Sepsis in intensive care patients: challenges in diagnosis and prognostication

Abstract

Sepsis is a syndrome that arises when the body’s response to a severe infection injures its own tissues. It is a major and increasing cause of in-hospital morbidity and mortality. Despite recent advances in the management of sepsis, the morbidity and mortality caused by sepsis remain unacceptably high. Patient populations with sepsis are heterogeneous with regard to the implicated pathogen, the underlying infectious disease and their genetic profile, which makes it demanding to improve the outcome of sepsis. In particular, the diagnosis of sepsis and the prognostication of sepsis patients remain challenging. For these reasons, the Molecular Diagnosis and Risk Stratification of Sepsis (MARS) data collection initiative and biobank study was started in two Dutch tertiary intensive care units in 2010. Its aim is to provide solutions for the current shortcomings in the diagnosis and prognostication of critically ill patients with presumed sepsis. This thesis describes some of the results of this initiative. The first part of this thesis focuses on the diagnostic challenges in critically ill patients with sepsis. Chapter 2 describes the influence of minor variations in the definition and measurement of SIRS criteria and organ failure on the observed incidences of sepsis, severe sepsis and septic shock. The interobserver agreement for classifying infections of various subtypes according to Centers for Disease Control and Prevention criteria in a mixed ICU population is described in Chapter 3. Chapter 4 provides an assessment of the likelihood of infection in patients who were treated for sepsis upon admission to the ICU, and quantified the association between plausibility of infection and mortality. Chapter 5 presents a validation study of a novel surveillance system for complications of mechanical ventilation as an alternative to the current surveillance of ventilator-associated pneumonia. In the second part, we describe some important complications of sepsis in critically ill patients and provide possible solutions for the improved prognostication of patients with sepsis. In Chapter 6 we present a comparative analysis of the incidence and attributable mortality of secondary infections in patients admitted to the ICU with or without sepsis. Chapter 7 presents the attributable mortality caused by an important complication of sepsis, delirium, in critically ill patients. The incidence and outcomes of atrial fibrillation in a cohort of critically ill patients with sepsis is provided in Chapter 8. Chapter 9 presents a model that predicts the evolution of disease for individual patients admitted for sepsis, by estimating daily probabilities of progression to death or multiple organ failure after one week in the ICU, using routinely available parameters. A synthesis of the results and general discussion is provided in Chapter 10.