MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
SULIANEGARA ET ALli
EXPERIMENTAL REPLACEMENT OF THE CERVICAL OESOPHAGUS WITH A FREE TRANSVERSE COLON AUTOGRAFT USING MICROVASCULAR ANASTOMOSES R. Q L i A N E G A R A ’, PETER RYAN BERNA,RD. MC C. O’BRIEN 3, BRIANT. COLLOPY‘ AND ROBERT M. PiANTA Microsurgery Research Unit, Section of General Surgery, Department of Anatomical Pathology, St Vincent’s Hospital, Melbourne Microsurgical revascularization In the neck has Improved the survival rate of free small-bowel grafts used to replace the cervical oesophagus. Colon vessels are larger than those in the jejunum, and even with naked-eye technique good results have been reported lollowing free colon grafting of the oesophagus. The authors thought it might be possible to achieve even better results using microvascular anastomoses in the neck, and In this way replaced the cervical oesophagus in each of 15 dogs with a free transverse colon autotransplant. Four dogs died during operation and lour within three days, two of the latter with graft necrosis. The next seven dogs survived eight days or more, and all their colon grafts were viable. Where patent arteries were demonstrated (three cases) the mucosa and myenteric plexuses of the graft survived. In ail four dogs where the arteries were thrombosed or not demonstrated, the mucosa was absent, myenteric plexuses were absent in three, and one graft was grossly stenosed. Among these seven dogs there were three fistulae, one where graft arterler were patent and two where they were not; immedlate anastomotic healing Was not guaranteed by graft survlval or arterial patency.
SINCEBillroth (1872) first described resection of the cervical oesophagus, many ingenious methods have been devised to replace it. The techniques used up to the present decade have been well reviewed by Harrison and McKelvie (1966), and include staged skin grafts, local non-alimentary tubes, pedicled alimentary grafts, and finally free alimentary grafts. Non-revascularized free intestinal grafts often survive (Iskeceli. 1962; Stoner et alii, 1972 but late stricture is invariable. Ballinger et alii (1962) and Lillehei et alii (1963a) showed that free revascularized intestinal grafts functioned perfectly without nerve or lymphatic connections, but when Eastcott and Simpson in 1954 (Eastcott, 1964) first replaced human cervical oesophagus with a free sigmoid colon graft revascularized by naked-eye large vessel First St Vincent s Indonesian Fellow in General Surgery and Research Fellow in Microsurgery Surgeon to Inpatients Director Microsurgical Research Unit Surgeon lo Outpatients Pathology Registrar later Surgical Registrar Reprints Mr Peter Ryan, 20 Collins Street Melbourne. Victoria 3000. (Dr R Sulianegara wasunfortunately killed in arnotoracctdentafterhlsreturn to Indonesia)
AUST N.Z. J. SURG VOL 48 - No.4. AUGUST,1978
anastomosis in the neck (inferior mesenteric artery to external carotid artery), they were frustrated by failure of the venous anastomosis (inferior mesenteric to internal jugular) in all three cases; they later reported one success (Eastcott and Simpson, 1965). The first successful use of a free colon graft in man was reported by Nakayamaetalii (1962). Table 1 and Table 2 list reports of cervical oesophageal replacement with free intestinal grafts revascularized in the neck using naked-eye technique. Influenced by the success of Jacobson and Suarez (1960) with small-vessel anastomoses, Som and Nussbaum (1969), Peters et alii (1971), McKee and Peters (1977) (personal communication), and Macleod (1977) (personal communication) have used magnification to revascularize free jejunal autografts with good results (Table 1). Colon has some advantages over jejunum (Table 3) as a free graft in the neck, and reports of free colon transplants with naked-eye vessel anastomosis in the neck have been favourable (Table 2 ) . It seems surprising that no one has tried to improve these results with microvascular technique, yet we found no report of attempts at such transplants using the operating microscope. 463
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
TABLE^ Reports of Free Jejunal Grafts used to Replace the Human Cervical Oesophagus or Pharynx. Total
Successful
Seidenberg et alii (1959) Roberts and Douglass (1961)
1
1
2
2
Iskeceli (1962)
2
0
Harrison and McKelvie (1964) Jurkiewicz (1965)
1
1
2
2
Staley and Scanlon (1967)
1
0
Dando et all) (I 967)
2
2
Som and Nussbaum (1969)
6
3
Nakamura ef all! (1975) WcKee and Peters (1977, personal communication)
13
10
*
11
9
4
3
‘Macleod (1977. personal communication)
33(73%)
45
‘Microvascular Technique
SULIANEGARA ET ALll
place for free autograft replacement of the upper oesophagus. I n view of the demonstrated superiority of microvascular technique in srnallvessel anastomoses (Jacobson and Suarez, 1960; O’Brien et alii, 1970a, 1970b, 1971; Hayhurst and O’Brien, 1975), we decided to apply this method to experimental replacement of the canine cervical oesophagus with a free colon autograft. MATERIAL AND METHODS Fifteen mongrel dogs were used weighing between 6 and 27 kg (average weight 16.0 kg). The first five dogs had no bowel preparation, but the remaining 10 were given 2 gm of framycetin (Soframycin) orally on each of the two days before operation (Ryan et alii, 1969). Dogs were anaesthetized with 30 rng/kg of intravenous pentobarbitone sodium (Nembutal). A segment of transverse colon 9 to 16 cm long was resected via a midline coeliotomy and colon continuity reestablished with a single layer of interrupted atraumatic 4/0 silk all coats anastomotic sutures.
Early carcinomas of the hypopharynx and cervical oesophagus may not require total oesophagectomy; local defects due to trauma and posttraumatic strictures may require no more than local resection; and not all patients with cervical or pharyngeal cancer have a colon orastomach which can be safely mobilized on its own vessels for anastomosis to a pharyngostome. We decided that in such clinical circumstances there would be a TABLE^ Reports 01 Free Colon Grafts used to Replace the Human Cervical Oesophagus or Pharynx. Total McKelvie (1964) Nakayama ef alii (1964)
Successful
1
1
21
16
Eastcott and Simpson (1965)
4
1
Chrysopathis (1966)
6
4
Maillet at alii (1966)
1
1
Noirclerc el elii (1976)
2
1 24(69%)
35
TABLE3
Comparison of Human Jejunum and Colon as Free Transplants in the Neck ~
Jejunum ~
Calibre Arteries Resistance to gastric juice
Colon ~
~
= oesophagus
= pharynx
1-2 mm
3.0 mm
++
+
+
++
Oxygen requirements
++
+
Sterility Ease in obtaining graft
++
+
Propulsive power
464
FIGURE1: Anastomosed colon graft in situ at conclusion of operation. Artery and vein anastomoses indicated by arrows. Dog a.
AUSTN.Z. J. SURG.,Voc 48 - No. 4. AUGUST. 1978
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
SULIANEGARA E T ALll
The radial mesocolic artery of the graft was 0.8 to 1.2mm in diameter and the vein 1.0 to 1.7mm in diameter. The graft lumen was washed out with normal saline and the radial mesocolic artery with lOOml of saline containing 1,000 units of heparin. The graft was placed in a solution of 1.2 megaunits of penicillin benzathine (Bicillin) in 250 ml of normal saline at room temperature. Via a midline incision in the neck a segment of cervical oesophagus 7 to 10 cm long was resected from just below the cricoid cartilage. The colon graft was joined to pharyngo-oesophagus above and to oesophagus below in isoperistaltic fashion with a single layer of interrupted atraumatic 4/0 silk all coats sutures. The radial mesocolic vein was then anastomosed end-to-end to the transected internal jugular vein or anterior facial vein tributary depending upon which had a corresponding diameter. The radial mesocolic artery was anastornosed in most cases end-to-end to the muscular branch of the cranial thyroid artery which has a corresponding calibre. When the microarterial clamp was released, the bloodless contracted shortened spastic graft became pink, soft and relaxed, the mesocolic arcade was filled, and the vein almost instantly filled with blood. The graft then regained its original size (Figure 1). The total ischaemic time ranged from three to four hours. After observing the patent vessels for 10 to 15 minutes, the neck wound was closed. After operation each dog was given 1 megaunit of penicillin intramuscularly daily for five days. The
first five dogs were given fluids on the first and solid food on the third postoperative day. The last ten dogs were given intravenous fluids for three days, oral fluids on the third and fourth, and solid food on the fifth postoperative day. All surviving dogs were subjected to barium swallow X-ray examination and arteriography by left common carotid cannulation at various times after operation and then sacrificed at once. The common carotid artey was then perfused with fine barium suspension (Micropaque) and the colon graft and its vessels examined histologically. RESULTS These are shown in Table 4. Deaths-Four dogs (1,2,5.6) died of anaesthetic overdose associated with prolonged operation and four because of defects in operative technique ($4, 7) or postoperative care (Dog 8 was fed solid food on the first postoperative day). Graft fistulae.-Colon graft anastomotic leakage occurred in three of the remaining seven dogs. These dogs were eating satisfactorily and did not die, but were sacrificed. It seems likely that their fistulae would otherwise have healed. In none of the dogs was any abdominal colon anastomotic leak found at autopsy. Colon graft viability and function.-In all of those seven dogs surviving more than a week (Nos. 9 to 15) the colon graft was viable. Good function was confirmed by barium swallow X-ray examination (Figure 2) in four of these dogs (Nos. 11, 12,13,15).
TABLE4 Analysis of Results No.
Outcome
Arterial Anastomosis
Colon Graft
Colon Graft
Colon Graft Myenteric Plexuses Mucosa
Colon Graft Anastomoses
Operation death
Patent
Not examined
Not examined
Operation death
Patent
Not examined
Not examined
Operation death
Disrupted
Not examined
Not examined
Operation death
Disrupted
-
Not examined
Not examined
Died 1 day
Patent
Viable
Not examined
Not examined
Satisfactory
Died 1 day
Patent
Viable
Not examined
Not examined
Satisfactory
Died 40 hrs
Vein thrombosed
Necrotic
Not examined
Not examined
Necrotic
Sacrificed 3 days
Unknown
Necrotic
Not examined
Not examined
Total breakdown
Sacrificed 0 days
Patent
Viable
Present
Present
Proximal and distal leak
10
Sacrificed 11 days
Unknown
Viable, chronic ischaemia
Absent
Absent
Distal leak
11
Sacrificed 2 weeks Patent on arteriogram
Present
Absent
Healed
12
Sacrificed 4 weeks ,Artery thrombosed Viable. chronic ischaemia
Absent
Absent
Distal leak
13
Sacrificed 5 weeks Patent
Present
Present
Healed
14
Sacrificed 7 weeks Artery thrombosed Viable. chronic ischaemic and stenosis
Absent
Absent
Healed
15
Sacrificed 10 weeks Patent
Present
Present
Healed
Viable, acute patchy ischaemia
Viable
Viable
AUST N.Z. J. SUAG..VOL. 48 - No 4. AUGUST,1978
465
SULIANEGARA ET ALll
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
Cineradiography was not employed, so no observation was made of graft peristalsis. The colon graft in the seven-week dog (No. 14) was stenosed (diameter 7mm). This dog alone was emaciated. Vessel anastomoses.-Carotid arteriograms carried out in three dogs (11, 13 and 15) demonstrated patent arterial anastomoses with a fine mesocolic network of vessels (Figure 3). Histological examination after Micropaque perfusion (Figure 5) showed patent arteries in two (13 and 15), thrombosis of the artery with partial recanalization in one (14), and arterial thrombosis in one case (1 1). In the last case thrombosis must have occurred after arteriography and explains the histological “acute ischaemia” (vide infra). Crossing-over of vessels from graft to distal oesophagus was demonstrated in Dog 15 (Figure4). The arterial anastomosis was not found in the specimen of Dog 12, so there was no histological confirmation of its patency. Neither arteriography
FIGURE2 Barium swallow - barium traversing colon graft Arrows indcate upper and lower anastomoses Dog 13.
FIGURE 3 Carotid arteriogram. showing main colon graft artery (see arrow) and fine mesocolic vessels Dog 11
466
FIGURE 4 Barium (Micropaque) perfusion via common carotid artery, showing ”crossing-over” of vessels into distal oesophagus D o g 15
AUST
Nz J
SURG VOL
48 - NO 4, AUGUST 1978
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
SULIANEGARA ET ALll
ischaemia with loss of colon mucosa and myenteric plexuses in all three (Figure 7) and macroscopic vide supra). stenosis in one (No. 14 Relationship of colon graft anastomotic healing to its blood supply.-Leakage was found at the distal graft-oesophagus junction in two of the three cases of chronic ischaemia (Nos 10 and 12), but when a large fistula due to both proximal and distal g r a f t - o e s o p h a g u s a n a s t o m o t i c breakdown occurred in Dog 9, both artery and vein anastomoses were found patent at reexploration. Obviously initial anastomotic healing is not merelya function of overall blood supply.
-
DISCUSSION Early failures were clearly due to lack of experience leading to prolonged operation with anaesthetic overdose and long ischaemia time. On the other hand, the colon grafts tolerated ischaemia for three to four hours at room temperature, unlike the small-bowel and stomach autotransplants of Lillehei et alii, (1963a, 1963b), and without the continuous perfusion used by Green and Som (1966a, 1966b) for small-bowel autografts. Nakayama et alii (vide supra) had made the point that colon requires less oxygen than stomach or small bowel; the extra ischaemia time thus afforded
FIGURE 5: Photomicrogram of’ artery anastomosis showing patent lumen and suture material in wall. Dog 13.
nor Micropaque perfusion was carried out in Dogs 9 and 10. At reexploration the vessels were found patent in Dog 9, but no vessels were found in Dog 10. We conclude that arteries were definitely patent in three (9, 13 and 15), partly patent in one (14)’ and thrombosed in three (10, 11 and 12). The vein anastomosis was found patent at reexploration in Dog 9, but venous anastomoses were not found in Dogs 10 to 15. Colon graft rnucosa and rnyenteric plexuses.-In those three dogs where patent arterial anastomoses had been demonstrated (9, 13 and 15) histological examination showed complete preservation of normal structures, including colon rnucosa and myenteric nerve plexuses (Figure 6). Where the arterial anastomosis apparently thrombosed just before sacrifice of the animal (No. l l ) , there was patchy acute necrosis of the colon wall and the colon mucosa was absent, but myenteric plexuses were still visible. In those three dogs in which graft arteries were thrombosed Or absent (10,12 and 14), there was histological evidence of chronic graft AUST.N.Z. J. SURG., VOL. 48 - No. 4, AUGUST, 1978
F,GURE 6: Photomicrogram of oesophagus.colon graft anastomosis, showing perfect survival of colon mucosa and myenteric nerve plexus ganglion cells. Dog 15.
467
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
FIGURE7: Photomicrogram of oesophagus-colon graft anastomosis, showing loss of colon mucosa and nerve plexus ganglion cells. Dog 14.
makes its use a more practical proposition. In our opinion ischaemia time might have been further reduced by carrying out neck dissection before the abdominal operation or by simultaneous neck and abdominal dissection using two surgical teams. Peters et alii (1971) achieved an ischaemia time of 30 to 35 minutes by anastomosing the vessels first, but we found there was too much danger of disrupting the vessels during subsequent graft anastomosis (as happened in the case of Dogs 3 and 4). Nine of 11 dogs surviving operation had viable functioning colon grafts (Nos 5-15. Table 4). yet only in five were patent arterial anastomoses found at autopsy or reexploration. Arterial anastomoses must have remained patent in the early postoperative period in the other four dogs (Nos, 10, 11, 12 and 14), or the grafts would not have survived at all. Presumably thrombosis occurred after c o n n e c t i o n s h a d been e s t a b l i s h e d w i t h neighbouring blood vessels sufficient to keep the graft alive. Histological quality of graft survival, however, correlated closely with arterial supply. Where arteries remained patent, colon mucosa and plexuses survived. It is to be expected that graft muscle contraction will be more satisfactory if 468
SULIANEGARA ET ALll
intrinsic myenteric plexuses are viable. Eastcott (1964). Nicks (1967), Som and Nussbaum (1969) and Stoner et alii (1972) have emphasized the importance of avoiding venous thrombosis in order to maintain graft viability. We did not identify vein anastomoses at reexploration in most dogs, but assume that they were satisfactory where the graft was viable. In Dog 7 venous thrombosis was recognized and that graft was necrotic. It may be that veins thrombose when arterial inflow is poor. We believe that the use of microvascular technique is the most important way of avoiding venous thrombosis. Anastomotic leakage occurred in two ischaemic grafts, but also in one where the overall blood supply was good and the vessels patent. Here leakage may have been due to a local cause such as a single tight suture. It has been shown that experimental bowel anastomoses which fail at one spot may elsewhere be of adequate strength (Ryan and O’Brien, 1971), suggesting that anastomoses often fail because of local, not general, ischaemia. Good overall blood supply of a graft is no guarantee against leakage at one spot. Furthermore, leakage at one spot should not be a disaster provided that there is adequate external drainage; the resulting fistula should heal if there is no general ischaemia of the graft with consequent stenosis.
CONCLUSIONS We conclude that careful microvascular technique achieves adequatevascularization of free colon transplants used to replace cervical oesophagus, with satisfactory function. Though we have found that grafts survived even where vessels did not remain patent, we believe our most important conclusion is that survival of specialized structures in the graft - colon mucosa and myenteric plexuses - absolutely depends upon persistent patency of the arterial anastomoses. Finally, we suggest that good overall graft blood supply, while not always preventing an immediate anastomotic leak, gives the best guarantee that it will ultimately heal. REFERENCES BALLINGER. W.F.. II, CHRISTY, M.G. and ASHBY. W.B. (1962), Surgery, 52: 151. BILLROTH.TH (1872). Langenbecks Arch. klin. Chir., 13: 65. CHRYSOPATHIS,P. (1966). Brit. J. Surg., 53: 122. DAIDO.R.et alii (1967). (lshoku Transpl. J.) quoted by Mullens and Pezacki. EASTCOTT.H.H.G., (1964), Lancet, 2: 1182. Lancet, 1: 1068. EASTC0TT.H.H.G.and SIMPSON.J.F., ( I = ) , GREEN,G.E.and S0M.M.L. (1966a). Surgery, 60: 1017. GREEN.G.E.and S0tdM.L. (1966b), Surgery, 60: 1017. HARRISON.D.F.N. and MCKELVIE,P. (1966) in Modern Trends in Surgery Butterworth. London, edited by Irvine. W.. vol. II; 276.
AUST. N.Z. J. SURG..VOL. 48 - No. 4. AUGUST,1978
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
HAYHURST.J.W.~~~O'BRIEN,B.M Brit. CC J-plast. . ( ~ ~ ~Surg., ~),
28: 128. (1962).Surgery, 51:496. ISKECELI.O.K. JACOBSON.J.H. I I and SUAREZ,E.L. (1960),Surg. Forum: 243. JURKIEWICZ. M.J. (1965). Plast. reconstr. Surg., 36: 509. LILLEHEI.R.C.,GOLDBERG.S., GOOTT.~ and . LONGERBEAM.J.K. (1963a).Arner. J. Surg., 105: 58. LILLEHEILRC, LONGEREEAM. J.K. and SCOTT.W.R.(1963b). J. Arner. med. Ass., 183: 861. MC KELVIE.~. (1964).Ch. M. Thesis, Manchester. quoted by Harrison, D.F.N. and McKelvie. P. 1966. MAILLET. P.. GALLARD.J.and SISTERON.A. (1966).Mem. Acad. Chrr., 92: 257. (1971).Int. J. Surg.. 55: 157. MULLENS.J.E. and PEZACKI.Z.J. NAKAMURA. T.. INOKUCHI. K. and SUGIMACHI, K. (1975).Jap. J. Surg., 5: 92. NAKAYAMA. K.. TAMIYA,T., YAMAMOTO. S. (1962),Surgery,
52: 918. N A K A Y A M K. A , YAMAMOTO. K. TAMIYA. T., MAKINO. H.. ODAKA. M.. OHWADA, M. and TAKAHASHI. H. (1964),Surgery.
55: 796. NICKS.R.(1967).Brit. J. Surg., 54: 124. NOIRCLERC, M.. MACMEJAC. C., PAOLI,J.M. and D0R.J. (19671, Ann. Chir. Thor. Car,, 6: 154.
AUST N.Z. J. SURG, VOL 48 - No 4, AUWST, 1978
SULIANEGARA ET ALll
0'BRIEN.B. McC., HENDERS0N.P.N. and CROCK.G.W.(1970a), Med. J. Aust.. 1: 712. O'BRIEN. B.. Mc C.. HENDERSON, P.N., BENNETT.R.C. and CROCK.G.W..(1970b).Med. J. Aust.. 1: 722. OBRIEN. 8. M C C., HENDERSON. P.N.. BAXTER. T.J. and MILLER.G.D.H. (1977)in "Transactions of the Fifth International Congress of Plastic and Heconstructive Surgery", Butterworth. Australia, edited by Hueston. J.T.: 578. PETERS, C.R.. Mc KEE.D.M. and BERRY,B.T. (1971).Amer. J. Surg.. 121: 675. ROBERTS.R.E. and DOUGLASS.F.M. (1961).New. Engl. J. Med.,
264: 342. RYAN.~.. STRATFORD.B.C. and DIXSON.S.(1969). AusT.N.Z. J. SURG.. 39: 105
i4YAN.P. and O'BRIEN.J.(1971).Abstract of 44th Annual General Scientific Meeting of Royal Australasian College of Surgeons, Sydney: 24. SEIDENBERG. B., ROSENAK, S.S.,HURWITT.E.S. and SOM.M.L. (1959).Ann. Surg., 149: 162. SOM.M.L. and NUSSBAUM, M . (1969).Otolaryngol. Clin. N. Amer., 2: 631. STALEY, C.J. and SCANLON.E.F.(1967).Surg. Clin. N. Amer..
47: 215. STONER, J.C.. THOMAS, G.K. and ALEO,D.C. (1972). Arch. Otolaryngol., 95: 141.
469
SULIANEGARA ET ALli
EXPERIMENTAL REPLACEMENT OF THE CERVICAL OESOPHAGUS WITH A FREE TRANSVERSE COLON AUTOGRAFT USING MICROVASCULAR ANASTOMOSES R. Q L i A N E G A R A ’, PETER RYAN BERNA,RD. MC C. O’BRIEN 3, BRIANT. COLLOPY‘ AND ROBERT M. PiANTA Microsurgery Research Unit, Section of General Surgery, Department of Anatomical Pathology, St Vincent’s Hospital, Melbourne Microsurgical revascularization In the neck has Improved the survival rate of free small-bowel grafts used to replace the cervical oesophagus. Colon vessels are larger than those in the jejunum, and even with naked-eye technique good results have been reported lollowing free colon grafting of the oesophagus. The authors thought it might be possible to achieve even better results using microvascular anastomoses in the neck, and In this way replaced the cervical oesophagus in each of 15 dogs with a free transverse colon autotransplant. Four dogs died during operation and lour within three days, two of the latter with graft necrosis. The next seven dogs survived eight days or more, and all their colon grafts were viable. Where patent arteries were demonstrated (three cases) the mucosa and myenteric plexuses of the graft survived. In ail four dogs where the arteries were thrombosed or not demonstrated, the mucosa was absent, myenteric plexuses were absent in three, and one graft was grossly stenosed. Among these seven dogs there were three fistulae, one where graft arterler were patent and two where they were not; immedlate anastomotic healing Was not guaranteed by graft survlval or arterial patency.
SINCEBillroth (1872) first described resection of the cervical oesophagus, many ingenious methods have been devised to replace it. The techniques used up to the present decade have been well reviewed by Harrison and McKelvie (1966), and include staged skin grafts, local non-alimentary tubes, pedicled alimentary grafts, and finally free alimentary grafts. Non-revascularized free intestinal grafts often survive (Iskeceli. 1962; Stoner et alii, 1972 but late stricture is invariable. Ballinger et alii (1962) and Lillehei et alii (1963a) showed that free revascularized intestinal grafts functioned perfectly without nerve or lymphatic connections, but when Eastcott and Simpson in 1954 (Eastcott, 1964) first replaced human cervical oesophagus with a free sigmoid colon graft revascularized by naked-eye large vessel First St Vincent s Indonesian Fellow in General Surgery and Research Fellow in Microsurgery Surgeon to Inpatients Director Microsurgical Research Unit Surgeon lo Outpatients Pathology Registrar later Surgical Registrar Reprints Mr Peter Ryan, 20 Collins Street Melbourne. Victoria 3000. (Dr R Sulianegara wasunfortunately killed in arnotoracctdentafterhlsreturn to Indonesia)
AUST N.Z. J. SURG VOL 48 - No.4. AUGUST,1978
anastomosis in the neck (inferior mesenteric artery to external carotid artery), they were frustrated by failure of the venous anastomosis (inferior mesenteric to internal jugular) in all three cases; they later reported one success (Eastcott and Simpson, 1965). The first successful use of a free colon graft in man was reported by Nakayamaetalii (1962). Table 1 and Table 2 list reports of cervical oesophageal replacement with free intestinal grafts revascularized in the neck using naked-eye technique. Influenced by the success of Jacobson and Suarez (1960) with small-vessel anastomoses, Som and Nussbaum (1969), Peters et alii (1971), McKee and Peters (1977) (personal communication), and Macleod (1977) (personal communication) have used magnification to revascularize free jejunal autografts with good results (Table 1). Colon has some advantages over jejunum (Table 3) as a free graft in the neck, and reports of free colon transplants with naked-eye vessel anastomosis in the neck have been favourable (Table 2 ) . It seems surprising that no one has tried to improve these results with microvascular technique, yet we found no report of attempts at such transplants using the operating microscope. 463
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
TABLE^ Reports of Free Jejunal Grafts used to Replace the Human Cervical Oesophagus or Pharynx. Total
Successful
Seidenberg et alii (1959) Roberts and Douglass (1961)
1
1
2
2
Iskeceli (1962)
2
0
Harrison and McKelvie (1964) Jurkiewicz (1965)
1
1
2
2
Staley and Scanlon (1967)
1
0
Dando et all) (I 967)
2
2
Som and Nussbaum (1969)
6
3
Nakamura ef all! (1975) WcKee and Peters (1977, personal communication)
13
10
*
11
9
4
3
‘Macleod (1977. personal communication)
33(73%)
45
‘Microvascular Technique
SULIANEGARA ET ALll
place for free autograft replacement of the upper oesophagus. I n view of the demonstrated superiority of microvascular technique in srnallvessel anastomoses (Jacobson and Suarez, 1960; O’Brien et alii, 1970a, 1970b, 1971; Hayhurst and O’Brien, 1975), we decided to apply this method to experimental replacement of the canine cervical oesophagus with a free colon autograft. MATERIAL AND METHODS Fifteen mongrel dogs were used weighing between 6 and 27 kg (average weight 16.0 kg). The first five dogs had no bowel preparation, but the remaining 10 were given 2 gm of framycetin (Soframycin) orally on each of the two days before operation (Ryan et alii, 1969). Dogs were anaesthetized with 30 rng/kg of intravenous pentobarbitone sodium (Nembutal). A segment of transverse colon 9 to 16 cm long was resected via a midline coeliotomy and colon continuity reestablished with a single layer of interrupted atraumatic 4/0 silk all coats anastomotic sutures.
Early carcinomas of the hypopharynx and cervical oesophagus may not require total oesophagectomy; local defects due to trauma and posttraumatic strictures may require no more than local resection; and not all patients with cervical or pharyngeal cancer have a colon orastomach which can be safely mobilized on its own vessels for anastomosis to a pharyngostome. We decided that in such clinical circumstances there would be a TABLE^ Reports 01 Free Colon Grafts used to Replace the Human Cervical Oesophagus or Pharynx. Total McKelvie (1964) Nakayama ef alii (1964)
Successful
1
1
21
16
Eastcott and Simpson (1965)
4
1
Chrysopathis (1966)
6
4
Maillet at alii (1966)
1
1
Noirclerc el elii (1976)
2
1 24(69%)
35
TABLE3
Comparison of Human Jejunum and Colon as Free Transplants in the Neck ~
Jejunum ~
Calibre Arteries Resistance to gastric juice
Colon ~
~
= oesophagus
= pharynx
1-2 mm
3.0 mm
++
+
+
++
Oxygen requirements
++
+
Sterility Ease in obtaining graft
++
+
Propulsive power
464
FIGURE1: Anastomosed colon graft in situ at conclusion of operation. Artery and vein anastomoses indicated by arrows. Dog a.
AUSTN.Z. J. SURG.,Voc 48 - No. 4. AUGUST. 1978
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
SULIANEGARA E T ALll
The radial mesocolic artery of the graft was 0.8 to 1.2mm in diameter and the vein 1.0 to 1.7mm in diameter. The graft lumen was washed out with normal saline and the radial mesocolic artery with lOOml of saline containing 1,000 units of heparin. The graft was placed in a solution of 1.2 megaunits of penicillin benzathine (Bicillin) in 250 ml of normal saline at room temperature. Via a midline incision in the neck a segment of cervical oesophagus 7 to 10 cm long was resected from just below the cricoid cartilage. The colon graft was joined to pharyngo-oesophagus above and to oesophagus below in isoperistaltic fashion with a single layer of interrupted atraumatic 4/0 silk all coats sutures. The radial mesocolic vein was then anastomosed end-to-end to the transected internal jugular vein or anterior facial vein tributary depending upon which had a corresponding diameter. The radial mesocolic artery was anastornosed in most cases end-to-end to the muscular branch of the cranial thyroid artery which has a corresponding calibre. When the microarterial clamp was released, the bloodless contracted shortened spastic graft became pink, soft and relaxed, the mesocolic arcade was filled, and the vein almost instantly filled with blood. The graft then regained its original size (Figure 1). The total ischaemic time ranged from three to four hours. After observing the patent vessels for 10 to 15 minutes, the neck wound was closed. After operation each dog was given 1 megaunit of penicillin intramuscularly daily for five days. The
first five dogs were given fluids on the first and solid food on the third postoperative day. The last ten dogs were given intravenous fluids for three days, oral fluids on the third and fourth, and solid food on the fifth postoperative day. All surviving dogs were subjected to barium swallow X-ray examination and arteriography by left common carotid cannulation at various times after operation and then sacrificed at once. The common carotid artey was then perfused with fine barium suspension (Micropaque) and the colon graft and its vessels examined histologically. RESULTS These are shown in Table 4. Deaths-Four dogs (1,2,5.6) died of anaesthetic overdose associated with prolonged operation and four because of defects in operative technique ($4, 7) or postoperative care (Dog 8 was fed solid food on the first postoperative day). Graft fistulae.-Colon graft anastomotic leakage occurred in three of the remaining seven dogs. These dogs were eating satisfactorily and did not die, but were sacrificed. It seems likely that their fistulae would otherwise have healed. In none of the dogs was any abdominal colon anastomotic leak found at autopsy. Colon graft viability and function.-In all of those seven dogs surviving more than a week (Nos. 9 to 15) the colon graft was viable. Good function was confirmed by barium swallow X-ray examination (Figure 2) in four of these dogs (Nos. 11, 12,13,15).
TABLE4 Analysis of Results No.
Outcome
Arterial Anastomosis
Colon Graft
Colon Graft
Colon Graft Myenteric Plexuses Mucosa
Colon Graft Anastomoses
Operation death
Patent
Not examined
Not examined
Operation death
Patent
Not examined
Not examined
Operation death
Disrupted
Not examined
Not examined
Operation death
Disrupted
-
Not examined
Not examined
Died 1 day
Patent
Viable
Not examined
Not examined
Satisfactory
Died 1 day
Patent
Viable
Not examined
Not examined
Satisfactory
Died 40 hrs
Vein thrombosed
Necrotic
Not examined
Not examined
Necrotic
Sacrificed 3 days
Unknown
Necrotic
Not examined
Not examined
Total breakdown
Sacrificed 0 days
Patent
Viable
Present
Present
Proximal and distal leak
10
Sacrificed 11 days
Unknown
Viable, chronic ischaemia
Absent
Absent
Distal leak
11
Sacrificed 2 weeks Patent on arteriogram
Present
Absent
Healed
12
Sacrificed 4 weeks ,Artery thrombosed Viable. chronic ischaemia
Absent
Absent
Distal leak
13
Sacrificed 5 weeks Patent
Present
Present
Healed
14
Sacrificed 7 weeks Artery thrombosed Viable. chronic ischaemic and stenosis
Absent
Absent
Healed
15
Sacrificed 10 weeks Patent
Present
Present
Healed
Viable, acute patchy ischaemia
Viable
Viable
AUST N.Z. J. SUAG..VOL. 48 - No 4. AUGUST,1978
465
SULIANEGARA ET ALll
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
Cineradiography was not employed, so no observation was made of graft peristalsis. The colon graft in the seven-week dog (No. 14) was stenosed (diameter 7mm). This dog alone was emaciated. Vessel anastomoses.-Carotid arteriograms carried out in three dogs (11, 13 and 15) demonstrated patent arterial anastomoses with a fine mesocolic network of vessels (Figure 3). Histological examination after Micropaque perfusion (Figure 5) showed patent arteries in two (13 and 15), thrombosis of the artery with partial recanalization in one (14), and arterial thrombosis in one case (1 1). In the last case thrombosis must have occurred after arteriography and explains the histological “acute ischaemia” (vide infra). Crossing-over of vessels from graft to distal oesophagus was demonstrated in Dog 15 (Figure4). The arterial anastomosis was not found in the specimen of Dog 12, so there was no histological confirmation of its patency. Neither arteriography
FIGURE2 Barium swallow - barium traversing colon graft Arrows indcate upper and lower anastomoses Dog 13.
FIGURE 3 Carotid arteriogram. showing main colon graft artery (see arrow) and fine mesocolic vessels Dog 11
466
FIGURE 4 Barium (Micropaque) perfusion via common carotid artery, showing ”crossing-over” of vessels into distal oesophagus D o g 15
AUST
Nz J
SURG VOL
48 - NO 4, AUGUST 1978
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
SULIANEGARA ET ALll
ischaemia with loss of colon mucosa and myenteric plexuses in all three (Figure 7) and macroscopic vide supra). stenosis in one (No. 14 Relationship of colon graft anastomotic healing to its blood supply.-Leakage was found at the distal graft-oesophagus junction in two of the three cases of chronic ischaemia (Nos 10 and 12), but when a large fistula due to both proximal and distal g r a f t - o e s o p h a g u s a n a s t o m o t i c breakdown occurred in Dog 9, both artery and vein anastomoses were found patent at reexploration. Obviously initial anastomotic healing is not merelya function of overall blood supply.
-
DISCUSSION Early failures were clearly due to lack of experience leading to prolonged operation with anaesthetic overdose and long ischaemia time. On the other hand, the colon grafts tolerated ischaemia for three to four hours at room temperature, unlike the small-bowel and stomach autotransplants of Lillehei et alii, (1963a, 1963b), and without the continuous perfusion used by Green and Som (1966a, 1966b) for small-bowel autografts. Nakayama et alii (vide supra) had made the point that colon requires less oxygen than stomach or small bowel; the extra ischaemia time thus afforded
FIGURE 5: Photomicrogram of’ artery anastomosis showing patent lumen and suture material in wall. Dog 13.
nor Micropaque perfusion was carried out in Dogs 9 and 10. At reexploration the vessels were found patent in Dog 9, but no vessels were found in Dog 10. We conclude that arteries were definitely patent in three (9, 13 and 15), partly patent in one (14)’ and thrombosed in three (10, 11 and 12). The vein anastomosis was found patent at reexploration in Dog 9, but venous anastomoses were not found in Dogs 10 to 15. Colon graft rnucosa and rnyenteric plexuses.-In those three dogs where patent arterial anastomoses had been demonstrated (9, 13 and 15) histological examination showed complete preservation of normal structures, including colon rnucosa and myenteric nerve plexuses (Figure 6). Where the arterial anastomosis apparently thrombosed just before sacrifice of the animal (No. l l ) , there was patchy acute necrosis of the colon wall and the colon mucosa was absent, but myenteric plexuses were still visible. In those three dogs in which graft arteries were thrombosed Or absent (10,12 and 14), there was histological evidence of chronic graft AUST.N.Z. J. SURG., VOL. 48 - No. 4, AUGUST, 1978
F,GURE 6: Photomicrogram of oesophagus.colon graft anastomosis, showing perfect survival of colon mucosa and myenteric nerve plexus ganglion cells. Dog 15.
467
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
FIGURE7: Photomicrogram of oesophagus-colon graft anastomosis, showing loss of colon mucosa and nerve plexus ganglion cells. Dog 14.
makes its use a more practical proposition. In our opinion ischaemia time might have been further reduced by carrying out neck dissection before the abdominal operation or by simultaneous neck and abdominal dissection using two surgical teams. Peters et alii (1971) achieved an ischaemia time of 30 to 35 minutes by anastomosing the vessels first, but we found there was too much danger of disrupting the vessels during subsequent graft anastomosis (as happened in the case of Dogs 3 and 4). Nine of 11 dogs surviving operation had viable functioning colon grafts (Nos 5-15. Table 4). yet only in five were patent arterial anastomoses found at autopsy or reexploration. Arterial anastomoses must have remained patent in the early postoperative period in the other four dogs (Nos, 10, 11, 12 and 14), or the grafts would not have survived at all. Presumably thrombosis occurred after c o n n e c t i o n s h a d been e s t a b l i s h e d w i t h neighbouring blood vessels sufficient to keep the graft alive. Histological quality of graft survival, however, correlated closely with arterial supply. Where arteries remained patent, colon mucosa and plexuses survived. It is to be expected that graft muscle contraction will be more satisfactory if 468
SULIANEGARA ET ALll
intrinsic myenteric plexuses are viable. Eastcott (1964). Nicks (1967), Som and Nussbaum (1969) and Stoner et alii (1972) have emphasized the importance of avoiding venous thrombosis in order to maintain graft viability. We did not identify vein anastomoses at reexploration in most dogs, but assume that they were satisfactory where the graft was viable. In Dog 7 venous thrombosis was recognized and that graft was necrotic. It may be that veins thrombose when arterial inflow is poor. We believe that the use of microvascular technique is the most important way of avoiding venous thrombosis. Anastomotic leakage occurred in two ischaemic grafts, but also in one where the overall blood supply was good and the vessels patent. Here leakage may have been due to a local cause such as a single tight suture. It has been shown that experimental bowel anastomoses which fail at one spot may elsewhere be of adequate strength (Ryan and O’Brien, 1971), suggesting that anastomoses often fail because of local, not general, ischaemia. Good overall blood supply of a graft is no guarantee against leakage at one spot. Furthermore, leakage at one spot should not be a disaster provided that there is adequate external drainage; the resulting fistula should heal if there is no general ischaemia of the graft with consequent stenosis.
CONCLUSIONS We conclude that careful microvascular technique achieves adequatevascularization of free colon transplants used to replace cervical oesophagus, with satisfactory function. Though we have found that grafts survived even where vessels did not remain patent, we believe our most important conclusion is that survival of specialized structures in the graft - colon mucosa and myenteric plexuses - absolutely depends upon persistent patency of the arterial anastomoses. Finally, we suggest that good overall graft blood supply, while not always preventing an immediate anastomotic leak, gives the best guarantee that it will ultimately heal. REFERENCES BALLINGER. W.F.. II, CHRISTY, M.G. and ASHBY. W.B. (1962), Surgery, 52: 151. BILLROTH.TH (1872). Langenbecks Arch. klin. Chir., 13: 65. CHRYSOPATHIS,P. (1966). Brit. J. Surg., 53: 122. DAIDO.R.et alii (1967). (lshoku Transpl. J.) quoted by Mullens and Pezacki. EASTCOTT.H.H.G., (1964), Lancet, 2: 1182. Lancet, 1: 1068. EASTC0TT.H.H.G.and SIMPSON.J.F., ( I = ) , GREEN,G.E.and S0M.M.L. (1966a). Surgery, 60: 1017. GREEN.G.E.and S0tdM.L. (1966b), Surgery, 60: 1017. HARRISON.D.F.N. and MCKELVIE,P. (1966) in Modern Trends in Surgery Butterworth. London, edited by Irvine. W.. vol. II; 276.
AUST. N.Z. J. SURG..VOL. 48 - No. 4. AUGUST,1978
MICROVASCULAR OESOPHAGUS-COLON REPLACEMENT
HAYHURST.J.W.~~~O'BRIEN,B.M Brit. CC J-plast. . ( ~ ~ ~Surg., ~),
28: 128. (1962).Surgery, 51:496. ISKECELI.O.K. JACOBSON.J.H. I I and SUAREZ,E.L. (1960),Surg. Forum: 243. JURKIEWICZ. M.J. (1965). Plast. reconstr. Surg., 36: 509. LILLEHEI.R.C.,GOLDBERG.S., GOOTT.~ and . LONGERBEAM.J.K. (1963a).Arner. J. Surg., 105: 58. LILLEHEILRC, LONGEREEAM. J.K. and SCOTT.W.R.(1963b). J. Arner. med. Ass., 183: 861. MC KELVIE.~. (1964).Ch. M. Thesis, Manchester. quoted by Harrison, D.F.N. and McKelvie. P. 1966. MAILLET. P.. GALLARD.J.and SISTERON.A. (1966).Mem. Acad. Chrr., 92: 257. (1971).Int. J. Surg.. 55: 157. MULLENS.J.E. and PEZACKI.Z.J. NAKAMURA. T.. INOKUCHI. K. and SUGIMACHI, K. (1975).Jap. J. Surg., 5: 92. NAKAYAMA. K.. TAMIYA,T., YAMAMOTO. S. (1962),Surgery,
52: 918. N A K A Y A M K. A , YAMAMOTO. K. TAMIYA. T., MAKINO. H.. ODAKA. M.. OHWADA, M. and TAKAHASHI. H. (1964),Surgery.
55: 796. NICKS.R.(1967).Brit. J. Surg., 54: 124. NOIRCLERC, M.. MACMEJAC. C., PAOLI,J.M. and D0R.J. (19671, Ann. Chir. Thor. Car,, 6: 154.
AUST N.Z. J. SURG, VOL 48 - No 4, AUWST, 1978
SULIANEGARA ET ALll
0'BRIEN.B. McC., HENDERS0N.P.N. and CROCK.G.W.(1970a), Med. J. Aust.. 1: 712. O'BRIEN. B.. Mc C.. HENDERSON, P.N., BENNETT.R.C. and CROCK.G.W..(1970b).Med. J. Aust.. 1: 722. OBRIEN. 8. M C C., HENDERSON. P.N.. BAXTER. T.J. and MILLER.G.D.H. (1977)in "Transactions of the Fifth International Congress of Plastic and Heconstructive Surgery", Butterworth. Australia, edited by Hueston. J.T.: 578. PETERS, C.R.. Mc KEE.D.M. and BERRY,B.T. (1971).Amer. J. Surg.. 121: 675. ROBERTS.R.E. and DOUGLASS.F.M. (1961).New. Engl. J. Med.,
264: 342. RYAN.~.. STRATFORD.B.C. and DIXSON.S.(1969). AusT.N.Z. J. SURG.. 39: 105
i4YAN.P. and O'BRIEN.J.(1971).Abstract of 44th Annual General Scientific Meeting of Royal Australasian College of Surgeons, Sydney: 24. SEIDENBERG. B., ROSENAK, S.S.,HURWITT.E.S. and SOM.M.L. (1959).Ann. Surg., 149: 162. SOM.M.L. and NUSSBAUM, M . (1969).Otolaryngol. Clin. N. Amer., 2: 631. STALEY, C.J. and SCANLON.E.F.(1967).Surg. Clin. N. Amer..
47: 215. STONER, J.C.. THOMAS, G.K. and ALEO,D.C. (1972). Arch. Otolaryngol., 95: 141.
469
