It is uncertain whether daily routine chest radiographs (routine CXR) truly affect daily management of critically ill patients. In the present study we determined the clinical efficacy of routine CXRs, by comparing two periods with different CXR-strategies: for 6 months (period I) routine CXRs were made in all ICU-patients until discharge/death; in addition CXRs were made on indication (i.e. admission to ICU, clinical deterioration, introduction of any invasive devices). In a second period of 6 month (period II) only on demand CXRs were made.
Questionnaires were completed for all CXRs, addressing indication and expected findings. The presence of pulmonary abnormalities (atelectasis, major infiltrates, any pneumothorax, pulmonary congestion or significant pleural fluid), and abnormal position of any invasive device (tube, central venous lines) was scored by the radiologist. Furthermore, it was determined whether the abnormalities truly initiated a change in therapy (i.e., order for cultures/start of antimicrobial therapy, ultrasound/pleurocenthesis, and change (in position) of malpositioned devices). Mortality and length of stay in the 2 periods were compared. Statistics: non-parametric testing.
During the period I, 5118 CXR were made in 885 patients (3396 routine and 1722 on demand CXR). Of all routine CXR, only 221 revealed an unexpected abnormality (6.5%) compared to 188 of the on demand CXRs (10.9%) (P < 0.0001); these findings included (on the routine CXR:) 1.2% malposition of tube –1.1% central venous lines –1.2% atelectasis. In only 74 (2.2%) of all routine CXRs these abnormalities resulted in a change in therapy/replacement of devices. In period II, the number of on demand CXR was slightly higher compared to period I: 1701 CXR were made in 687 patients. Of all on demand CXRs, 239 (14.1%) revaled unexpected abnormalities. Mortality and length of stay was similar in the two study periods.
Routine CXRs have minimal value in guiding management decisions.
We propose to abandon routine CXRs in ICU-patients.
Marcus Schultz, None.