The Japanese Journal of Surgery (1992) 22:226- 232
~
SURGERYTODAY © Springer-Verlag 1992
Hemodynamic Changes After Resection of Thoracic Duct for En Bloc Resection of Esophageal Cancer MASAYUKI IMAMURA, l YUTAKA SHIMADA, 1 TAKEHIRO KANDA, 1 TOKIHARU MIYAHARA, MITSUAKI HASHIMOTO, 1 TAKAYOSHI TOBE, 1 TOSHIYUKI ARAI, 2 and YosH~o HATANO2 ~First Department of Surgery, and 2Department of Anesthesiology, Faculty of Medicine, Kyoto University, Kyoto, Japan
Abstract: An en bloc resection of esophageal cancer is one of the most radical forms of esophagectomy, and includes the resection of the thoracic duct, but a relatively high hospital motality rate has been reported. There is very little knowledge on the pathophysiological changes after resection of the thoracic duct. We examined 24 patients who underwent en bloc resection. Some patients developed severe tachycardia or shock postoperatively which subsided after a massive infusion of plasma. Analysis of the fluid balance revealed that much more fluid was necessary during surgery and the postoperative 24 h than in patients treated by a standard esophagectomy. Postoperative lymphangiography or CT revealed abnormal collateral lymphatics around the kidneys or in the pelvic cavity. This suggests the development of the lymphaticovenous shunts, which differed depending on the anatomy of each patient. One patient with chronic hepatitis developed uncontrollable ascites. These are important findings which can hopefully reduce the high rate of hospital death after this operation. Key Words: en bloc resection for esophageal cancer,
postoperative complications, resection of thoracic duct, lymphatico-venous shunt
Introduction
En bloc resection for esophageal cancer is one of the most radical forms of esophagectomy. It was first proposed by Skinner, 1 who resected the azygos vein, pericardial sac, parietal pleura and thoracic duct along with the thoracic esophagus. Recently this technique has been performed in Europe and in Japan, but the hospital motality rate has been relatively high. Little
Reprint requests to: Masayuki Imamura, MD, First Department of Surgery, Faculty of Medicine, Kyoto University, 54 Shogoin Kawaracho, Sakyoku, Kyoto 606, Japan (Received for publication on Dec. 12, 1990)
is known about the pathophysiological changes after resection of the intrathoracic thoracic duct. We examined the hemodynamic changes in 24 patients whose intrathoracic thoracic ducts were resected during esophagectomy for esophageal cancer, and had lymphangiography performed on three patients three months, one year and three years after surgery. Here was report the results of these studies and some observations which may explain the high hospital mortality rate after this surgery.
Patients and Methods
Twenty four patients who had en bloc resection of the esophagus for intrathoracic esophageal cancer (Group A), and 21 patients who had a standard esophagectomy for intrathoracic esophageal cancer without resection of the thoracic duct (Group B) were examined. All of the patients were operated on by the first author of this manuscript. En bloc resection was performed on one patient with Tis, three with T1, nine with T2, two with T3 and nine with T4. A preoperative study revealed no abnormality of liver function except in one patient who had chronic hepatitis. A thoracic duct preserving esophagectomy was performed on one patient with Tis, five with T1, nine with T2, four with T3 and two with T4. In all of the patients a right thoracotomy was performed through the fifth intercostal space, and the upper intrathoracic esophagus was cut after it had been freed from the posterior mediastinum, and the lymph nodes of the posterior mediastinum were dissected either with or without resection of the intrathoracic portion of the thoracic duct. The distal intrathoracic esophagus was wrapped in a nylon bag. The chest was closed, and the abdomen was opened through an upper midline incision. The intrathoracic esophagus was then pulled out through the esophageal fissure after dis-
M. Imamura et al.: Resection of Thoracic Duct
227
section of the lymph nodes along the intraabdominal esophagus. After division of the left gastric artery, a part of the lesser curvature of the stomach was resected on the line from the third branch of the right gastric artery to the top of the gastric fundus, as described by Akiyama, to make a gastric tube. 2 The lymph nodes along the cervical esophagus, the carotid sheath and the supraclavicular portion were then dissected, and anastomosis of the esophagus with the gastric tube which was passed through the retrosternal space was performed by an EEA stapler as described elsewhere. -~ During thoracotomy, all patients were intubated with a left-sided endotracheal twin-lumen tube, and the lung on the operated side was ventilated by high frequency positive pressure ventilation with a Servo 900B ventilator, while the lung on the nonoperated side was ventilated with usual large tidal volume ventilation, as described elsewhere. 4 Anesthesia and intraoperative fluid control were conducted by the same chief anesthesiologist.
Case 1.
65 y.o. VPB I~ ~ ,,, ; ;
During 7"
6-
5"
male
operation
Infusion ---
• ~
Pulse bpm B.P. mmHg ' 160
i
The rate of the intraoperative infusion of electrolyte solutions was about 8 ml/kg/hr during esophagectomy. In the 37 patients who had underwent an esophagectomy during the past three years, the fluid balance both during and after surgery was calculated; 22 patients were in Group A and 15 in Group B. Intraoperative fluid balance was calcurated by subtracting blood loss and urinary output from the volume of electrolyte solution and blood infused during surgery. Postoperative fluid balance was calcurated by subtracting the volume of urinary output and of thoracic and abdominal drainage from the volume infused electrolytes, plasma and blood. Fluid balance per kilogram of body weight per hour was calculated by dividing the water-balance by body weight (kg) and the duration of the operation (hours). Lymphangiography was performed in three patients three months, one year or three years after surgery. Lipiodol®(Kodama Co. Ltd., Osaka, Japan) was injected into a foot lymphatic, and thoracic or abdominal X rays were taken one hour, 6h, 12 h and 24h later. Thoracic and abdominal CT scans were also performed 6h and 12h after the injection of Lipiodol®. The nutritional state of patients surviving more than three years after surgery was estimated by the per cent ideal body weight. Statistical analysis was performed by Wilcoxon's test.
• 140
!
Results •
120
4-
100
Intraoperative and Postoperative Clinical Status After En Bloc Resection
3"
80
In 23 of the 24 Group A patients, retroperitoneal edema was observed when the abdomen was opened following the thoracic procedures. In three patients in Group A, a tachycardia of more than 160 bpm and a fall in blood pressure were observed within 24 hours postoperatively; in two of these patients (Cases 1 and 2) severe tachycardia with shock or arrythmia occurred (Figs 1 and 2). These were corrected by a rapid infusion of plasma. In 13 patients in Group A, tachycardia of more than 110bpm continued for more than three days. None of the Group B patients had any retroperitoneal edema, and tachycardia more than 100bpm which continued for more than three days was observed in only four of the 21 patients.
,Balance ~////~
• 60
~
~'////,
o
40
~ O--9h
I 9-
~ ~27h 18 - 27h
18h
Excretion
1Oh
Fig. 1. Hemodynamic changes in Case i for 27 h after en bloc resection for esophageal cancer. Even with sufficientamounts of intraoperative infusion of free water (8.6l ml/kg per h), he developed tachycardia of more than 120bpm 10h postoperatively, followed by shock. Rapid infusion of plasma restored blood pressure and pulse rate to normal. VPB, Ventricular premature beatsi B.P., blood pressure
Case Reports Case 1. A 65 year-old male, weighing 51 kg, 165 cm in height, underwent en bloc resection of the intrathoracic
228
M. Imamura et al.: Resection of Thoracic Duct Case 2.
59 y.o.
male • Pulse bpm ~ B.F. mmHg 180
PAT, AF During operation 6.5h
160 140
L 5
i
120
Infusion 100
I
f/~
"//~ ~//Z
~elan¢
'30
v
• 20 • 10
0__
• 100
:crel
• 20O - 300
~/,
60 4O
//PAP~
~/'//~
Case 2. A 59 year-old male, weighing 42.5 kg, 161 cm in height, underwent en bloc resection of the esophagus for cancer of the middle third esophagus (T2, N1, M0). The operation took 61/zh. During surgery 3,150ml of electrolyte solution and 950 ml of plasma were infused intravenously; blood loss was 745 ml and urinary output was 750ml, thus, fluid balance during surgery was 2,605mi and water-balance per kg of body weight/h was 9.43mL/kg/hour. One h and ten min after the operation, arrhythmia and tachycardia of more than 140bpm developed, followed by atrial fibrillation. A rapid infusion of plasma 320mi/h with 2rag of verapamil HCL, and 0.5 mg of digoxin lowered the pulse rate to 100 bpm two h later. He had no tachycardia. He received radiation of 58Gy to the neck and mediastinum and was discharged 66 days after surgery.
CVP-
1 ah
¢00
Urine (ml/h)
Fig. 2. Hemodynamic changes in Case 2 after en bloc resection for esophageal cancer. He was infused with sufficient amounts of plasma or electrrolyte solution during surgery, and fluid balance was 9.43 ml/kg h. However, he developed shock within l h after surgery. It was strange that urinary output was more than 100ml/h before the development of shock. With rapid infusion of plasma, he recovered from shock and has never experienced shock since then
esophagus for cancer of the intrathoracic middle third esophagus (T2, N2, M0). The operation took ten hours, during which 5,150ml of electrolyte solution, 1,000 ml of blood and 500 ml of plasma were infused, and blood loss was 1,315ml and urinary output was 900 ml, thus intraoperative fluid balance was 4,435 ml and fluid balance per kg of body weight/h was 8.61 ml/ kg/hour. He developed tachycardia of more than 140bpm with ventricular premature beats l l h after surgery, followed by shock. After massive infusions of plasma 320 ml/h and of electrolyte solution, 120 ml/h, the pulse rate fell to about 80bpm and the blood pressure returned to normal. He received postoperative prophylactic radiation to the neck and mediastinum and was discharged 46 days after surgery.
Case 3. A 69-year-old male underwent en bloc resection of the esophagus for cancer of the lower third of the esophagus. When we opened his abdomen following the intrathoracic procedures, including resection of the intrathoracic part of the thoracic duct, we found that his liver was enlarged and its surface was slightly nodular. Histopathological diagnosis of the liver biopsy was chronic hepatitis. Tachycardia of more than 140bpm began within 24 h after surgery. Rapid infusion of plasma (320 ml/h for three h) and digoxin kept his pulse rate between 100 and 120bpm. The volume of drainage from the right thorax continued to be more than one L per day until the sixth postoperative day. It decreased to 80ml/day on the 14th postoperative day. Oral feeding started on the 10th postoperative day, when ascites started to increase and furosemide or potassium canrenoate was injected every day for more than two months. He finally died of peritonitis and renal failure two months after surgery, though all of the other patients in both Groups were able to be discharged after postoperative radiation.
Analysis of Fluid Balance both During and After Surgery The mean values and standard deviations of the intraand postoperative fluid balance per kilogram of body weight per hour are shown in Table 1 and 2 along with the mean values and standard deviations of body weight and duration of surgery. During the operation the fluid balance was significantly greater for Group A patients than for Group B patients (P < 0.05). From the second to the sixth operative day, there was no significant difference in the fluid balance per kilogram of body weight/h between the two Groups.
M. Imamura et al.: Resection of Thoracic Duct
229
Table 1. Fluid balance during and after surgery in groups A and B
Group A (22) (en bloc resection)
Group B (15) (intact thoracic duct)
21 : 1
11:4
51.4 + 8.9
54.4 _+ 11.7
Sex (male:female) Body weight (kg)
During surgery Duration of surgery (h) Infused solution (ml) Blood loss (ml) Urine (ml) Fluid balance (ml) (ml/kg per h)
8.74 6,052.4 1,209.4 874.1 3,895.7 9.10
+ + + + + +
1.39 1,403.0 622.6 351.4 923.6* 2.35*
8.41 4,819.0 1,235.8 844.7 2,666.9 6.05
+ 1.33 _+ 1,041.5 _+ 451.4 _+ 434.8 _+ 948.0* _+ 2.22*
Time (h) Infused solution (ml) Urine and exudate (ml) Fluid balance (ml) (ml/kg per h)
12.24 1,750.1 1,309.2 430.7 0.65
+ 1.54 + 727.7 _+ 597.5 + 523.8 + 0.77
13.27 1,677.9 1,223.7 444.2 0.70
__+1.00 _+ 5.3.7 _+ 4.7.6 _+ 5.8.5 _+ 0.80
Firstpostoperative day Infused solution (ml) Urine and exudate (ml) Fluid balance (ml) (ml/kg per h)
2,354.1 2,418.6 -64.4 -0.09
+ 588.9 + 806.1 _+ 1143.4 + 0.97
2,057.2 2,085.5 -30.1 0.0
_+ 478.4 _+ 478.5 _+ 725.4 _+ 0.54
48 h After start of surgery Fluid balance (ml) (ml/kg per h)
4,262.1 + 1,280.1" 1.74 + 0.51'
On day of surgery
3,081.0 _+ 1,107.2" 1.21 _+ 0.44*
* P < 0.01
Table 2. Postoperative changes of fluid balance in groups A
and B Group A (23)
Group B (15)
Fluid balance (ml) 48 h After surgery 2nd p.o.day 3rd p.o.day 4th p.o.day 5th p.o.day 6th p.o.day
4,262.1 -158.4 331.3 608.0 770.0 757.1
+ 1,280.1" _+ 758.0 + 436.3 + 628.4 _+493.5 _+ 419.7
3,081.0 -231.8 -128.2 335.7 628.1 756.7
Fluid balance (ml/kg per h) 1.74 + 0.51' -0.14 _+ 0.65 0.28 _+ 0.40 0.52 _+ 0.57 0.65 _+ 0.47 0.65 + 0.30
48 h After surgery 2nd p.o.day 3rd p.o.day 4th p.o.day 5th p.o.day 6th p.o.day
+ 1,107.2" + 564.0 + 858.4 _+ 612.7 + 548.9 + 515.5
1.21 + 0.44* -0.18 _+ 0.44 -0.08 _+ 0.69 0.27 _+ 0.51 0.53 _+ 0.47 0.63 + 0.30
~P < 0.01 p.o.day, Postoperative day
the ligated intrathoracic thoracic duct (Fig. 3), but abnormal accumulation of Lipiodol R on the capsules of the kidneys was visualized by CT scan 6 h after the injection of Lipiodol R in a patient examined one year postoperatively (Fig. 4). In another patient who was examined three years postoperatively, abnormal lymphatic pathways were visualized in the right pelvic cavity (Fig. 5). These findings suggest the development of retroperitoneal abdominal lymphatico-venous shunts after resection of the intrathoracic thoracic duct. The nutritional state of five patients who survived more than three years after resection of the thoracic duct was good. Body weight, appetite and bowel m o v e m e n t s were not different from those of the patients treated with standard esophagectomy for esophageal cancer. Neither ascites nor e d e m a of the lower extremities developed in these patients (Table 3).
Discussion
Analysis of Lymphatic Vessels by CT or L ymphangiography ACT scan during the lymphangiography of three patients in G r o u p A did not reveal any lymphatic collaterals in the mediastinum or chest wall above
The en bloc resection of the esophagus for esophageal cancer was introduced by Skinner I in 1983. H e resected the thoracic duct as well as the azygos vein and part of the pericardium, aiming at a radical en bloc resection with the idea that these tissues have developed em-
230
M. Imamura et al.: Resection of Thoracic Duct
Fig. 3. CT scans during lymphangiography of a patient who has been alive 1 year since en bloc resection of the esophagus. CT scan 12h after the injection of Lipiodol into a foot lymphatic vessel revealed no development of lymphatic collaterals in the chest wall or the mediastinum cranial to the ligated portion of the thoracic duct (black arrow) above the chyle cistern. Stagnation of Lipiodol in the intra-abdominal lymphatics is clearly seen in the lower two films. Post op., Postoperatively
Fig. 4. CT scan during lymphangiography in a patient 1 year after en bloc resection of the esophagus. 6 h after the injection of Lipiodol into a foot lymphatic vessel an accumulation of Lipiopdol was seen around the capsules of the kidneys (white arrows). No collateral pathways can be seen above the ligated portion of the thoracic duct. The abdominal lymphatics were not dilated, and some of them seemed to go toward the kidneys (black arrows). Post op., Postoperatively
M. Imamura et al.: Resection of Thoracic Duct
231
Table 3. Nutritional status of patients surviving more than 3 years after resection of the thoracic duct
Number 1 2 3 4 5
Age (years)
Histol. state
Years
Sex
55 49 60 75 53
M F M M F
3 1 1 3 2
6 4 3.8 3.5 3.3
Height (cm)
Weight (kg)
Nutrit. state
Appetite
165 142 168 185 140
50.0 44.3 61.2 46.0 39.1
fair fair fair slender fair
good good good good good
Ascites
Diarrhea
0 0 0 0 0
0 0 0 0 0
Histol., Histological; Nutrit., nutritional; M, male; F, female
Fig. 5. Lymphangiography of a patient 3 years after en bloc resection of the esophagus. The film was taken 6 h after the injection of Lipiodol. Abdominal lymphatics are not dilated, and abnormal lymphatic pathways can be seen in the right pelvic cavity suggesting the presence of a lymphatico-venous shunt in the pelvic cavity (lower three white arrows in the left photo). An accumulation of Lipiodol at the level of the second lumbar vertebra suggests opening of the lymphatics into the inferior vena cava through a lymphatico-venous shunt (upper white arrow in the left photo and black arrow on the right photo). Post op., Postoperatively bryologically from the same part of the foregut. He described his technique to be a modificaton of Logan's approach, 5 and his operative mortality improved from Logan's 21 per cent to 11 per cent, but he experienced a number of early postoperative complications: troublesome cardiac arrhythmias, massive retention of pleural effusion, and cardic failure. Siewert et al. 6 reported a low rate of mortality, 6 per cent, after en bloc resection, but also a relatively high rate of postoperative complications, including low-output-syndrome and peritonitis due to infected lymphatic fistulae. In order to lower the postoperative cardiovascular and lymphatic omplications after en bloc resection, it seems necessary to better elucidate the hemodynamic changes which take place.
Only a few reports have previously described the changes of lymphatic flow or the hemodynamic changes in humans after resection of the thoracic duct above the chyle cistern. Job conducted a detailed study on the lymphatico-venous communications in the rat by injecting Berlin blue gelatin and India ink. 7 He proved that in the rat there are a few lymphatico-venous communications other than with the subclavian vein: connections with the inferior vena cava in 48 per cent, with the renal vein in 8 per cent, with the portal vein in 27 per cent and rarely with the iliolumbar vein. In 1963 Threefoot et al. studied the same subject with plastic corrosion models in rats and described lymphaticovenous communications of the left renal vein, adrenal vein, vena cava, as well as the left jugular vein. He also studied the human lymphatico-venous communications between the inferior vena cava and the lymphatics by postmortem lymphangiography, and described communications at a few levels below the right renal vein. s Neyazaki et al. used lymphangiography to study collateral lymphatico-venous communications after obstruction of the thoracic duct in dogs. 9 He observed on immediate enlargement of mesenteric lymph nodes and dilatation of the lymphatics following central lymphatic obstruction which persisted for at least three weeks. He visualized the development of lymphaticovenous communications of the left renal vein, inferior vena cava and the inguinal or iliac lymph nodes in dogs and assumed that the establishment of lymphaticovenous communications relieved the obstruction by restoring the intralymphtic pressure to normal. Our study demonstrated the development of lymphatic collateral pathways around the kidneys and the iliac cavity more than one year after resection of the thoracic duct. These findings suggest the existence and dilatation of lymphatico-venous communications in the retroperirenal region or in the iliac cavity, which at many points coincide with the results of the studies performed with imaging techniques in rats and dogs. The fact that in all except one patient retroperitoneal edema was observed following resection of the thoracic duct suggests that in most humans lymphatico-venous
232 anastomoses are not large enough to compensate for the abrupt obstruction of the thoracic duct, but in a few persons, lymphatico-venous anastomoses might become naturally wide enough to compensate for the abrupt obstruction of the thoracic duct. In Case 1 and Case 2, severe shock took place immediately after en bloc resection of the esophagus when the standard rate (8ml/kg/h) of infusion of electrolyte solution or plasma was used during surgery. At that time we first suspected cardiac failure, because we did not appreciate the great decrease in circulating plasma volume after the resection of the thoracic duct. Massive infusions of plasma and electrolyte solution counteracted shock in these patients. These experiences led us to study the fluid balance of patients who underwent en bloc resection. Our analysis of the hemodynamic changes during surgery and within 48 hours after resection of the thoracic duct clearly showed that in order to maintain normal postoperative hemodynamics, significantly larger amounts of plasma or electrolyte solutionr are necessary after en bloc resection than in patients whose thoracic duct are not resected (P < 0.01). We also observed that after en bloc resection, data of blood gas analysis within 12h after surgery were better than that in patients who had the usual type of esophagectomy, probably due to intravascular hypovolemia. In spite of intravascular hypovolemia, urinary output immediatey after the surgery was increased (Case 2). Urinary osmotic pressure was within normal ranges. We cannot explain this phenomenon, which led us to a delayed diagnosis of hypovolemia. After 48h, large intravenous infusions were no longer necessary to maintain a normal hemodynamic state. This early recovery from the abrupt abnormal hemodynamic changes induced by resection of the thoracic duct may be related to the fact that these patients are not permitted any oral food intake for at least ten days. After the intravascular hypovolemia has been corrected by rapid infusion of plasma, standard amounts of infusion of plasma or electrolytes seem to be able to maintain a stable hemodynamic state. In the patient with chronic hepatitis, however, (Case 3), ascites increased progressively after the start of oral food intake. This retention of ascites could not be
M. Imamura et al.: Resection of Thoracic Duct controlled and we had to thus restrict oral intake and continue intravenous hyperalimentation. He eventually died of peritonitis and renal failure. Thus, in patients with hepatic failure, such as chronic hepatitis or cirrhosis, postoperative fluid management and nutrition are very difficult problems. We think that en bloc resection of the esophagus should not be performed in these patients. Resection of the thoracic duct does not seem to influence the hemodynamics or the lymphatic dynamics in the late postoperative days, since the nutritional state of patients who have survived more than three years postoperative does not differ from that of patients who have undergone a standard esophagectomy. Therefore, we conclude that en bloc resection for esophageal cancer should be performed in selected patients with no liver disease under the careful management for the hemodynamic changes due to resection of the thoracic duct.
References 1. Skinner DB (1983) En bloc resection for neoplasms of the esophagus and cardia. J Thorac Cardiovasc Surg 85:59-71 2. Akiyama H, Tsurumaru M, Watanabe G, et al. (1984) Development of surgery for carcinoma of the esophagus. Am J Surg 147:9-16 3. Imamura M, Ohishi K, Tobe T (1987) Retrosternal esophagogastrostomy with the EEA stapler. Surg Gynocl Obst 161:364-366 4. Imamura M, Yanagibashi K, Tobe T et al. (1988) Transthoracic resection of esophageal cancer in patients with pulmonary dysfunction - - usefulness of high frequency ventilation during thoracotomy. Ann Surg 208:601-605 5. Logan A (1963) The surgical treatment of carcinoma of the esophagus and cardia. J Thorac Cardiovasc Surg 46:150-161 6. Siewert JR, Holsher AH, Roder J, Bartels H (1988) En-bloc Resection der Speiserohre beim Oesophaguscarcinom. Langenbecks Arch Chir 373:367-376 7. Job TT (1918) Lymphatico-venous communications in the common rat and their significance. Am J Anat 24:467-485 8. Threefoot SA, Kent WT, Hatchett BF (1963) Lymphaticovenous and lymphaticolymphatic communications demonstrated by plastic corrosion models of rats and postmortem lymphangiography in man. J Lab Clin Med 61:9-22 9. Neyazaki T, Kupic EA, Marshall WH, Abrams HL (1965) Collateral lymphatico-venous communications after experimental obstruction of the thoracic duct. Radiology 85:423-432
~
SURGERYTODAY © Springer-Verlag 1992
Hemodynamic Changes After Resection of Thoracic Duct for En Bloc Resection of Esophageal Cancer MASAYUKI IMAMURA, l YUTAKA SHIMADA, 1 TAKEHIRO KANDA, 1 TOKIHARU MIYAHARA, MITSUAKI HASHIMOTO, 1 TAKAYOSHI TOBE, 1 TOSHIYUKI ARAI, 2 and YosH~o HATANO2 ~First Department of Surgery, and 2Department of Anesthesiology, Faculty of Medicine, Kyoto University, Kyoto, Japan
Abstract: An en bloc resection of esophageal cancer is one of the most radical forms of esophagectomy, and includes the resection of the thoracic duct, but a relatively high hospital motality rate has been reported. There is very little knowledge on the pathophysiological changes after resection of the thoracic duct. We examined 24 patients who underwent en bloc resection. Some patients developed severe tachycardia or shock postoperatively which subsided after a massive infusion of plasma. Analysis of the fluid balance revealed that much more fluid was necessary during surgery and the postoperative 24 h than in patients treated by a standard esophagectomy. Postoperative lymphangiography or CT revealed abnormal collateral lymphatics around the kidneys or in the pelvic cavity. This suggests the development of the lymphaticovenous shunts, which differed depending on the anatomy of each patient. One patient with chronic hepatitis developed uncontrollable ascites. These are important findings which can hopefully reduce the high rate of hospital death after this operation. Key Words: en bloc resection for esophageal cancer,
postoperative complications, resection of thoracic duct, lymphatico-venous shunt
Introduction
En bloc resection for esophageal cancer is one of the most radical forms of esophagectomy. It was first proposed by Skinner, 1 who resected the azygos vein, pericardial sac, parietal pleura and thoracic duct along with the thoracic esophagus. Recently this technique has been performed in Europe and in Japan, but the hospital motality rate has been relatively high. Little
Reprint requests to: Masayuki Imamura, MD, First Department of Surgery, Faculty of Medicine, Kyoto University, 54 Shogoin Kawaracho, Sakyoku, Kyoto 606, Japan (Received for publication on Dec. 12, 1990)
is known about the pathophysiological changes after resection of the intrathoracic thoracic duct. We examined the hemodynamic changes in 24 patients whose intrathoracic thoracic ducts were resected during esophagectomy for esophageal cancer, and had lymphangiography performed on three patients three months, one year and three years after surgery. Here was report the results of these studies and some observations which may explain the high hospital mortality rate after this surgery.
Patients and Methods
Twenty four patients who had en bloc resection of the esophagus for intrathoracic esophageal cancer (Group A), and 21 patients who had a standard esophagectomy for intrathoracic esophageal cancer without resection of the thoracic duct (Group B) were examined. All of the patients were operated on by the first author of this manuscript. En bloc resection was performed on one patient with Tis, three with T1, nine with T2, two with T3 and nine with T4. A preoperative study revealed no abnormality of liver function except in one patient who had chronic hepatitis. A thoracic duct preserving esophagectomy was performed on one patient with Tis, five with T1, nine with T2, four with T3 and two with T4. In all of the patients a right thoracotomy was performed through the fifth intercostal space, and the upper intrathoracic esophagus was cut after it had been freed from the posterior mediastinum, and the lymph nodes of the posterior mediastinum were dissected either with or without resection of the intrathoracic portion of the thoracic duct. The distal intrathoracic esophagus was wrapped in a nylon bag. The chest was closed, and the abdomen was opened through an upper midline incision. The intrathoracic esophagus was then pulled out through the esophageal fissure after dis-
M. Imamura et al.: Resection of Thoracic Duct
227
section of the lymph nodes along the intraabdominal esophagus. After division of the left gastric artery, a part of the lesser curvature of the stomach was resected on the line from the third branch of the right gastric artery to the top of the gastric fundus, as described by Akiyama, to make a gastric tube. 2 The lymph nodes along the cervical esophagus, the carotid sheath and the supraclavicular portion were then dissected, and anastomosis of the esophagus with the gastric tube which was passed through the retrosternal space was performed by an EEA stapler as described elsewhere. -~ During thoracotomy, all patients were intubated with a left-sided endotracheal twin-lumen tube, and the lung on the operated side was ventilated by high frequency positive pressure ventilation with a Servo 900B ventilator, while the lung on the nonoperated side was ventilated with usual large tidal volume ventilation, as described elsewhere. 4 Anesthesia and intraoperative fluid control were conducted by the same chief anesthesiologist.
Case 1.
65 y.o. VPB I~ ~ ,,, ; ;
During 7"
6-
5"
male
operation
Infusion ---
• ~
Pulse bpm B.P. mmHg ' 160
i
The rate of the intraoperative infusion of electrolyte solutions was about 8 ml/kg/hr during esophagectomy. In the 37 patients who had underwent an esophagectomy during the past three years, the fluid balance both during and after surgery was calculated; 22 patients were in Group A and 15 in Group B. Intraoperative fluid balance was calcurated by subtracting blood loss and urinary output from the volume of electrolyte solution and blood infused during surgery. Postoperative fluid balance was calcurated by subtracting the volume of urinary output and of thoracic and abdominal drainage from the volume infused electrolytes, plasma and blood. Fluid balance per kilogram of body weight per hour was calculated by dividing the water-balance by body weight (kg) and the duration of the operation (hours). Lymphangiography was performed in three patients three months, one year or three years after surgery. Lipiodol®(Kodama Co. Ltd., Osaka, Japan) was injected into a foot lymphatic, and thoracic or abdominal X rays were taken one hour, 6h, 12 h and 24h later. Thoracic and abdominal CT scans were also performed 6h and 12h after the injection of Lipiodol®. The nutritional state of patients surviving more than three years after surgery was estimated by the per cent ideal body weight. Statistical analysis was performed by Wilcoxon's test.
• 140
!
Results •
120
4-
100
Intraoperative and Postoperative Clinical Status After En Bloc Resection
3"
80
In 23 of the 24 Group A patients, retroperitoneal edema was observed when the abdomen was opened following the thoracic procedures. In three patients in Group A, a tachycardia of more than 160 bpm and a fall in blood pressure were observed within 24 hours postoperatively; in two of these patients (Cases 1 and 2) severe tachycardia with shock or arrythmia occurred (Figs 1 and 2). These were corrected by a rapid infusion of plasma. In 13 patients in Group A, tachycardia of more than 110bpm continued for more than three days. None of the Group B patients had any retroperitoneal edema, and tachycardia more than 100bpm which continued for more than three days was observed in only four of the 21 patients.
,Balance ~////~
• 60
~
~'////,
o
40
~ O--9h
I 9-
~ ~27h 18 - 27h
18h
Excretion
1Oh
Fig. 1. Hemodynamic changes in Case i for 27 h after en bloc resection for esophageal cancer. Even with sufficientamounts of intraoperative infusion of free water (8.6l ml/kg per h), he developed tachycardia of more than 120bpm 10h postoperatively, followed by shock. Rapid infusion of plasma restored blood pressure and pulse rate to normal. VPB, Ventricular premature beatsi B.P., blood pressure
Case Reports Case 1. A 65 year-old male, weighing 51 kg, 165 cm in height, underwent en bloc resection of the intrathoracic
228
M. Imamura et al.: Resection of Thoracic Duct Case 2.
59 y.o.
male • Pulse bpm ~ B.F. mmHg 180
PAT, AF During operation 6.5h
160 140
L 5
i
120
Infusion 100
I
f/~
"//~ ~//Z
~elan¢
'30
v
• 20 • 10
0__
• 100
:crel
• 20O - 300
~/,
60 4O
//PAP~
~/'//~
Case 2. A 59 year-old male, weighing 42.5 kg, 161 cm in height, underwent en bloc resection of the esophagus for cancer of the middle third esophagus (T2, N1, M0). The operation took 61/zh. During surgery 3,150ml of electrolyte solution and 950 ml of plasma were infused intravenously; blood loss was 745 ml and urinary output was 750ml, thus, fluid balance during surgery was 2,605mi and water-balance per kg of body weight/h was 9.43mL/kg/hour. One h and ten min after the operation, arrhythmia and tachycardia of more than 140bpm developed, followed by atrial fibrillation. A rapid infusion of plasma 320mi/h with 2rag of verapamil HCL, and 0.5 mg of digoxin lowered the pulse rate to 100 bpm two h later. He had no tachycardia. He received radiation of 58Gy to the neck and mediastinum and was discharged 66 days after surgery.
CVP-
1 ah
¢00
Urine (ml/h)
Fig. 2. Hemodynamic changes in Case 2 after en bloc resection for esophageal cancer. He was infused with sufficient amounts of plasma or electrrolyte solution during surgery, and fluid balance was 9.43 ml/kg h. However, he developed shock within l h after surgery. It was strange that urinary output was more than 100ml/h before the development of shock. With rapid infusion of plasma, he recovered from shock and has never experienced shock since then
esophagus for cancer of the intrathoracic middle third esophagus (T2, N2, M0). The operation took ten hours, during which 5,150ml of electrolyte solution, 1,000 ml of blood and 500 ml of plasma were infused, and blood loss was 1,315ml and urinary output was 900 ml, thus intraoperative fluid balance was 4,435 ml and fluid balance per kg of body weight/h was 8.61 ml/ kg/hour. He developed tachycardia of more than 140bpm with ventricular premature beats l l h after surgery, followed by shock. After massive infusions of plasma 320 ml/h and of electrolyte solution, 120 ml/h, the pulse rate fell to about 80bpm and the blood pressure returned to normal. He received postoperative prophylactic radiation to the neck and mediastinum and was discharged 46 days after surgery.
Case 3. A 69-year-old male underwent en bloc resection of the esophagus for cancer of the lower third of the esophagus. When we opened his abdomen following the intrathoracic procedures, including resection of the intrathoracic part of the thoracic duct, we found that his liver was enlarged and its surface was slightly nodular. Histopathological diagnosis of the liver biopsy was chronic hepatitis. Tachycardia of more than 140bpm began within 24 h after surgery. Rapid infusion of plasma (320 ml/h for three h) and digoxin kept his pulse rate between 100 and 120bpm. The volume of drainage from the right thorax continued to be more than one L per day until the sixth postoperative day. It decreased to 80ml/day on the 14th postoperative day. Oral feeding started on the 10th postoperative day, when ascites started to increase and furosemide or potassium canrenoate was injected every day for more than two months. He finally died of peritonitis and renal failure two months after surgery, though all of the other patients in both Groups were able to be discharged after postoperative radiation.
Analysis of Fluid Balance both During and After Surgery The mean values and standard deviations of the intraand postoperative fluid balance per kilogram of body weight per hour are shown in Table 1 and 2 along with the mean values and standard deviations of body weight and duration of surgery. During the operation the fluid balance was significantly greater for Group A patients than for Group B patients (P < 0.05). From the second to the sixth operative day, there was no significant difference in the fluid balance per kilogram of body weight/h between the two Groups.
M. Imamura et al.: Resection of Thoracic Duct
229
Table 1. Fluid balance during and after surgery in groups A and B
Group A (22) (en bloc resection)
Group B (15) (intact thoracic duct)
21 : 1
11:4
51.4 + 8.9
54.4 _+ 11.7
Sex (male:female) Body weight (kg)
During surgery Duration of surgery (h) Infused solution (ml) Blood loss (ml) Urine (ml) Fluid balance (ml) (ml/kg per h)
8.74 6,052.4 1,209.4 874.1 3,895.7 9.10
+ + + + + +
1.39 1,403.0 622.6 351.4 923.6* 2.35*
8.41 4,819.0 1,235.8 844.7 2,666.9 6.05
+ 1.33 _+ 1,041.5 _+ 451.4 _+ 434.8 _+ 948.0* _+ 2.22*
Time (h) Infused solution (ml) Urine and exudate (ml) Fluid balance (ml) (ml/kg per h)
12.24 1,750.1 1,309.2 430.7 0.65
+ 1.54 + 727.7 _+ 597.5 + 523.8 + 0.77
13.27 1,677.9 1,223.7 444.2 0.70
__+1.00 _+ 5.3.7 _+ 4.7.6 _+ 5.8.5 _+ 0.80
Firstpostoperative day Infused solution (ml) Urine and exudate (ml) Fluid balance (ml) (ml/kg per h)
2,354.1 2,418.6 -64.4 -0.09
+ 588.9 + 806.1 _+ 1143.4 + 0.97
2,057.2 2,085.5 -30.1 0.0
_+ 478.4 _+ 478.5 _+ 725.4 _+ 0.54
48 h After start of surgery Fluid balance (ml) (ml/kg per h)
4,262.1 + 1,280.1" 1.74 + 0.51'
On day of surgery
3,081.0 _+ 1,107.2" 1.21 _+ 0.44*
* P < 0.01
Table 2. Postoperative changes of fluid balance in groups A
and B Group A (23)
Group B (15)
Fluid balance (ml) 48 h After surgery 2nd p.o.day 3rd p.o.day 4th p.o.day 5th p.o.day 6th p.o.day
4,262.1 -158.4 331.3 608.0 770.0 757.1
+ 1,280.1" _+ 758.0 + 436.3 + 628.4 _+493.5 _+ 419.7
3,081.0 -231.8 -128.2 335.7 628.1 756.7
Fluid balance (ml/kg per h) 1.74 + 0.51' -0.14 _+ 0.65 0.28 _+ 0.40 0.52 _+ 0.57 0.65 _+ 0.47 0.65 + 0.30
48 h After surgery 2nd p.o.day 3rd p.o.day 4th p.o.day 5th p.o.day 6th p.o.day
+ 1,107.2" + 564.0 + 858.4 _+ 612.7 + 548.9 + 515.5
1.21 + 0.44* -0.18 _+ 0.44 -0.08 _+ 0.69 0.27 _+ 0.51 0.53 _+ 0.47 0.63 + 0.30
~P < 0.01 p.o.day, Postoperative day
the ligated intrathoracic thoracic duct (Fig. 3), but abnormal accumulation of Lipiodol R on the capsules of the kidneys was visualized by CT scan 6 h after the injection of Lipiodol R in a patient examined one year postoperatively (Fig. 4). In another patient who was examined three years postoperatively, abnormal lymphatic pathways were visualized in the right pelvic cavity (Fig. 5). These findings suggest the development of retroperitoneal abdominal lymphatico-venous shunts after resection of the intrathoracic thoracic duct. The nutritional state of five patients who survived more than three years after resection of the thoracic duct was good. Body weight, appetite and bowel m o v e m e n t s were not different from those of the patients treated with standard esophagectomy for esophageal cancer. Neither ascites nor e d e m a of the lower extremities developed in these patients (Table 3).
Discussion
Analysis of Lymphatic Vessels by CT or L ymphangiography ACT scan during the lymphangiography of three patients in G r o u p A did not reveal any lymphatic collaterals in the mediastinum or chest wall above
The en bloc resection of the esophagus for esophageal cancer was introduced by Skinner I in 1983. H e resected the thoracic duct as well as the azygos vein and part of the pericardium, aiming at a radical en bloc resection with the idea that these tissues have developed em-
230
M. Imamura et al.: Resection of Thoracic Duct
Fig. 3. CT scans during lymphangiography of a patient who has been alive 1 year since en bloc resection of the esophagus. CT scan 12h after the injection of Lipiodol into a foot lymphatic vessel revealed no development of lymphatic collaterals in the chest wall or the mediastinum cranial to the ligated portion of the thoracic duct (black arrow) above the chyle cistern. Stagnation of Lipiodol in the intra-abdominal lymphatics is clearly seen in the lower two films. Post op., Postoperatively
Fig. 4. CT scan during lymphangiography in a patient 1 year after en bloc resection of the esophagus. 6 h after the injection of Lipiodol into a foot lymphatic vessel an accumulation of Lipiopdol was seen around the capsules of the kidneys (white arrows). No collateral pathways can be seen above the ligated portion of the thoracic duct. The abdominal lymphatics were not dilated, and some of them seemed to go toward the kidneys (black arrows). Post op., Postoperatively
M. Imamura et al.: Resection of Thoracic Duct
231
Table 3. Nutritional status of patients surviving more than 3 years after resection of the thoracic duct
Number 1 2 3 4 5
Age (years)
Histol. state
Years
Sex
55 49 60 75 53
M F M M F
3 1 1 3 2
6 4 3.8 3.5 3.3
Height (cm)
Weight (kg)
Nutrit. state
Appetite
165 142 168 185 140
50.0 44.3 61.2 46.0 39.1
fair fair fair slender fair
good good good good good
Ascites
Diarrhea
0 0 0 0 0
0 0 0 0 0
Histol., Histological; Nutrit., nutritional; M, male; F, female
Fig. 5. Lymphangiography of a patient 3 years after en bloc resection of the esophagus. The film was taken 6 h after the injection of Lipiodol. Abdominal lymphatics are not dilated, and abnormal lymphatic pathways can be seen in the right pelvic cavity suggesting the presence of a lymphatico-venous shunt in the pelvic cavity (lower three white arrows in the left photo). An accumulation of Lipiodol at the level of the second lumbar vertebra suggests opening of the lymphatics into the inferior vena cava through a lymphatico-venous shunt (upper white arrow in the left photo and black arrow on the right photo). Post op., Postoperatively bryologically from the same part of the foregut. He described his technique to be a modificaton of Logan's approach, 5 and his operative mortality improved from Logan's 21 per cent to 11 per cent, but he experienced a number of early postoperative complications: troublesome cardiac arrhythmias, massive retention of pleural effusion, and cardic failure. Siewert et al. 6 reported a low rate of mortality, 6 per cent, after en bloc resection, but also a relatively high rate of postoperative complications, including low-output-syndrome and peritonitis due to infected lymphatic fistulae. In order to lower the postoperative cardiovascular and lymphatic omplications after en bloc resection, it seems necessary to better elucidate the hemodynamic changes which take place.
Only a few reports have previously described the changes of lymphatic flow or the hemodynamic changes in humans after resection of the thoracic duct above the chyle cistern. Job conducted a detailed study on the lymphatico-venous communications in the rat by injecting Berlin blue gelatin and India ink. 7 He proved that in the rat there are a few lymphatico-venous communications other than with the subclavian vein: connections with the inferior vena cava in 48 per cent, with the renal vein in 8 per cent, with the portal vein in 27 per cent and rarely with the iliolumbar vein. In 1963 Threefoot et al. studied the same subject with plastic corrosion models in rats and described lymphaticovenous communications of the left renal vein, adrenal vein, vena cava, as well as the left jugular vein. He also studied the human lymphatico-venous communications between the inferior vena cava and the lymphatics by postmortem lymphangiography, and described communications at a few levels below the right renal vein. s Neyazaki et al. used lymphangiography to study collateral lymphatico-venous communications after obstruction of the thoracic duct in dogs. 9 He observed on immediate enlargement of mesenteric lymph nodes and dilatation of the lymphatics following central lymphatic obstruction which persisted for at least three weeks. He visualized the development of lymphaticovenous communications of the left renal vein, inferior vena cava and the inguinal or iliac lymph nodes in dogs and assumed that the establishment of lymphaticovenous communications relieved the obstruction by restoring the intralymphtic pressure to normal. Our study demonstrated the development of lymphatic collateral pathways around the kidneys and the iliac cavity more than one year after resection of the thoracic duct. These findings suggest the existence and dilatation of lymphatico-venous communications in the retroperirenal region or in the iliac cavity, which at many points coincide with the results of the studies performed with imaging techniques in rats and dogs. The fact that in all except one patient retroperitoneal edema was observed following resection of the thoracic duct suggests that in most humans lymphatico-venous
232 anastomoses are not large enough to compensate for the abrupt obstruction of the thoracic duct, but in a few persons, lymphatico-venous anastomoses might become naturally wide enough to compensate for the abrupt obstruction of the thoracic duct. In Case 1 and Case 2, severe shock took place immediately after en bloc resection of the esophagus when the standard rate (8ml/kg/h) of infusion of electrolyte solution or plasma was used during surgery. At that time we first suspected cardiac failure, because we did not appreciate the great decrease in circulating plasma volume after the resection of the thoracic duct. Massive infusions of plasma and electrolyte solution counteracted shock in these patients. These experiences led us to study the fluid balance of patients who underwent en bloc resection. Our analysis of the hemodynamic changes during surgery and within 48 hours after resection of the thoracic duct clearly showed that in order to maintain normal postoperative hemodynamics, significantly larger amounts of plasma or electrolyte solutionr are necessary after en bloc resection than in patients whose thoracic duct are not resected (P < 0.01). We also observed that after en bloc resection, data of blood gas analysis within 12h after surgery were better than that in patients who had the usual type of esophagectomy, probably due to intravascular hypovolemia. In spite of intravascular hypovolemia, urinary output immediatey after the surgery was increased (Case 2). Urinary osmotic pressure was within normal ranges. We cannot explain this phenomenon, which led us to a delayed diagnosis of hypovolemia. After 48h, large intravenous infusions were no longer necessary to maintain a normal hemodynamic state. This early recovery from the abrupt abnormal hemodynamic changes induced by resection of the thoracic duct may be related to the fact that these patients are not permitted any oral food intake for at least ten days. After the intravascular hypovolemia has been corrected by rapid infusion of plasma, standard amounts of infusion of plasma or electrolytes seem to be able to maintain a stable hemodynamic state. In the patient with chronic hepatitis, however, (Case 3), ascites increased progressively after the start of oral food intake. This retention of ascites could not be
M. Imamura et al.: Resection of Thoracic Duct controlled and we had to thus restrict oral intake and continue intravenous hyperalimentation. He eventually died of peritonitis and renal failure. Thus, in patients with hepatic failure, such as chronic hepatitis or cirrhosis, postoperative fluid management and nutrition are very difficult problems. We think that en bloc resection of the esophagus should not be performed in these patients. Resection of the thoracic duct does not seem to influence the hemodynamics or the lymphatic dynamics in the late postoperative days, since the nutritional state of patients who have survived more than three years postoperative does not differ from that of patients who have undergone a standard esophagectomy. Therefore, we conclude that en bloc resection for esophageal cancer should be performed in selected patients with no liver disease under the careful management for the hemodynamic changes due to resection of the thoracic duct.
References 1. Skinner DB (1983) En bloc resection for neoplasms of the esophagus and cardia. J Thorac Cardiovasc Surg 85:59-71 2. Akiyama H, Tsurumaru M, Watanabe G, et al. (1984) Development of surgery for carcinoma of the esophagus. Am J Surg 147:9-16 3. Imamura M, Ohishi K, Tobe T (1987) Retrosternal esophagogastrostomy with the EEA stapler. Surg Gynocl Obst 161:364-366 4. Imamura M, Yanagibashi K, Tobe T et al. (1988) Transthoracic resection of esophageal cancer in patients with pulmonary dysfunction - - usefulness of high frequency ventilation during thoracotomy. Ann Surg 208:601-605 5. Logan A (1963) The surgical treatment of carcinoma of the esophagus and cardia. J Thorac Cardiovasc Surg 46:150-161 6. Siewert JR, Holsher AH, Roder J, Bartels H (1988) En-bloc Resection der Speiserohre beim Oesophaguscarcinom. Langenbecks Arch Chir 373:367-376 7. Job TT (1918) Lymphatico-venous communications in the common rat and their significance. Am J Anat 24:467-485 8. Threefoot SA, Kent WT, Hatchett BF (1963) Lymphaticovenous and lymphaticolymphatic communications demonstrated by plastic corrosion models of rats and postmortem lymphangiography in man. J Lab Clin Med 61:9-22 9. Neyazaki T, Kupic EA, Marshall WH, Abrams HL (1965) Collateral lymphatico-venous communications after experimental obstruction of the thoracic duct. Radiology 85:423-432
