Study
Patients
Outcome
TTE
THE
Relative risk for THE
Quality of evidence
Goldminc et al. (1993) [12]
TTE, n = 35
Anastomotic leak
9 %
6 %
0.66 (p = NS)
Level 1B
THE, n = 32
Recurrent laryngeal nerve paralysis
3 %
3 %
1.00 (p = NS)
OR time (hours)
6
4
p = NR
Median overall survival
NR
NR
p = NS
Chu et al. (1997) [13]
TTE, n = 19
Blood loss (L)
0.7
0.7
p < 0.001
Level 2B
THE, n = 20
OR time (hours)
3.5
2.9
p < 0.001
Hospital stay (days)
27
18
p = NS
In-hospital mortality
0
15 %
p = NS
Recurrence
32 %
20 %
0.62 (p = NS)
Median overall survival (years)
1.1
1.3
p = NS
Jacobi et al. (1997) [14]
TTE, n = 16
OR time (hours)
5.5
3.2
p = 0.005
Level 1B
THE, n = 16
Blood loss (L)
2.3
1.0
p = 0.003
Pulmonary complications
50 %
25 %
0.5 (p)
Anastomotic leak
13 %
13 %
1.0 (p = NS)
Thirty-day mortality
6 %
6 %
1.0 (p = NS)
1-year overall survival
77 %
70 %
p = NS
Hulscher et al. (2002) [15]
TTE, n = 114
Pulmonary complications
57 %
27 %
0.47 (p < 0.001)
Level 1B
THE, n = 106;
Anastomotic leak
16 %
14 %
0.88 (p = NS)
80 % distal esophageal,
Vocal cord paralysis
21 %
13 %
0.62 (p = NS)
Chyle leak
10 %
2 %
0.20 (p = 0.02)
20 % cardia;
OR time (hours)
6.0
3.5
p < 0.001
96 % adenocarcinoma
Blood loss (L)
1.9
1.0
p < 0.001
Ventilator time (days)
2
1
p < 0.001
Hospital stay (days)
19
15
p < 0.001
In-hospital mortality
4 %
2 %
0.50 (p = NS)
R0 resection
71 %
72 %
1.01 (p = NS)
Number lymph nodes
31
16
p < 0.001
Recurrence
50 %
58 %
p = NS
Median disease-free survival (years)
1.7
1.4
p = NS
Median overall survival (years)
2.0
1.8
p = NS
Median QALY (years)
1.8
1.5
p = NS
In 1997, Chu and colleagues from Hong Kong reported a randomized clinical trial including patients with lower-third adenocarcinomas that accrued from 1990 to 1994 [13]. Patients were similarly excluded based on use of neoadjuvant therapy, advanced malignancy, concomitant malignancy, pulmonary function, and poor general condition. Investigators demonstrated no significant differences in postoperative complications, although anastomotic leak and chyle leak were not specified. No significant differences in mortality and tumor recurrence were seen during follow-up, although time-to-event analysis was not performed.
Jacobi and colleagues from two German institutions [14] focused on the effects of the type of esophageal resection on perioperative cardiopulmonary function. They randomized 32 patients with resectable esophageal cancer presenting between 1992 and 1995 to either THE or en bloc TTE with cervical anastomosis. Outcome measures included pulmonary arterial catheter measurements, perioperative complications, and survival. There were fewer pulmonary complications after THE, although intraoperative cardiopulmonary parameters did not correlate with complications. Similar rates of leak and postoperative mortality were noted in both groups, with no survival difference seen at a mean follow-up of 1 year.
In 2002, Hulscher and colleagues [15] reported a randomized trial that accrued from two high-volume academic centers in the Netherlands between 1994 and 2000, comparing THE (n = 106) with TTE with extended en bloc lymphadenectomy and cervical esophagogastrostomy (n = 114). Patients had histologically-confirmed adenocarcinoma of the mid to distal esophagus, or gastric cardia adenocarcinoma involving the distal esophagus, and had no distant metastases, celiac or cervical lymph node metastases, with resectable, local disease. Exclusion criteria included neoadjuvant therapy, prior cancer, and involvement of the stomach that would preclude reconstruction with a gastric conduit. Endpoints included morbidity, mortality, survival, and incremental costs per quality-adjusted life-years. Investigators found a significantly higher morbidity associated with transthoracic resection, with longer hospital length of stay, longer ICU stays, and higher costs. There was no significant difference between groups in terms of in-hospital mortality. Although disease-free and overall survival curves trended toward better outcomes with extended transthoracic resection, there were no significant differences between the groups. Further follow-up confirmed no significant overall survival difference between THE and TTE, with 5-year survival rates of 34 and 36 %, respectively [16].
Boshier and colleagues performed a meta-analysis of English-language studies up to 2010 comparing THE with TTE, including the prior randomized controlled trials [17]. A total of 52 studies comprising nearly six thousand patients were included. Lymph node yield was greater during TTE versus THE, although heterogeneity was significant. Operative time and postoperative length of stay were less for THE versus TTE. There were no significant differences in the overall incidence of cardiac complications, although respiratory complications were significantly higher in the TTE group. Early mortality was significantly greater after TTE compared to THE, without significant heterogeneity. Lower rates of anastomotic leak and vocal cord paralysis were noted after TTE compared to THE. Overall analysis of 5-year survival showed no significant differences between procedures.
Chang and colleagues [18] published the largest population-based study examining esophagectomy using the Surveillance, Epidemiology, and End Results (SEER)–Medicare linked database from 1992 to 2002, identifying 225 patients who underwent THE and 643 patients who underwent TTE. Mortality was lower after THE compared to TTE, at 7 % versus 13 % (p = 0.009). Five-year survival was similar for patients after adjusting for stage, patient, and provider factors. Overall, THE may confer an early survival advantage, but long-term survival does not appear to differ by approach. Indeed, the largest reported series of transhiatal esophagectomies (n = 2,007) described an in-hospital mortality of 1 %, a recurrent laryngeal nerve injury rate of 2 %, chylothorax rate of 1, and a 9 % leak rate across the most recent 944 patients, suggesting exceptional results in expert hands [10].
Technique of Anastomosis
Hand-sewn versus stapled esophagogastric anastomosis has been examined in 12 randomized trials and 2 meta-analyses, demonstrating similar anastomotic leak rates between techniques [19–23]. Subgroup analysis has suggested that the use of circular staplers leads to greater stricture rates compared to hand-sewn anastomoses [20]. Walther and colleagues from Sweden confirmed no differences in morbidity, mortality, hospital stay, anastomotic diameter, postoperative weight, or overall survival when comparing cervical hand-sewn versus stapled intrathoracic anastomoses, demonstrating that site of anastomosis does not adversely affect outcome [21].
Conduit Route
Seven randomized controlled trials studied the route of conduit transposition after transhiatal esophagectomy, randomizing between the anterior and posterior mediastinum [24–26]. Bartels and colleagues randomized patients presenting between 1986 and 1989 preoperatively to THE with anterior and posterior reconstruction, studying 96 patients after excluding those with colonic interpositions or who had not achieved an R0 resection [24]. Patients with anteriorly placed conduits had a significantly longer ICU stay, greater reduction in stroke volume index, and greater cardiac complications, with a non-significant increase in hospital mortality compared to those patients who received reconstruction in an orthotopic position. However, further data including a meta-analysis suggests that, while there are trends to improved outcomes with posterior gastric reconstructions, there are no significant differences in complications, mortality, and functional outcomes based on route.
The Role of Minimally Invasive Esophagectomy
To answer the question if minimally-invasive esophagectomy (MIE) confers benefit, Biere and European colleagues performed a multicenter randomized trial comparing TTE with MIE for resectable esophageal cancer [27]. Patient accrual occurred between 2009 and 2011, with exclusion of patients with cervical esophageal cancers, metastatic disease, cT4 disease, or other malignancies. Only hospitals who had performed more than 30 esophagectomies per year could participate, with a requirement that surgeons had experience performing at least 10 MIE and could also perform the procedure open. Open TTE was performed using right thoracotomy and laparotomy with a cervical or intrathoracic anastomosis. MIE was performed using a right thoracoscopy in a prone position, with laparoscopy and a cervical or intrathoracic anastomosis. Reconstruction was using a gastric conduit, with 65 % of randomized patients undergoing cervical esophagogastrostomy. The investigators studied primary outcomes focusing on pulmonary infections, with secondary outcomes including length of stay, quality of life, lymph node retrieval, intraoperative data, and postoperative complications. A total of 56 patients were randomized to open TTE and 59 to MIE, with all patients having completed neoadjuvant chemoradiation (92 %) or chemotherapy (8 %). Tumors were of distal or gastroesophageal junction in 55 % of patients. Among MIE patients, 14 % were converted to open technique in either the abdomen or chest.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
