associated lung injury in patients without acute lung injury at the onset of mechanical ventilation. Crit Care Med 2004;32:1817–24
12. Gajic O, Frutos-Vivar F, Esteban A, Hubmayr RD, Anzueto A. Ventilator settings as a risk factor for acute respiratory distress syndrome in mechanically ventilated patients. Intensive Care Med 2005;31:922–6
13. Schreiber T, Boyle W. Lung injury caused by mechanical ventilation. Contemp Crit Care 2005;3:1–12
14. Choi G, Wolthuis EK, Bresser P, Levi M, van der Poll T, Dzoljic M, Vroom MB, Schultz MJ. Mechanical ventilation with lower tidal volumes and positive end-expiratory pressure prevents alveolar coagulation in patients without lung injury. Anesthesiology 2006;105:689–95
15. Wrigge H, Zinserling J, Stuber F, von ST, Hering R, Wetegrove S, Hoeft A, Putensen C. Effects of mechanical ventilation on release of cytokines into systemic circulation in patients with normal pulmonary function. Anesthesiology 2000;93:1413–7
16. Wrigge H, Uhlig U, Zinserling J, Behrends-Callsen E, Ottersbach G, Fischer M, Uhlig S, Putensen C. The effects of different ventilatory settings on pulmonary and systemic inflammatory responses during major surgery. Anesth Analg 2004;98:775–81
17. Moore FA, Moore EE, Read RA. Postinjury multiple organ failure: role of extrathoracic injury and sepsis in adult respiratory distress syndrome. New Horiz 1993;1:538–49
18. Rotstein OD. Modeling the two-hit hypothesis for evaluating strategies to prevent organ injury after shock/resuscitation. J Trauma 2003;54:S203–6
19. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985;13:818–29
20. Harrell FE Jr, Lee KL, Califf RM, Pryor DB, Rosati RA. Regression modelling strategies for improved prognostic prediction. Stat Med 1984;3:143–52
21. Iami K, van Dyk D. Causal inference with general treatment regimes: generalizing the propensity score. J Am Stat Assoc 2004;99:854–66
22. R Development Core Team. R: a language and environment for statistical computing.Vienna, Austria: R Foundation for Statistical Computing, 2009
23. Watson GA, Sperry JL, Rosengart MR, Minei JP, Harbrecht BG, Moore EE, Cuschieri J, Maier RV, Billiar TR, Peitzman AB. Fresh frozen plasma is independently associated with a higher risk of multiple organ failure and acute respiratory distress syndrome. J Trauma 2009;67:221–7
24. Mizock BA, Falk JL. Lactic acidosis in critical illness. Crit Care Med 1992;20:80–93