872. PROPHETIC: Predicting Pneumonia in Hospitalized Patients in the ICU - A Model and Scoring System
Session: Oral Abstract Session: Respiratory Infections: An Update
Thursday, October 4, 2018: 2:30 PM
Room: W 2002
Background: Prospectively identifying patients at highest risk for hospital-acquired and ventilator-associated bacterial pneumonia (HABP/VABP) by implementing a risk assessment scoring tool may help focus prevention efforts, optimize the screening process to improve clinical trial feasibility, and enhance development of new antibacterial agents.

Methods: Within the intensive care units (ICU) of 28 United States hospitals, between February 6, 2016 and October 7, 2016, patients hospitalized >48 hours and receiving high levels of respiratory support were prospectively followed for meeting the definition of HABP/VABP recommended in U.S. FDA draft guidance. Patient demographics, medical comorbidities, and treatment exposures were recorded. The association between candidate risk factors and odds of developing HABP/VABP was evaluated with a multivariable logistic regression model. Risk factors were selected using backward selection with α=0.1 for model inclusion. A web-based scoring system was developed to estimate the risk of HABP/VABP from the risk factors identified.

Results: 5101 patients were enrolled, of whom 1005 (20%) developed HABP/VABP. 4613 patients were included in the model, excluding 488 (10%) with HABP/VABP at or before enrollment. There are 15 variables included in the model. APACHE II admission score >20 (p<0.001, OR 2.14, 95% CI 2.00 - 2.29), admission diagnosis of trauma (p< 0.001, OR 3.31, 95% CI 1.90 - 5.74), frequent oral or lower respiratory tract suctioning (p< 0.001, OR 2.33, 95% CI 1.81 - 2.99), and receipt of enteral nutrition (p<0.001, OR 2.31, 95% CI 1.69 - 3.16) were the key drivers of increased pneumonia risk. The model demonstrated excellent discrimination (bias-corrected C-statistic 0.861, 95% CI 0.843, 0.880). The web-based scoring system can be accessed via this link: https://ctti-habpvabp.shinyapps.io/web_based_tool/.

Conclusion: Using a web-based scoring system, ICU patients at highest risk for developing HABP/VABP can be accurately identified. Prospective implementation of this tool may assist in focusing additional prevention efforts on the highest risk patients and enhance new drug development for HABP/VABP.

Stephen P. Bergin, MD1, Adrian Coles, PhD2, Sara B. Calvert, PharmD3, John Farley, MD, MPH4, Jonas Santiago, PharmD, MS4, Marcus J. Zervos, MD5, Ana Cecilia Bardossy, MD5, Marin Kollef, MD, FACP, FCCP6, Michael J. Durkin, MD6, Matthew Sims, MD, PhD7, Claire Greenshields, RN, BSN7, Badih A. Kabchi, MD8, Helen K. Donnelly, RN, BSN9, John Powers III, MD10, Pamela Tenaerts, MD, MBA3, Peidi Gu, MD2, Vance G. Fowler Jr., MD1,2 and Thomas L. Holland, MD1,2, (1)Duke University, Durham, NC, (2)Duke Clinical Research Institute, Durham, NC, (3)Clinical Trials Transformation Initiative (CTTI), Durham, NC, (4)U.S. FDA, Silver Spring, MD, (5)Henry Ford Health System, Detroit, MI, (6)Washington University School of Medicine, St. Louis, MO, (7)Beaumont Health System, Royal Oak, MI, (8)East Carolina University, Greenville, NC, (9)Northwestern University, Chicago, IL, (10)George Washington University School of Medicine, Washington, DC


S. P. Bergin, CTTI: Investigator and Scientific Advisor , Research support and Travel to study related meetings .

A. Coles, CTTI: Investigator and Scientific Advisor , Salary .

S. B. Calvert, CTTI: Employee , Salary .

J. Farley, None

J. Santiago, None

M. J. Zervos, CTTI: Investigator , Research support .

A. C. Bardossy, CTTI: Investigator , Research support .

M. Kollef, CTTI: Investigator , Research support .

M. J. Durkin, CTTI: Investigator , Research support .

M. Sims, CTTI: Investigator , Research support .

C. Greenshields, CTTI: Investigator , Research support .

B. A. Kabchi, CTTI: Investigator , Research support .

H. K. Donnelly, CTTI: Collaborator and Scientific Advisor , Research support and Salary .

J. Powers III, None

P. Tenaerts, CTTI: Employee , Salary .

P. Gu, CTTI: Collaborator , Research support and Salary .

V. G. Fowler Jr., CTTI: Investigator and Scientific Advisor , Research support and Salary . Merck: Consultant , Grant Investigator and Scientific Advisor , Consulting fee , Grant recipient and Research support . Cerexa/Actavis/Allegan: Grant Investigator , Grant recipient . Pfizer: Consultant and Grant Investigator , Consulting fee and Grant recipient . Advanced Liquid Logics: Grant Investigator , Grant recipient . NIH: Investigator , Grant recipient , Research support and Salary . MedImmune: Consultant and Grant Investigator , Consulting fee and Grant recipient . Basilea: Consultant and Grant Investigator , Consulting fee and Grant recipient . Karius: Grant Investigator , Grant recipient . Contrafect: Consultant and Grant Investigator , Consulting fee and Grant recipient . Regeneron: Grant Investigator , Grant recipient . Genentech: Consultant and Grant Investigator , Consulting fee and Grant recipient . Achaogen: Consultant , Consulting fee . Astellas: Consultant , Consulting fee . Arsanis: Consultant , Consulting fee . Affinergy: Consultant , Consulting fee . Bayer: Consultant , Consulting fee . Cerexa: Consultant , Consulting fee . Cubist: Consultant , Consulting fee . Debiopharm: Consultant , Consulting fee . Durata: Consultant , Consulting fee . Grifols: Consultant , Consulting fee . Medicines Co.: Consultant , Consulting fee . Novartis: Consultant , Consulting fee . Novadigm: Consultant , Consulting fee . Theravance: Consultant , Consulting fee and Speaker honorarium . xBiotech: Consultant , Consulting fee . Green Cross: Consultant , Speaker honorarium .

T. L. Holland, CTTI: Investigator and Scientific Advisor , Research support and Salary .

Findings in the abstracts are embargoed until 12:01 a.m. PDT, Wednesday Oct. 3rd with the exception of research findings presented at the IDWeek press conferences.