Tracheoinnominate Artery Fistula After Tracheostomy: A Systematic Literature Review of Associations With Mortality

Abstract

Tracheoinnomate artery fistulas (TIF) are a rare but deadly complication of tracheostomy. This study summarizes TIF cases in the literature and seeks to understand associations with mortality. MEDLINE was searched (up to May 2021) for studies reporting individual clinical and demographic characteristics of patients with TIFs after tracheostomy. Exclusion criteria were: no tracheostomy tube in place at the time of fistula or additional procedures performed at the tracheostomy site. This study was performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. A total of 122 TIF patients from 18 case series and 46 case reports were included. The median age was 40.9 years (range: 4-85, interquartile range: 20.9-60.9 years) and 53.3% (64 of 120) were male. Trauma was the leading indication for tracheostomy (26.2%, 32 of 122), followed by head/neck cancer (20.5%, 25 of 122) and structural malformation (19.7%, 24 of 122). Mortality characteristics are summarized in the Table. Overall mortality rate was 64.8% (79 of 122), with a 15.4% higher rate in females than males (P = .077). There was no significant difference in the median number of days between tracheostomy and TIF detection between those who survived their TIF vs those who died (46 days vs 23 days, P = .077). There was a significant difference in mortality between cases that initially presented with vs without sentinel bleeding (50% [25 of 50] vs 75% [54 of 72], P = .005). The presence or absence of frank hemorrhage also showed a significant difference in mortality (59.6% [53 of 89] vs 78.8% [26 of 33], P = .035). Mortality differed in whether or not the tracheostomy cuff was overinflated for temporary hemostasis during resuscitation (52.2% [35 of 67] vs 80% [44 of 55], P < .001). A total of 27 cases received endovascular management, 2 of which ultimately needed open surgery. No difference in mortality was found between endovascular and surgical management, yet both of these were superior to no treatment (48.6% [36 of 74] vs 89.6% [43 of 48], P < .001). Of the 49 cases definitively managed with open surgery, there were 36 innominate artery resections, 15 tracheal reconstructions, and 12 innominate artery reconstructions. Mortality is a major concern and resuscitation paired with endovascular or surgical intervention is imperative. Investigating sentinel bleeds in a timely fashion and intervening upon hemorrhage with temporary cuff overinflation may lead to improved outcomes.TableMortality characteristicsMortalityP valueDied64.7 (79/122)Sex Male57.8 (37/64).077 Female73.2 (41/56)Age, groups in years 4-2563.8 (23/36).756 26-5061.4 (27/44) >5069.0 (29/42)Sentinel bleed Present50 (25/50).005 Absent75 (54/72)Hemorrhage Present59.6 (53/89).035 Absent78.8 (26/33)Cuff over inflation Performed52.2 (35/67)<.001 Not performed80 (44/55)Endovascular vs surgical management Endovascular48.0 (12/25).938 Surgical49.0 (24/49)Definitive treatment vs no treatment Endovascular or Surgical48.6 (36/74)<.001 No treatment89.6 (43/48)Values represented as percentage and count (condition present/total). Open table in a new tab