A Prospective Randomized Comparison of Transhiatal and Transthoracic Resection

for Lower-third Esophageal Carcinoma Kent-Man Chu, MB, BS, FRCS(Ed), Simon Y. K. Law, MB, Bchir, FRCS(Ed),

Manson Fok, MB, BS, FRCS(Ed), John Wong, PhD, FRCS(Ed), FRACS, FACS, Hong Kong, China

BACKGROUND: The question whether transhiatal (TH) or transthoracic (lT) resection is most suited for the extirpation of an esophageal cancer re- mains unresolved. The present study compared the two approaches in a prospective randomized manner.

PATIENTS AND METHODS: Thirty-nine patients with carcinoma of the lower third of the esophagus who were clinically fit for either TH or TT resec- tion were prospectively randomized to TH (20 pa- tients) and TT (19 patients) resection. Patients of the two groups were comparable in age, sex, pre- operative tumor staging, and pulmonary and car- diac risks for surgery.

RESULTS: There was no significant difference in the amount of blood loss between the two groups although intraoperative hypotension (systolic ~80 mm Hg) occurred more frequently in the TH group (P <O.OOl). The mean operating time for the TH and lT groups were 174 minutes and 210 minutes, respectively (P <O.OOl). There was no difference in postoperative ventilatory require- ments, cardiopulmonary complication rates, and, mean hospital stay between the two groups. There was no 30-day mortality in either group but there were 3 hospital deaths in the TH group from bronchopneumonia (2 patients) and disseminated malignancy (1 patient). The median survival rates were 18 and 13.5 months, respectively, for the TH and lT groups (f = NS).

CONCLUSIONS: Although there was no demonstra- ble statistical difference in results between TH and TT approaches, the lT approach is preferred as it allowed for a more controlled opera- tion. Am J Sorg. 1997;174:320-324. 83 1997 by Excerpta Medica, Inc.

R esection remains the primary treatment for esopha- geal carcinoma because of its superior and more du- rable quality of swallowing, as compared with non-

operative treatment modalities. In general, the choice of

From the Department of Surgery, The University of Hong Kong, From the Department of Surgery, The University of Hong Kong, Queen Mary Hospital, Hong Kong, China Queen Mary Hospital, Hong Kong, China

Requests for reprints should be addressd to John Wong, PhD, Requests for reprints should be addressd to John Wong, PhD, Department of Surgery, The University of Hong Kong, Queen Mary Department of Surgery, The University of Hong Kong, Queen Mary Hospital, Pokfulam Road, Hong Kong, China. Hospital, Pokfulam Road, Hong Kong, China.

Manuscript submitted April 24, 1996 and accepted in revised Manuscript submitted April 24, 1996 and accepted in revised form September 3, 1996. form September 3, 1996.

operative approach depends on the tumor location, stage of disease, the fitness of the patient, and the experience of the surgical team. The question whether tramhiatal (TH) or transthoracic (TT) resection is most suite’d for the extir- pation of an intrathoracic esophageal cancer remains un- resolved. Both approaches have their supporters. Advocates of TH resection claim the advantages of avoiding a thora- cotomy and its associated respiratory complications, a shorter operating time, and the benefits of placing the upper esophageal anastomosis in the cervical region where a longer resection margin is obtained and ,an anastomotic leakage is less lethal. Supporters of the TT .resection object to the TH approach because it abandons established prin- ciples in cancer surgery, and because the declared advan- tages are either unproven or are achieved in the TT ap- proach.’

A near-total esophagectomy with cervical anastomosis can be achieved by either approach. On the other hand, only the TT approach provides the opportunity to accu- rately stage the disease, and so allow prediction of prognosis and selection of appropriate adjuvant treatment. The TT approach also enables comparisons of results of manage- ment between institutions or between different manage- ment protocols to be made. Moreover, TH dissection of the esophagus is blunt if not blind and could result in tumor rupture and contamination, chylothorax, significant medi- astinal bleeding, airway injury, or recurrent nerve paresis.‘.’

In our practice, patients with high pulmonary risks for a thoracotomy and whose lesions were prim.arily located at the upper or lower thirds of the esophagus were selected for the TH approach, while for more advanced lesions and those in the middle third of the esophagus, a TT approach was nearly always necessary. Although there were numerous retrospective studies comparing TH and TT resection for esophageal cancer, there was only one prospective random- ized trial comparing the two options.3 The :present prospec- tive randomized trial comparing TH with TT resection adds to the proper evaluation of the surgical approaches.

PATIENTS AND METHODS Between March 1990 and November 1994, 463 patients

presented to the Department of Surgery of the University of Hong Kong at Queen Mary Hospital with newly diag- nosed cancer of the esophagus. Of these 463 patients, 344 patients had squamous cell carcinoma of the intrathoracic esophagus distributed to the upper third (30 patients), mid- dle third (233 patients), and lower third (81 patients), re- spectively. All patients were subjected to preoperative as- sessment, which included complete blood counts, liver and -J

)20 0 1997 by Excerpta Medica, Inc. 0002-9610/97/$17.00

All rights reserved. PII SOOO2-9610(97)00105-O

TRANSHIATAL OR TRANSTHORACIC RESECTlON FOR ESOPHAGEAL CANCEWCHU ET Al

renal function tests, arterial blood gases, chest radio~aphy, flexible esophagoscopy with hiopsy, flexible hronchoscopy, barium swallow, cardiopulmonary assessment including lung function tests and electrocardiogram, and ultrasonogp raphy of liver.

After preoperative assessment of the 81 patients with car- cinoma of the lower third of the esophagus, 42 patients were excluded from the study with reasons listed in TabIe I. The remaining 39 patients were prospectively randomized to TH (20 patients) and TT (19 patients) resection. Patients of the two groups were comparable in age, gender, preopera- tive tumor staging, and pulmonary and cardiac risks for sur- gery (Table II). Pulmonary risk was based on history of chronic lung disease, clinical examination, radiology, and special tests. Patients were considered to have moderate pulmonary risk if they had known chronic lung disease with forced expiratory volume at 1 minute (FEV,) of less than 70%. For cardiac assessment, routine cardiac consultation was required and echocardiogram performed on selected pa- tients, with risks again based on history, examination, and special investigation if necessary. The operative technique has been described elsewhere.4 Operative and postoperative data were collected prospectively. The main outcome mea- sures were postoperative complications, mortality, and, sur- vival. Operating time was measured from skin incision to completion of wound closure. Hospital stay was measured from the date of operation to the date of discharge. Hospital mortality was defined as death within the same hospital admission for surgery. Survival was measured from the date of operation.

All continuous values were expressed as mean + SE of mean unless otherwise stated. Univariate analysis was per- formed by Mann-Whimey U test for continuous variables and by chi-square test (with Yates’ correction when appro- priate) for categorical variables. Fisher’s exact test was used instead of chi-square test if any expected cell value in a 2 x 2 cable was less than 5. Statistical analysis was performed with standardized biomedical statistical software (SPSS/ PC+, SPSS, Chicago, Illinois). Statistical significance was taken when P <0.05.

RESULTS Operative Results

All patients underwent the planned procedure (Table III). None of the patients in the TH group required con- version to thoracotomy. Airway injury did not occur in ei+ ther groups. One patient in the TH group developed sig- nificant hemorrhage of 2 L from the aorta during the transmediastinal mobilization of esophagus. The bleeding was conrrolled successfully by packing without the need for thoracotomy.

There was no significant difference in the amount of blood loss between the two groups. On the other hand, intra- operative hypotension, defined as systolic blood pressure he-

low 80 mm Hg, occurred more frequently in the TH group (P <O.OOl). The number of episodes of hypotension per patient was more frequent in the TH group (P <O.OOl) although the mean duration of these hypotensive episodes, while twice as long in the TH group, were not different between the two groups (I’ = 0.34).

The mean operating time was significantly shorter in the TH group than in the TT group (P <O.OOl). Much of the

TABLE I

Reasons for Excluding 42 Patients with Carcinoma of the

Lower Third of the Esophagus from the Study

Number of Reason Patients

Preoperative treatment given Neoadjuvant chemotherapy trial 8 Received radiotherapy elsewhere 2 Received chemoirradiation elsewhere 1

Advanced malignancy Surgical bypass 5 Endoscopic intubation 4 Conservative management 2

Poor general condition Endoscopic intubation 2

Poor pulmonary function Transhiatal resection 4

Previous Billroth II gastrectomy 4 Concomitant malignancy elsewhere

Oropharynx 1

Hypopha~nx 1 Right breast 1 Sigmoid colon 1

Thoracoscopic approach 2 Presence of tuberculous pleural effusion 1 Refused surgery 3

Total 42

TABLE II

Characteristics of Patients Undergoing Transhtatal (TH) or

Transthoracic (TlJ Resection

TH(n=20) lT(n=19) P

Age 60.7 + 1.8 63.9 r’ 1.1 NS Gender (M/F) ia/2 17/2 NS Preoperative staging

Early 4 2

Moderately advanced 13 14

Locally advanced 3 :3 Disseminated 0 0 NS

Pulmonary risk Normal ia 16 Moderate 2 2

High 0 0 NS Cardiac risk

Normal 17 Ia Moderate 3 1

High 0 0 NS

NS = not significant.

additional time required in the TT group was spent in tum- ing and redraping the patient, and in opening and closing

the chest wound. As the resection of the esophagus was extended up to the cervical region in the TH group, the proximal resection margin in the TH group was significantly longer than the TT group (I’ <0.002).

Postoperative Results Postoperative events were similar in both groups (Table

IV). One patient in the TH group and 4 patients in the TT group required postoperative mechanical ventilation (P

THE AMERICAN JOURNAL OF SURGERY@ VOLUME 174 SEPTEMBER 1997 321

ANSHIATAL OR TRANSTHORACIC RESECTION FOR ESOPHAGEAL CANCER=]

= NS). There was no difference between the two groups in terms of pulmonary complications, which included sputum retention, aspiration, bronchopneumonia, pleural effusion, and respiratory failure. The same proportion of patients in each group developed atria1 fibrillation. One patient of each group developed transient hoarseness from injury to recur- rent laryngeal nerve. Two patients in the lT group required re-exploration through the thoracotomy incision; 1 for per- sistent air leak from a small lung laceration and the other for necrosis of the lesser curve of the intrathoracic stomach. Resection of the intrathoracic stomach, cervical esopha- gostomy, and gastrostomy were performed as a salvage pro- cedure. The patient survived, and intestinal continuity was restored 2 months later by colonic interposition. None of the patients in either group developed anastomotic leakage. One patient in the TH group had infection of his abdominal wound and 2 patients in the TT group had infection of their chest wounds (P = NS). Because of 3 more patients re- quiring mechanical ventilation and 2 patients requiring re- exploration, the mean hospital stay of the TT group was longer than the TH group (I’ = NS).

Mortality and Survival There was no 30-day mortality in either group (Table V).

However, there were 3 hospital deaths in the TH group. Two of them died of bronchopneumonia, 1 on day 53 and the other on day 94. Both patients were ex-smokers with normal preoperative pulmonary assessment. Their operative procedures were uneventful with operating times of 200 and 180 minutes; respectively. The longer operating time in the former patient was due to tumor infiltration of aorta and diaphragm. Their operative blood losses were 700 and 650 mL, respectively. They developed postoperative sputum re- tention and succumbed despite tracheostomy, frequent bronchoscopic toilets, and vigorous chest physiotherapy. The third patient developed malignant pleural effusion with confirmation by cytology postoperatively and gradually suc- cumbed on day 75.

The mean follow-up periods of both groups were similar. At the time of analysis, 4 patients of the TH group and 6 patients of the TT group have developed recurrence (P = NS). The sites of recurrence were similar in the two groups. The time to develop recurrence in the two groups of pa- tients was not different statistically. Two patients of each group developed benign anastomotic strictures that re- quired endoscopic dilatation (P = NS). The median sur- vival rates were 16 and 13.5 months, respec:tively, for the TH and TT groups (P = NS).

COMMENTS Following the descriptions of transhiatal approach5 and

abdominal right-side chest approach617 for esophagec- tomy, there have been studies evaluating these two op- tions in managing patients with carcinoma of esophagus. Almost all studies to date (including two from the au- thors’ institution) were retrospective,‘-I5 making them less than adequate for proper and unbiased comparison. Many of these retrospective studies included patients over a relatively long period of time so that factors other than the surgical approach, like improvement in peri- operative and postoperative support, may have contrib- uted to the patient outcome. These associated temporal

-

322 THE AMERICAN JOURNAL OF SURGERY@ VOLUME 174

TABLE Ill

Comparison of lntraoperative Factors and Complications for the Transhiatal (TH) and Transthoracic (TTJ Groups

TH (n = 20) lT (n = 19) P

Urgernt thoractomy 0 -- -

Bronchial tear 0 Cl NS Major hemorrhage 1 (5%) Cl NS lntraoperative

hypotension (~80 mm Hg) 12 (60%) 1 (5%) <O.OOl

Mean episodes 1.85 2 0.39 0.17 !: 0.17 to.001 Mean duration (min) 0.90 t 0.2 0.44 iz 0.4 NS

Mean blood loss (ml) 724 2 58 671 ir 47 NS Mean operating time (min) 174 i 6 210 1: 7 <O.OOl Mean upper margin (cm) 9.1 ? 0.6 6.3 11 0.4 co.002

NS = not significant.

TABLE IV

Comparison of Postoperative Factors and Complications for the Transhiatal (TH) and Transthorac.ic (lT)

(n y20) (n yl9) P

Mechanical ventilation required 1 (5%) 4 (21%) NS Atrial fibrillation 3 (15%) 3 (15.8%) NS Pulmonary complications

Sputum retention 10 (50%) es (42%) NS Aspiration 0 (0%) 1 (5%) NS Bronchopneumonia 2 (10%) Cl (0%) NS Pleural effusion 8 (40%) 6 (32%) NS Respiratory failure 2 (10%) 1 (5%) NS

Hoarseness 1 (5%) 1 (5%) NS Lesser curve necrosis 0 (0%) l* (5%) NS Persistent air leak 0 (0%) 1 l (5%) NS Wound infection 1 (5%) S! (1 1%) NS Mean hospital stay (days) 18 2 2.2 27 t 5.0 NS

* Re-expioration required. MS = not significant.

changes are likely as few centers manage a sufficient num- ber of patient to conduct a prospective ra.ndomized trial over a relatively short period of time. There has been only one prospective randomized trial by Goldminc et al3 comparing the two approaches in 67 patients with car- cinoma of the intrathoracic esophagus. The study con- cluded that both approaches were equally effective.

Because the level of tumor may affect the conduct of the operation and hence its results, the present prospec- tive randomized trial included only patients with lower third squamous cell carcinoma of the esophagus when either approach is justified. In addition, -patients had to be fit enough for either operation before they were in- cluded. Not unexpectedly, the sample siz’e was relatively small. Given the small sample size, there is a possibility of a type II error. We believed, however, that the results were quite representative.

None of the patients in the TH group required conversion to thoracotomy. This was in accordance with previous re- ports that virtually all intraoperative conversion occurred with middle third cancers because of its adherence to me-

SEPTEMBER 1997

diastinal structures. p ” A art from a patient with a serious hemorrhage in the TH group, there were no significant in- traoperative mishaps in both groups of patients. One major difference between the two groups, however, was the more frequent hypotensive episodes in the TH group. This was likely to be the result of the operator’s hand and forearm in the posrerior mediastinum, reducing venous return as well as compressing the heart since the amount of blood loss was similar in both groups. Such a finding has also been dem- onstrated in our earlier retrospective study.“’ The other ma- jor difference was the shorter operating time in the TH group, as explained earlier. It might seem that the shorter operating time would be beneficial to patients with fimited cardiopulmonary reserve. There were, however, no differ- ences between the TH and the TT groups in terms of post- operative pulmonary complications. The 30-day mortality rates and hospital mortality rates were also comparable be- tween the two groups. The same finding was noted by Gold- mint et aL3 On the other hand, patients in the present study were fit to withstand either TH or TT resection, so these results might not be appropriately extrapolated to patients with higher pulmonary risk. Indeed, Bolton et al” dem- onstrated a lower pulmonary complication rate in TH te- section even though the patients had significantly worse mean preoperative American Society of Anesthesiologists (ASA) risk class in comparison with patients undergoing TT resection.

Our retrospective analysis also revealed that pulmonary- related deaths appeared to occur less frequently in TH re- section than TT resection for patients with higher pulmo- nary risks for surgery.‘” Therefore, while there was no difference in pulmonary morbidity in patients who were fit for either approach, it seems advisable to consider the TH approach in patients with high pulmonary risk. Neverthe- less, with the emergence of thoracoscopic esophagec- tow, 18-*’ it may replace the TH approach in patients with lower-third cancers with high pulmonary risk. With tho- racoscopy, esophageal dissection is performed under visual guidance, thus avoiding the risks of blind mobilization. Moreover, it has the potential of reducing the thoracotomy- related pulmonary complications. In patients with obliter- ated pleural cavity, however, this technique is not appli- cable.

The advantage of the n approach is to allow a more complete clearance of involved lymph nodes and, theoret- ically, provide a better chance for long-term survival. Re- sults of nonrandomized and retrospective studies had been controversial in this respect.“*“*‘” For curative resections of middle~third tumors, our previous retrospective analysis did show a significantly better survival after TT resection in comparison with TH resection.z7 For lower-third tumors, much of the TH dissection is done under vision, and lymph node clearance is therefore comparable to the TT approach. This may explain why the present prospective randomized study revealed no difference in the occurrence and the sites of recurrence as well as survival between the two groups. Randomized study by Goldminc et al3 also demonstrated similar survival between the two approaches irrespective of the lymph node status. One may also argue that in patients with possible residual mediastinal disease after TH resec- tion, postoperative radiotherapy may circumvent local re- currence . 24 Postoperative radiotherapy, however, was asso-

TABLE V

Comparison of Mortality, Recurrence, and Survival Figures for the Transhistal (TH) and Transthorecic (rr) Gro&

TH l-r

30-day mortality

In = 20) (n=lQ) P

0 (0%) 0 (0%) NS Hospital mortality Mean follow-up (months) Tumor recurrence

Time of appearance (months)

Sites of recurrence lntrathoracic stomach Tracheobronchial tree

Upper anastomosis Madiastinum Abdominal

Benign anastomotic stricture Median survival (months)

3 (15%) 13.7 f 3.4

4 (20%) 13.0 t 5.5

2 1

0

2 (100%) 16

0 (0%) NS 15.13 i- 3.0 NS

6 (32%) NS 8.l t 3.6 NS

0 2

2(tl%) NS 13.5 NS

ciated with significant morbidity and mortality from irradiation injury of the stomach and should not be given without good indication, which may include incomplete resection.24

Within the limitation of this study, we did not demon- strate any significant difference in results when using either approach for extirpation of a lower-third esophageal carci- noma in a selected group of fit patients. It did appear, how- ever, that the TT approach is preferable as it allowed for a more controIled operation with fewer adverse intraopera- tive cardiovascular events. We would reserve the TH ap- proach for carcinomas of the upper and lower thirds of the esophagus where the dissection of the tumor scan be per- formed under direct vision, and for patients with strong contraindications to thoracotomy, such as the elderly and patients with limited cardiopulmonary reserve, or for pa- tients with an obliterated pleural cavity.16 For more ad- vanced lesions and for tumor in the middle third of the esophagus, a TT approach is nearly always necessary. The value of thoracoscopic esophagectomy awaits further studies.

1. Wang 1. Transhiatal oesophaeectomv for carcinoma of the tho- racic oes&hagus. Br J Surg. i986;73:89-90.

2. Tilanus HW, Hop WCJ, Langenhorst BL, van Lanschor JJB. Eso- phagectomy with or without thoracotomy. Is there arty difference? J Thuc Curdiwac Surg. 1993;105:898-903. 3. Goldminc M, Maddern G, Le-Prise E, et al. Oesophagectomy by a transhiatal approach or thoracotomy: a prospective randomized trial. Br J Surg. 1993;80:367-370. 4. Fok M, Wong J. Cancer of the oesophagus and gastric cardia. Standard oesophagectomy and anastomotic technique. Ann Chir Gynuecot. 1995;84:1i’P-183. 5. Grey-Turner G. Some experiences in the surgery of the esoph- agus. NEJM. 1931;205:657. 6. Tanner NC. The present position of carcinoma of the oesoph- agus. Postgrad Med.!. 1947;23:109-139. 7. Lewis 1. The surgical treatment of carcinoma of the oesophagus with special reference co a new operation for growths of the middle third. Br J Surg. 1946;34:18-31.

THE AMERICAN JOURNAL OF SURGERY* VOLUME 174 SEPTEMBER 1997 323

ANSHIATAL OR TRANSTHORACIC RESECTI

3. Bolton JS, Ochsner JL, Abdoh AA. Surgical management of

zsophageal cancer. A decade of change. Ann Surg. 1994;219:475- 480. 3. Daniel TM, Fleischer KJ, Flanagan TL, et al. Transhiatal eso- phagectomy: a safe alternative for selected patients. Ann Thorns

Surg. 1992;54:686-689. 10. Fok M, Siu KF, Wong J. A comparison of transhiatal and trans- thoracic resection for carcinoma of the thoracic esophagus. Am J Surg. 1989;158:414-419. 11. Fok M, Law S, Stipa F, et al. A comparison of transhiatal and

transthoracic resection for oesophageal carcinoma. Endoscopy. 1993;25:660-663. 12. Gotley DC, Beard J, Cooper MJ, et al. Abdominocervical (tran- shiatal) oesophagectomy in the management of oesophageal carci- noma. Br J Surg. 1990;77:815-819.

13. Mathisen DJ, Grill0 HC, Wilkins E Jr, et al. Transthoracic eso- phagectomy: a safe approach to carcinoma of the esophagus. Ann Thorac Surg. 1988;45:137-143.

14. Orringer MB, Marshall B, Stirling MC. Transhiatal esophagec- tomy for benign and malignant disease. J Thorax Cardiouasc Surg. 1993;105:265-276.

15. Pat M, Basoglu A, Kocak H, et al. Transhiatal versus transtho- racic esophagectomy for esophageal cancer. J Thorax Cnrdiovasc Surg. 1993;106:205-209.

16. Fok M, Wong J. Oesophageal cancer treatment: curative mo- dalities. Eur J Gastroenterol Hepatol. 1994;6:676-683. 17. Bolton JS, Sardi A, Bowen JC, Ellis JK. Transhiatal and trans- thoracic esophagectomy: a comparative study. J Surg Oncol.

1992;51:249-253. 18. Liu HP, Chang CH, Lin PJ, Chang JP. Video-assisted endo- scopic esophagectomy with stapled intrathoracic esophagogastric

anastomosis. World J Surg. 1995;19:745-747. 19. Cuschieri A. Thoracoscopic subtotal oesophagectomy. Endosc Surg Allied Technol. 1994;2:21-25. 20. Gossot D, Fourquier P, Celerier M. Thoracoscopic esophagec- tomy: technique and initial results. Ann Thorax Surg. 1993;56:667-

670. 21. Skinner DB. En bloc resection for neoplasms of the esophagus and cardia. J Thorax Cnrdiovasc Surg. 1983;85:59--71. 22. Horstmann 0, Verreet PR, Becker H, et al. Transhiatal oeso- phagectomy compared with transthoracic resection and systemic lymphadenectomy for the treatment of oesophageal cancer. Eur J

Surg. 1995;161:557-567. 23. Wong J. Surgery in esophageal cancer: how radical should it be. Dig Surg. 1993;10:164-166. 24. Fok M, Sham JST, Choy D, et al. Postoperative radiotherapy for carcinoma of the esophagus: a prospective, randomized con-

trolled study. Surgery. 1993;113:138-147.

324 THE AMERICAN JOURNAL OF SURGERY@ VOLUME 174 SEPTEMBER 1997