Extravascular lung water in patients with severe sepsis: a prospective cohort study

Critical Care 2005, 9:R74-R82

Introduction

Severe sepsis occurs at a rate of 250.000 - 750.000 cases a year in the USA. The hallmark of sepsis is increased capillary permeability, which is characterised in the lung by the accumulation of extra vascular lung water (ELVW). In the current study the authors prospectively evaluate the accumulation of EVLW in sepsis and try to describe the heterogeneity of EVLW in patients who developed ARDS. Furthermore they also describe the effects of alcohol abuse on EVLW accumulation and outcome.

Study

Single thermal indicator determination of cardiac output, EVLW and volumetric parameters (PiCCO, Pulsion Medical Systems) was performed in consecutive patients within 72 hours meeting the accepted criteria for severe sepsis. Patients were considered to have elevated EVLW if any measurement was > 10 ml/kg. ARDS was defined according to the American-European consensus conference. ARDS severity was classified using the lung injury score (LIS). The lung permeability index was calculated by relating the EVLW to the intrathoracic blood volume and reflects the permeability of the alveolar-capillary barrier.

A total of 29 patients with severe sepsis were included in the study and 15 (52%) fulfilled the AECC criteria for ARDS. The population included predominantly Afro-Americans (24/29) and 11 patients were HIV positive. Chronic alcohol abuse was present in 13 patients and 5 developed ARDS. Of the 16 non-alcoholic patients, 10 developed ARDS. At the time of enrollment, the median EVLW for all patients was 8.5 ml/kg. Median EVLW was greater in non-survivors than in survivors (14 ml/kg versus 8.0 ml/kg, p < 0.0001), and death was associated with greater EVLW over time. No significant longitudinal differences in oxygenation between survivors and non-survivors were found. EVLW correlated poorly but significantly with measures of lung injury (r2 for PaO₂/FiO₂ ratio, chest radiograph score and LIS 0.27, 0.28 and 0.18 respectively). In non-ARDS patients EVLW was above normal in just over half of the patients (8/14) and these patients were significantly more hypoxic. Patients with ARDS had an increased lung permeability index compared to non-ARDS patients (1.18  ± 0.45 versus 0.60  ± 0.31) but 4/15 ARDS patients did not have increased EVLW during the course of the study. ARDS patients with chronic alcohol abuse had an increased permeability index compared to non-alcoholic ARDS patients (1.73  ± 0.33 versus 1.20  ± 0.47). Using multivariate analysis corrected for differences in fluid intake, chronic alcohol abuse predicted for a higher EVLW in ARDS.

Discussion

This study was designed as a prospective observational study. The abstract and the introduction are confusing as the proposed research questions are actually different. The design of the study is not very well suited for answering all the proposed questions and it appears that some of the questions were raised after the completion of the study. As a proper sample size calculation is missing the study may have been severely underpowered to answer all the questions. The definition of chronic alcohol abuse is broad and suggests that an individual will have permanent changes in alveolar-capillary permeability. Chronic alcohol abuse appears to be an important confounder in this study and it would have been better if it had been an exclusion criterion. Furthermore, the study population contained a high proportion of Afro-Americans and a high number of HIV infected individuals. Therefore, the results cannot be easily applied to the Western European situation. Besides the concerns with the design of this study other important questions remain unanswered. The weak correlation between EVLW and other indexes of lung injury and the fact that 4 patients with ARDS never had increased EVLW remains unexplained. Hydrostatic capillary pressure resulting from fluid resuscitation is probably very important. An adequately designed and powered study is necessary to understand the role of EVLW in patients with severe sepsis and ARDS.

M. Hijmering, Fellow Intensive Care
J.G. van der Hoeven, Intensivist