Materials and methods

The records of 305 consecutive patients who underwent microvascular free flap reconstruction from 1995 to 2005 were reviewed retrospectively. Demographic variables were collected in addition to diagnosis, tumour site, stage, cancer therapy including radiation/chemotherapy and surgery, medical co-morbidities, flap donor and recipient site, postoperative complications and flap survival. The primary study group consisted of patients with malignant head and neck tumours that had undergone surgical salvage with microvascular free flap reconstruction for persistent or recurrent disease following definitive radiation or concomitant chemoradiation as part of their initial treatment modality. An additional group of demographically matched patients who underwent microvascular free flap reconstruction for a variety of problems and who never received radiation were randomly selected to serve as a control group. Patients with inadequate records, a history of osteoradionecrosis and those who received postoperative radiation therapy (re-irradiation) or chemotherapy were excluded from the study. For purposes of comparison, the study cohort was divided, according to radiation treatment: group 1 (control) no radiation; group 2 (XRT) radiation alone; group 3 (ChemoXRT) concomitant chemoradiation. Descriptive statistics were recorded and an analysis of variance was performed to identity differences between the three groups. Fisher’s exact test was used to evaluate the null hypothesis that the occurrence of a particular outcome measure was the same for all three patient groups. The primary outcome measures were type and rate of various postoperative complications. Study cohort characteristics are summarized in Table 1 . A new variable, ‘problem’ was created, which was true if there was partial or complete flap failure. Partial flap loss was defined as complete or partial necrosis of the skin paddle for the fibula osteocutaneous flap or partial necrosis for the anterolateral thigh flap. In order to determine if some combination of data could be found which improved the prediction of a complication, a logistic regression analysis was performed on a number of patient variables, such as sex, radiation status, infection and other complications, and reoperation.

XRT generally consisted of megavoltage photon and electron treatment delivered by means of a standard 3-field isocentric technique. Patients receiving chemotherapy were generally treated with cisplatin. All patients were reconstructed immediately with various microvascular free flaps using standard microinstrumentation. The microvascular anastomosis was performed under a surgical microscope of 10× power, utilizing microinstruments with a low closing pressure and 9-0 suture. The arterial anastomosis was generally performed in end-to-end fashion into a branch of the external carotid artery. The venous anastomosis was generally performed in end-to-side fashion into the internal jugular vein. No systemic heparin or colloid solution was administered routinely, although all patients were treated with aspirin to minimize platelet aggregation postoperatively. During the first 24 h postoperative period, patients were monitored with hourly flap checks using visual and hand-held Doppler examination. Any significant change in turgor, colour or Doppler signal prompted return to the operating room. Salvage reconstructive procedures were variable depending on the prognosis of the patient.