Abstract and Introduction
Intensive care unit (ICU) acquired pneumonia is one of the most common and morbid health care-associated infections. Despite decades of work developing and testing prevention strategies, ICU-acquired pneumonia remains stubbornly pervasive. Pneumonia prevention studies are difficult to interpret because all are at risk of bias due to the subjectivity and poor specificity of pneumonia definitions. Interventions associated with improvements in objective outcomes in addition to pneumonia, such as length of stay or mortality, should therefore be prioritized. Avoiding intubation, minimizing sedation, implementing early extubation strategies, and mobilizing patients do appear to improve some of these objective outcomes. Many of our other assumptions about how best to prevent ICU-acquired pneumonia, however, have recently been challenged. Elevating the head of the bed is supported by very little randomized trial data. Early reports suggested that subglottic secretion drainage may decrease time to extubation and ICU length of stay, but more recent analyses refute these findings. Novel endotracheal tube cuff designs do not clearly lower pneumonia rates. A large randomized trial of selective digestive decontamination in ICUs with high baseline rates of antimicrobial resistance did not identify any benefit. Oral care with chlorhexidine may increase mortality risk and stress ulcer prophylaxis may facilitate pneumonia. Early data on probiotics suggest a possible effect but there is no clear signal yet that they shorten duration of mechanical ventilation or lower mortality. Ventilator bundles on balance do appear to be beneficial but it is not clear which components are most important nor how best to implement them. This article will review recent studies that have challenged, refined, or complicated our understanding of how best to prevent ICU-acquired pneumonia.
Pneumonia is the most common and morbid hospital-acquired infection. It is associated with a crude mortality rate of approximately 30%, an attributable mortality rate of 8 to 12%, and prolongs hospital length of stay by approximately 6 days.[2–4] Hospital-acquired pneumonia extends length of stay more than hospital-acquired bloodstream infections and urinary tract infections and is associated with more years of life lost and years of disability than any other health care-associated infection.[5,6] Most pneumonia surveillance and prevention studies to date have focused on patients on mechanical ventilation. There is increasing appreciation, however, that nonventilated patients are also at high risk for nosocomial pneumonia. The absolute risk of nosocomial pneumonia is substantially lower for nonventilated patients than ventilated patients (1–2 vs. 5–10%) but the adjusted mortality rate for nonventilator hospital-acquired pneumonia (NV-HAP) is equal to or greater than the mortality rate for ventilator-associated pneumonia (VAP)[8,9] and nonventilated patients account for numerically more cases of nosocomial pneumonia at the hospital level by virtue of their greater numbers.[1,10]
The frequency and morbidity of pneumonia in the ICU compel providers to implement robust prevention plans. Our knowledge of how best to prevent nosocomial pneumonia, however, is patchy and incomplete. Some of our bedrock assumptions about how best to prevent pneumonia have been recently been challenged (e.g., oral care with chlorhexidine), the evidence base for some widely practiced interventions remains surprisingly sparse (e.g., head of bed elevation), one persistent component of many hospitals' bundles may facilitate pneumonia (e.g., stress ulcer prophylaxis), our most powerful potential prevention strategy remains mired in controversy despite multiple high-quality trials (i.e., selective digestive decontamination), and rigorous studies of some very promising interventions have failed to include pneumonia as an outcome (e.g., minimizing sedation and spontaneous awakening and breathing trials).
The problem is compounded by the ongoing difficulty the field faces with accurately identifying pneumonia since the lack of a sensitive and specific definition exposes all prevention trials to risk of bias and complicates their interpretation.[11–15] This is true of prospective cohort studies (including before–after and time-series analyses of ventilator bundle implementations) since there is a risk that well-meaning surveyors will subconsciously apply the subjective components of pneumonia definitions more strictly over time leading to a specious impression of lower pneumonia rates.[16,17] It is also true of double-blinded randomized controlled trials since most interventions designed to prevent nosocomial pneumonia work in ways that are circular with pneumonia definitions: the interventions decrease the frequency of positive respiratory cultures and/or volume of respiratory secretions leading to a drop in perceived pneumonias in the intervention arm of the study; but pneumonia clinical criteria correlate imperfectly with histological pneumonia so the drop in observed pneumonias may not correspond to a true decrease in invasive pneumonias. Both these potential sources of error allow for the possibility that investigators can report dramatic decreases in pneumonia rates that may in fact be spurious.
One practical solution to the measurement dilemma is to evaluate the impact of pneumonia prevention strategies on objective outcomes that are less susceptible to measurement bias in addition to pneumonia rates. These can include duration of mechanical ventilation, ICU length of stay, antibiotic utilization, ventilator-associated events, costs, and mortality. A salutary effect on one or more objective outcomes in addition to lower pneumonia rates provides corollary evidence that an intervention is beneficial for patients and that the observed reduction in pneumonia corresponds to a true decrease in serious disease.
This review will focus on selected pneumonia prevention strategies where recent studies have helped modify, and in some cases overturn longstanding beliefs about how best to prevent pneumonia (Table 1). In many cases, these studies have added more nuance and ambiguity than clarity about how best to prevent pneumonia in critically ill patients.
Semin Respir Crit Care Med. 2019;40(4):548-557. © 2019 Thieme Medical Publishers