Though the toll of sepsis is known to be enormous – it is estimated to cost the U.S. health care system $24.3 billion each year, and is the nation’s third-leading killer, behind heart disease and cancer – the true magnitude of incidence of and death from the illness remains unknown. There is substantial variability in these numbers, depending on the method used to identify the condition in patients treated at hospital across the United States, according to a new study from the Perelman School of Medicine at the University of Pennsylvania. The authors say these discrepancies limit the potential to improve treatment for the condition.
The findings, which will be published in the May issue of Critical Care Medicine, examined the capture rate of cases of severe sepsis when applying four previously validated methods for identifying sepsis to an inpatient database cataloging 8 million hospital stays each year from 1,050 hospitals in 44 states. The research team found that yearly incidence varied as much as 3.5-fold depending on the method used to capture them. For example, in 2009 the total number of cases of severe sepsis nationally ranged from 894,013 to 3,110,630 when using the various methods. Mortality from severe sepsis varied two-fold depending on the definition used, ranging from 15 percent to 30 percent. Using one method, 25 percent of patients showed cardiovascular dysfunction attributable, in part, to sepsis while using another method, 43 percent exhibited sepsis-related cardiovascular dysfunction. Complicating matters further, the International Classification of Diseases (also known by the abbreviation ICD), the World Health Organization’s health care classification system, has separate codes for the related conditions of sepsis, severe sepsis, and septic shock.
Without a dependable method of identifying sepsis, the Penn authors say, it will be difficult for physicians and researchers to assess clinical interventions and compare outcomes between hospitals and regions. Consequently, effective treatments may be missed because consensus on what is being measured is lacking.
“Early recognition and timely intervention can significantly reduce sepsis-associated deaths,” said the study’s lead author, David F. Gaieski, MD, an associate professor of Emergency Medicine. “As therapies for sepsis evolve, accurate and consistent estimates of sepsis incidence are critical for proper distribution of limited health care resources.”
The source of the discrepancies is the comprehensiveness of the definition of sepsis. More inclusive definitions capture more cases with lower severity; more exclusive definitions capture fewer cases but with higher severity.
Patients with sepsis have a robust systemic inflammatory response to bacteria or other infectious pathogens. Organ dysfunction occurs, and frequently blood pressure drops, resulting in shock. Major organs and body systems, including the heart, kidneys, liver, lungs, and central nervous system, can stop working properly because of poor blood flow and inflammation. Ultimately, this cascade of complications causes 17 percent of all hospital deaths in the United States.
“Many hospitals are using a more proactive approach to identifying sepsis, by implementing alerts based on lab results, vital signs, and other data in an attempt to lower their sepsis mortality,” said the study’s senior author, Brendan G. Carr, MD, MS, an assistant professor of Emergency Medicine, Surgery, and Epidemiology. “Pay-for-performance initiatives could also be developed to target clinical interventions for patients with severe sepsis. But for these efforts to be truly effective, a uniform definition of sepsis and a universal means by which to measure outcomes are essential.”
Additional Penn authors of the study are J. Matthew Edwards, MD and Michael J. Kallan, MS.
The study was supported, in part, by grants from the Beatrice Wind Gift Fund and by National Institutes of Health.
University of Pennsylvania School of Medicine