Chapter 11 Intensive Care Unit

251

CHAPTER 11 • Intensive Care Unit Imaging

bronchopleural fistula, a complication that most com- monly develops within 8 to 12 days of surgery. If a fistula develops earlier, failure of the bronchial clo- sure should be suspected, prompting consideration of reoperation. Bronchopleural fistulas tend to displace the mediastinum to the contralateral side, an unusual occurrence during uneventful postoperative recovery. Small residual air spaces may remain for up to a year following pneumonectomy and do not necessarily imply the presence of a persistent fistula. Very rapid postoperative filling of the hemithorax suggests infec- tion, hemorrhage, or malignant effusion. Fistulous Tracts Fistulas between the trachea and innominate artery develop most frequently when a tracheal tube angu- lates anteriorly and to the right in a patients with a low tracheostomy stoma, persistent hyperextension of the neck, or asthenic habitus. Because of this association, anteriorly directed tracheal tubes should be repositioned. Fistulas also may form between the trachea and esophagus during prolonged ET intu- bation. These usually occur at the level of the ET cuff, directly behind the manubrium. Predisposing factors include cuff overdistention, simultaneous presence of a nasogastric tube, and posterior angu- lation of the tracheal tube tip. The sudden occur- rence of massive gastric dilation in a mechanically ventilated patient provides an important clue. A radiographic contrast agent may be introduced into the esophagus after cuff deflation or tube removal in an attempt to confirm the presence of the fistula. Pulmonary Embolism Although the plain CXR rarely if ever diagnoses PE, it is quite useful to detect other conditions in the differential diagnosis including CHF, pneumo- thorax, and aspiration. Despite limited diagnostic utility, large emboli may give rise to suggestive find- ings: ipsilateral hypovascularity, pulmonary artery enlargement, and (rarely) abrupt vascular cutoff. Local oligemia (the Westermark sign) may be seen early in the course of PE, usually within the first 36 hours. “Hampton’s hump,” a pleural-based tri- angular density caused by pulmonary infarction, is seldom seen. About 50% of patients with PE have an associated pleural effusion. For critically ill ICU patients with suspected thromboembolism, it often makes sense to begin

the evaluation with a Doppler examination of the limbs. If the US exam reveals what appears to be fresh clot in any deep vein, the diagnosis of “throm- boembolism” is established, other tests are unneces- sary as anticoagulation is indicated. It has become clear that for ICU patients not only are the legs a potential source of clot, but the neck and arms are as well. Roughly half of all CVCs in place for a week or more are associated with at least a partially occlusive thrombus, and approximately 15% of these patients have concurrent PEs. The initial use of limb US has several advantages, including avoidance of con- trast exposure and travel from the ICU, as well as lower cost and limited interpretive turnaround time. If the US is negative but the clinical suspicion of PE remains high, ventilation/perfusion ( V / Q ) scan- ning or contrasted chest CT may be performed. The rarity of a normal CXR diminishes the value of V / Q scanning in the critically ill. Nonetheless, normal perfusion scans are very helpful, and abnormal scans help guide the angiographic search for emboli if the systemic contrast needed for CT is contraindicated by renal dysfunction. The sensitivity and specificity of chest CT for the diagnosis of PE are now well established. In the right clinical context, it is safe to assume that a large filling defect seen in the pulmo- nary circuit of a technically adequate study repre- sents clot (i.e., high specificity). Primary tumors of the pulmonary artery, primary lung tumors, cancers metastatic to the mediastinum, nonneoplastic medi- astinal adenopathy, hydatid disease, and mediastinal fibrosis rarely can mimic PE. By contrast, because the sensitivity of CT varies among institutions, and even in the best centers is not 100% for subsegmen- tal clots, a negative CT should not be regarded as definitive data excluding PE. Sensitivity is optimized by a scanner with many rows of detectors, quick acquisition time, optimal contrast injection tech- nique and gating, adequate breath-holding by the patient, and experienced interpretation of optimally reconstructed images including three-dimensional views. Although there is controversy about the impor- tance of subsegmental clots in healthy patients, in critically ill patients with impaired cardiopulmonary reserve, it is probably inadvisable to overlook such emboli. Echocardiogram may reveal right ventricular and pulmonary arterial dilation. If the Doppler US is negative, V / Q not practical, echocardiography unre- vealing, and CT nondiagnostic while clinical suspi- cion remains high, angiography is the next, seldom taken step. Frequently, CT and catheter angiography