To evaluate the yield of routine daily chest X-rays in the Intensive care patients
All routine daily chest radiographs performed in a 21-bed medical-surgical ICU were included. The following data was collected prospectively: demographics, clinical presentation, ventilatory mechanics, sedation and vasoactive medications. All radiographs were reviewed by the ICU consultant during morning X-ray rounds. Nothing was done to influence the ordering of radiographs. We documented new radiographic changes, such as the development of new infiltrates, atelectasis, pnuemothorax or edema, variations in the position of tubes and lines and any intervention based on these changes.
A total of 308 consecutive chest radiographs were included. The mean APACHE II score was 19.8 ± 6.5. New radiographic changes were observed in 92 instances (29.8%). The most frequent changes were malposition of endotracheal tube (36.9%), pulmonary edema (18.4%) and atelectasis (11.9%). Two or more findings were observed in 16.3% .Pneumothorax was found in only 4.3% (1.2% of all chest radiographs). Major findings were observed in only 13.9%. A change in therapy was made on the chest radiograph in 79 instances (25.6%). The most frequent therapeutic interventions were repositioning of endotracheal tube (51.8%), starting diuretics (16.1%) and several interventions (starting antibiotics, diuretics, and chest physiotherapy) in 11.3%. There was no significant difference in the percentage of radiographic changes between intubated vs. non-intubated patients (32.2% vs. 24.5%, p=0.26) or between patients with Glasgow coma scale of less than 10 vs. those above 10 (34% vs. 26% p=0.41).CONCLUSIONS: Routine daily chest radiographs revealed new findings in only 29.8% and resulted in management change in 25.6% of instances. Radiographic changes and interventions were comparable among intubated and non-intubated patients. Life threatening findings (such as pneumothorax) were rather rare.
These findings question the practice of routine chest radiographs in ICU patients. It is possible that many of these tests can safely be avoided. Randomized trial comparing routine versus as necessary ordering behavior is necessary in this field.
O. Dabbagh, None.