Emory Program in Critical Care

The Emory Program in Critical Care is the outgrowth of a long-standing interest in intensive care, which grew from our broad and diverse clinical foundation spanning multiple intensive care units (ICU’s) in the Emory-affiliated hospitals.  Our research interests began with acute lung injury and acute respiratory distress syndrome – related conditions that cause severe hypoxemic respiratory failure and too often result in morbid complications, death or prolonged disability. 

From this beginning, we have broadened our interests to include various etiologic conditions that cause acute respiratory distress syndrome, such as sepsis, trauma, pneumonia and blood product transfusion.  Each of these areas has subsequently intersected with other areas of research, to investigate factors that influence the outcome of sepsis or ARDS, such as chronic alcohol abuse, cancer, diabetes and HIV. 

Investigators in our group were the first to identify chronic alcohol abuse as a significant risk factor for developing ARDS and other organ dysfunction associated with sepsis, and for chronic alcohol abuse contributing to worse outcomes among critically ill patients. In addition, our group identified the association between diabetes mellitus and reduced incidence of acute respiratory distress syndrome, and we continue to investigate the clinical and biological factors that contribute to this association.

Research teams in our group were also the first to conduct large-scale assessments that characterize the epidemiology of sepsis on a longitudinal and a national basis.  As part our health services portfolio we have also explored disparities that exist in various critical illness, such as sepsis.  Investigators in our group have laid the foundation for better identification and treatment of patients with venous thromboembolic disease and ventilator associated pneumonia.  In translational studies, our group has found substantial alterations in pulmonary antioxidant status that lead to altered permeability of alveolar-capillary barrier and accumulation of pulmonary edema, and thus contribute to the development of acute respiratory distress syndrome.  

Our work in the areas outlined below has spanned from simple observational studies to national epidemiologic investigations and health services research to multi-centered randomized, controlled clinical trials. 

To join our Critical Care research group, contact Dr. Greg Martin, or to learn more about our NIGMS-funded T32 training program in Critical Care, contact Dr. Craig Coopersmith.

Learn more abut the Emory Critical Care Center.

Epidemiology and Health Services Research in Critical Care

All members of the Emory Program in Critical Care (EPICC) group have studied the epidemiology of various critical illnesses, most often examining these conditions from a national perspective using administrative data.  Collaborations with investigators at the Centers for Disease Control and Prevention (CDC) facilitated our early work in this area, and these have continued to grow.  In addition to the work described in this section, we have also generated important data related to venous thromboembolism, cardiac arrest, pulmonary artery catheterization.  

Learn more about Epidemiology and Health Services Research in Critical Care at Emory.

Structure, Process and Outcome in Critical Care

Based on the principles of Donabedian, the quality and value of ICU health care are affected by the structure and processes inherent locally to an ICU and more generally to the hospital and the organization in which the ICU operates.  For example, a “closed” ICU organizational structure is generally observed to produce better outcomes at a lower cost than an “open” ICU, and introduction of a “daily goals of care” process has been shown to decrease surgical ICU length of stay.  Importantly, however, there are conflicting reports, such as improvements in outcomes for patients cared for in an ICU with newly developed protocols even if the protocols were not applied to the patients of interest. In order to better understand the heterogeneity in organizational structure and processes across intensive care units and how these factors contribute to relevant clinical outcomes, Dr. Jonathan Sevransky, Dr. William Checkley and Dr. Greg Martin launched CIOS, the Critical Illness Outcomes Study (NCT01109719) under the aegis of the US Critical Illness and Injury Trials Group (USCIITG). 

The CIOS study was designed as a multi-centered ecological study of patients cared for in critical care units across many types.  Overall there were approximately 6,500 patients enrolled from 69 ICU’s geographically spread across the United States. The CIOS study has generated multiple manuscripts, such as:

  • Describing the complexity of executing a multi-centered study where data is collected on both an organizational level (structure and processes) and at a patient level (data collected on individual patients cared for in study ICU’s).
  • Discussion and comparison of the both the consistent and variable processes in conducting multi-site studies in critical care, particularly as it relates to interpretation of ethical regulations by institutional review boards.
  • Identifying the associations between better clinical outcomes and structure and process variables such as the nurse:patient ratio and incorporation a daily plan of care into ICU rounds.
  • The relationship between protocols and outcomes at the patient level in medical, surgical and mixed ICU’s.


Fluid Therapy in Acute Lung Injury

Fluid therapy is one of the most common daily interventions in the ICU. Competing fluid goals are common in critically ill patients, and optimal fluid management may not onlyfluid acute lung injury improve physiology but also improve relevant clinical outcomes.  Dr. Greg Martin and collaborators at Vanderbilt University and Wake Forest University have conducted three seminal clinical studies in this area: one observational cohort study and two interventional clinical trials.  These studies have suggested significant benefits for improving oxygenation and potentially for shortening the duration of mechanical ventilation for hypo-oncotic patients with acute lung injury (ALI).  The results may be summarized as follows:

  1. In a retrospective cohort study of 455 severe sepsis patients at risk for ALI, reductions in oncotic pressure were highly and independently predictive of the development of ALI and subsequent adverse clinical outcomes (prolonged mechanical ventilation, greater mortality). 
  2. In a randomized, controlled clinical trial of 37 hypo-oncotic patients with ALI, treatment with a combination of albumin and furosemide improved oxygenation and hemodynamic stability compared to dual placebo.
  3. In a randomized, controlled clinical trial of 40 hypo-oncotic patients with ALI, treatment with a combination of albumin and furosemide improved oxygenation and hemodynamic stability compared to placebo plus furosemide (diuretic monotherapy).

Further analyses from these studies reveal two important findings.  The first is that changes in oxygenation were not well predicted by clinical measures of oncotic pressure or fluid balance (i.e. diuretic effects), but they were statistically associated with albumin dosage.  Second, additional laboratory studies confirmed an important biological effect of albumin as a thiol donor that altered systemic antioxidant balance in critically ill patients with ALI.

To further understand the role of colloids in optimally managing fluid balance in ALI patients, we have a current trial, Phase III Study Comparing Albumin and Hetastarch Therapy in Acute Lung Injury (also known as “Cardiopulmonary Evaluation of Albumin and Starch Effects in ALI” and as NCT00796419 in ClinicalTrials.gov) that is currently enrolling hypo-oncotic patients with ALI to determine:

  1. The comparative effect on oxygenation of albumin vs. hetastarch treatment
  2. The physiological effects on pulmonary edema (“extravascular lung water”) of treatment with albumin vs. hetastarch
  3. The pulmonary and systemic antioxidant effects of albumin and hetastarch therapy and their associations with clinical outcomes
  4. Relevant clinical outcomes comparing albumin and hetastarch therapy, including ventilator-free days, ICU length of stay and 30 day and 180 day subject survival

Diabetes and Acute Lung Injury

Acute lung injury (ALI) is a common condition occurring in as many as 25% of mechanically ventilated patients and is associated with significant morbidity and mortality.  The inflammatory response in ALI is driven by activation of key signaling pathways that involve alveolar macrophage (AM) responses, activation of transcription factors, oxidative stress, and cytokine release.  Predisposing factors for sepsis (age, chronic liver disease, HIV, cancer) may not equally contribute to the risk of ALI, as not all critically ill patients develop ALI.  Of the disorders commonly associated with the development of ALI, sepsis carries the highest risk at approximately 40%.  Dr. Annette Esper has investigated specific co-morbid conditions that are disproportionately observed in the sepsis population.  Diabetes mellitus (DM) is a common co-morbid condition in critically ill patients, present in 17% of patients with sepsis.  A prospective cohort study conducted by our research group was the first to document that diabetic patients who develop sepsis are at lower risk for developing ALI compared to non-diabetics.  To broaden our understanding of diabetes and its influence on acute organ dysfunction with sepsis, Dr. Esper aggregated multi-year national hospitalization data and reported that sepsis patients with diabetes were consistently less likely to develop respiratory failure than sepsis patients without diabetes.  These findings have now been independently reproduced in at least two prospective cohort studies.

diabetes acute lung

To investigate further this association between DM and ALI, Dr. Esper designed a prospective cohort study of patients with severe sepsis in order to elucidate clinical factors, such as anti-diabetic medications and glycemic control, as factors influencing the development of ALI in diabetics.   A priori, the scientific hypothesis was that specific clinical factors such as prehospitalization glucose control, or exposure to specific anti-diabetic medications such as insulin and thiazolidinediones (TZDs), would impact the development of ALI and would lead to further studies on better defining the mechanisms by which these factors alter ALI risk. 

diabetes acute lung chestThis unexpected finding that people with diabetes mellitus have a decreased incidence of ALI during sepsis is intriguing. Although there are many potential explanations for this apparently paradoxical association, there are compelling reasons to hypothesize that increased activation of the peroxisome proliferator-activated receptor gamma (PPARg) in the AMs of individuals with diabetes could limit acute alveolar injury and promote more rapid resolution of inflammation within the alveolar space.  Our preliminary data suggest that PPARg activity, a key regulator of AM activation, is increased in otherwise healthy diabetics.  If this is true, then the relative PPARg activation in the AM should correlate with the risk of ALI in diabetics, but also in nondiabetics as well.  Most importantly, PPARg activation represents an attractive therapeutic target that could decrease the incidence or severity of acute lung injury in diabetics and nondiabetics with sepsis or other acute stresses. To further investigate the role of PPARg, we are currently conducting a study to determine the association between AM PPARg activity and ALI in patients with severe sepsis, independently of underlying diabetes; and to characterize the effect of PPARg activity on AM phenotype and function in these patients.

The Intersection of Sepsis and Cancer

By integrating large-scale data sources, including the CDC’s National Hospital Discharge Survey and the National Cancer Institute’s Surveillance, Epidemiology and End-Results (NCI SEER) cancer database we were able to identify cancer as the chronic comorbidity with the highest rate of contracting sepsis – even higher than for patients with HIV or diabetes.  For patients with cancer, the risk of sepsis is 10x higher than for patients without cancer.  Importantly, cancer is also the most common comorbidity in septic patients with nearly 100,000 cases annually. Importantly, cancer is also the comorbidity associated with the highest risk of death in patients with sepsis.  As with other conditions that we have studied, there are racial and gender disparities in incidence and outcome with sepsis for patients with cancer.

Dr. Craig Coopersmith has created a murine model that replicates the increased mortality seen in septic cancer patients compared to previously healthy patients who develop sepsis and our groups are together studying possible mechanisms through which pre-existing cancer increases mortality when a host develops sepsis. Based upon preliminary data, both the immune system and gut integrity appear to be involved, and Dr. Coopersmith’s lab is examining both in multiple models of sepsis with multiple tumor lines to determine whether results are generalizable or specific to either type of sepsis or type of cancer.


HIV Disease, Sepsis and Lung Host Immunity

Acute lung disease is a leading cause of death in individuals infected with HIV despite highly active anti-retroviral therapy (HAART).  In addition, acute respiratory distress syndrome (ARDS) persists as a significant challenge to critical care physicians.  Premature immune senescence or “aging of the immune system” increases the risk of pneumonia in HIV, and despite the fact that pneumonia is a leading cause of ALI/ARDS, there is little known of HIV-1 infection and the risk of ALI/ARDS. 

Dr. Sushma Cribbs, Dr. Jenny Han and Dr. David Guidot have been interested in the pathobiologic mechanisms underlying HIV immune dysfunction in the lung.  There is significant evidence to suggest that phagocytosis and other innate immune functions are impaired in experimental animal models of HIV-1, thereby increasing the risk of lung infection. At Emory University, our group has examined granulocyte-macrophage colony stimulating factor (GM-CSF) in animal models of HIV and found that not only is GM-CSF essential for alveolar macrophage (AM) function and maturation, but chronic HIV infection results in alterations in GM-CSF receptor (GM-CSFr) expression.  If shown to be the case in human models as well, strategies to improve these molecular mechanisms could be effective therapies to enhance pulmonary host defense in the setting of ALI.


Additionally, there is compelling evidence that immune function is highly dependent on adequate levels of zinc (Zn), and dietary Zn supplementation can enhance host defenses in the airway.  Our group at Emory University, including Dr. Pratibha Joshi, has found that Zn levels in HIV-1 transgenic rats are significantly lower when compared to levels in wild-type rats. In addition, pharmacological induction of Zn deficiency causes defects in AM GM-CSFrβ expression and phagocytosis, resembling defects that were seen in animal models of HIV. Furthermore, Zn supplementation increases intracellular Zn levels, intracellular GM-CSFrβ expression and bacterial phagocytic function, comparable to those for wild-type AM.  This evidence suggests that Zn bioavailability in the airway modulates immune function and antioxidant defenses in AMs, and that Zn supplementation maybe a novel strategy to help protect against microbial infections in individuals with HIV and respiratory failure, thus improving outcomes in patients with ALI/ARDS.

Currently, Dr. Sushma Cribbs and others have clinical trials underway to prospectively determine the relationship between HIV and the development of ALI by examining the immunostimulatory role of Zn and markers of AM function in ALI patients with and without HIV.  Additionally, we are evaluating the efficacy of Zn supplementation to improve Zn levels and AM function in ALI patients with and without HIV.

Stem Cells in Sepsis and Acute Lung Injury

Sepsis is a significant public health problem that affects nearly 1,000,000 people every year in the United States.  aliIt is an acute inflammatory response to an infection, leading to systemic illness and organ dysfunction, which is a primary determinant of survival.  However, the development of organ dysfunction is highly variable in sepsis patients and not predicted by clinical or physiological variables. Of the mechanisms that lead to sepsis-related organ dysfunction and death, alterations in vascular function are a primary contributor.  Early descriptions of endotoxin induced vascular reactions showed that changes in vascular hemodynamics were influential in accounting for the poor responses of animals to endotoxin.  Since then, many experimental studies have reported and discussed the microvascular alterations that occur in sepsis and more recently, alterations in microvasculature have been reported in human subjects with sepsis and have correlated with severity of disease.  Damage to the endothelium can result in multi-organ failure, which can lead to increased severity of sepsis, and ultimately increase mortality. Stem cells from the bone marrow may contribute to endothelial repair, migrating from the bone marrow to the peripheral circulation where they contribute to vascular repair.  When injected into animal models of ischemia, stem cells are incorporated into sites of neovascularization and have even improved outcomes in critically ill patients. 

sepsisOur group at Emory University includes Dr. Sushma Cribbs, Dr. Greg Martin, Dr. W. Robert Taylor, among other collaborators throughout the University as well as external to Emory.  We have been interested in bone marrow derived progenitor or stem cells for many years, both in experimental animal models and clinically relevant human studies.  We have shown that infusion of bone marrow derived mesenchymal stem cells abrogates lung injury in a murine LPS sepsis model and that circulating endothelial progenitor cells correlate prognostically with organ dysfunction in a cohort of sepsis patients and with survival in a cohort of acute lung injury patients.  Understanding the function of stem cells in sepsis and sepsis-related organ dysfunction (such as acute lung injury), and the relationship to various cytokines and growth factors, may provide innovative diagnostic and therapeutic approaches in this devastating disease.

Vitamin D in Critically Ill Patients

Dr. Tom Ziegler, Dr. Vin Tangpricha and Dr. Greg Martin as investigators in EPICC were among the first to investigate the relationship of vitamin D to antimicrobial peptide expression (particularly LL-37 or cathelicidin) in humans.  In this study, critically ill subjects had significantly lower vitamin D levels and cathelicidin/LL-37 levels compared to healthy controls and there was a strong association between the two mediators.

vitamin 3

This initial work laid the foundation for large-scale epidemiology studies of vitamin D in acutely and critically ill patients and prospective cohort studies to determine the relationship of vitamin D and antimicrobial peptides to infectious and other clinical outcomes in critically ill patients.  In a prospective cohort of medical ICU patients, Dr. Jordan Kempker found no relationship between vitamin D and infectious complications during hospitalization.  Expanding from an institutional cohort towards large scale epidemiology, Dr. Kempker expanded into health services research by integrating data from the National Health and Nutritional Examination (NHANES) with hospitalization data from the Centers for Medicare/Medicaid Services (CMS) to study the association between vitamin D and infectious or non-infectious hospitalizations.  Our group has also completed an NHLBI-funded randomized controlled trial of vitamin D supplementation in ICU patients (NCT01372995) and demonstrated that high dose vitamin D therapy can be safely increase vitamin D levels even in critically ill patients and this treatment may lead to improvements in organ function and important patient-centered clinical outcomes.




Vitamin C for Treatment of ARDS

The lung is an important target of inflammatory mediators in severe sepsis and increased pulmonary NF-kappa-B drives inflammatory mediators in severe sepsis.  Reactive oxygen species (ROS) oxidize vital proteins and activate redox-sensitive pathological signaling pathways, and this exuberant ROS synthesis also damages cells and host tissues and contribute to the pathogenesis of ARDS.  ardsgrowing body of evidence supports the notion that vitamin C alters the pathogenesis of sepsis and sepsis-induced organ dysfunction.  For example, low ascorbate concentrations are common features of patients with sepsis and correlate inversely with multiple organ failure and mortality.  Further, prior data obtained from a phase I human safety study suggests that high doses of vitamin C can be administered intravenously with little or no adverse events. At Emory and Grady Memorial Hospital, Dr. Martin and Dr. Sevransky are co-investigators in an NHLBI-sponsored study with colleagues at Virginia Commonwealth University, the Cleveland Clinic and the University of Wisconsin to conduct a phase II multi-center study of high dose intravenous vitamin C in patients with sepsis-induced ARDS (NCT02106975).

Chronic Alcohol Abuse in Critical Care

Alcohol is one of the most commonly abused substances in the world and chronic alcohol abuse is present in as many as 10% of all hospital admissions.  Alcohol is involved in more than 50% of traumatic injuries, according to some reports, and is a relatively frequent condition for patients in the ICU.  A comprehensive prospective cohort study from our group documented chronic alcohol abuse to occur in approximately 40% of admissions to a medical ICU and to be associated with adverse clinical outcomes.  Investigators in our group were the first to identify chronic alcohol abuse as an important factor in ALI/ARDS.  People who chronically abuse alcohol are more than two times more likely to develop ALI/ARDS compared to non-alcoholics, and also twice as likely to die if they develop ARDS.  Our group has also identified that chronic alcohol abuse predisposes to the development of multiple organ dysfunction syndrome (MODS) in patients with sepsis.  Within the Emory Alcohol and Lung Biology Center, our colleagues have documented tremendous alterations in local pulmonary antioxidant status that lead to altered permeability of the alveolar-capillary barrier.  These abnormalities more readily permit the accumulation of pulmonary edema, as shown in humans with ALI/ARDS who have greater quantities of pulmonary edema fluid compared to non-alcoholics, and thus contribute to the worse outcomes we and other have observed with ALI/ARDS.

alcohol and lung biology