ADONIS 030654569100044C
British Journal of Obstetrics and Gynaecology February 1991, Vol. 98, pp. 184-188
The free-flap vaginoplasty ; a new surgical procedure €orthe treatment of vaginal agenesis VICHOLAS JOHNSON, RICHARD JAMES LILFORD, 4NDREW BATCHELOR Summary. The ideal operation for a young woman born without a vagina would be a one stage procedure, creating a functionally normal vagina without cosmetically unattractive scars, without the need for subsequent dilatation, stents or obturators. This goal was achieved with a free flap vaginoplasty using a full thickness skin graft taken from the scapula region. The blood supply of the graft was maintained by microvascular anastomosis of the graft pedicles to vessels in the groin. The operation has been performed in three young women who were born with uterine hypoplasia and vaginal agenesis. We experienced no unexpected complications, the procedure was well tolerated and left our patients with a good length, fully functional vagina. However, the operation is a major undertaking and needs to be performed by those with expertise in plastic surgery as well as in gynaecology.
Surgeons have deviscd many ingenious operitive procedures which creatc a vagina for women born without one (Johnson & Lilford 1990). However, they all have disadvantages. The ideal operation would provide a vagina that s of good length, does not need to be dilatated, Rill not cicatrize nor stenose, does not mutilate .he vulva and does not lcave an ugly donor site. To avoid constriction and contraction the surZically created rccto-vesical cavity (potential ieo-vagina) needs to be lined with a full thickJess tissuc flap with an integral blood supply. Bowel flaps have beerr uscJ for the purpose but skin has many advantages. An epithclial skin lining avoids the need for intra-abdominal surgcry snd is not associated with a mucous discharge or Departments of Obstetrics & Gynaecdogy and Plastic Surgery, St James’s University Hospital, Leeds, England N. JOHNSON R.J. LILFORD A. BATCHELOR Correspondence: N . Johnson
184
an unsightly introitus. Myocutaneous pedicle skin flaps from the medial aspect of the thigh have a precarious bloodsupply (Mathes & Nahai 1984) and tlaps taken from the abdominal wall involving rectus abdominus (Lilford et al. 1989) are very bulky and can only be transposed after extensive pelvic dissection behind the bladder. To avoid these complications we decided to use a free skin graft. Wc describe the first cases of vaginal agenesis treated with a free flap graft consisting of full thickness skin obtained from a scapular site. Patients and methods All three patients had congenital uterine hypoplasia and vaginal agenesis; (Rokitansky syndrome; Hauser & Schreiner 1961). They were aged between 17 and 19 and formed part of a group of patients who did not want to practise self dilatation. They were offered a choice bctwcen the Mclndoe procedure (Mclndoe & Bannistcr 1938), tissuc cxpansion vaginoplasty (Lilford et al. 1988) and the free flap operation.
Free-flap vaginoplasty
185
They were aware of the experimental nature of this surgery and that it would be a lengthy and major operation. The frcc flap vaginoplasty is a single operation involving four stages. Stage I Spiral
Under anaesthesia the patient is placed in the lithotomy position, draped and swabbed and an inverted u-shape incision is made over the dimple of the Miillerian pit. A recto-vesical cavity is created by blunt finger dissection making a space that will accommodatc the lcngth of two fingers. The mid-line sagittal fibrous septum may have to be divided with scissors at the top of the neo-vagina. Venous bleeding responds to tamponadc.
seam
Stuge 2
The patient is then positioned supine and the right groin cleansed and draped (for the right handed suryeon). A 5 cm skin incision is made one finger breadth below the inguinal ligament in the line of the skin crease, exposing the femoral vessels. A long and wide calibre vein and an artery with few branches, originating close to the femoral vessels, are identified and mobilized from their distal attachments (usually the superficial femoral circumflex vessels suffice). Once sufficient vessel length has been obtained to avoid tension at the future anastomosis with the
Fig. 2. Making a pouch from the skin flap. The skin is inncrrnost and thc pouch is closed at the medial end to bccomc the vault of thc nco-vagina. Thc latcral aspcct of the flap will become the introitus supplied by the circumflex scapular vessels (CSV).
graft vessels, the distal end is clipped temporarily. The vessels from the graft will be passed under a skin bridge in the groin to meet the mobilized femoral branches. This sub-cutaneous tunnel is created at this stage by blunt disscction starting at the medial edge of the groin incision to emerge at the perineal incision. The wound is then temporarily closed and the patient turned on her left side. Stage 3
Fig. 1. Raising a free skin flap from the back. The
circumflex scapular vascular pedicle is mobilized back as far as the axillary artery.
A free skin flap is raised from the back overlying the scapula using the technique of Mayou et al. (1982) (Figure 1).This vascular pedicle is mobilized to its origin at the axillary artery and the vessels are divided as close as possible to the axilla. Haemostasis is secured and the donor site closed over a drain. The in-vitro skin flap measures 18 cm from left to right and 10 cm from top to bottom corner. It is trimmed of excess fat and fashioned into a pouch that will form the new vagina (Figure 2). The seam is spiralled and the pouch is closed at the opposite end from the origin of the vascular pedicle. The opening of the pouch is fashioned into an oval extending obliquely into a beak at the site of the seam. The
186
N. Johnson et al.
patient is then placed in the Lloyd-Davis position.
cutaneous heparin is uscd to minimize thrombosis but prophylactic antibiotics are not used.
Stage 4
Results
Only the right groin and the vaginal introitus are exposed. The neo-vaginal pouch, skin innermost, is then placed into the rectovesical cavity. The vascular pedicle is directed towards the right, the vaginal pouch is checked to ensure that it is not twisted and then the opening is secured to the introitus. To prevent an annular scar at the site of the anastomosis the apex of the original inverted u incision is extended posteriorly and mediolaterdlly to create a small Y-plasty that accommodates the beak of the pouch. The vascular pedicle is tunneled underneath the subcutaneous bridge in the groin to emerge in the inguinal wound (Figure 3). The patient is then positioned supine and the groin wound is reopened. Using microvascular surgical techniques, the previously mobilised branches of the femoral vessels are anastomosed to those supplying the neo-vagina. The groin incision is then closed.
Three young women with vaginal agenesis have been treated with a free flap vaginoplasty. They knew that the procedure was experimental but wcre keen to avoid the two stage vaginoplasty and did not want to be subjected to self-dilatation that is invariably rcquired with the MeIndoc procedure. Post-operatively, the tissue perfusion of the first patient was monitored with a reflectance pulse oximeter. After 20 min a pulsatile signal was obtained which corresponded to the patient’s heart rate and the oxygen saturation recorded by the oximeter was94%. The signal strength was poor and post-operative monitoring was not continued. Apart from postoperative pyrexia that lasted for 48 h, there were no complications. The first patient has returned to another country and is reported to be enjoying coitus and claims to be asymptomatic. The second and third cases have also had excellent results. The vaginal introitus appears quite normal. Both of the remaining patients have vaginas of at least 4 inches in length (it is possible to reach the vault by digital examination) but the sccond patient has a rather tight stenotic ring at the site of the introital anastomosis. This will admit one, but not two fingers, and may require a Z-plasty before intercourse can begin. The last patient has no such narrowing and claims to be enjoying intercourse.
Post-operative cure
The recto-vesical cavity is drained with closed system suction drainage and the bladder drained suprapubically until the patient can micturate comfortably. Post-operatively they are nursed with their thighs flexed and extension is forbidden. Sub-
Discussion
\ Fig. 3. The neo-vaginal pouch is placed into the rcctovesical cavity and the circumflex scapular vessels (CSV) are passed under the skin bridge and anastomosed to vesscls in the groin.
The large number of surgical procedures (Johnson & Lilford 1990) that can be offered to a woman born with vaginal agenesis is a tribute to the operator’s ingenuity and imagination but it implies that the perfect operation has yet to be described. The classical treatment of repeated self dilatation (Frank 1938) and the bicycle seat modification (Ingram, 1981) are tedious and rarely achieve a full length vagina if the woman has no more than a Mullerian dimple. Williams’ pouch (Williams 1964 & 1970) leaves the young woman with a short and abnormally angled vagina. The alternative to a pseudo-vagina is to create a neo-vagina by surgically opening the recto-vesieal space. This space has to be lined to prcvent adhesion of the vaginal walls. Split skin (Abbe 1899, Mclndoe & Bannister 1938), full
Free-pap vaginoplasty
thickness skin, peritoneum (Tamaya etal. 1984), amnion (Dhall 1984) gut supplied by a vascular pedicle, devitalized gut and even rabbit gut have all been used (Cairns & De Villiers, 1980). Unfortunately none of these linings is ideal. Amnion cannot be used nowadays because of the fear of transmitting AIDS. Our current knowledge of heterografts prevents us repeating the work described by Ciriechieke in 1892 (see Cairns & De Villiers, 1980) with rabbit gut. Split skin grafts create a neo-vagina that will contract unless persistent regular dilatation is carried out. Contraction and stenosis and the need for obturators, moulds and dilators can best be avoided by lining the recto-vesical space with viable tissue carrying its own blood supply. A segment of interrupted bowel with its mesentry can be transposed into the recto-vesical space. This major procedure can leave the patient with an ugly stoma at the introitus that looks like a colostomy and through which mucosa may prolapse. The alternative to pedicle bowel grafts is a full thickness skin flap taken from the perineum (Wee & Joseph 1989, Song etal. 1982), thigh (McGraw et al. 1976), buttock, abdomen (Lilford et al. 1989; Tobin & Day, 1988; McGraw et al. 1988) or skin that has been newly created using tissue expansion techniques (Lilford er al. 1988). Perineal flaps limit the length of the potential vagina and mutilate the vulva, flaps from gracilis, the buttock or the groin have a notoriously precarious blood supply and leave the woman with long thigh, buttock or groin scars. Skin from the abdomen involves major surgery and encroaches upon the integrity of the abdominal wall. Tissue expansion techniques can create sufficient labial skin to line a neo-vagina but this involves a twostage procedure with prolonged hospitalisation. A free graft has the potential advantages of offering a young woman with no vagina a one stage operation to create a functional anatomically normal vagina. It is suitable for young women with vaginal agenesis (with or without uterine hypoplasia) and for women who have lost their vagina following radiotherapy or surgery for vaginal cancer. It lines a recto-vesical cavity with full thickness skin, sparing the patient the tedium of self dilatation and limiting the risk of contraction. The potential donor site is not limited by its distance from the pelvis and can be selected to be cosmetically unobtrusive. In our experience, the free flap vaginoplasty fulfills all of these promises. However, the opcration is a major undertaking. The viability of any
187
free flap can never be guaranteed and the women are left with three wounds instead of one. The procedure involves a four stage operation; the creation of a recto-vaginal space, mobilizing adequate donor vessels, isolating a graft and making it into a vaginal pouch and finally placing the pouch into the recto-vesical space and uniting the vessels supplying the pouch to those in the groin. The potential complications are related to each stage. In general the woman is at risk of thrombosis associated with any major pelvic surgery which takes over 3 h. The first stage may be complicated by venous haemorrhage from the pelvic venous plexus or by perforation of the rectum or bladder. The first will respond to tamponade and the second should only occur as a result of surgical mishap. The second stage involves exposing the femoral vessels and puts the patient at risk of deep venous thrombosis. However, this risk is probably no more than theoretical and will not be greater than the similar risk associated with a long saphenous ligation employed for the treatment of varicose veins. The complication of the third stage depends upon the chosen site of donor skin. Isolating skin with sufficient vessel length from over the scapula involves dissection close to the axillary vessels and nerves with potential of causing damage to either of these structures. The risks and complications of the final stage include leakage and subsequent groin haematoma from the microsurgical anastomosis. Insufficient vessel length may necessitate vein interposition graft, thus increasing the risk of graft failure and subsequent necrosis and infection. Failure of the microsurgical anastomosis would be difficult to detect especially as ischaemia would initially develop at the vault. Although others (Salova & Horton, 1988) have claimed that devitalized full thickness skin can safely be left in the recto-vesical space, the potential risk of infection from a mass of dead tissue between the rectum and bladder must, in our opinion, limit such management. Urinary infection from the catheter, which must be left in situ until the introital oedema has subsided, must be considered to be a minor but common complication of any vaginoplasty. Other potential complications associated with poor technique may include introital stcnosis or a perineal hood. Late complications will be related to changes in the new vagina. Dysplasia and even carcinoma have been reported in skin transplanted into the vagina and there is no
188
N.Johnson et al.
reason to assume that a free flap would not have the potential to be similarly afflicted. As the new vaginal skin has associated adipose tissuc, obesity may be associated with increased fat dcposition around the vagina, limiting the potential space for the bladder, ureters and even the rectum. The increasing bulk of the graft may cause displacement of the ureter with consequent obstruction, particularly as the ureters may be abnormally positioned. Finally, the vagina will be dry and if the donor skin originates from the hair-bcaring area then there may be discharge associated with the hair follicles. Whilst in theory a free flap vaginoplasty has much to recommend it, and in our experience the operation fulfilled our expectations, the procedure is a major surgical undcrtaking lasting over 3 h with significant potential risks. The advantages it offers over other techniques must be offset against these disadvantages. In our opinion it should be limited to women who want a one stage operation to create a fully functional and cosmetically normal vagina without the rcstrictions of self dilatation. Women with a sizeable Mullcrian pit should be offered dilatation with or without Ingram's moulds. Obstruction of the lower vagina (with haematacolpos) should be treated using a pcrineal skin flap (Wcc & Joseph, 1989). Young women with total vaginal agenesis and the mental fortitude to withstand tissue expansion should be offered Lilford's tissue expansion vaginoplasty and those who live in regions of the world without expertise in vaginoplasty or plastic surgery should be offcrcd either a Williams' vulvovaginoplasty or a bowel transplant. References Abbc R. (1898) A new mcthod of creating a vagina in a case of congcnital abscncc. Medical Record of New York 836-838. Cairns T. S. & dc Villicrs W. (1980) Vaginoplasty. S Afr Med J 57, 50-5.5. Dhall K. (1984)Amnion graft for thc trcatmcnt of congenital abscncc of the vagina. Br J Obstct Gynuecol 91,279-282. Frank K. T. (1938) The formation of an artificial vagina without operation. Am J Obstet Gynecol35, 1053-1055. Hauser G. A. & Schrcincr W. E. (1961) Das MaycrRokitansky-Kuster Syndrome Schweiz Med Wochenschr 12,281. lngram J. M. (1981) Thc bicycle seat stool and the
treatment of vaginal agenesis and stenosis: A preliminary report Am J Obstet Gynecol140,867-873. Johnson N. & Lilford R. J. (1990) The surgical treatment of gynaecological congcnital malformations. In Progress in Obstetrics und Gynaecology 8, (Studd J, ed), Churchill Livingstone, Edinburgh pp. 413-431, Lilford R. J., Sharpe D. T. & Thomas D. F. M. (1988) Usc of tissue cxpansion techniques to crcatc skin flaps for vaginoplasty. Case rcport. Br J Obsrer Gynaecol95,402-407. Lilford R. J., Johnson N. & Batchelor A. (1989) A new operation for vaginal agenesis; construction of a nco-vagina from a rectus abdominus musculocutaneous flap. Rr J Obstet Cynnecol%,1089-1094. Mathes S . J. & Nahai F. (1984) Muscle and myocutancous flaps. In The Unfavourable Result in Plastic Surgery: Avoidance and 'lieofrnent. (Goldwyn R. M. cd), 2d ed, Litle Brown, Boston, U.S.A. pp. 1 1 1 - 1 1.5. Mayou B. J., Whitby D. & Jones B. M. (1982) Thc scapular f l a p a n anatomical and clinical study. Br J Plast Surg 35,8-13. McGraw J. B. & Kurtzman L. (1988) Vaginal and pclvic rcconstruction with distally based rcctus abdominis myocutaneous flaps; discussion. Plast Reconst Surg 81,71-73. McGraw J . B., Masscy F. M., Shanklin K. D. & Horton C. E. (1976)Vaginal reconstruction with gracilis myocutancous flaps. PIast Reconsf surg 58, 176-183. Mclndoc A. H. &Bannister J. B. (1938) An operation for the cure of congcnital abscnce of thc vagina. J Obstet Cynaecol Emp 45,490-494. Salova R. C. & Horton C. E. (1988) Utilizing full thickness skin grafts for vaginal reconstruction. Clin Plast Surg 15,443-448. Song R., Wang X.L Zhou G . (1982) Reconstruction of thc vagina with scnsory function. Clin Plast Surg 9 , 105-108. Tamaya T., Yamamoto T., Nakata Y., Ohno Y. & Okada H. (1984) Thc use of pclvic peritoncum in thc construction of a vagina: 10 Cases. Asia Oceunia J Obstet Gynaecol 10, 439-443. Tobin G . R. & Day T. G. (1988) Vaginal and pclvic rcconstruction with distally bascd rectus abdominis myocutancous flaps. Plast Reconst Surg 81, 62-70. Wec J . T. K. & Joscph V. T. (1989)A new technique of vaginal reconstructicw using neurovascular pudcndal-thigh flaps; A prcliminary rcport. Plavf Reconst Surg 83, 7 0 1 - 7 0 . Williams E. A . (1970)Vulvo-vaginoplasty. Proc R Soc Med 63, liW. Williains E. A. (1964) Congcnital absence of the vagina; a simple operation for its relief. J Obstet Gynaecol Br Common 71,511-516.
Received 30 April I990 Accepted 20 July 1990
British Journal of Obstetrics and Gynaecology February 1991, Vol. 98, pp. 184-188
The free-flap vaginoplasty ; a new surgical procedure €orthe treatment of vaginal agenesis VICHOLAS JOHNSON, RICHARD JAMES LILFORD, 4NDREW BATCHELOR Summary. The ideal operation for a young woman born without a vagina would be a one stage procedure, creating a functionally normal vagina without cosmetically unattractive scars, without the need for subsequent dilatation, stents or obturators. This goal was achieved with a free flap vaginoplasty using a full thickness skin graft taken from the scapula region. The blood supply of the graft was maintained by microvascular anastomosis of the graft pedicles to vessels in the groin. The operation has been performed in three young women who were born with uterine hypoplasia and vaginal agenesis. We experienced no unexpected complications, the procedure was well tolerated and left our patients with a good length, fully functional vagina. However, the operation is a major undertaking and needs to be performed by those with expertise in plastic surgery as well as in gynaecology.
Surgeons have deviscd many ingenious operitive procedures which creatc a vagina for women born without one (Johnson & Lilford 1990). However, they all have disadvantages. The ideal operation would provide a vagina that s of good length, does not need to be dilatated, Rill not cicatrize nor stenose, does not mutilate .he vulva and does not lcave an ugly donor site. To avoid constriction and contraction the surZically created rccto-vesical cavity (potential ieo-vagina) needs to be lined with a full thickJess tissuc flap with an integral blood supply. Bowel flaps have beerr uscJ for the purpose but skin has many advantages. An epithclial skin lining avoids the need for intra-abdominal surgcry snd is not associated with a mucous discharge or Departments of Obstetrics & Gynaecdogy and Plastic Surgery, St James’s University Hospital, Leeds, England N. JOHNSON R.J. LILFORD A. BATCHELOR Correspondence: N . Johnson
184
an unsightly introitus. Myocutaneous pedicle skin flaps from the medial aspect of the thigh have a precarious bloodsupply (Mathes & Nahai 1984) and tlaps taken from the abdominal wall involving rectus abdominus (Lilford et al. 1989) are very bulky and can only be transposed after extensive pelvic dissection behind the bladder. To avoid these complications we decided to use a free skin graft. Wc describe the first cases of vaginal agenesis treated with a free flap graft consisting of full thickness skin obtained from a scapular site. Patients and methods All three patients had congenital uterine hypoplasia and vaginal agenesis; (Rokitansky syndrome; Hauser & Schreiner 1961). They were aged between 17 and 19 and formed part of a group of patients who did not want to practise self dilatation. They were offered a choice bctwcen the Mclndoe procedure (Mclndoe & Bannistcr 1938), tissuc cxpansion vaginoplasty (Lilford et al. 1988) and the free flap operation.
Free-flap vaginoplasty
185
They were aware of the experimental nature of this surgery and that it would be a lengthy and major operation. The frcc flap vaginoplasty is a single operation involving four stages. Stage I Spiral
Under anaesthesia the patient is placed in the lithotomy position, draped and swabbed and an inverted u-shape incision is made over the dimple of the Miillerian pit. A recto-vesical cavity is created by blunt finger dissection making a space that will accommodatc the lcngth of two fingers. The mid-line sagittal fibrous septum may have to be divided with scissors at the top of the neo-vagina. Venous bleeding responds to tamponadc.
seam
Stuge 2
The patient is then positioned supine and the right groin cleansed and draped (for the right handed suryeon). A 5 cm skin incision is made one finger breadth below the inguinal ligament in the line of the skin crease, exposing the femoral vessels. A long and wide calibre vein and an artery with few branches, originating close to the femoral vessels, are identified and mobilized from their distal attachments (usually the superficial femoral circumflex vessels suffice). Once sufficient vessel length has been obtained to avoid tension at the future anastomosis with the
Fig. 2. Making a pouch from the skin flap. The skin is inncrrnost and thc pouch is closed at the medial end to bccomc the vault of thc nco-vagina. Thc latcral aspcct of the flap will become the introitus supplied by the circumflex scapular vessels (CSV).
graft vessels, the distal end is clipped temporarily. The vessels from the graft will be passed under a skin bridge in the groin to meet the mobilized femoral branches. This sub-cutaneous tunnel is created at this stage by blunt disscction starting at the medial edge of the groin incision to emerge at the perineal incision. The wound is then temporarily closed and the patient turned on her left side. Stage 3
Fig. 1. Raising a free skin flap from the back. The
circumflex scapular vascular pedicle is mobilized back as far as the axillary artery.
A free skin flap is raised from the back overlying the scapula using the technique of Mayou et al. (1982) (Figure 1).This vascular pedicle is mobilized to its origin at the axillary artery and the vessels are divided as close as possible to the axilla. Haemostasis is secured and the donor site closed over a drain. The in-vitro skin flap measures 18 cm from left to right and 10 cm from top to bottom corner. It is trimmed of excess fat and fashioned into a pouch that will form the new vagina (Figure 2). The seam is spiralled and the pouch is closed at the opposite end from the origin of the vascular pedicle. The opening of the pouch is fashioned into an oval extending obliquely into a beak at the site of the seam. The
186
N. Johnson et al.
patient is then placed in the Lloyd-Davis position.
cutaneous heparin is uscd to minimize thrombosis but prophylactic antibiotics are not used.
Stage 4
Results
Only the right groin and the vaginal introitus are exposed. The neo-vaginal pouch, skin innermost, is then placed into the rectovesical cavity. The vascular pedicle is directed towards the right, the vaginal pouch is checked to ensure that it is not twisted and then the opening is secured to the introitus. To prevent an annular scar at the site of the anastomosis the apex of the original inverted u incision is extended posteriorly and mediolaterdlly to create a small Y-plasty that accommodates the beak of the pouch. The vascular pedicle is tunneled underneath the subcutaneous bridge in the groin to emerge in the inguinal wound (Figure 3). The patient is then positioned supine and the groin wound is reopened. Using microvascular surgical techniques, the previously mobilised branches of the femoral vessels are anastomosed to those supplying the neo-vagina. The groin incision is then closed.
Three young women with vaginal agenesis have been treated with a free flap vaginoplasty. They knew that the procedure was experimental but wcre keen to avoid the two stage vaginoplasty and did not want to be subjected to self-dilatation that is invariably rcquired with the MeIndoc procedure. Post-operatively, the tissue perfusion of the first patient was monitored with a reflectance pulse oximeter. After 20 min a pulsatile signal was obtained which corresponded to the patient’s heart rate and the oxygen saturation recorded by the oximeter was94%. The signal strength was poor and post-operative monitoring was not continued. Apart from postoperative pyrexia that lasted for 48 h, there were no complications. The first patient has returned to another country and is reported to be enjoying coitus and claims to be asymptomatic. The second and third cases have also had excellent results. The vaginal introitus appears quite normal. Both of the remaining patients have vaginas of at least 4 inches in length (it is possible to reach the vault by digital examination) but the sccond patient has a rather tight stenotic ring at the site of the introital anastomosis. This will admit one, but not two fingers, and may require a Z-plasty before intercourse can begin. The last patient has no such narrowing and claims to be enjoying intercourse.
Post-operative cure
The recto-vesical cavity is drained with closed system suction drainage and the bladder drained suprapubically until the patient can micturate comfortably. Post-operatively they are nursed with their thighs flexed and extension is forbidden. Sub-
Discussion
\ Fig. 3. The neo-vaginal pouch is placed into the rcctovesical cavity and the circumflex scapular vessels (CSV) are passed under the skin bridge and anastomosed to vesscls in the groin.
The large number of surgical procedures (Johnson & Lilford 1990) that can be offered to a woman born with vaginal agenesis is a tribute to the operator’s ingenuity and imagination but it implies that the perfect operation has yet to be described. The classical treatment of repeated self dilatation (Frank 1938) and the bicycle seat modification (Ingram, 1981) are tedious and rarely achieve a full length vagina if the woman has no more than a Mullerian dimple. Williams’ pouch (Williams 1964 & 1970) leaves the young woman with a short and abnormally angled vagina. The alternative to a pseudo-vagina is to create a neo-vagina by surgically opening the recto-vesieal space. This space has to be lined to prcvent adhesion of the vaginal walls. Split skin (Abbe 1899, Mclndoe & Bannister 1938), full
Free-pap vaginoplasty
thickness skin, peritoneum (Tamaya etal. 1984), amnion (Dhall 1984) gut supplied by a vascular pedicle, devitalized gut and even rabbit gut have all been used (Cairns & De Villiers, 1980). Unfortunately none of these linings is ideal. Amnion cannot be used nowadays because of the fear of transmitting AIDS. Our current knowledge of heterografts prevents us repeating the work described by Ciriechieke in 1892 (see Cairns & De Villiers, 1980) with rabbit gut. Split skin grafts create a neo-vagina that will contract unless persistent regular dilatation is carried out. Contraction and stenosis and the need for obturators, moulds and dilators can best be avoided by lining the recto-vesical space with viable tissue carrying its own blood supply. A segment of interrupted bowel with its mesentry can be transposed into the recto-vesical space. This major procedure can leave the patient with an ugly stoma at the introitus that looks like a colostomy and through which mucosa may prolapse. The alternative to pedicle bowel grafts is a full thickness skin flap taken from the perineum (Wee & Joseph 1989, Song etal. 1982), thigh (McGraw et al. 1976), buttock, abdomen (Lilford et al. 1989; Tobin & Day, 1988; McGraw et al. 1988) or skin that has been newly created using tissue expansion techniques (Lilford er al. 1988). Perineal flaps limit the length of the potential vagina and mutilate the vulva, flaps from gracilis, the buttock or the groin have a notoriously precarious blood supply and leave the woman with long thigh, buttock or groin scars. Skin from the abdomen involves major surgery and encroaches upon the integrity of the abdominal wall. Tissue expansion techniques can create sufficient labial skin to line a neo-vagina but this involves a twostage procedure with prolonged hospitalisation. A free graft has the potential advantages of offering a young woman with no vagina a one stage operation to create a functional anatomically normal vagina. It is suitable for young women with vaginal agenesis (with or without uterine hypoplasia) and for women who have lost their vagina following radiotherapy or surgery for vaginal cancer. It lines a recto-vesical cavity with full thickness skin, sparing the patient the tedium of self dilatation and limiting the risk of contraction. The potential donor site is not limited by its distance from the pelvis and can be selected to be cosmetically unobtrusive. In our experience, the free flap vaginoplasty fulfills all of these promises. However, the opcration is a major undertaking. The viability of any
187
free flap can never be guaranteed and the women are left with three wounds instead of one. The procedure involves a four stage operation; the creation of a recto-vaginal space, mobilizing adequate donor vessels, isolating a graft and making it into a vaginal pouch and finally placing the pouch into the recto-vesical space and uniting the vessels supplying the pouch to those in the groin. The potential complications are related to each stage. In general the woman is at risk of thrombosis associated with any major pelvic surgery which takes over 3 h. The first stage may be complicated by venous haemorrhage from the pelvic venous plexus or by perforation of the rectum or bladder. The first will respond to tamponade and the second should only occur as a result of surgical mishap. The second stage involves exposing the femoral vessels and puts the patient at risk of deep venous thrombosis. However, this risk is probably no more than theoretical and will not be greater than the similar risk associated with a long saphenous ligation employed for the treatment of varicose veins. The complication of the third stage depends upon the chosen site of donor skin. Isolating skin with sufficient vessel length from over the scapula involves dissection close to the axillary vessels and nerves with potential of causing damage to either of these structures. The risks and complications of the final stage include leakage and subsequent groin haematoma from the microsurgical anastomosis. Insufficient vessel length may necessitate vein interposition graft, thus increasing the risk of graft failure and subsequent necrosis and infection. Failure of the microsurgical anastomosis would be difficult to detect especially as ischaemia would initially develop at the vault. Although others (Salova & Horton, 1988) have claimed that devitalized full thickness skin can safely be left in the recto-vesical space, the potential risk of infection from a mass of dead tissue between the rectum and bladder must, in our opinion, limit such management. Urinary infection from the catheter, which must be left in situ until the introital oedema has subsided, must be considered to be a minor but common complication of any vaginoplasty. Other potential complications associated with poor technique may include introital stcnosis or a perineal hood. Late complications will be related to changes in the new vagina. Dysplasia and even carcinoma have been reported in skin transplanted into the vagina and there is no
188
N.Johnson et al.
reason to assume that a free flap would not have the potential to be similarly afflicted. As the new vaginal skin has associated adipose tissuc, obesity may be associated with increased fat dcposition around the vagina, limiting the potential space for the bladder, ureters and even the rectum. The increasing bulk of the graft may cause displacement of the ureter with consequent obstruction, particularly as the ureters may be abnormally positioned. Finally, the vagina will be dry and if the donor skin originates from the hair-bcaring area then there may be discharge associated with the hair follicles. Whilst in theory a free flap vaginoplasty has much to recommend it, and in our experience the operation fulfilled our expectations, the procedure is a major surgical undcrtaking lasting over 3 h with significant potential risks. The advantages it offers over other techniques must be offset against these disadvantages. In our opinion it should be limited to women who want a one stage operation to create a fully functional and cosmetically normal vagina without the rcstrictions of self dilatation. Women with a sizeable Mullcrian pit should be offered dilatation with or without Ingram's moulds. Obstruction of the lower vagina (with haematacolpos) should be treated using a pcrineal skin flap (Wcc & Joseph, 1989). Young women with total vaginal agenesis and the mental fortitude to withstand tissue expansion should be offered Lilford's tissue expansion vaginoplasty and those who live in regions of the world without expertise in vaginoplasty or plastic surgery should be offcrcd either a Williams' vulvovaginoplasty or a bowel transplant. References Abbc R. (1898) A new mcthod of creating a vagina in a case of congcnital abscncc. Medical Record of New York 836-838. Cairns T. S. & dc Villicrs W. (1980) Vaginoplasty. S Afr Med J 57, 50-5.5. Dhall K. (1984)Amnion graft for thc trcatmcnt of congenital abscncc of the vagina. Br J Obstct Gynuecol 91,279-282. Frank K. T. (1938) The formation of an artificial vagina without operation. Am J Obstet Gynecol35, 1053-1055. Hauser G. A. & Schrcincr W. E. (1961) Das MaycrRokitansky-Kuster Syndrome Schweiz Med Wochenschr 12,281. lngram J. M. (1981) Thc bicycle seat stool and the
treatment of vaginal agenesis and stenosis: A preliminary report Am J Obstet Gynecol140,867-873. Johnson N. & Lilford R. J. (1990) The surgical treatment of gynaecological congcnital malformations. In Progress in Obstetrics und Gynaecology 8, (Studd J, ed), Churchill Livingstone, Edinburgh pp. 413-431, Lilford R. J., Sharpe D. T. & Thomas D. F. M. (1988) Usc of tissue cxpansion techniques to crcatc skin flaps for vaginoplasty. Case rcport. Br J Obsrer Gynaecol95,402-407. Lilford R. J., Johnson N. & Batchelor A. (1989) A new operation for vaginal agenesis; construction of a nco-vagina from a rectus abdominus musculocutaneous flap. Rr J Obstet Cynnecol%,1089-1094. Mathes S . J. & Nahai F. (1984) Muscle and myocutancous flaps. In The Unfavourable Result in Plastic Surgery: Avoidance and 'lieofrnent. (Goldwyn R. M. cd), 2d ed, Litle Brown, Boston, U.S.A. pp. 1 1 1 - 1 1.5. Mayou B. J., Whitby D. & Jones B. M. (1982) Thc scapular f l a p a n anatomical and clinical study. Br J Plast Surg 35,8-13. McGraw J. B. & Kurtzman L. (1988) Vaginal and pclvic rcconstruction with distally based rcctus abdominis myocutaneous flaps; discussion. Plast Reconst Surg 81,71-73. McGraw J . B., Masscy F. M., Shanklin K. D. & Horton C. E. (1976)Vaginal reconstruction with gracilis myocutancous flaps. PIast Reconsf surg 58, 176-183. Mclndoc A. H. &Bannister J. B. (1938) An operation for the cure of congcnital abscnce of thc vagina. J Obstet Cynaecol Emp 45,490-494. Salova R. C. & Horton C. E. (1988) Utilizing full thickness skin grafts for vaginal reconstruction. Clin Plast Surg 15,443-448. Song R., Wang X.L Zhou G . (1982) Reconstruction of thc vagina with scnsory function. Clin Plast Surg 9 , 105-108. Tamaya T., Yamamoto T., Nakata Y., Ohno Y. & Okada H. (1984) Thc use of pclvic peritoncum in thc construction of a vagina: 10 Cases. Asia Oceunia J Obstet Gynaecol 10, 439-443. Tobin G . R. & Day T. G. (1988) Vaginal and pclvic rcconstruction with distally bascd rectus abdominis myocutancous flaps. Plast Reconst Surg 81, 62-70. Wec J . T. K. & Joscph V. T. (1989)A new technique of vaginal reconstructicw using neurovascular pudcndal-thigh flaps; A prcliminary rcport. Plavf Reconst Surg 83, 7 0 1 - 7 0 . Williams E. A . (1970)Vulvo-vaginoplasty. Proc R Soc Med 63, liW. Williains E. A. (1964) Congcnital absence of the vagina; a simple operation for its relief. J Obstet Gynaecol Br Common 71,511-516.
Received 30 April I990 Accepted 20 July 1990
