Acute lung injury : incidence and predictors of outcome with special reference to inhaled nitric oxide

Abstract

This work started out as a critical examination of the initial clinical experiences with inhalation of nitric oxide (INO) to patients with acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) in Scandinavia 1991-1994. During this examination, we found that epidemiological data and prognostic factors were lacking in Scandinavia for these patient groups, making it difficult to evaluate new therapies. In addition, no studies addressing genotoxicity of INO in humans were found.

Objectives: 1. Retrospectively determine mortality and identify potential risks with INO in patients with ALI/ARDS treated outside clinical studies in Scandinavia 1991-1994, and evaluate possible long-term adverse effects on pulmonary function in survivors. 2. In a human test-model study potential genotoxic, adverse effects after a 2-hour INO exposure. 3. Prospectively determine incidence and 90-day mortality of acute respiratory failure (ARF), and ALVARDS in Scandinavia. 4. Compare the compatibility of two proposed definitions of ARDS. 5. Prospectively determine whether evaluation of respiratory variables recorded in the first week of ICU care, could predict outcome in ARF and ARDS.

Measurements and results: We identified and analyzed all patients (n=56) treated with INO in Scandinavia 1991-1994. Survivors (n=16) were tested with pulmonary function tests > 8 months after the acute episode. To assess genotoxicity, ten healthy volunteers were exposed for INO with scoring of chromosome aberrations in peripheral blood lymphocytes prior to and after exposure. We also performed an 8-week prospective epidemiological cohort study in Sweden, Denmark and Iceland gathering data on 1,515 patients with ARF, and identified patients with ALI (n=287) and ARDS (n=221) using the American-European Consensus Conference on ARDS criteria. We calculated Kaplan-Meier survival curves and tested for significance by log-rank test. Factors independently contributing to mortality were analyzed using a Cox regression model. Patients included in Sweden and Iceland (n=789) were followed for the first week of ICU care. The independent contribution to outcome of collected respiratory variables were tested in a logistic regression model.

Conclusions: No major alterations in the repair process of the lung due to INO therapy as measured by static and dynamic spirometry were found. However, an adverse effect on the diffusion capacity of carbon monoxide cannot be ruled out. No genotoxic adverse effect was found after short-term INO in healthy human volunteers, but such an effect must still remain a possible concern in patients treated with INO.

The overall incidence of ARF is 77.6, of ALI 17.9 and of ARDS 13.5/100.000/yr. in Sweden, Denmark and Iceland with no difference in 90-day mortality (41%) between the groups using conventional ICU resources. The used definition of ARDS and the proposed ARDS definition using the Lung injury score identify two different groups of patients that only marginally overlap. P/F ratio trend, as well as other commonly used isolated respiratory variables cannot be used to predict 90-day mortality in patients (n=128) with early ARDS (< 24 hrs). In patients with ARF (n=661) not fuIfilling ARDS criteria, P/F ratio trend is more closely related to mortality although the predictive value must be interpreted with caution.