Barium swallows after free jejunal transfer: should they be performed routinely?

Fistula formation after free jejunal transfer for pharyngoesophageal reconstruction is a serious complication with potentially critical consequences. Barium swallow is used postoperatively to check for anastomotic competence before feeding but has been unreliable as a predictor of leak at our institution. The objective of this study was to evaluate the role of routine postoperative barium swallow in 41 consecutive jejunal transfers. Thirty-nine patients who underwent 41 consecutive free jejunal transfers had a routine barium swallow performed between postoperative days 12 and 17. Radiologic findings and clinical outcome were evaluated and correlated. All barium swallows were reviewed by a single experienced radiologist in a blinded fashion. One total and one partial flap failure necessitated a second free jejunal transfer. Pharyngocutaneous fistulae developed after nine free jejunal transfers, of which the barium swallow was normal in four (44 percent) and showed a leak in five (56 percent). In the 32 free jejunal transfers with no clinical leaks, 6 (19 percent) had radiologic leakage of contrast. Thus, barium swallow was normal in 30 patients and showed leakage in 11 patients. Normal barium swallow correlated with uncomplicated clinical course in 26 of 30 cases. In the remaining four cases (13 percent), however, a delayed fistula developed, which was secondary to flap necrosis in one case (negative predictive value 87 percent). On the other hand, radiologic leaks corroborated clinical fistula in 5 of 11 cases (45 percent), whereas no fistula developed in 6 cases (positive predictive value 46 percent). Of the five patients with clinical fistulae, four had early leaks (within 1 week), and the barium swallow did not provide additional information. The fifth patient developed a delayed leak 2 weeks after the barium swallow. Review of these barium swallows at the time of this study reversed the initial report of leakage in three patients, improving the predictive value to 63 percent. These patients had an uncomplicated clinical course. The positive predictive value of clinical assessment alone was 63 percent. We conclude that barium studies following free jejunal transfers can be difficult to interpret, but an experienced radiologist can improve their accuracy. A normal barium swallow, however, does not ensure an uneventful clinical course. Similarly, radiologic leaks do not imply a clinical complication of fistula. Clinical judgment should therefore be exercised in initiating oral intake after free jejunal transfer. Barium swallow should be used only as an adjunct to aid in patient management.