[Do clinical or physiopathological parameters exist for rationalizing the demand for thoracic radiograms in patients admitted to an intensive care unit].

To rationalize the demand of chest radiography in Intensive Care Unit patients, 423 radiographs of 35 patients were studied prospectively. The radiologic findings were compared with clinical data, blood gas analysis and laboratory data. No significant correlation was demonstrated between pulmonary radiographic changes and patient clinical conditions classified as unchanged, improving and worsening. In contrast, a good correlation (R = 0.65; p < 0.001) was observed between the chest radiographic score of pulmonary changes and the PaO2/FiO2 ratio. Changes in this parameter may be a good reason to perform bedside chest radiographs, whereas this exam appears rather useless when PaO2/FiO2 ratio is unchanged.