Abstract
A best evidence topic was written according to a structured protocol. The question addressed was whether routine chest radiography is indicated following chest drain removal in patients undergoing cardiothoracic surgery. A total of 356 papers were found using the reported searches; of which, 6 represented the best evidence to answer the clinical question. The authors, date, journal, study type, population, main outcome measures and results are tabulated. Reported measures were mean duration of drains left in situ, timing of drain removal, pathology detected on chest radiographs (CXRs), interventions following imaging and clinical assessment, complications in patients not undergoing routine CXRs and the cost saving of omitting routine CXRs. One large cohort study reported the detection of pathology in 79% of clinically indicated CXRs in comparison to 40% of routine CXRs (P = 0.005). Ninety-five per cent of the non-routine CXR cohort remained asymptomatic and required no intervention. One large observational study reported the detection of new pneumothoraces in 9.3% of patients, 70.3% of which were barely perceptible. Intervention following CXR was required in 0.25% and only one medium-sized pneumothorax would have been potentially missed without CXR. Another large observational study reported intervention following CXR in 1.9% and the presence of relevant clinical signs and symptoms to be a significant predictor of major intervention (P < 0.01). A smaller observational study reported no pathology detected or intervention following CXR in 98% and the cost saving of omitting a single CXR at £10 000 per annum. Another small observational study reported only 7% of CXRs to be clinically indicated with a false-positive rate of 100%, and a false-negative rate of 7% in CXRs not clinically indicated. The smallest study reported no complications in the non-CXR cohort and only one patient undergoing intervention in the routine CXR cohort. We conclude that there is evidence that routine post drain removal CXR provides no diagnostic or therapeutic advantage over clinically indicated CXR or simple clinical assessment. The best evidence studies reported the detection of pathology on routine CXR ranging from 2 to 40% compared with 79% in clinically indicated CXRs (P = 0.005). Whilst the rate of intervention following routine CXR was as high as 4% in the smallest study, clinical signs and symptoms suggestive of pathology were a significant predictor of major re-intervention (P < 0.01).